Previous attempts to correlate in vivo pyridoxine-responsiveness with in vitro assays of cystathionine beta-synthase activity in synthase-deficient homocystinuric patients have been only partially successful. All such studies, however, have been conducted with extracts of cultured skin fibroblasts grown in medium containing a high concentration (1,000 ng/ml) of pyridoxal. Having recently shown that such growth conditions may obscure important aspects of enzyme-coenzyme interactions by saturating most synthase molecules with their cofactor, pyridoxal 5'-phosphate, we have established conditions for growth of cells in pyridoxal-free medium. Under these conditions, intracellular pyridoxal 5'-phosphate fell by >95%, and saturation of cystathionine beta-synthase apoenzyme with pyridoxal 5'-phosphate decreased from a predepletion value of 70% to <10%. When such depleted cells were grown in media containing pyridoxal concentrations ranging from 0 to 1,000 ng/ml, cellular pyridoxal 5'-phosphate reached a maximum of 30 ng/mg cell protein at a medium pyridoxal concentration of 100 ng/ml. Maximal saturation of aposynthase with coenzyme in control cells was reached at a medium pyridoxal concentration of 10 ng/ml. In contrast, maximal saturation of residual aposynthase in cells from an in vivo responsive patient was achieved at a medium pyridoxal concentration of 25-50 ng/ml, whereas that from cells from an in vivo unresponsive patient was reached at 100 ng/ml. Estimates of the affinity of control and mutant cystathionine beta-synthase for pyridoxal 5'-phosphate in cell extracts supported the differences observed in intact cells. The apparent K(m) of cystathionine beta-synthase for pyridoxal 5'-phosphate in extracts of depleted cells from four in vivo-responsive patients was two to four times that of control. In contrast, the K(m) for pyridoxal 5'-phosphate in two lines from in vivo nonresponsive patients was 16- and 63-fold normal. These results suggest that cystathionine beta-synthase activity in cells from patients containing a mutant enzyme with a moderately reduced affinity for pyridoxal 5'-phosphate can be increased by pyridoxine supplements in vivo, whereas that from patients whose enzyme has a more dramatically reduced affinity for the coenzyme cannot be so modulated because of limits on the capacity of such cells to accumulate and retain pyridoxal 5'-phosphate.