DOI: https://doi.org/10.21203/rs.3.rs-1529055/v1
Natural hybridization plays a significant role in evolution of the genus of Rosa, e.g., Kushui Rose (R. Sertate x R. Rugosa) is always considered as a natural hybrid of R. sertata and the cultivated rose R. rugosa. In order to test the hybrid origin of Kushui Rose, we conducted molecular sequences analysis and morphological traits examination in its related taxa in the genus. Although Principal component analysis (PCA) of Morphological characteristics revealed Kushui Rose was intermediate between R. rugosa and R. willmottiae, and also and R. sertata, but the unambiguous sequence additivities of nuclear DNAs (ITS and GAPDH) as well as the incongruence between nuclear and chloroplast gene trees revealed the hybrid origin of Kushui Rose from R. rugosa and R. willmottiae rather than R. sertata. Further, the cluster relationships in concatenated chloroplast (cpDNA) trees (psbA-trnH, trnL-F, and trnS-G) implied that R. willmottiae and R. rugosa respectively acted as its maternal parent and paternal parent during its homoploidization speciation. Here our study could provide a more reasonable orientation for further study in conservation biology of the hybrid species as well as in its agronomy traits with heterosis effect.
The genus Rosa is widely distributed in the cold temperate and subtropical regions of Asia, Europe, North Africa and North America, with more than 200 species, of which 95 species have been found in China (Flora of China Editorial Committee of Chinese Academy of Sciences 1985). Kushui Rose, a deciduous shrub of the genus, is narrowly distributed in Yongdeng County, Gansu Province, China, and has been cultivated for 200 years (Ma et al. 1990; Niu et al. 2015; Zhang et al. 2018). Kushui Rose is one of the popular varieties of oil-bearing rose in China and widely cultured in the local agriculture, because it has the characteristics of unique fragrance and high oil content, which can reach 0.04% when extracted (Wang et al. 2012; Wang et al. 2012; Guo et al. 2017; Wang 2018). In addition, it has anti-oxidation (Liu et al. 2018) and vitro anti-complementary (Cheng et al. 2019) effects, and its flower buds have served as folk medicine in China for a long time (Jin 2000). In conclusion, Kushui Rose is a valuable plant resource possessing ornamental, medicinal, edible and economic values.
Kushui Rose was usually regarded as a natural hybrid of R. sertata and R. rugosa which was originally identified by Dejun Yu and Cuizhi Ku, thus scientifically named R. sertata × R. rugosa Yu. et. Ku (Ma et al. 1985), and also named Rosa rugosa ‘KuShui’ (Niu et al. 2017). Kushui Rose usually bears no fruit, and the garden individuals are mostly cloned by asexual reproduction which was responsible for its relative stability of genetic traits and the homogeny of germplasm resource (Niu et al. 2017). Natural hybridization is assumed to play an important role in plant speciation (Rieseberg et al. 2000; Hegarty et al. 2005). It was well accepted that there are no strong internal barriers to the gene flow between various species within the genus Rosa (Atienza et al. 2005) and natural hybridization plays a very important role in the evolution of the genus Rosa (Bai et al. 2009). Therefore, it is critical to clarify whether Kushui Rose originated from a hybridization event before it would be broadly utilized and effectively conserved in practice (Zhang et al. 2020). However, except for the solo record about the hybrid origin of Kushui Rose, there was no relevant research data to support or deny it so far.
In fact, it is always a challenge to identify the species of homoploid hybridization (Rieseberg et al. 2000). In general, morphological associations or intermediate characteristics between suspected hybrid and their parents were detected to identify natural hybridization (Rieseberg et al. 1993; Horie et al. 2012). Whereas, morphological characteristics are affected by both genes and environment, lack common standards, and morphological intermediacy can arise through convergent evolution or the retention of ancestral character states (Rieseberg et al. 2000; Gross 2012; Wang 2017). In contrast with morphological features, molecular data can directly reflect the internal genetic relationship among species (Yang et al. 2020). Among them, sequences alignment and phylogenetic analysis are the most common method to identify natural hybridization, especially interspecific hybridization (Wang 2017). The biparental inheritance property of nuclear genome facilitates inference of hybrid origin by using it in phylogenetic analysis. Meanwhile, the mostly maternal inheritance of chloroplast genome with small size and less recombination provides a powerful means for identifying the female parent of hybrid species (Joseph et al. 1988; Soltis et al. 1992; Wang 2017).
