Computational design used to engineer and construct a protein that could sense the nerve agent soman

Duke University Medical Center biochemists have used computational design to engineer and construct a protein that could sense the nerve agent soman. They said their achievement constitutes a proof-of principle that such engineered proteins can be made to detect nerve agents such as sarin and other toxic substances.

Such proteins could be incorporated into detectors, which might resemble smoke detectors and could be widely deployed as early-warning alarms, weapons monitors or in the decontamination process after an attack. The detector could not only warn of the presence of the nerve agent, but act as a continuous monitor of its levels.

Led by Professor of Biochemistry Homme Hellinga, Ph.D., the researchers reported their achievement in a paper published online May 17, 2004 in the Proceedings of the National Academy of Sciences. Besides Hellinga, other co-authors of the PNAS paper were Malin Allert, Shahir Rizk and Loren Looger. Their research is sponsored by the Defense Advanced Research Projects Agency.

In the PNAS paper, Helling and his colleagues described how they had designed a protein that detects a surrogate for soman, called pinacolyl methyl phosphonic acid (PMPA), which has the same basic chemical structure as soman, but is less toxic. Soman is a nerve agent first invented by the Germans before World War II and manufactured in large quantities by the former Soviet Union.

In developing the PMPA detector, the Duke researchers used the same general design technique that they had previously used to tailor proteins to sense glucose, lactate, TNT and the brain chemical serotonin.

They began with proteins, called "periplasmic binding proteins," from the gut bacterium E. coli. These proteins are normally part of the bacterium's chemical-sensing system by which it detects nutrients. Such protein receptors detect their target molecule via an "active site" that has a precise complementary shape and binding properties that fit only that molecule, called a "ligand" -- like a key fitting a lock.

Basically, the computational design process developed in Hellinga's laboratory involves redesigning the normal protein's "lock" to fit a very different molecular key. The computational process narrows down to a manageable few the vast number of possible mutations in the normal protein and their corresponding structures, to fit a particular molecule. Once the designs are narrowed down, the biochemists construct the proteins and test them for selectivity and binding properties.

"We chose PMPA because it is a commercially available surrogate of soman and is a breakdown product of the nerve agent," said Hellinga. "The design technique we used can be readily applied to any nerve agent. Also, the design challenge is quite similar to those we faced in designing proteins to detect TNT and other compounds."

Hellinga and his colleagues designed the PMPA-detecting protein not only to highly selectively bind to PMPA, but also to signal that binding by means of an attached fluorescent molecule. Thus, the protein can be incorporated into a detector that would sense a change in fluorescence of the protein as an indicator of the presence and concentration of the nerve agent.

According to Hellinga, the specificity and affinity of the PMPA detector proteins they created are sufficient for development of first-generation detectors. However, the protein must still be made more robust to function stably over long periods of time. Thus, the researchers are experimenting with corresponding proteins from thermophilic bacteria -- known for the robustness of their proteins -- that live in hot springs. The biochemists are also launching efforts to design proteins to detect other nerve agents, including sarin.

"One particularly important aspect of this computational design technique is that it can be done very rapidly," said Hellinga. "It takes at most a day to calculate a set of candidate structures and perhaps a week to construct them. So, in the event of the deployment of a new chemical threat, it might take in principle only weeks to develop a sensor system for it. We're now working to develop an automated laboratory process into a system to fabricate such proteins."

The researchers are working with Nomadics, Inc. of Stillwater, Okla. To develop sensors based on their advances, and they plan further commercialization of the design and synthesis technologies.