Characterization of antibodies as probes for structural and biochemical studies of tektins from ciliary and flagellar microtubules

Doublet microtubules of sea-urchin sperm flagella can be fractionated in two ways. First, extraction of doublet tubules with Sarkosyl detergent produces resistant ribbons of three protofilaments composed of tubulin (Meza et al. 1972; Witman, 1970; Witman et al. 1972a,b) and a subset of microtubule proteins that do not comigrate with tubulin (Linck, 1976). Second, doublet microtubules can be fractionated by a combined extraction with Sarkosyl-urea into extended filaments, 2—6 nm in diameter, principally composed of three equimolar polypeptide groups with apparent molecular weights (M_r) of 47 000, 51 000 and 55 000 (Linck & Langevin, 1982; Linck et al. 1985). These filament proteins are collectively named tektins. Initial findings indicated the tektins are similar in their properties to mammalian intermediate filament proteins (Linck & Langevin, 1982; Linck & Stephens, 1987).

In previous studies antibodies were prepared against an equimolar mixture of the three major tektins and used to study tektin location in axonemal microtubules. The antibodies (referred to as anti-tektins) stained ciliary and flagellar axonemes fixed with methanol for immunofluorescence microscopy (Amos et al. 1985; Linck et al. 1985). By immunoelectron microscopy the anti-tektins labelled 2–3 nm diameter fibrils at the ends of the Sarkosyl ribbons and similar fibrils remaining after extraction of doublet microtubules with Sarkosvl-urea. The results suggested that these intermediate filament-like proteins exist as filaments or possiblv protofilaments in the Sarkosyl-resistant domain of the A-microtubule wall (Amos et a!. 1986; Linck, 1982; Linck & Langevin, 1982).

For our present studies it seemed preferable to have separate, polyclonal antibody preparations against each of the three major tektins. For example, specific anti-tektins might provide information about the identity and relatedness of the tektins and permit the localization of each tektin within the axoneme. This paper reports the preparation of such antibodies against individual tektins from two sea-urchin species and characterizes their crossreactivities with sea-urchin tektins and proteins from other sources by immunoblotting and immunofluorescence microscopy. Preliminary and additional results appear elsewhere (Linck et al. 1986; Steffen & Linck, 1987; unpublished).

Ethylenediamine tetraacetate (EDTA), sodium dodecyl sulphate (SDS), polyacrylamide gel electrophoresis (PAGE), isoelectric focusing (IEF), two-dimensional (2-D), phos-phate-buffered saline (PBS: 10mM-phosphate, 0·15M-NaCl, 2·7mM-KCl, pH 7·4), PBS-Tween-NaN₃ (PTN: PBS containing 0·05% Tween 20 and 0·02% NaN₂), bovine serum albumin (BSA), room temperature (RT).

Echinoderms: Lytechimis piel us and Strongylocentrotus purpuratus. Molluscs: Aequipecten irradions. Chickens: Leghorn, white, female, 7 months old. Rabbits: New Zealand, white, female, 4–5 kg.

Biochemical reagents deemed important in this work were obtained from the following sources: SDS, electrophoresis purity reagent from Bio Rad (Richmond, CA, USA); sodium dodecyl sarcosinate (Sarkosyl), Ciba-Geigv Corp. (Greensboro, NC); ampholytes, LKB, pH 5·7, Cat. no. 1809–121, and pH 3·5–10, Cat. no. 1809–101 (Bromma, Sweden); 4-chloro-1 -naphthol, Aldrich Chemical Co. (Milwaukee, WI); rabbit anti-(.S. purpuratus egg)-tubulin, Polysciences, Inc. (Warrington, PA); fluorescein-conjugated goat anti-rabbit IgG, Cooper Biomedical (Malvern, PA); biotin-conjugated goat anti-rabbit IgG, Sigma Chemical Co., Cat. no. B9642 (St Louis, MO); avidin-rhodamine, Sigma, Cat. no. A3026; rabbit anti-chicken desmin, DA KO, Cat. no. A611, Lot no. 096b (Santa Barbara, CA) and Miles Scientific, Cat. no. 65793–1, Lot no. A200 (Naperville, IL); luciferin, luminol, and 4-methylumbelliferone, Sigma, Cat. nos L6882, A8511 and M1508, respectively; ¹²³1-labelled goat anti-rabbit IgG, New England Nuclear (Boston, MA).

