A DNase I Binding/Immunoprecipitation Assay for Actin*

THEJOURNAL

OF BlOLoGlCAl. CHEMISTRY

Vol. 256. No. 12. I s u e of dune 25, pp. 6291-6295, 1981 Prtnled in C.S A

A DNase I Binding/ImmunoprecipitationAssay for Actin* (Received for publication, October 21, 1980, and in revised form, March 16, 1981)

Michael C. SnabesS, Aubrey E. Boyd, 1118, Robert L. Parduel, and Joseph Bryan11 From the Departments of Medicine and Cell Biology, Baylor College of Medicine, Houston, Texas77030

An actin assay which employs the competition between labeled and unlabeled rabbit skeletal muscle actin for DNase I has been developed. Iodination of actin by the method of Bolton and Hunterresults in the incorporation of approximately 0.5 mol of 125-iodine/ mol of actin, This I2'I-actin retained the ability to bind to DNase I and inhibit enzymatic activity. The Iz5Iactin-DNase complex can be precipitated by the ad& tion of a monospecific rabbit antibody to DNase I. The efficiency of this immunoprecipitationstep i s improved by the use of a second sheep anti-rabbit y-globulin. Using this immunoprecipitation assay, there is a linear displacement of the DNase I-bound '251-actinby rabbit skeletal muscle actin standards or by the actin present in tissue and cell extracts. Using 17.5 ng of DNase I and approximately 500 pg of '2SI-actin,50% inhibition of binding was obtained with 23 ng of unlabeled actin. Reducing the amount of DNase I to 2 ng results in 50% inhibition of binding with 4 ng of unlabeled actin and an increase in the estimated sensitivity of the assay from 1.7 to 0.24 ng. The slopes of the displacement curves generated with both vertebrate and invertebrate non-muscle actins are parallel to rabbit skeletal muscle actin. This observation indicates approximately equal actin-DNase I binding affinities and suggests a high degree of conservation of the actin-DNase I binding site. The assay is useful for measuring the pools of F- and G-actin in a wide range of cells.

munoassay. This approachoffers sensitivity inthe nanogram/ ml range and the possibility of analyzing large numbers of

samples. Actin antibodies are known to exhibit wide crossreactivity when using indirect immunofluorescence as an assay (8). Morgan et al. ( 7 ) note that small differences in primary structure of actins from various species or even from different tissues result in marked differences in actin binding to antiactins. It has therefore been difficult to quantitate absolute amounts of non-muscle actin using radioimmunoassay. T o develop an actin assay, we have utilized the competition of unlabeled rabbit skeletal muscle actin and'"I-actin for DNase I. Bound and free actin are separated by a two-step immunoprecipitation of the actin-DNaseI complex. The assay canbe performed in 2 days, on large numbers of samples, provides picogram sensitivity, and demonstrates parallel displacement using actin from many types of tissues and species. We use the method hereto survey actin in several tissue types. EXPERIMENTAL PROCEDURES

