Exploring Genetic Diversity and Phylogenetic Relationships in Camellia reticulata Cultivars Using Novel Low-Copy Nuclear Gene Markers

Abstract

This study delves into the exploration of genetic diversity and phylogenetic relationships within Camellia reticulata cultivars, providing a vital reference for horticultural research on this hetero-hexaploid species. Utilizing available transcriptome data from C. reticulata, novel low-copy nuclear gene sequences were successfully identified. With tailored primer design, these genes were amplified and sequenced from 100 C. reticulata cultivars originating from Kunming, Chuxiong, Dali, and Tengchong in China. Five distinct low-copy nuclear gene sequences were found to collectively span 3481 bp, showcasing 71 polymorphic mutation sites (P_i = 0.0077) and 91 haplotypes (H_d = 0.9974). The genetic diversity among cultivars from the four provenances ranked Tengchong > Dali > Kunming > Chuxiong. AMOVA analysis revealed that 96.50% of the genetic variation exists within the provenances. Low genetic distance and differentiation (F_st = 0.0199) were observed among cultivars from the four provenances. Ninety-nine cultivars were clustered into four clades, corresponding to the three ancestors of C. reticulata (diploid C. reticulata, C. pitardii, and C. saluenensis). The novel low-copy nuclear gene sequence markers developed in this study provide an effective tool for analyzing genetic diversity, phylogenetic relationships, and origination of C. reticulata cultivars.

1. Introduction

Camellia reticulata Lindl. (known as Yunnan camellia), an exclusive Theaceae species native to China, is a national key second-level protected plant and proudly serves as the official city flower of Kunming [1]. As the tallest ornamental Camellia species worldwide, C. reticulata can ascend to an impressive height of 28.1 m. Renowned for its sizable flowers reaching up to 22 cm in diameter, vibrant colors, and a diverse array of cultivars (with 823 registered on the World Camellia Plant Variety Registry Database) [2], these camellias are extensively used as potted and landscaping plants throughout Yunnan, China. With a rich history spanning over 1300 years, they have transcended national borders and become cherished ornamentals in numerous countries [3,4].

Despite their widespread cultivation, it is noteworthy that C. reticulata has predominantly been grown in controlled environments, such as greenhouses, deviating from their native outdoor habitats. This is attributed to their origin in the mild regions of southwest China (Figure 1), specifically at altitudes ranging from 1000 to 3200 m [5]. Although more than 100 cultivars have been successfully cultivated, they are susceptible to freezing damage below –5 °C, and their growth is impeded beyond 32 °C [1], restricting the outdoor planting range for C. reticulata. To overcome these limitations and expand the outdoor cultivation footprint, there is a compelling need to develop new cultivars endowed with enhanced heat and cold resistance.

Understanding the genetic diversity of C. reticulata is important for the development of new cultivars. Recent investigations into the genetic diversity of C. reticulata cultivars have encompassed both morphological markers [2] and molecular studies. While AFLP (Amplified Fragment Length Polymorphism) markers have been employed to explore genetic diversity and structure [6], challenges arise from the hexaploid nature of C. reticulata, which also includes octaploid and decaploid cultivars [7,8]. Traditional molecular markers, such as AFLP, ITS (Internal Transcribed Spacer), and SSR (Simple Sequence Repeat), may present complications, such as double bands during amplification and multi-peaks during sequencing in polyploid groups [9,10,11]. It has been reported that the phylogenetic relationships of C. reticulata cultivars, based on ITS copies from clones, are confounded by the presence of numerous ITS pseudogenes [12]. Prior attempts utilizing ITS sequences for genetic diversity and phylogenetic relationships faced challenges due to incomplete consistency evolution and the necessity for high-quality DNA and specific PCR conditions [4,12,13,14].

Recognizing the constraints of traditional markers, this study advocates for the utilization of low-copy nuclear genes. Characterized by ease of acquisition and parental inheritance traits, low-copy nuclear genes offer an optimal choice for studying hybrid and polyploidy populations, particularly suited for phylogenetic analysis [9,15,16,17]. Notable examples include the use of the single-copy nuclear gene waxy for molecular phylogeny analysis of Camellia [13], and the utilization of the low-copy nuclear gene RPB2 (RNA polymerase II) in the authentication of C. chekiangolosa and its closely related species [18]. Moreover, the phylogeography of C. taliensis [19] and the population genetic structure and phylogeography of C. flavida [20] have been successfully explored using the low-copy nuclear gene PAL (Phenylalanine Ammonia Lyase).

