Genetic Structure in 21 Remnant Phoebe Sheareri Populations in Southern China: Implications for Genetic Resource Conservation

Background: Phoebe sheareri (Lauraceae) is a valuable and endemic tree species in China, with limited large natural communities remnant. Genetic diversity and differentiation analysis are essential to manage their conservation and utilization. To provide a conservation and utilization strategy of P. sheareri based on sound genetic diversity and differentiation data. Results: We found medium level of genetic diversity and low inbreeding. Nei's gene diversity and Shannon’s information index value showed medium genetic diversity in nature populations of P. sheareri, which was higher than other two Phoebe species. AMOVA showed the genetic differentiation among populations was signicantly, and 21.2% of genetic variation was among populations. Bayesian clustering, obtained with STRUCTURE, grouped the populations into four genetic clusters, whereas UPGMA analysis distinguished three main groups approximately in line with the geographic area of occurrence. Conclusion: Based on the study results, the establishment of gene conservation units must be considered in nature conserves in order to protect the genetic diversity of the species, and the proposal of sampling strategies for ex situ conservation and reforestation.


Background
Phoebe sheareri (Hemsl.) Gamble (Lauraceae) is a valuable tree species unique to China. Its hard wood, with smooth cut surfaces and beautiful texture is highly valued for quality furniture and engraving material. P. sheareri is an important original plant species of the commercial timber known as "golden nanmu", with the most beautiful wood vein and highest cold tolerance among Pheobe species. The habitat of this tree species usually was warm and humid environment, which is commonly found in broadleaf forests of mountain valleys below 1,000 m (Liu 1957). P. sheareri is naturally distributed in Zhejiang, Jiangxi, Guizhou, Fujian, Guangdong, Guangxi, Hunan and Hubei provinces in southern China, within the range 24.50-32.05°N, 106.23-121.45°E (Fig. 1). However, due to excessive harvesting, lack of protection, and weak natural regeneration capabilities (Chen et al. 2013), P. sheareri natural stocks are in decline, with limited natural communities remnant (Chen et al. 2014). This fragmentation of habitats will hinder the gene ow between two populations, increasing the possibility of the reduction of genetic diversity. With the aim to ensuring the genetic diversity from being decreasing, it is urgent to the proposal of an effective protection strategy for P. sheareri. The understanding of the genetic diversity and structure of P. sheareri natural population is the prerequisite to making conservation strategy.
Genetic diversity is an essential component of biodiversity, and is of great signi cance for understanding the origin of species, predicting adaptability, and estimating the distribution of genetic resources (Muriira et al. 2018). Genetic markers are the main tools for studying genetic diversity; these can include phenotypes, cytology, biochemistry, and molecular markers (Ge et al. 1988; Qin et al. 2016). The advantage of expressed sequence tags microsatellite (EST-SSR) markers is that their polymorphism is associated with the transcribed sequences, re ecting the variation in the expressed region of the genome. EST-SSR markers also have the advantages of good stability, high polymorphism, co-dominant inheritance, and easy operation. It has been widely applied to study genetic diversity in many plant species (Varshney et al. 2005). Fraxinus chinensis Roxb. (Wu et al. 2016) were conducted to understand their genetic diversity and genetic structure of natural populations.
However, the genetic diversity and structure of natural residual P. sheareri populations were unclear, which hinders the proposal of protection strategy in P. sheareri. In the present study, with the aim to proposing a conservation strategy for this endangered species P. sheareri, 32 pairs of polymorphic EST-SSR primers were applied to analyze 21 natural populations involving in 428 individuals. Firstly, the genetic diversity and structure of 21 populations were evaluated. Secondly, the genetic differentiation within and between populations was also evaluated and combined with the geographic distance and climatic index to reveal the potential force driving the present structure. Lastly, the conservation strategy for this valuable tree species has been proposed.