Materials-DNase, DNA, gelatin, sodium azide, and bovine serum albumin fraction V were purchased from Sigma, Bolton and Hunter reagent' was purchased from Amersham. Rabbit IgG fraction I1 was purchased from Miles Laboratory (Elkhard, IN). Freund's adjuvant was purchased from DIFCO Laboratory (Detroit, MI). Actin from sea urchin eggs, isolated from Trzpneustes gratilla, was a gift from Dr. Robert Kane of the Pacific Biomedical Research Center. University of Hawaii. Purification of Actin-Rabbit skeletal muscle actin was purified by the method of Spudich and Watt(9) and was stored as F-actina t 4 "C with 0.058 sodiumazide as a preservative. We have standardized the active actin in these preparations using the DNase I inhibition assay of Blikstad et al. (4) after depolymerization to G-actin. A detailed understanding of the role of actin in cellular Preparation of Cell a n d Tissue Extracts for Actin Measurementprocesses requires a means of quantitating changes in the Cells and tissues for actin measurements were collected on ice and number and stateof assembly of actin molecules. Early assay placed incold lysis buffercontaining 0.1%Triton X-100 in 90 mM KCI, methods took advantageof the relative abundanceof cellular 0.1 mM ethylene glycol bis(P-aminoethyl ether)N,N,N',N'-tetraacetic acid, 1.0 mM NaN.>,10 mM PIPES (pH 6.9). The suspension was then actin and used densitometry of stained electrophoretic gels to vortexed or homogenized. The lysate was centrifuged a t 15,000 X g study changes in actin content (1-3). Blikstad et al. (4) ex- for 5 rnin in an Eppendorf microfuge and the supernatant was colploited the unique property of soluble or G-actin to bind to lected for measurement of soluble actin. For measurementof F-actin, the pellet was resuspended in lysis buffer plus an equalvolume of 1.5 DNase I with high affinity (5, 6) and inhibit the enzymatic hydrolysis of DNA to measure actin. The DNase I inhibition M Gdn-HC1, 1.0 M Na-acetate, 1.0 mM CaCI2, and 1.0 mM ATP (4). assay measures both soluble (G) and filamentous (F) actin. After incubation on ice for 10 min to depolymerize actin filaments, the samplewas centrifuged at 15,000 X g for 5 min and the supernatant This procedure requires lengthy individual spectrophotomet- was used for analysis of actin. ric determinations of DNA hydrolysis but can reliably meaPreparation of Antisera-Electrophoretically pure DNase I sure actin in the microgram/& range. Recently, Morgan et (Sigma grade EP), was solubilized in 0.9% NaCl,emulsified in an al. ( 7 ) have reported attempts to develop an actin radioim- equal volume of Freund's complete adjuvant, and injected intradermally and subcutaneously into 2-month-oldNew Zealand white rab* The costs of publication of this article were defrayed in part by bits. Four hundred pg of DNase I were injected initially. At 6-week the payment of page charges. This article must therefore be hereby intervals, 200 pg were injected subcutaneouslyin incomplete Freund's marked "adoertzsement" in accordance with 18 U.S.C. Section 1734 adjuvant. Anti-rabbit IgG was obtained by injecting a sheep using a similar injection schedulewith 2.0 mg of rabbit IgG as the antigen. solely to indicate this fact. f Supported by NationalInstitutes of HealthTrainingGrant Radioiodination of Actin-Purified rabbit skeletal muscle F-actin ____. T32AM7348. 0 Recipient of National Institutes of Health Grant AM23033 and a ' T h e abbreviations used are:BoltonandHunterreagent, (Ngrant from the Juvenile Diabetes Foundation. succinimidyl-3(4-hydroxy, 5-["~1]iodopheny1)proprionate); PIPES, 1Supported by National Institutes of Health Grant CA22610 to 1,4-piperazinediet.hanesulfonic acid; '"I-actin, '"I-labeled rabbit skelDr. Bill Brinkley. etal muscle actin; DNase I, deoxyribonuclease I; Gdn-HC1, guanidine I( Recipient of National Institutes of Health Grant GM26091. hydrochloride. _ I -