In light of the existing abundance of C. reticulata cultivars and the scarcity of available low-copy nuclear gene markers, this study aims to meticulously select low-copy nuclear genes utilizing available transcriptome data. The subsequent screening process identified stable primers with high-quality amplification, facilitating a comprehensive exploration of the genetic diversity and phylogenetic relationships among C. reticulata cultivars. The insights gained from this endeavor are anticipated to serve as a valuable reference for the collection, preservation, identification, and innovative breeding of C. reticulata cultivars.

2. Materials and Methods

To assess the genetic diversity within the C. reticulata cultivars, a total of 100 cultivars were collected from Kunming Botanical Garden, Chuxiong Elu Park, and Tengchong Laifengshan Camellia Garden. This diverse collection encompassed 34 cultivars from Kunming, 19 from Chuxiong, 12 from Dali, and 35 from Tengchong (Table 1, Figure 1), as documented by the Kunming Association for Science and Technology [1] and the World Camellia Plant Variety Registry Database [2]. To enrich the analysis, we also included a set of closely related diploid species, including C. reticulata, C. saluenensis, C. yunnanensis, C. pitardii, C. tuberculate, and C. japonica obtained from Kunming Botanical Garden. Fresh and healthy adult leaves were carefully collected from all species, thoroughly dried with discolored silicone, sealed in plastic zip lock bags, and stored at −20 °C for DNA extraction.

Table 1. Sample information of Camellia reticulata cultivars.

Provenance	Cultivars
Kunming (34 cultivars)	‘Shizitou’ (8x), ‘Juban’(Figure 2b, 8x), ‘Songzilin’ (8x), ‘Liuye Yinhong’, ‘Damanao’ (6x), ‘Maye Taohong’ (8x), ‘Yujie’, ‘Yipinhong’, ‘Xiaoguiye’ (6x), ‘Zuijiaohong’, ‘Xiaotaohong’, ‘Houye Diechi’, ‘Baoyuhong’, ‘Caiyun’, ‘Taohongpao’ (6x), ‘Dianchi Xiuqiu’, ‘Mudancha’ (8x), ‘Dianchi Mingzhu’ (8x), ‘Zipao’ (8x), ‘Zaomudan’ (8x), ‘Saitaohong’ (8x), ‘Guiye Yanghong’ (6x), ‘Dayinhong’ (8x), ‘Maye Yinhong’, ‘Dataohong’, ‘Baiyi Zaotaohong’ (8x), ‘Lianrui’, ‘Fenhudie’, ‘Dayulan’ (6x), ‘Yanhe’, ‘Jinrui Furong’, ‘Yulancha’, ‘Hongwancha’, ‘Jingancha’ (8x)
Chuxiong (19 cultivars)	‘Dandinghe’ (8x), ‘Zibao’ (8x), ‘Miyilu’ (8x), ‘Chuxiongcha’, ‘Baize’ (8x), ‘Fozuolian’ (6x), ‘Zehe’ (6x), ‘Ziyu’, ‘Lifang’, ‘Weixihong’, ‘Zifen’, ‘Luchengchun’ (6x), ‘Ziyan’, ‘Liuye Meihong’, ‘Yundie’, ‘Jinrui Dahong’, ‘Jianye Diechi’, ‘Zilian’, ‘Weichu’
Dali (12 cultivars)	‘Daguiye’ (6x), ‘Saijuban’, ‘Yanhong Songzike’, ‘Meihong Guiye’ (6x), ‘Tongzimian’ (8x), ‘Baozhucha’ (6x), ‘Dalicha’, ‘Duxin Dalicha’, ‘Hentiangao’ (10x), ‘Pumencha’ (6x), ‘Songzike’ (6x), ‘Xiguiye’
Tengchong (35 cultivars)	‘Fentongcao’ (6x), ‘Mudankui’ (Figure 2a), ‘Fengshancha’, ‘Yunzhen’, ‘Wujiao Xiuqiu’ (6x), ‘Hehua Xianzi’ (8x), ‘Tuanye Diechi’ (6x), ‘Shuimeiren’ (6x), ‘Biyu’, ‘Yunfeng’ (8x), ‘Honghua Youcha’ (8x), ‘Fenzhen Mudan’, ‘Huahun’, ‘Jinrui Diechi’, ‘Xiaojiaojiao’, ‘Fenzhaoyun’, ‘Yushizi’, ‘Xianyecha’, ‘Yumeiren’, ‘Hongmei’, ‘Xiyingchun’, ‘Taohong Mudan’, ‘Yunhuacha’, ‘Fentianjiao’, ‘Saierqiao’, ‘Taihe Mudan’, ‘Manao Dahongcha’, ‘Jiaohe’, ‘Naochun Xiuqiu’, ‘Jiaoxiaohong’, ‘Fenhong Xiuqiu’, ‘Yanghong Songzike’, ‘Yulin’, ‘Nansongzi’, ‘Heidahong’

