Bio-engineer plants to absorb more CO2?

Our roadmap towards ‘net zero’ requires 20-30GTpa of carbon offsets using nature based solutions, including reforestation and soil carbon. This short note considers whether the task could be facilitated by bio-engineering plants to sequester more CO2. We find exciting ambitions, and promising pilots, but the space is not yet investable.


What is bio-engineering? In 2016, scientists at DuPont gene-edited maize to grow more effectively in dry conditions. In 2017, researchers at the University of Oxford introduced a maize gene into rice plants, to increase the number of photosynthetic chloroplasts surrounding leaf veins. In 2019, scientists at Huazhong Agricultural University gene-edited rice to tolerate higher soil salinity. These are examples of bio-engineering: modifying the genetic code of plants for practical purposes.

How could it help? The world’s land plants absorb 123GTpa of carbon each year through photosynthesis. 120GTpa is re-released through respiration and decomposition. The result is a net sink of 3GTpa. For contrast, total anthropogenic carbon emissions are 12GTpa. It follows that small changes in the natural carbon cycle could materially shift carbon balances, per our climate model below.

The limitations of photosynthesis. Photosynthesis uses sunlight to convert CO2 into plant-sugars. It is only 1-5% inefficient, suggesting great potential for improvement. It is also vastly complex, comprising over 170 separate sub-stages. Amidst the complexity, RuBisCO is the most crucial limitation.

The limitations of RuBisCO. RuBisCO is an enzyme that catalyzes the reaction between CO2 and RuBP during photosynthesis. However, the RuBisCO enzyme is imprecise. It evolved at a time when the world’s atmosphere contained much lower oxygen concentrations. Unfortunately, under present atmospheric conditions, 20-35% of RuBisCO’s catalytic activity reacts O2 with RuBP, instead of CO2. The resultant products cannot continue their biochemical journey into becoming sugars. Instead, they are broken down in the process of photorespiration. Photorespiration uses up c30% of the total energy fixed by photosynthesis, and re-releases CO2 into the atmosphere. Photorespiration lowers agricultural yields by 20-40%.

What if RuBisCO could be helped to fix more CO2 and less oxygen? One way to do this is to increase the atmospheric concentration of CO2 in greenhouses, which can increase crop yields by c30%, per our note below. Another way is through bio-engineering.

Realizing Increased Photosynthetic Efficiency (RIPE) is a research institute funded by the Bill and Melinda Gates Foundation, UK foreign aid, the USDA and academic institutions. It aims to generate higher crop yields per unit of land, using bioscience. After ten years of research, RIPE has recently modified tobacco plants with genes from green algae and pumpkin plants to reduce the energy penalties from photorespiration. The result is that these modified tobacco plants grew 40% larger. A follow-up study may achieve plants that are 60% larger. Similar modifications are also being tested on soybeans and cowpea plants.

Researchers at the University of Wurzburg have also modelled metabolic pathways that may increase the photosynthetic efficiency of plants, potentially by as much as 5x, with results published in 2020. The work uses synthetic CO2-fixating carboxylases, RuBisCO from cyanobacteria, and additional methods of preventing fixed CO2 from being re-released. Experiments are planned to test the work in tobacco plants and thale cress.

Increasing photosynthetic efficiency and crop yields could be a crucial help, lowering the land intensity of crop production, which covers 1.7bn hectares of the globe today (data below). For comparison, our target of 15GTpa of reforestation will require 1.2bn hectares of land, hence any material reductions in cropland requirements will be helpful.

Sequestering more of the CO2. 50-95% of the carbon that is stored in natural eco-systems is not stored in biomass above ground, but in the soil. An emerging set of agricultural practices that restore soil carbon are explored in our research note below. But another option is to ‘program’ plants to grow deeper, larger roots, which push more carbon into soils.

The Land Institute in Salina, Kansas has developed a grain called Kernza. It is derived from an ancestor of wheat. It is perennial, rather than requiring yearly replanting. Its roots reach 3-6x further into the soil than conventional wheat, which connotes 3-6x more carbon storage, and also promotes drought resistance. It is being grown across 2,000 acres today.

