Small-scale enzyme research could have big biomass implications

By Kris Bevill | January 28, 2011

Biofuels goals are more often than not measured by the millions of gallons, but some of the biggest accomplishments are first achieved on a much smaller scale. Researchers at the U.S. DOE’s Great Lakes Bioenergy Research Center, based at the University of Wisconsin and Michigan State University, are three years into a research project focused on identifying new enzymes for cellulosic ethanol production and creating enzyme cocktails that will produce more ethanol at lower costs than what is currently possible.

 Jonathan Walton, a professor of plant biology at Michigan State University and associate director of MSU’s activities at the research center, is leading the team of researchers. In order to more quickly identify ideal enzymes, the team first developed an automatic process that can efficiently evaluate various enzymes and determine what proportions of enzymes combine to produce the most sugars from biomass samples such as corn stover, switchgrass, miscanthus and distillers grains. The machine, called the GLBRC Enzyme Platform, or GENPLAT, is vastly more efficient than its human alternative and is capable of determining ideal enzyme proportions within 48 hours of receiving the sample. Up to 96 glucose assays can be calculated in one fell swoop using the technology developed by Walton’s research team, far more efficiently than by human researchers, Walton said. Considering the thousands of yet un-researched proteins that could be valuable to ethanol production, GENPLAT’s ability to process information could rapidly speed the rate of discovery and help researchers meet their goal of reducing enzyme requirements by ten-fold in less time than would otherwise be possible.

For now, research is conducted at a milligram-scale. Walton’s group starts by adding combinations of pure, characterized enzymes and then works toward more complex mixtures including unknown proteins from many microbial sources. Enzyme combinations are optimized for different feedstocks and different pretreatments. “We start with our core set and systematically add additional components,” he said. “From there, it’s a process of elimination. If a new component doesn’t make a difference, which most don’t, it gets set aside. If it does improve the process it gets added to the core set and becomes the platform to find additional enzymes.” It’s a painstaking process, but small improvements can quickly add up. Walton said the team has achieved glucose yields of 85 percent with their synthetic mixtures, but their ultimate goal is to get the same yield using ten-fold less protein. This would make the cost of enzymes for lignocellulosic biomass conversion economically realistic.

In order to accomplish this, the team’s short-term goal is to determine which enzymes are important for ethanol production and which can be left out of the equation. “For example, we can take 18 pure enzymes and feed them into GENPLAT and the answer we get back is for optimal glucose you need ‘x’ percentage of this enzyme and ‘x’ percentage of that enzyme,” he said. “Sometimes we get back the answer that an enzyme makes no contribution whatsoever and can be left out. And that’s helpful. Right now, the commercial enzymes available for biomass deconstruction are basically one-size-fits-all. What we’re trying to do is make more custom cocktails that would be suitable and optimal for very specific substrates. With GENPLAT, we’re able to figure out what enzymes you need for any particular biomass substrate that might be coming in the door of an ethanol plant.”

In the future, as more ethanol facilities become multi-feedstock-friendly, the need for specialized enzyme cocktails will become even more valuable. Walton said he envisions this research being scaled up eventually to a point where an ethanol facility can receive a shipment of biomass and, two days later, have data from a GENPLAT system that details exactly what enzymes are required to produce the maximum amount of sugars.

Producing individual pure enzymes for ethanol production is not cost effective, of course, but Walton said once the individual enzymes are identified, which is the focus of his team’s research, fungi could be engineered to produce a compromise of one to three of the enzymes of interest in large batches. “It’s a matter of time,” he said. “We have all the tools to engineer fungi to have them make whatever we feel like having them make. If you had 12 fungi each making mainly a single protein, you could package that and ship it to the factory and make a blend right there on the spot.” Much more research and development needs to be conducted before the industry gets to that point, he said. Research at GLBRC is currently being conducted at the milligram-scale, “so we have to scale that up about a billion-fold if we want to have an impact on ethanol production in the country,” Walton said.