Note: the cell ploidy of cultivars was labeled according to the literature published by Xu et al. [7].

Table 1. Sample information of Camellia reticulata cultivars.

Provenance	Cultivars
Kunming (34 cultivars)	‘Shizitou’ (8x), ‘Juban’(Figure 2b, 8x), ‘Songzilin’ (8x), ‘Liuye Yinhong’, ‘Damanao’ (6x), ‘Maye Taohong’ (8x), ‘Yujie’, ‘Yipinhong’, ‘Xiaoguiye’ (6x), ‘Zuijiaohong’, ‘Xiaotaohong’, ‘Houye Diechi’, ‘Baoyuhong’, ‘Caiyun’, ‘Taohongpao’ (6x), ‘Dianchi Xiuqiu’, ‘Mudancha’ (8x), ‘Dianchi Mingzhu’ (8x), ‘Zipao’ (8x), ‘Zaomudan’ (8x), ‘Saitaohong’ (8x), ‘Guiye Yanghong’ (6x), ‘Dayinhong’ (8x), ‘Maye Yinhong’, ‘Dataohong’, ‘Baiyi Zaotaohong’ (8x), ‘Lianrui’, ‘Fenhudie’, ‘Dayulan’ (6x), ‘Yanhe’, ‘Jinrui Furong’, ‘Yulancha’, ‘Hongwancha’, ‘Jingancha’ (8x)
Chuxiong (19 cultivars)	‘Dandinghe’ (8x), ‘Zibao’ (8x), ‘Miyilu’ (8x), ‘Chuxiongcha’, ‘Baize’ (8x), ‘Fozuolian’ (6x), ‘Zehe’ (6x), ‘Ziyu’, ‘Lifang’, ‘Weixihong’, ‘Zifen’, ‘Luchengchun’ (6x), ‘Ziyan’, ‘Liuye Meihong’, ‘Yundie’, ‘Jinrui Dahong’, ‘Jianye Diechi’, ‘Zilian’, ‘Weichu’
Dali (12 cultivars)	‘Daguiye’ (6x), ‘Saijuban’, ‘Yanhong Songzike’, ‘Meihong Guiye’ (6x), ‘Tongzimian’ (8x), ‘Baozhucha’ (6x), ‘Dalicha’, ‘Duxin Dalicha’, ‘Hentiangao’ (10x), ‘Pumencha’ (6x), ‘Songzike’ (6x), ‘Xiguiye’
Tengchong (35 cultivars)	‘Fentongcao’ (6x), ‘Mudankui’ (Figure 2a), ‘Fengshancha’, ‘Yunzhen’, ‘Wujiao Xiuqiu’ (6x), ‘Hehua Xianzi’ (8x), ‘Tuanye Diechi’ (6x), ‘Shuimeiren’ (6x), ‘Biyu’, ‘Yunfeng’ (8x), ‘Honghua Youcha’ (8x), ‘Fenzhen Mudan’, ‘Huahun’, ‘Jinrui Diechi’, ‘Xiaojiaojiao’, ‘Fenzhaoyun’, ‘Yushizi’, ‘Xianyecha’, ‘Yumeiren’, ‘Hongmei’, ‘Xiyingchun’, ‘Taohong Mudan’, ‘Yunhuacha’, ‘Fentianjiao’, ‘Saierqiao’, ‘Taihe Mudan’, ‘Manao Dahongcha’, ‘Jiaohe’, ‘Naochun Xiuqiu’, ‘Jiaoxiaohong’, ‘Fenhong Xiuqiu’, ‘Yanghong Songzike’, ‘Yulin’, ‘Nansongzi’, ‘Heidahong’

Note: the cell ploidy of cultivars was labeled according to the literature published by Xu et al. [7].

