Nuclei from bovine thymus contain a high level of partially latent protein phosphatase 1 (PP-1). More than 90% of this PP-1 is associated with the insoluble chromatin/matrix fraction and can be extracted with 0.3 M NaCl. The salt extract also contains three heat- and acid-stable inhibitory proteins of PP-1 that can be resolved on Mono Q. We have purified two of these nuclear inhibitors of PP-1 (NIPP-1a and NIPP-1b) until homogeneity. They are acidic proteins (pI = 4.4) with a molecular mass of 18 kDa (NIPP-1a) and 16 kDa (NIPP-1b) on SDS-PAGE. Judged from the larger molecular mass that was deduced from gel filtration (35 kDa), NIPP-1a and NIPP-1b appear to be asymmetric or dimeric proteins. The nuclear inhibitors totally inhibited the phosphorylase phosphatase activity of PP-1, but even at a 250-fold higher concentration they did not affect the activities of the other major serine/threonine protein phosphatases (PP-2A, PP-2B, and PP-2C). NIPP-1a and NIPP-1b inhibited the catalytic subunit of PP-1 with an extrapolated Ki of about 1 pM, which is some three orders of magnitude better than the cytoplasmic proteins inhibitor 1/DARPP-32 and modulator. The nuclear inhibitors were not inactivated by incubation with protein phosphatases that inactivate inhibitor 1 and DARPP-32. Unlike modulator, they were not able to convert the catalytic subunit of PP-1 into a MgATP-dependent form. Remarkably, the extent of inhibition of PP-1 by NIPP-1b depended on the nature of the substrate. The phosphorylase phosphatase and casein phosphatase activities of PP-1 were completely blocked by NIPP-1b, whereas the dephosphorylation of basic proteins was either not at all inhibited (histone IIA) or only partially (myelin basic protein). These data may indicate that the acidic NIPP-1b is inactivated through complexation by basic proteins. Indeed, nonphosphorylated histone IIA antagonized the inhibitory effect of NIPP-1b on the casein phosphatase activity of PP-1. Our data show that the nucleus contains specific and potent inhibitory proteins of PP-1 that differ from earlier described cytoplasmic inhibitors. We suggest that these novel proteins may control the activity of nuclear PP-1 on its natural substrate(s).