The following procedures are described in detail as referenced: SDS-PAGE was performed according to Laemmli (1970). 2-D IEF/SDS-PAGE was modified from that of O’Farrell (1975) as described by Linck & Langevin (1982). IEF pH gradients were measured from duplicate gels focused without protein and cut into 16 mm × 5 mm segments; two matched segments were equilibrated by shaking in 1 ml 0·1 M-NaCl for 1 h at RT ; pH was measured with a Radiometer PHM 82 meter with a combined electrode, type GK2321-C or GK2322-C (Copenhagen, Denmark). Immunoblotting was modified from Towbin et al. (1979) as follows: electrotransfer was performed in 10% methanol, 0’1% SDS, 0·192 M-glycine, 0·-025 M-Tris HC1, at 25 mA overnight, and then for 1 h under the same conditions minus SDS. Nitrocellulose strips were stained as indicated with one of the following: (1) Amido Black; (2) an appropriate dilution of rabbit anti-tektin IgG followed by peroxidase-conjugated goat anti-rabbit IgG, followed by addition of substrate, using either 4-chloro-l-naphthol or luciferin/luminol/4-methyl-umbelliferone (the latter according to Laing, 1986); or (3) ¹²⁵1-labelled goat anti-rabbit IgG.

The following were purified and fractionated according to previously published procedures: sea-urchin sperm flagellar axonemes (Gibbons & Fronk, 1972); Sarkosyl-resistant protofilament ribbons (Linck, 1976); sea-urchin embrvo ciliary axonemes (Stephens, 1977); chicken gizzard desmin (Geisler & Weber, 1980, as modified by Linck & Langevin, 1982). Molluscan gill cilia were a gift from R. E. Stephens and were prepared according to Linck (1973) and Stephens (1983).

Tektin filaments were prepared, as previously described (Linck & Stephens, 1987), by twice extracting flagellar doublet microtubules of S. purpuratus or L. pictus seaurchin sperm flagella with 0·5% Sarkosyl, 2·0 M-urea, 50 mM-Tris HCl, 50 mM-lysine, 1 mM-EDTA, pH8 0 at 4°C. The pellets of tektin filaments were resolved by preparative SDS-PAGE into the three major tektin polypeptide bands (47, 51 and 55 (×10³) M_r). The proteins were visualized by precipitation in the gel with 0·5M-KC1, cut out, electroeluted, dialysed against deionized water, and extracted with 100% acetone. Chicken gizzard desmin was similarly purified.

Preimmune sera were obtained from rabbits by taking 35–50ml blood from the ear vein, clotting at 37°C for 1 h, removing the clot and centrifuging at 10 000g’ for 10 min. Serum from each rabbit was checked by SDS-PAGE immunoblotting and immunofluorescence to determine whether it crossreacted with sea-urchin flagellar axoneme proteins. Rabbits not immune to axonemal proteins were bled weekly, until at least 45 ml of preimmune serum had been collected. Crude IgG was then isolated by two precipitations in 40% ammonium sulphate; the final crude IgG was dialvsed against PBS and stored frozen in samples at —80°C. For immunization, approximately 0·2 mg of a given purified tektin was sonicated into 1·1 ml of 1 mM-Tris-HC1, 0·1 mM-EDTA, pH 7·8; this sample was then emulsified with an equal volume of Freund’s complete adjuvant. Animals were shaved along the vertebral column and 0·1 ml of adjuvant mixture was injected at each of 10 sites on each side. Animals were boosted after 5 weeks with a similar preparation of antigen in incomplete adjuvant. In the sixth week, and weekly thereafter, the animals were bled and IgG purified in the manner described above.