In this study, we used morphological and molecular data of Kushui Rose and its related taxa in Rosa, including two nuclear regions and three chloroplast regions, to trace its hybrid origin and the putative parents. Remarkably, we obtained an unusual pathway of hybrid origin of Kushui Rose in evolution of the genus, and provide theoretical basis for the better development, utilization and conservation of Kushui Rose.
The materials included Kushui Rose and its related plants of the Rosa, totaling 34 species/variety. Part of material specimens were deposited at the Herbarium of Northwest Normal University (NWNU), and all sequences of R. beggeriana, R. fedtschenkoana, R. saturate and R. praelucens, as well as partial sequences for the remaining species, were downloaded from NCBI. New sequences were submitted to GenBank and detailed material information was listed in Table 1.
Plant materials (species/variety) | Location | |
---|---|---|
Sect. Cinnamomeae DC. | ||
R.sertata × R.rugosa Yu. et. Ku (KSRR, KuShui Rose) | Yongdeng, Gansu | |
Rosa acicularis Lindl. | Pingliang, Gansu | |
Rosa bella Rehd. | Gannan, Gansu | |
Rosa caudata Baker | Tianzhu, Gansu | |
Rosa moyesii Hemsl. | Gannan, Gansu | |
Rosa davurica Pall. | Yongdeng, Gansu | |
Rosa davidii Crep. | Tianshui, Gansu | |
Rosa rugosa Thunb. ‘YanXia’ | Yongdeng, Gansu | |
Rosa rugosa Thunb. ‘ZiZhi’ | NWNU, Lanzhou, Gansu | |
Rosa rugosa Thunb. ‘FengHua’ | NWNU, Lanzhou, Gansu | |
Rosa laxa Retz. | Longnan, Gansu | |
Rosa willmottiae Hemsl. (TZ) | Tianzhu, Gansu | |
Rosa willmottiae Hemsl. (JSX) | Yongdeng, Gansu | |
Rosa giraldii Crep. | Gannan, Gansu | |
Rosa sertata Rolfe. | Qilian Mountain, Gansu | |
Rosa sweginzowii Koehne. | Gannan, Gansu | |
Sect. Pimpinellifoliae DC. | ||
Rosa omeiensis Rolfe | Yongdeng, Gansu | |
Rosa xanthina Lindl. | Yongdeng, Gansu | |
Rosa xanthina Lindl. f. normalis | NWNU, Lanzhou, Gansu | |
Sect. Synstylae DC. | ||
Rosa henryi Bouleng. | Longnan, Gansu | |
Rosa rubus Levl. | Longnan, Gansu | |
Rosa helenae Rehd. | Longnan, Gansu | |
Rosa multiflora Thunb. var. carneaThory | Longnan, Gansu | |
Rosa multiflora Thunb. var. cathayensisThory | Longnan, Gansu | |
Sect. Laevigatae Crep. | ||
Rosa laevigataMichx. | Jiujiang, Gansu | |
Sect. Chinenses DC. ex Ser. | ||
Rosa chinensis Jacq. | Longnan, Gansu | |
Sect. Rosa | ||
Rosa gallica L. | Yongdeng, Gansu | |
Rosa damascene Mill. | Yongdeng, Gansu | |
Sect. Mierophyllae Crep. | ||
Rosa roxburghii Tratt. | Guangyuan, Gansu | |
Sect. Bracteatae Thory | ||
Rosa bracteate Wendl. | Garze, Sichuan | |
Sect. Banksianae Lindl. | ||
Rosa banksiae Ait. | Longnan, Gansu |
Total DNA was extracted from silica-gel dried leaves using the modification of the Doyle and Doyle (1987) cetyltrimethylammonium bromide (CTAB) protocol. The extracted product was tested for DNA concentration by a nucleic acid detector.
Two nuclear regions (ITS and GAPDH) and three chloroplast fragments (psbA-trnH, trnL-F, trnS-G) were amplified by PCR and gene cloning. Primers were listed in Table 2. PCR amplification were performed in 25 µL total amplification system with 12.5 µL Ex Taq (Takara Bio, Shiga, Japan), 1 µL of each primer, 8.5 µL of ddH2O, 1 µL of genomic DNA. The annealing temperature was 50℃ for ITS, 54℃ for GAPDH, 51℃ for the psbA-trnH, 54℃ for the trnL-F and the trnS-G. Part of the amplification products were used for sequencing, and the other part was detected by 1.5% agarose gel electrophoresis, and positive clones were screened through purification, ligation, transformation, and colony PCR. Each material was screened for 5 monoclonal samples and sent to GENEWIZ (Tianjin, China) to perform bidirectional sequencing.
Gene region | Primer | Primer sequence (5’-3’) |
---|---|---|
ITS | ITS-F | TATGCTTAAAYTCAGCGGGT |
ITS-R | AACAAGGTTTCCGTAGGTGA | |
GAPDH | GPDX7-F | GATAGATTTGGAATTGTTGAGG |
GPDX11-R | GACATTGAATGAGATAAACC | |
psbA-trnH | PsbA-F | GTTATGCATGAACGTAATGCTC |
trnH-R | CGCGCATGGTGGATTCACAATCC | |
trnL-F | trnLc | CGAAATCGGTAGACGCTACG |
trnFf | ATTTGAACTGGTGACACGAG | |
trnS-G | trnS | GCCGCTTTAGTCCACTCAGC |
trnG | GAACGAATCACACTTTTACCAC |
The sequenced data of DNA fragments were read and edited using Chromas software (Version 2.6.5, http://technelysium.com.au/wp/chromas/). Online blast tools (Boratyn et al. 2013) were used to align sequences. Dnasp5.0 (Librado et al. 2009) was used to calculate variation base sites of all samples and analyze the haplotype sequences.
For nuclear and cpDNA dataset, MEGA7 (Kumar et al. 2016) was used to model selection and ML analyses. ML bootstrap support values were summarized from 5000 bootstrap trees.
In order to compare morphological data, 17 characters of 29 species of genus Rosa were measured. A total of 17 characters comprising 14 qualitative characters and 3 quantitative characters. In which qualitative characters were scored, and 10 individuals from each species were selected to count the mean value of each quantitative character. Principal component analysis (PCA) can summarize and visualize multiple morphological data of individuals to facilitate subsequent identification of species relationships (Carney et al. 2000). The morphological data were selected for standardization and principal component analysis. All statistical calculations used the R (Version 4.10) for Linux.
The ITS matrix contained 31 sequences and 603 base pairs (bp) DNA sequences, of which 89 mutation sites (14.8%) and 31 parsimony informative sites (5.1%). The ML tree constructed based on ITS sequences showed that two sequences of Kushui Rose (KSRR-H1 and KSRR-H2) were divided into two branches respectively (Fig. 1). The KSRR-H1 was clustered with part of taxa from R. sects. Cinnamomeae and formed an inner clade with R. rugosa. Another sequence with R. davidii, R. willmottiae (JSX), R. giraldiiand and R. willmottiae (TZ) from R. sects. Cinnamomeae were clustered in a clade. According to the results of ITS sequences alignment between Kushui Rose and its possible parent species, 5 sites of Kushui Rose showed perfect additivity (Table 3, ITS). Each Kushui Rose individual had two nucleotides, one of each of the parents.
The GAPDH dataset included 29 sequences and 697 aligned nucleotides that contained 111 mutation sites (15.9%) and 41 parsimony informative sites (6.3%). The phylogenetic relationships of GAPDH (Fig. 2) were similar to ITS tree. The KSRR-H1 was resolved as sister to R. rugosa ‘YanXia’ with 75% bootstrap value in clade C1. The KSRR-H2, R. willmottiaeand and R. giraldiiare clustered in a small clade. For GAPDH sequence alignment, 7 sites of Kushui Rose also showed perfect additivity (Table 3, GAPDH).
Three chloroplast regions (psbA-trnH, trnL-F, and trnS-G) were used in this study. The concatenated matrix of 23 taxa and 3255 bp, with 187 (5.7%) mutation sites, and 91 (2.8%) parsimony informative sites. The phylogenetic analyses results showed that Kushui Rose and R. willmottiae (TZ) formed a small branch E1, and the bootstrap value was 86 (Fig. 3). Meanwhile, the sequence similarity between Kushui Rose and R. willmottiae (TZ) was high without significant differences.
Species |
ITS |
GAPDH |
|||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
26 |
60 |
85 |
111 |
122 |
177 |
413 |
59 |
186 |
228 |
288 |
291 |
630 |
659 |
||
Kushui Rose |
T/C |
C/T |
C/T |
C/T |
G/A |
T/- |
C/- |
C/T |
C/T |
G/T |
T/C |
G/C |
A/G |
C/G |
|
Rosa rugosa ‘YanXia’ |
T |
C |
C |
C |
G |
T |
C |
C |
C |
G |
T |
G |
A |
C |
|
Rosa rugosa ‘ZiZhi’ |
T |
C |
C |
C |
G |
T |
C |
C |
C |
G |
T |
G |
A |
G |
|
Rosa willmottiae (TZ) |
T |
T |
T |
T |
A |
- |
C |
T |
T |
T |
C |
C |
G |
G |
|
Rosa willmottiae (JSX) |
T |
T |
T |
T |
A |
- |
C |
T |
T |
T |
C |
C |
G |
G |
|
Rosa giraldii |
T |
T |
T |
T |
R |
T |
C |
T |
Y |
T |
C |
C |
R |
G |
|
Rosa davidii |
T |
T |
T |
T |
A |
- |
C |
T |
Y |
T |
C |
C |
R |
G |
|
Rosa sertata |
T |
T |
T |
T |
A |
T |
C |
T |
T |
G |
C |
G |
A |
G |
In morphological data, we found that Kushui Rose had many morphological characters between R. willmottiae and R. rugosa ‘YanXia’. Similarly, there were a lot of characters in Kushui Rose that were the same as those of the suspected parents. PCA of the phenotypes of 29 Rosa species showed that the first two main principal components accounted for 21.3% and 12.6% respectively. Figure 5 showed that species of the same genus with the same color tended to cluster together. Kushui Rose was located between R. willmottiae and R. rugosa.
Table 4
Morphological characteristics of KSRR and its related species
Characteristics |
Kushui Rose |
R. rugosa‘YanXia’ |
R. willmottiae |
R. sertata |
|
---|---|---|---|---|---|
Twig color |
Taupe |
Taupe |
Taupe |
Rufous |
|
Prickles |
Straight/Curved |
Straight/Curved |
Straight, Thin/Curved |
Straight/ Stingless |
|
The number of leaflets |
5–7 |
5–9 |
7–9 |
7–11 |
|
Leaflet apex |
Rounded |
Acute |
Rounded |
Rounded |
|
Leaflet margin |
Serrate |
Serrate |
Serrate |
Serrate |
|
Leaflet shape |
Elliptic |
Elliptic |
Obovate |
Elliptic |
|
Rachis thorns |
No |
No |
No |
Yes |
|
Inflorescence type |
Solitary |
Solitary |
Solitary |
Solitary |
|
Petal color |
Purple-red |
Purple-red |
Pink |
Pink |
|
Petal apex |
Emarginate |
Emarginate |
Emarginate |
Emarginate |
|
Petal count |
23 |
38 |
5 |
5 |
|
Whether has peduncles |
No |
No |
No |
No |
|
Coverage of flowers |
0.69 |
0.20 |
0.20 |
0.60 |
|
Calyx tube shape |
Cuplike |
Cuplike |
Cuplike |
Urceolate |
|
Sepals |
Erect |
Erect |
Erect |
Erect |
|
Whether the sepals are exfoliated |
No |
No |
Yes |
No |
|
Stipule shape |
Lanceolate |
Oval |
Oval, Lanceolate |
Auricular |
