Propyl-cannabinoids ("the varins")--CBDV, THCV, CBGV, CBCV, etc.--are structurally similar to their pentyl counterparts (CBD, THC, CBG, CBC), but have 3 carbon side chains instead of 5 (C3 vs. C5) and are exceedingly rare in cannabis. Both propyl and pentyl cannabinoids are produced by the known cannabinoid synthases responsible for general chemotype, but propyl compounds have a different upstream precursor (divarinolic acid vs. olivetolic acid). Knowing the promise of propyl cannabinoids, we started searching for varin-rich plants when our company was founded in 2015. The greatest impediment to that search was the lack of reference standards that enable laboratories to identify varin-rich individuals. This changed in 2018, when several Oregon labs started testing for THCVA, enabling us to more rapidly screen large populations of type I plants rumored to contain elevated amounts of the compound--these included Durban Poison, California Orange, Skunk #1, old Haze lines curated by generations of growers, dozens of Jack the Ripper crosses, and any other plants brought to our attention by citizen-cannascientists. We had to start with type I plants because CBDVA and CBGVA standards were not available, a fact that plays an important role in this larger story of our varin breeding program and will be elucidated below. After screening thousands of plants, we learned a number of things about the chemical analysis of early leaves and natural varin production levels in pentyl-dominant cannabis at different life-stages--the most important finding, however, was that those varieties rumored to contain propyl cannabinoids really don't. Not all received wisdom stands up to analytical scrutiny! Haystack needles are hard to come by when searching for varins, but once we finally found one in late 2018, the project shifted from exploratory to a systematic breeding effort--something our team is really good at--at the same time as CBDVA standards became more widely available to commercial labs.

As we have said before, success like this is simple in cannabis breeding. Step 1. Start thousands of seeds in year 1, grow them to sexual maturity (2-3 feet tall), then test each one. If lucky, you will find 2-3 plants that contain trace amounts of the cannabinoid you seek (if not, back to step 1). Step 2. Make clones, isolate two copies, and reverse one to self-pollinate the other. 3 months later, collect seed. Step 3. Grow out said seed and repeat the first step, keeping only the plants expressing slightly higher concentrations of the compound you seek. Step 4. Repeat steps 1-3 until you have achieved maximum production of your compound of interest (maxaa are variety-specific). Step 5. Outcross the best maxaa plant with your desired chemotype (CBD, CBG, CBC, or THC), then repeat steps 1-3 again 3 more times. Voila. CBDV, CBGV, CBCV, and THCV plants magically become stabilized.

Once the initial varin-rich germplasm was identified, we used it to pollinate a large number of our established production mothers, as well as other plants in our R&D program. More details on this process can be found in this presentation. The F1 seed created in those initial crosses were then grown out and chemotyped via HPLC. The top 5 varin-containing phenotypes from each population were then self-pollinated in isolation chambers and analyzed at final flowering. The individuals with the best flavor, structure, and propyl content were selected, F2(S1) seeds collected, and then the real work began.

Where the F1 population is relatively uniform in medium-to-low-varin content, the F2 generation contains the maximum diversity in chemotype imaginable--to identify a "keeper" requires large population sizes and well-timed testing after plants attain sexual maturity. On average, appropriately varin-rich plants only appeared 1.3% of the time in these large populations. In other words, to find 10 plants in any given population with the "right" chemotype requires an initial population of 1300--and for chemical testing purposes, they must be 2-3 feet tall before assessment. To say that these projects required substantial time, space, and resources to develop would be a tremendous understatement--and it is important to remember we were only able to get to this phase after years of searching for the initial germplasm.

Once identified via HPLC, F2(S1) "keepers" were subjected to another round of inbreeding to produce F3(S2) seeds, where all resulting individuals contain high levels of propyl cannabinoids and are approximately 88% homozygous (the equivalent of an F10 full-sib breeding strategy). These line-bred and stabilized seeds serve as the foundation of our new varin-rich mothers (who undergo non-GMO ploidy transformation to become tetraploids) and day-neutral pollen donor lines.

The main question that emerged from the first, long phase of our varin development program was: why is the rate of "keepers" so low in the F2(S1) population? The answer to this question did not emerge until very recently and required the most comprehensive genomic investigation of cannabis to date. We focused our efforts on one specific F2 population to identify the genetic associations with propyl-rich traits. First, we deep-sequenced the original parental lines and the selected F1 using PacBio's Seqwell II HiFi platform, then used whole genome shotgun sequencing for over 300 F2 siblings to generate data for a Genome Wide Association Study (GWAS). Concurrently, the same population was subjected to analysis using the Eurofins "CannSNP90" SNP array, phenotyped by our plant breeders, and subjected to additional chemical analysis (HPLC). By triangulating the resulting data from both projects with our in-house whole genome library, we have successfully identified the relative genomic locations of varin production genes and are now able to design molecular markers to accelerate future breeding efforts (learn more by watching the CannSNP90 webinar here). This radical advancement allows us to screen seedlings in future projects instead of mature plants, which saves considerable time, energy, and space over our initial process (though it was mandatory to develop the markers). But back to that originating question--why is the keeper rate so low in F2s? The answer: recombination (or, in this case, lack thereof).

Our whole genome sequencing efforts have suggested that the genomic regions associated with varin-production in the population under investigation are located on two separate chromosomes (3 and 7). The major cannabinoid synthases (THCAS, CBDAS) are present on chromosome 7, which means there is a strong association in this population between an active THCAS gene and propyl-production characteristics. In other words, a recombination event must occur for this trait to be successfully introgressed in a CBD or CBG dominant plant; this means a small portion of chromosome 7 becomes unlinked from its former home near THCAS and reattaches in the same region of chromosome 7 near CBDAS or our inactive THCAS used to create CBG chemotypes . These recombination events are very rare and help explain why it is so difficult to create a CBDV-rich variety when the initial germplasm used for trait introduction is a type I plant.

If this sounds like a foreign language, that's OK--it's our job to understand what it means, what the implications are for a breeding program, and how to stabilize these traits using genomically-informed traditional breeding practices. For those who understand, it's obvious that this particular project is a significant cannabis breeding achievement. We learned SO much valuable information about relative recombination rates in cannabis from this study and look forward to sharing the results when published later in 2021--the findings have profound implications for future breeding efforts and trait introgression in cannabis.

What does this mean for farmers? CBDV is a rare, high value cannabinoid that will be in demand once available to consumers and product manufacturers. For the first time in the history of cannabis farming, triploid, field-ready, CBDV-rich plants are possible and worrying about final compliance testing and cross-pollination are relegated to the dust-bin of history. Final flowers range from 40%-55% varin compounds as a proportion of the total cannabinoid fraction, which allows high-content plants to remain below the total 0.3% THC threshold (THCV and THCVA are NOT included in total THC).

The other unique aspect of growing these plants is the relative expression of propyl cannabinoids throughout the flowering process: varin compounds predominate early in the flowering cycle and only equalize with their pentyl counterparts around harvest. This means the total amount of THC present is substantially less than THCV during critical pre-harvest compliance testing. The time-course data provided below highlights this phenomenon; using the standard, USDA-recommended sampling protocol (28 days before harvest), you should expect to see 1.9x more CBDV than CBD and 2.8x more THCV than THC--with no chance of compliance failure. These ratios trend towards equal 1:1 profiles, but do so over the life-course of the plants.