6291

6292 (4-8 mg/ml) was depolymerized by dilution to 2.0 mg/ml with 0.1 M borate saline (pH 8.4) followedby additionof an equal volume of cold depolymerization buffer (4). Approximately 5.0pgof this actin was radioiodinated by the method of Bolton and Hunter (10). "'I-actin was separated from free "'I-Bolton and Hunter reagent by chromatography a t 4 "C on a Sephadex G-75 column (1.5 x 40 cm) equilibrated previously with the assay buffer (5 mM phosphate, 50 mM NaCl (pH 7.5), 0.1% (w/v) gelatin, 0.2 mM ATP, 0.2 mM CaC12,0.05% NaN:I).We usually use only the threepeak tubes and will refer to this material as '"I-actin. Actin Assay-We have used two assays which differ in the amount of DNase I employed, either 17.5or 2.0 ng/assay tube. Actin standards or experimental samplesare aliquoted into disposable glass tubes (12 X 75 mm) and brought to a volume of 500 p1 with the assay buffer. For the less sensitive assay, 17.5 ng of DNase I in 100 pl of assay buffer is added followed by 500 pgof '"I-actin in 100 PI. This mixture is then incubated at 4 "C for 2 h before addition of 100 pl of a 1:10 dilution of anti-DNase I rabbit serum. After 4 h at 4 "C, 100 pi of a 1: 2 dilution of sheep anti-rabbit IgG serum is added. After another 1 h at 4 "C, 3.0 ml of phosphate-buffered saline is added. The tubes are centrifuged at lo00 X g for 30 min, the supernatants are decanted, and the pellets are counted in a gamma counter. The more sensitive assay uses 2 ng of DNase I incubated for 4 h, 100 p1 of 1:200 dilution of the anti-DNase I serum for 12 h, and 100 pl of 1:8 dilution of the second step serum incubated for 8 h. The data from the binding assay have been analyzed as follows. Nonspecific binding is considered to be the radioactivity which is not displaced by a 10,000-fold excess of unlabeled actin (5pg/tube). Specific binding in the absence of competitor, defined as Bo, is equal to the difference between the total bound counts and the nonspecific counts divided by the total L"51-actin counts in the assay tube.Specific binding in the presence of competitor, defined as B, is equal to the difference between the counts bound and the nonspecific counts divided by the total "'1-actin counts originally in the assay tube. The statistical analysis and estimates of the unknown actin concentrations were performed using the computer program of Duddleson et al. (11, 12).This program does a logit transformation, tests for linearity, and analyses for parallelism of standards and unknowns by an analysis of covariance.

FIG. 1. Elution profile of an actin iodination mixture on a Sephadex G-75 column. Actin (6.14 pg) was iodinated by the method of Bolton andHunter (12) and chromatographedon a column previously equilibrated with 5.0 mM sodium phosphate, 50 mM NaC1, 0.3 mM ATP, 0.2 mM CaCh 0.1% gelatin, and 0.05% NaN., at pH 7.5. Thirty-drop fractions were collected, actin was determined by the DNase I inhibition assay, and 125-Iodine was assayed in a gamma counter. The specific activity of the fractions in t h e f i s t peak was calculated andis expressed as the ratio of counts per min/ng of actin. 0,125-Iodine counts; bars, amounts ofactin. The dataabove the first peak are thespecific activities of each fraction.

RESULTS

Iodination of Actin-A Sephadex G-75 elution profile of rabbit skeletal muscle actin radioiodinated by the method of Bolton and Hunter (10) is shown in Fig. 1. In five separate iodinations, an average of 20.0 f 4.5% of the '251-Boltonand Hunter reagent was transferred to 5.3 rt 0.3 pg (rtS.E., N = 5) of actin. The mean specific activity was 1.0 f 0.3 x 10:' Ci/ mmol or 0.49 k 0.15 molecules (+S.E., N = 6) of iodine/ molecule of actin. The shoulder on the leading edge of the actin peak is probably aggregated actin or residual F-actin. This material inhibits DNaseI poorly and gives a 6-fold higher nonspecific binding than the actin in the major peak. The actin recovered in the peak is approximately 130% of the initial sample as estimated using the DNase 1 inhibition assay. The average specific activity of this material (Fig. 1) was 35,500 cpm/ng of actin. There was no detectable DNase I inhibitory activity in the second iodine peak. Preliminary Characterizationsof the DNaseI AntibodyThe immunodiffusion pattern of the anti-DNase I antibody showed a single precipitation band. The DNase I antibody has also been used in the indirect fluorescence staining procedure described by Wang and Goldberg (13) and localizes added DNase I on stress fibers and actin cables in cells grown in monolayer culture. Characterization of the Immunoprecipitation ReactionsIn order to optimize the conditions for the actin assay, we characterized the immunoprecipitationreactions. We routinely use 300-500 pg of Iz5I-actin or approximately 10,000 cpm/assay tube. Fig. 2 illustrates the results of increasing the concentration of ant,i-DNase I with a constant concentration of ""I-actin: the dashedlines show the extent of specific actin binding (Bo)a t various actin/DNase I ratios in the absence of

