1 INTRODUCTION

Delineation of genetic diversity of endangered species has become a primary task for conservation biologists due to increasing global environmental change and anthropogenic impacts (Crandall et al., 2000; Spielman et al., 2004). Although numerous studies have been conducted on endangered terrestrial flora and fauna, research on the conservation and management of endangered seaweeds is still limited (Brodie et al., 2009; Couceiro et al., 2011). A major problem in addressing this paucity of literature is the deficiency of data on taxonomy, genetics, and phylogeography (Brodie et al., 2009). Conservation genetics provides vital theoretical and practical guidelines to maintain population sustainability and avoid endangered species from becoming extinct (Fagan and Holmes, 2006), as exemplified in the endangered red seaweed Ahnfeltiopsis pusilla (Couceiro et al., 2011). Maintaining a high level of genetic diversity is crucial for the sustainability of viable populations (Reed and Frankham, 2003), including the identification of evolutionarily significant units (ESUs) (Li and Ge, 2002). However, ESUs are sometimes difficult to manage because of their large geographical distribution (Li et al., 2020). The concept of management units (MUs) is to determine the appropriate population unit for the sufficient targeting of management and monitoring programs toward independent populations (Moritz, 1994).

The brown alga Silvetia siliquosa (Tseng et Chang Serrão, Cho, Boo and Brawly) is a perennial and functionally constructive species in the intertidal ecosystem (Mineur et al., 2015). This species was previously known as Pelvetia siliquosa and was transferred into a new genus, Silvetia, after phylogenetic reassessment (Serrão et al., 1999). The historical distribution range of S. siliquosa is restricted to the Yellow-Bohai Sea (Tseng and Chang, 1953, 1958) and the south and west coasts of Korea (Yoo and Kim, 2003; Oh et al., 2005; Park et al., 2007; Song et al., 2011; Hwang et al., 2015; Han et al., 2016). Interestingly, the S. siliquosa morphology is greatly affected by its dwelling environment; it exhibits more branches and larger fronds in a wave-sheltered habitat than in a wave-exposed habitat (Tseng and Chang, 1953). The genus Silvetia is monoecious. The species of this genus exhibit only sporophytic generation in their life history without an independent gametophytic stage, and the gametophytic generation is replaced by the conceptacles, antheridium, and oogoniums (Cho et al., 2001). From early October to late November when seawater temperature ranges from 14 ℃ to 18 ℃, S. siliquosa naturally undergoes spermatozoids and releases eggs, which develop into juvenile sporophytes after in vitro fertilization (Huang et al., 2008). Although preliminary research has investigated the biological characteristics of S. siliquosa, its population genetic connectivity and genetic structure remain poorly understood.

Population biomass and distribution range of S. siliquosa in East Asia have declined dramatically since the 1990s due to habitat fragmentation and anthropogenic impacts (Fig. 1; Tseng and Chang, 1958; Baek et al., 2007). Thus, this species has been listed as endangered in the ecological region of Yellow-Bohai Sea, with a high extinction risk (Kim et al., 2012). The market price of S. siliquosa as a commercial seaweed is approximately 10 USD/kg wet weight, which is much higher than those of other edible seaweeds such as Saccharina japonica and Undaria pinnatifida (Gao et al., 2017). Therefore, overfishing is a crucial reason for the rapid decline of population resources (Hwang et al., 2015). Additionally, temperature is a key factor that restricts the growth and development of S. siliquosa. The temperature in its growth habitat generally does not exceed 25 ℃ in summer (Tseng and Chang, 1958). When temperature is between 21 ℃ and 23 ℃ from early August to September, S. siliquosa releases spermatozoids and eggs from the conceptacles; however, they do not survive (Huang et al., 2008). After October, when the water temperature drops to 18.4 ℃, some fertilized eggs attach, germinate, and grow well (Huang et al., 2008). Global warming is estimated to be approximately 1.0 ℃ (±0.2 ℃) above pre-industrial levels, and the warming level is likely to reach 1.5 ℃ between the year 2030 and 2052 if it continues to increase at the current rate (Intergovernmental Panel on Climate Change, 2018). Therefore, the rising seawater temperature may exert a disastrous effect on the growth, development, and geographical distribution of S. siliquosa. The restoration and protection of S. siliquosa are an imperative task due to the severe diversity loss and range contraction.

