Plastics are perfectly designed for every imaginable need. Some are strong and steadfast, others flexible and elastic. Plastics by definition can be molded into any possible shape and seem to be available for any plausible purpose. They are inexpensive to make and lightweight to transport. In fact, after millennia of human ingenuity, plastics are perhaps the ideal material. Except they’re not.
Plastics are cheap to produce because they are byproducts from the production of fossil fuel-based gasoline, and there are associated costs. Fossil-fuel resources are limited, the production of plastics creates greenhouse gases that contribute to climate change, and when plastic products are thrown away—as millions are every single day—they litter landfills and oceans, where they will persist for hundreds of years.
Humans are faced with the stark reality that these materials, which seemed to magically transform modern life, are in fact a danger to the planet and its inhabitants. Plastic particles from deteriorating garbage are contaminating drinking water, showing up in the stomachs of sea creatures, and making it up the food chain and onto our dinner plates. And, although awareness is rising along with the prevalence of re-usable water bottles, straws, grocery bags, and lunch containers, the plastics production lines have not slowed. Why? Because plastics are optimized for each product and purpose, and society is struggling to find suitable alternatives.
New strategies are needed to address this challenge. Humans cannot realistically remove plastics from their lives, but meaningful action is possible nonetheless. As many scientific solutions have been found by turning to nature for help, biologists at Los Alamos are working to build better plastics by engaging living sources of carbon, such as plants, and workhorse microorganisms, such as bacteria and algae. Their strategy combines experimental biology, chemistry, bioengineering, genomics, and machine learning to optimize every aspect of the plastic-production process. The goal is simple: the plastics of the future will be bio-based, biodegradable, and of course, perfectly designed for every imaginable need.
Making carbon cycle
Every minute of every day, one million plastic bottles are purchased. Water bottles, soda bottles, shampoo bottles, dish soap bottles… one million, every minute. These bottles are composed mostly of carbon, hydrogen, and oxygen atoms arranged as a long chain of identical sub-units called monomers, together forming a polymer. The monomer building blocks come mainly from fossil fuels, which in turn come from plants and animals that were decomposed and fossilized millions of years ago.
The strong carbon bonds that are created during the polymerization process make plastic difficult to break apart, which is one reason it is so useful. As a result, discarded plastics in landfills and oceans just break down into smaller and smaller pieces of plastic, known as microplastics, and don’t easily decompose back into their monomer building blocks to be usefully incorporated into either the ecosystem or the economy.
Plastics can be recycled once or twice, but overall very few plastic products are recycled at all—less than 10 percent by some estimates. Furthermore, most are downcycled into items of lesser economic value such as furniture stuffing or carpet. If plastics could be upcycled into higher-value items, such as sports equipment or auto parts, there would be incentive to invest in recycling infrastructure and to manufacture items with recycled material.
Alternatively, corn and sugar cane can be used to generate “bioplastic” polymers, such as polylactic acid (PLA), that can be made into cups, cutlery, and containers. These and other bioplastics don’t use fossil fuels; however, their production competes with food resources and the polymers are still difficult to break down. Traditional recycling facilities cannot process most bioplastics, so they must be taken to industrial composting facilities, which are not widely available.
“Just because something is made biologically does not mean it is going to degrade faster,” says Babetta Marrone, Los Alamos biologist. “There may be some greenhouse-gas advantage, but they’re not contributing to a more circular economy that could free up the carbon for other applications.”
The idea of a circular economy is to make products that do not become waste, but rather, after their intended use is complete, they feed back into the system to become something else, or even the same product again. A bioeconomy is this same concept but driven by materials that are derived from biological sources or made through biological processes. These economic ideas mimic nature’s carbon cycle: Most carbon-based materials such as wood and cotton are naturally recycled in the environment when microorganisms decompose them and move carbon molecules from one form to another. When it comes to plastic, however, the carbon cycle is stuck: plastic just stays plastic for too long.
But not all plastic needs to last forever. While some plastics, like the dashboard of a car, must continue to be strong for decades, single-use plastics don’t need to be as robust. The challenge is to figure out the components of a bioeconomy for plastics: What plastic-like materials can be competitively made for various purposes and also degrade when appropriate? Microorganisms are key to nature’s cycle of breaking down and reconstructing carbon—and as such they will be key to a future bioeconomy as well.