gt

Anti-UNsrs I

FIG. 2. The titration of the 'ZSI-actinDNase I complex with rabbit anti-DNase I serum. Five hundred pgof Iz5I-actin were incubated with increasing concentrations of DNase I in a totalvolume of 600 pl of assay buffer for 24 h. Varying amounts of anti-DNase I were added andthe incubationcontinued for 24h. Second step antibody, anti-rabbit y-globulin, was added to the indicated samples and all the samples were incubated for an additional 12 h before processing as described under "Experimental Procedures." The data are expressed as the percent specific binding (BOX 100) uersus microliters of added anti-DNase I serum. --, with second step antibody; - - -, without this antibody. 0,0.5 ng of DNase I; 0,2.0 ng of DNase I; 0, 5.0 ngof DNase I; 10.0 ng of DNase 1.

.,

the second step antibody; the solid lines are thespecific actin binding after the addition of a second immunoprecipitation step using anti-rabbit IgG. In the absence of the second step antibody, approximately20%of the '"1-actin DNase I complex is precipitated at the highest anti-DNase concentration used. Inclusion of second step antibody precipitates about 60% of

6293

Actin Assay the specific radioactivity at this DNase I concentration and increases the efficiency of precipitation a t all the concentrations used. By increasing the DNase I to approximately 100 ng, using 100 pl of the anti-DNase I and an excess of the second antibody, it is possible to precipitate 90% of the total 125 I-actin, indicating that most of the traceris in a nativeform which will bind with DNase I. The efficiency of precipitation of the specifically bound counts by the second step antiserum was further characterized using increasing amounts of second step antibody and a constant amount of anti-DNase I (Fig. 3). The effect of added carrier rabbitserum is also illustrated. In the absence of normal rabbit serum thereis an optimum between 10 and 20 pl of the second step antiserum. With addition of carrier serum there is a broad optimum between 50 and 100 plof added second step antiserum. Similar results were obtained with twice the concentration of the anti-DNaseI serum (data not shown). The resultsindicate there is no significant advantage in using carrier serum. We have estimated the time required for completion of the various binding reactions by mixing the reagents for a given period of time and terminating the reaction by addition of the next reagent. The results are shown in Fig. 4. The time course of ‘251-actinbinding to DNase I is known to be rapid (6) and, using this crude approach,the reaction is 75% complete at our earliest time point for the 17.5-ng assay. There is then a slow increase in the extent of binding over the next 90 min with completion of the reaction within 2 h. The reaction of antiDNase I with the complex is slower but is complete within 2 h. With the lower protein concentrations used for the 2.0-ng assay, both the reactions are slower, but the reactions appear to be complete within 12 h at 4 “C. Displacement of ”‘I1-actinfrom DNase I by Different Actins-Fig. 5 illustrates the competition or displacement curves using ’251-actinand the actin in various cell extracts as the competitor. The data are given as a percentage, bound actin in the presence of competitor ( B ) over the total bound actin in the absence of competitor (Bo),plotted against the dilution of each extract. Thedisplacement curves are parallel, indicating that the affinities of actin from rat pituitary cells, human platelets, hamster pancreatic islets, sea urchin eggs, and rat aorta are essentially equal. Replotting the data in log-logit form gives a seriesof lines, which are linear and parallel when analyzed using the statistical tests in the Duddleson et al. (11, 12) program.

