Preparation of antiserum against a tryptic fragment (fragment A) of dynein and an immunological approach to the subunit composition of dynein.

An improved method for purifying the tryptic fragment (Fragment A) of flagellar ATPase (dynein) from sea urchin spermatozoa is described. The preparation appears homogeneous as judged by ultracentrifugation, electrophoresis on polyacrylamide gels, and immunological techniques. The molecular weight of undenatured Fragment A was determined to be 400,000 and 370,000 by the two methods of disc electrophoresis on polyacrylamide gel and sedimentation equilibrium, respectively. The fragment dissociated into two principal polypeptide chains with molecular weights of 190,000 and 135,000 when heated in the presence of sodium dodecyl sulfate. Antiserum against dynein was prepared in rabbits using purified Fragment A from the sea urchin Anthocidaris crassispina as an antigen. The specificity of this serum toward Fragment A and toward dynein was determined by double diffusion in agarose, by inhibition of ATPase activity, and by sodium dodecyl sulfate-electrophoresis of the antigen-antibody complex. This antiserum also reacted with the enzymes from two other species of sea urchin, Pseudocentrotus depressus and Hemicentrotus pulcherrimus. Analysis of the precipitated antigen-antibody complex showed that the antiserum reacted specifically with the "high molecular weight" polypeptide seen in sodium dodecyl sulfate-polyacrylamide gel electrophoresis of crude dynein fractions. This finding supports previous reports that this band derives from dynein ATPase. In our preparations, this "high molecular weight" dynein band appeared single.


Preparation of Antiserum against a Tryptic Fragment (Fragment A) of Dynein and an Immunological
Approach to the Subunit Composition of Dynein* (Received for publication, January 13,197jj KAZUO OGAWA~ AND HIDEO MOHRI § From the Department of Biology, Tokyo Metropolitan University, Setagaya-ku, Tokyo, and the Biological Institute, University of Tokyo, Meguro-ku, Tokyo, Japan An improved method for purifying the tryptic fragment (Fragment A) of flagellar ATPase (dynein) from sea urchin spermatozoa is described. The preparation appears homogeneous as judged by ultracentrifugation, electrophoresis on polyacrylamide gels, and immunological techniques. The molecular weight of undenatured Fragment A was determined to be 400,000 and 370,000 by the two methods of disc electrophoresis on polyacrylamide gel and sedimentation equilibrium, respectively. The fragment dissociated into two principal polypeptide chains with molecular weights of 190,000 and 135,000 when heated in the presence of sodium dodecyl sulfate.
Antiserum against dynein was prepared in rabbits using purified Fragment A from the sea urchin Anthocidaris crassispina as an antigen. The specificity of this serum toward Fragment A and toward dynein was determined by double diffusion in agarose, by inhibition of ATPase activity, and by sodium dodecyl sulfate-electrophoresis of the antigen-antibody complex. This antiserum also reacted with the enzymes from two other species of sea urchin, Pseudocentrotus depressus and Hemicentrotus pulcherrimus.
Analysis of the precipitated antigen-antibody complex showed that the antiserum reacted specifically with the "high molecular weight" polypeptide seen in sodium dodecyl sulfate-polyacrylamide gel electrophoresis of crude dynein fractions. This finding supports previous reports that this band derives from dynein ATPase. In our preparations, this "high molecular weight" dynein band appeared single.
The flagellum of sea urchin sperm consists of a membraneenclosed axoneme composed of a cylinder of nine doublet tubules surrounding a central pair of single tubules. Each doublet tubule bears two parallel rows of projections, called arms, which project asymmetrically from the A-tubule of each doublet toward the B-tubule of the adjacent doublet. According to the observations of Summers and Gibbons (l), it may be expected that the arms contain the flagellar ATPase protein and that their enzymatic site interacts with the B-tubule of the neighboring doublets. By chemical dissection of the flagellum, Gibbons' group (2)(3)(4) showed that the arms contain the flagellar ATPase protein named dynein. Dynein solubilized from the flagella by dialysis at low ionic strength in the presence of EDTA can recombine with the extracted flagella if added back in the presence of divalent cations (5,6). Tryptic §To whom reprint requests should be addressed at the Biological Institute, University of Tokyo, Meguro-ku, Tokyo, Japan. digestion of dynein produces a smaller well defined protein fragment with ATPase activity. This tryptic fragment, named Fragment A, lacks the ability to rebind to extracted flagella (7). The evidence that the ATPase activity of Fragment A is activated by the protein(s) in the B-tubule fraction more than by the protein(s) in the A-tubule fraction suggests that ATPase sites of the arms face the B-tubule of the neighboring doublet (7), but it has not yet been determined exactly where the site of ATPase activity is located on the arms.
