Previously it was shown that when condensed chromatin from several different types of cell is stained with uranyl-lead and examined in thin sections in the electron microscope, the stain is distributed into a dot-dash pattern arising from threads, with lesser-staining intermediate areas. We now show that when a section through chicken erythrocyte chromatin is stained with ethanolic phosphotungstic acid (PTA) the stain distribution is homogeneous. This shows that the lesser-staining regions after uranyl-lead, cannot be an overlap artifact. We conclude that the stains and hence the molecules in chromatin are distributed between 2 phases, an o- and an e-phase, so called because the structural units in chromatin are arranged in an orderly way at the surface of the nucleus and give rise to oddly (o) and evenly (e) numbered bands. Measurements of electron density per unit thickness, proportional to the number of stain molecules per unit volume, are made in thin sections through erythrocytes and reticulocytes from adult hen, 4-day-old chicks and 17-day embryos. The results indicate differences in the packing of the molecules in chromatin and further show that the e-phase is quite likely to have a higher DNA to protein ratio than the o-phase. After uranyl-lead stain the visibility of the dot-dash pattern in cells from adult hen is relatively low due, we propose, to closer packing. In micrographs through condensed chromatin treated with uranyl-lead the eye selects out only the densely stained dots and dashes, width 17 nm. When erythrocyte chromatin is partially or completely disrupted in various ways, threads 25-30 nm then become visible. We propose that condensed chromatin in intact cells contains structural units which consist of a central element, width 17 nm previously referred to as the unit thread, forming the e-phase, surrounded by a cylindrical shell forming the o-phase. This socalled superunit thread is similar in width, about 25-30 nm, to that reported by other workers in preparations of chromosomes spread on water surfaces. The hypothesis therefore helps explain what appeared to be discrepancies in thread dimensions. Certain other ultrastructural features of erythrocyte nuclei are also reported which are either pertinent to the general aim of this study, namely the way in which nucleoproteins fold up in chromosomes, or to biochemical studies, to be reported shortly, in which attempts are made to locate the proteins removed from isolated erythrocyte nuclei during subsequent washing in salt solutions.
The mechanism of thrombin-induced platelet factor 4 secretion
