p-Azidosalicyl-5-amino-6-phenoxybenzimidazole photolabels the N-terminal 63-103 amino acids of Haemonchus contortus beta-tubulin 1.

Benzimidazoles (BZ) are broad spectrum anthelmintics thought to exert their effects by interacting with and disrupting the functions of microtubules. However, direct biochemical evidence for binding between BZ and tubulin has not been shown nor is it known what sequences in tubulin interact with BZ. In this study, a photoactive analogue of 2-acetamido-5-(3-aminophenoxy)benzimidazole that has biological activity similar to other benzimidazoles was synthesized and used to photoaffinity label cell lysates from the parasitic nematode of sheep Haemonchus contortus. The photoactive analogue, 2-acetamido-5-[3-(4-azido-3-I-salicylamido)phenoxy]benzimidazole or I-ASA-BZ, was shown to photolabel a 54-kDa protein that was specifically immunoprecipitated with anti-tubulin monoclonal antibodies. Tubulin photoaffinity labeling by I-ASA-BZ was also inhibited with molar excess of various BZ analogues and colchicine. Interestingly, I-ASA-BZ photoaffinity-labeled the β- and not the α-subunits of tubulin. Proteolytic digestion of I-ASA-BZ-labeled tubulin with Staphylococcus aureus V8 proteinase revealed one major peptide with an apparent molecular mass of 3.5 kDa. Exhaustive digestion of I-ASA-BZ-labeled β-tubulin with trypsin resulted in two fractions containing radioactive peptides. Protein sequencing of the high performance liquid chromatography-purified tryptic ASA-BZ-photolabeled peptides identified the N-terminal 63-77 and 78-103 sequences as the BZ binding domain.

addition to their anti-tumoral and anti-fungal activity (3)(4)(5). It is presently believed that BZ exert their cytotoxic effects by binding to and disrupting the functions of the microtubule system (6 -9). The implication of tubulin as target for BZ has been supported by drug binding studies using enriched extracts for helminth and mammalian tubulin (6 -8). Moreover, competitive drug-binding studies using mammalian tubulin have shown that BZ compete for colchicine binding and inhibit the growth of L1210 tumor cells in vitro (10 -12). However, BZ display selective toxicity toward nematodes when administered as anthelmintics and are not toxic to the host (1). Such selective toxicity is in contrast to the effect of BZ on the in vitro polymerization of mammalian tubulin and the growth of tumor cells (3)(4)(5). Differences in both the affinity between host and parasite macromolecules for BZ (13,14) and the pharmacokinetics of BZ within the host and the parasite have been suggested as responsible for BZ selective toxicity (15). Therefore, the nature of BZ selective toxicity remains unclear. The direct identification of BZ receptor(s) in nematodes would help clarify this uncertainty.
The frequent application of BZ in the control of parasitic nematodes has led to the rapid selection of BZ-resistant populations (16,17). Recently, mutations in tubulin genes have been correlated with the development of BZ resistance (18 -20), and mutations conferring BZ resistance have been mapped to the locus encoding the ␤-tubulin (21)(22)(23)(24)(25)(26)(27). However, without direct biochemical evidence for binding between tubulin and BZ, it is not clear if the mutations in the ␤-tubulin genes that confer resistance to BZ affect direct BZ binding to ␤-tubulin. In addition, we have shown recently (5) that BZ are substrates for the P-glycoprotein drug efflux pump that mediates the multidrug resistance phenotype in tumor cells (28,29) and in some parasites (30,31). Therefore, an enhanced drug efflux mechanism in resistant H. contortus could confer resistance to BZ.
