Issue Date: December 3, 2012
To Make Better Biofuels, Scientists Mine Microbiomes
Ya Ya and Le Le are a couple of giant pandas at the Memphis Zoo who don’t realize they are participating in an important science experiment. Once in a while a person will come into their space at the zoo, grab their poo, and scurry away. Ya Ya and Le Le shrug off the intrusions and go back to what they do best: eating bamboo.
The results coming from their backsides could benefit the pandas. Scientists studying the microbial community in the animal’s gastrointestinal tract are gaining a better understanding of panda nutrition, which could help improve the health and reproduction of the lovable and endangered creatures.
But people have still another motive for these panda poo investigations. They want to figure out which symbiotic microbes living in the panda gut are deconstructing the tough, fibrous lignocellulose that makes up the plant cell walls of bamboo. By analyzing DNA they glean from the stool samples, scientists are working backward to identify the microbes and pinpoint genes that code for their cellulose-busting enzymes.
And researchers aren’t stopping at pandas: They are searching for such enzymes in a variety of vegetarian animals. Ultimately, they want to cut and paste the genetic blueprints for such enzymes into other microorganisms, giving the engineered microbes the ability to break apart biomass into soluble sugars that can be fermented into biofuels.
It remains to be seen whether such technologies will eventually lead to an economically viable cellulosic biofuel market. But those challenges aren’t stopping biochemist Ashli Brown and members of her Mississippi State University research group. They watch Ya Ya and Le Le eat and wait for them to poo. When the time comes, they alert Memphis zookeepers to move in for a search and recover operation.
And it’s not such a dirty job. “Panda pellets are fairly desiccated and fibrous,” Brown says. “They are like mini hay bales. For anyone with experience working with animal fecal material, I can assure you panda poo has a fairly pleasant smell and is probably the nicest to work with.”
To find useful enzymes in the pellets, Candace L. Williams in Brown’s group takes them, grinds them up, performs extractions, and then uses traditional anaerobic bacterial culture methods to grow and identify bacterial colonies. The microbes found clinging to incompletely digested bamboo in panda poo can be difficult to cultivate in a lab, like any microbe when it’s out of its element. Brown’s group inevitably misses a few of them.
To circumvent that problem, Brown’s team has added another tool to its arsenal: metagenomics. This relatively new DNA sequencing approach can reveal the diverse array of microbes that live in a given environment, such as the panda gut. With metagenomics, scientists can take a census of all of an environment’s microbes—collectively known as a microbiome—without having to culture them.
The researchers chemically dice up the DNA extracted from the panda poo into pieces of about 100 to 600 base pairs and sequence each piece, Williams explains. They then take the thousands of short sequences generated from the microbiome, align them to find overlapping regions, and piece them together to isolate the 16S ribosomal RNA genes. This gene codes for a component of the ribosome, the machinery used to build all the proteins in a cell. It has a relatively short sequence—about 1,500 base pairs long—and each microbial species has a unique one.
The team then turns to computational algorithms to sort through the billions of base pairs sequenced to match the 16S rRNA and all the other genes in the microbiome to the individual microbes. They also use publicly available bioinformatics database tools to compare microbial diversity between multiple microbiomes to determine what all the different microbes in the panda gut do and which enzymes they produce.
In the end, because the sequencing process and data analysis are not 100% accurate, the Mississippi State team is validating its metagenomics data by using the polymerase chain reaction to amplify the microbial cellulase genes they discover in panda poo.
The Mississippi State team is not alone in its quest. By using a combination of metagenomics, whole-genome sequencing, enzyme activity assays, and sugar and fatty acid composition analyses, scientists have been profiling the gut microbiomes of a variety of animals to begin to understand how they degrade plant biomass, says biophysicist Susannah G. Tringe, head of metagenomics at the Department of Energy’s Joint Genome Institute (JGI).
“We now have a pretty standard approach to metagenomics projects,” Tringe says. “DNA is a common currency no matter where it comes from, allowing us to study microbiomes in soil, water, plants, or animals, including humans.”