In the present study, we obtained partial DNA sequences of three gene markers [mitochondrial cytochrome oxidase I subunit (cox1) + intergenic spacer (IGS) and ribosomal internal transcribed spacer II (ITS2)] of S. siliquosa from China and Korea. The present study had the following main goals: (i) to describe population genetic diversity and differentiation and identify MUs, (ii) to reveal population genetic connectivity in the Yellow-Bohai Sea by estimating gene flow. Our findings may provide the basic guidelines for the government and policy-makers to formulate conservation and management plans for S. siliquosa in East Asia.

2 MATERIAL AND METHOD 2.1 Sampling, DNA extraction, polymerase chain reaction and sequencing

Historically, S. siliquosa was widely distributed along the Shandong and Liaodong peninsulas of China and the western and southern coasts of Korea. However, its distribution range continuously declined from the mid-1990s onward (Song et al., 1996; Lee et al., 1997; Baek et al., 2007). From 2014 to 2018, we collected S. siliquosa specimens (n=145) from the Shandong peninsula and the southwest coast of Korea (Table 1 & Fig. 2). We randomly collected ≥12 individuals form localities where the specimens were commonly distributed and easily accessible, with an interval transect > 10 m. We only found a small patch of biomass (ca 40 cm×30 cm) in Rushan, China. Thus, only one individual was collected. A 3–5-cm tip of apical tissue was excised from each individual collected from all the localities and stored in silica gel for molecular analysis.

Total genomic DNA was extracted using a Fast Pure Plant DNA Isolation Mini Kit (Vazyme Biotech Co., Ltd., Nanjing, China). Mitochondrial cox1, 23S/trnK IGS (mtIGS), and nuclear ITS2 were chosen based on previous phylogenetic and phylogeographical studies of the order Fucales (Hu et al., 2007; Li et al., 2017a, b). Three primer sets (Supplementary Table S1) were developed based on the available sequences of S. siliquosa (GenBank accession number: AF102960.1), the congeneric Silvetia compressa (HQ990495.1), and Fucus vesiculosus (AY494079.1) by using the Primer-BLAST Program in the National Center for Biotechnology Information database. PCR amplification was performed in 50 μL of reaction volume, containing 25 μL of 2×Taq plus Master Mix II (Vazyme Biotech Co., Ltd., Nanjing, China), 2 μL of forward primer (10 μmol/L), 2 μL of reverse primer (10 μmol/L), 1 μL of template DNA, and 20 μL of ddH₂O. The thermo-cycling parameters were set according to the instructions of 2×Taq Plus Master Mix II, and the annealing temperatures were set at 58 ℃ for cox1, 55 ℃ for mtIGS, and 60 ℃ for ITS2. Electrophoresis, purification, and sequencing were conducted following the available protocols (Hu et al., 2007, 2011). Sequences were aligned and manually edited using Bioedit v7.2.5 (Hall, 1999) and MEGA v7.0 (Kumar et al., 2016).

2.2 Population genetic diversity and differentiation

Nuclear ribosomal ITS2 and concatenated mitochondrial cox1+IGS were used separately for the following analysis. Population genetic diversity was estimated using DnaSP v6 (Rozas et al., 2017) by assessing the number of haplotypes (N_h), haplotype diversity (h), nucleotide diversity (π), and the number of polymorphic sites (S). Pairwise population genetic differentiations were estimated by computing F_ST by using ARLEQUIN v3.5 (Excoffier and Lischer, 2010). The significance of covariance components for all results were tested using 104 permutations. Analysis of molecular variance (AMOVA) was computed in ARLEQUIN to partition genetic diversity among and within populations, and the significance was assessed through 104 random permutations. To assess the evolutionary relationships among mitochondrial haplotypes or ribosomal ribotypes, parsimony median-joining networks were generated using Network v10 (Bandelt et al., 1999).

