Subgenome composition but not ploidy is associated with local ocean salinity
Studies in other systems indicate that an increase in ploidy tends to be associated with increased salt tolerance both within and between closely related species (Chao et al., 2013; Wu et al., 2019; Yang et al., 2014). However, when ploidy is considered alone, there is no detectable difference in local ocean salinity between diploid and triploid P. vaginatum (Fig. 5). Instead, copy number of the P. vaginatumsubgenome may be the driving factor. Gene dosage effects can have a dramatic influence on phenotypes and increased copy number of salt tolerance genes has been implicated in the evolution of salt tolerance in halophytes (Oh, Dassanayake, Bohnert, & Cheeseman, 2012). Variation in gene copy number can also occur on a whole genome level in populations of hybrids that vary in the number of subgenomes they received from each parental species. Other studies have investigated reciprocal allotriploid hybrids (e.g. 1:2 vs 2:1), similar to those we describe here, which are often more similar to the parent from which they received two copies (Betto-Colliard, Hofmann, Sermier, Perrin, & Stöck, 2018; Harvey, Fjelldal, Solberg, Hansen, & Glover, 2017; Tan et al., 2016; Yao, Gray, Auger, & Birchler, 2013).
Our results support the hypothesis that the other unknown subgenome present in the coarse-textured ecotype is not as adapted for saline environments as is P. vaginatum (Fig 5). Plants with two copies of the P. vaginatum subgenome (fine-textured, and 2:1 triploids) were found in locations where the local ocean salinity was on average ~15% higher than those with only one copy (1:1 diploids and 1:2 triploids). The effect of increasing counts of the unknown subgenome was weaker and in the opposite direction (slightly less ocean salinity with each copy). We cannot say for certain whether this weaker effect is due to increased ploidy or subgenome dosage effects, or both. If it is a ploidy effect then it is the opposite of predictions that polyploids should be more salt tolerant.
The large effect of the number of P. vaginatum subgenomes is consistent with our finding that coarse-textured hybrids are likely “general-purpose genotypes” because their local ocean salinity is not only lower, but also more variable. Furthermore, unlike most other studies of reciprocal triploids, we can rule out maternal subgenome effects because all of the coarse-textured hybrids carry the same cpDNA haplotype, implying the same maternal parent species for each cross.
Although the trend we observe between P. vaginatum subgenome copy number and local salinity is striking, environmental associations are necessarily correlational in nature and must be confirmed with controlled salt tolerance assays and transcriptomics. It is also possible that unknown demographic or introduction history could contribute to these patterns. Additionally, our sea surface salinity dataset does not have the resolution to identify the microhabitat variation likely experienced by plants in the wild (for example, in an estuary with a steep salinity gradient). While higher resolution salinity data exist for some coastal areas, temporal and methodological variation between datasets make them unsuited for comparison across the geographic scale of this study.