To further characterize the biochemical and molecular basis for the action of BZ, it was of interest to show directly the receptor(s) of BZ using a photoaffinity labeling assay. This approach has been used to identify the receptor and the ligand binding domain on such a receptor for many ligands including several antimitotic drugs. Photoactive analogues of vinblastine, colchicine, rhizoxin, and taxol have been used previously to demonstrate their direct binding to tubulin and to identify their photolabeled sequences (32)(33)(34)(35). Moreover, and in support of the latter approach, a recent study (36) using high resolution images of taxol-bound tubulin has provided structural evidence for taxol binding domain that is consistent with earlier results using a photoactive analogue of taxol (35). In this report, we describe the synthesis of the photoactive-radioactive analogue of 2-acetamido-5-(3-aminophenoxy)benzimidazole, 125 I-ASA-BZ, and its use in a photoaffinity labeling assay. The results of this study show for the first time the direct binding of BZ to tubulin. In addition, we show that the 125 I-ASA-BZ binding domain is localized to a 36-amino acid sequence (Ala 63 -Lys 103 ) in the N-terminal of ␤-tubulin. The implication of these findings with respect to tubulin drug binding sites and BZ species selectivity will be discussed.
Synthesis and Iodination of Photoactive Analogue of BZ-Numerous BZ and pro-BZ analogues have been synthesized over the years. Substitution in the imidazole carbamate has been shown to reduce the drug efficacy against parasitic nematodes. For example, hydrolysis of the methyl carbamate group of BZ to an amino group results in loss of microtubule inhibitory activity as well as reduced efficacy (15). In contrast, alkyl or aromatic additions to the 5-or 6-position of the BZ ring generally result in more potent anthelmintics. Therefore, all the chemical modifications that were carried out on the amino group of 2-acetamido-5-(3-aminophenoxy)benzimidazole were distal (6-substitute group) to this critical region as shown in Fig. 1. 2-Acetamido-5-(3aminophenoxy)benzimidazole has been shown to be efficacious against gastrointestinal nematodes in farm animals (37). Moreover, the photoactive analogue 2-acetamido-5-[3-(4-azido-3-125 I-salicylamido)phenoxy]benzimidazole (or ASA-BZ) was also a good inhibitor of [ 3 H]methyl benzimidazole binding to nematode proteins in our study (data not shown). ASA-BZ was therefore a suitable candidate for the identification of BZ binding protein and BZ binding site(s). ASA-BZ was synthesized by reacting 2-acetamido-5-(3-aminophenoxy)benzimidazole and NHS-ASA in dimethylformamide following previously published procedures (38). Briefly, 10 mg of 2-acetamido-5-(3-aminophenoxy)benzimidazole were dissolved in 250 l of dry dimethylformamide, and 5 l of triethylamine were added to the mixture. NHS-ASA (15 mg in 250 l of dimethylformamide) was added, and the above reaction was stirred at room temperature for 48 h in the dark. Phosphate buffer at pH 8.0 was added to hydrolyze any unreacted NHS-ASA, and the mixture was loaded onto SepPak cartridge (Millipore) and washed with 20% methanol. Bound products were eluted with 100% methanol, vacuum-dried, and ASA-BZ was extracted with ethyl acetate. The solvent was removed by vacuum drying, and the residue was dissolved in 500 l of methanol; ASA-BZ was further purified by high performance liquid chromatography (hplc) using a reverse phase column (Vydac 201HS54 C18). The chromatographic procedure consisted of a 30-min gradient of 20 -100% acetonitrile in 0.025 M ammonium acetate buffer, pH 5.5, at a flow rate of 1 ml/min and detection at 292 nm. Under these conditions, 2-acetamido-5-(3-aminophenoxy)benzimidazole, NHS-ASA, and ASA-BZ had retention times of 16.84, 18.80, and 21.30 min, respectively. Purification of ASA-BZ was done in the absence of UV detection to avoid photodestruction. ASA-BZ was iodinated by the method of Hunter and Greenwood (39). The purity of the 125 I-ASA-BZ sample was determined by thin layer chromatography on silica plates using water:chloroform: formic acid (10:90:1). Upon autoradiography, one major spot was observed and subsequently eluted with methanol. Fig. 1 shows the chemical structures of the 2-acetamido-5-(3-aminophenoxy)benzimidazole (or amino-BZ) and 2-acetamido-5-[3-(4-azido-3-125 I-salicylamido)phenoxy]benzimidazole (or 125 I-ASA-BZ).
