Microbes paired for biological gas-to-liquids (Bio-GTL) process

Forecasts predict that diesel consumption will continue to grow over the next 30 y, surpassing gasoline in total volume (1). Growing demand is motivating the development of alternative sources of fuels for the heavy transportation sector. Biodiesel is the primary alternative to petrodiesel and is already the second-most abundant biologically derived transportation fuel in the United States. Most biodiesel is produced via transesterification of lipids isolated from oil crops, such as palm, canola, or soybean, which are limited in supply. In PNAS, Hu et al. demonstrate an integrated biological process for converting synthesis gas to triacylglycerides (TAGs) that could be used as an alternative biodiesel feedstock (2). The process pairs two microbes to take advantage of their unique metabolisms (Fig. 1). An anaerobic bacterium converts gaseous CO, CO₂, and H₂ to acetate in a first stage. Aqueous acetate is then fed to an engineered oleaginous yeast for production of lipids in a second stage. The process leverages a potentially abundant, low-cost feedstock to produce liquid transportation fuels analogous to established gas-to-liquid (GTL) processes that use heterogeneous catalysts; only here, the catalysts are living organisms.

Fig. 1.

A comparison of chemical GTL and new Bio-GTL synthesis of diesel fuels. Both approaches use synthesis gas that can be made from a wide range of sources, including steam reforming of natural gas and gasification of carbonaceous materials. Renewable feedstocks, such as biomass or biogas, can also be used. Chemical GTL processes use Fischer-Tropsch synthesis and catalytic cracking (red arrow) to produce diesel hydrocarbons. The Bio-GTL process described in Hu et al. (2) uses an acetogen to convert syngas to acetate in an anaerobic process. Acetate is fed to oleaginous yeast, which produces TAGs in an aerobic process. TAGs would then be transesterified with established technologies to produce biodiesel (blue arrow).

One of the major challenges to increasing biodiesel use is the limited supply of feedstocks. One solution is to identify and cultivate additional oil crops, such as algae or Jatropha, which can achieve high oil yield and productivity. A second alternative is to use microbes to convert renewable sugars into lipids (3) or biodiesels (4) directly. Unfortunately, despite significant research and development, neither approach has overtaken the cultivation of existing oil crops. Many challenges facing large-scale cultivation of lipid-rich algae—including water requirements, bioreactor design, and containment—remain unsolved (5). Oil crops are most productive in nutrient-rich soils, which leads to tensions between oil production, food cultivation, and maintenance of native tropical habitats. The conversion of sugars to biodiesel and other oleochemicals has been improved by implementing metabolic engineering strategies with elegant regulatory circuits, but yields remain well below theoretical limits (6). In contrast, oleaginous microbes, when triggered with the appropriate nutrient conditions, can convert sugars into TAGs in high yields and acceptable productivities (7). As with current processes, biodiesel from sugar will be limited by feedstock supply. Corn is currently the dominant source of sugars for fermentation to ethanol in the United States. However, the sustainability of corn-ethanol has been challenged, motivating the development of cellulosic biomass as a feedstock for biofuel-producing fermentations (8). The Department of Energy and Department of Agriculture have estimated that 1–1.5 billion tons of biomass will be available for conversion to biofuels (9), but this is equal to only 30–50% of the petroleum energy content used by the United States each year for transportation. So even if the challenges associated with biomass recalcitrance and high-yield conversion of lignocellulosic hydrolysates can be solved, there is insufficient feedstock to replace current petroleum demand. Therefore, new feedstocks and alternative routes must be explored.

In addition to overall supply limitations, the cost of feedstocks presents an economic challenge. Recently, low petroleum prices (<$40 per barrel), relatively high sugar prices (∼$0.15 per lb), and uncertainty in the price per ton of biomass have hurt the outlook of biofuel production in the short term. Simultaneously, the supply of natural gas has dramatically increased, leading to record low prices. One only needs to look at a satellite image of North America at night—the Bakken Formation in North Dakota is as brightly lit as a major US city—to see the immense volume of natural gas that is being flared on a nightly basis. Gas is flared because cost-effective alternative uses are not available and venting methane-rich shale gas would be many-fold worse to the environment than the CO₂ generated from flaring. The low cost and abundant supply of natural gas has motivated research to upgrade methane and other natural gas components to liquid transportation fuels or other higher-value chemicals (10, 11). Chief among these are liquefied natural gas, methanol, and GTL hydrocarbons. Liquefied natural gas sells for two to eight times the cost of pipeline natural gas, indicating how costly it is to produce. Low-cost, partial oxidation of methane to methanol is a holy grail of catalysis and established methods are currently costly to implement. Microbes have evolved the ability to activate and partially oxidize methane but the rates of methane assimilation are very poor (12). Research to improve methane monooxygenase is ongoing, as are academic and industrial efforts to engineer methanotrophs to convert methane into higher-value chemicals, including liquid transportation fuels (13, 14).