Field study and experimental materials
Following an in situ survey of the natural distribution of P. sheareri, 21 natural populations with representative habitat conditions were selected for sampling, from which we collected a total of 428 samples. Sampling trees were separated by a minimum distance of 30 m. The 21 sampled populations included one population in Jiangsu province (ZJS), 12 in Zhejiang province (XH, TMS, THY, CH, LHT, ZJB, LJ, CA, NH, SC, SY, and QY), and two populations from each of the Jiangxi, Hunan, Guizhou and Guangxi provinces (populations WY, LC; ZJ, XN; FJS, XS; and ZY, LS, respectively), representing the major natural P. sheareri habitats of China (Table 1, Fig. 1). We collected 15 pieces of young leaves from each tree; sampled leaves were quickly placed in ziplock bags, stored in a cooler on ice, and brought to the laboratory, where they were frozen at − 40 °C until DNA extraction. The Global Positioning System (GPS) was used to record the latitude, and the longitude. The main climatic factors of each sampling site were obtained through the China Meteorological Science Data Sharing Center platform (Table 1).  P. sheareri is a protected species, wood cores or destructive sampling was forbidden to determine the age of individuals. Therefore, we analyzed population age structure according to diameter class, combing with the sapling height, and basal diameter (D) was used as a standard to categorize the diameter classes (Liu et al. 2015). Recording to Qu et al (1952), the rst diameter class (I) was de ned as D ≤ 2.5 cm and H ≤ 0.33 m; the second diameter class (II) was de ned as D ≤ 2.5 cm and H > 0.33 cm; the third diameter class (III) was de ned as 2.5 cm < D ≤ 7.5 cm; the fourth diameter class (IV) was de ned as 7.5 cm < D ≤ 22.5 cm; and the fth diameter class (V) was de ned as D > 22.5 cm. Then, the plotted diagrams of population diameter-class structure were plotted (Qu et al. 1952).

Dna Extraction
Approximately 100 mg of young leaves per sample were used for genomic DNA extraction using the CTAB Plant Genomic DNA Rapid Extraction Kit (Aidelai Biotechnology Company, Beijing, China). DNA quality and concentration were determined using Nanodrop 2000 system, and DNA integrity was evaluated using 1% agarose gel electrophoresis. DNA was diluted to a nal concentration of 25 ng/µL and stored at − 20 °C prior to EST-SSR ampli cation.

Est-ssr Primer Screening
A total of 6958 SSR loci were detected in P. sheareri transcriptome data using MISA (http://pgrc.ipkgatersleben.de/misa/misa.html), and batch-developed EST-SSR primers were produced using Primer 3.0, as described previously (Lu et al. 2018). Ninety-four pairs were randomly selected and synthesized by the Nanjing Kingsray Biotechnology Co. Primers, then were selected using 10 different P. sheareri DNA sources, and polymorphisms were used in this study (Table S1).

Data Analyses
The DNA fragments ampli ed by the SSR primers were recorded as alleles based on the comparison with external standards, and a bp data matrix was obtained. Use DateTrans1.0 software to convert the data into the required form for subsequent analysis ( The Bayesian clustering analysis was conducted to evaluate the genetic structure of P. sheareri populations using STRUCTURE version 2.3 (Daniel et al. 2003). This method estimates the number of genetic clusters (K) in the data and estimates the ancestry of individuals in these clusters (Daniel et al. 2003). The admixture model with correlated allele frequencies was applied to Bayesian analysis. Models were tested for K-values (testing from K = 1 to K = 10), and the model was run with 10 independent stimulations for each K, and a burn-in period of 100,000 iterations and 100,000 Markov chain Monte Carlo (MCMC) repetitions (Daniel et al. 2003). The most likely K value was determined using Structure Harvester based on both the log likelihood and the maximum ΔK (Earl et al. 2012).
Due to the mutation rate of the microsatellite sequence and the effective content of the population, the accuracy of the phylogenetic tree by UPGMA (Unweighted Pair Group Method with Arithmetic means) is slightly higher than that of Bayesian analysis (Takezaki et al. 1996;Nei et al. 2000). The UPGMA tree was constructed based on Nei's genetic distance using PopGene 32 (Yeh et al. 2000).
Geographic distances between populations were estimated with Earth Explorer 6.5, the Mantel test between genetic distance and geographic distance, annual precipitation difference, annual mean temperature difference was detected by XLSTAT software (XLSTAT 2017).