Preparation of Tubulin-enriched Fractions-Adult worms were recovered from the mucosa of the abomasum and washed extensively with 0.8% NaCl, 0.15% NaHCO 3 at 37°C. Worms were ground into a fine powder with a mortar and pestle while being maintained in liquid nitrogen. The worm powder was resuspended in 0.025 M MES buffer, pH 6.5, containing 2 mM phenylmethylsulfonyl fluoride, 0.5 g/ml leupeptin, and 0.7 g/l pepstatin. The homogenate was centrifuged at 100,000 ϫ g for 1 h at 4°C, and the pellet was discarded. The clear supernatant was used in the binding assays. Recombinant ␤-tubulin was produced in Escherichia coli using ␤12-16 cDNA from H. contortus and purification of the recombinant protein was as described previously (40). Protein concentration was determined using bovine serum albumin as standard (41).
Photoaffinity Labeling and Immunoprecipitation of Tubulin-125 I-ASA-BZ was incubated with cytosolic proteins (100 g) or recombinant ␤12-16 (5 g) in a final volume of 50 l at 37°C for 30 min in the dark. 125 I-ASA-BZ was added in 5 l of 20% dimethyl sulfoxide to give a final concentration of 50 nM. For inhibition of photoaffinity labeling, protein fractions were preincubated with a molar excess of cold BZ analogues before the addition of 125 I-ASA-BZ. At the end of the incubation period (30 -60 min), samples were placed on ice for 10 min followed by a 10-min UV irradiation at 254 nm (UV Stratalinker 1800, Stratagene). 125 I-ASA-BZ-photolabeled tubulin from H. contortus (200 g) or recombinant H. contortus ␤-tubulin (20 g) from E. coli were immunoprecipitated with anti ␣and ␤-tubulin antibodies (5 g each) or an irrelevant second antibody (IgG 2a ) as described previously (5).

SDS-Electrophoresis and Western
Blotting-Photoaffinity-labeled samples were mixed with 50 l of 2 ϫ Laemmli sample buffer, boiled for 2 min, allowed to cool down, and resolved in 10% polyacrylamide gels (42). In order to separate ␣ and ␤ subunits, a 4 -16% gradient gel was prepared and run as described previously (43) using reduced and carboxymethylated proteins. Electrophoresis gels were fixed in 10% acetic acid and 40% methanol and stained with Coomassie blue for 1 h, destained, dried, and exposed to an XAR-5 film with an intensifying screen. For Western blot analysis, photoaffinity-labeled proteins were electrophoresed and transferred onto nitrocellulose paper in a Trisglycine buffer system in the presence of 20% methanol as described previously by Towbin et al. (44). The nitrocellulose membrane was blocked in 3% bovine serum albumin in phosphate-buffered saline and incubated with anti-chicken ␣and ␤-tubulin monoclonal antibodies (1 g/ml) overnight at 4°C. The nitrocellulose membrane was washed several times in phosphate-buffered saline and incubated with peroxidase-conjugated goat anti-mouse IgG (1:1000 dilution; Sigma) for 2 h at room temperature. The positive signal was visualized by autoradiography using the Luminol ECL method (Amersham). The above anti-␣/␤tubulin monoclonal antibodies were previously shown to bind to tubulin from H. contortus (45).
Protease Cleavage of 125 I-BZ-tubulin-The photoaffinity-labeled protein which corresponded to tubulin was excised and digested with Staphylococcus aureus V8 protease (10 -20 g/gel slice) in the well of a 15% Laemmli gel as originally described by Cleveland et al. (46). For trypsin digestion, 2 mg of recombinant ␤12-16 tubulin was photoaffinity-labeled with 50 nM 125 I-ASA-BZ. Photolabeled tubulin was reduced with dithiothreitol and alkylated with iodoacetamide before the addition of trypsin. The digestion was carried out in 0.05 M (NH 4 ) 2 SO 4 , pH 8.0, for 24 h at a trypsin to protein ratio of 1:25 (w/w). The resultant tryptic digest was dried by speed vacuum and prepared for hplc analysis.
