Two important structures in the membranes of many cells are aquaporins on cell surfaces and gap junctions between tightly apposed cells in multicellular systems. As shown in Table 1, neither is present at any stage between the MII oocytes and four- to six-cell embryos. The early eight-cell embryo also shows no evidence of the presence of gap junctions or AQP 3; however, its membranes may possibly contain AQP 9. At this stage, the surfaces of the eight blastomeres are stacked on each other but not structurally joined. Later in the eight-cell stage, the membranes that form the boundary between two blastomeres now form tight junctions, and gap junctions appear. Simultaneously, AQP 3 now becomes detectable in the membranes facing the outside medium.
Gap junctions almost universally form across the dual bilayer constituting the tight junctions between cells. Most gap junctions are formed by a family of connexin proteins, and in the center of these junctions is a transmembrane pore that varies from 4 to 12 A in radius depending on the connexin that forms them. According to Saez et al., the pores are hourglass in shape. The outer surfaces have radii of 12-20 A that narrow to 7.5 A in the center. Reading here
Aquaporins were discovered some 17 yr ago. They too are transmembrane protein complexes that form pores in many animal and plant cells (although by no means all) in the portions of plasma membranes that are exposed to the external medium. These pores are also hourglass shaped. In AQP 1, the diameter tapers from 15 A at the extracellular side to 2.8 A at the constriction, a width that can just accommodate a single water molecule. If IIF in a blastomere facing the medium is caused by contact with external ice, it is difficult to see how anything in the nature of an ice crystal could pass through the 2.8-A neck unless some force widened the constriction. It is conceivable that external ice in its attempt to grow into the supercooled water inside the cells could provide such a force.
But when ice forms in one blastomere of a morula, the much wider pore in the gap junction might serve as the route by which the ice in this first blastomere nucleates the supercooled water in adjoining blastomeres. This appears to be the case on both theoretical and experimental grounds. The value of 0°C for the melting point of pure ice is the value for a planar ice crystal, i.e., one with an infinite radius of curvature. However, if the ice crystal is convex, its melting point, on thermodynamic grounds, will be suppressed. This is equivalent to saying that the chemical potential of water at the surface of a convex crystal is higher than that at the surface of a planar crystal at the same temperature.