Results
The population characteristics of P. sheareri The diameter-class of 21 populations showed varied distribution patterns (Fig.S1). The class I seedlings were predominantly in the Population LS; the class II saplings had proportional advantage in the Population ZJS, CH, and WY; the class III young trees were predominantly in the Population THY, ZJB, QY, and LC; the class IV middle-aged trees had dominant position in the Population TMS, CA, LJ, XH, NH, SC, SY, ZJ, XN, FJS, XS, and ZY, especially this class tree accounting for above 50% in Population LJ and ZY; while the class V big-aged trees were predominantly in the Population LHT. However, none class V tree was observed in nine populations such as ZJS, THY, and ZJB (Fig.S1).

Polymorphisms Of The Est-ssr Markers
A total of 428 samples from 21 P. sheareri populations were ampli ed using 32 pairs of EST-SSR primers, and 105 alleles were detected. The value of Na ranged from 2 to 5 (mean = 3.219), and  The genetic structure among P. sheareri populations Using a ΔK value to determine a reasonable K, the ΔK value reached a maximum when K = 4 (Fig. 2a). The 21 natural populations of P. sheareri could then be divided into four distinct groups. The rst category included ZJS, XH, TMS, THY, CH, LHT, ZJB, LJ, and CA; the second category comprised NH, SC, WY, and LC; the third category included SY, QY, and XN; and the fourth was composed of ZJ, FJS, XS, ZY, and LS (Fig. 2b).
Fst and Nm among P. sheareri populations The AMOVA indicated larger genetic variation within populations than among them (Table 3). Nevertheless, genetic differentiation among populations was extremely signi cant (P < 0.001). Fst analysis indicated that 21.2% of genetic variation was among population (Table 3). Similar results were obtained when calculating by PopGene software (Fst = 0.227, Table S3). The inbreeding coe cient (Fis) was predicted to be low, only 0.036. The average Nm calculating by all SSR loci was 1.322, and the Nm derived from above formula involving in Fst was 0.927, suggesting that there is relatively low gene ow among P. sheareri populations.

Genetic Distance Among Populations And Cluster Analysis
Nei's genetic distance and Nei's genetic identity among the 21 natural populations of P. sheareri were shown in Table S4, and they ranged from 0.077 to 0.492 and from 0.612 to 0.926, respectively. TMS and THY had the largest genetic identity and the smallest genetic distance, indicating that their kinship was closer than that of other populations. Genetic identity was smallest between the SY and LS populations and their genetic distance was the largest, indicating higher genetic differentiation between two populations.
Based on the Nei's genetic distance between populations, UPGMA cluster analysis was performed on the 21 populations (Fig. 3).

Mantel test between genetic distance and geographical distance and climate difference
Mantel test analysis showed a signi cant correlation between geographic and genetic distance among the P. sheareri populations (r = 0.624; P < 0.0001; Fig. 4). It indicated that the geographic distance observed among populations were key factors in uencing genetic differentiation. The Mantel test showed that the genetic distance was insigni cantly positively correlated with temperature difference (r = 0.114, P = 0.098, Supplementary Fig.S2a), but signi cantly correlated with the precipitation difference (r = 0.204, P = 0.003, Supplementary Fig.S2b). Furthermore, the mantel test showed that the genetic distance was signi cantly positive correlated with geographic distance among western populations (r = 0.525, P < 0.0001), whereas insigni cantly with precipitation difference (r = -0.027, P = 0.784, Supplementary Fig.S3). In the case of eastern populations, neither precipitation difference nor geographic distance was signi cantly with genetic distance ( Supplementary Fig.S3).

