A Technology That Fits the Genome Like a Glove

By Jason Socrates Bardi

People with thick hair know that one of the best ways to comb it is with their fingers, but could it also be the best way of combing the gene-thick strings of the human genome?

Combing the genome for genes that underlie biology and disease is a hot topic now that the human genome has been solved, with the final draft to be published next month. Now three scientists from The Scripps Research Institute (TSRI) are proposing that one of the easiest ways to identify genes and explore genetic pathways in the genome is to, well, run their fingers through it.

Research Associates Pilar Blancafort and Laurent Magnenat and Professor Carlos Barbas III, who holds the Janet and Keith Kellogg II Chair in Molecular Biology at TSRI, have published a paper in this month's Nature Biotechnology that describes a new technique for looking for genes with a combinatorial library of zinc finger proteins.

Zinc finger proteins contain particular zinc finger domains that each specifically bind to a particular three-base-pair sequence of DNA—a codon. Zinc fingers are a common protein motif in nature because they bind to DNA. They come in various shapes and sizes, but they all chelate a zinc ion in their binding domain, and they all have a long alpha helix that inserts into the major groove of DNA, making contact with the bases.

Several years ago, Barbas found he could string several of these zinc finger domains together into a "hand" to create a highly selective specificity for a longer, more unique sequence of DNA. Using phage display and oligonucleotide hairpins (short, single-strand pieces of DNA that twist into tiny helices to which the zinc fingers can bind), the Barbas laboratory selected for zinc fingers that bound to sequences of interest and in cases where selections failed, they designed the desired protein domain.

A Gene Regulation Library

One of the sequences of interest are those "regulatory" regions within genes, where proteins known as transcription factors bind and turn the genes on or shut them off. By fusing zinc fingers with other, repressor or activator domains, Barbas and his colleagues found a way to design transcription factors to specifically down- and up-regulate genes for which they knew the regulatory regions.

In a series of reports a few years ago, Barbas demonstrated the efficacy of using polydactyl zinc finger proteins to bind to two 18-base-pair sequences in the regulatory regions of the protooncogenes ERBB-2 and ERBB-3. These two genes are involved in human cancers, particularly breast and ovarian cancers, and show increased expression in cancerous cells.

Now Blancafort, Magnenat, and Barbas have extended their technique to generate a library of zinc finger proteins that can be used to discover genes and to turn on or off virtually any gene in the genome.

The team used a combinatorial strategy to generate a library of nearly 100 million zinc finger protein variants from previously optimized zinc finger domains into multimodular 3- or 6-zinc finger proteins, and they have the ability to deliver up to 10 million at a time into cells using a retroviral vector.

Many of the sites where these fingers bind do not lead to regulation of genes, but with many more zinc fingers than there are genes in the human genome, they have many more chances to hit any one particular gene. If one zinc finger does bind within a promoter region of a gene, it can be linked to a promoter or repressor protein and become a regulator that activates or suppresses the gene to which that zinc finger binds.

In their study, the researchers applied the libraries of zinc fingers to cells and selected cells for the expression of surface markers that were up or down regulated by the zinc fingers. In this way, they were able to select for those that bound to the regulatory region of the target gene.

To read the article, "Scanning the human genome with combinatorial transcription factor libraries" by Pilar Blancafort, Laurent Magnenat, and Carlos F. Barbas III, please look at the March 2003 issue of Nature Biotechnology, page 269, or see:




Human carcinoma cells retrovirally infected with a large combinatorial library of artificial transcription factors. Click to Enlarge