When the retinal Schiff base in chymotryptically cleaved bacteriorhodopsin is reduced to a secondary retinylamine by prolonged exposure to 10% (wt/vol) sodium cyanoborohydride, at pH 10, in the absence of light, =45% of the retinal is found linked to , and the remainder is scattered over various sites on the large chymotryptic fragment, including the physiological site at Lys-216. The retinal-binding site is destroyed or blocked by the reduction conditions, but the bacteriorhodopsin lattice remains intact. The results demonstrate that artifactual linkage to Lys-40/41 is possible under special conditions. Under these conditions, the e-amino groups of Lys-40/41 show an enhanced ability to form retinylidene linkages with the retinal released by the physiological linkage site at Lys-216, due to some combination of close proximity to the normal linkage site, and increased reactivity with respect to other lysine Eamino groups. The results are of interest for the characterization of the two newly discovered rhodopsin-like proteins, halorhodopsin and slow rhodopsin.The determination of the primary protein (1, 2) and structural gene (3) sequences of bacteriorhodopsin (bR), the lightdriven proton pump from the purple membrane (pm) of Halobacterium halobium, has led to a reinvestigation of the retinal-binding site by chemical reduction and peptide analysis. This resulted in a reassignment for the binding site from the E-amino group of to the E-amino group of Lys-216 (5-9). Additional chemical modification (10) and resonance Raman (11) studies have demonstrated conclusively that the binding site does not change during the photoreaction cycle as originally postulated by Ovchinnikov's group (12). The reason for the original misassignment of the binding site by Bridgen and Walker is obvious in retrospect (7). However, the same arguments cannot explain the more recent observations of reductive retinal linkage to Lys-40/41 (7,9,12). Lemke and Oesterhelt (6) consider these to be in error due to a previously unnoticed proteolytic cleavage of bR by NaBH4. We show here that this explanation does not hold, at least for our results, and that under some conditions of reduction, two-thirds of the retinal are indeed found bound to Lys-40/41.More important, we have previously argued that while binding to Lys-40/41 is a preparation artifact, it might still be used to obtain important information about the tertiary structure of bR. We have, therefore, repeated and extended our earlier experiments. Unfortunately, the new observations considerably weaken the structural argument, but they demonstrate an unusually high reactivity of Lys-40/41 for Schiff base formation and/or reduction and are of considerable interest for the interpretation of experiments trying to identify the two additional retinal pigments recently discovered in H. halobium (13)(14)(15) Repenerated, Chymotryptically Cut, [3H]Retinal-Labeled pm ( H-RG-CT-pm). RG-CT-pm was prepared essentially as described (7, 17), except that hydroxylamine-bleached pm was cut at a c...