Using two-dimensional polyacrylamide gels stained with Coomassie blue we have studied the protein composition of the nuclear matrix obtained from mouse erythroleukemic nuclei kept at 0 degrees C throughout the isolation procedure to prepare the high ionic strength resistant fraction (control matrix) or stabilized in vitro or in vivo by different procedures prior to subfractionation (ie 37 degrees C incubation of isolated nuclei; sodium tetrathionate exposure of purified nuclei; heat shock of intact cells). When the matrix obtained from 37 degrees C incubated nuclei was compared with the control matrix, striking differences in the polypeptide pattern were seen if the protein was obtained in both cases from an equivalent number of nuclei. On the other hand, if the same amount of protein for both the samples was applied to the gels the differences were less evident. Sodium tetrathionate stabilization of isolated nuclei and heat shock of intact cells produced a matrix protein pattern that was very similar and differed from that of the in vitro heat-exposed matrix. Using specific polyclonal antisera, we demonstrate that nucleolar proteins B23/numatrin and C23/nucleolin were very abundant in the matrix obtained from chemically-treated nuclei or in vivo heat-stabilized nuclei but were recovered in very small amounts (B23) or completely absent (C23) in the matrix prepared from nuclei heated to 37 degrees C in vitro. Differences were seen also in the recovery of nuclear lamins, and especially lamin B, that was poorly represented in the sodium tetrathionate-stabilized matrix. The results demonstrate that in mouse erythroleukemia cells the increased recovery of nuclear matrix protein that is seen after in vitro heating of isolated nuclei is predominantly due to an additional recovery of the same types of polypeptides that are detected also in the absence of such a treatment. The data also indicate that in vivo heat shock of intact cells produces a nuclear matrix protein pattern that is more similar to the pattern seen after stabilization of purified nuclei with sodium tetrathionate and differs significantly from that obtained by exposing nuclei to 37 degrees C in vitro, unlike to that what previous reports have indicated.