To generate low-copy nuclear gene markers, the genome coding sequences (CDS) from transcriptome data (SRX1343024) [21] were downloaded and filtered to select single isoform mRNA strands for ease of processing. Genes within the range of 500–1100 bp were chosen to ensure they could be easily amplified using traditional PCR methods. Next, the transcriptome reads were cleaned to remove adapter sequences, low-complexity sequences, contamination, and PCR duplicates [16,17]. Sequences that match plastid, chloroplast, ribosomal, transposon, or mitochondrial loci in the National Center for Biotechnology Information (NCBI) database (http://www.ncbi.nlm.nih.gov/, accessed on 9 December 2023) were then filtered out. Finally, sequences that show hits to RepeatMasker (http://repeartmasker.org/, accessed on 11 December 2023) were removed to ensure the selected gene markers were specific to the low-copy nuclear genes of interest.

Primer pairs were designed using Primer 4.0 [22] with a focus on conservative coding sequences. The primer selection phase involved the utilization of eight samples, including three randomly selected cultivars and the aforementioned five congeneric species.

Total DNA from C. reticulata leaf samples weas extracted using an improved CTAB method [23]. The PCR reaction mixture consisted of 25 µL Taq premixture, 1 µL DNA, 22 µL ddH2O, and 1 µL primers. The PCR reaction program followed a sequence of steps: initial denaturation at 98 °C for 2 min, 35 cycles including denaturation at 98 °C for 10 s, annealing at 60 °C for 10 s, and extension at 72 °C for 20 s, and final extension at 72 °C for 5 min. PCR products were sequenced using the ABI 3730 sequencer (Applied Biosystems, Beijing). Subsequently, mRNAs associated with amplified gene partial sequences were identified through an online Blast search (https://blast.ncbi.nlm.nih.gov/, accessed on 19 December 2023).

Upon sequencing with selected primers, low-copy gene sequences were compared through the MAFFT method [24] executed by the GeneiousR10 software. After alignment, crucial genetic parameters, including the number of haplotypes (h), haplotype diversity (H_d), and nucleotide diversity (P_i), were calculated using the DnaSP5.1 software [25]. The distribution of genetic variation within and between provenances was analyzed with the software Arlequin3.0 [26]. Analysis of Molecular Variance (AMOVA) was performed, and the genetic differentiation coefficient F_st was calculated. Gene flow (N_m) was estimated using the formula N_m = (1 − F_st)/4F_st [27]. The model-based cluster analysis was performed using the program STRUCTURE v2.3.1 [28]. The optimum number of clusters (K) was processed and identified by STRUCTURE HARVESTER through comparing log probabilities of data for each value of K [29]. The output of structure analyses was visualized using the software CLUMPP v1.1.2 [30] and DISTRUCT v1.1 [31].

The Kimura 2-parameter model was employed to calculate the genetic distance that was used to construct a Neighbor-Joining (NJ) phylogenetic tree, wherein the bootstrap method was used for testing through 1000 replicates. The Median-joining method [32], facilitated by Network5.0 (http://www.fluxus-engineering.com/sharenet.htm, accessed on 29 December 2023), was used for a comprehensive analysis of haplotype data, including the determination of phylogenetic relationships.

3. Results

3.1. Identification of Low-Copy Nuclear Genes of C. reticulata

A total of 966 partial sequences of low-copy genes were obtained through an integrated analysis of the C. reticulata transcriptome. Following the selection process, 30 low-copy nuclear gene sequences were identified, ranging from 538 to 1038 bp in length. Among the corresponding 30 primer pairs, 18 failed to amplify targets or exhibited faint bands, and seven displayed a problematic bimodal pattern during sequencing. Consequently, five primer pairs emerged with clear amplification bands and single-peaked sequencing outcomes (

Table 2. Primer pairs selected for the amplification of the low-copy nuclear genes in Camellia.

