A few small changes to the structure of an enzyme can significantly increase that enzyme's ability to withstand high temperatures. This could one day make these otherwise fragile proteins that catalyze chemical reactions useful for a variety of industrial and research applications, from cancer treatments to environmental cleanup.
The trick is deciding which of the billions of possible changes will make the enzymes more heat-resistant without affecting their function.
Using a novel approach, researchers in Dr. Barry Stoddard's lab in the Basic Sciences Division, collaborating with scientists at the University of Washington and Washington State University, have demonstrated the effectiveness of a computational method for modifying an enzyme to remain stable at higher temperatures while retaining natural function. Another test showed that bacteria depending on the enzyme for life at elevated temperatures fared better with the modified enzyme than those with the original. The researchers reported their results in the May 6 issue of the journal Science. They plan next to use their technique to improve the stability and function of other enzymes used commonly in lab and industrial processes.
Improving an enzyme's ability to withstand heat or other stressful conditions (such as in detergents or organic solvents) is of great potential value to a variety of fields, ranging from basic science to bioremediation. The researchers believe that their technique should be applicable to designing any enzyme with a known protein structure.
"The most important aspect of the research is the demonstration that a computational algorithm can accomplish the stabilization of an enzyme," Stoddard said. "This is a way of accomplishing something that is of great general interest to biotechnology."
Among other possible applications, Stoddard notes that enhanced enzymes designed by computer might eventually be used to break down pollutants or potential biowarfare agents. Improved thermostability could also theoretically be used to enhance protein therapy agents used to fight cancer.
To test the technique for such future applications, the researchers focused their research on an enzyme with a well-determined structure. The structure for the enzyme (yeast cytosine deaminase) had been worked out in an earlier study by Stoddard's lab, using a technique called x-ray crystallography.
Next, using a software program called Rosetta that was developed by collaborator Dr. David Baker at the UW, the researchers pinpointed small changes to just three of the enzyme's 158 amino acids. As it turned out, these small changes were enough to make the modified enzyme stand up to temperatures higher than normal by 10º C.
"We tried to be as surgical as possible," said UW graduate student Aaron Korkegian, who carried out the experiments in Stoddard's lab. "We didn't want to have some sort of large-scale change because a large-scale change probably would have translated to loss of activity."
An enzyme's activity depends on its specific three-dimensional shape. Its shape in turn depends on its linear sequence of amino acid building blocks and how this linear sequence folds into three-dimensional space. Because an enzyme needs to have a specific shape to be active in speeding or otherwise controlling a reaction, certain changes in the sequence of amino acids can result in an inactive enzyme — like a bent key no longer able to open its lock.
Meanwhile, an unstable enzyme is one that is prone to unfolding or falling apart under certain conditions such as raised temperatures.
Maintaining the activity or function of the enzyme while improving stability was the main goal of the investigation, Korkegian said. "We were testing if these protocols could be used to not only thermostabilize proteins — which had been shown before — but to thermostabilize an enzyme without losing the activity, which hadn't been shown before."
Although the computational technique is limited to enzymes with previously determined structures and functions, the method offers great potential for improving enzymes much more efficiently than has been possible in the past.
Advantages over existing techniques
The existing method for improving an enzyme, called directed evolution, involves a laborious and expensive lab procedure of introducing a vast number of random mutations and screening them for success. Through successive applications of this technique, one eventually obtains an improved enzyme. However, the process is time-consuming, may require high-throughput lab equipment, and depends upon chance to produce results.
In contrast, the computational method is much more directed. "What our computational design allows us to do is to simulate evolution," Korkegian said. The computer program narrows the possibilities to a few, he said, "so that instead of billions of things we're trying, we're trying four or five things."
While it's true that the new method requires foreknowledge of an enzyme's structure, Stoddard said this shouldn't prevent researchers from using it to improve potentially useful enzymes. "Any enzyme that people are interested enough in, they will know or be working on the structure," he said.
Applying the technique
Having shown the value of the computational method for improving their model enzyme, the researchers now plan to collaborate with the Massachusetts company New England Biolabs to explore improvements for some of the company's commercial enzymes.
"What's next is to take a few enzymes used for fairly common lab purposes and to make them more active and stable at higher temperatures," Stoddard said.
In some cases, they will try to stabilize enzymes that may be of use but have so far been unmarketable because they break down too easily.
Stoddard and Baker also aim to explore applications for the computational method other than stabilization. They plan to direct research toward the computational design of active sites for reactions without known enzymes and the design of binding interfaces between proteins and nucleic acids.
For now, though, the demonstration using the model enzyme serves as an elegant example of the technique's potential for streamlining enzyme design. As Korkegian points out, the computer chose three changes that improved thermostability for this enzyme on its first try. "We didn't pick this enzyme because we knew it was going to work. And yet it did — and pretty quickly."