hplc Separation and Sequence Analysis of Photoaffinity-labeled Peptides-The tryptic peptides were resuspended in 0.1% trifluoroacetic acid in water and resolved by reverse phase hplc (Vydac 201HS54 C18 RP column). The chromatographic procedure consisted of an 80-min gradient of 0 -80% acetonitrile with 0.1% trifluoroacetic acid at a flow rate of 1 ml/min and detection at 214 nm. Fractions were collected and monitored for radioactivity. The fractions containing the peak of radioactivity was further purified by hplc using a shallower acetonitrile gradient with 200-l fractions. A single peak of well resolved peptide was submitted for N-terminal sequencing at the Protein Sequencing Facilities at Queen's University in Kingston, Ontario. Amino acid sequencing of peptides was performed according to the method of Edman and Begg (47) using an Applied Biosystems Gas-Phase Model 470A sequenator according to the procedure described by Flynn et al. (48).

RESULTS
Photoaffinity Labeling of Tubulin with 125 I-ASA-BZ-To determine the BZ receptor(s) in parasitic nematodes, we have synthesized a photoactive analogue of BZ ( Fig. 1) and used it in a photoaffinity labeling assay. The sheep nematode H. contortus was used as a source of tubulin. Fig. 2A, lane 1, shows a Coomassie Blue staining of the cytosolic proteins from adult worms of H. contortus resolved on SDS-PAGE. Photoaffinity labeling was performed by incubating 100 g of cytosolic proteins with 50 nM 125 I-ASA-BZ followed by UV-irradiation and SDS-PAGE (see "Experimental Procedures"). The results in Fig. 2B show a single protein with a molecular mass of approximately 54 kDa that was specifically photolabeled with 125 I-ASA-BZ (lane 2). The incubation of cytosolic proteins with 125 I-ASA-BZ, but without UV-irradiation, did not result in the photolabeling of the 54-kDa protein (Fig. 2B, lane 1). The latter results demonstrate that the 54-kDa protein in H. contortus cytosolic fraction is covalently cross-linked by ASA-BZ. To confirm the specificity of 125 I-ASA-BZ binding to the 54-kDa protein, cytosolic proteins were incubated with 50 nM 125 I-ASA-BZ in the presence of increasing molar concentrations of the unmodified amino-BZ. The results in lanes 2-6 of Fig. 2B show the photoaffinity labeling of 54-kDa protein in the absence (lane 2) and in the presence of 1-50 M concentration of the amino-BZ (lanes 3-6). The photolabeling of the 54-kDa protein appears to be specific since the presence of excess unmodified amino-BZ inhibited its photolabeling with 125 I-ASA-BZ in a dose-dependent manner. Furthermore, the addition of higher concentrations of 125 I-ASA-BZ (50 -200 nM) to cytosolic extracts (100 g) showed saturable photoaffinity labeling of a 54-kDa protein (Fig. 2C). Taken together, these results show that the photolabeling of the 54-kDa protein with 125 I-ASA-BZ in cytosolic extracts from H. contortus is specific and saturable.
To confirm the identity of the 54-kDa ASA-BZ-photolabeled protein, 125 I-ASA-BZ-photolabeled cytosolic fractions were incubated with anti-tubulin monoclonal antibodies (mAbs), and the immunoprecipitated proteins were resolved on SDS-PAGE. The results in Fig. 3 show a 54-kDa protein immunoprecipitated from H. contortus with anti-tubulin mAbs (lane 3), but not with an irrelevant IgG 2a (lane 1). Similar results were also obtained when a cell extract from E. coli, which overexpress the H. contortus ␤12-16 tubulin gene, was photoaffinity-labeled with 125 I-ASA-BZ and immunoprecipitated with anti-tubulin mAbs or an irrelevant IgG 2a (Fig. 3, lane 4 or 2, respectively). These results confirm the identity of the 54-kDa protein as tubulin and provide the first direct evidence for BZ binding to tubulin.
