Enzymatic conversion of all-trans-beta-carotene to retinal by a partially purified enzyme from rabbit, rat, and human neonatal intestinal mucosa has been demonstrated. The enzymatic product was characterized based on the following evidence. First, the product gave rise to its O-ethyl oxime by treatment with O-ethylhydroxylamine with an absorption maximum at 363 nm in ethanol characteristic of authentic retinal (O-ethyl) oxime. High-performance liquid chromatography of this derivative yielded a sharp peak with a retention time of 7.99 min, corresponding to the authentic compound. The enzyme blank and boiled enzyme blank failed to show any significant HPLC peaks corresponding to retinal (O-ethyl) oxime or retinal or retinol. Second, the mass spectrum of the O-ethyl oxime of the enzymatic product was identical to that of authentic retinal (O-ethyl) oxime (m/z 327, 45%; m+ and m/z 282, 100%, methoxy). Third, the 14C radioactivity persisted to constant specific activity even after repeated crystallization of the retinal (O-ethyl) oxime isolated from the enzyme reaction with purified beta-[14C]carotene. Fourth, the enzymatic product exhibited an absorption maximum at 370 nm in light petroleum characteristic of authentic retinal. Furthermore, it was reduced by horse liver alcohol dehydrogenase to retinol with an absorption maximum at 326 nm in light petroleum. This retinol was enzymatically esterified to retinyl palmitate by rat pancreatic esterase with a retention time of 10 min on HPLC, corresponding to authentic retinyl palmitate. Thus, the enzymatic product of beta-carotene cleavage by the partially purified intestinal enzyme has been unequivocally confirmed to be retinal. Similarly, enzymatic conversion of all-trans-beta-carotene to retinal by an intestinal mucosal enzyme from autopsy samples of human neonates has also been demonstrated. Based on the observed activities among intestinal samples from 12 premature infants, the BCC enzyme activity ranged from 3.3 to 1210 pmol/mg mucosal protein/hr. However, the observed activities in the human autopsy samples may be markedly underestimated, presumably because of marked loss of enzyme activity from the time of death to the time of assay. Therefore, the true activity of the enzyme can be assessed only after the extent of the loss of its activity on storage of the human samples can be accurately measured. Nonetheless, the demonstration of BCC enzyme activity in human neonates shows that beta-carotene may be an important source of vitamin A nutrition during gestation.