To confirm the location of dynein ATPase in flagella, and to obtain information about the location of the ATPase site on the arms, we have developed a direct enzyme-antibody technique using an antiserum against the enzymatic fragment of the dynein molecule. This ATPase-containing fragment (Fragment A) of dynein was used as the antigen because we hoped to use the antiserum as a probe to study the localization and function of the enzymic portion of the molecule. In this paper, we report the specificity of this antiserum for dynein ATPase as determined by double diffusion in agarose (Ouchterlony's test), by inhibition of ATPase activity, and by Na dodecyl-SO, polyac. rylamide gel electrophoresis of the antigen-antibody complex. In order to obtain a specific antiserum, it was first necessary to develop a new procedure for purifying Fragment   Method of Yphantis-Equilibrium centrifugation was carried out in a centrifuge installed with interference optics using the method of Yphantis (13). The result of centrifuging a typical preparation of Fragment A is presented in Fig. 4, and clearly shows the preparation to be homogeneous in molecular weight. The line in the figure represents the theoretical line for a molecular weight of 370,000. This value of the molecular weight of Fragment A agrees fairly well with the value of 400,000 obtained by the method of Hedrick and Smith.

Na Dodecyl-SO, Electrophoretic
Pattern of Fragment A When Fragment A was treated with Na dodecyl-SO, in a boiling water bath as described under "Methods," the preparation gave two main protein bands on electrophoresis in polyacrylamide gels in the presence of Na dodecyl-SO, (Fig.  5A). The slower and faster bands were designated as the F, and F, peptide chains, respectively. Preliminary data from qualitative scanning of gels with a Toyo scanning densitometer showed that the F, band is more intense by 20 to 40% than the F, band. Fig. 5B shows the Na dodecyl-SO, electrophoretic profile of a mixture of Fragment A, and marker proteins including myosin, phosphorylase b, serum albumin, and lactate dehydrogenase.
The two peptide chains, F, and F,, were located between myosin and phosphorylase b. The molecular weights of F, and F, peptide chains were calculated to be 190,000 and 135,000, respectively, from the relationship between the molecular weights and the relative mobilities of the markers, together with those of Fragment A, on Na dodecyl-SO,-polyacrylamide gels of three different concentrations (3.5, 4, and 5%).
In order to examine the possibility that the F, peptide might be derived from the F, peptide as a result of more extensive trypsin digestion, the purified Fragment A preparation  (13). The logarithms of the fringe displacements are plotted against yz, where y is the distance in centimeters to the axis of rotation. Fragment A, 0.99 mg of protein/ml in Tris-EDTA solution containing 50 mM KCl, was run with a rotor speed of 9.341 rpm; the temperature was kept at 18". The line in the figure represents the theoretical line for a molecular weight of 370,000. treatment as described under "Methods," an aliquot was subjected to Na dodecyl-SO, electrophoresis. The reaction mixture containing trypsin pretreated with an equal amount of the inhibitor served as control. These experiments showed no indication that the I?, peptide was transformed to F, by more extensive digestion. Instead, the two minor bands below the F, and F, bands in Fig. 5A became relatively more intense after the extended digestion, although the F, and F, bands were still present. It is probable that the minor bands are derived from F, and F, components by further tryptic digestion. These results suggest that Fragment A consists mainly of two polypeptide components, F, and F,. When Fragment A was subjected to electrophoresis on 10% gel, a faint band, designated as peptide X, with the apparent molecular weight of 42,000, was often observed. Since the intensity of the X band varied greatly between different preparations of Fragment A, it may have been an impurity.   The left portion of Fig. 6A shows the precipitin profile of Fragment A with serial P-fold dilutions of antiserum (No. 8). When the antiserum was diluted 16-fold, the precipitin line could no longer be observed directly. However, Amido black staining of the agarose increased the sensitivity and made it possible to detect the precipitin line even using 64-fold diluted antiserum. The right portion of Fig. 6A represents the pattern of antiserum with serial 2-fold dilutions of Fragment A. In this experiment, the precipitin line could be observed so long as the amount of antigen was greater than 2.5 pg. Our preparation of antiserum (No. 64) having the highest titer of antibody was obtained by repeated booster injections, and it formed a visible precipitin line even with 64-fold diluted antiserum (Fig. 6B) and of Fragment A, it was concluded that it contains specific antibodies against these enzymes. Curves A, B, and C in Fig. 8 show the time dependence of ATPase inhibition by three different amounts of antiserum. When the amount of antiserum added was large, the ATPase activity was rapidly and almost completely inhibited (Curve C), but smaller amounts of antiserum required a longer incubation period to achieve their maximum inhibitory effect (Curve A). Therefore, overnight incubation at O-4' was chosen as a standard incubation time in determining the inhibitory activity of the antisera. Fig. 9 shows the residual ATPase activity of Fragment A from A. crassispina and P. depressus after addition of serial P-fold dilutions of antiserum. The residual activity was determined as described under "Methods." For purposes of defining the inhibitory activity of the antiserum toward various enzymes, the titer of antibody was defined as (number of units of ATPase activity in the uninhibited mixture per volume of antiserum (in microliters) required to give 50% inhibition of this ATPase activity). unit).
