Igor Dawid and Thomas Sargent explain how they developed subtractive mRNA hybrization to find genes expressed by different cell types.

Hi, I'm Igor Dawid. Hi, I'm Igor Dawid and I'm Thomas Sargent. We were interested in "differentiation" genes. Even though every cell in an organism has the same set of DNA, there are many different cell types. We developed a technique called subtractive mRNA hybridization that allowed us to quickly find the genes expressed by different cell types. Let us show you how this works. We used frogs as the experimental system. Frogs have specific stages of development. We reasoned that by looking at the differences in the mRNA population between these stages, we would be able to find the genes expressed in the different cell types. For example, the blastula is a ball of undifferentiated cells, whereas the ectoderm, endoderm, and mesoderm cell layers develop in the gastrula stage. The genes that specify the differentiated cell types must turn on at the gastrula stage or just before. To test this idea, we isolated mRNA from the different stages. Then, using reverse transcriptase, we made DNA complements — cDNA — of the mRNA collected from the gastrula stage. CLICK TO ADD REVERSE TRANSCRIPTASE Since cDNA is a complement, it can hybridize to matching mRNA. After digesting away the mRNA template, we mixed the gastrula cDNA with blastula mRNA; the mRNA found in both stages will end up as a cDNA-mRNA hybrid pair. GASTRULA cDNA + BLASTULA mRNA We passed the mixture through a hydroxyapatite column which binds the double-stranded cDNA-mRNA hybrids. Thus, we effectively subtracted the mRNAs common to both stages, and were left with a population of cDNAs unique to the gastrula stage. HYDROXYAPATITE COLUMN We separated these cDNAs and then inserted them into plasmids. These recombinant plasmids made up a library of clones from which we were able to generate a steady supply of any one of the gastrula-specific cDNAs for further study. We can use the unique cDNAs as radioactive probes on what we called developmental dot blots. We spotted mRNAs collected from the different stages and times of development onto a special type of paper — nitrocellulose. Then we incubated the strips with the radioactive cDNAs, which hybridize to mRNA spotted onto the strips. When we used these strips to expose photographic film, a dark spot appeared where the radioactive cDNA had hybridized. As you can see, when we lined the strips up, we have a visual record of the stages and times of development of these genes. Using the radioactive cDNAs, we can also do the same type of hybridization with a tissue sample. This techinque, in situ hybridization, (in situ is Latin for "in state") lets us see where mRNA is expressed within an organism. This is an in situ of cpg-15 mRNA expression in the developing frog eye. cpg-15 is a gene needed for neural circuitry development, and is expressed in retinal ganglion cells (RGC). RGCs project neural processes to the brain. A radioactive probe that hybridizes to cpg-15 mRNA shows the zone of expression (silver dots). Hi, I'm Pat Brown. I developed a technique where cDNAs can be embedded onto glass slides. Using these DNA arrays, we can do large-scale expression studies. Growth stage-specific mRNAs are isolated, and then reverse transcribed to give unique cDNA populations. These are directly embedded onto specially-coated glass slides. The slides are coated with poly-lysine, which is positively charged. DNA is negatively charged, so the cDNA "sticks" to the slide through an ionic interaction. The cDNA can still interact with a DNA probe. We used arrays to obtain gene expression profiles of cancer. Diffuse large B-cell lymphoma (DLBCL) is a common lymphoma — cancer of the lymph nodes. We found sub-types of DLBCL that correlated with survival rates. We made a chip with a genes expressed by the lymph nodes and those important in cancer biology. A total of 17,856 cDNA genes were printed onto what we called "lymphochip." Remember, each square on the chip corresponds to a different cDNA. Then, we made cDNAs from different DLBCL tumors. We labelled one set of cDNA with a red fluorescent tag; the other with a green tag. We incubated the arrays with the tagged cDNAs, which bound to the matching genes printed on the array. Since we knew the positions of the genes on the DNA array, we could figure out the levels of gene expression based on the color signal. If the gene was only expressed in DLBCL1 cells, the square was red. Similarly, if the gene was only expressed in DLBCL2 cells, the square was green. If the gene was expressed equally in both cells, the square was yellow. Thus, we identified two sub-types of DLBCL — GC B-like and Activated B-like DLBCL. These sub-types have different responses to therapy, and with this type of diagnosis, more tailored treatment can begin for patients. This type of tailored treatment is called "pharmacogenomics." We can also analyze the differences in gene expression for these two very similar lymphomas. This may give us a better understanding of how cancers work, and hopefully develop better therapies and cures. Hi, I'm Stephen Fodor. Whereas Pat Brown's DNA arrays use cDNAs, I developed a technique called GeneChip® probe arrays where I can build the sequences I want to screen. These GeneChip® arrays are printed on special glass. DNA sequences are built up using light-directed chemical synthesis. First, a substrate with a nucleotide is fixed onto the chip at specific positions. The nucleotide has a protecting group (X) that blocks polymerization. This protector group is photolabile and is released on exposure to UV light. Without the protector, polymerization and chain build-up occur. We add a filter to the chip so that only some of the nucleotides are exposed to light. These deprotected groups are then free to add the next nucleotide to the chain. By alternating the position of the filter, we can build a GeneChip® with an array of different sequences about 20 nucleotides long. When we add a cDNA probe to the GeneChip®, we can simultaneously assay tens of thousands of different sequences at the same time. Since this entire process is done with a computer, we can quickly pinpoint the matching sequences, which can then be matched to available DNA sequences in gene databases. In this example, the sequence matches BRCA-2, which has been implicated in causing human breast and ovarian cancer. On the same Genechip®, we can include sequences of known BRCA-2 mutations, and thus have a diagnostic test for women at risk for these cancers. Eventually, everyone can have a personal genetic profile imprinted onto a GeneChip®. Specific sequences unique to a person, known as single nucleotide polymorphisms (SNPs), can be the ultimate DNA fingerprint. Genetic disease mutations can also be imprinted and used for diagnosis or more tailored drug therapies.

thomas sargent, mrna population, gastrula stage, undifferentiated cells, blastula, recombinant plasmids, cell layers, endoderm, mesoderm, dawid, hybridization, hybrids, genes, frogs, differentiation, clones