Affinity purification was carried out as follows: 10mg of L.pictus tektin filaments were solubilized in 2% SDS, 5% 2-mercaptoethanol, 19·2 mM-glycine, 2·5 mM-Tris HC1, pH 6·8. After boiling for 2 min, the sample was dialysed exhaustively against deionized water, then freeze-dried. The SDS-dcnaturcd tektin filament preparation was conjugated to cyanogen bromide-activated Sepharose (Pharmacia, Sweden), according to the manufacturer’s recommended procedure. After thoroughly washing each affinity column with PBS, immune sera were passed through a 5-ml bed, continuously cycling the IgG for 1 h at RT. The column was then washed extensively with PBS; the wash fraction was recovered and found to contain negligible anti-tektin activity by SDS-PAGE immunoblotting procedures. Bound antibody was released by elution with 0·2M-glycine, pH 2·3, collected in 1-ml fractions and immediately neutralized with 1ml 0·2 M-Tris-base. Antibody was then dialysed against 1:10 (v/v) diluted PBS, divided into samples, freeze-dried and stored at — 80°C. The affinity-purified anti-tektins were reconstituted by adding deionized water to yield a normal PBS concentration. A separate affinity column was maintained for each anti-tektin.

Diluted samples of sperm were attached to poly-L-lysine-coated glass coverslips, fixed in methanol at ∽20°C for 30 min, washed three times (10 min each) in PTN, treated in PTN-1% BSA for 20 min, incubated with primary antibody in PTN-BSA overnight at RT or 2h at 37°C, washed three times (10 min) in PTN, incubated with secondary antibody in PTN-BSA for 1 h at 37°C, and washed three times (10min) in PTN. Goat anti-rabbit IgGs conjugated to fluorescein or biotin were used as secondary antibodies. Specimens stained with biotin conjugates were washed three times (10 min) with PTN, treated with poly-L-lysine (0·l mg ml⁻¹) in PTN for 10min to reduce non-specific DNA-avidin binding, incubated in avidin-rhodamine in PTN-BSA for 45 min at 37°C, and washed three times (10min) in PTN. All specimens were then mounted in paraphenyline diamine (1 mg ml⁻¹) in 90% glycerol, 0·1 M-TrisHCI, pH 9.

Eor unfixed specimens the methanol step was omitted from the protocol described above. Demembranated axonemes were prepared for this purpose by extracting whole sperm in 1% Triton X-100, 10 mM-Tris HCl, 0·15 M-KC1, 5 mM-MgSO₄, 0·5 mM-EDTA, pH 8, for 20 min at 4°C, and centrifuging at 1000g for 5 min to pellet most of the heads. Axonemes in the supernatant fraction were pelleted at 9000g for 6 min, re-extracted with the Triton solution, and then washed twice with the solution minus Triton.

Specimens were examined with a Zeiss UEM microscope equipped with Neofluar 25×/0·8 NA and 63×/1·4 NA objectives. Photographs were taken using either Kodak Tri-X or hvpersensitized Technical Pan 2415 35mm film (Lumicon, Livermore, CA, USA), the latter being developed according to the recommended procedure.

Tektin filaments were obtained by twice extracting purified flagellar doublet microtubules from 5. purpuratus or L. pictus with 0·5% Sarkosyl — 2 M-urea. These conditions were determined, as described by Linck & Stephens (1987), by varying the concentration of both Sarkosyl and urea to optimize the purity and yield of the three major polypeptides (M_r = 47, 51 and 55(× 10³)), referred to in this paper as the 47K M_r, 51K M_r, and 55K M_r tektins. SDS-P.AGE analysis of such tektin filament preparations is illustrated in Fig. 1. The filament proteins were preparativelv separated by SDS-PAGE for use as immunogens; each tektin band was cut out according to molecular weight, avoiding the zones between the bands, clcctroeluted and analysed for purity by SDS-PAGE. As seen in Fig. 1, the three molecular weight fractions were pure.

As described in Materials and methods, preimmunc IgG and immune IgG fractions were prepared, adjusted to twice the original serum IgG concentration, and stored at — 80°C. Affinity-purified antibodies against S. purpuratus and L. pictus tektins (referred to here as anti-tektins) were obtained using SDS— denatured tektin filaments from L. pictus as the affinity probe in both cases. A different affinity column was maintained for each anti-tektin. The characterization of the affinity-purified antibodies is reported below, primarily for .S. purpuratus. The results were essentially identical for L. pictus.

To a first approximation, the specificities of the anti-tektins were determined by one-dimensional SDS-PAGE immunoblot analysis of axoncmal proteins. As shown in Fig. 2, the anti-tektins recognized only proteins with the same apparent molecular weights as the three major tektins. Antibodies to the 55K M_r tektin (anti-55) specifically stained a 55K M_r. band, the anti-51 stained a 51K M_r band, and the anti-47 stained a 47K M_r band but also crossreacted significantly with the 55KM_r band. The anti-tektins were never seen to stain polypeptides with molecular weights higher than 55K M_r, and rarely did they stain lower molecular weight polypeptides. By 1-D SDS-PAGE immunoblotting the specificities of antibodies raised against L. pictus tektins were found to be qualitatively and quantitatively similar to those from S. purpuratus, as reported elsewhere (Steffen & Linck, 1987; unpublished).

The specificities of the anti-tektins were next analysed by 2-D SDS-PAGE immunoblotting against purified tektin filament proteins and are illustrated in Fig. 3 for S. purpuratus. The isolelectric points and nominal molecular weights of the tektin filament proteins are given in Table 1. Anti-55 strongly stained the principal 55KM_r polypeptide a (see Fig. 3) and an adjacent ≈55K M_r spot b; the anti-55 also stained polypeptide spot x at ≈45K M_r, pl 6–4 that occasionally, but rarely, appeared in filament preparations. Anti-51 specifically stained the two main 51KM_r polypeptides a′ and b′ ; although not apparent in the 2-D blot of Fig. 3, anti-51 also stained an ≈53K M_r polypeptide, which is evident in the original tektin filament preparations (Fig. 1). Anti-47 strongly stained the major 47K M_r polypeptide c″ and a weaker spot d″; in addition, it crossreacted weakly, but consistently, with the main 55K 47 _r polypeptide a. These results should be compared with those for the Sarkosyl ribbon proteins presented below.

To look more closely at the complex mixture of proteins in the tektin—tubulin region, axonemes were extracted with 0·5% Sarkosyl to reduce the amount of tubulin present. The resulting Sarkosyl-resistant protofilament ribbons were then analysed by 2-D immunoblotting (Fig. 4). The isoelectric points of the Sarkosyl protofilament ribbon proteins are given in Table 1. The relative specificities and crossreactions of the anti-tektins were observed to be essentially identical to those found in the tektin filament preparations. Anti-55 stained only the main 55K M_r spot a and adjacent spots b and c; no other polypeptides stained with anti-55 (in particular, polypeptide x in Fig. 3 was not observed). Anti-51 specifically stained two 51KM_r spots (a′ and b′). Anti-47 stained four polypeptides (a″, b″, c″ and d″) and crossreacted with the principal 55K M_r. spot a. The anti-tektins did not stain the tubulins or the 77K/83K M_r polypeptide pair. Two alpha and one beta tubulin subunits were identified by anti-tubulin staining.

Each of the anti-tektins was examined for its staining distribution in whole, fixed sperm by immunofluorescence microscopy. Representative results are shown in Figs 5 and 6. As with the immunoblot data, the results for L. pictus and S. purpuratus were essentially identical. All three anti-tektins stained the entire length of the sperm flagellum; also, all three antibodies more faintly stained thin-tip regions, ≈4μm in length, that presumably correspond to the A-tubules extending past the termination of the B-tubules (Figs 5F, 6). In addition to the flagellum, several other structures were specifically stained. All anti-tektins stained the basal body and acrosome regions (Figs 5, 6). Furthermore, anti-47 stained an area coincident with the mitochondrion.