2.3 Population connectivity

Gene flow is a vital quantitative indicator of inter-population connectivity that facilitates the conservation and management of endangered species. We used DnaSP v6 (Rozas et al., 2017) to evaluate gene flow between populations according to the methods described by Hudson et al. (1992) and assessed whether population connectivity is affected by surface currents in the Yellow-Bohai Sea. In the method, each polymorphic site was considered a separate locus and the F_ST from the frequencies of alleles at each locus in different geographic locations was figured out. According to F_ST, the gene flow N_m was measured by using the formula N_m=(1−F_ST)/NF_ST, where N_m denotes the number of migrants per generation and F_ST is the population fixation index, N=2 or 4 for mtDNA or nuclear DNA, respectively (Hudson et al., 1992). If N_m > 1, gene flow can resist the effect of genetic drift and prevent population differentiation, whereas if N_m < 1, genetic drift will become the dominant factor for divided population genetic structure (Slatkin, 1985).

3 RESULT 3.1 Molecular diversity

Concatenated mitochondrial cox1+IGS fragments were obtained for 136 individuals with an aligned length of 921 bp, including 525 bp of cox1 and 396 bp of mtIGS. The concatenated dataset contained two parsimony informative sites and yielded three haplotypes. Of these haplotypes, H1 was the most abundant and shared by 102 specimens, which accounted for 75% of all individuals (Table 1 & Fig. 2), whereas H2 and H3 were site-specific. The three haplotypes were unevenly distributed among the sampling sites. H1 dominated along the Shandong peninsula and the west coast of Korea, whereas H2 dominated in the populations of Jindo Island, Korea (JIN). The population of Gwanmaedo, Korea (GWA), was characterized by the richest haplotype (h=0.571± 0.014) and nucleotide diversity (π=0.062±0.000; Table 1).

We obtained 143 sequences, with an aligned length of 427 bp containing two parsimony informative sites, which yielded 3 ribotypes for ITS2. Of these ribotypes, R1 was shared among the five locations (except for Jindo Island), whereas R2 and R3 were site-specific and mainly distributed along the south coast of Korea (Table 1 & Fig. 2). Similarly, the populations from Gwanmaedo, Korea, harbored the highest ribotype (h=0.530±0.006) and nucleotide diversity (π=0.126±0.000), whereas other populations exhibited low or no genetic diversity.

3.2 Population genetic structure and differentiation

Haplotype network indicated no genetic structure in six S. siliquosa populations (Fig. 2). Concatenated cox1+mtIGS indicated that the populations JIN and GWA significantly diverged from others, with F_ST values ranging from 0.444 to 1.000 (Fig. 3; Supplementary Table S2). Despite relatively short geographical distance between JIN and GWA, the F_ST value between them was as high (0.925). No genetic differentiation (F_ST=0.000) was observed between the populations from China (LII and RUS) and the west coast of Korea (YEO and CHA). ITS2 revealed a similar pairwise F_ST pattern among populations, with the populations from the southern coast of Korea (JIN and GWA) highly diverged from others (F_ST=0.364–1.000). Moderate divergence (F_ST=0.069) was detected between the LII and YEO populations; however, this difference was statistically non-significant (P=0.121).

We defined the six populations as one group to partition genetic variance between and within populations. AMOVA results based on cox1+mtIGS indicated that most of the genetic variance (95.09%) occurred among populations, whereas only 4.91% occurred within populations (Φ_ST=0.950 9, P < 0.000 1; Table 2). Similarly, ITS2 exhibited that genetic variation among populations accounted for 86.89% of the total variance (Φ_ST=0.868 9, P < 0.000 1; Table 2).