Animal metagenomics studies like that of the panda serve as a foundation for biofuel development, Tringe says. “The Human Genome Project, for example, hasn’t actually cured any diseases itself. But it has given scientists a much faster way to understand the genetics of individual diseases and potential treatments. Now that we are getting cellulosic enzyme metagenomics data, it’s creating opportunities for a variety of downstream studies.”
Like pandas, cows are unwitting participants in these gut metagenomics projects. Enzymes in the cow’s rumen, the primary grass-digesting chamber in the cow’s multistomach GI tract, allow the ruminants to process massive amounts of grass. In 2011, Tringe was part of a multi-institution team that conducted a metagenomics study of the cow rumen (Science, DOI: 10.1126/science.1200387). The researchers took 268 billion bases they sequenced and whittled them down to 27,755 genes that encode for carbohydrate-active enzymes. They ultimately identified 51 enzymes with cellulose-degrading activity and are now testing them for their catalytic abilities.
Another animal with potentially useful cellulase enzymes is the hoatzin, a chicken-sized, reddish-brown bird with a spiky crest that hails from South America. The hoatzin is unusual among birds in that it primarily eats leaves. Like other birds it has a crop, a built-in storage bin that is part of the digestive system—a kind of prestomach. But in the hoatzin, the crop functions as a biomass fermenter, similar to the cow’s rumen. In fact, the hoatzin is nicknamed the stink bird because its droppings smell like cow manure.
Studies on the hoatzin microbiome, led by Filipa Godoy-Vitorino and María G. Domínguez-Bello of the University of Puerto Rico, Río Piedras, detected hundreds of microbes. The researchers also carried out a comparative analysis of the hoatzin and cow microbiomes.
“Despite being phylogenetically distant, adult hoatzins and pasturing cows share specific digestive microbes,” says Godoy-Vitorino, who is now at JGI. After classifying the microbial species, the researchers found seven that were unique to the hoatzin, she says. They are now classifying the hoatzin cellulose-degrading enzymes.
And then there are pandas. “If you can use a microbe that efficiently breaks down biomass naturally and as efficiently as pandas do, and convert them into sugars that could be converted to oils and other chemicals, production costs for alternative fuels would be cut tremendously,” Mississippi State’s Brown says. “That would be fantastic.”
Pandas have to be efficient at breaking down biomass—they eat as much as 30 lb of bamboo each day, and not much else, Brown notes. Bamboo makes up a whopping 99% of the giant panda’s diet in the wild. The creatures have unusual adaptations to accommodate, she says. For one, pandas have a pseudothumb that allows them to grab bamboo and strip away leaves and stems. They also have well-developed teeth, jaws, and chewing muscles, designed more for masticating like a cow than ripping meat like a carnivore.
But the panda’s digestive system isn’t built like a cow’s, Brown continues. Whereas cows and other ruminants typically have a long GI tract with a multichambered stomach system designed to digest grasses, she says, the panda sports a short GI tract and a single-chambered stomach. The combination of bamboo gluttony and short GI tract means the panda processes bamboo quickly, in only about four hours from front to rear; in cows, it’s a one- to three-day process.
Brown’s research has been guided by the panda genome, which was sequenced and reported in 2009 by an international consortium led by scientists in China (Nature, DOI: 10.1038/nature08696). The Chinese team assembled the genome of Jingjing, a panda at the Beijing Zoo named after one of the mascots of the 2008 Olympic Games.
The study revealed a genetic basis for the panda’s love of bamboo. It turns out pandas have mutations resulting in loss of function of the gene that codes for umami taste receptors, which recognize glutamate, Brown notes. These mutations appear to have robbed pandas of the ability to savor the taste of meat and other high-protein foods, she explains, pushing them toward their diet of readily available bamboo.
Regardless of the panda’s food preferences, it is genetically speaking a carnivore, Brown says. Bamboo-loving pandas and dogs, the only other carnivore sequenced to date, have 99% of their genes in common.