Discussion
Genetic diversity in P. sheareri Both Nei's and I were used to re ect the level of genetic diversity, such that greater values indicate higher genetic diversity within the population. In this study, the average values of Nei's and I were 0.376 and 0.576, respectively, indicating that natural populations of P. sheareri contain rich genetic diversity. There is an inseparable relationship between the genetic diversity of a species and its living habits and life history characteristics (Hamrick et al. 1979). High genetic diversity has also been detected in other subtropical tree species, which is likely a consequence of life history traits of these trees, such as a long life span and a predominantly outcrossing mating system (Hamrick et al. 1996;Petit et al. 2006). Genetic diversity is also affected by geographical distribution, population size, and climate change caused by glaciers (Angela et al. 2012). Populations that are continuously distributed over a large area have more opportunities to maintain the level of allelic diversity than niche populations (Michele et al. 2014). P. sheareri is currently distributed in southern China, with a relatively wide distribution region, which is consistent with the medium genetic diversity. In addition, P. sheareri harbored higher genetic diversity than other Phoebe species, such as P. chekiangensis (Ding et al. 2015) and Phoebe bournei (Hemsl.) Yang ). It was consistent with the wider distribution region in P. sheareri than other two Phoebe species.
Genetic structure among P. sheareri populations Analysis of genetic diversity and population structure is of great signi cance for plant molecular breeding and protection of genetic resources. Fst is an effective way to measure genetic differentiation and gene ow between populations (Peng et al. 2017 Previous studies showed that woody species with large geographic ranges and outcrossing mating system usually harbored more abundant genetic variation within population than those of among populations (Hamrick et al. 1992). P. sheareri with the ten-years old began owering, which is the hermaphrodite ower with diverse volatile secondary compounds and insect pollination. The levels of inbreeding was expected to low in P. sheareri populations, as estimated by Fis = 0.036 (Table S3) Geographic isolation restricts gene ow. The structuring of diversity within and among populations is expected to be related to effective population size and gene ow (Hamrick et al. 1992). Gene ow between populations is thought to be existed when Nm > 1, indicating that alleles are distributed among different populations, reducing the probability of genetic drift (Slatkin 1987). The average Nm among P. sheareri populations were 1.322 (Table S3), and the Nm calculating from Fst was 0.927, suggesting that there is relatively low gene ow among P. sheareri populations. This value is smaller than those reported for P. chekiangensis (1.992) (Ding et al. 2015) and other widespread species such as B. luminifera In situ survey of the natural distribution of P. sheareri showed that there are fragmented biomes in the distribution range.
Previous studies showed that the fragmented biomes were the interaction of the factors, such as biological characteristics, habitat heterogeneity and arti cial disturbance (Liu et al. 2015).
The key factor limiting population development was low natural regeneration ability. The reproduction of P. sheareri is characterized by the production of many seeds, while the germination was affected by soil moisture condition. The germination rates of seeds are high (75%) under suitable moisture, such as under big trees. However, the seedling growth was signi cantly affected by light condition, exempli ed by when the young trees were older than three years, the shading limited the growth, even causing the death (Chen et al. 2013). Only a few seedlings of P. sheareri survive to become adults, increasing the di culty to enlarge the population size. It was consistent with the varied distribution patterns of diameterclass in 21 populations (Fig.S1).
In addition, further research found that the genetic distance between populations was insigni cantly correlated with the temperature difference but was signi cantly correlated with precipitation difference, indicating that precipitation has a role in population differentiation. It was consistent with biological characteristics of P. sheareri, which enjoys humid environment and has a certain cold tolerance, so the difference in water is more signi cant for population differentiation.
Conservation strategy of P. sheareri natural population An important component of biodiversity is the diversity of forest genetic resources. Due to the deterioration of the ecological environment and the frequent intensi cation of human activities, the natural population of P. sheareri has been gradually decreased and is fragment-shaped, which is not conducive to maintaining the genetic diversity of P. sheareri. The ability of a population or species to evolve and adapt to the environment depends on its level of genetic diversity. Low levels of genetic diversity are not bene cial for areas of increased distribution and may increase the likelihood of disease or pests ). Environmental conditions and species characteristics are also considered key factors that may affect genetic diversity. Therefore, understanding the value of genetic diversity and population genetic differentiation is critical to identifying current threats to conservation and elucidating the mechanisms for protecting endangered species (Petit et al. 2010). As a tree species with high ecological and economic value, it is urgent to formulate a reasonable and effective protection strategy. In the present study, the natural population of P. sheareri has a medium degree of genetic diversity, with obvious genetic differentiation. The diversity level within the population is much higher than that of among the populations. Therefore, in situ conservation is the main strategy for protection. We found that the distribution patterns of diameter-class varied signi cantly in 21 populations (Fig.S1). For these populations locating in nature reserves and harboring more than one hundred individuals, such as TMS, FJS, and THY harboring relative high genetic diversity, the key is to effectively protect the native forest ecosystem. None class V tree was observed in THY, and other eight populations, indicating that the arti cial disturbance might cause limitations for the population self-propagation. So we suggest that P. sheareri should be included in the endangered species protection list Redbook, thus more people know this species and increase people's protection awareness. In the nature reserves, such as THY, TMS, FJS, and LS, the P. sheareri population and its biodiversity, and ecological environment should be further studied, then we can expand the population and its distribution are by thinning forest stand, cultivating seedling reintroduction, and helping population growth. Although ZJ harbored less than 200 individuals, the genetic diversity index ranked second. Five diameter-class individuals were detected and the class IV trees and class II saplings were predominantly in this population. We should improve the population environment to help the class I and class III individuals' growth, to sustain the genetic diversity and structure in ZJ.
Due to the seed germination of P. sheareri requires certain shade conditions and has certain di culties, in addition to in situ conservation, we can collect P. sheareri seeds in the main distribution region. Through arti cial breeding, seedlings with three-years old were planted in different places, such as the original population, thereby expanding the weak P. sheareri populations, and can also establish offspring test and genetic resource orchards to protect the genetic diversity of P. sheareri.
On the other hand, the existing individuals by through vegetative propagation and their offsprings can be collected in the gene reservation center of P. sheareri. In addition, seed orchards, and collection nurseries of superior tree should be established to preserve the diversi ed genotypes to ensure requirement of long-term breeding in P. sheareri.

