Fourth in an occasional series on the bioeconomy.
Down on a farm in Illinois, his forearm stuck inside the noisome gut of a living and otherwise unperturbed brown cow, Matthias Hess, a German-born microbiologist and geneticist, felt far removed from the white hum of his biology lab.
Hess had been fishing in the cow's rumen, its largest stomach, for a nylon mesh sack resembling an oversized teabag. The stink of vomit mixed with rotten eggs and fertilizer. Working through a permanent rubberized hole carved into the heifer's side, Hess waited for its half-digested slop to churn, freeing his hand. Then he pulled out the teabag, which three days earlier he had stuffed with pulverized prairie grass.
The rumen is like a huge bathtub, he said, holding about 50 large soda bottles' worth of fluid redolent with bacteria. Relying on these symbiotic microbes, cows eat up to 150 pounds of grass a day, a food inedible to most animals, including humans. Hess was after those microbial secrets, and the placid heifer was happy to oblige.
"You can just do your experiment," he said later. "The cows don't really care."
Unglamorous as it may sound, Hess and his fellow researchers are at the forefront of one of the defining scientific pursuits of our time. It's a hunt that, if successful, could reshape the world's landscape, sending biofuel prices through the floor and allowing a drastic reduction in the country's oil use. It's a long campaign, and the opposition is all around.
It is a war, at its most fundamental, against plants.
For eons, plants have locked the sun's energy into complex strands of sugar, used to build their stems and leaves. These chains are far different from table sugar or grain starch; they cling together, providing the meat of tree trunks and cotton strands. They are the most abundant organic material on the planet, and one of the most hunted.
As long as plants have built up these complex sugars, life in all its forms, from microbes to mastodons, has sought ways to unleash that energy. Since plants can't run, and live for hundreds of years, they have built remarkable defenses, wrapping their cellulose, as the sugars are called, in a sort of barbed wire that, to this day, defies human degradation.
From a series of low-slung buildings in Walnut Creek, Calif., east of Oakland and nestled at the base of Mount Diablo, the Department of Energy's Joint Genome Institute (JGI), fresh from sequencing the human genome, has pursued for the past half-decade the DNA of microbes known to unwind these barbed wires. Marshaling these genetic resources is one of the institute's top priorities, on parallel with its cancer research.
It may sound strange, but the cellulose work is like Lewis Caroll's "Through the Looking-Glass," said Eddy Rubin, JGI's director and Hess' former boss. At one point, Alice hustles after the Red Queen, but she never gains any ground. Likewise, plants have kept their lead over all comers, Rubin said.
"It's been this Red Queen kind of event," he said. "This molecule has evolved ways to prevent water and pests from getting in. It's kept on building new structures. ... These plants have evolved for tens of millions of years to prevent their breakdown."
If biofuels are going to make up a large amount of of the U.S. fuel mix, scientists will need to chase down this Red Queen. Only then can the incipient bioeconomy, now based on energy-intensive corn, decouple from its competition with food. At their most hopeful, scientists envision fields of ultra-productive grasses and trees growing on degraded land, the plants providing the energy for a new American century.
It's a vision that has existed for decades. And a dream that won't be easily reached.
Skeptical about mandates
A new modesty has descended on the biofuel world. Several years ago, overconfident researchers and policymakers predicted that they would soon develop tools to cheaply break down plant walls, creating "cellulosic" ethanol. These predictions begat policy, with congressional biofuel mandates, passed in 2007, calling for 250 million gallons of cellulosic ethanol to be used this year, rising to 500 million gallons in 2012.
It didn't happen. This year, the government lowered its cellulosic mandate to 6.6 million gallons, and in late June U.S. EPA proposed requiring between 3.5 million and 12.9 million gallons of the fuel for 2012, making up less than 0.010 percent of the country's fuel supply (Greenwire, June 22). Despite mandates and subsidies, science can only be pushed so quickly. It is a recalcitrant problem, said Steve Koonin, DOE's undersecretary for science, at a meeting earlier this year.
"There's a long route from where we are currently with [the] technology, which is in fact is coming down in cost and scalability, and deploying it out and actually meeting those mandates," Koonin said. "At least some number of people are increasingly skeptical that we will make it."