3.3 Gene flow

Gene flow based on mitochondrial and nuclear markers yielded similar results. We observed a weak gene flow (N_m) between the GWA population in Korea and others (cox1+mtIGS: 0.041–0.626; ITS2: 0.105–0.437; Table 3). Similarly, either none or weak gene flow was observed between the JIN and other populations (cox1+mtIGS: 0.000–0.041; ITS2: 0.000–0.105; Table 3). Although ITS2 revealed a high gene flow (N_m=3.373) between the LII and YEO populations, the result was statistically non-significant. Moreover, no gene flow was observed among the three populations (YEO, CHA, and LII) based on both mitochondrial and nuclear datasets. Finally, we observed that the gene flow between the YEO and the CHA population was negative (N_m= -22.98). If the sum of differences among sequences from different populations is less than the sum of differences among sequences within a single population, it will result in N_m < 0 (Hudson et al., 1992). When the inter-population level of N_m is less than 0, it can be assumed that no gene exchange occurs between populations (Hudson et al., 1992).

4 DISCUSSION

In the present study, mitochondrial cox1+IGS and nuclear ITS2 consistently revealed a strikingly low population genetic diversity of the endangered S. siliquosa in East Asia, indicating a high risk of extinction. Based on population-level connectivity and genetic differentiation, we designated four MUs in East Asia for better conservation and restoration of natural S. siliquosa resources.

4.1 Low genetic diversity and range contraction

A high level of genetic diversity is essential for maintaining the evolutionary potential and viability of species. However, mitochondrial cox1+IGS and nuclear ribosomal ITS2 detected few genotypes, indicating that genetic variation in the S. siliquosa populations from China and Korea is strikingly low. Most threatened populations have lower average heterozygosity, evolutionary potential, and reproductive fitness than those of their non-threatened counterparts (Spielman et al., 2004). Thus, the low genetic diversity in S. siliquosa may aggravate the decline of species fitness and adaptability, eventually leading to the decrease in reproduction and increase in mortality. This pattern of no or low local genetic diversity has also been reported in other endangered algae such as Aegagropila linnaei (Soejima et al., 2009) and Ahnfeltiopsis pusilla (Couceiro et al., 2011). The endangered S. siliquosa populations are likely to become extinct because the assessment of some known extinct cases suggests that species extinctions can be preceded by a complete loss of genetic diversity (Johnson and Dunn, 2006).

