DOI: https://doi.org/10.21203/rs.3.rs-2575763/v1
Identification of natural hybrids considered as endangered species is of vital importance in biodiversity conservation and taxonomy, as natural hybrids will usually waste the conservation resource and obscure the divergences between distinct species. During the field surveys in the Sanqing Mountain, we found the endangered endemic species, Ilex sanqingshanensis, strictly co-occurred with I. ficoidea and I. pernyi and then supposed a hybrid origin for this taxon. Combing the molecular analyses of ITS and cpDNA (petA-psbJ + psbA-trnH) with the morphological analyses of eight leaf characters, we confirmed this taxon to be a hybrid between I. ficoidea and I. pernyi and accepted it as I. × sanqingshanensis. Despite the presence of intermediacy in morphology, this hybrid is sharply distinct from the two parents in all tested traits, misleading the botanists to treat it as a species. Considering the inadequacies of morphological distinctions in distinguishing holly hybrids, we have emphasized the necessity of molecular evidence for erecting Ilex species.
Natural hybridizations commonly arise in vascular plants across many different families and floras when infraspecific populations or closely related species come into contact (Ellstrand et al. 1996; Rieseberg 1997; Whitney et al. 2010; Kadereit 2015). These processes play a crucial role in the formation and maintenance of species (Seehausen 2004; Arnold and Martin 2009; Soltis and Soltis 2009; Nolte and Tautz 2010; Abbott et al. 2013). It is estimated that at least a quarter of plant species are involved in hybridization (Mallet 2005).
Whereas there are undoubtedly species of hybrid origin (Barrier et al. 1999; Rieseberg 2006; Meier et al. 2017; Lamichhaney et al. 2018; Wang et al. 2021), it is inappropriate to assign taxonomic rank to each production of hybridization (Marczewski et al. 2016). To date, a large body of endemics spanning a relatively narrow range were proven to be hybrids rather than species (Wiegleb and Kaplan 1998; Zha et al. 2010; Shin et al. 2014; Shi et al. 2016; Zhang et al. 2020 a, b; Lyu et al. 2021; Ao et al. 2022). These hybrids usually have poorer fitness than parental species, and depend on the repeated hybridization between parental species to maintain the populations. As a consequence, they are ordinarily considered as “negative assets” in biodiversity estimation and conservation (Allendorf et al. 2001; Jackiw et al. 2013). For taxonomy, accurate species delimitation is the bedrock and the guarantee. Natural hybrids have long been seen as “troublemakers” by taxonomists as they usually show an intermediate state in a part of characters probably making the morphological divergences between the parents less obvious (Stebbins 1957; Wagner 1969; Dejaco et al. 2016). Thus, it is essential to uncover the hybrid status of potential “disguisers” which should not be attributed with species rank.
Due to the co-occurrence with potential parental species, natural hybrids are often noticed during field investigations at first. Morphological intermediacy further divulges clues of hybridizations for its common application in the identification of natural hybrids (Marczewski et al. 2016). As the development of molecular approaches, there are more available tools to help to unmask hybrids, including but not limited to incomplete ITS (internal transcribed spacer region) concerted evolution (Grimm & Denk 2008; Kou et al. 2017), cytonuclear disequilibrium (Hodkinson et al. 2002; Yu et al. 2014), microsatellite polymorphisms (SSRs; Schroeder & Fladung 2010; Zhang et al. 2020b), heterozygous alleles in single or low-copy nuclear genes (Liao et al. 2015, 2021), and single nucleotide polymorphisms (SNPs; Väli et al. 2010; Zheng et al. 2021). Among these approaches, ITS allied with several plastid makers has a broader application because of the virtues of both high practicality and simplicity, for instance, in the genus Ilex L. (Son et al. 2009; Shi et al. 2016).
