It has long been assumed that Ca2+ are translocated from the cytosol to the cell nucleus by a long distance to activate transcription machinery buried deep in the nucleoplasm. However, this model has been recently challenged. When HeLa cells were loaded with fluo-3, highly fluorescent spots of approximately 2 microns in diameter were observed in the cell nucleus while the fluo-3 signals were low in their neighbouring nucleoplasm as determined by confocal microscopy. These fluorescent spots were devoid of but usually associated with chromatin on their boundary. When cells were stimulated by ionomycin (1 microM), the fluo-3 fluorescence in these spots increased faster than that in their neighbouring nucleoplasm. In another experiment, optical sections with hot spot(s) were used to construct 3-D images to study the morphology of the hot spots. Views of reconstruction from different angles indicated that the hot spots formed a tubular structure with a connection to the nucleocytoplasmic interface. Moreover, injection of calcium green-dextran (70 kDa), a Ca(2+)-sensitive indicator conjugated with an inert molecule of large molecular size, into the cytosol leads to a formation of signals also in a tubular shape inside the nucleoplasm. This suggests that the 'channels' are real inside the nucleus and they are derived from an invagination of the double-membraned nuclear envelope. Taken together, our results indicate (1) tubular structures are found inside the cell nucleus; (2) they are extended from the cytosol into the nucleus through the invagination of the double membraned nuclear envelope; (3) molecules of molecular size up to 70 kDa could penetrate into these 'tunnels'; (4) Ca2+ can be released or transported into the cell nucleus through these tubular structures after ionomycin stimulation; and (5) the structures are usually associated with chromatin.