There are many ways to determine an enzyme’s function; for example, targeted mutagenesis and protein structure studies can be employed. However, one approach that has perhaps not been explored enough is combining ancestral sequence reconstruction with producing and testing these reconstructed enzymes in the laboratory.

Research conducted in Wageningen demonstrates the power of this approach. Robin van Velzen’s research group at Wageningen University & Research has been studying cannabinoid oxidocyclases for some time. These enzymes are responsible for producing cannabinoids such as (−)-trans-Δ9-tetrahydrocannabinolic acid (THCA), cannabichromenic acid (CBCA) and cannabidiolic acid (CBDA) in cannabis. These cannabinoids are converted into the bioactive substances THC, CBC and CBD when heated.

Despite extensive research into these enzymes, the researchers were left with two questions. Firstly, they wondered about the evolutionary history of the various products within this enzyme family. Secondly, they wanted to know what determines which product the enzyme produces. To address these issues, postdoctoral researcher Cloé Villard, who is now working at the University of Amsterdam, joined forces with Van Velsen, combining her biotechnological expertise with his phylogenetic knowledge.

Little Red Riding Hood

First, they used ancestral sequence reconstruction to calculate what the enzymes probably looked like millions of years ago. “You can compare it to textual criticism, for example, when trying to determine the original version of a story based on the various existing versions,” says Villard. ‘You collect as many stories as possible which have changed slightly over time in each country, and try to reconstruct the original story based on the differences and similarities.’

‘We do the same thing with DNA,’ says Villard. In order to identify the ancestral enzymes, she collected as many DNA sequences as possible from cannabinoid oxidocyclases and related enzymes. These included Berberine Bridge-Like enzymes found in hops, which are closely related to cannabis. Villard then used a phylogenetic tree to determine the most likely nucleotide sequence in the ancestral enzymes.

The group was particularly interested in three ancestral enzymes: the most recent ancestor of THCA synthase and CBCA synthase; the ancestor of all known cannabinoid synthases; and the ancestral enzyme of the ancestor of all cannabinoid oxidocyclases and their closest relatives present in hops. Villard ‘resurrected’ these enzymes by synthesising their DNA and then expressing them in yeast. She then studied the activity of the resurrected enzymes. Only the oldest ancestral enzyme had difficulty converting CBGA (cannabigerolic acid) into cannabinoids.

Different perspective

Villard also examined hybrid intermediate forms of successive ancestral enzymes by gradually adding changed amino acids to the oldest variant. She thus gave the ancestral enzyme shared with hops the active centre of the last common cannabinoid oxidocyclase of cannabis. She then added the missing amino acids from the younger ancestor, finally replacing the cannabis-specific region where an essential cofactor binds.

The first hybrid enzyme converts CBGA into CBCA only, which none of the modern cannabinoid oxidocyclases can do. Subsequent hybrid enzymes convert CBGA into CBCA, THCA and CBDA, with the latter doing so slightly faster.

‘This allowed us to view the enzyme from a different perspective. We were no longer considering what we could add to the enzyme, but rather how it developed,’ says Villard. ‘Going back in time gave us new insights, which made it possible to focus on parts of the enzyme that we had previously overlooked entirely.’

Easier

Professor Todd Barkman, who researches protein evolution in plants at Western Michigan University in the US, thinks the study is cool. ‘This study shows quite convincingly that the current enzymes for the production of the cannabinoids THCA, CBCA, and CBDA could have originated from a single ancestral enzyme,’ he says. ‘In addition, they show that, after the duplication of this single enzyme in the more recent ancestors of cannabis, these daughter enzymes became increasingly specialised.’

This also explains why researchers have been unable to convert THCA synthase into CBDA synthase. As well as the active centre, the rest of the THCA synthase enzyme focuses on producing THCA.

This study provides new insights and has implications for the industrial production of cannabinoids. Yeast found it much easier to produce the ancestral enzymes in large quantities than their modern descendants. Furthermore, the researchers were able to easily modify the ancestral enzymes. “This opens the door to targeted industrial production of cannabinoids,” says Barkman.

Villard, C.I. et al. (2025) Plant Biotechnology Journal, DOI: 10.1111/pbi.70475