Ilex, the sole genus of Aquifoliaceae, consists of at least 14 sections (Yang et al. 2022) and holds more than 600 species as well as a lot of interspecific hybrids both naturally occurring and cultivated (Galle 1997; Powell et al. 2000; Loizeau et al. 2005; Chen et al. 2008). To date, the confirmed inartificial Ilex crosses have only been found to be intra-sectional; however, most natural hybrids are concentrated in the largest section, I. sect. Ilex, which contains over 100 species and has a center of diversity in East Asia (Yang 2020). During the last 40 years (especially the 1980s), I. sect. Ilex had experienced a rapid growth of species number from ca. 50 to over 100. These newly described species commonly are endemics. They usually have narrow native range and are listed in the ICUN red list of endangered species (ICUN 2022), e.g., I. sanqingshanensis W.B.Liao, Q.Fan & S.Shi (Fig. 1) which only occurs in the Sanqing Mountain, eastern China. During the field investigations conducted in 2018 and 2020, we however found that I. sanqingshanensis always co-occurs with the other two members of I. sect. Ilex, i.e., I. ficoidea Hemsl. and I. pernyi Franch. Specifically, I. sanqingshanensis only grows at the elevation of 1300–1600 m where I. ficoidea and I. pernyi have converged (Fig. 1). The discoveries in situ reminds us of the probable hybrid origin of I. sanqingshanensis.
In this study, we analyzed a multi-gene dataset including ITS and two chloroplast DNA (cpDNA) regions (petA-psbJ and psbA-trnH) and a morphological dataset covering eight leaf traits. We aim to: (1) test the hypothesis of the hybrid origin of I. sanqingshanensis, (2) fix the parental species if it is indeed a product of hybridization, and (3) give a reasonable identity to the target, a hybrid species or just a hybrid.
Ilex sanqingshanensis grows in the Sanqing Mountain, ShangRao City, Jiangxi Province, eastern China, with only single population discovered. During the field investigations in 2020, we totally found 22 individuals (labeled as S1–S22 successively) of I. sanqingshanensis, 19 individuals may grow from seeds while the other three (S4, S5 and S10) seem to be clonal sprouts. Thus, we only sampled the 19 individuals. We also collected 20 and 19 individuals of the putative parents (I. ficoidea and I. pernyi), respectively, keeping two individuals of the same species over 100 m apart. The other 17 hollies (e.g., I. cornuta Lindl. & Paxton and I. latifolia Thunb.) occurring in the Sanqing Mountain were also taken into account (sampled or downloaded the sequences from GenBank). Species delimitation was in accordance with the descriptions given by Flora of China (Chen et al. 2018), Shi et al. (2015), and Yang (2020).
Three DNA loci, the internal transcribed spacer (ITS) and the chloroplast psbA-trnH and petA-psbJ regions, were sequenced for phylogeny reconstruction, with the primers 17SE and 26SE (Sun et al. 1994), psbAF and trnHR (Sang et al. 1997a), and petA and psbJ (Shaw et al. 2007) employed, respectively. The detailed protocols of DNA extraction, PCR amplification, and DNA sequencing followed that of Jiang et al. (2017) and Yang et al. (2017). For all the individuals of I. sanqingshanensis, direct sequencing produced the superimposed chromatograms and unreadable peaks on most sites at the ITS maker, hence we implemented cloning sequencing to purify the PCR products. We conducted ligation reactions with a pMD19-T&A cloning kit (Takara, Dalian, China) and selected at least ten positive clones for each individual for sequencing. Voucher information and GenBank accession numbers of the sequences are shown in Table S1.
Raw chromatograms were evaluated in Sequencher 5.4.6. Then sequences were aligned in MAFFT (Katoh and Standley 2013) with manual adjustment if necessary. The ITS dataset comprised 102 sequences of Ilex sanqingshanensis, the putative progenitors I. ficoidea and I. pernyi, and the other 17 Ilex species which were also distributed in the Sanqing Mountain based on specimens’ records. Clones that showing recombined sequence were discarded to exclude noise from PCR reactions (Bradley and Hillis 1997; Sang et al. 1997b). Ilex viridis Champ. ex Benth. from the section Paltoria (the basal lineage of Ilex revealed by Yang et al. 2022) was treated as outgroup. Because of the poor intra-sectional revolution of the chloroplast makers (Manen et al. 2010; Shi et al. 2016), the combined cpDNA dataset only contains the sequences of I. sanqingshanensis, I. ficoidea, and I. pernyi with their co-occurring relatives I. litseifolia Hu & T.Tang (section Lioprinos), I. pedunculosa Miq. (Lioprinos), and I. viridis (Paltoria) selected as outgroups.
