Designer Zinc Finger Proteins

By Jason Socrates Bardi

For the past several years, Carlos Barbas, who, at 36, holds the Janet and Keith Kellogg II Chair in Molecular Biology, has been pursuing one line of research aimed at answering a simple—but important—question: can one design proteins to regulate the expression of any human gene?

With zinc fingers, a common structural element found in proteins, says Barbas, the answer is “yes.”

Barbas has found a set of these small zinc finger protein "motifs" that each specifically bind to a particular three base pair sequence of DNA—a codon. By stringing several of these zinc fingers together, he can create a multiple zinc finger protein that can bind any sequence of interest, including unique regulatory regions, and to which he can fuse repressor and activator proteins to specifically down- and up-regulate those genes.

Finding Zinc Fingers that Bind

Barbas has accomplished this feat using the same technology that he developed for his antibody research—phage display.

In his antibody research, which is a separate, thriving area within his laboratory, Barbas isolates catalytic antibodies—human immunoglobins that have useful chemical activity in addition to their antigen specificity.

For instance, Barbas and his colleagues designed the first commercially available human catalytic antibody, 38C2, which can be made to bind certain markers on a cancer cell and catalyzes reactions there. This is useful, because patients can be injected with the antibody and then treated with a highly powerful anti-cancer agent they otherwise could not tolerate. Many such agents are too toxic to administer, but they can be rendered nontoxic by slightly modifying their chemical structure. But the catalytic antibodies, hinged to surface of the cancerous cells, can modify them back to their potent form once they reach the cells.

Barbas and his colleagues find these antibodies through phage display.

Phage display is a method of generating billions of protein variants and selecting for those that bind to a particular target. In the technique, a protein is fused to a viral coat protien of the phage—a filamentous virus that infects bacteria. Then the virus is allowed to reproduce in culture, where it copiously makes new copies of itself. Billions of variants of a single protein can be studied in this way.

Since the phage virus displays these proteins on the surface of the virions, it makes them easy to select for in vitro by simply passing the viral stew over a stationary phase containing the target substrate. Those that can bind will, and the ones that bind the best will bind the tightest.

"It became obvious to me that we should use this approach to evolve proteins that bind specific DNA sequences," says Barbas. "If we could do that, then we could do almost anything in the genome that we wanted to—turn genes on or off."

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.

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 bind their target codons with at least 100-fold greater affinity than they bind other sequences that are a single base change different.

Since three base pairs is the length of a codon and there are 43=64 possible ways to combine the four types of bases into a codon, Barbas needed to develop 64 such zinc fingers to cover all possibilities—a goal which is now in sight.

"We have just a few zinc fingers [left] to find," he says.

Regulating the Regulators

Finding the fingers was only the first step, however. The next was to combine several of these zinc finger proteins into a "hand," so that they have a highly selective specificity for a longer, more unique sequence of DNA. Barbas knew he had to combine a minimum of six or seven fingers together in order to recognize 18 to 21 base pairs.

"That DNA address is long enough that it is potentially found only once in a genome," says Barbas.

Then he had to fuse on the activator or repressor domains that would do the work of turning the gene on or off. A particular six zinc finger combination fused with a repressor should silence the gene’s expression, and the same six zinc fingers fused to an activator should increase expression.

And he had to design switches that would regulate the activation or repression of the action of small molecules which, when bound to the protein, would prevent the zinc fingers from binding to the DNA. And he had to design reporter systems—green fluorescent protein and other markers—so that the binding and to and regulation of the genes could be studied in the laboratory.

The ability to regulate the regulators is particularly compelling, because it increases the possibilities for clinical control. For instance, if one wanted to treat growth hormone deficiency, one could presumably deliver a transcription factor that would enhance transcription of growth hormone. But more importantly, one could couple the transcription factor with a switch that would activate or inactivate the gene whenever a pill was swallowed.

"It can be done either way," says Barbas.

Another possibility is to use a tissue-specific switch or a switch that would be sensitive to endogenous signals. Glucose levels, for instance, could conceivably be a signal that turns on or off a gene that encodes for a protein that increases the production of insulin.

These possibilities are all down the road, but Barbas has already designed several successful zinc finger transcription factors to regulate several different genes.

"What we’re doing now," he says, "is applying this approach to many types of diseases. We have encoded in our own genes the solutions to many diseases."

"What awaits is just a way to switch on those critical genes."

Shutting Down the Cancer Genes erbB-2 and erbB-3

In a series of reports last year, Barbas demonstrated the efficacy of using multiple zinc finger proteins to bind to two 18 base pair sequences in the promotor 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.

By fusing a set of zinc fingers that bound to the regulatory regions of these genes with the repressor protein KRAB, they were able to shut off expression of erbB-2 and erbB-3. Then, by fusing the transactivator protein VP64, they were able to dramatically increase the transcription of the genes.