The US Department of Energy also has a Laboratory of Environmental Molecular Sciences, aiming to increase carbon transfer into the soil. One team has developed a strain of rice that emits less methane, as it contains a gene from barley, reducing the carbon that the plant moves underground, which in turn reduces the carbon that can be metabolized by anaerobic bacteria. Studies are underway to reverse the process and increase the carbon that crops move underground.

The Salk Institute for Biological Studies is based in La Jolla, California. It is undertaking the most elaborate program to bioengineer crops and other plants, to sequester up to 20x more CO2 than conventional crops. Deploying these plants across 6% of the world’s agricultural lands are said to potentially offset 50% of global CO2 emissions.

Salk’s Harnessing Plants Initiative started in 2017 and aims to grow “ideal plants” with greater efficiency at pulling CO2 from the air, deeper roots that store more carbon underground, and other superior agricultural properties. One pathway is to promote production of suberin, the carbon-rich polymer in cork (but also found in melon rinds, avocado skins and plant roots). This is waxy, water-resistant compound that degrades very slowly, thus remaining in the soil for centuries.

In 2019, Salk’s team discovered a gene, which determines whether roots will grow shallow or deep. It is called EXOCYST70A3, and affects the distribution of the PIN4 protein. PIN4 modulates the transport of auxin, a hormone that regulates root architecture. Different alleles of EXOCYST70A3 can increase root depth and plant resistance.

Technical readiness is the challenge for all of the bio-engineering methods discussed above. We generally begin integrating technologies into our models (first with high risking, later with lower risking) once they have surpassed TRL7. No bio-engineering method is there yet. Salk received a $35M grant in 2019, to accelerate its work, but prototype crop variants (corn, soybean, rice) are still not foreseen for five years. More pessimistically, scientists at RIPE have said it could take 15-years to deploy enhanced crops in the field. So while we will track this technology, it is not yet moving our models.

Greenhouse gas: use CO2 in agriculture?

Enhancing the concentration of CO2 in greenhouses can improve agricultural yields by c30%. It costs $4-60/ton to supply this CO2, while $100-500/ton of value is unlocked. Shell and ABF have already under-taken projects, while industrial gas and monitoring companies can also benefit. But the challenge is scale. Around 50Tpa of CO2 is supplied to each acre of greenhouses. Only c10% is sequestered. So the total CO2 sequestration opportunity may be limited to around 50MTpa globally.

This 8-page note explains the opportunity, progress to date and our conclusions.

Biofuels: better to bury than burn?

The global bioethanol industry could be disrupted by a carbon price. Between $15-50/ton, it becomes more economical to bury the biofuel crop, rather than convert it into biofuels. This would remove 8x more CO2 per acre, at a lower total cost. More conventional oil could be decarbonized with offsets. Ethanol mills and blenders would be displaced. The numbers and implications are outlined in this 12-page report.


Nature-based solutions to climate change need to double annual CO2 uptake from plants in our models of decarbonization, using forests and fast-growing grasses (pages 2-3).

We profile the bioethanol industry, which is already using fast-growing grasses to offset 2Mbpd of liquid fuels. But our models suggest the economics, efficiency and CO2 intensities are weak (pages 4-6).

A first alternative is to reforest the land used to grow biofuels, which would carbon-offset 1.5x more oil-equivalents than producing biofuels (pages 7-8).

A more novel alternative is to bury the biomass, such as sugarcane or other fast-growing grasses, which could sequester 8x more CO2, with superior economics at $15-50/ton CO2 prices (pages 9-11).

Company implications are summarized, suggesting how the ethanol industry might be displaced, and quantifying the CO2 intensity of incumbents (page 12).

Disrupting Agriculture: Energy Opportunities?

Precision-engineered proteins are on the cusp of disrupting the meat industry, according to an exceptional, 75-page report, published recently by RethinkX. The science is rapidly improving, to create foods with vastly superior nutrition, superior taste and superior costs, by the early-2020s.

The energy opportunities are most exciting to us, after reading the report. If RethinkX’s scenarios play out, we estimate: direct CO2 savings of 400MTpa, enough to offset 10% of US oil demand; 2bcfd of upside to US gas demand; and enough land would be freed up to decarbonise all of US oil demand, or increase US biofuels production by 6x to c6Mbpd.

We would be delighted to introduce clients of Thunder Said Energy to the reports’ authors, Catherine Tubb and Tony Seba. Please contact us if this is useful.