The molecular phylogenetic relationships within Ilex were reconstructed based on maximum likelihood (ML) and Bayesian inference (BI) approaches. The optimal nucleotide substitution models for ML and BI analyses, namely GTR + G for ITS dataset and TPM1uf + I for chloroplast dataset, were found with jModeltest 2.1.7 using the Bayesian information criterion (Darriba et al. 2012). The ML analyses were conducted in the RAxML-HPC2 (Stamatakis 2014) at the CIPRES Science Gateway (http://www.phylo.org; Miller et al. 2010) with 1000 bootstrap replicates followed by a search for the best-scoring tree in a single run. BI analyses were performed by MrBayes v.3.2.6 (Ronquist et al. 2012) with running 10,000,000 generations, sampled every 1000 generations, and the first 25% of the trees discarded as burn-in.
Flower and fruit traits were not measured because of two reasons: (1) the flowers of Ilex ficoidea and I. pernyi show few differences in shape, size, color and even indumentum (Yang 2020); (2) there were five I. sanqingshanensis’s individuals bearing fruits in 2020, however, only one tree bore red mature drupes and the other four hold immature fruits. Therefore, we measured eight leaf characters in this study, i.e., blade length, blade width, ratio of blade length/width, leaf area, number of spine or serration, spine or serration length, petiole length, and number of secondary veins. One healthy second-year branch per each sampled individual of the three target taxa were kept for making specimen and later assessment. The specimens were then digitized and standardized by the scanner (EPSON WF-C5790). ImageJ software was employed to examinate the target traits. Three well-preserved leaves per specimen were randomly selected for examination. Detailed measurements and voucher information are available in Table S2. The values of the eight characters were normalized by mean and standard deviation. Comparisons of eight leaf characters among I. ficoidea, I. pernyi, and I. sanqingshanensis were conducted. Significant differences between any two taxa were identified using the least significant difference (LSD) test. Box charts and Principal Component Analysis (PCA) were employed to visualize the differences among the three target species by Origin2018 (Moberly 2018) and PAST software ver. 4.11(Hammer et al. 2001), respectively. To avoid collinearity in PCA, Pearson correlation analyses were performed in the IBM SPSS Statistics 26 to weed out highly correlated characteristics (blade length and leaf area). Principal Component Analysis (PCA)
The aligned matrix of the ITS maker were 715 bp in length. Due to the highly divergences of Ilex species at the ITS region, we only displayed the sequence variations among I. sanqingshanensis, I. ficoidea, and I. pernyi herein. The putative parents were strict monomorphic intraspecies. There were nine nucleotide substitutions (at the site of 41, 71, 168, 189, 195, 444, 473, 509, and 518 bp) and four 1-bp insertion/deletions (47, 570, 613, and 701 bp) between them (Table 1). For I. sanqingshanensis, all 19 individuals exhibited exact additivity at these sites. Nevertheless, the clones S12_C3, S16_C3, S17_C8, and S22_C9 showed single exclusive nucleotide substitution at the site of 48 (G for S12a, A for all the other sequences), 590 (G for S16b, A for all the others), 651 (C for 17c, T for all the others), and 565 bp (C for S22b, T for all the others), respectively. The aligned sequences of psbA-trnH and petA-psbJ had a total of 900 bp in the three focal taxa. No sequence variations were detected within each taxon. The sequence of I. sanqingshanensis was identical to that of I. pernyi while different from I. ficoidea in three sites (Table 1).