Primer Code	Primer Sequence (5′-3′)	Amplicon Length (bp)	Associated mRNAs	GenBank No
c1179-F c1179-R	ATCGCCAACAGAAACAACACGC ATTCACATCTAATGAGCGAAGGTTG	700	Disease resistance protein RPM1	MH257926
c10316-F c10316-R	CTCCCAACCCATCGTCCTTT CCTTGCCGCTCTTGCAATC	1038	Endo-1,3(4)-beta-glucanase	MH257929
c10214-F c10214-R	TGGAAGCTCGGCAATACCAG CCTTTGCGTTCATGGGCATT	538	Myosin-binding protein	MH257931
c11230-F c11230-R	TCTTGTGAGGTTAAGAGGGTTT CTTGGACATTATCATTGGAGCA	800	S-receptor-like serine/threonine protein kinase	MH257925
c11847-F c11847-R	ACATCGAAGAGCATGGCACA GCCCCAAACTAGCACTCTCT	786	Myb-like transcription factor	MH257932

). These five distinct low-copy nuclear gene sequences, registered in GenBank (MH257911-48), have been identified as partial coding sequences associated with unique mRNAs (Table 2).

3.2. Nucleotide and Haploid Diversity of C. reticulata Cultivars

Among five low-copy genes, C1179 showed the most single nucleotide polymorphism (SNP, 20), while C10214 showed the least number of polymorphic markers. The cumulative length of five amplified sequences was 3481 bp. A total of 71 mutation sites were identified (Figure A1), accounting for 2.04% of the overall sites. The overall nucleotide diversity (Pi) within C. reticulata cultivars was determined to be 0.0077, accompanied by an average nucleotide difference (k) of 26.65. Among four provenances, Tengchong displayed the highest nucleotide genetic diversity (Pi = 0.0084), followed by Dali (Pi = 0.0081), Kunming (Pi = 0.0068), and Chuxiong (Pi = 0.0066).

Locus variation analysis revealed a total of 91 haplotypes (Figure A1), with 35 from Kunming, 14 from Chuxiong, 10 from Dali, and 36 from Tengchong (

Table 3. Genetic diversity of Camellia reticulata cultivars from different provenances.

Provenance	Number of Cultivars	Average Number of Different Nucleotides (k)	Diversity of Nucleotide (Pi)	Number of Haplotypes (h)	Diversity of Haplotype (Hd)	Variance of Haplotype Diversity (Vh)	Standard Deviation of Haplotype Diversity (Sh)
Kunming	34	23.5544	0.0068	35	1.0000	0.0001	0.0070
Chuxiong	19	22.8070	0.0066	14	0.9825	0.0007	0.0260
Dali	12	28.0833	0.0081	10	1.0000	0.0027	0.0520
Tengchong	35	29.2404	0.0084	36	0.9958	0.0001	0.0070
Total	100	26.6465	0.0077	91	0.9974	0.0000	0.0020

). The collective haplotype diversity (H_d) attained an impressive 0.9974. Within four provenances, Dali and Kunming exhibited the highest haplotype diversity (H_d = 1.0000), closely followed by Tengchong (H_d = 0.9958) and Chuxiong (H_d = 0.9825).

AMOVA further revealed that a substantial 96.50% of the genetic variation within C. reticulata cultivars is distributed within the provenances, with only 3.50% accounting for the genetic variation between provenances (

Table 4. Analysis of molecular variance (AMOVA) of Camellia reticulata cultivars.

Source of Variation	df	SSD	MSD	Variance Component	Total Variance (%)	p-Value
Among populations	3	71.8090	23.9630	0.4719	3.50	<0.0010
Within populations	99	1247.1911	12.5980	12.9916	96.50	<0.0010
Total	102	1319.0001	36.5610	13.4635	100

).

3.3. Genetic Distance, Genetic Structure and Differentiation of C. reticulata Cultivars

Within four provenances, the genetic distance between provenance pairs is consistently comparable, ranging from 0.0070 to 0.0080 (

Table 5. Genetic distance (below) and differentiation (above) between Camellia reticulata cultivars.

