Inteins aren't what you'd call “helping enzymes.” In other words, they don't assist other proteins in the reactions that transform them from their primary states into the protein complexes that enable them to perform their unique biological functions. Instead, they remove themselves entirely from the proteins of which they are components–and then splice the remaining parts together to form a whole molecule.

“You can think of a protein as a kind of ribbon,” says Phil Shemella, a graduate student in the department of physics at Rensselaer Polytechnic Institute. “At two locations, the ribbon is cut, and the middle–the intein–falls away.” These inteins are autocatalytic–they already contain the enzyme necessary to accomplish the cleaving and splicing that forms the new protein structure.

“We sometimes call them 'selfish proteins,' or 'selfish DNA,'” says Saroj Nayak, an assistant professor of physics at RPI. Nayak is Shemella's advisor and a principal investigator in a multidiscplinary investigation into why inteins behave the way they do. “Inteins are by themselves–they don't care about the rest of the physical area, they don't need anything else, they're entirely self-contained.”

Splicing and cleaving

Inteins have been observed in some 100 proteins found in around 30 single-celled organisms, including some yeast, bacteria, some thermophilic microorganisms often found around deep-sea vents, and human pathogens–including the bacterium that causes tuberculosis. In fact, current knowledge about inteins has already been exploited in researching new antibiotic treatments for the widespread communicable disease. Nayak, Shemella, and their collaborators at RPI and the New York State Department of Health's Wadsworth Center are working toward the creation of a nanoswitch that would be useful for turning biological processes on and off.

However, not much is really known about the exact mechanism by which the intein excises itself–a process researchers refer to as “splicing and cleaving.” It is known that the intein breaks the protein up into groups of around fifteen amino acids each. But which amino acids are responsible for splicing the protein remnants back together is unclear.

“There are a lot of arguments about why certain amino acids are shared at certain locations for proteins to have the splicing behavior,” says Shemella, whose own work focuses on examining one specific area of one particular amino acid grouping, or motif. “From the literature–that's the best understanding we have of the mechanism–we know that the conserved residues must be important for cleaving. We know that we can alter three amino acids and prevent the cutting from occurring on the left-hand side, and we know that the process speeds up at lower pH levels, but we don't know why.”

To answer this question, Nayak, Shemella, and their collaborators, Georges Belfort and Shekhar Garde of the chemical and biological engineering department at RPI and Marlene Belfort and Vicky Derbyshire, molecular biologists at the Wadsworth Center, are using molecular scale modeling on NCSA's Tungsten to simulate intein behavior and the environment that would trigger splicing and cleaving. Their quantum mechanical/ molecular mechanical (QMMM) method uses the Gaussian03 code to combine quantum mechanics with classical molecular dynamics; the quantum mechanics component allows them to begin the simulation ab initio, without requiring experimental parameters.

Inside the protein molecule

During the simulation, Nayak and Shemella run calculations on the entire molecule and the solution surrounding it (water) using classical molecular dynamics methods, while they run calculations on the reactions within the molecule itself using quantum mechanical methods. It's important for studying this reaction, because at the same time that the dynamics is going on there's a chemical reaction going on, so you have to have both quantum mechanics and molecular mechanics,” explains Nayak. “Ideally, you should use quantum mechanics for the complete system, but practically that's not possible, because the system's too large. But you need the system to be too large, because you want to give the proteins all the freedom that they need, and you want to mimic reality as much as possible.” The size of a typical run, according to Nayak, is around 6,000 to 7,000 atoms.

Gaussian03 runs on anywhere from one to eight processors at a time for a total of around 40 hours, and Shemella and Nayak have been pleased with the overall performance. “Tungten has been pretty nice, pretty fast,” Shemella says. “It's well-maintained, too.” To ensure that their simulations run smoothly, they have been in close contact with Dodi Heryadi of NCSA's User Support and Consulting Group. “NCSA does a really excellent job in supporting whatever we need,” says Nayak. “It's really remarkable that they take the time to get the software, to help us set up. I don't think it would have been possible to do this in our home institution–the infrastructure and the time it would have taken just would not have been possible at our university, or at many universities.”

Using a snapshot of the protein structures from the molecular dynamics simulation performed by Garde's group, the collaborators ascertain what the possible key aspects (for example, temperature, acidity, etc.) might be that trigger the intein's mechanism, alter them, and run more simulations accordingly. They have narrowed the field down to five possible mechanisms. “That's where the fun begins,” says Nayak. “We can take our results to the experimentalists [Belfort at RPI and Belfort and Derbyshire at Wadworth], and they can do experiments involving the mutations we've predicted. And if we have a two-way cross-check, you have more confidence of things going the way they're supposed to.”

The multidisciplinary nature of the collaboration to investigate intein mechanisms brings a wide variety of scientific goals into synchronicity. The combined goal, from the standpoint of molecular biology, chemical engineering, and physics, is to exploit the intein to create molecular nanoswitches that respond to stimuli such as light and heat for use in biotechnology. But Shemella identifies an even more basic goal. “Any time you can figure out what makes a mechanism work, why it works, how it works — that's important.”