-o D

n

.-0

-

L

( 6 antl-DNae

IC

ARGG

I

2ol 1 , , , 1 , , , , 6 , , , 1 2 4 6 3 6 9 1 2 3 6 9 1 2 Hrr. Prmr to Next Step

FIG. 4. Time course of binding of the various reactants. The reaction sequence for the assay is the formation of an actin-DNase I complex followed by formation of a complex with anti-DNase I. Finally, anti-rabbit y-globulin binds to theanti-DNase. Panel a, time course of the actin-DNase I reaction; Panel 6 , the reactions with antiDNase I; and Panel e, the reaction with the sheep anti-rabbit yglobulin (ARGG). Each panel gives the results for the 2-ng assay (0) and the 17.5-ng assay (A).Each reaction used 500 pg of 12’II-actin. The 2-ng experiments used 100 pl of anti-DNase I at a 1:200 dilution and 100 p1 of the sheep serum at a 1:8 dilution. The 17.5-ng experiments used 100 pl of a 1:lO dilution of anti-DNase serum and 100 p1 of a 1:2 dilution of the sheep serum. Each reactionwas terminated at the times indicated by addition of the next reactant then carried through the complete two-step immunoprecipitation assay. A value for specific binding (BO)was determined at each point and the data are plotted as percentages of the maximal value for each experiment.

1.56 3.9

3.1 7.8

6.25 12.5 15.6 31.25

25 60.5

50 125

100 250

200 500

ng/tube rl/tube

Muscle actin was prepared as described under “Experimental Procedures” andwas “standardized” using the DNase I inhibition assay (4). The muscle actin data (0)are plotted as nanograms of actin/tube. The other samples are plotted as microliters of extract/tube. B and Bo are defined under“ExperimentalProcedures.”Initial dilutions were done on each extract in order to have the data on comparable scales. 0, rat GHJ cells; W, human platelets; 0, hamster pancreatic islets; A,sea urchin eggs; A,rat aorta.

30

z

=

f

DNw

FIG. 5. Displacement of I2’I-actin from DNase I by rabbit skeletal muscle actin and by actin from cell and tissue extracts.

40

m .E

A

20

(0

10 5

50

100

150

200

diSheep-ARGG

FIG. 3. The titration of the actin-DNase complex with sheep anti-rabbit y-globulin. Iz5I-actinwas incubated with 2 ng of DNase

I for 12 h to form a complex. Normal rabbit serum (5 or 10 pl) and 0.5 p1of anti-DNase I were added and incubated for an additional 12 h. Varying amounts of sheep anti-rabbit y-globulin were added to the assay tubes and incubated for a final 24 h. The per cent specific binding is equivalent to the BO value defined under “Experimental Procedures.” 0, 0.5 pl of anti-DNase I without normal rabbit serum; 0, 0.5 pl of anti-DNase I plus 5 pl of normal rabbit serum; A, 0.5 pl of anti-DNase I plus 10 pl of normal rabbit serum.

Fig. 6 compares different actins and also ‘illustrates the agreementbetween the DNase I inhibition assay and the binding/immunoprecipitation assay. The actin concentrations for purified sea urchin egg actin, purified rabbit skeletal muscle actin, and human platelet extracts were determined using the DNase I inhibition assay and then analyzed using the binding/immunoprecipitation assay. The data for both the high and low sensitivity assays are given, plotted asthe amount of actin against the percent B/BO.The results confvm the parallel displacement by the different actins. The data also provide a measure of the interassay variability. In principle, if there were no variability, the data points from all