The residual activities of dynein from A. crassispina, P. depressus, and H. pulcherrimus after addition of serial 2-fold dilutions of antiserum are shown in Fig. 10. In the experiment with H. pulcherrimus, we used a low ionic strength dynein extract from fresh sperm. The antiserum was No. 8 in all cases. The titer of antibody was 0.014, 0.019, and 0.035 for the dynein from H. pulcherrimus, P. depressus, and A. crassispina, respectively.
These results show that the titer of antibody in antiserum prepared against Fragment A from A. crassispina is significantly lower for dynein from the other species of sea urchin than for that from A. crassispina.
Proportion of ATPase Activity Resulting from Serum-sensitive ATPase in Glycerinated Flagella-When antiserum was added to dynein (Fig. 10) or Fragment A (Fig. 9), the ATPase activity of the mixture was not completely inhibited. Two to 5% of the original activity remained in the supernatant, although the precipitate of antigen-antibody complex showed no ATPase activity. With a low ionic strength dynein extract from sperm, 20% of the ATPase activity remained after addition of a saturating amount of antiserum (Fig. 10). Since this antiserum was prepared against Fragment A, only the form of dynein which, when digested with trypsin, gave rise to Fragment A might be expected to be inhibited by the antiserum. To examine this point further, we estimated that proportion of the total flagellar ATPase activity results from the dynein. Flagella were prepared from glycerinated sperm of A. crassispina, stored at -15" for 1 year, by the method of Ogawa and Mohri (5). When the flagella with a total ATPase activity of 10.6 units were dialyzed against Tris-EDTA solution, their total activity was reduced to 6.4 units. The activities of the supernatant and the precipitate obtained after centrifugation were 5.0 and 1.6 units, respectively.
When these fractions were mixed with antiserum for 1 day, the inhibition was 68 and 51% in the supernatant and precipitate, respectively. Accordingly, the proportion of the total ATPase activity in glycerinated flagella resulting from dynein ATPase was at least 66%. After 4 days preincubation with antiserum, the amount of inhibition was as high as 82%. Some of the ATPase activity that was not inhibitable by antiserum may have been due to enzymes associated with the mitochondria or cell membranes (19 its antigen-antibody complex, respectively. The profile of antigen-antibody complex using the purified Fragment A is shown in Fig. 11C. The upper two bands indicated by arrows correspond to the F, and F, peptides of Fragment A. Since the formation of antigen-antibody complex was coupled with the disappearance of the ATPase activity of Fragment A, this confirms that these polypeptides are derived from the component responsible for the ATPase activity of Fragment A.
Subunit Composition of Dynein-We now can identify the principal polypeptide subunit of dynein on Na dodecyl-SO,polyacrylamide gel even in the case of crude fraction of dynein. Na dodecylS0, electrophoretic patterns of the antigen-antibody complex of a low ionic strength dynein extracts from fresh sperm and from glycerinated sperm stored at ~ 15" for 1 year are shown in Fig. 12, C and D, respectively. Two major bands are observed on the gels in addition to the bands deriving from the antiserum.
The patterns of the fresh and glycerinated sperm appear identical. The top band corresponds to the band described previously as the dynein A component (8) or A (a)-dynein (4,9). The lower band with a molecular weight of 54,000 indicated by arrow coincided with the tubulin band. Three to four faint bands slightly below the top dynein band were always observed on Na dodecyl-SO,-gels of the antigenantibody complex. It seems likely that the tubulin-like protein and the several other minor proteins are not subunit components of dynein and that their presence in the antigen-antibody precipitate is due to co-precipitation of denatured protein formed during the incubation with antiserum. The molecular weight of the top band corresponding to the dynein subunit was estimated from its position after co-electrophoresis of dynein-antibody complex and Fragment A on 4% Na dodecyl-SO,-polyacrylamide gel. A tentative value of 320,000 for the molecular weight of this band was determined by comparison with the positions of the bands corresponding to F, (M, 190,000), F, (M, 135,000), and reduced y-globulin (M, 60,000 and 20,000).