Several of the large firms specializing in microbe engineering -- Dupont, Novozymes and DSM -- periodically tout advances in the cellulose process. Last month, Dupont launched a new "cocktail" of proteins to break down plant walls and declared its intent to build a cellulosic ethanol plant in Iowa. (The firm currently operates a demonstration plant in Tennessee.) Meanwhile, DSM announced an advance in processing sugars into ethanol, once they are liberated from plant walls.
This steady tick of upbeat news has continued for several years, causing wariness that microbe-based approaches remain far from commercial viability. It is a caution seen in a recent spate of loan guarantees issued by the Obama administration, which largely passed over bugs to focus on startups that cook cellulose into fuel ingredients (E&ENews PM, Jan. 20). (Earlier this month, however, DOE did announce its first support for a microbe-based cellulosic refinery.) These cooking systems rely on more proven technology, but also require huge economies of scale, which could drive up their fossil fuel demands.
"People have talked about cellulosic ethanol for a while," said Wes Bolsen, head of government affairs at Coskata, which received a $250 million conditional loan guarantee from the government for its gasification-based system. "Coskata is delivering it."
Neither technology has landed a hard blow on the other, said Philippe Lavielle, the chief of business development at Genencor, the Dupont division focused on microbial engineering. Dupont doesn't need the government's guarantees, though it does need its fuel mandates and subsidies; if its first refinery fails, the company won't abandon its plans, he said.
"I think the two [technologies] will play for a while," Lavielle said. "The two will find their niche over time. My guess is there will be five to six technologies that will eventually find their space."
While these companies jockey in public, sequestered in their labs, public and private scientists are continuing the latest front of what has been a long-running campaign. They are scouring the world's genetic resources, sequencing DNA at a furious pace. They are probing exotic salts able to slink through plants' barbed wires. And they are tinkering with DNA, pushing the limits of synthetic biology.
The potential is out there, if it can only be found, JGI's Rubin said.
"Walk through a forest," he said. "There's not a few thousand years of leaves accumulating. There are microbial communities in the forest floor. They are not letting that energy go to waste."
Fittingly, the fight to break down plants began during another campaign: World War II.
As U.S. troops spread through the Pacific, they soon discovered their equipment could not stand the heat. Boots, sand bags, tarps -- all would succumb to jungle rot, causing scarce cargo space to be lost to replaced equipment. It was an untenable situation, causing billions of dollars in losses, and the military wanted to know why.
The Army sent a young Harvard mycologist, Lawrence White, to the Pacific. White followed a grim schedule, retracing the sites of former battles, collecting the many varieties of rot -- white and brown, wet and dry. One rot, found eating through a pup tent on Bougainville Island, in the Solomons, seemed promising. Along with hundreds of others, he sent the strain, QM6a, to Philadelphia for analysis.
Growing QM6a on heavy cotton fabrics to test its strength, military scientists soon found the fungus, a variety of Trichoderma, produced the proteins needed to tear apart plant walls with a single-minded intensity. Two of these scientists, Elwyn Reese and Mary Mandels, made study of QM6 their lives' work.
Mandels created a mutant of QM6, now named Trichoderma reesei, that produced four times the typical amount of degrading proteins, and hopes rose that the mutant would soon unlock a source of nearly unlimited sugars from agricultural waste and trees.
The year was 1982.
Thirty years on, T. reesei remains the source of nearly all the industrial proteins -- enzymes -- used to break down plant walls, mostly in the paper business. One of the first studies JGI undertook was to sequence its genome. By exploring its DNA, they hoped to identify a bounty of enzymes. What they discovered was a very simple genome, with limited variety, said Jim Bristow, the institute's deputy director.
It was disappointing. "We assumed it would be a gold mine," he said.
With that dead end, the institute broadened its search.
Many animals have developed some ability to degrade dead plants, hosting colonies of bacteria in their digestive tracts. Wallabies do it. Sheep do it. Even the hoatzin, an ungainly, foul-smelling South American bird, feeds on leaves, Rubin said, which is a bit "like running a 747 with a steam engine." All are utterly dependent on their bugs judiciously deploying enzymes to free up energy.
"If I sterilized your [human] stomach, you wouldn't be happy, but you wouldn't die of starvation," Rubin said. "If you did that to a termite or cow, it would die immediately."
