Genetic structure and trait variation within a maple hybrid zone underscore North China as an overlooked diversity hotspot

Genetic structure of the parental population

Based on the lnPD and ΔK values obtained using STRUCTURE, we identified two genetic groups within the DHS Acer population (Supplementary Fig. S1). The q value from STRUCTURE analysis represents the proportion of ancestral origin²⁸ (Fig. 2a). Among the 70 individual trees, 72.9% were assigned a q value smaller than 0.1 or larger than 0.9, thereby signifying a typical bimodal distribution (Fig. 2b). Individuals with q value greater than 0.9 and consistent genetic origin from the NEA region were defined as the NEA lineage (hereafter “NEA-DHS”), whereas those with values less than 0.1 and with consistent genetic origin from the SEA region were defined as the SEA lineage (hereafter “SEA-DHS”). Individuals with intermediate q value between 0.1 and 0.9 were defined as hybrid genetic types (hereafter “Hybrid-DHS”). Accordingly, we identified 27 SEA-DHS (38.6%), 24 NEA-DHS (34.3%), and 19 Hybrid-DHS (27.1%) (Fig. 2b).

Both principal coordinates analysis (PCoA) and NewHybrids analysis revealed patterns similar to those obtained using the STRUCTURE analysis. The PCoA results indicated that the SEA-DHS and NEA-DHS based on STRUCTURE assignment were divergent and clustered along the first axis of the ordination plot, whereas the Hybrid-DHS individuals were found in an intermediate position of the ordination space between the SEA-DHS and NEA-DHS groups (Fig. 2c).

Newhybrids analysis indicated that 65.7% of individuals were of the parental type (30% SEA-DHS and 35.7% NEA-DHS), and 4.3% were F₂ hybrids. Among the remaining unclassified 21 individuals, 12 could be classified as F₂ individuals, and two appeared to be NEA-DHS backcross types with a posterior probability criterion of 0.55.

A comparison of the assignment results obtained based exclusively on the DHS dataset (DHS-only: 70 individuals and 11 loci), and that of the entire species range combined with data from a previous phylogeographic study (whole-range: 1278 individuals and 6 loci, Guo et al.¹⁰) revealed broadly similar patterns. Based on the whole-range dataset, 27% and 31% of the individuals were SEA-DHS and NEA-DHS, respectively (Supplementary Fig. S2). In this regard, given that for STRUCTURE assignment analysis, a larger number of assessed loci has been reported to be more effective than increasing the number of sampled individuals^29,30, we have only reported the results obtained based on the DHS-only dataset.

Genetic analysis of the offspring population

Among the 410 analyzed seeds, 198, 170, and 42 were collected from the SEA-DHS, Hybrid-DHS, and NEA-DHS maternal trees, respectively (Supplementary Table S1). Similar to the parental population, a bimodal pattern was observed in the offspring population (Fig. 2b, Supplementary Table S1). The NEA-DHS maternal trees produced 97.6% pure NEA-DHS seeds, and the remaining seeds (2.4%) were of the backcross type with a high genetic proportion of NEA (q > 0.85). The SEA-DHS maternal trees produced 72.2% pure SEA-DHS seeds, 10.1% pure NEA-DHS seeds, and 17.7% Hybrid-DHS seeds (Fig. 2a). Almost all seeds (q value > 0.5) produced by the SEA-DHS were obtained from a single tree, which was identified as SEA-DHS based on the DHS-only dataset, although it was indicated to be Hybrid-DHS based on the whole-range dataset. The Hybrid-DHS maternal trees produced 17.6% pure SEA-DHS seeds, 57.6% pure NEA-DHS seeds, and 24.7% hybrid seeds.

