Commandeering cellular machinery: recognition mechanism to detect small molecules
Researchers have learned how to commandeer the complex machinery that cells use to recognize and respond to such important molecules as steroid hormones, thyroid hormones and vitamin D.
The development could provide a foundation for a new family of biologically-based mechanisms able to detect common drugs, chemical weapons and other small molecules. By allowing manipulation of this cellular protein machinery – known as nuclear receptors – the technique could also lead to new methods for producing enzymes and important pharmaceutical compounds. "We are hijacking these nuclear receptors for a new set of purposes," explained Donald Doyle, assistant professor in the School of Chemistry and Biochemistry at the Georgia Institute of Technology. "We want to change the nuclear receptors themselves so they dont recognize what they normally recognize, and instead recognize the small molecules we want to detect. That would allow us to develop a new type of sensing mechanism."
A paper published in the September 27 – October 1, 2004 issue of the journal Proceedings of the National Academy of Sciences describes how Doyles research team – which also included Lauren Schwimmer, Priyanka Rohatgi, Bahareh Azizi and Katherine Seley – modified one type of nuclear receptor to bind a drug compound to which it previously did not respond. Based on this success, the researchers hope to demonstrate broader application with other small molecules.
The work was sponsored by the Research Corporation, the Seaver Foundation and the National Science Foundation.
Nuclear receptors are ligand-activated transcription factors contained in cells. When activated by specific small molecules, the nuclear receptors initiate a complex process that results in gene expression. Because these receptors play a vital role in controlling cellular response to these small molecules, scientists have been attempting to understand them and solve their molecular structures – with a goal of creating pharmaceuticals able to turn them off or on. The research often involves placing nuclear receptors into yeast cells, which because they do not have nuclear receptors of their own, allow scientists to isolate the activities of specific receptors.
For their study, Doyles team chose the nuclear receptor retinoid x receptor (RXR) whose molecular structure has been well documented by other researchers. Using a technique known as structure-based codon randomization, the Georgia Tech researchers used their knowledge of RXRs structure to modify the 20 different amino acids that make up the receptor pocket. The goal was to create a library containing 32,768 different variations in the hope of creating – and then finding – a few re-engineered receptors that would have the ability to bind to a molecule known as LG335. (LG335 is structurally similar to Targretin®, a pharmaceutical developed to bind RXR. However, LG335 does not effectively bind unmodified RXR.)
Though the goal was 32,768 variants, the method actually produced approximately 380,000 samples – which would have been impossible to test using conventional screening methods. However, Doyle and his research team used a new protein engineering technique known as chemical complementation that allowed the variants to be tested in parallel – a task akin to finding genetic needles in thousands of haystacks simultaneously.
Using the technique, the researchers placed each variant, which had been inserted into yeast cells, onto petri plates containing LG335. A small number of yeast cells (less than 0.1 percent of the total) grew into visible colonies, suggesting they might contain nuclear receptors for the LG335.
To verify that the colonies were growing in response to LG335 and not some other compound contained in the yeast, samples were then smeared onto another petri plate that did not contain LG335. Any colonies that grew there were discarded. From their initial 380,000 candidates, the researchers ended up with about a dozen nuclear receptors whose recognition pockets had been re-engineered to respond to the LG335.
"Using this technique, we dont have to evaluate each member of our library," Doyle said. "The yeast actually do the work for us. If only one in a million respond to the compound on the plate, the yeast will form a colony around that one. This allows us to do highly parallel and rapid screening to find a few functional receptors in a large collection of nonfunctional receptors."
The receptors found by the researchers varied in their responsiveness, with some significantly more sensitive than others. Some performed the function of on-off switches, while others were more like dimmer switches, responding in proportion to the amount of LG335 bound to them.
Doyles team also studied their re-engineered receptors in mammalian cells – and found they behaved much the same as in yeast. The re-engineered receptors could be used in gene therapy against cancer, or as research tools to investigate gene function. But more importantly, the new receptors serve as proof of principle for the protein engineering technique used to produce them. Doyle envisions using that technique to produce other nuclear receptors that could be the basis for sensing arrays in which a variety of receptors, each sensitive to a different compound, could be used to quickly analyze an unknown agent. A sensor array might also be used in a hospital emergency room to rapidly test for chemical agents in the blood of an unconscious patient. "If we could take the receptors, express them and put them into a device where there is a color change or another signal produced, we could potentially detect small molecules in a robust way that could complement or replace other detection technologies," he explained. The same technology could also be used to produce enzymes, and to regulate metabolism in cells, Doyle said.
Having demonstrated the ability to re-engineer one nuclear receptor to respond to a small molecule to which it previously did not bind, the research team next wants to demonstrate that the technique could apply to other small molecules. "Now we have to see how far we can push this and how many small molecules we can accommodate with this technique," Doyle said. "We are trying to generalize this approach to genetic selection. There is a lot of diversity we can work with in terms of different binding pockets and shapes, so this is only the first step."
John Toon | EurekAlert!