Cosexuality Reduces Pollen Production and Fitness in Cannabis sativa L

Abstract

Cannabis sativa L. is cultivated globally for its cannabinoid-dense inflorescences. Commercial preference for sinsemilla has led to the development of methods for producing feminized seeds through cross-pollination of cosexual (masculinized) female plants. Although the induction of cosexuality in Cannabis plants is common, to date, no work has empirically tested how masculinization of female Cannabis plants impacts male flowering, pollen production, pollen fitness, and related life-history trade-offs. Here, we cultivated a population of Cannabis plants (CFX-2) and explored how the route to cosexuality (drought vs. chemical induction) impacted flowering phenology, pollen production, and pollen fitness, relative to unsexual male plants. Unisexual males flowered earlier and longer than cosexual plants and produced 223% more total pollen (F_2,28 = 74.41, p < 0.001), but per-flower pollen production did not differ across reproductive phenotypes (F_2,21 = 0.887, p = 0.427). Pollen viability was 200% higher in unisexual males and drought-induced cosexuals (F_2,36 = 189.70, p < 0.001). Pollen non-abortion rates only differed in a marginally significant way across reproductive phenotypes (F_2,36 = 3.00, p = 0.06). Here, we demonstrate that masculinization of female plants impacts whole-plant pollen production and pollen fitness in Cannabis sativa.

Keywords: Cannabis sativa; cosexual; dioecy; phenology; pollen; sexual lability.

Figure 1

Conceptual image showing the experimental design. Each experimental group experienced one of three routes to masculinity: karyotype-based (unisexual males), drought-induced cosexuals, or chemically induced cosexuals. Created with BioRender.com.

Figure 2

Flowering phenology of unisexual males, drought-induced cosexual plants, and chemically induced cosexual plants. Unisexual males flowered significantly earlier (F_2,36 = 79.52, p < 0.001) and longer (F_2,36 = 68.61, p < 0.001) than cosexual plants. The end of the flowering period differed significantly (F_2,36 = 15.21, p < 0.001) between chemically induced cosexuals and both unisexual males and drought-induced cosexuals.

Figure 3

Correlations between pollen characteristics (pollen abundance, non-abortion rate, and viability) and related phenological traits (stared, duration, and end of anthesis) across three reproductive phenotypes. Blue lines indicate the direction of association.

Figure 4

Plots of mean values (±SD) of pollen characteristics across three reproductive phenotypes. Unisexual males differed in their pollen abundance (F_2,28 = 74.41, p < 0.001) from both drought-induced (Tukey’s HSD, p < 0.001) and chemically induced (Tukey’s HSD, p < 0.001) cosexual plants. Chemically induced cosexual plants differed in their pollen viability (F_2,36 = 189.70, p < 0.001) from both drought-induced cosexual plants (Tukey’s HSD, p < 0.001) and unisexual males (Tukey’s HSD, p < 0.001).

Figure 5

Image of plants from the experiment showing differences in floral morphology used to assign sex. Unisexual male plants (a) developed dense clusters of male inflorescences. Unisexual female plants (b) developed colas with protruding stigmas. Cosexual female plants (c) initially developed female flowers, and after undergoing one of two experimental treatments, later developed male inflorescences at apical branching junctions.