"We’re also using the approach to target sickle-cell anemia," says Barbas.

Sickle-cell anemia is a chronic disease prevalent in individuals of African, Mediterranean, or Southwest Asian origin where the oxygen binding protein hemoglobin inside red blood cells have a single amino acid mutation that causes them to polymerize, which in turn causes the red blood cells to become misshapen into "sickles." This condition can lead to clotting, vascular damage, thrombosis, and extensive tissue and bone damage in adults.

All humans have a redundant hemoglobin gene, however. The fetal hemoglobin gene is always normal in adults with the sickle cell mutation, and as a result, people do not suffer from sickle-cell anemia as infants. At some point in early development, though, the responsibility for producing the body’s hemoglobin switches from the fetal gene to the adult gene. A number of drugs already exist that increase the ratio of fetal to adult hemoglobin in the body, but they are all rather toxic.

"If we can make a genetic switch that turns on fetal hemoglobin," says Barbas, "then we can deliver it the same way we can deliver other genes [into stem cells]." Because zinc finger transcription factors are rather small proteins, several of them can be inserted at once, and another switch could be delivered to simultaneously turn off the bad or defective adult hemoglobin gene.

This sort of therapy would seek to insert the gene for the zinc finger regulatory protein into stem cells of a patient suffering from sickle-cell anemia. Then these stem cells would grow into red blood cells that would express the normal, fetal hemoglobin.

The advantage of a stem cell-based therapy is that it could be done once and the gene correction would be in place forever.

A Potential Strategy Against HIV

This sort of therapy could prove useful as a way of treating people who are infected with HIV as well. The Barbas laboratory has collaborated with Associate Professor Bruce Torbett on an effort to develop a method for delivering therapeutic genes into patients' stem cells. See A Vaccine Factory Inside Each Cell. .

In that effort, Barbas and his group isolated an antibody, and its gene, that is specific for CCR5—a coreceptor that is essential for viral entry. They then designed a peptide "anchor" on one end that retains it in the cell’s endoplasmic reticulum. These intracellular antibodies have been termed "intrabodies." Once the intrabody gene is incorporated in a cell, the cell would express the intrabody that would then grab the CCR5 and keep it from getting to the surface of the cell, effectively preventing HIV from entering the cell.

But HIV has no proofreading mechanism to correct mistakes when it transcribes itself, and this makes the virus notoriously error prone, encoding one mutation with every copy. This error rate is one of the confounding problems of HIV therapy, of course, because mutations in the gene products that are themselves targets of the therapy can reduce the affinity—and therefore the efficacy—of the drug. For instance, HIV can stop using the CCR5 coreceptor and switch to the CXCR4 coreceptor, which cannot be knocked out.

However, mutations that knock out the expression of essential genes are lethal to the virus. Fortunately, certain parts of the viral genome do not tolerate mutations—they are completely conserved.

Barbas now believes the same stem cells can also be modified so that if the virus does mutate and does get in, it can still be blocked through the action of zinc finger transcription factors designed to target these conserved regions.

For instance, the cells could be equipped with proteins that, when expressed, could bind to the conserved long terminal repeat regions of the HIV genome, which are the ends that integrate into the host cell's DNA. Zinc fingers bound at these sites should prevent the DNA from inserting.

"We're looking towards building those fingers now," says Barbas.

Another conserved region is the primer binding site, which is the 21 base pair site at the front of the HIV genome where reverse transcription of the viral RNA into viral DNA starts. After a virion enters a cell and uncoats, the RNA must attach to a human tRNA in order to turn on the reverse transcription. In fact this is the most conserved region in the HIV genome. "Since the human tRNA is not changing, the virus doesn't change," says Barbas. "If the primer binding site changes, then the virus cannot replicate in human cells."

Targeting this site with a zinc finger fused to a repressor protein should stop viral replication.

"Ideally, we will be able to make multiple blockades of the HIV virus," Barbas says. The CCR5 blockers would prevent the HIV from entering cells. The LTR binding fingers, if Barbas is successful in designing them, could prevent the virus from integrating into the host cell genome, and the primer binding site repressors could shut down replication.

"So if we don't get it coming in, at the surface, or before it gets into the nucleus," says Barbas, "we’ll inhibit stop it once it tries to make viral RNA."


Go back to News & Views Index


Principal investigator Carlos Barbas’ s interest in designing proteins grew out of his work on antibodies.











“We have encoded in our own genes the solutions to many diseases. What awaits is just a way to switch on those critical genes.”

—Carlos Barbas III














Three views of zinc fingers bound to DNA. Top panel shows the space filling model of the zinc fingers (in blue) bound to the double stranded DNA (in red and orange). Middle panel shows space filling model with backbone ribbon. Bottom panel shows a partial space filling model of areas of contact with side chains shown as worms.









For more information, visit: