Meghan Richter, Laboratory Production Specialist at Open Biosystems, loads lentiviral shRNA clones onto a Tecan Genesis robot, for production of a cancer shRNAmir array. (Photo by Marc Berlow)
By Angelo DePalma

During the last ten years RNA has revealed itself as much more than a simple messenger, or intermediary between DNA and protein. RNA can act as a catalyst, binding site for small molecules, regulator of gene expression, structural component of ribosomes, and more. Because of its seeming involvement in every biological process, RNA has been termed the “dark matter” of the cell.

Messenger RNA (mRNA) contains, in addition to coding regions, non-coding sequences, which at one time were considered “nonsense” or molecular junk. “But over the years researchers have found interesting things among this junk,” says Ron Breaker, Ph.D. of Yale University (New Haven, CT). Riboswitches, which regulate gene expression, are the latest genetic control elements identified within junk RNA.

Most genetic control factors are proteins, which recognize chemical signals and subsequently bind to genes, activating or deactivating them. RNA riboswitches have the same capability. Breaker is working on seven riboswitches that regulate genes in bacteria by directly transmitting chemical information to a gene without the intervention of a protein. “Biologists used to think that proteins were the active components in cells,” Breaker notes. “Riboswitches have changed the accepted view of gene control elements.”

Riboswitches form when regions of untranslated RNA fold onto themselves, through regions of local complementarity, into molecular recognition pockets. When molecular signals, typically small-molecule metabolites, reach a critical concentration they enter these recognition sites. Riboswitch triggers include coenzyme B12, thiamine pyrophosphate, flavin mononucleotide, S-adenosylmethionine, guanine, adenine, and lysine. Conformational changes in the RNA leads to formation of an intrinsic transcription terminator: mRNA synthesis stops, and with it the gene’s ability to make protein.

Riboswitches are often uncovered serendipitously. For years scientists knew that purine concentrations regulated purine synthesis. In the late 1980s scientists found a protein that turned off purine synthesis by binding adenine and docking to the genome upstream of purine metabolism, stopping transcription in its tracks. A similar effect was observed for guanine, but scientists could not find the relevant protein. Breaker discovered that the mystery control factor was a riboswitch that recognized guanine through base pairing to a single cytidine. When Breaker mutated this residue to uracil, the system became sensitive to adenine. “This was a terrific example of how a minor RNA mutation results in a vastly different sensor that ‘sees’ a completely different molecule,” Breaker notes.

Last year, Breaker and colleagues formed a biotech company to exploit riboswitches for antibacterial drug discovery. He has discovered, for example, that several antibiotics whose mechanisms of action were poorly understood work by mimicking the natural ligand for riboswitches, causing metabolic pathways critical to a bacterium’s survival to shut down. One such drug, pyrithiamin, selectively targets riboswitches in organisms that normally bind a derivative of vitamin B1 (thiamin).

BLOCK-iT Pol II miR RNAi expression vector kits (Invitrogen Corp.) combine artificial miRNAs and the tissue-specific options of pol II promoters with the traditional advantages of RNAi vectors.
Running Interference

For some time after their discovery in plants, in 1990, RNA interference (RNAi) and short-interfering RNA (siRNA) were little more than scientific curiosities. The field took off after researchers found siRNAs in mammals in the mid-90s. More elegant (and practical) than antisense, RNAi has become the gene silencing technique of choice. Since RNAi work is computation-intensive, it has drawn the attention of the informatics community.

Computer scientist professor Lenwood Heath, Ph.D. of Virginia Tech (Blacksburg, VA) is part of an interdisciplinary group of computer scientists, biochemists, and students working at Tech and with collaborators at New York University on computational modeling for RNAi. The goal is to model and predict organism-wide RNAi effects.

Heath’s biologist colleagues verify the models in the nematode, C. elegans. Meanwhile, he is building computational models that integrate phenotype, genotype, protein and gene expression, protein structure, and other data. One member of Heath’s team, a computer scientist, investigates the 2D and 3D structures of siRNA and proteins involved in gene silencing, to determine which factors best predict the gene knockout effect. Other group members apply more discrete computational tools such as data mining, artificial intelligence, and probability, combining them into a unifying model. Researchers believe consensus among these approaches will provide better predictive power.

“We’re particularly interested in organism-wide effects,” Heath explains, “especially those occurring downstream of silencing that affect the entire organism, for better or for worse.”

The Virginia Tech group is targeting the C. elegans gene that controls tolerance to drying out. Researchers hope to knock down that gene and look for effects in other characteristics. Eventually, Heath would like to predict those downstream effects at the level of cells and perhaps the entire organism. These effects may be adaptive, as when other genes compensate for the silenced gene, or destructive, for example if a critical metabolic pathway relies on the protein whose production is blocked by the knockdown.