Provenance	Kunming	Chuxiong	Dali	Tengchong
Kunming	–	0.0911	−0.0183	0.0331
Chuxiong	0.0070	–	−0.0040	0.0471
Dali	0.0070	0.0070	–	−0.0019
Tengchong	0.0080	0.0080	0.0080	–

). Notably, Tengchong exhibited a higher genetic distance from the other provenances. The genetic differentiation coefficient of C. reticulata cultivars stands out between Chuxiong and Kunming, registering the highest value at 0.0911. In contrast, the genetic differentiation coefficients between Dali and the remaining are relatively low.

The optimum number of clusters (K) was processed and identified by STRUCTURE, which was K = 5 (Figure 3), indicating that the Camellia reticulata cultivars were mainly composed of 5 gene pools (represented in red, blue, yellow, peach, and green, respectively). As shown in Figure 3, on the whole, the proportions of five gene pools in the four provenances are relatively uniform, indicating that the C. reticulata cultivars from four provenances have similar genetic structures. Among four provenances, Kunming showed relatively high proportion of peach gene pool (33%), while Tengchong showed relatively high proportion of red gene pool (34%).

The total genetic differentiation coefficient F_st for C. reticulata cultivars stood at 0.0351, with a corresponding N_st of 0.0199 (p < 0.05), falling below the threshold of 0.02. This suggests a relatively low genetic differentiation among diverse provenances of C. reticulata cultivars. The gene flow index N_m was 6.8827, significantly exceeding 1.0, pointing to a frequent exchange of genes between distinct provenances.

3.4. Phylogenetic Relationships of C. reticulata Cultivars

Cluster analysis divided 100 C. reticulata cultivars into four groups (Figure 4). Notably, three of these groups displayed a close relationship with the diploid ancestors of the allogeneic hexaploid C. reticulata. These three groups were specifically linked to the diploid C. reticulata, encompassing a total of 47 cultivars (19 from Kunming, 9 from Chuxiong, 5 from Dali, and 14 from Tengchong), the diploid C. pitardii, comprising 38 varieties (13 from Kunming, 9 from Chuxiong, 5 from Dali, and 11 from Tengchong), and the diploid C. saluenensis, featuring a collective count of 14 varieties (4 from Kunming, 1 from Chuxiong, and 9 from Tengchong). The fourth and final branch only has one cultivar, namely ‘Taihe Mudan’, originating from Tengchong.

Haploid network analysis showed that haplotypes H12 and H80 were located at the center of the network, while haplotype H53, distinguished by its extensive distribution and highest frequency, played a prominent role in shaping the network’s structure (Figure 5). As shown in Figure 5, there were no obvious diffusion trends among C. reticulata cultivars in any specific direction. Instead, a mixed and monogenic distribution pattern was observed in the chart, strongly indicating a close genetic kinship among cultivars from different provenances.

4. Discussion

4.1. Genetic Diversity of C. reticulata Cultivars

In comparing our AMOVA analysis results (96.50%) with those from previous studies utilizing AFLP markers [6], a remarkable consistency emerges, with the earlier study reporting 96.31% genetic variation within populations. This phenomenon can be explained from the following two aspects. Firstly, the genetic diversity within the wild populations of C. reticulata are relatively high [33], while cultivars are mainly selected from different wild variations. Secondly, the long-term breeding work may enhance genetic diversity within the provenances. Furthermore, the genetic distance between distinct provenances in our study (0.0070–0.0080) aligns closely with AFLP marker results (0.0077–0.0300) [6]. This robust alignment underscores the efficacy of low-copy nuclear gene markers developed in our study for a comprehensive analysis of C. reticulata cultivar genetic diversity.

Our study reveals an order in the genetic diversity of C. reticulata cultivars among different provenances: Tengchong > Dali > Kunming > Chuxiong. Contrastingly, a previous AFLP markers-based study suggested Kunming > Dali > Tengchong > Chuxiong [6]. This discrepancy likely stems from variations in sample composition. The AFLP study incorporated 190 samples, including 56 from Tengchong, 37 from Chuxiong, 28 from Dali, and 66 from Kunming [6], but only 85 of these were distinct C. reticulata cultivars, some of which were collected across different provenances. In the present study, we focused on a selected cohort of 100 cultivars, with a more balanced representation from Tengchong (35), Chuxiong (19), Dali (12), and Kunming (34).