The above results in Fig. 3 (lane 4) suggest that ␤-tubulin expressed in E. coli is photoaffinity-labeled with ASA-BZ; hence, BZ can interact with recombinant ␤-tubulin monomer(s). However, it is unclear whether ␣-tubulin is also photolabeled by 125 I-ASA-BZ. To determine if one or both tubulin subunits are photoaffinity-labeled with ASA-BZ, cytosolic extracts from H. contortus were photolabeled with 125 I-ASA-BZ, separated on gradient SDS-PAGE (4 to 14%), and transferred to nitrocellulose membrane for Western blot analysis with anti-␣and/or ␤-tubulin mAbs. The results in lane 3 of Fig. 4 show a single polypeptide detected with anti-␤-tubulin mAb. Lane 4 of Fig. 4 shows the two subunits of tubulin when the sample identical with lane 3 is probed with both anti-␣and ␤-tubulin mAbs. These results demonstrate that the photolabeled protein (Fig. 4, lanes 1 and 2) co-migrates with ␤-tubulin subunit (lane 3). In addition, the photoaffinity-labeled subunit in H. contortus (lane 1) co-migrated with the recombinant ␤-tubulin from E. coli extracts. Similar results were also obtained when purified mammalian brain tubulin was photoaffinity-labeled and fractionated on gradient SDS-PAGE to separate ␣and ␤-tubulin subunits (data not shown) except that the mammalian ␣migrates slower than the ␤-tubulin as reported previously (43).  2-6, respectively). Mebendazole, a potent BZ analogue, was very effective in inhibiting the photoaffinity labeling of tubulin with 125 I-ASA-BZ (lane 3). Thiabendazole, a BZ analogue that is structurally distinct from BZ carbamates, was virtually without effect (lane 6). Oxfendazole, a sulfoxide metabolite of fenbendazole, was less inhibitory to the photoaffinity labeling of tubulin by 125 I-ASA-BZ ( Fig. 5; lanes 4 versus  3). Colchicine, previously shown to interact at or near BZ binding site(s) in mammalian tubulin (9), was not a good inhibitor of ASA-BZ photoaffinity labeling of tubulin (lane 5). Taken together, these results suggest that the binding of ASA-BZ to tubulin occurs at a physiologically relevant BZ binding domain(s).

The Monomer ␤-Tubulin in E. coli Encodes a Similar BZ Binding Domain as Native Tubulin Dimer from H. contortus-
The results in Fig. 3 (lane 4) show that ␤12-16 tubulin monomer from cell extracts of E. coli is photoaffinity-labeled by To determine if 125 I-ASA-BZ interacts with similar domain(s) in native and recombinant ␤12-16 tubulin, photolabeled tubulins were subjected to Cleveland mapping using S. aureus V8 proteinase. Fig. 7 shows a V8 digest of recombinant ␤12-16 tubulin (lane 1) and that of native tubulin from H. contortus (lane 2). These results show a similar peptide map for native and recombinant ␤-tubulin with one major photoaffinity-labeled peptide migrating with an apparent molecular mass of ϳ3.5 kDa. In addition to the 3.5-kDa photoaffinity-labeled peptide, a minor photoaffinity-labeled peptide (ϳ5.3 kDa) was also detected in the above V8 map. However, the latter peptide may represent an incomplete digest of the 3.5-kDa peptide or a different photoaffinity-labeled site. The other major radioactivity signal in Fig. 7 co-migrates with the SDS-PAGE loading dye and may contain smaller peptides and free 125 I-ASA-BZ. Furthermore, partial cleavage of 125 I-ASA-BZ-photolabeled native and recombinant ␤-tubulin using increasing concentrations of V8 protease (0.001 g-10.000 g/gel slice) did not reveal differ-  (lanes 3 and 4). ences in the number or the electrophoretic mobility of the photolabeled peptides (data not shown). Taken together, these results suggest that 125 I-ASA-BZ binding to recombinant and native tubulin is similar. In addition, the identity of the 54-kDa photoaffinity-labeled protein as ␤-tubulin in H. contortus is further confirmed since identical peptide maps were obtained when recombinant ␤12-16 tubulin and the 54-kDa (unresolved ␣and ␤-tubulin) photoaffinity-labeled proteins were digested with V8 proteinase.  Fig. 8A show the separation of ␤-tubulin tryptic peptides by hplc using reverse-phase chromatography. Analysis of the above fractions for radioactivity indicated the presence of two radioactive peaks in fractions 45 and 54, respectively (Fig. 8A). Both fractions were collected and further purified using a shallower gradient of acetonitrile. Fig. 8B shows hplc tracing and radioactivity profile for 125 I-ASA-BZphotolabeled peptides in fractions 45 and 54 following several runs of purification (panels a and b or c and d, respectively). The amino acid sequence of the resultant tryptic peptides was determined by Edman degradation (see "Experimental Procedures"). The sequence of the tryptic peptide in fraction 45 was shown to encode the following amino acids H 2 N-Ala-Val-Leu-Val-Asp-Leu-Glu-Pro-Gly-Thr-Met-Asp-Ser-Val-Arg-COOH. The latter sequence corresponded to residues 63-77 of H. contortus ␤12-16 tubulin. Fraction 54 was shown to contain a larger tryptic peptide of 26 residues long (78 -103; H 2 N-Ser-Gly-Pro-Phe-Gly-Ala-Leu-Phe-Arg-Pro-Asp-Asn-Phe-Val-Phe-Gly-Gln-Ser-Gly-Ala-Gly-Asn-Trp-Ala-Lys-COOH). Interestingly, the amino acid sequence of the second tryptic peptide mapped immediately following the N-terminal of the first tryptic peptide. In an effort to determine the amino acids that are specifically modified by 125 I-ASA-BZ, samples were removed following each cycle of the Edman degradation, and their radioactivity was quantitated. A significant amount of radioactivity was found to be associated with the third and the fourth cycle of the Edman degradation of peptides 45 and 54, respectively. Thus, the amino acids Leu 65 and Phe 80 represent the modification sites by ASA-BZ in peptides 45 and 54, respectively.

DISCUSSION
In this study we show that a photoactive analogue of BZ interacts directly and specifically with tubulin from H. contortus. The specificity of 125 I-ASA-BZ toward tubulin was confirmed by the inhibition of photoaffinity labeling in the presence of BZ analogues. These results are consistent with earlier predictions that the high affinity BZ binding to nematode homogenate is due to tubulin (6 -8, 14, 50). Furthermore, the observations that thiabendazole, which lacks the methyl carbamate and the sulfoxide BZ analogue, oxfendazole, competes poorly for BZ binding to tubulin agree with previous drug binding studies (6 -8, 49). Colchicine, which is structurally unrelated to BZ, was less inhibitory to the photolabeling of  B (panels a and b or c and d) shows the absorbance at 214 nm and the radiolabel within the resolved tryptic peptides in fractions 45 and 54. tubulin with 125 I-ASA-BZ. An earlier study (6) has shown that colchicine reduces the binding of BZ to both mammalian tubulin and nematode homogenate. Moreover, BZ binding to fungal extracts was competitively inhibited by colchicine and oncodazole and not by other unrelated anti-microtubule agents (51). However, binding of colchicine to tubulin is characteristically different from that of BZ compounds, and the selectivity of the latter for lower eukaryotic protein has not been demonstrated for colchicine. Several studies (6,51) have suggested that (a) both BZ and colchicine bind to the same site or (b) the binding of colchicine induces conformational changes in tubulin that are unfavorable to BZ binding.