Instead of trying to control the oxidation of methane, natural gas hydrocarbons can be converted into synthesis gas: mixtures of H₂, CO, and CO₂, via steam reforming at high temperature in the presence of a heterogeneous catalyst. Synthesis gas can also be created by gasification (high-temperature steam reforming) of other carbonaceous materials, including coal, biomass, and municipal waste. Fisher-Tropsch synthesis can then be used to upgrade the gases into liquid hydrocarbons, as was done in Germany during World War II. The integrated syngas generation and catalytic upgrading to diesel fuels is termed gas-to-liquids, or GTL. Mature GTL technologies have existed for many years, but commercialization requires a substantial financial investment to build large plants that capitalize on economies of scale (14). Work is ongoing, but GTL technologies have not yet displaced significant volumes of petrodiesel.

In PNAS, Hu et al. (2) demonstrate a novel biological GTL (Bio-GTL) process by combining lipid synthesis in an engineered yeast with acetogenesis (metabolism of CO₂ to acetate) in a bacterium. Their approach takes advantage of two organisms that have evolved to specialize in individual GTL steps. Moorella thermoacetica is a model acetogen that uses the Wood-Ljungdahl pathway to assimilate CO₂ and generate acetyl-CoA. This central intermediate can be used for biosynthesis or be consumed to create ATP and acetate (15). The reducing power required for CO₂ assimilation comes from H₂ via hydrogenase activity and is used to drive a set of electron transport reactions that result in a proton gradient across the cell membrane used for ATP synthesis. Like fermentation, acetogenesis has evolved to couple cell survival with product (acetate) formation, thereby enabling high carbon and electron yield of acetate. Similarly, the oleaginous yeast, Yarrowia lipolytica, has evolved the ability to use available carbon feedstocks to synthesize (with high yield) and store large amounts of lipids when growth is limited by inorganic nutrient starvation. Strains of Y. lipolytica have been engineered to store up to 90% of dry weight as lipid (3). The approach of combining evolved specialists is a sharp contrast to the more frequently observed synthetic biology strategy of combining desired traits into a single organism [e.g., a consolidated bioprocessing microbe (16)]. Here, the individual strains have not only evolved efficient metabolic pathways but also the physiology required to handle unique environmental conditions (e.g., tolerance of low pH or high oil content). These complex traits are not always compatible and rarely easily transferred between organisms. It will be interesting to see if this study motivates efforts to combine other promising traits from distinct microbes into engineered consortia.

Hu et al. demonstrate a novel biological GTL (Bio-GTL) process by combining lipid synthesis in an engineered yeast with acetogenesis (metabolism of CO₂ to acetate) in a bacterium.

A large amount of classic biochemical engineering was required to achieve the semicontinuous, integrated Bio-GTL process. Through characterization of the acetogenesis phase, Hu et al. (2) identify sets of reactor conditions that could be used to maximize biomass generation and specific acetate productivity. When optimal conditions were sequenced, the highest volumetric acetate productivity was achieved. Biochemical engineering textbooks show students how continuous cultivation is vastly superior to batch in terms of productivity, yet biotechnology is dominated by batch and fed-batch processes caused by other constraints, such as strain stability, operating costs, and the scale of typical bioprocesses. Here, hollow-fiber membrane filtration was applied to recycle cells and continuously remove spent media in a pseudoperfusion scheme aimed at maximizing lipid titer and productivity. This design enabled maintenance of optimal acetate levels in the first stage, continuous transfer of carbon source to the second stage, and accumulation of lipids to high titers. Although lipid titer and productivity were less than that achieved in separate fed-batch experiments, a few key variables, including nutrient levels in the lipid synthesis bioreactor, were identified as opportunities to further improve the process. Continuous operation was stably maintained for a period of 100 h, equivalent to the time required for the acetogen to reach its maximum cell density. At this stage lipids were still accumulating, which suggests that performance has not yet peaked. Further development could examine the limit of continuous operation in terms of acetogen time on stream and the potential of continuous removal of lipid-rich cells. If successful, this approach could motivate similar strategies, where acetate can be converted to other metabolites, providing a biorefining route from natural gas to higher-value products. Similarly, this demonstration may help motivate renewed interest in developing continuous bioprocesses to meet the growing demand for biomanufactured products.

Although technically sound, the demonstrated Bio-GTL process (2) has many technoeconomic and environmental performance hurdles to overcome before it can be successfully deployed at commercial scale. A complete process model and life cycle assessment is needed to provide a fair comparison of the trade-offs between it, current biodiesel sources, algae, sugar-to-lipid, and alternative Bio-GTL strategies that fix CO₂ or oxidize methane directly. The most straightforward deployment of this Bio-GTL approach is to use natural gas as a feedstock. If capture technologies and small-scale syngas production routes can be made economically viable, Bio-GTL could help reduce the amount shale-gas flaring. Better yet, biogas generated from anaerobic digestion or syngas generated from biomass gasification could provide a sustainable feedstock and reduce net CO₂ release. One advantage Bio-GTL may have is the ability to use producer gas, a syngas mixture also containing nitrogen, which is much cheaper to produce than the clean syngas streams used in current GTL processes. Although gasification technologies are continually improving, energy requirements are not small and the scale of the gasifier and available feedstock supply will need to be matched with the nutrient and water requirements for the downstream fermentors. If the economics work out for natural gas, it will be interesting to see if other feedstocks—biomass, municipal waste, or coal—would also be economically viable. A full technoeconomic and life cycle assessment of this attractive technology will be very revealing.