A truly detailed map of gene expression patterns can be obtained using in situ hybridization. Here, a labeled antisense mRNA probe (a DNA or RNA sequence that is complementary to the sequence of a specific mRNA) is hybridized with the mRNA in the organ itself. Antisense mRNA is made from a cloned gene in which the gene is reversed with respect to a promoter within the vector. The mRNA transcribed from such a gene encodes a sequence complementary to the normal (“sense”) mRNA made by that gene. Such antisense mRNA can be used as a probe, since it will recognize the “sense” mRNA in the cell.

For in situ hybridization, the antisense RNA is labeled either by being made radioactive or by being attached to a dye, allowing the probe to be visualized. Thus, one makes a sequence-specific stain that will label only those cells that have accumulated mRNAs of a particular sequence. When radioactive probes are used, embryos or organs are first fixed to preserve their structure and to prevent their mRNA from being degraded. The fixed tissues are then sectioned for microscopy and placed on a slide. When the labeled sequence is added, it binds only where the target mRNA (to which it is complementary) is present. After any unbound probe is washed off, the slide is covered with a transparent photographic emulsion for autoradiography. By using dark-field microscopy (or computer-mediated bright-field imaging), the reduced silver grains can be shown in a color that contrasts with the background stain. Thus, we can visualize those cells (or even regions within cells) that have accumulated a specific type of mRNA.

In whole-mount in situ hybridization, the entire embryo (or a part thereof) can be stained for certain mRNAs. This technique, which uses dyes rather than radioactivity, allows researchers to look at entire embryos (or their organs) without sectioning them, thereby observing large regions of gene expression. Figure 1 shows an in situ hybridization performed on a whole chick embryo that was fixed without being sectioned. (The embryo was also permeabilized by lipid and protein solvents so that the probe could get in and out of its cells.) The probe used in this experiment recognized the mRNA encoding Pax6 in the chick embryo.

The probe shown in Figure 1 was labeled not with a radioactive isotope, but with a modified nucleotide. To create this probe, a region of the cloned Pax6 gene was transcribed into mRNA, but with an important modification: in addition to the standard uridine triphosphate (UTP), the nucleotide mix also contained UTP conjugated with digoxigenin. Digoxigenin—a compound made by particular groups of plants and not found in animal cells—does not interfere with the coding properties of the resulting mRNA, but does make it recognizably different from any other RNA in the cell.

The digoxigenin-labeled probe was incubated with the embryo. After several hours, numerous washes removed any probe that had not bound to the embryo. Then the embryo was incubated in a solution containing an antibody against digoxigenin. The only places where digoxigenin should exist is where the probe bound (i.e., where it recognized its mRNA), so the antibody sticks in those places. This antibody, however, is not in its natural state. It has been conjugated covalently to an enzyme, such as alkaline phosphatase. After repeated washes to remove all the unbound enzyme-conjugated antibody, the embryo was incubated in another solution that was converted into a dye by the enzyme. The enzyme should be present only where the digoxigenin is present, and the digoxigenin should be present only where the specific complementary mRNA is found.

Figure 1   Schematic of whole-mount in situ hybridization localizing Pax6 mRNA in early chick embryos. A digoxigenin-labeled antisense probe hybridizes to a specific mRNA. Alkaline phosphatase-conjugated antibodies to digoxigenin recognize the digoxigenin-labeled probe. The enzyme is able to convert a colorless compound into a dark purple precipitate. (After Li et al. 1994.)