A framework for documenting homoploid hybrid speciation
This strategy was initially applied to the documentation of the
homoploid hybrid origin of Ostryopsis intermedia (Wang et al.,
2021; also see commentary by Owens, 2021). Ostryopsis is a member
of the birch family (Betulaceae) and includes three species of shrubs –O. davidiana , O. intermedia , and O. nobilis – all
of which are endemic to China. The three species are mostly allopatric,
although O. intermedia and O. nobilis occasionally
co-occur in Southwestern China. As implied by its name, O.
intermedia is intermediate phenotypically relative to the other two
members of the genus, and an earlier analysis of admixture based on 20
microsatellites and 11 nuclear loci supported a hybrid origin
approximately 0.5-1.2 million years ago (mya), associated with
Quaternary climate oscillations (Liu et al., 2014).
In their more recent and comprehensive investigation of the system, Wang
et al. (2021) demonstrated that O. intermedia was reproductive
isolated from its parental species (criterion 1 of Schumer et al. 2014),
more robustly validated the hybrid origin of O. intermedia(criterion 2), and directly linked the origin of reproductive barriers
to hybridization (criterion 3). While we primarily focus on the latter
component of the study, the findings of the first two components provide
useful background information. Briefly, Wang et al. (2021) found that
the parental species were primarily isolated by two premating barriers
(differential habitat adaptation and flowering time), whereas the hybrid
was isolated from one parent (O. davidiana ) by habitat (mainly
soil iron concentration) and the other parent (O. nobilis ) by
flowering time. There also was a partial crossability barrier between
the hybrid and parents, with total cumulative isolation exceeding 0.87
in all comparisons, even without considering ecogeographic isolation.
Next, Wang et al. (2021) used whole genome sequencing data and standard
phylogenomic and admixture analyses (e.g., PhyloNet and ABBA-BABA tests)
to more robustly document the hybrid origin of O. intermedia , as
well as the timing of admixture. Consistent with the previous study, the
two parental species were found to have diverged ~9.8
mya, with their hybrid originating approximately 1.8 mya. While this
analysis was largely confirmatory of earlier work, such phylogenomic
studies are useful because they can detect previously unsuspected cases
of admixture (e.g., Green et al., 2010, and refine (Nevado et al., 2020;
Owens et al., 2023) or even reject (Zalmat et al., 2021; Goulet-Scott et
al., 2021) earlier hypotheses of homoploid hybrid speciation.
To show that hybridization contributed to the evolution of reproductive
barriers isolating O. intermedia from its parental species, Wang
et al. (2021) employed the following strategy (Figure 1). First, they
searched for candidate genes underlying the key habitat (soil iron
concentrations) and flowering time barriers. For tolerance to high soil
iron concentrations, a combined population genomic data set of O.
intermedia and O. nobilis was compared against O.
davidiana , whereas for flowering time, population genomic data ofO. intermedia and O. davidiana was compared againstO. nobilis . Candidate isolation genes were expected to be under
positive selection as indicated by Hudson–Kreitman–Aguade´ tests
(Hudson et al., 1987), outliers with respect to divergence (in the top
2.5% of number of fixed non-synonymous mutation sites), and have
functional annotations consistent with the putative control of variation
in soil tolerance or flowering time. Three candidate genes (ZIP5,
FRO4, and 018656 ) satisfied these criteria for the comparison ofO. intermedia + O. nobilis to O. davidiana ; all
were annotated as being involved in iron reduction. Similarly, for the
comparison of O. intermedia + O. davidiana to O.
nobilis, three candidate genes (LHY , PIE1 , andFPA ) were under positive selection and had annotations consistent
with their involvement in the control of flowering.
Second, Wang et al. (2021) analyzed the ancestry of O. intermediaalleles for each of these six genes. As expected, alleles forZIP5, FRO4, and 018656 were mainly derived from O.
nobilis , whereas those for LHY , PIE1 , and FPA were
mainly derived from O. davidiana . However, there were several
instances in which a few genotypes of O. intermedia were
heterozygous for alleles from both parental species, possibly due to
post-speciation introgression.
Lastly, Wang et al. (2021) validated the function of alleles for three
of the candidate premating barrier genes (FRO4 , ZIP5 , andLHY ) using a combination of gene expression analyses, enzymatic
activity assays (for FRO4 and ZIP5 ), and functional
complementation of mutations in Arabidopsis thaliana. In
transgenic Arabidopsis plants, alleles of FRO4 andZIP5 from O. intermedia or O. nobilis outperformed
those from O. davidiana under high iron concentrations, with
opposite effects under low concentrations, strongly supporting their
involvement in adaptation to different soil iron concentrations.
Likewise, transgenic plants with LHY alleles from O. intermediaor O. davidiana flowered earlier than those with O.
nobilis alleles, further confirming their contribution to the flowering
time barrier isolating O. intermedia from O. nobilis .