RT-PCR generally works on a “one gene in one experiment” basis, which means that the “whole picture” of differential gene expression between cells or tissues is difficult to obtain. In the past several years, a new technology, called DNA microarrays, has allowed scientists to monitor changes in the transcription of thousands of genes simultaneously (Wan et al. 1996). Microarrays combine the technology of RT-PCR with high-speed robotics. First, one takes mRNA from a tissue and, using reverse transcriptase, converts the mRNAs into their complementary DNAs. Each individual cDNA is then cloned, denatured, and amplified by the polymerase chain reaction Each of the resulting cDNA clones serves as a probe and is robotically printed onto glass slides in a particular order. The slides are subsequently hybridized to two “targets” with different fluorescent labels. These targets are pools of cDNAs that have been generated after isolating mRNA from cells or tissues in two states that one wishes to compare. For instance, if the aim is to compare cell type A with cell type B, one takes mRNAs from both cell types, converts the mRNAs into cDNAs, and labels the cDNAs from cell type A with fluorescein (green) and the cDNAs from cell type B with rhodamine (red) (Figure 1). The two DNA pools would then be equally mixed, and the mixture placed onto each spot of probe DNA. The resulting fluorescent intensities are produced using a laser confocal fluorescent microscope, and ratio information is obtained following image processing. By comparing the fluorescent intensities, one can tell if a particular cDNA (and hence, the mRNA of interest) is present in higher amounts in one cell type or another.
For instance, one can look at the genes active in the prospective dorsal part of a frog blastula and compare them with the genes active in the future ventral portion of the same blastula (Altmann et al. 2001). Alternatively, one can look at entire animals at different stages of their development to see which genes are active at each stage (White et al. 1999). This allows one to focus one’s research on those genes whose expression differs between the two sets of cDNAs.
One can use microarrays even without knowing what gene to look for. The difference in gene expression will indicate which genes are important in the particular tissue. For instance, in Chapter 14 we will discuss the different gene expression in the cells of male (XY) and female (XX) mouse brains. Researchers did not have to guess which genes would be expressed differently—the microarrays were enough to tell them which genes were differentially active in XX and XY brains, with the results then being confirmed by RT-PCR.
A less expensive modification of the microarray is the macroarray. The technology is similar, but the spots are larger (about 1 mm as compared to 250μ). This means that macroarrays can be visually interpreted without a microscope, and that radioisotopes as well as fluorescent labels can be used. For instance, in Chapter 7, we saw that Nodal genes are important in specifying the different regions of mesoderm in Xenopus and zebrafish embryos. One of the genes in zebrafish is Xenopus Nodal related-1 (Xnr1). To see if Xnr1 is activating particular genes, Naoto Ueno’s laboratory injected Xnr1 mRNA into some early embryos but not others. They then harvested mRNAs from the animal caps of the Xnr1-secreting and control embryos. (The animal cap would not be exposed to Xnr1 under normal development.) After converting the mRNAs into radioactively labeled cDNAs, Ueno and colleagues hybridized these probes to slides containing bound DNA. One macroarray was hybridized with radioactive cDNAs from the control animal caps; an identical macroarray was hybridized with radioactive cDNAs from the Xnr1-treated caps. One of their results is shown in Figure 2, in which the gene for chordin is shown to be expressed in the treated but not in the untreated animal cap.