The development of BZ resistance in nematodes and other BZ-sensitive organisms has often been associated with changes in the genes encoding ␤-tubulin. Our data provide direct evidence that indeed ␤-tubulin is the acceptor protein for BZ in parasitic nematodes. This conclusion is further bolstered by the observation, in this study, that recombinant ␤12-16 tubulin monomers are specifically photoaffinity-labeled by ASA-BZ. Thus, ␤-tubulin alone appears to bind BZ while the ␣-subunits are less essential for BZ binding to microtubule. This finding is not exclusive to BZ, as photoactive analogues of several antimitotic drugs (taxol, colchicine, and rhizoxin) that bind to and inhibit the functions of microtubule have been shown to bind ␤and not ␣-tubulin (33)(34)(35). This is in contrast with other antitubulin drugs (e.g. vinblastine) that photoaffinity label both ␣and ␤-tubulin (32). The significance of the observed difference in antimitotic drug interactions with tubulin subunits is currently not clear.
The cleavage of 125 I-ASA-BZ-photolabeled ␤-tubulin with trypsin yielded two labeled peptides. N-terminal sequencing of 125 I-ASA-BZ-photolabeled tryptic peptides has localized the BZ binding domain to a span of 36 amino acids (Ala 63 to Lys 103 ) in ␤-tubulin. The assignment ASA-BZ binding to this region of ␤-tubulin is consistent with our V8 protease mapping results which showed one major 125 I-ASA-BZ-photolabeled peptide of ϳ3.5 kDa on SDS-PAGE. Analysis of ␤-tubulin amino acid sequence for all possible V8 cleavage sites revealed a 34-amino acid peptide (Ser 75 to Glu 108 ) with a calculated molecular mass of ϳ3.7 kDa. This peptide would contain a few amino acids from the first tryptic photolabeled peptide and the complete sequence of the second. Other V8 peptides that contain the 125 I-ASA-BZ-photolabeled amino acid (e.g. Leu 65 ) from the first tryptic peptide in fraction 45 would be too small to detect on SDS-PAGE and would migrate with the dye front.
The photolabeling of two tryptic peptides by 125 I-ASA-BZ was of the same intensity as determined from the radiolabel associated with each peptide. Moreover, comparison of the amino acid sequences of the two 125 I-ASA-BZ-photolabeled peptides showed no apparent sequence identity to support the possibility of two similar binding domains. Thus, the photoaffinity labeling of ␤-tubulin at two sites is likely due to rotational freedom about the ASA moiety that allows the photoreactive group in ASA-BZ to cross-link more than one sequence in native protein. In this respect, it was shown recently that two different sequences in ␤-tubulin are photolabeled by different photoactive analogues of taxol (p-azidobenzoyl-or m-azidobenzoyltaxol; Refs. 34 and 52). The photolabeling of the two domains (i.e. amino acids 1-31 and 217-231) in ␤-tubulin by m-azidobenzoyl-and p-azidobenzoyltaxol is thought to be due to differences in the position of the photoreactive groups (52). Although further characterization of the taxol binding domain is required, in general, the photolabeling of several distant sites in a protein is compatible with the three-dimensional nature of a drug binding site. Consequently, it is conceivable that other photoactive analogues of BZ could cross-link differ-ent sites in ␤-tubulin. Future studies using molecular and structural approaches to define the BZ binding domain are required to determine if the above photolabeled peptides are part of the BZ binding site in ␤-tubulin.
The binding of BZ to the N-terminal of ␤-tubulin brings to three the number of anti-microtubule agents whose binding maps to this region (33,35). However, both taxol and colchicine interact with two regions of ␤-tubulin (i.e. amino acids 1-46 and 214 -247). These two domains although far removed from each other in the primary sequence, are thought to come together in the folded protein to form the drug binding site (33). It is interesting that although taxol stabilizes while colchicine depolymerizes microtubule, these two drugs appear to interact with common domains (i.e. amino acids 1-46 and 214 -247). Thus, the location of a BZ binding site to N-terminal quarter suggests that this region may be critical for the assembly of microtubule. Accordingly, the efficacy of major classes of antimicrotubule drugs may be dependent on their ability to interact with this region. It is important to note that GTP, required for microtubule assembly, interacts with tubulin at the N-terminal half and the requirement of GTP for BZ binding has been suggested although not clearly shown (6). One of the GTP binding sites (amino acids 63-69) is contained within the 125 I-ASA-BZ photoaffinity-labeled domains, and this may explain the GTP requirement for BZ binding to tubulin. The binding of taxol is known to reduce the need for GTP during microtubule assembly. The poor ability of colchicine to inhibit photoaffinity labeling of tubulin by 125 I-ASA-BZ could be explained by the fact that these two drugs bind adjacently but not the same domains.