Of particular interest was a specific anti-51 staining of the outline of the sperm head or nuclear envelope; we refer to this region microscopicallv as the sperm head envelope. In the case of L. pictus sperm, anti-L. pictus 51KM_r (affinity-purified with L. pictus tektin filaments) stained the sperm head envelope as brightly as the flagellum (Fig. 5E). In the case of S. purpuratus sperm, non-affinity-purified IgG against the S. pttipur-atus 5IK M_r tektin also stained the head envelope brightly (Fig. 6), whereas no staining was observed with preimmune IgG (not shown). However, following affinity purification of the anti-S. purpuratus tektin with L. pictus tektin filaments, the head envelope staining in .S, purpuratus sperm was substantially reduced or eliminated (Fig. 5B).

The anti-tektins were also examined for their ability to stain unfixed axonemes (Fig. 7). Both anti-47 and anti-55 stained the unfixed axonemes brightly, while anti-51 stained them extremely weakly. All three anti-tektins did, however, show bright staining at the ends of the broken flagellar axonemes, 5–30μm in length. The unfixed axonemes also stained brightly with antitubulin (not shown).

The antibodies to flagellar tektins were examined for their abilities to crossreact with ciliary proteins from echinoderms and molluscs (Fig. 8). Ciliary axonemes were isolated from S. purpuratus blastulae and gastru-lae embryos and from molluscan (A. irradions) gill tissue and probed with non-affinity-purified antibodies against S. purpuratus flagellar tektins. In both cases cilia were stained, as observed by immunofluorescence microscopy (not shown). In SDS-PAGE immunoblots of X. purpuratus cilia the anti-tektins crossreacted with polypeptides of molecular weights similar to those in flagella and showed identical specificities; i.e. the anti-55 stained only a 55K M_r band, the anti-51 primarily stained a 51K M_r. band (and more weakly 47K, 53K and 55KM_r bands), and the anti-47 prominently stained the 47K M_r band but crossreacted with the 55K M_r band.

In the case of molluscan gill cilia the immunoblot staining was more complex. The anti-55 crossreacted with polypeptide bands with molecular weights of approximately 52K, 54K and 56K M_r; anti-51 crossreacted with 52K and 55KM_r bands; and anti-47 crossreacted with 45K, 49K, 52K and 55K M_r bands.

Anti-tektins were not found to crossreact with purified chicken gizzard desmin by 1-D SDS-PAGE immunoblot; similarly, polyclonal antibodies to desmin (those prepared in our laboratory and those commercially available from DAKO and Miles) were not found to crossreact with sea-urchin flagellar tektins, even though sensitive methods of secondary immuno-stain-ing were employed, including the ^l25I radiolabel and the peroxidase—luminol—luciferin methods.

Echinoderm (sea-urchin) sperm flagellar doublet microtubules contain a set of related polypeptides called tektins that are the major components of the 2–6 nm Sarkosvl-urea-insoluble filaments (Linck et al. 1985; Linck & Langevin, 1982). Three major tektins with apparent molecular weights of 47K, 51K and 55K have been characterized and found to be related, but distinct, gene products; they are different from tubulin but similar to intermediate filament proteins in their structural, electrophoretic and solubility properties (Amos et al. 1986; Beese, 1984; Linck & Langevin, 1982; Linck & Stephens, 1987; cf. Steinert et al. 1985). Previous studies with antibodies raised against undenatured tektins indicated that the antigenic sites of the tektins are masked by their structural organization within the microtubule, except where thin filaments protrude at the ends (Linck et al. 1985). The goal of our present efforts has been to prepare antibodies to each of the tektins for use as specific probes for further molecular studies. We discuss here the specificities, crossreactivities and immunofluorescence staining of these anti-tektins.

The specificity of each anti-tektin is largely restricted to the original immunogen. Anti-55 is the most specific, recognizing onlv the 55K M_r. band in 1-D immunoblots of flagellar axoneme proteins. In 2-D immunoblots of tektin filaments or Sarkosvl ribbons, anti-55 recognizes three polypeptides with narrowly defined molecular weights and isoelectric points (spots a, b and c in Figs 3, 4). These three polypeptides comigrate at ≈55K M_r on 1-D SDS-PAGE, and all would have been included in the preparation used to raise the anti-55 antibodies. We have not determined whether these polypeptides are distinct or related proteins. One insight into the structure of the 55KM_r tektin may come from the staining by anti-55 of spot x at —45 K M_r and pl 6–4 (Fig. 3). This spot is usually not present in the preparations and is not recognized by anti-47 or anti-51 ; therefore, it may represent a proteolytic fragment of the 55K M_r tektin. If this is the case, then the pl shift of the major fragment from 7 0 to 6–4 implies that the — 10K M_r cleavage fragment would be a highly basic terminal domain.