In the 1920s, French microbiologist Maurice Lemoigne studied the organism Bacillus megaterium, which naturally produces a plastic-like polymer and uses it to store energy. Pseudomonas putida is a bacterium discovered in the 1980s that can metabolize benzene, toluene, and other aromatic hydrocarbons that are hazardous to the environment. In 2016, Japanese researchers isolated another bacterium that can break down and metabolize plastic; they found it outside a bottle-recycling facility.
Envisioning that microorganisms could be harnessed to degrade plastic waste is one thing, but optimizing them for a large-scale clean-up—picture the 1.6 million-square-mile plastic patch in the Pacific Ocean—is a challenge. Furthermore, scientists have studied multiple microorganisms that naturally produce carbon-building blocks similar to those found in transportation fuels and plastics. However, to compete with fossil fuel-based production, these organisms would need to thrive under industrial conditions.
“The natural enzymes that break down plastic waste are not efficient and need to perform faster,” explains Taraka Dale, biologist and Biomass and Biodiversity team leader at Los Alamos. Through bioengineering, scientists can improve enzymatic activity—such as enabling enzymes to perform at the high temperatures needed for industry—and fine-tune cellular production of various building blocks for new plastics. However, in order to succeed, they need sensitive and accurate ways to “see inside the cells” to determine which enzymes are working properly.
Dale and a number of Los Alamos colleagues are partners in two Department of Energy-funded consortia—the Agile BioFoundry and BOTTLE (Bio-Optimized Technologies for keeping Thermoplastics out of Landfills and the Environment)—focused on developing bio-based manufacturing and deconstructing plastic waste. For both consortia, one of the main contributions from the Los Alamos team is a revolutionary biosensor, called Smart Microbial Cell Technology, that allows the scientists to screen bacteria and select the varieties that perform best.
As part of the Agile BioFoundry, the Los Alamos team develops these performance-screening tools for new strains of bacteria that have been identified as good strains for biomanufacturing. The consortium adopted P. putida because of its tolerance for extreme conditions and its ability to digest aromatic hydrocarbons such as toluene into a useful molecule called muconate, which the bacterium naturally produces during metabolism.
Muconate can be converted to adipic acid, which is a building block of nylon. BioFoundry partners at the National Renewable Energy Laboratory (NREL) began engineering P. putida to improve its ability to use glucose from plants, instead of toluene, as food to make the bacterium produce the building blocks for bio-derived nylon. Next, the Los Alamos team contributed its biosensor as a reliable, sensitive way to figure out which candidate cells were performing best. The biosensor is added to all candidate cells and causes the ones that produce the most muconate to “self-report” by glowing green (see graphic above).
The plastics of the future will be bio-based and biodegradable.
“Using the biosensor, we can screen thousands of variables in one tube at one time,” says Dale. “And in this application, all of them are reporting production of the target molecule, muconate.” In their most recent studies with NREL and Oak Ridge National Laboratory, the team reported three-fold improvements over other systems to create bio-derived muconate.
Furthermore, because the biosensor can be tailored to report many types of activity within a cell, it is being used for many different projects. As part of BOTTLE, the biosensor will be used to develop microbes that deconstruct traditional plastic into its building blocks, which can then be upcycled into higher-value products.
The elephant in the room
One of the first plastics ever developed was invented in 1868 as a way to save the elephants. At the time, the game of billiards was gaining popularity and people feared the ecological impact of making more and more billiard balls out of ivory. Hoping to win a $10,000 award for finding an ivory substitute, an inventor named John Wesley Hyatt made billiard balls out of something completely different: cellulose nitrate, or celluloid. In an attempt to save one precious natural resource, an entirely new class of materials was invented.
Something disruptive is needed again. Using microbes to make plastics or degrade plastics is one solution to the plastics problem; however, many agree that a better solution would be to develop completely novel kinds of bioplastics that are intentionally designed to degrade at the end of their lives. Furthermore, new types of plastic molecules may even have advantages over fossil fuel-derived ones.
