Advertisement

If you have an ACS member number, please enter it here so we can link this account to your membership. (optional)

ACS values your privacy. By submitting your information, you are gaining access to C&EN and subscribing to our weekly newsletter. We use the information you provide to make your reading experience better, and we will never sell your data to third party members.

ENJOY UNLIMITED ACCES TO C&EN

Biological Chemistry

DNA-programmed Organic Synthesis

DNA encodes small-molecule synthesis and offers potential to evolve drugs

by CELIA M. HENRY, C&EN WASHINGTON
January 31, 2005 | A version of this story appeared in Volume 83, Issue 5

DNA is considered the blueprint of life because of its role as the material of genes. Now, researchers are showing that DNA can be the blueprint for the synthesis of libraries of nonnatural small organic molecules as well.

Just as DNA allowed life to evolve, it could eventually allow researchers to evolve their way to drugs or other biologically interesting small organic molecules. Evolution would involve selecting members of a population that meet particular criteria and then "breeding" the "winners" with one another, says biochemist Pehr A. B. Harbury of Stanford University. "In order to do that with small molecules, you need to have some kind of blueprint that can be propagated forward and can be crossed with other blueprints," Harbury says.

DNA sequences can be used to program small-molecule organic synthesis in two general ways: by bringing two reactants close enough to one another to increase their effective concentration or by directing the partitioning of reactants in a twist on traditional split-and-mix combinatorial chemistry. The first approach is usually called DNA-templated synthesis. The second method is known as DNA display. Both have as the end result a small molecule attached to a DNA molecule that encoded its synthesis.

DNA-programmed synthesis has been pioneered by two academic groups--led by David R. Liu at Harvard University (DNA-templated synthesis) and Stanford's Harbury (DNA display)--and by a group at the Danish company Nuevolution led by Henrik Pedersen. Pedersen's work in this area has been published only in the form of patents.

In DNA-templated synthesis, the DNA controls the reactivity of the reactants. The template is a piece of single-strand DNA that is 30 to 100 nucleotides long and consists of three or four 10- to 20-nucleotide-long coding regions--codons--and a starting material tethered to the end. Reagents are attached to 10- to 20-nucleotide pieces of single-strand DNA that are complementary to the codons. The reagents are at low enough concentrations--nanomolar to micromolar--that off-template reactions are unlikely to occur. When complementary strands bind, the reagents and starting material are brought close enough together that they experience effective concentrations in the millimolar to molar range. The increase in effective molarity allows multiple reactions, even ones that would normally not be possible in a conventional organic synthesis format, to occur in a single solution without interference [Angew. Chem. Int. Ed., 41, 4104 (2002)].

CLOSE UP
[+]Enlarge
Credit: COURTESY OF DAVID LIU
In DNA-templated synthesis, reagents at nanomolar or micromolar concentrations are attached to short strands of DNA. When the DNA sequences are not complementary (top), the concentration is too low for the reagents to react. If the sequences are complementary, the reagents experience an increase in their effective concentration to the millimolar or molar range and the reaction occurs.
Credit: COURTESY OF DAVID LIU
In DNA-templated synthesis, reagents at nanomolar or micromolar concentrations are attached to short strands of DNA. When the DNA sequences are not complementary (top), the concentration is too low for the reagents to react. If the sequences are complementary, the reagents experience an increase in their effective concentration to the millimolar or molar range and the reaction occurs.

In earlier versions of DNA-templated synthesis predating Liu's and Pedersen's independently developed work, the products mimicked the structure of DNA itself. "We decided to test whether DNA-templated synthesis was sufficiently general that you could make an arbitrary synthetic structure that didn't look like DNA," Liu says. The first DNA-templated reaction that his group carried out was a thiol maleimide addition [J. Am. Chem. Soc., 123, 6961 (2001)].

Liu and his coworkers have since expanded their repertoire to hundreds of different reactions and dozens of reaction types, including carbon-heteroatom coupling and carbon-carbon bond formation. "We're convinced that DNA-templated synthesis is a general way to make synthetic structures" that don't resemble DNA, Liu says.

In addition to single-step reactions, Liu and his coworkers have demonstrated that they can use DNA to program multistep organic synthesis. For their first multistep synthesis, they performed an amine acylation, followed by a Wittig olefination and a thiol addition, resulting in a series of branched thioethers [J. Am. Chem. Soc., 124, 10304 (2002)]. They've since used multistep DNA-templated synthesis to prepare heterocycles [J. Am. Chem. Soc., 126, 5090 (2004)] and libraries of synthetic small-molecule macrocycles [Science, 305, 1601 (2004)].

WHEREAS DNA-TEMPLATED synthesis is based on proximity, DNA display uses DNA to segregate molecules into subpopulations [C&EN, July 12, 2004, page 23; PLoS Biol., 2, 1015, 1022, and 1031 (2004)]. To achieve the segregation, the polystyrene bead typically used in combinatorial chemistry is, in a sense, replaced with a "polyanion handle" supplied by DNA, Harbury says. He has used DNA display to make a library containing more than a million peptides and has taken that library through two rounds of selection to find a ligand that binds to a protein target.

