Using a single nucleotide polymorphism association study in 52 men with prostate cancer receiving docetaxel, we found that individuals carrying two copies of the CYP1B1*3 polymorphic variant had a poor prognosis after docetaxel-based therapies compared with individuals carrying at least one copy of the CYP1B1*1 allele (30.6 versus 12.8 months; P = 0.0004). The association between CYP1B1*3 and response to therapy was not observed in similar subjects receiving non-taxane-based therapy (P = 0.18). The systemic clearance of docetaxel was also unrelated to CYP1B1 genotype status (P = 0.39), indicating that the association of CYP1B1*3 with clinical response is not due to docetaxel metabolism. To explain these results, we hypothesized that an indirect gene-drug interaction was interfering with the primary mechanism of action of docetaxel, tubulin polymerization. We therefore conducted tubulin polymerization experiments with taxanes in the presence or absence of certain CYP1B1 estrogen metabolites, which are known to bind to nucleophilic sites in proteins and DNA, that revealed the primary estrogen metabolite of CYP1B1, 4-hydroxyestradiol (4-OHE2), when oxidized to estradiol-3,4-quinone strongly inhibits tubulin polymerization. The 4-OHE2 is also formed more readily by the protein encoded by the CYP1B1*3 allele, validating further our data in patients. Furthermore, estradiol-3,4-quinone reacted in vitro with docetaxel to form the 4-OHE2-docetaxel adduct. This pilot study provides evidence that CYP1B1*3 may be an important marker for estimating docetaxel efficacy in patients with prostate cancer. This link is likely associated with CYP1B1*3 genotype-dependent estrogen metabolism.