Putting evolution to use

Biosensors exploit the natural tendency of proteins to recognize and react to other molecules. Ron Breaker theorized that the RNA in proteins could be isolated and serve as a “switch” that targets specific molecules. The central image at left represents a binary RNA switch as it goes through the chemical machinations necessary to recognize a molecule—it must bind to two different molecules before the ribozyme can cleave. Biosensors could be used to diagnose patients, monitor for bioterror weapons and detect metabolites in clinical samples, contaminants in food or pollutants in water.

When Ronald R. Breaker, Ph.D., associate professor of molecular, cellular and developmental biology, wanted to develop a high-tech tool for detecting everything from infectious agents like HIV in the blood to contaminants such as arsenic in water, he figured that the place to start was about 3.5 billion years ago, when life began.

Some have theorized that RNA, the active component in genes, and not DNA, the genetic library, composed the first life forms. Based on that theory, Breaker supposed that RNA would have needed to act like a simple switch when it came into contact with another molecule, sending out signals in order to control metabolism. He and his team put that evolutionary theory to work in a test tube to “back-engineer” RNA-based molecular switches. As he reported in a paper published in Nature Biotechnology in April 2001, the switches worked. In fact, they worked so well that some believe they could be key to one day developing small, easy-to-use biosensors for detecting everything from tumor cells to toxic chemicals.

RNA molecular switches could be used to construct a kind of dashboard panel to detect when sought-after molecules are present. Such biosensors might be developed to detect contaminants in food, pollutants in water, metabolites in clinical samples or biological-warfare agents on the battlefield.

Yale licensed the RNA molecular-switch technology to a Cambridge, Mass., biotechnology company that Breaker helped found, Archemix Inc., which is developing possible commercial applications of the biosensor technology. In his own laboratory, he is continuing to explore the science behind molecular switches. Contributing Editor Marc Wortman asked Breaker to explain how biosensors work and to discuss the challenges of developing the technology.

What is a biosensor?

A biosensor is a device that uses biological molecules (usually proteins) to detect chemicals or other biological molecules. Each of our cells holds a complex mixture of chemical compounds, and all of these compounds are created and used in a highly orchestrated fashion. To do this, our cells mostly use intricately folded proteins to selectively recognize many of these compounds and to make use of them in very specific ways. Scientists found that they could remove one or more of these proteins from their natural setting and use them as biological components of sensor devices.

Perhaps the best-known target for a biosensor is glucose. For many diabetics, frequent monitoring of the concentration of glucose in their blood is necessary. However, glucose is only one type of molecule within the very complex mixture of chemical and biological compounds present in blood. One way to “see” glucose in this vast sea of other compounds is to use an enzyme called glucose oxidase. This enzyme can identify glucose among all the other compounds and cause it to react with oxygen to produce hydrogen peroxide. Then a second enzyme called catalase destroys hydrogen peroxide in a process that can be monitored electronically. Higher levels of glucose in a drop of blood will give higher electronic signals, and thus one can determine precisely how much glucose is present.

How difficult is it to detect the presence of a potentially harmful biological agent, such as anthrax or Salmonella, in the water or air or on the surface of things?

Detecting infectious agents such as anthrax, Salmonella, or HIV is very difficult, particularly if the contamination is low and if you want the results fast. The challenge largely centers on the number of targets you want to see. Glucose concentrations are typically measured in “millimolar” units, which means that there are more than a billion trillion glucose molecules in a pint of blood. There are so many glucose molecules that it is quite easy to detect and accurately measure them. However, HIV concentrations could be present in “zeptomolar” units, where only a few hundred virus particles might be present but still able to cause disease. Biological agents such as bacteria and viruses can make billions of copies from a single infectious particle. There might be only a few HIV particles in a pint of blood, but they need to be detected in order to ensure that donated blood is safe for transfusion.

What are the usual methods?

Detection of biological infections has routinely been achieved by growing cultures from samples taken from patients, or by looking for the production of antibodies.

More recently, scientists have exploited two technologies—enzyme-linked immunosorbent assays (ELISA) and polymerase chain reaction (PCR)—to expand the power of biosensors and bioanalytical methods.