In Harbury's system, the DNA molecules are made of 20-nucleotide coding regions separated by 20-nucleotide noncoding regions. Hybridization of the coding regions to complementary DNA immobilized on a resin segregates the molecules, which are then transferred to another resin where reaction takes place. For multistep reactions, the process is repeated for each subsequent reaction. In between reactions, the DNA sequences are pooled and resegregated.

The DNA in Harbury's method serves as "zip codes that tell, on a molecule-by-molecule basis, where the molecules should go to get their chemistry done," says Gerald F. Joyce, a professor of chemistry and molecular biology at Scripps Research Institute. "Then everybody comes back to the post office, gets shuffled up, and the next set of digits in the zip code get read to tell you where to go for your second synthesis." The DNA serves as a "paper trail" telling what reactions were done.

ENCODED DNA
[+]Enlarge
Credit: © 2004 D. R. HALPIN AND P. A. B. HARBURY
display uses DNA molecules consisting of a series of coding regions (colored rectangles) to direct the split-pool synthesis of a small molecule attached to the end of the DNA (colored circles).
Credit: © 2004 D. R. HALPIN AND P. A. B. HARBURY
display uses DNA molecules consisting of a series of coding regions (colored rectangles) to direct the split-pool synthesis of a small molecule attached to the end of the DNA (colored circles).

BOTH METHODS have constraints on their reaction conditions. Liu's method requires conditions that are compatible with DNA hybridization, meaning that the pH can't be too high or too low and the temperature can't be too high. It generally requires aqueous solutions, but Liu has had some success working with organic solvents. Neither method works well at extremely low pH, which would cause DNA to lose purine bases.

Joyce believes that the true strength of both DNA-based methods is the promise of being able to evolve small-molecule products based on mutations of the coding regions. The molecules can compete against one another to find the best ones for a given purpose, and the best "become the breeding stock for the next population," Joyce says. "That's why I'm so excited about these technologies. They can allow a truly Darwinian search." Neither Liu nor Harbury has yet demonstrated molecular evolution, but Joyce says he wouldn't be surprised to see both of them do so by the end of this year.

Harbury, however, believes it is premature to talk about using evolution. "That process of introducing mutagenesis or recombination is only relevant when the complexity of the library is greater than the number of molecules in the test tube," he says.

One of the challenges in using DNA-directed methods to evolve small-molecule drugs is keeping molecular masses below 500 daltons, the limit for most drugs and druglike molecules. "At a six-step synthesis, each monomer has to be less than 100 Da, which is getting pretty small, so you're stuck in the four- to six-step synthesis range," Harbury says. "If you go through the math, you'll see that you need thousands of building blocks per step" to get enough diversity to need evolution.

DNA-directed synthetic methods are already starting to be used for drug discovery. At Nuevolution, both the DNA-templated synthesis and a simple split-and-mix DNA-tagging method are used in combination with affinity selection methods to discover new drug leads. Nuevolution's suite of technologies is called Chemetics because of its fusion of synthetic chemistry and molecular genetics. "We see the two approaches as being complementary," Pedersen says.

Pedersen considers DNA-templated synthesis complex and resource intensive. Thus, the company's first efforts are based on the less complex split-and-mix DNA-tagging method. "We focus on drug discovery, so our main focus has been the development of robust and efficient drug discovery technologies allowing the generation of large libraries of high-quality, not challenging or fancy, chemical reactions. We try to make our compounds as druglike and as diverse as possible but still be below 500 Da. To do that, it is a big advantage to have access to these alternative chemistries and to a diverse set of druglike building blocks. For that, we use the simpler DNA-encoding schemes."

Nuevolution uses its Chemetics approaches to generate large libraries. Last year, the company made and screened a library of 100 million compounds in just a few weeks, Pedersen says.

At first, pharmaceutical companies were skeptical that the approach would work, but as Nuevolution shows that it can generate large libraries and identify potential targets from the libraries, they are starting to accept that it might work, Pedersen says. That the approach works with reactions routinely used in medicinal chemistry has been a highly convincing argument for the feasibility, he adds.

"We needed to convince them on the chemistry side more than the biology side. They always believed in the encoding scheme," he says

Pedersen's lofty goal for the technology is to use large libraries to skip the lead optimization step in drug discovery by simply sorting through a diverse set of ligands. Achieving that is at least two to four years in the future, Pedersen believes.

Nuevolution is not the only company using such methods. Liu has also formed a company, called Ensemble Discovery, to apply DNA-templated synthesis to drug discovery and reaction discovery. The company, located in Cambridge, Mass., raised $15 million in its first funding round.

Anthony C. Forster, an assistant professor of pharmacology at Vanderbilt University Medical Center, is excited by the prospect of using such methods, especially Harbury's method, to generate "gigantic" libraries. "I've maintained for a long time that larger libraries are both theoretically and practically more useful for screening ligands in the sense that the more structures you can screen, the greater the probability that you'll find a specific fit to your target," he says. "I see so many diseases for which there's so little progress being made in terms of treatment, despite the fact that we have an abundance of molecular targets. Drug discovery just can't keep pace. One of the limitations is the difficulty in identifying high-affinity, quality lead compounds."

Article:

This article has been sent to the following recipient:

0 /1 FREE ARTICLES LEFT THIS MONTH Remaining
Chemistry matters. Join us to get the news you need.