“What we took away from the study of the panda genome and our study of the panda’s digestive microbiome is that the enzymes for bamboo digestion must all come from symbiotic microorganisms living in the gastrointestinal tract,” Brown says. “And these bacteria have to be highly specialized to efficiently break down bamboo so quickly. Pandas really are marvels in their ability to have adapted to the highly fibrous bamboo diet. It seems that the panda recruited the microorganisms based on a functional need, rather than an evolutionary process.”
As for biofuel prospects, among the hundreds of microbe species inside pandas, the Mississippi State researchers discovered 17 cellulose-degrading microbes. They are still trying to identify the exact species of them all, but some of them have also been found in other biomass-digesting animals. Brown’s team has done preliminary work culturing the newfound microbes, feeding them cellobiose, the primary disaccharide in cellulose. They are finding that some of the microbes convert the sugar to fatty acid methyl esters for energy storage.
“We see considerable cellobiose consumption and quite a mix of different triglycerides and free fatty acids forming,” she says. Those that yield primarily oils or alcohols are likely to be more amenable to producing biofuels, she adds. Her group is collaborating with Mississippi State chemical engineer Darrell Sparks Jr. and colleagues to explore the possibility of transplanting genes unearthed from the panda gut into the yeast Rhodotorula glutinis to enhance its oil-producing ability. For example, they are hoping to engineer yeast strains that can feed off waste lignin in paper mills or residual organic carbon in wastewater at treatment plants.
“There is now a ton of sequence data that’s been generated hinting at putative cellulolytic enzymes,” JGI’s Tringe says. “I am not sure how many more animal microbiomes we need to explore. The real need now is to understand enzyme functions better to move the data downstream to make them useful for industry.”
To that end, chemical engineer Trent R. Northen, director of array-based assays at the Joint BioEnergy Institute (JBEI), is leading work on methods for rapid analysis of microbiome enzymes to determine which types of bonds the enzymes are forging or breaking. JBEI is one of three DOE bioenergy centers focused on biofuel research.
Chemists traditionally use high-performance liquid chromatography often coupled to mass spectrometry for this task, Northen says. But the throughput is limited by the chromatography. JBEI has designed assays that eliminate this time-consuming step.
“We are trying to address the mismatch between the rate we can sequence genomes and the rate we can characterize gene functions,” Northen says. For this function-based screening, Northen and his team use nanostructure-initiator mass spectrometry (NIMS). Northen helped develop this technique, originally designed with biomedical research and drug discovery in mind, when he was in physical chemist Gary Siuzdak’s group at Scripps Research Institute.
In NIMS, a laser beam vaporizes initiator molecules, typically fluorinated siloxanes, which are trapped in nanometer-sized pores on a silicon surface, Northen explains. As the initiator molecules are ablated, they carry away sample molecules that have been deposited on top of the pores, desorbing the molecules for mass analysis.
Northen and his colleagues have done more than 50,000 assays on 200 glycoside hydrolases—a family of enzymes that includes cellulases—originating from JGI’s cow rumen, insect, and other metagenomics projects. The assays pit oligosaccharides found in biomass against different enzymes under different temperature, pH, and salt concentration conditions. The analysis determines how the enzymatic reactions unfold by identifying the products formed when oligosaccharide bonds are modified.
Accumulating this type of functional information in public databases will be a resource for synthetic biologists and others, Northen relates. It will enable them to look through databases, find a specific reaction catalyzed by a specific enzyme under specific conditions, download a sequence, synthesize the gene, and clone it into a microorganism to make the biofuel ingredients or other products they want, he says.
“There are tremendously exciting possibilities with this approach,” Northen continues. “We don’t yet have a predictive understanding of biology. You can’t design enzymes or biosynthetic pathways like a computer chip and have them work. We need ways of characterizing them—the equivalent of an oscilloscope or a voltmeter. That’s where these high-throughput screening tools to characterize the libraries of genes and enzymes fit in.”
But even as hundreds of researchers and millions of research dollars have been dedicated to enzyme bioprospecting, there is one thing still missing: a viable market for cellulosic biofuels. If scientists like Brown come up with improved technologies for processing the biomass, a viable market will require more, including a practical and economical source for the biomass, companies to produce biofuels, and consumer demand.