Genetic analyses of BZ resistance in H. contortus have identified three possible amino acid substitutions, Phe 76 3 Val, Phe 200 3 Tyr, and Ile 368 3 Val, that could lead to resistance (53,54). Similarly, four amino acid substitutions at His 6 , Val 165 , Glu 198 , and Phe 200 in benA ␤-tubulin gene of Aspergillus nidulans are thought to confer resistance to BZ (25,27). The change at position 200 was strongly favored as the one most likely candidate for causing BZ resistance since tyrosine is observed in other BZ-resistant organisms and in mammalian tubulin. Our study, however, identified amino acids 63-103 as the BZ binding domain that spans one of the mutations implicated in drug resistance (i.e. Phe 76 3 Val). Further biochemical evidence will be required to confirm the effects of these three amino acid changes in BZ-tubulin interactions. As amino acid changes that may lead to resistance need not be found on the BZ binding domain or photolabeled peptides in the linear sequence, it is conceivable that any of these mutations could be important in conferring BZ resistance. We 2 show that Phe 200 3 Tyr substitution leads to reduced photoaffinity labeling of ␤12-16 tubulin by 125 I-ASA-BZ. However, since several ␤-tubulin genes exist in parasitic nematodes, drug resistance in vivo may involve a combination of several isoforms.
BZ show a remarkable safety when used as anthelmintics in the treatment of many veterinary and human helminthiases. This is surprising since BZ also inhibit mammalian microtubule formation in vitro (6,55). In fact, the efficacy of different BZ analogues correlates well with their microtubule inhibitory potencies (25). Thus, the molecular basis for BZ selectivity is unknown. However, several factors may contribute to the selective toxicity of BZ. The binding of BZ to parasite tubulin is stable to charcoal adsorption and is stronger than that of mammalian tubulin that is readily removed by charcoal adsorption (7,13,14). A comparison of the amino acid sequence of ASA-BZ-photolabeled peptides (N-terminal 63-103) of H. contortus to bovine or human ␤-tubulin show two different amino acid residues (Ala 85 3 Gln and Leu 86 3 Ile) that could confer weaker binding of BZ to mammalian tubulin. Alternatively, the observed safety of BZ as anthelmintics may be unrelated to BZ-tubulin binding but due to differences in metabolism or detoxification pathways. For example, the rapid and extensive metabolism of BZ into less toxic metabolite (e.g. sulfoxides and sulfones) by the liver microsomal enzymes (56, 57) may account for some lack of host toxicity. Parasites, on the other hand, lack these metabolic pathways and are killed by BZ. In addition, we have shown recently that BZ are substrates for the P-glycoprotein transporter in multidrug-resistant tumor cells (5). It may be speculated that P-glycoprotein which is overexpressed in several normal tissues and organs (58,59) could mediate the transport of BZ. Thus, P-glycoprotein rather than differences in tubulin amino acid sequence may mediate the observed safety of BZ to the host. The latter speculation is interesting since ivermectin, a potent anthelmintic agent, is also remarkably safe to the host (60), and its accumulation in normal tissues is affected by the presence of P-glycoprotein (61). The latter results were elegantly demonstrated using homologous recombination to inactivate the class I P-glycoprotein gene in mice (61). Consequently, P-glycoprotein-deficient mice showed a dramatic increase in ivermectin accumulation and toxicity in comparison to mice with normal P-glycoprotein expression (61).