Flowering phenology

The sexual system of Acer has four phenotypes: duodichogamous, protogynous, protandrous, and male³¹. Hence, there are three functional sex types: (1) “Male I” flowers open earlier than “Female” flowers, with mature stamens, no style, and ovary; (2) “Female” flowers have mature pistils, short filaments, and indehiscence anthers; (3) “Male II” flowers open later than “Female” flowers, with mature stamens, ovaries, and separated stigmas. Duodichogamy is characterized by “Male I,” “Female,” and “Male II” types; protandry by “Male I” and “Female” types; and protogyny by “Female” and “Male II” types³¹.

During the flowering season, we monitored a total of 10,074 flowers produced by 29 trees (Fig. 2d), among which one tree (SEA-DHS) was protandrous, four trees (three Hybrid-DHS and one NEA-DHS) were protogynous, and the remaining 24 trees were duodichogamous. We observed that the blooming phenology of SEA-DHS and NEA-DHS differed significantly to most assessed phenological indices, with a single exception being a marginally significant difference in the peak blooming time of Male I (Table 1). Compared with NEA-DHS, SEA-DHS were characterized by significantly later flowering phenology, with Male I commencement and cessation of blooming being on average two and three days later, respectively. Similarly, the commencement, peak, and cessation of Female occurred later by averages of 4, 4, and 5 days, respectively, whereas those of Male II occurred later by 5, 4, and 5 days, respectively. Furthermore, the duration of blooming was significantly longer in the SEA-DHS group than in the NEA-DHS group by three days. In the case of Hybrid-DHS, the values obtained for all assessed phenological indices were intermediate between those of the two parental types. Among these, the values of the six indices differed significantly from one or the other parental types, with the majority (5/6) differing from those of the SEA-DHS. Thus, phenologically, Hybrid-DHS appeared to be closer to NEA-DHS.

However, despite the differing phenology of the SEA-DHS and NEA-DHS, we observed instances of overlap in the blooming periods of male or female flowers in one genetic type with those of flowers of the opposite sex in another genetic type. For example, the peak of Female among NEA-DHS (11.67 ± 0.67) was found to coincide with the peak of Male I (11.44 ± 1.06; p = 0.879) in SEA-DHS. Similarly, Female blooming in the SEA-DHS peaked (16.11 ± 1.09) just 1 d after the peak of Male II (15.50 ± 0.43) in the NEA-DHS (p = 0.667), which at this time still retained an abundance of male flowers in bloom. In contrast, we detected no overlapping phenology with respect to the blooming of Male I of NEA-DHS or Male II of SEA-DHS with the Female in another genetic type.

Morphological variation of leaves and fruit

Leaves Among the eight leaf indices, all except InfectionRatio were significantly different between lineages. Generally, the leaves of NEA-DHS were found to have seven lobes, whereas those of SEA-DHS were typically five lobed (Lobes#), thereby contributing to significantly larger leaves in NEA-DHS than in SEA-DHS (TotalArea). Furthermore, NEA-DHS leaves had shorter and wider central lobes (CentralLength and CentralWidth), as well as an earlier and narrower inflection of the central lobes (InflectionLength and InflectionWidth), compared with those of SEA-DHS (Table 2). Six indices had correlation coefficients of less than 0.7, which were used for principal component analysis (PCA) analysis (Supplementary Table S2). The first two axes of the PCA were found to explain 63.7% of the variation in leaf morphology (Fig. 3a), with InflectionLength, CentralLength, and CentralRatio contributing the most to the first axis (38.2%), whereas TotalArea contributed the most to the second axis (25.5%) (Supplementary Table S3). The leaves of SEA-DHS and NEA-DHS plants were largely clustered in separate groups (Fig. 3a). However, all indices were continuous variables with large overlaps between the lineages (Table 2). For example, NEA-DHS had a significantly larger leaf area (21.06–88.70 cm²) than SEA-DHS (11.34–70.09 cm²). The shape of the central lobe is another major leaf trait that distinguishes between the two species. NEA-DHS had a shorter and wider central lobe (CentralRatio:0.67–2.49), while SEA-DHS had a longer and narrower central lobe (CentralRatio:0.9–3.46).