Homoploidyhybridization seems to play an important role to form new species by reassembling genomes without changes in ploidy number in the genus of Rosa, such as R. pseudobanksiaeandpart of edible roses (Zhang et al. 2020; Cui et al. 2022). According toChinese distinguished plant taxonomists Junde Yu and Cuizhi Ku’s observation, Kushui Rose was proposed as the product of natural hybridization betweenR. rugosa and R. sertatais, and it was scientifically named asR. Sertate × R. RugosaYu. et. Ku(Ma et al. 1985).Unfortunately,the two scientistshadn’t published any available data about the topic so left it still an open question.If Yu & Ku’s proposal is true, Kushui might originate from a homoploidization event, becauseKushui Rose (Zhao et al. 2007), R. rugosaand R. sertatais (Fang 2020) were all identified as diploid in karyotype analysis, with the number of chromosomes 2n=14.
In fact, at Yu & Ku’s time, only depending on morphological traits to identify a hybrid species without ploidy is a very difficult task (Rieseberg 1997;Rieseberg et al. 1990). Generally regarding, there are some intermediate morphological characters between hybrid species and their parents. Kushui Rose exhibited a lot of morphological characters that are of the intermediate to R. rugosa and R. sertata, such as the number of leaflets and petal count, Fig. 5. However, the characters in Kushui Rose could be also observed intermediate to R. rugosa and R. willmottiae, such as the number of leaflets, petal count, coverage of flowers and stipule shape. Its twig color, petal apex and calyx tube shape were consistent with R. willmottiae and R. rugosa ‘YanXia’, Fig. 5. So only based on the morphological characters there is no clear answer for the question whether Kushui Rose was a hybrid species between R. rugosa and R. sertatais, or between R. rugosa and R. willmottiae. On the other hand, Male sterility due to meiosis failure to form vigorous microspores was recently observed in Kushui Rose, quite consistent with its assumed identity of hybrid species (Wang et al. in press).
Species with hybrid origin usually exhibit additivity of parental nuclear genomes, therefore the additivity of nuclear gene sequences alignment and the incongruency in further phylogenetic analysis are the most important evidences for identifying homoploidy hybridization (Sun 2003; Zhang 2005). The present nuclear sequences alignment results indicate that ITS and GAPDH sequences of Kushui Rose showed perfect additivity at 5 and 7 sites between R. rugosa‘YanXia’ and R. willmottiae (TZ, while Kushui Rose exhibited additivity at only 4 sites between R. rugosa‘YanXia’ and R. sertata. These results indicate strongly that Kushui Rose as arisen through hybridization of R. rugosa‘YanXia’ and R. willmottiae (TZ). In nuclear trees, two haplotypes from Kushui Rose were closely related with R. rugosa ‘YanXia’ and R. willmottiae (TZ) respectively (Fig. 2 and Fig. 3), while R.sertata did not formed a monophyletic clade with Kushui Rose, suggesting that Kushui Rose origins from two different parents,R. rugosa ‘YanXia’ and R. willmottiae (TZ). In addition, only R. willmottiae (TZ) with Kushui Rose clustered an inner clade in phylogenetic analyses ofcpDNAsequences, and neither R. rugosa ‘YanXia’ nor R.sertata clustered closely to Kushui Rose, in which R.sertata formed a single branch. The significant differences between nuclear and chloroplast trees are considered to be strong evidence for hybridization events (Costea et al. 2010; Gruenstaeudl et al. 2012). Although the inconsistency of nuclear gene tree and chloroplast gene trees might also be derived from the other historical events such as lineage sorting (Rieseberg 1997), in this study, multiple unlinked loci producing phylogenetic trees with the similar topological feature could give a relatively solid conclusion of homoploid hybridization to some extent.