Seeking to harvest this potential, some five years ago JGI researchers went to Costa Rica to sample termite guts. The insects, the size of the date stamped on a penny, were difficult to work with, and the DNA sequencing was not quite up to the task, resulting in a study that contained many incomplete gene fragments. One of those fragments caught Rubin's eye.
"The closest related gene," he said, "was found in cows."
'Chop, chop, chop'
West of Walnut Creek, Seema Singh spends much of her day thinking about salt.
An energetic researcher at DOE's Joint BioEnergy Institute (JBEI) in Emeryville, Calif., one of the country's leading biofuel research outfits, Singh is seeking to reduce the sheer amount of enzymes needed to transmute grass into simple sugars.
For decades, one solvent after another has failed. High temperatures and harsh acids could work, but such processes, common in the paper industry, are expensive and energy intensive.
"People have been going crazy," Singh said. "Is there anything that can dissolve it?"
JBEI believes it has found an answer in ionic liquids, salts that remain fluid at room temperature rather than crystallizing into the familiar table seasoning.
Already produced at low volumes by the German chemical giant BASF for use in batteries, these liquids attack plant walls like little else, Singh and her co-workers have found.
"I want to break sugar like this," Singh said in an interview at JBEI's labs earlier this year. She sliced her hand downward like a graceful hatchet. "Chop, chop, chop."
Understanding the challenge of these chops requires a foray into the plant itself.
Take a stem. It's divided between three interlocked chains of chemicals. Two-thirds of these strands are the complex sugars cellulose and hemicellulose. Despite their names, they are far from similar. Cellulose runs straight, similar to the starch found in corn kernels, except that every other sugar is inverted, mucking up enzymatic machinery. These straight chains can be incredibly useful: Cotton fibers are nearly pure cellulose.
Hemicellulose is more like a repetitive shrub, branching out every which way. It is a maddening molecule, varying between plants. "But it does have a sort of order to it," said Dominique Loque, another JBEI scientist. "The sugars are not randomly added."
Wrapped around these two strands is the plant's ultimate defense: lignin.
Half concrete and half barbed wire, lignin encircles the sugars, digging into nooks and crannies and bonding heavily with their strands. It repels water, microbes -- pretty much anything. It gives the plant its rigidity; without lignin, trees would become droopy, flopping in the breeze.
Revelations about lignin are near constant. It accounts for a third of the planet's organic carbon, and it is the precursor to coal. Pure lignin is nearly impossible to isolate from its sugars; nearly any craft used to split the two invariably changes their structure. As one paper put it, the study of lignin is like archaeology -- instantly destructive.
Getting rid of lignin is the most important step to cheap cellulosic fuels.
"The No. 1 priority is to delignify," Singh said.
Unlike most solvents, ionic liquids have such a powerful affinity for sugars that they can overcome many of lignin's bonds. Currently, JBEI can remove 80 percent of the lignin from a plant wall, freeing nearly all of the trapped sugars. Adding water to the solution then drags the sugars out, leaving water-adverse lignin behind. Boil the water and the ionic liquid is recycled, ready to work again. JBEI has repeated the process six times, finding its efficiencies never wane.
Recycling is vital. Ionic liquids aren't cheap, but partially that's because there has never been a large demand for them. At first, JBEI could only purchase microliters. They're up to gallons now, and BASF promises to increase production. It is the chicken-and-egg problem of innovation: no one explored ionic liquids because they were expensive; they were expensive because no large-scale facility produced them.
The promise of these liquids is so great -- cutting down expensive enzyme demands by up to 100 times -- that researchers have begun searching for cellulose degraders known to operate in hypersaline conditions. (Salt disrupts the classic T. reesei proteins, one of their many drawbacks.) For example, JGI scientists published a study late last month of a microbe, discovered in the Great Salt Lake, that for some reason produces saline-friendly enzymes.
The problem is far from solved. The 20 percent of lignin that remains hooked onto the sugars still fights the good fight, inhibiting enzymes and slowing down what should be an hourlong process to eight hours. That is far from fast enough for the cut-rate margins of the fuel business. The lignin is holding its ground.
"It looks so hopeful and real to me," Singh said, "in another five years or so."
BP's $500M bet
Calvin Hall has always been an idiosyncratic place.