Samples |
Variable sites |
|||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
ITS |
psbA-trnH |
petA-psbJ |
||||||||||||||
41 |
47 |
71 |
168 |
189 |
195 |
444 |
473 |
509 |
518 |
570 |
613 |
701 |
182 |
372 |
144 |
|
I. ficoidea (R1–R20) |
G |
G |
G |
T |
G |
A |
T |
C |
C |
G |
gap |
gap |
T |
C |
T |
T |
I. pernyi (M1–M19) |
A |
gap |
T |
A |
A |
G |
A |
G |
A |
A |
C |
G |
gap |
A |
G |
C |
I. sanqingshanensis (S1-C2, S2-C1, S3-C7, S6-C10, S7-C10, S8-C7, S9-C6, S11-C7, S12-C3, S13-C1, S14-C5, S15-C7, S16-C3, S16-C4, S17-C3, S18-C8, S19-C1, S20-C5, S21-C1, S22-C3) |
A |
gap |
T |
A |
A |
G |
A |
G |
A |
A |
C |
G |
gap |
A |
G |
C |
I. sanqingshanensis (S1-C3, S2-C6, S3-C6, S6-C6, S7-C2, S8-C5, S9-C3, S11-C3, S12-C4, S13-C8, S14-C10, S15-C1, S16-C2, S17-C1, S17-C8, S18-C5, S19-C7, S20-C3, S21-C9, S22-C9) |
G |
G |
G |
T |
G |
A |
T |
C |
C |
G |
gap |
gap |
T |
For the phylogenetic analyses based on ITS and cpDNA, the ML tree and the BI tree were largely identical in topology. As a result, only the ML trees were presented with the posterior probabilities (PP) from BI analysis indicated (Fig. X). On the ITS tree, all six members of Ilex sect. Ilex (I. cornuta, I. latifolia, I. shukunii Yi Yang & H.Peng, and the three focal taxa) formed a clade with high support values (0.98/74). Ilex cornuta, I. latifolia, and I. shukunii were strongly supported as independent species (Fig. 2). In contrast, the clones of I. sanqingshanensis were divided into two groups, one clustered with I. pernyi (1/98) and the other with I. ficoidea (1/82). On the cpDNA tree, I. sanqingshanensis and I. pernyi formed a clade with I. ficoidea left out (Fig. 2).
All the eight leaf characters showed that Ilex sanqingshanensis’s measures fell in between I. ficoidea and I. pernyi (Fig. 3). Based on the LSD test analyses (p < 0.05), I. sanqingshanensis significantly differs from I. ficoidea and I. pernyi in all tested traits (Fig. 3A–H). In principal component analysis, the three tested taxa were identified as three distinct groups by the scatter plots of the first two principal components (PCs), with PC1 accounted for 79.65% of total variances and PC2 for 10.24%, respectively (Fig. 3I).
For most hollies at a given region in subtropical Asia, there are ample opportunities for natural hybridization because of the spatial sympatry, flowering overlap, and pollinator sharing (Galle 1997; Chen et al. 2008; Tsang and Corlett 2005; Shi et al. 2016; Yang 2020). However, the confirmed natural hybrids of Ilex were all derived from the intra-sectional crosses (Galle 1997; Yang et al. 2022). For instance, I. × meserveae S.Y.Hu, I. × wandoensis C.F.Mill. & M.Kim, and I. × dabieshanensis K.Yao & M.P.Deng come from the natural crosses I. aquifolium L. × I. rugosa F.Schmidt (Hu 1970), I. cornuta × I. integra Thunb. (Son et al. 2009)d cornuta × I. latifolia (Shi et al. 2016), respectively, all the progenitors are from the type section. In the Sanqing Mountain, we found other five members of I. sect. Ilex: I. cornuta, I. latifolia, and I. shukunii share the area below an elevation of 800 m, I. pernyi resides the region over 1300 m, and I. ficoidea occurs in the whole mountain. Due to the strict co-occurrence with I. ficoidea and I. pernyi, we supposed a hybrid origin for I. sanqingshanensis.
