A transverse section of stem wood from the researchers’ greenhouse-grown poplar tree.
This image shows a transverse section of stem wood from the researchers’ greenhouse grown poplar tree.
Ilona Peszlen

Humans have domesticated numerous crops, but trees have largely escaped us. This is because breeding new tree varieties takes many decades.1 Now, a team has combined machine learning and advanced CRISPR technology to create trees with desirable traits for both industry and the environment in a fraction of the time. 

Wood consists of about 25 percent lignin,2 but the paper and fiber industries remove it because it lowers the quality of their products—for instance, by yellowing paper that comes in contact with air. Lowering the lignin content in the trees themselves would reduce the need for that process, said genome editing specialist Rodolphe Barrangou at North Carolina State University (NCSU) and coauthor of the paper, which was recently published in Science.3

Barrangou and his colleagues used a predictive machine learning model to identify genes in poplar trees (Populus trichocarpa) that they could modify to create ideal traits for fiber productions, including increase cellulose-to-lignin ratio. Among the thousands of editing strategies offered by the model, the team chose seven, which all involved altering between three to six genes. 

To test the strategies experimentally, the team used multiplex CRISPR, a genetic engineering technique that targets multiple genes simultaneously to generate 174 engineered tree lines. After growing inside a greenhouse for six months, the CRISPR trees showed significant improvements in desirable wood properties compared to wild type trees. In the most drastic cases, the lignin content was reduced by 29 percent and the cellulose-to-lignin ratio increased by 228 percent.

See also “A Pioneer of The Multiplex Frontier”  

“[This was] certainly pushing the boundaries in terms of what has been done,” said co-author Jack Wang, a forest biotechnologist also at NCSU. Wang and Barrangou are co-founders of TreeCo, a company enhancing trees with CRISPR. No one has targeted so many genes concurrently in trees before, he said.

The genetic changes, however, meant that many of the edited trees grew much slower. Nevertheless, the team’s analysis suggested that the CRISPR-edited wood would likely boost fiber production efficiency and reduce the carbon footprint, since the energetic and chemical input required to remove excess lignin would decrease.

The process of removing lignin from wood is very demanding “in terms of materials, solvents energy, pressure,” and generates chemical waste, said Leuphana University of Lüneburg’s Vânia Zuin Zeidler, who studies sustainable chemistry and wasn’t involved in the work. That makes this study very relevant, especially considering the climate crisis, she said.

But the approach may be less convincing once the trees go outside, said Wout Boerjan, a plant biotechnologist at Ghent University who also didn’t participate in this study but has previously collaborated with Wang. Greenhouse conditions can be very different from the field, so he worries that growth would be further affected. 

See also “Certain Tree Species Are More Susceptible to Death by Lightning”  

The team acknowledged that risk. Doing this in the lab is just one initial milestone, Barrangou said, and he and his colleagues intend to test their new tree lines in the field. This is also why they created multiple lines for testing. 

“It’s great to have these prototypes,” said Boerjan about the large genetic variation achieved by the team. “I hope some of these lines will give reasonably good performance in the field.” 

References

  1. Lantz CW. Genetic Improvement of Forest Trees. In Bonner FT, Karrfalt RP. The Woody Plant Seed Manual, USDA FS Agriculture Handbook 727. 2008:39-56. 
  2. Novaes E, et al. Lignin and biomass: A negative correlation for wood formation and lignin content in trees. Plant Physiol. 2010;154(2): 555–561.
  3. Sulis DB, et al. Multiplex CRISPR editing of wood for sustainable fiber production. Science. 2023;381(6654):216-221.