Actin Assay

6294

three samples would be superimposable. We argue that the sedimentation of the tissue homogenate a t 15,000 X g for 5 displacement curves in Fig. 6 are not significantly different for min. The soluble fractionis measured as G-actin; the pellet as the following reasons. For the worst case, the 2-ng assay, the filamentous or cytoskeletal actin. The results indicate that slopes are 2.72,2.63, and 2.69 formuscle, seaurchin,and actin is about 1.2% of the total liver protein and that, under platelet actins,respectively. These values are not significantly these solution conditions, which maintain muscle actin in the different when analyzedby theDuddleson et al. (11, 12) F-form, most of the actin in theliver, 82%, is in the G-form. program. The averagecoefficient of variation at the 50%point A comparison of diaphragm, heart, and aortic smooth musfor five separate standard displacement curves 9.4%. was Using cle from the rat shows strikingdifferences in the actin distrithis average value and the 4.3-ng mean value a t 50% for this bution. Actin is approximately 10% of the total proteinof the particular assay,we calculate the standard deviation to 0.4 be diaphragm muscle, less than1% of which is nonsedimentable, ng. The 95% confidence limits are, therefore, 4.3 f 0.8 ng. or G-actin, under these conditions. Similar results, with about Both the sea urchin and platelet50% mean values fall within 7% nonsedimentable actin, were obtained with the gastrocnethese limits. This is not a completely rigorous comparison mius muscle. since we have not included the possible variability of the Actin represents 643% of the total proteinin aortic smooth initial actin determinationsusing the DNase inhibition assay. muscle and about 3% of the protein in cardiac muscle. In the Finally, Fig. 6 illustrates the difference in sensitivity be- heart, approximately 90% of the actin is filamentous by our tween the two assay conditions. The 17.5-ng assay has 50% definition, whereas in the aorta, less than 30% of the actin is displacement at 23.7 f 3.4 ng of actin ( G . E . , N = 8); the sedimentable. It is notclear if the extent of assemblyin assaywithgreater sensitivity,2ng of DNase I, has 50% smooth muscle is a function of the solution conditions used displacement a t 2.9 f 0.3 ng (fS.E., N = 5) of actin. The for assay, for example, the calcium ion concentration as has average assay sensitivity for the 17.5-ng assay defined from been shown for platelets (14) or is a function of the state of the assay median variance ratio(13)is 4.9 f 1.5 ng (-+S.E.,N contraction. = 5) while the limit of detection, defined as 1 S.E. from the The results for one tissue culture cell, Chinese hamster buffer control tube, is1.73 0.33 ng (+S.E., N = 5). The 2-ng ovary cells, are in agreement with the dataof Blikstad et a2. DNase I assay has a sensitivity of0.24 ng and a limit of (4) with about 35-40% of the actin in the filamentous form. detection of 61 pg. The actual values for total actidcell are 5-6 pg/cell, about Distribution of Actin in Several Tissues-In a preliminary 4% of the totalcelI protein. These values are for a population survey we have used the actin assay to quantitate actin in of confluent cells representing all stages of the cell cycle. We liver, heart, skeletal muscle, vascular smooth muscle (aorta), stress that the results given in Table I represent essentially and Chinese hamster ovary cells. The results are given in all the actin in the preparations but that the G- and F-values Table 1. The G- and F-fractions are arbitrarily defined by may be dependent upon thelow speed centrifugation criteria which we used to "define" G- and F-actin. The principal problem is that the low speed supernatants may contain some unsedimented F-actin. We have tested this explicitly for the Chinese hamster ovary cell preparations. The average actin concentration in two independent samples of the low speed supernatant was 6.5 pg/ml. Recentrifugation of this supernatant at 100,000 X g for 120 min produced ahigh speed supernatant with an actin value of 6.4 pg/ml. We have reported a similar result using sea urchin coelomocytes and the DNase I inhibition (18).