In this study, the hypothesis has been clearly confirmed by phylogenetic analyses. On the ITS phylogenetic tree, the clones of I. sanqingshanensis split into two groups inserted in I. pernyi and I. ficoidea clades (rather than other hollies also distributing in the Sanqing Mountain), respectively. It clearly uncovers that I. sanqingshanensis is the production of hybridization between I. ficoidea and I. pernyi. On the cpDNA tree, all individuals of I. sanqingshanensis cluster with I. pernyi rather than I. ficoidea. It further exposes the maternal role of I. pernyi and the reverse of I. ficoidea.
For hybrid speciation, hybridization is only the first stage. If the hybrid population is towards enhancing fitness, occupying novel niche, and developing reproductive isolation, speciation and diversification may arise (Hegarty and Hiscock 2005; Baack and Rieseberg 2007; Lamichhaney et al. 2018; Wang et al. 2021). In other words, hybrid speciation is not simply the production of F1 hybrids or backcrosses between the parental species, as it requires to erect a third, distinct species (Counterman 2016). In taxonomy, the crossing class dominated by F1s or backcrosses should be treated as a hybrid rather than hybrid species (Wagner 1969).
To Ilex sanqingshanensis, there is a lot of evidence against its status of hybrid species. In the hybrid zone, both I. ficoidea and I. pernyi are common and have a high population density, while I. sanqingshanensis is extremely rare and sparse (only 22 individuals in total were discovered during the field investigations in the Sanqing Mountain). The thin population reflects the low fitness or fertility of I. sanqingshanensis in comparison with the parental species. Moreover, sequence analyses in this study showed that all sampled 19 hybrid individuals were of additivity at all fixed sites diverging between I. pernyi and I. ficoidea, implying that these individuals could be F1s. Additionally, morphological analyses also indicated the domination of F1s in I. sanqingshanensis. In a hybrid population, backcrossing can obscure and even melt the morphological divergence between distinct species because backcrossed offspring more resemble the parents in comparation with F1s (Wagner 1969; Soltis and Soltis 2009). In this study, I. sanqingshanensis was identified as a distinct group without any overlaps with I. pernyi or I. ficoidea by PCA, suggesting the absence of backcrossed individuals in the sampled population. Thus, this “endangered species” should actually be accepted as a nothospecies, I. × sanqingshanensis.
Expectedly, morphological analyses have also disclosed the hybrid origin of I. × sanqingshanensis, as I. × sanqingshanensis shows an intermediacy between I. ficoidea and I. pernyi in all tested morphological characters (Fig. 3). However, I. × sanqingshanensis distinctly differs from the two parental species in all of the eight leaf traits, making it highly peculiar and distinguishable in morphology. Consequently, the novel characters derived from the fusion of hugely divergent parental species had misled botanists to treat the hybrid as a new species and overlook the clues of hybridization (Shi et al. 2015). In the genus Ilex, such confusion caused by hybridization was common, e.g., I. × attenuata Ashe (Galle 1997), I. × dabieshanensis (Shi et al. 2016; Yang and Peng 2019), and I. chengkouensis C.J.Tseng, I. miguensis S.Y.Hu, and I. zhejiangensis C.J.Tseng ex S.K.Chen & Y.X.Feng (Yang 2020). As Ilex shows weak reproductive isolation intra section and has a large number of hybrids (Galle 1997; Yang 2022), only presenting morphological distinctions is not enough to support the establishment of Ilex species. Given the effectivity of molecular data in identifying hybrids of Ilex (Son et al. 2009; Shi et al. 2016; and this study), we herein underline the indispensability of molecular evidence in erecting Ilex species.
Acknowledgements
This study is supported by the Natural Science Foundation of Jiangxi, China (Grant No. 20212BAB215008), the Science and Technology Project of Jiangxi Provincial Department of Education (Grant No. GJJ200410), the National Natural Science Foundation of China (Grant No. 32260048), and the Modern Agriculture Project of Jiangsu Province (Grant No. BE2021307).
No conflict of interest exits in the submission of this manuscript, and the manuscript has been approved by all co-authors. The work is original, and has neither been published previously, nor is under consideration for publication elsewhere, in whole or in part. There are not financial interests that are directly related to the work submitted for publication.