Lack of specificity — cross-reactivity with multiple mRNAs — makes RNAi effects unpredictable. “We’re still unsure, even in an organism as simple as C. elegans, of how the various factors regulating gene expression, protein translation, and all the other functions of a cell, fit together,” Heath told Bioscience Technology. “We have putative pathways and have clues from gene expression microarrays. But those are only snapshots of transcription that don’t tell us how things are connected. We have no models we can look up to and say, ‘If I flip this switch it will have this effect.’”

HCS (high content screening) WorkCell (Thermo Electron Corp.) combines instrumentation and software to enable complete sample preparation and scheduled analysis for applications in stem cell research, cell health and RNAi.
Free and Easy

Open Biosystems (Huntsville, AL) has sold whole-genome tools since its inception four years ago. Recently, the company has offered RNAi reagents under a business model that resembles that of the open-source software industry. Similar to how Red Hat Software markets the free Linux operating system, Open Biosystems serves as the industrial nexus for commercializing academic research-level RNAi methods and reagents. The company works closely with researchers to develop the tools then optimizes them, sells them, and provides customer support. As end-users improve on these tools, they recycle back into the company, are optimized and re-released.

Among the RNAi reagents for human, mouse, and rat genomes are short hairpin RNAs, which is a DNA-based RNAi platform. The delivery platform is viral based, so genotype is forever. “This allows researchers to make transgenic mice that express the RNAi molecule,” says chief technical officer Troy Moore.

Open Biosystems has combined short hairpins with a lentivirus platform. “Lentivirus has been highly touted in the gene therapy area, but not in research tools,” notes Moore. “We’re hoping to change that because it’s really a unique delivery platform. Lentivirus allows you to go into primary as well as non-dividing cell types, such as neurons. Previously, it was impossible to get short interfering RNAs into those cell types.

According to Moore many early problems with open-access software stemmed from the absence of a viable business model. That has changed as companies created value by providing materials, support, and innovation. Open Biosystems does the same with its RNAi products. It has recently instituted an open access RNAi program that provides universities with copies of the firm’s entire short hairpin library plus ongoing, extended updates. That allows university researchers to expand on the technology or provide services based on it to colleagues. “We provide them the content,” says Mr. Moore, “what they do with it is up to them.”

Analysis of mammalian kidney cells infected with lentiviral shRNAmir. (Courtesy Open Biosystems; photo by Marc Berlow.)
More Than Just Selling Molecules

RNAi has changed biology forever, and with it the business of biological reagents and services. Dharmacon (Lafayette, CO) evolved from a vendor of synthetic RNA at its founding, in 1995, to what William Marshall, Ph.D., VP of Technology Assessment terms a “knowledge provider” for RNA research. The company was founded by Stephen Scaringe to commercialize RNA synthesis technology invented by Scaringe and his mentor, Prof. Marvin Caruthers at the University of Colorado. Based on the 5’-silyl-2’-orthoester protecting group, this new RNA chemistry was instrumental in the first large-scale (for RNA), high-yield synthesis of pure RNA under mild conditions.

So when RNAi burst onto the scene in the late 1990s, Dharmacon was ready. At first the company merely sold synthetic siRNAs to customer specifications. Today, the company provides a “canned” solution for gene knockout work. Customers provide the gene, and Dharmacon sells them a “smart pool” of siRNA molecules guaranteed to silence that gene. The company also develops RNAi drugs with development partners.

Dharmacon has founded, with non-profit academic groups, the RNA Global Initiative, which conducts genome-wide screens for gene functions. Central to this effort is a project to characterize the effects of chemotherapy one gene at a time. Marshall calls this “probably one of the biggest events in biology.” The “druggable genome” consists of only 5,000 to 7,000 genes out of the 22,000 known human genes. “We don’t really know what the other genes are doing or if they’re druggable,” he says. Knowing if knockdown of a gene intensifies or inhibits chemotherapy could change how cancer is treated while providing diagnostics for predicting patient outcomes.

Some members of the Initiative have achieved “absolutely fascinating” results, Marshall says. “Pharmaceutical early adopters have small-molecule drugs entering the clinic this year based on genes identified and validated using this technology in 2003.”

What’s next in store for RNA? Marshall mentioned glycomics and metabolomics — the contribution of carbohydrates and other small-molecule cofactors, respectively, to RNA function. Marshall also identified RNA that binds directly to proteins, and elucidation of multiple antisense transcripts operating simultaneously to regulate genes through some sort of RNAi-based mechanism. “RNA opens op another layer of things to consider,” Marshall mused. “It’s like subatomic physics. As we gain the capability of finding smaller and smaller particles we do it, and we keep finding the next layer. I think a similar process is happening in all of science."