Once thought to be static platforms for fundamental chemical messages, nucleic acids are known today to be as varied and dynamic as many of the rest of the cells' molecules.
Chemists at U.C. Berkeley are looking beyond Watson and Crick's 1954 model of nucleic acid structure, using the scientific equivalents of a magnifying glass and a fine-tooth comb to search for clues that relate form to function.
Perhaps one of the most dramatic revelations that has been made about nucleic acids in the past 40 years is the fact that ribonucleic acid (RNA) and its close cousin, the double-stranded molecule deoxyribonucleic acid (DNA), are strikingly different.
"In the past, people used to think of RNA as just a copy of DNA," said Professor of Chemistry Sung-Hou Kim. "Since then, people started discovering that RNA can assume many different structures and that they have many different functions."
According to Kim, the complexity of RNA function is not well understood on a structural level. Progress in this area is slow, Kim said, because RNA's extra hydroxyl group changes its properties, making it difficult for scientists to isolate sufficient quantities of material in high purity to study.
"It is very difficult to do chemistry of RNA," said Kim. "DNA chemistry is now quite well worked out, but RNA chemistry is lagging behind. Because of that, there is very little structural information about RNA."
What scientists do know about the structure of these molecules, according to Professor of Chemistry Ignacio Tinoco, is that they are folded into complex tertiary formations more like proteins than DNA, which is constructed linearly.
This difference can be attributed to the fact that RNA is composed of combinations of single-stranded and double-stranded regions while DNA is purely a double-stranded molecule. DNA bases associate with complementary bases on the adjacent strand forming a helical rod, as Watson and Crick first demonstrated. RNA bases form Watson and Crick base pairs as well, but can also pair to other bases along the same strand, causing the molecule to fold back on itself in a series of loops and knots.
"Every base has been found in a biological molecule to be paired with every other base," Tinoco said. "Everybody is trying to learn about these non-Watson/Crick base pairs and what role they play in the different RNA structures."
Tinoco's research group is examining these structural elements as they occur in approximately 30-nucleotide segments. Typically these segments contain such things as "hairpin loops," "bulges" and "pseudo-knots."
Tinoco's group has a particular interest in the pseudo-knot structure because it is the component of RNA retroviruses that allows them to cause frameshifting during translation of the viral RNA into protein.
"We have made a lot of mutations [in the RNA] to try to relate the sequence and the structure to its frameshifting function," Tinoco said.
Kim's group, in collaboration with Dr. Stephen Holbrook of LBL, is studying another structural element called an "internal loop," which takes its name from scientists' early predictions that RNA duplexes containing stretches of non-complementary bases will contain looped regions where the mismatched bases exist.
"It turns out the shape doesn't look like this," said Kim. "There's no actual loop. When you look at it, it looks pretty much like a normal duplex except that this area where it is non-complementary makes unusual hydrogen bonds."
Kim said he first observed non-Watson/Crick base pairing in RNA in 1974, when he solved the structure of transfer RNA, one of the smallest biologically-functional RNA molecules.
"Finding the three-dimensional structure of tRNA was an important milestone for RNA structure because that was the first discovery of any RNA structure," Kim said. "The structure taught us a lot about the complexity of RNA, but it was just a beginning glimpse."
Hoping to follow up on that success, Kim's group is focused on using x-ray crystallography to solve the structure of another intriguing RNA molecule called "hammerhead."
Hammerhead is a ribozyme -- an enzyme composed of RNA. Typically, enzymes are proteins, which are made of amino acids. Hammerhead, which has the ability to cleave other RNA molecules, is an example of an RNA that demonstrates the extreme diversity of this class of nucleic acids.
"That's what has made the RNA structure much more exciting -- its catalytic activity and the knowledge that RNA interacts with lots of other things just the way proteins do," Tinoco said.
In order to visualize the structure of hammerhead in its active state, Kim has mutated the ribozyme so that the rate of its enzymatic reaction is reduced to a point where the molecule can be crystallized.
According to Professor of Chemistry David Wemmer, the discovery of such broad variation in RNA structures is exciting but not surprising, considering the fact that it is a single-stranded molecule capable of forming intricate shapes. More surprising is the structural variation discovered in DNA, a molecule conventionally envisioned as a fixed duplex.
"There's a lot of local structure variation and some larger scale variation," said Wemmer, adding that there are sometimes joints within pieces of DNA, cross-linked structures which produce cruciforms, and hairpin turns in the helical lattice.
"There's a good deal of evidence that such things are involved in regulation in various ways," Wemmer said.
One example, said Wemmer, is a phenomenon known as supercoiling, where forces in the cell try to twist the DNA along its long axis.
"It turns out that when you form this kind of structure, you release some of the supercoiling energy," Wemmer said.
Supercoiling and other subjects that are based on the notion that DNA is an elastic rod are of particular interest to Professor of Chemistry John Hearst.
"One of the principal questions that my lab has spent time on is the issue of how much DNA is unwound, what actually is moving, and how the torsional tension that is associated with these processes is released by things called topoisomerases," Hearst said.
Hearst is specifically interested in the motion of the enzyme RNA polymerase, which is responsible for transcribing the DNA template into an RNA molecule. His work has established that the polymerase moves along the DNA helix like a nut does on a bolt and that, as this occurs, the DNA is unwound in a small loop ahead of the polymerase and wound up again behind it, creating torsional stress in the DNA.
According to his "continuous" model of transcription, DNA is driven rotationally at about four turns per second relative to the polymerase, which is fixed in place.