From these results of Na dodecyl-SO, electrophoresis of the antigen-antibody complex, it is concluded that the storage of sperm in 50% glycerin has no effect on the dynein structure, the principal subunit of the dynein molecule has a weight of about 320,000, and the principal peptide chains of Fragment A are F, (M, 190,000) and F, (M, 135,000).

DISCUSSION
One of the characteristic features of flagellar ATPase (dynein) from sea urchin spermatozoa is that the enzyme can recombine with the outer microtubules in the presence of divalent cations (5,6). When dynein is digested with trypsin, a protein designated as Fragment A is produced which retains the ATPase activity but lacks the ability to recombine with the microtubules.
Preparations A, crude Fragment A (17 pg of protein). It was prepared by adding trypsin to a low ionic strength extract. The protein ratio of the extract to trypsin was 12:l. After incubation for 3 hours at 20", the reaction was terminated with trypsin inhibitor. B, antigenantibody complex of crude Fragment A. C, antigen-antibody complex of purified Fragment A. If the complex formation is complete, 8.5 Fg of Fragment A is applied on the gel. A few faint bands above F, band were derived from the tryptic intermediates of dynein. 1 The protein bands derived from antiserum were indicated by the bars. Electrophoresis was run on 4% gel. peptide chains, F, and F,, with molecular weights of 190,000 and 135,000, respectively.
Assuming that the intensities of the bands are proportional to the mass of protein therein, the slightly greater intensity of the F, band suggests that Fragment A contains one F, chain and one F, chain. The fact that F, peptide is not convertible to F, peptide by more extended digestion seems to support this hypothesis. In this case, the molecular weight of the undissociated F, + F, fragment would be 325,000. The discrepancy between this value and the 370,000 to 400,000 molecular weight of undenatured Fragment A might plausibly be ascribed to the presence of some heterogenous small peptides in Fragment A, possibly resulting from some nicking of the peptide chains during digestion.
Recently, several groups have observed that both low ionic strength dynein extracts from cilia and flagella (8,9,20,21) and whole flagella (4) give two "high molecular weight" polypeptide bands on Na dodecyl-SO,-gels. Linck (8) originally interpreted that the upper (A) and the lower (B) polypeptide bands as corresponding to ATPase-and structuralcomponents of dynein, respectively, but Kincaid et al. (4) and Burns and Pollard .(9) have shown that one form of dynein ATPase is composed of only the A-polypeptide, and that the B-polypeptide is derived from a separate protein (B-or p dynein) which may or may not have ATPase activity. Our results show that the high molecular weight polypeptide seen in the Na dodecyl-SO,-electrophoresis of a low ionic strength ' Unpublished observation. It was concluded that the top band (indicated by arrow) corresponds to dynein (see Text). The tubulin band (M, 54,000) indicated by the lower arrow was observed below y-globulin band (M, 60,000). The other main bands indicated by the bars were derived from the antiserum. Four per cent gel was used.
extract from cilia and flagella is certainly a subunit component of dynein ATPase. However, since we observed only a single electrophoresis band in this region, while several other groups (8,9,20,21) have observed two bands in low ionic strength extracts, either our electrophoresis system failed to resolve the two bands, or the protein corresponding to the second band was not solubilized from the flagella under our dialysis conditions. The molecular weights reported for the A polypeptide have ranged from 320,000 (this paper) to 560,000 (21). The reason for this disagreement seems to be that there are few marker proteins with known subunit molecular weights above 200,000, and that an accurate calibration curve cannot be drawn by using polymers of smaller peptide chains. Therefore, a different technique may need to be used to determine the molecular weight of the dynein subunit.
When anti-dynein serum was added to purified dynein or to Fragment A, the ATPase activity of the mixture was not completely inhibited.
In the case of a low ionic strength extract from fresh sperm, the ATPase activity which was not inhibited by anti-dynein serum amounted for 20% of the total. A similar proportion of antiserum-insensitive ATPase activity was observed in glycerinated flagella. These results suggest the possible presence of another ATPase different from dynein ATPase in flagella. According to Watanabe and Falvin (22), Chlamydomonas flagella contain two forms of dynein with sedimentation coefficients of 12 S and 18 S, and also a Ca2+-ATPase with a sedimentation coefficient of 3.0 S. Blum (23) has reported that Tetrahymena cilia contain two ATPases, one extractable and the other nonextractable at low ionic