In the 1960s, a pioneering chemist, Melvin Calvin, considered the father of photosynthesis science, built his new lab at the University of California, Berkeley, in the round. A squat tube free of right angles, it was constructed in a spirit of open, collaborative science, and so there is some irony that the building now houses the Energy Biosciences Institute (EBI).
Funded by a $500 million grant from BP PLC, the institute has been controversial since the ink on its contract dried. It can be confounding at first glance: While BP has first rights to any advance EBI develops, the lab continues in Calvin's open, collaborative spirit, fostering interdisciplinary collaboration -- it often works with JGI and JBEI -- and publishing its research in peer-reviewed journals.
It is also home to one of the biggest recent steps toward cellulosic ethanol.
Traditional ethanol is made when brewer's yeast feeds on glucose, the sugar found in corn starch and cellulose. Yeast, the industry's powerhouse microbe, is a finicky eater. Much like a child refusing broccoli when a bowl of ice cream is in sight, yeast won't consume branched hemicellulose sugars while glucose is around. Getting the bug to feed simultaneously on the sugars is a pressing concern.
According to EBI's director, Chris Somerville, the challenge is over. Finished.
"We solved the problem of using all the sugars during fermentation," Somerville said on a recent tour of Calvin Hall. "We've figured out how to engineer strains to simultaneously [use] all the sugars with very high rates. It's good enough that the company moved it into commercialization already."
When Somerville says "the company," of course, he means BP. The oil major broke ground this year on its first cellulosic ethanol refinery in Highlands County, Fla., using a relative of sugar cane that is light on the sugar and thick in the stem. (They call it "energy cane"; the crop is already planted.) The facility will produce 35 million gallons of cellulosic ethanol a year beginning in 2013, BP says.
The sugar advance had its origins in work funded by the National Institutes of Health. The agency has poured about $100 million into research on a pink bread mold, Neurospora crassa, that has long served as one of medicine's model organisms. Some 6,000 papers had been published about the fungus, but fewer than 10 had anything to do with how the fungus fed on rotting logs.
Using tools developed for Neurospora medical work, Jonathan Galazka, a doctoral student at Berkeley, found that one set of newly discovered proteins transported attached pairs of glucose into the cell for digestion. Seeing Galazka's results, eventually published in Science last year, a BP scientist suggested that he combine his yeast with a strain developed by Yong-Su Jin, a professor at the University of Illinois, Urbana, who had spent 10 years getting yeast to eat branched hemicellulose sugars.
Over a feverish six months, scientists from four labs synthesized the yeast to carry both sets of genetic tools. Testing it, they found the yeast produced much more ethanol than a bug fed either individual sugar; it was burning through both energy sources. In effect, rather than making the hemicellulose -- the broccoli -- more attractive, they had made the glucose -- the ice cream -- less appealing. They published their work this past December, just after Christmas (Greenwire, Jan. 3).
Even this new yeast strain, which eliminates a whole step and cuts enzyme costs by a third, won't make BP's Florida plant profitable. BP has accepted that it will lose money on the biorefinery, which is "almost certainly not going to work well," Somerville said. And it will be expensive.
"They're putting down $400 million for only 35 million gallons a year of capacity," Somerville said. "Let's call that $10 per annual gallon. That's about two to three times as high as the corn ethanol guys."
Somerville expects that BP's second plant could work quite well, though it is unlikely to come online until later this decade. Perhaps by 2020 the field could be profitable. While this is not a reality that the government and media are eager to hear, some caution is justified.
"You got to understand that it's a big new industrial process," he said. "It's a really expensive proposition to put that much steel in the ground. And once you put it in the ground, it's there. You can't just pick it up and put the money somewhere else."
Beyond the lab
Once, before it dipped into cow guts, the Joint Genome Institute hummed with industrial activity.
A decade ago, DNA sequencing required a series of large Rube Goldberg devices operating in tandem, a spectacle that would consume floorspace at reckless speed. Groups of schoolchildren came to the institute, staring at the machines in awe.
Today, DNA can be read at high speed by anonymous white boxes whirring on a lab bench. Large sections of JGI's buildings, once devoted to sequencing, now sit vacant. The lab suggests school groups visit JBEI, on the other side of the Berkeley hills. Visually, there's a lot more going on there.
This decline in specialized machinery, driven by the push to cheaply read human DNA, has pushed the JGI to focus on everything that comes after sequencing: assembling genomes previously unknown to science and, especially, interpreting the streams of genetic data now available. As any geneticist will tell you, there is a vast gulf between reading DNA and understanding it.