Conclusion
Page 13/20 The genetic diversity and genetic structure of 21 natural population of P. sheareri involving in 428 individuals were evaluated using 32 pairs of EST-SSR primers. The results showed that P. sheareri had a medium genetic diversity, indicating that those populations with high genetic diversity should be conserved in situ. However, there was signi cant genetic differentiation among populations, and the gene ow was relatively low, indicating that gene ow between populations might be blocked and genetic drift might have been existed. Genetic structure revealed that signi cant genetic differences among populations were observed, which correlated with geographical distances. It suggested that habitat fragmentation and geographical isolation, and precipitation difference contribute to the genetic differentiation of natural populations in P. sheareri.

Declarations
Ethics approval and consent to participate Not applicable.
Consent for publication: Not applicable.
Availability of data and material: We have deposited the raw data into the public repository, and the DOI was https://doi.org/10.6084/m9. gshare.12332552.v4 Competing interests: The authors declare that they have no con ict of interest.   Genetic structure analysis of P. sheareri populations. a) The relationship between the number of clusters (K) and the corresponding ΔK statistic was calculated from ΔK according to structural analysis. b) Results of the structure analysis of P. sheareri populations when K = 4. Each individual is represented by a single vertical bar, which is partitioned into four different colors, each representing a genetic cluster; colored segments show the estimated ancestor ratio of the individual to each genetic cluster.  Mantel test on the correlation between Nei's genetic distance and geographic distance (km) among P. sheareri populations. Genetic distance was signi cantly positively correlated with geographic distance.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.