Shape, not bonds, may drive DNA synthesis

The enzyme that fashions the double helix of DNA works like a reliable, predictable matchmaker, lining up the molecular units of the two strands on the basis of the hydrogen bonds between them -- or so scientists have long thought. New research suggests, however, that the enzyme arranges this marriage by recognizing the shape of those units rather than the bonds between them.

 The four bases that make up DNA pair off depending on how well they form hydrogen bonds with each other: Adenine sticks to thymine, and cytosine fastens to guanine. Those pairings hold complementary strands of DNA together in a double helix.

"Clearly, [the bases] form those hydrogen bonds once the DNA is made," says Eric T. Kool of the University of Rochester in New York, but he asks how the enzyme, DNA polymerase, recognizes which bases go together.

 To test hydrogen bonding's role, Kool and his colleagues looked at how a molecule impersonating thymine affected DNA polymerase's ability to make DNA. They synthesized difluorotoluene, a molecule that has the same size and shape as thymine but that is "terrible" at forming hydrogen bonds with adenine, says Kool. When incorporated into DNA, difluorotoluene makes the double helix unstable.

 The researchers substituted the molecule for thymine in single strands of DNA, which serve as templates for the synthesis. DNA polymerase moves along the template, selecting the appropriate bases and linking them together to form the second strand of the double helix.

 To the group's surprise, the enzyme matched adenine to difluorotoluene without problems. "The molecule by itself is really bad at base pairing, yet the enzyme very faithfully puts adenine opposite it," Kool says. "I expected it to be a bad substrate, but the reverse was true." The synthesis proceeded at about the same rate as if thymine had been in the template, Kool and his colleagues report in the Feb. 26 Journal of the American Chemical Society.

 Moreover, the rate was about 100 times better than that for templates with sites that contain no base. Previous studies had shown that DNA polymerase tends to select adenine for empty spots.

 Other researchers have tested base mimics that don't form hydrogen bonds, Kool says, but those molecules also had different shapes and sizes. "Size effects can be easily as large as hydrogen-bonding effects."

 One unanswered question, says Ronald Breslow, a chemist at Columbia University, is whether the fluorine atoms in the impostor form some kind of bond with adenine. Synthesizing a compound without fluorine would clear up that issue, he adds. Kool says that difluorotoluene is the closest possible analog to thymine; other compounds would introduce differences in size or shape once again.

 The Rochester team now plans a complementary experiment, testing whether, in the absence of free thymine, the enzyme will choose difluorotoluene to insert opposite adenine in the template. They've synthesized mimics for other bases as well, so testing those should provide further information about the role that hydrogen bonds play.