Biomass supplies are ample, according to DOE estimates. It projects that by 2030 the U.S. could produce some 1.1 billion dry tons of bioenergy crops or agricultural residues annually without upsetting food, animal feed, and fiber production—enough to replace about 30% of the country’s petroleum consumption.
With that in mind, enzyme producers, chemical and oil companies, and biofuel start-ups are forming partnerships as they pick and choose technology platforms to make cellulosic biofuels. But no companies have yet to achieve commercial-scale production (C&EN, Oct. 8, page 22).
Any new technology that emerges from animal microbiome mining would need to improve upon the proprietary enzyme systems already in companies’ arsenals. For example, enzyme company Novozymes is already marketing cellulase enzymes from Trichoderma reesei,a fungus originally discovered because it was degrading cotton military uniforms and canvas tents in the South Pacific during World War II. The company, which sequenced T. reesei with JGI, has slipped some of the fungal genes into other microorganisms to produce enzymes it is selling to a handful of cellulosic biofuel companies around the globe.
But biofuel producers are still looking for enzyme systems that can reduce or eliminate the need for harsh chemical pretreatments—acids, ammonia, or ionic liquids—currently required in cellulosic biofuel production processes. Pretreatment serves to open up the lignocellulose matrix to give enzymes more room to work. Ultimately, cocktails of 30 to 40 enzymes, tailored to accommodate different types of biomass, will be needed by biofuel producers to efficiently convert untreated biomass into fermentable sugars and on into biodiesel or biogasoline.
Scientists anticipate that the mounds of data coming out of the animal metagenomics studies will provide replacement or supplemental genes needed to improve current biofuel production systems to make them cost-competitive. But so far, neither metagenomics researchers nor enzyme producers have reached out to each other to collaborate on the gut microbiome results.
“There is potentially great synergy between what the Joint Genome Institute and academic institutes are doing to discover genes that code for cellulase enzymes and what enzyme companies are doing with expression technology to take new enzyme technology to the marketplace,” says Mark A. Emalfarb, president and chief executive officer of enzyme company Dyadic International. “But we need to do a better job helping each other bring these discoveries to market.”
Dyadic has spent 20 years developing a robust enzyme expression system based on a genetically modified soil fungus. The company has licensed its technology to other companies to make biofuels, and it’s developing its own line of biofuel cellulases.
“This industry is not for the faint of heart,” Emalfarb says. “Expression technology takes years and millions of dollars to develop. No company today produces enzymes that are cost-effective enough for economically viable production of cellulosic biofuels.”
Beyond the science and process engineering challenges, policy decisions out of the control of scientists impact consumer demand for biofuels, Emalfarb adds.
The biofuel industry faces a formidable foe in fossil fuels that are cheaper and will remain readily available for years to come. Increased federal fuel-efficiency standards will put biofuels at a further disadvantage. And although the federal Renewable Fuel Standard, which gradually increases the amount of cellulosic biofuels required to be produced each year, continues to create market demand for new enzymes, the federal government has declined to mandate that carmakers produce vehicles that can accept a range of biofuel/fossil-fuel blends.
Emalfarb suggests the cellulosic biofuel industry needs to become more engaged in the outcome of the metagenomics studies to prevent the research on pandas, cows, hoatzins, and other creatures from sitting on a shelf collecting dust.
In Memphis, oblivious to these dilemmas, Ya Ya and Le Le keep chowing down on bamboo. Mississippi State’s Brown and her team, probing pellets for more enzymatic clues, are more aware of what’s at stake.
“Our work suggests that metagenomics is useful for enzyme bioprospecting,” Brown states. “Hopefully we are helping standardize DNA sequencing technology and high-throughput screening technology to make this approach more credible. And hopefully in doing that we can make a difference for the pandas and for biofuels.”
- Chemical & Engineering News
- ISSN 0009-2347
- Copyright © American Chemical Society