Optimal conditions are described for measurement of cystathionase activity in long term lymphoid cell lines. In 21 control lines established from normal subjects, cystathionase (EC 4.4.1.1) specific activity was 25.8 ± 1.7 (mean ± SE) nmole cysteine/hr/mg protein. Extracts of three lymphoid lines (NB-36, NB-95, and NB-77) established from three vitamin B6-responsive patients with primary cystathioninuria had activity of 3.6–7.3; from a B6-unresponsive patient (NB-68) had no detectable activity; from five obligate heterozygotes for cystathioninuria had activity of 11.2–18.9.Two, or possibly three, different modifications of the cystathionase molecule could be demonstrated in cultured cells from the patients with primary cystathioninuria, based on the effects of added pyridoxal phosphate (PLP) in the enzyme assay, on the extent of reaction with rabbit antihuman hepatic cystathionase, and on the ability to compete with normal lymphoid cell line enzyme extract for the antibody combining sites. When PLP is not added to the assay system, the normal enzyme extract still had 89% of its activity in 1.0 mM PLP; on agar double diffusion analysis it gave a band of identity with normal human hepatic cystathionase; the precipitin band had cystathionase activity, but inhibition by antibody could be shown in solution. Lymphoid line extract from the B6-unresponsive patient had no detectable activity in the absence or presence of PLP, did not form a precipitin band, and did not compete with normal enzyme extract for the antibody combining sites. Thus, synthesis of apocystathionase is absent or significantly reduced; alternatively, the protein produced has lost its antigenic determinants