With regard to Hybrid-DHS, the leaves were morphologically intermediate between those of the two parental types (Fig. 3a), as were the values of the assessed morphological trait indices (Table 2).

Fruits 11 indices of fruits were significantly different between lineages. NEA-DHS tend to be characterized by smaller fruits (FruitLength and FruitWidth), seeds (SeedLength, SeedWidth and JunctionWidth), and fruit wings (WingLength and WingWidth). Moreover, the seed wings of NEA-DHS fruits are typically oriented at an obtuse angle, whereas those of SEA-DHS fruits tend to be aligned at a right angle (FruitAngle). The length ratio of the wing and seed (Wing:Seed) was larger in NEA-DHS than in SEA-DHS (1.24 vs 1.06, respectively, Table 2). Eight indices had correlation coefficients of less than 0.7, which were retained for PCA analysis (Supplementary Table S4). The first two axes of the PCA explained 58.4% of the variation in fruit morphology (Fig. 3b), with JunctionWidth and SeedLength contributing the most to the first axis (35.1%), whereas SeedRatio and WingRatio contributed the most to the second axis (23.3%) (Supplementary Table S3). The fruits of SEA-DHS and NEA-DHS plants were largely clustered in separate groups, with most fruits of SEA-DHS having negative values in Axis 1, while most fruits of NEA-DHS having positive values (Fig. 3b). Both JunctionWidth and SeedLength in Axis 1 reflect the size of the seed. NEA-DHS had smaller seed (SeedLength: 0.63–1.21 cm, SeedWidth:0.43–0.75 cm), while larger seed in SEA-DHS (SeedLength:0.79–1.49 cm, SeedWidth:0.49–0.93 cm). All indices were continuous variables with large overlaps between the lineages (Table 2).

The morphology of Hybrid-DHS fruits was generally intermediate between that of the two parental types (Fig. 3b), as reflected in the values of the different morphological traits. The exceptions in this regard were FruitLength, WingLength, as well as two ratio indices (SeedRatio and WingRatio), with hybrid trees typically producing longer fruit with longer fruit wings (Table 2).

Ecological niche divergence between NEA and SEA

We found a positive correlation between q value from Structure analysis and altitude (Pearson’s r = 0.83, p < 0.0001) (Fig. 2d). Furthermore, we found significant differences in the altitudinal distributions of the three genotypes. For example, NEA-DHS was primarily distributed at the hilltop of the transect (altitude > 670 m), whereas SEA-DHS was clustered at the foothill (< 600 m), with a few individuals scattered at higher altitudes. The Hybrid-DHS group was found at intermediate altitudes between the parental groups, that is, from the midpoint of the transect to the hilltop (570–730 m).

At the spatial scale of the species range, habitats of NEA and SEA lineages were significantly divergent in 11 out of 19 bioclimatic variables (Supplementary Table S5). Among the 11 temperature-related variables (Bio1–Bio11), only Mean Temperature of Warmest Quarter (Bio10) was not significant between the NEA and SEA habitats. Among the eight precipitation-related variables (Bio12–Bio19), only Precipitation Seasonality (Bio15) was significant between the lineages. Six variables had correlation coefficients of less than 0.7, which were used for PCA analysis (Supplementary Table S5). The first two axes accounted for 82.3% of the variance, with Mean Diurnal Range (Bio2) and Temperature Seasonality (Bio4) contributing the most to the first axis (57.19%), whereas Mean Temperature of Wettest Quarter (Bio8) contributed the most to the second axis (25.11%). The NEA and SEA habitats were mainly divergent along the first axis (Supplementary Fig. S3). Compared with SEA habitats, most NEA habitats were characterized by lower temperatures and larger seasonality of both temperature and precipitation (Supplementary Table S5, Supplementary Fig. S3).

Source: Ecology - nature.com