Hearst said this model is particularly valid when a sequence that codes for a membrane-bound protein is transcribed because in this situation, the polymerase is anchored to the cell's membrane.
However, said Hearst, when DNA is transcribed in the cytosol of the cell, there are a different set of topological problems to resolve.
"If you want the RNA to emerge on one side of the DNA and not be whipping around like a propeller, then at some point, there either has to be a strand cross-over, where the RNA is lifted up over the non-coding DNA strand, or there has to be an enzyme that cuts and reseals," said Hearst. "Otherwise, you get entanglement."
Although there are various schools of thought in regard to this issue, the true mechanism remains unknown.
Another aspect of DNA research that intrigues many scientists is the subject of how proteins interact with DNA to regulate the activity of genes. Wemmer's group is also currently examining how DNA structure plays a role in the molecule's affinity for small molecule ligands.
The group is examining the binding of a naturally-occurring, peptide-like molecule to the minor groove of DNA -- the narrowest of two grooves formed by the twists of the double helix. The molecule they are using has been observed to bind only to regions of DNA that are rich in adenine (A) and thymine (T) base pairs. Other researchers have found that such regions contain specific functional groups jutting into the groove with which the molecule interacts. But recently, Wemmer has discovered another pivotal characteristic needed for binding.
"The most important factor is that when there is a run of four or more A's and T's together," said Wemmer, "the narrow groove is narrower than normal. It clamps down and this molecule slides in with a very tight fit."
Scientists' attempts to change the sequence-specificity of the protein so that it would bind to areas rich in guanine (G) and cytosine (C) bases have failed in the past. According to Wemmer, this is because the researchers only manipulated the ligand in terms of its reactivity to the functional groups found in G/C-rich regions.
"If you realize that A/T's have a very narrow groove and G/C's have a rather wide groove, you have to take that into account in your design," Wemmer said.
Using this missing piece to the puzzle, Wemmer was able to combine what was known about the interactions of these molecules with his knowledge of groove width to fit two ligands side by side into a G/C-rich stretch of the minor groove.
"We know what functional groups are on the DNA and we know the relationship of the sequence of the DNA to the width of the groove," said Wemmer. "So, we know how to design a ligand that will fit and have functional group complementarity."
According to Wemmer, the design scheme is general enough so that one could essentially customize any molecule to bind a specific sequence of DNA.
"We've got half a dozen different combinations that have been tested and they all work so well that we're very confident that this extended idea will work," Wemmer said .
Hearst's group is also studying molecules that bind to DNA, specifically the class of photochemicals known as psoralens. Psoralens intercalate into the DNA structure and, when irradiated with ultraviolet light, react with the DNA, covalently tying its two strands together.
Taking advantage of psoralens' ability to easily penetrate cellular barriers, Hearst's group has used this reaction for the past 20 years as a way to probe in vivo DNA structure.
"This is a chemical way of taking a photograph at any particular instant in time because you turn the psoralens on with light and they give a fingerprint of where the DNA is reactive and where it is not," said Hearst. "This is now being used as a generalized probe to give information about the large-scale arrangement of DNA in the cell."
Along with Professor of Chemistry Henry Rapoport, Hearst was able to synthesize improved, more reactive psoralens that were used to map the complex hairpins in ribosomal RNA.
Hearst's research has pushed psoralens into the biomedical industry spotlight.
"We demonstrated very early that you could inactivate almost all viruses using this chemistry because psoralens get into both RNA and DNA viruses," said Hearst. "Once you've reacted the nucleic acid with the psoralen, you basically inactivate the virus."
Hearst is currently putting this research to the test, using psoralen technology as the basis for a company called Steritech, which is developing products to sterilize blood fractions.
"Everything that is pathogenic in blood contains nucleic acid. So if you knock out the nucleic acid in a gentle way, ...you can make it sterile and still retain the function of the blood fraction," Hearst said.
In fact, because of the fundamental nature of nucleic acids, most DNA and RNA research is bound to spur new biotechnological and medical innovations in the near future.
Wemmer said that his success in selectively targeting a ligand to a short DNA sequence is the first step towards a day when custom-designed molecules could quickly deliver various reactive groups to specific genes to manipulate their activity or dampen the effects of mutations. However, Wemmer acknowledged that such technology is "still a dream," adding that there are many steps that come between basic research and a safe clinical drug. The next step in his research will involve designing chains of customized molecules that recognize longer DNA sequences.
RNA research has also yielded techniques that can be used to regulate the activity of nucleic acids, according to Tinoco. Many biotechnology companies are currently taking advantage of "anti-sense" technology, where one synthesizes a strip of RNA that is complementary to a specific region on an RNA molecule. When the anti-sense strand binds to the RNA, it prevents other enzymes from binding and translating that particular region. Such technology could be used to deactivate retroviruses such as HIV, the virus that causes AIDS.
Some researchers are also developing an analogous technique for DNA called "triple-strand formation," where a complementary base sequence is synthesized that can form hydrogen bonds to specific areas of the DNA duplex. The complementary sequence prevents cellular proteins from binding to the DNA and can also deliver oxidizing agents that cut the DNA at a particular point.
Ribozymes also show promising potential for biomedical applications. Tinoco said that ribozymes are especially efficient genetic tools because they can perform a particular function many times at a specific location in the cell.
However, Kim said that scientists need to collect more information about ribozymes and other RNA molecules before they can be effectively used.
"All of these interventions require knowledge of the structure," said Kim, adding that such information will allow researchers to design ribozymes for specific biological tasks.
"RNA is so fundamental, it is going to touch on almost everything you want to do in the cell," Kim said. "The medical applications are just a small part of it."