In 2018, Marrone convened a multi-disciplinary team of scientists at Los Alamos to explore this challenge in a project called BioManIAC (BioManufacturing with Intelligent Adaptive Control). Inspired by MANIAC, the pioneering first computer at Los Alamos, Marrone and her team are using biology, chemistry, and machine learning to create a process for finding entirely new plastics. Their goal is to identify monomer building blocks that are made using photosynthetic microbes such as algae and that readily return to the ecosystem as the plastic degrades, eliminating the need to collect and recycle waste materials.
Why the switch to algae? As Dale puts it: “Instead of growing plant biomass in order to feed and grow bacteria, it is more sustainable to use photosynthetic microbes directly.” Algae can grow outdoors using minimal infrastructure and non-potable or saline water, and they can even use waste carbon dioxide (CO2) from a nearby industrial plant, rather than simply venting the greenhouse gas into the atmosphere. As long as they have these ingredients—sunlight, water, and CO2—the algae use photosynthesis to produce carbon building blocks.
Bio-derived molecules have diverse functionalities, so many new kinds of polymers are possible. The challenge lies in deciding what makes a good bioplastic, then matching physical characteristics with their chemical makeup, and then matching those chemicals with the appropriate biological pathways.
“We can use machine learning to accelerate the process of biopolymer discovery, design, and development,” says Marrone. “We want to be able to say ‘I want to design out brittleness and design in elasticity, but at the same time make the product degrade faster’.”
Built to last… but not forever
The BioManIAC team is composed of experts in three scientific disciplines: chemistry, biology, and machine learning. Working together, the team members have begun to evaluate a few specific building-block molecules and the corresponding pathways for producing novel bioplastics.
To test their approach, the team is using cyanobacteria, which are simple, well-studied organisms that, like algae, use photosynthesis to make carbon-based molecules for energy storage. Some of these storage molecules, called polyhydroxyalkanoates (PHAs), are already considered desirable for bioplastic production because they are biodegradable and can be used to make polymers with a wide range of plastic-like characteristics. However, much is still unknown about what combination of different PHA monomers is required for which specific plastic traits. The goal of the chemistry team is to evaluate these possibilities.
“Our team will systematically create and observe each monomer combination for physical and mechanical properties,” says Los Alamos polymer chemist Carl Iverson.
Specifically, Iverson’s team is screening PHA monomers for their thermal properties, such as melting point and glass transition temperature, which is the point at which a material transitions between a glassy and brittle state to a rubbery, flexible one. The team is also examining mechanical properties by doing puncture and elongation tests and measuring resistance to tearing. In these first BioManIAC experiments, Iverson is looking for polymers that can replace polyethylene and polypropylene, both of which are used for single-use plastics like grocery bags and cutlery.
To accelerate identification of the best candidates, Los Alamos machine-learning expert Ghanshyam Pilania and his team are combining Iverson’s experimental data with previously published data to develop, train, and validate a predictive machine-learning model that connects polymer chemical structure to specific physical properties. The team also plans to identify optimal biological synthesis and culture conditions for PHA production. The machine-learning model is adaptable and can be iteratively refined and improved as new data become available, guiding the next experiments. This way, optimization occurs in a fast, efficient manner.
“Machine learning helps us systematically navigate this multi-parameter, multi-objective optimization problem,” explains Pilania. “We have a wish list of properties for biopolymers such as strength, flexibility, and biodegradability, but more often than not, these properties have trade-offs and sometimes strongly conflicting relationships. You can’t have them all optimized at once. The goal is to find the best-case scenarios hidden in this polymer-chemical space.”
Using additional experiments, the team is examining what environmental conditions are necessary to biodegrade the candidate polymers, what the polymers will degrade into, and whether or not those molecules will have a negative environmental impact. The goal is to have new plastics that could degrade into monomers fairly quickly in a landfill environment, allowing the building blocks to feed back into the natural carbon cycle. Alternatively, if collected, these plastics could be composted in an industrial environment and the building-block nutrients could be extracted and fed to algae, bacteria, or plants to be incorporated into new plastic products.