In sum, the morphological analysis in this study together with the previous study of male sterility in cytological level support Yu & Ku’s proposal of Kushui Rose as a hybrid species. On the other hand, our further molecular biological analysis, including additivity among nuclear DNA sequences alignment as well as their incongruent positions in the same phylogeny, and the incongruency of nuclear gene tree and chloroplast gene trees, evidenced that Kushui Rose was the product of natural homoploid hybridization between R. rugosa and R. willmottiae, rather than between R. rugosaand R. sertata. Thus, its scientific name should be R. willmottiae× R. rugosa.
Chloroplast genome is usually maternally inherited and is not affected by hybridization factors in angiosperm, which is often used to identify the maternal origin of the hybrid.In phylogenetic analyses of the cpDNA sequence, R. willmottiae (TZ) was clustered with Kushui Rose, therefore we further believed that R. willmottiae (TZ) and R. rugosa ‘YanXia’ should be the female parent and male parent, respectively.
In terms of distribution area, Kushui Rose distributes in Kushui, Yongdeng, Gansu. R. willmottiae distributes in Gansu, Sichuan, Shaanxi and Qinghai, which grows in shrub, hillside or roadside at an altitude of 1300-3150 meters, and flowering period is May-June.R. rugosa distributes throughout China, and flowering period is May-June (Flora of China Editorial Committee of Chinese Academy of Sciences 1985). The habitat and distribution area of R. willmottiae overlap with R. rugosa, and the flowering period is the same. In addition,we also found R. Willmottiae and R. Rugosa in the area near Kushui town. Accordingly, we speculated that R. willmottiae and R. rugosa have frequent gene exchanges in Kushui town or upstream areas, and eventually form a natural hybrid - Kushui Rose.
Wang (in press) proved that Kushui Rose and R. rugosa have cross-compatibility through hybridization experiments, but there has been no report on cross hybridization between R. rugosa and R. willmottiae. Kushui Rose has cross-compatibility with R. rugosa, which means that it has not formed reproductive isolation between paternal parent. Other studies have found that Kushui Rose is unable to bear fruit by self-crossing, let alone obtain fertile seeds (Ma et al. 1985; Wang et al. inpress). Where the two species meet, these hybrids may form a stable hybridization zone that serves as an effective barrier with the parent species. At the same time, backcrossing with parent or hybrid dysgenesis may prevent the process of homoploid hybrid species formation. In general, only after a group overcomes seed sterility and produces reproduction and niche isolation with its parents can the offspring of these hybrids form a new self-evolving lineage (Zhang et al. 2020). Combined with the analysis of this study, it is shown that Kushui Rose has not formed an obvious genetic pedigree.
The reproductive isolation between Kushui Rose and its parent species has not been established, which means that Kushui Rose may have backcrossed with its sympatric parents. Although introgression may play a creative role in the evolution of species, it may also have many negative consequences, or even lead to the extinction of species (Cabria et al. 2011). The tracing of hybrid origin of Kushui Rose in this study is the basis for conservation not only because of its many values, but also can provide the direct raw material for hybrid breeding, to provide the possibility of further development of new varieties of hybrid (Zhang et al. 2020). Furthermore, the existing evidence cannot prove that Kushui Rose is the F1 generation hybrid of R. rugosa and R. willmottiae, and whether its formation has undergone introgression after hybridization, such as backcrosses, requires further studies.
Funding
This work was supported by National Natural Science Foundation of China (31060033) and State Key Laboratory of Plant Cell and Chromosome Engineering (PCCE-KF-2019-06).
Competing Interests
The authors have no relevant financial or non-financial interests to disclose.
Author Contributions
All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Ying Wang, Fan Zhao, Ling-Ling Da and Hui Zhang. The first draft of the manuscript was written by Fan Zhao and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Data Availability
The datasets generated during the current study are available in the GenBank repository.