Typically, F_st < 0.05 signifies a low level of genetic differentiation and suggests close relationships [6,34]. This pattern may be attributed to historical cultivation practices spanning 1300 years, during which cultivars were frequently introduced from their native provenances to new cultivation regions. These introduced cultivars then served as ancestors for breeding programs, associating the new cultivation regions as their provenances. This observation also likely explains the high gene flow (N_m = 6.8827), the high genetic variation (96.50%), and the similar genetic structures within populations.

4.2. Genetic Relationships among C. reticulata Cultivars

In this study, 99 C. reticulata cultivars were clustered in three clades (Figure 4), with each clade aligning with one of three diploid ancestors of C. reticulata, namely diploid C. reticulata, C. pitardii, and C. saluenensis. This clustering based on genetic relationships rather than provenances reaffirms the accuracy of recognizing diploid C. reticulata, C. pitardii, and C. saluenensis as ancestors of the hetero-hexaploid C. reticulata. Notably, the cultivar ‘Taihe Peony’ stands alone on a single branch, hinting at the potential involvement of other yet unknown parents in the genetic makeup of C. reticulata cultivars.

In terms of leaf morphology, ‘Xiaoguiye’ and ‘Xiguiye’ exhibit a striking resemblance to the diploid ancestor C. saluenensis. However, these two cultivars did not cluster with C. saluenensis but rather aligned with 2x C. reticulata and C. pitardii, respectively. This observation suggests that low-copy nuclear genes of allopolyploid C. reticulata stem from diverse ancestral species and, in this study, only a random copy from either source was amplified. Consequently, the five amplified low-copy nuclear gene copies of ‘Xiaoguiye’ and ‘Xiguiye’ predominantly correlated with two other ancestral species, diverging from the expected alignment with C. saluenensis.

Analyzing phylogenetic relationships in polyploids demands an extensive amplification of low-copy nuclear genes. The clone technology can be used to amplify numerous copies of a single sequence [12], while genome resequencing provides access to a large number of different low-copy nuclear genes [21]. In this study, the utilization of five low-copy nuclear gene markers across 100 cultivars resulted in a substantial dataset of 500 low-copy nuclear sequences. This abundance not only facilitates a more accurate study of phylogenetic relationships but also enables the tracing of the diverse ancestors of C. reticulata cultivars.

4.3. Evolutionary Dynamics of C. reticulata Cultivars

In addition to the well-known 6x cultivars, C. reticulata also encompasses 8x cultivars [7], such as ‘Honghua Youcha’ and ‘Zipao’ [35]. However, the origins of 8x cultivars remain a mystery.

Despite being recognized as an allopolyploid species involving 2x, 4x, and 6x variations, C. reticulata has recently been regarded as a part of polyploid complexes [5,36,37]. Genomic in situ hybridization (GISH) studies have revealed a fascinating evolutionary process. The diploid C. reticulata and C. pitardii gave rise to the allotetraploid C. reticulata. Subsequently, the allotetraploid C. reticulata and the diploid C. saluenensis led to the formation of the hexaploid C. reticulata in the Jinshajiang valley of south Sichuan [37]. The 6x C. reticulata then migrated from its origins to the Yunnan Province (Figure 1).

Interestingly, while 2x and 4x C. reticulata variants are confined to the Jinshajiang valley, no instances of 8x C. reticulata have been discovered in this region [38]. This raises the possibility that 8x C. reticulata may not originate through the duplication of the 4x variant. Furthermore, the widespread cultivation of 6x and 8x C. reticulata cultivars in Kunming, Chuxiong, Dali, and Tengchong, where 2x or 4x C. reticulata have not been documented, suggests that the 8x C. reticulata may have developed through additional hybrid polyploidization of the 6x variant during its southward diffusion. Notably, the unique clade ‘Taihe Mudan’ may be linked to unknown parents in this process.

5. Conclusions

This study successfully identified five pairs of primers suitable for amplifying low-copy nuclear gene sequences in Camellia. Our analysis encompassed 100 cultivars, revealing significant genetic diversity and low genetic differentiation of C. reticulata among four provenances. The majority of the cultivars (99) were categorized into three distinct clades, aligning with the diploid ancestors of the hetero-hexaploid C. reticulata. Newly developed low-copy nuclear gene markers in this study offer valuable tools for the genetic enhancement of C. reticulata and the advancement of related research endeavors involving Camellia species.