BIG DATA

We are approaching the end of peak oil production, after which the rate of global petroleum extraction from the earth will steadily decline. As we pass this threshold, the fossil fuels that have enabled us to cheaply drive our cars, heat our homes, build our towns and power our factories, will become increasingly expensive, hampering growth and radically changing our way of life.

This is reason enough to develop new energy sources, but there are other reasons to find alternatives to oil, coal, and natural gas, notably, the role of fossil fuels in global warming, and the fact that their use makes the U.S. reliant on unstable nations for our supply of crude oil.

While there are alternatives on the horizon, these options are either a long way from mainstream use (hydrogen power); not easily converted into transportation fuel (solar, wind or hydroelectric energy); or detrimental to food prices and supply (ethanol made from food crops).

Many scientists believe the best hope for the immediate future lies in biomass — organic matter that can be converted into ethanol. Whether in the form of crop and forestry by-products like corn stover, wheat straw, switch-grass or sawdust, biomass is plentiful, low-cost, and sustainable. However, ethanol currently produced from biomass is not cost-competitive with fossil fuels, which makes it a hard sell to consumers.

A major effort is underway among private companies and government research facilities to improve the efficiency of biomass ethanol, and thus reduce the price.

This image shows the binding module of this enzyme on a flat (1,0,0) crystalline surface of cellulose. Molecular dynamics simulations showed that this domain undergoes a rearrangement when introduced to the cellulose surface. This indicates that the domain has a mechanism for recognizing these cellulose surfaces.

“The current process of making ethanol from biomass is estimated to cost $2.61 a gallon and we’d like to reduce that cost down to $1.49 a gallon,” said Mark Nimlos, senior scientist at the National Renewable Energy Lab (NREL) in Golden, Colorado. “The research that we’re doing on Ranger at the Texas Advanced Computing Center is trying to make the enzymes that break down cellulose work faster. If we could do that, biomass ethanol would be cost competitive with other sources of transportation fuel.”

Producing ethanol from biomass is expensive because the enzymes that break down plant matter aren’t very efficient. Plants have cellulose walls that are difficult to digest, making them unattractive food and energy sources to most organisms.

And yet, some bacteria and fungi secrete enzymes that are perfectly capable of turning cellulose into energy. In the stomachs of cows and termites, for example, organisms exist that are experts at extracting nutrients from grass and wood. Among the most prodigious deconstructors of cellulose is fungal matter, like Trichoderma reesei, which has been at the center of much academic and industrial scrutiny. However, even T. reesei can’t convert biomass fast enough to compete with oil.

“We’re trying to make these enzymes more efficient, but we don’t know a lot about how they work,” Nimlos said. “It’s difficult to determine the mechanism experimentally by which these enzymes break down sugars because they’re so small and they work so slowly. But one of the tools that can help in developing an understanding of how they work is computational science.”

Nimlos is involved in a computational effort aimed at understanding aspects of the biomass problem from the molecular level up as part of one of the most advanced groups in the world working on the problem of biofuel production.

On Ranger, Nimlos is performing molecular dynamic simulations that imitate the behavior of a cellulase, the enzyme that breaks down cellulose as it turns plant matter into sugar, the precursor of ethanol.

The results from these simulations are fed to the experimental group at NREL who alter the amino acid sequence of the organism through selective mutation, Nimlos explained. “We do the computation and then another group does an experiment to validate our results, or vice versa. It’s a very powerful approach to trying to make improvements to the enzymes.”

By fully understanding the hydrolysis process, it may be possible to engineer super-efficient enzymes that can create a substantial amount of energy from what are essentially agricultural waste products.

In the last decades, scientists have discovered that the cellulase acts like a “protein machine,” breaking bonds in a processional manner, pulling out the strands, binding, and breaking the bonds in assembly-line-style. However, researchers realized that somewhere in the process there is a bottleneck — a slow step that limits the effectiveness and cost-efficiency of the transformation. Uncovering that bottleneck, and selectively replacing a few amino acids to speed up the bond-breaking, lies at the heart of NREL’s effort.

On Ranger, Nimlos applies CHARMM molecular dynamics software to simulate 20,000 to 50,000 atoms (a subset of the complete one-million-atom system) interacting in tandem to break down cellulose. By repeatedly calculating the force of each atom acting on the others, a picture emerges of the protein machine in action.

“We have some evidence that the protein actually changes when it binds to the surface,” a phenomenon known as induced fit docking, said Nimlos. “We think that may be an important part of how the enzyme works, in terms of being selective for cellulose and binding strongly enough to help disrupt the structure of the cellulose. “

Once researchers know the energetics of the systems, they will be able to discover what’s limiting the reaction rate of these enzymes. “Is it pulling the chain out of the surface? Is it feeding it to the catalytic domain? Is it the enzyme absorbing onto the surface?” Nimlos asked. “We need to understand what the rate-limiting step is so that we can try to make changes to the enzyme to speed up that particular step. That’s the biggest hurdle.”

The crucial bottleneck is identified not only as a stage in the molecular process, but is recognized in terms of specific amino acids that can be altered to create a new, super-efficient enzyme.

A picture of the entire cellulase enzyme called cellobiohydrolase I (CBH I) from the fungus Trichoderma reesei on top of a cellulose fibril. Molecular dynamics simulations of this enzyme complex on a large cellulose microfibril in a box of water were conducted. These simulations contained nearly 1 million atoms and as such represents one of the largest atomistic molecular dynamics simulations to date.

“We need to see how specific amino acids in the binding module affect the energetics of binding, and we can do that with computational modeling and then back it up with experiments,” Nimlos said. “Ranger allowed us to do a lot of simulations of this binding module at the catalytic surface, and we’ve already been able to map some of the energetics of that process.” [Some of these discoveries were reported in Cellulose magazine in 2008; see citation below.]

Nimlos’ computational simulations are part of a larger puzzle. In fact, many members of the biomass team, including John Brady at Cornell University, and Mike Crowley, Gregg Beckham, Yannick Bomble, Lintao Bu, James Matthew and Michael Himmel at NREL, are using Ranger to explore different aspects of the nanoscale behavior of the T. reesei.

With oil prices predicted to rise and a climate change disaster looming on the horizon, the pace of investment and discovery in alternative energy solutions is quickening, Nimlos says. He believes that in the near future, his group will fully understand the T. reesei enzymes, and know how to change it to make ethanol production more efficient and economical.

“We have a milestone for the program — to make some predictions that lead to an improvement in the next year or two,” Nimlos said. By 2030, the Department of Energy expects to derive 30 percent of transportation fuel from biomass, which means increasing the resources — both computational and human —devoted to solving the problem.

As in many fields of science, supercomputers have a crucial role to play in the race to a solution. By simulating what cannot be seen, mimicking the behavior of the world’s smallest biological systems, HPC plays a significant role in 21st century science.

“It’s in this country’s best interests to move away from imported oil for a lot of reasons — security, global climate change, economics — and I think there’s great potential for producing fuels from the cellulosic part of the biomass,” Nimlos said. “Furthermore, the Department of Energy has shown that they can make this displacement without affecting food prices, without affecting water usage, without adversely affecting the environment. This is something that our country needs to do to improve our security and our environmental impact and make a more sustainable transportation future.”

******************************************************************************************************************************

To learn more about biomass engineering at the National Renewable Energy Laboratory, visit the NREL research page.

Or read Zhong, L. Walker, R.C, Brady, J.W. et al. "Interactions of the Complete Cellobiohydrolase I from Trichodera reesei with Microcrystalline Cellulose IB.", Cellulose. 2008, 15, 261-273 (Full Text PDF)