Anti-51 is specific in 2-D immunoblots for two polypeptide spots (a′ and b′) with masses of 51 K M_r (Figs 3, 4). For the same reasons as with anti-55, we cannot define the nature of these two spots. It is clear, however, that anti-51 does not recognize the 51 K M_r polypeptide c′ that is present in Sarkosyl ribbons in an amount equal to the 51K M_r tektin a′ spot. Faint anti-51 staining of 47K, 53K and 55K M_r, bands (better seen in 1-D immunoblots, see Fig. 8C) coidd be attributed to contamination of the 51K 47,. antigen with these neighbouring polypeptides; however, the presence of an ≈53 K M_r band consistently in tektin filament preparations (Fig. 1) argues for the existence of at least one other bona fide filament component (Table 1).

In 2-D immunoblots anti-47 stains two to four polypeptides, 47KM_r spots a″, b″, c″ and d″, which vary slightly. Again, it is not known whether thèse spots are related. Interestingly, anti-47 crossreacts with the 55K M_r tektin spot a. It is unlikely that the 47K M_r antigen preparation was contaminated by 55K M_r antigens, since a substantial area (including the 51K M_r region) was cut out between these two tektins during the electrophoretic purification of the 47K 47,. band. This observed crossreaction thus suggests a degree of immunological (i.e. structural) similarity between the 47K and 55KM_r tektins, a conclusion reinforced by the fact that L. pictus and S. purpuratus yield identical results.

Sperm flagellar tektins were originally defined by their solubility properties and apparent molecular weights (Linck et al. 1985; Linck & Langevin, 1982). The polyclonal antibody studies we have reported here help to establish the identity of the three major molecular weight groups of tektins previously studied, but they also suggest that the tektins may be a more complex family of proteins. Of the polypeptides on 2-D immunoblots that stain with anti-tektins, spots 47K M_r-c″, 51KM_r-a′ and -b′, and 55K M_r-a correspond to the major components of the tektin filaments, but a less abundant 53K M_r component is also present. As stated earlier, we cannot yet determine the relationship of the other 2-D spots to the main tektins. Clearly, some polypeptides may be isoelectric variants of their parent tektin (e.g. 47K M_r-a″, b″ and d″ may be phosphorylated variants), others may be novel tektins (e.g. spot 51KM_r-c′ and the ≈53KM_r band), and others may not even be tektins. The potential complexity of the tektins is particularly evident in the crossreactions of anti-echinoderm tektins with molluscan ciliary proteins (Fig. 8E-H), in which anti-47 crossreacts with molluscan polypeptides b, c, d and e, anti-51 with bands b and c, and anti-55 with bands a, b and c. These crossreactions with molluscan proteins have been more carefully analysed and are to be reported elsewhere (R. E. Stephens, personal communication). Seaurchin anti-tektins have also been found to crossreact with tektin-like proteins in Chlamydomonas flagella (C. Silflow & K. Joyce, personal communication) and Drosophila testes and sperm (E. Raff & J. Hutchens, personal communication).

Immunofluorescence microscopy demonstrates that all three tektins are present throughout the length of flagellar and ciliary axonemes and in the basal body region (Figs 5–7). In a related study we have demonstrated that all three anti-tektins label basal bodies and each of the nine doublet microtubules of sperm flagella, and also label centrioles and/or centrosomes of Chinese hamster ovary cells (Steffen & Linck, 1987; unpublished). At the distal tip of the axoneme (Figs 5F, 6), a thinner segment, ≈4μzm in length, is also stained by the anti-tektins. In most species it is in this region of cilia and flagella that the B-tubules terminate and the A-tubules extend as singlet microtubules (Gibbons, 1961 ; Gibbons & Grimstone, 1960; Linck & Langevin, 1981; Satir, 1967; Stephens, 1970); thus, the anti-tektins are presumably staining the A-tubules. These observations are in agreement with our earlier results (Linck et al. 1985) and also with those of Stephens (1977, and personal communication), who showed that ‘Component 20’ (of Linck, 1976) is the 55K M_r tektin synthesized during sea-urchin embryo ciliogenesis in an amount consistent with that of a length-determining component of ciliary (and flagellar) A-tubules.

