We have studied the pattern of histone acetylation in intact rat hepatoma tissue culture (HTC) cells, in isolated HTC nuclei, and in chromatin prepared from these cells. The results have been compared with the histone acetylation observed in a reconstituted in vitro system consisting of a variety of purified soluble nucleosomal substrates, [3H]acetyl-CoA, and one of two different purified histone N-acetyltransferases. Acetylase A, a highly purified nuclear enzyme, catalyzed the acetylation of 1) nucleosomally bound histones in the order H4 > H2a = H2b > H3, and 2) free histones in the order H4 > H3 > H2b > H2a. Acetylase B, a cytoplasmic enzyme, modified only free histone H4, and it failed to acetylate histones in nucleosomes. The pattern of histone acetylation obtained by in vitro reaction of purified nucleosomes with the purified nuclear acetylase A differed considerably from the corresponding patterns obtained either by acetate labeling of intact cells, or by the acetyl-CoA labeling of nuclei and crude preparations of nucleosomes, as catalyzed by endogenous chromatin-bound acetylase(s). The most striking difference was in the relative preference for acetylation of histone H4 versus acetylation of histone H3: with the purified acetylase, histone H4 in nucleosomes was acetylated to a much greater extent than was histone H3, whereas the reverse preference was found with the endogenous acetylase(s). This result suggests that either a second nuclear acetylase enzyme, or a separate cofactor for acetylase A, is required for histone H3 acetylation in vivo. In support of this view, we find that the acetylation of histones H4, H2a, and H2b in nuclei is inhibited by urea, salt, or N-ethylmaleimide treatments to a very different extent than is the acetylation of histone H3. By comparing n-butyrate-treated HTC cells with untreated cells, classes of nucleosomes specially accessible and inaccessible to acetylation can be distinguished (Cousens, L. S., Gallwitz, D., and Alberts, B. M. (1979) J. Biol. Chem. 254, 1716-1723). Both types of special nucleosomal reactivities were present in isolated nuclei, but were lost as nucleosomes were purified from these cells. OUr data thus suggest the existence of labile specificity factors or structures, which guide the acetylase(s) to restricted groups of otherwise similar nucleosomes in vivo.