Although Iverson’s team is making test monomers in a chemistry lab, he explains that the chemical process isn’t scalable to industrial-level production—it’s too expensive and toxic. This is why a biological system for production is needed. Furthermore, the atoms in PHA molecules have a specific spatial arrangement (i.e., the angle at which the methyl group sits) that is very difficult to achieve in a chemistry lab.
“The biological process to make PHAs is exquisite,” says Iverson.
Instructions for production
Biology might be responsible for exquisite PHA production, but the process is hidden inside microscopic cells. In order to prepare cyanobacteria—and eventually algae—for industrial-scale production, scientists require a solid understanding of which genes and proteins are responsible for making PHA. Furthermore, if the chemistry team identifies specific PHA-like monomers that should be modified to make, for instance, a better plastic bag, then the biology team needs to decide if it’s biologically feasible to do so and how.
For decades, scientists have been able to sequence the genomes of living organisms to study the DNA blueprint that enables them to exist. The DNA broadly contains “coding” and “non-coding” regions. The former is made up of genes—sequences of nucleotides that provide instructions for assembling amino acids, which align and fold in unique ways to form proteins that carry out specific functions in cells. The latter is the DNA found between genes, which used to be considered junk but is now known to contain valuable information about gene regulation.
Coupling genes with their functions is a laborious experimental process, so often the first step in understanding a new genome is to search databases for sequences that match known genes. BioManIAC scientists can search the databases to find genes that are known to make PHA; however, if they want to discover new pathways, or even new monomer products, they must explore the unknown territory in both coding and non-coding regions. To do so in the most efficient manner and without spending excessive time in the lab, the team is employing machine learning.
The BioManIAC approach is to start by looking for small pieces instead of entire genes or proteins: Instead of looking for a whole fish, they are looking for anything that resembles a fin. With this approach in mind, the team divides an organism’s genomic data into equivalent-sized pieces called “k-mers”: sections of DNA that are k nucleotides long. For instance, if k=14, then the algorithm would determine all possible 14-nucleotide pieces in an entire genome of interest (including non-coding areas) and compare those segments to all known genes associated with PHA production. Any k-mer match would indicate possible new PHA-related information that should be further investigated.
The team is also looking through the genomic data for code that matches that of specific PHA-related “protein families.” Protein families are small groups of amino acids that work together for a specific function, such as in the active site of an enzyme. Again, by searching for protein families instead of whole proteins, the scientists are looking for a match of only a small, but critical, piece of data. In essence if it looks like a fin, then there is the possibility it might be from a fish, and perhaps the right kind of fish. If these protein families are found, they could lead to the discovery of new enzymes associated with PHA.
“We might first ask: Is the k-mer present in the data?” says Los Alamos biologist and Bioinformatics and Genomics team leader Shawn Starkenburg, “But then we go deeper: If the k-mer is present, does the organism also produce PHA? And next, is a specific protein family also present?” By putting these all together, the team is beginning to discover new information about PHA production in cyanobacteria. So far this approach has helped the machine-learning team identify three PHA-production gene candidates that the team is now in the process of studying experimentally.
When Hyatt invented celluloid in 1868 as a replacement for ivory billiard balls, he showed that a new material could be just as good, if not better, than the status quo. Celluloid was not a successful ivory replacement, but it opened the door to a new world of possibilities of polymer-based plastics, and billiard balls are now made with acrylic or plastic resins.
Traditional plastics seem today like the only material for many of life’s necessities, but history suggests the advantage of a more imaginative outlook. Through creativity and science, new materials could become competitive alternatives in the $500 billion plastics market. Marrone, Dale, and their colleagues are making headway using microorganisms to produce muconate and PHA, but this is just the beginning, as microbe-developed polymers are on the rise throughout the research and development community.
Dale explains that her team is already moving forward to more types of plastic replacements and that she has a new project to develop absorbent biopolymers, which could mean, for example, bio-based paints and diapers. One by one, Dale hopes this work will lead to bio-based alternatives to each current plastic—ultimately leading to a healthier outlook for the future of the planet.
“This is my dream,” she says with a smile. LDRD