*

DISCUSSION

L

I

1

I

1

1

3

10

30

100

The results presented in this paper demonstrate that a binding/immunoprecipitation assay for actin has been estabActin ng/Tube FIG. 6. Comparison of the displacement of '*'I-actin by pu- lished which utilizes the high affinity actin-DNase I interacrified muscle and sea urchin actins and by human platelet tion. Thisassay providesa simpleandrapidmethodfor actin. The amount of actin in each sample was estimated using the quantitation of filamentous and soluble actin. The sensitivity DNase I inhibition assay (4) and standard displacement curves were for measurement of actin is increased about 1000-fold from generated for each sample. -, 17.5-ng assay; - - -, 2-ng assay. The the 10 pg/ml reported for the DNaseI inhibition method (4). data are expressed as per cent BIBo. 0, muscle actin; A, human Using the binding/immunoprecipitation assay an individual platelet actin; U, sea urchin actin. can measure100 actin samplesin duplicate at several dilutions with a coefficient of variation of less t h m 10%in 2 days. Since TABLE I the standard curve is linear after logit transformation, the Actin content in rat tissues assay is amenable to the statistics used for radioimmunoasProh Actin/ G F G/F Actin" Tissue tein protein says. The Duddleson et al. (11, 12) program uses the F-test 97 % B (15) to analyze for linearity and parallelism. Using this test, mg Pg 1.2 4.5 all the actins assayed, which include a-, ,B-,and y-actins, are 23.4 82 18 Liver 280 10.0 0.7 99.3 0.007 1310 13.1 Diaphragm linear and the displacement curves parallel. are This indicates 15.6 6.6 93.4 0.07 Gastrocnemius 170 1.1 a high degree of conservation of the DNase I binding site on 0.12 11.0 89 2.9 Heart 280 9.8 actin. This finding is in agreement with theknown similarity 4.9 17 83 8.3 19 0.23 Aorta of actin amino acid sequences (24). 1.7 37 4.3 1.05 Chinese hamster 63 45 The statistical treatment also allows an estimation of the ovarv cells sensitivity and the limits of detection of the assay. These '' Actin was measured by binding/immunoprecipitation assay. Protein values are the sum of the protein measurements on the values and their estimations arediscussed in detail by others values soluble fraction (G-actin) and the nonsedimentable fraction used for (16, 17). In the resultswe stated the most conservative F-actin determination. given by the Duddleson et al. (11, 12) program. ~

Actin Assay In order to use the binding/immunoprecipitation assay to measure both G- and F-actin, it has been necessary to separate these forms before assay and todepolymerize the filamentous actin. The strategyemployed is to rapidly and gently lyse the cells with a nonionic detergent and then separatea cytoskeleta1 fraction by centrifugation at 15,000 X g for 5 min. The supernatant is either diluted directly for assay or is diluted after incubation with the Gdn-HC1 buffer; the F-actin in the cytoskeletal fraction is depolymerized in Gdn-HC1 before dilution for assay. This strategyis not unique but does attempt to minimize the potential effects of the redistribution of actin during the initial fractionation. Duringthe assay, the concentrations of actin are routinely in the nanogram/ml range, at least 100 times below the critical concentrations for assembly; therefore, the actinremains in a nonfilamentous form available to DNaseI. We have tested the low speed extracts for the presence of additional F-actin by recentrifugation at 100,000 X g for 120 min. In two cases, Chinese hamster ovary cells and sea urchin coelomocytes (18), we find no difference and suggest that, under the lysis conditions employed, the bulk of the filamentous actin inthese cells is present in a “cytoskeleton” which is sedimentable at relatively low g forces. It should be noted that the criterion for discriminating G- and F-actin is substantially different from that used by Blikstad et al. (4). The inhibition assay relies upon the substantially higher rate of enzyme binding to G-actin to measure soluble actin in the presence of F-actin. The values obtained with the actin binding assay agree reasonably well with published data obtained using very different methods. Gordon et al. (19) estimate actin to be 1.5% of total liver protein after purifying actin and correcting for recoveries. This is in excellent agreement with the 1.2% determined here. We estimate that 10% of the total protein from diaphragm and 15% for gastrocnemius is actin which agrees with the general figure quoted for skeletal muscle of about 20%. This number is based on literature values which have been calculated from older actin/myosin ratio data (20-23). Murphy et al. (20) estimated that actin was 25% of the total protein or porcine carotid, again using actin/myosinratio data. This seems like an unreasonably high value and, using aorta asa sourceof smooth muscle, we determined actinto be about 7% of the noncollagenous protein. The binding/immunoprecipitation assay consistently gives somewhat lower values than theactin/myosin ratio methodwhich relies on somewhat indirect estimates of the myosin content of a tissue and sodium dodecyl sulfate gels to get an actin/myosin ratio. The exact reasons for the differences are, however, not clear. In summary, thisassay provides the convenience and statistical advantages of a radioimmunoassay. Moreover, it is unnecessary to purify individual actins from the tissues or species of interest since all the a-,/3-, and y-actins tested to