Sometimes natural proteins are not available for a particular target; interestingly, the immune systems of animals can be used to produce new antibodies that selectively bind to new targets of interest. These antibodies are typically used in ELISA tests.

More recently, PCR tests have become commonplace. These tests rely on the power of DNA-making enzymes (DNA polymerase) to selectively amplify genetic fragments of infectious agents such as anthrax and HIV. DNA fragments of just a few infectious particles in a sample are used in PCR to make billions of copies, which can then be observed by any one of several different methods.

What are the problems or limitations of those current methods?

Speed, shelf life, and cost are of significant concern with all existing methods. ELISA and PCR assays take time to run and are usually labor intensive. Automation of some aspects of these tests can reduce the time needed to set up and interpret multiple tests, but the biochemical processes themselves also take time. Antibodies and many other proteins are notoriously unstable, and most biosensor kits have a shelf life that is measured in months—far too short to be of use to most consumers. Finally, proteins can be expensive to produce and store, which drives up the cost of making most biosensors.

Given concerns about bioterrorism, what efforts are under way in your laboratory and elsewhere to develop better biosensors?

Several science funding agencies had embarked on efforts to accelerate biosensor development even before recent events brought attention to bioterrorism. I think that this was a recognition that improvements were needed in detection technology, and also that recent advances in biological research indeed make possible great advances in biosensor sophistication.

Our laboratory took a rather unusual path to enter the area of biosensor research. We were testing a theory of how life began some 3.5 billion years ago. It is believed that an “RNA world” once existed, where all enzymes and other molecular components of primitive cells were made of ribonucleic acid, or RNA. Although this entire way of life has long since become extinct, we can perhaps use evolution in a test tube to recreate many of these long-lost RNA molecules. Through these efforts, we invented RNA switches that can be used as biosensor elements. For example, we have made many types of RNAs that self-destruct only when they come in contact with a specific target molecule. We recently assembled these on a prototype RNA biochip that can be used to detect toxic metals such as cobalt, drug compounds such as theophylline and natural compounds such as cyclic AMP and cyclic GMP.

What types of applications could they have?

We expect that this new type of biological switch will be used to make next-generation biosensor devices that detect a variety of chemical and biological agents in a single assay. Scientists have recently developed “gene chips” that can be used to see thousands of genes on one miniature platform. We imagine similar platforms that see genes, metabolites, drugs, toxins, biohazards and any other targets of interest all in a single assay.

Biosensor technology of this advanced type could be used to diagnose patients in a doctor’s office, help discover new treatments for disease, detect industrial contaminations and even aid in monitoring for chemical or biological attacks.

Are RNA molecules stable enough to serve as biosensor elements?

Nature has chosen DNA to store genetic information because each link in the chain is extraordinarily stable. In contrast, each link in RNA’s chain is about 100,000 times more likely to break. If this occurs in one of our switches (which are about 100 links in size), then its function is likely to be destroyed. Fortunately, RNA is sufficiently stable to provide a shelf life of several years. After immobilization, the RNA array can be stored for long periods of time simply by letting the surface air dry and storing it at room temperature. Of course this is assuming that the RNA is not being digested by contaminating enzymes. Even the oils covering our skin carry enzymes that can degrade RNA very rapidly. Again, we are fortunate because chemists have created many modified RNA links that resist destruction by nucleases. In other words, the science to create modified RNA switches that resist attack by chemicals and enzymes is already well developed.

Will we eventually have low-cost, easy-to-use biosensors in our homes and workplaces?

Without question. Glucose tests and pregnancy tests have already become routinely self-administered. If the technology continues to advance, I imagine that home diagnostic tests for diseases such as cancer and viral infections might become the first line of defense for health care.

What might they look like and how would they be used?

Perhaps not surprisingly, we need to look at our science fiction stories to give us some sense of the possibilities. Although biosensors of the near future might look more like home pregnancy tests than a tricorder from Star Trek, I think that handheld biosensor devices that can detect thousands of important targets are being envisioned by many in the field. They will contain some sophisticated electronics and some savvy computer algorithms for interpreting various signals and perhaps for giving courses of action to the user. But at their core, they are likely to contain an advanced form of biochip that is engineered to recognize and report the presence of thousands of targets. YM