Bridging this gulf is why Hess, now at Washington State University, came to JGI.
"I set out to really show that you can use sequencing to find something that can be applied," Hess said. "And that's exactly what we did."
Rubin sent Hess, then his post-doc, out to an experimental farm in Illinois. There, a dozen cows had holes grafted into their sides, a long-standing research tool; few digestive systems are as well studied. Hess slipped his grass-filled sack into the stomach and, after three days, removed the samples, which had gone "smooshy." Beyond his stuck arm, it was smooth research.
JGI scientists then randomly sequenced all of the microbial DNA found in these samples, regardless of species, giving a sense of the genetic abilities of the many bugs scientists would be unable to grow in the lab. All past studies searching for these proteins worked off, combined, some 0.20 gigabases of DNA data. A single rumen sample generated 268 gigabases of similar data, said Zhong Wang, one of JGI's computational genetics gurus.
"It's equal to sequencing a human genome a hundred times," Wang said.
For decades, the world has relied on a single species of Trichoderma for their enzymes -- a distinct lack of diversity. JGI's rumen study, published early this year in Science, generated some 28,000 genes that appeared to encode similar proteins. And these were full genes, not the fragments like the termite study, Rubin said.
It was a stunning diversity of proteins. "The past 40 years, people have been looking for these enzymes, and in this one study we double the numbers," he said.
The rumen team was interested in more than hypothesizing these proteins existed. They pulled 90 proteins from the rumen and tested them individually against grasses and model carbohydrates; 51 showed at least some activity against one of the feedstocks, suggesting at least half the team's flagged enzymes were real.
Perhaps most impressively from the point of pure science, the team was able to divine nearly complete genomes for several microbial species, out of an estimated 1,000 bugs expected in the rumen.
Grouping this assembled DNA into what they called "genome bins," they then conducted single-cell analysis that suggested the bins matched up well with their wild counterparts. It is a result that suggests scientists are close to being able to generate whole genomes from cluttered, wild samples -- an important point, when only 1 percent of the world's microbes can be grown in the lab.
"Right now, we generate hundreds of human genomes of data," Rubin said. "In the future, it will be thousands and tens of thousands. We won't have to grow these things in labs."
There are limits to where this genetic prospecting can go.
Imagine a fungus on a log. It digests the log only as quickly as it can use the sugars, and no faster. It wants to sustain itself, not the whole forest, said John Grate, the chief science officer of Codexis, a bioengineering firm that specializes in evolving known enzymes.
"The rotting log in a forest is not an industrial environment," he said. "A termite's gut is not an industrial environment. Natural evolution does not get you what industrialized mankind needs."
Even if no industrial-ready protein is ever discovered during JGI's bioprospecting, it can still provide the grist of discovery. Even Codexis' scientists, focused on artificial evolution, download their initial proteins from public online databases. Those databases have swelled with JGI's contribution, and any company is free to explore their potential.
There has been thousands of hits on the data, Rubin said.
"This has really overwhelmed people," he said.
JGI is not resting on its rumen laurels. It has dozens of sequencing projects under way, many at the behest of JBEI or EBI. They're looking at kangaroo and wallaby guts, and dung beetles, too. Then there's the shipworm mollusk, which carries wood-fond microbes; it is said shipworms marooned Columbus in the West Indies for a year.
As scientists continue their battle against plants, it is clear that few studies will alone shift the industry's expectations, Rubin said.
"There's going to be a bunch of incremental advances that will slowly make the cost calculations better," he said.
Despite the long ambitions humans have had to harness plants' energy, cellulose research, in its modern incarnation, is a young field, five or six years old. Like many alternative energy sources, funding existed for similar programs in the 1970s, only to be killed by cheap oil.
Rubin wonders where the science would be had the money stayed. "It'd be very different now," he said.
Mary Mandels, the military scientist who created the Trichoderma mutants still widely used today, saw the funding dry up from her field three decades ago. Someday, it would return, she said. It did, though not as soon she would have liked.
"At the moment there is plenty of oil, and interest in ethanol or chemicals from cellulose has declined," Mandels said in 1983. "No doubt, the pendulum will swing back soon."
Click here for a peek at JGI's rumen DNA data.