6295

date displace in parallel. Finally, DNase I is commercially available in an electrophoretically pure form and is highly antigenic. The use of DNase I and anti-DNase I obviate the need for characterization of individual specific actin antibodies. Acknowledgments-We want to thank Linda Miner and Dotty Florip for preparation of the manuscript. REFERENCES 1. Tregear, R. T., and Squire, J . M. (1973) J.Mol. Biol. 77,279-290 2. Bray, D., and Thomas, C. (1975) Biochem. J. 147, 221-228 3. Bullard, E., and Reedy, M. K. (1973) Cold Spring Harbor Symp. Quant. Biol. 37,423-429 4. Blikstad, I., Markey, F., Carlsson, L. Persson, T., and Lindberg, U. (1978) Cell 15,935-943 5. Berger, G., and May, P. (1967) Biochem. Biophys. Acta 139, 148161 6. Mannherz, H. G., Goody, R. S., Konrad, M., and Nowak, E. (1980) Eur. J.Biochem. 104, 367-379 7. Morgan, J. L., Holladay, C. R., and Spooner, B. S. (1980) Proc. Natl. Acad. Sei.U. S. A. 77, 2069-2073 8. Lazarides, E., and Weber, K. (1974)Proc. Natl. Acad. Sci. U. S. A . 71,2268-2272 9. Spudich, J. A., and Watt, S. (1971) J.Biol. Chem. 246,4866-4871 10. Bolton, A. E., and Hunter, W . M. (1973) Biochem. J. 133, 529538 11. Midgley, A. R., Jr., Niswender, G. D., and Rebar, R. W. (1969) Acta. Endocrinol. 142, (suppl.) 247-269 12. Duddleson, W. G., Midgley, A. R., Jr., and Niswender, G. D. (1972) Comput. Biomed. Res. 5,205-217 13. Wang, E., and Goldberg, A. R. (1978) J. Histochem. Cytochem. 26, 745-749 14. Carlsson, L., Markey, F., Blikstad, I., Persson, T., and Lindberg, U. (1979) Proc. Natl. Acad. Sci.U. S. A . 76, 6376-6380 15. Finney, D. J. (1964) Sfatistical Methods in Biological Assay, Hafner, New York 16. Ekins, R. P. (1978) in Radioimmunoassay and Related Procedures in Medicine (Hall, P. E., and Ekins, R. P., eds) Vol. 2, pp. 6-20, International Atomic Energy Agency, Vienna 17. Rodbard, D., Munson, P. J., and DeLeon, A. (1978) in Radioimmunoassay and Related Procedures in Medicine (Hall, P. E., and Ekins, R.P., eds) Vol. 1, pp. 469-504, International Atomic Energy Agency, Vienna 18. Otto, J., and Bryan, J . (1981) Cell Motil. 1, 179-192 19. Gordon, D. J., Boyer, J. L., and Korn, E. D. (1977)J.Biol. Chem. 252,8300-8309 20. Murphy, R. A., Herlihy, J. T., and Megerman, J. (1974) J. Gen. Physiol. 64, 691-705 21. Ebashi, S., and Nonomura, Y. (1973) in The Structure and Function ofMuscle (Bourne, G. H., ed) pp. 260-362, Academic Press, New York 22. Pollard, T. D. (1975)in Molecules and Cell Movement (Inoue, S., and Stephens, R. E., eds) pp. 259-286, Raven Press, New York 23. Stossel, T. P. (1978) Annu. Reu. Med. 29,427-457 24. Vandekerckhove, J., and Weber, K. (1978) Eur. J. Biochem. 90, 451-462