Demembranated, unfixed axonemes were also stained by anti-47 and anti-55, but weakly stained by anti-51. All three anti-tektins did, however, stain brightly the ends of the unfixed, broken axonemes (Fig. 7). In our previous studies (Linck et al. 1985), anti-tektins raised against undenatured tektin filaments did not label flagellar microtubules, unless the tubules were first fixed or disrupted. The positive staining of undisrupted flagellar microtubules reported here may relate to the fact that our individual anti-tektins were raised against SDS-denatured tektins. These anti-tektins may, therefore, recognize a greater number of antigenic sites exposed on the microtubule surface; the weaker staining with anti-51 may imply that the 51KM_r tektin is less exposed on the surface. The brighter staining of the ends of axonemes with all anti-tektins correlates well with our earlier observation that antibodies do label thin filaments protruding from the frayed ends of microtubules (Amos et al. 1986).

Of interest, also, is the observation that the 51K M_r tektin antibodies stain a region that we refer to as the sperm head envelope (Figs 5, 6), as intensely as they do the flagellum. Both non-affinity-purified anti-S. purpuratus and anti-L. pictus, affinity-purified against L. pictus tektins, stain the head envelope (compare Figs 5E, 6). However, affinity purification of anti-S. purpuratus-51 with L. pictus tektin filaments substantially reduces head envelope staining in S. purpuratus, conceivably because of differences in the immunological determinants of the envelope components in the two species. Since originally reporting sperm head envelope staining (Linck et al. 1985), noting that it might be artifactual, we have taken precautions to avoid nonspecific staining. Treatment of the specimens with polylysine was found to block non-specific DNA-avi-din binding. Without polylysine treatment, avidin-rhodamine stains the entire sperm head; after treatment, only the thin outline of the head envelope is stained with anti-51 and not with anti-47 or anti-55. We conclude that the crossreaction of anti-tektins with the envelope region is likely to result from the similarities between tektins and intermediate filament proteins (Amos et al. 1986; Beese, 1984; Linck & Langevin, 1982; Linck & Stephens, 1987) and between intermediate filament proteins and nuclear lamins (Fisher et al. 1986; McKeon et al. 1986). Recently, it has been reported that one monoclonal antibody against seaurchin tektins crossreacts with nuclear envelopes and lamins from cultured mammalian cells (Chang & Piperno, 1987).

The properties of tektins indicate their similarities to mammalian intermediate filament proteins. Chang & Piperno (1987) have recently reported that a monoclonal anti-tektin crossreacted with desmin and a polyclonal, non-affinity-purified anti-desmin crossreacted with a certain tektin. We investigated the possibility of crossreactions between the anti-tektins and chicken gizzard desmin and between anti-desmin and tektins; but no evidence of such crossreactions was observed, using either the peroxidase/4-chloro-l-naphthol method of immunoblot staining or more sensitive methods of detection (i.e. peroxidase/luminol-luci-ferin or ¹²⁵1 label). Recent studies on cultured mammalian cells, however, have demonstrated that one of our affinity-purified anti-tektins recognizes a keratin in HeLa cells (Steffen & Linck, 1987). Our finding that certain rabbits were preimmune to sea-urchin tektins and/or mammalian intermediate filament proteins urges prudence in the interpretation of non-purified, polyclonal antibody crossreacfivities between specific IF proteins and tektins.

This investigation was supported by USPHS grant GM 35648 (to R.W.L.) from the National Institute of General Medical Sciences. We are especially grateful to Richard Erickson and Bettina Mutter for expert assistance with the SDS-PAGE analysis and purification of proteins.