Several reasons may account for the low genetic diversity and distributional contraction of S. siliquosa. Ocean currents and inappropriate environmental conditions are crucial in reducing genetic diversity and distribution range of species. According to the hypothesis of Tseng and Chang (1958), the historical distribution of S. siliquosa was shaped by mainly oceanic currents in the Yellow-Bohai Sea. Ocean currents can produce highly homogeneous algal genetic lineage in a wide space as a medium for gene migration and exchange (Muñiz-Salazar et al, 2005; Mitarai et al, 2009; Li et al, 2017b). In this case, we detected no clear genetic divergence among populations in both China (LII and RUS) and along the west coast of Korea (CHA and YEO, Fig. 3; Supplementary Table S2), indicating that the genetic structure of the S. siliquosa population is influenced by ocean currents and leads to a high degree of genetic homogeneity. On the other hand, the two geographically proximate populations (JIN and GWA) from Jindo, Korea highly diverged from others with a weak genetic exchange (Fig. 2; Table 3), indicating that the two isolated populations may be subject to the decrease in genetic diversity (Wofford et al., 2005; Riley et al., 2006). Unsuitable environmental conditions, such as storms, affect the zygote attachment and the growth and development of juvenile thalli. S. siliquosa generally germinates in a wave-sheltered habitat, where it grows abundantly and the individual is large. On the other hand, S. siliquosa that grows on exposed rocks susceptible to wind and waves has fewer branches and smaller fronds (Tseng and Chang, 1958). S. siliquosa cultures in laboratory indicate that the zygotes should adhere to the substrate after fertilization because the attachments of the zygotes are not firm and it must be cultured in still water (Li et al., 2007). If it fails to adhere to the substrate, the fertilized eggs will wither up and discolor, and eventually fail to develop into young thalli (Yoon et al., 2003). Therefore, storm might interfere with the growth and development of S. siliquosa. Alternatively, seawater temperature can not only affect the release and survival of S. siliquosa sperm and egg (Huang et al., 2008), but also affect its growth, development, and recruitment (Hwang et al., 2015; Gao et al., 2017). The S. siliquosa receptacles develop and mature from March to August when seawater temperatures range from 8 ℃ to 26 ℃ (Hwang et al., 2015), and the most suitable water temperature for the release of sperm and egg is below 18.4 ℃ every year (Huang et al., 2008). However, the sea surface temperature (SST) in the Yellow-Bohai Sea has increased by 0.67–0.89 ℃ from 1982 to 2006 (Belkin, 2009). This increase in the SST may be partly responsible for the reduction in the geographical range of S. siliquosa in East Asia and hence the loss of genetic diversity. High-temperature stress resulted in the loss of genetic diversity and adaptability of Fucus serratus (Jueterbock et al., 2014), a brown alga closely related to S. siliquosa. Thus, the low level of genetic diversity further reduces individual fitness and population adaptability (Kaljund and Jaaska, 2010), which eventually results in small populations. Thus, future work should concentrate on the effect of increased temperature and irradiation on the fertilization, settlement, growth, and reproduction of S. siliquosa. Furthermore, S. siliquosa is commonly used as food and possesses medicinal value in China (Lee et al., 2003; Kang et al., 2005; Spavieri et al., 2010), and the current market price far exceeds other brown algae. Consequently, commercial interest results in overfishing of natural S. siliquosa resources, which accelerates the decline in its biomass and hence causes its extinction, which renders the species endangered with low genetic variation. Gadus morhus (Reynolds et al., 2005) and Aegagropila linnaei (Soejima et al., 2009) are two comparable cases in which overfishing made them endangered.

4.2 Population connectivity and the conservation of S. siliquosa

Understanding connectivity is the key to manage and conserve marine ecosystems (Bradbury et al., 2008). MUs provide a solution for independent populations affected by human activities (Palsbøll et al., 2007), particularly for the short-term management and monitoring of natural populations (Schwartz et al., 2007). If the degree of population connectivity is sufficiently low, each population should be monitored and managed separately (Taylor and Dizon, 1999). Alternatively, MUs can be assigned when the observed genetic divergence is significantly greater than a predefined threshold value (Palsbøll et al., 2007). In the present study, JIN and GWA from Jindo, Korea, exhibited higher levels of genetic differentiation than other populations, indicating a poor gene flow and low connectivity (N_m=0.041–0.105, F_ST=0.705–0.925, Table 3; Supplementary Table S1). Thus, these two populations (JIN and GWA) can be designated as two independent MUs. Additionally, a considerable geographical distance is present from the west coast of Korea to the Shandong peninsula. Absence of a significant genetic differentiation and weak gene flow imply that both the LII and RUS populations in China and the YEO and CHA populations along the west coast of Korea should be designated individually as a single MU.

The delineation of the underlying genetic data of the MUs provides a theoretical basis for the monitoring and management of specific target species and relevant conservation objectives. However, the strikingly low genetic variation and weak population connectivity in East Asia may lead to a low evolutionary potential of S. siliquosa. The self-sustainment of future populations may not be sufficient to support long-term survival, which necessitates human intervention to increase its abundance and distribution range. Transplantation provides technical support for increasing natural populations and restoring the degradation of natural habitats. The extensive transplantation experiments with brown seaweeds such as Fucus gardneri, Macrocystis pyrifera, and Cystoseira compressa (Stekoll and Deysher 1996; Hernández-Carmona et al., 2000; Susini et al., 2007) have been undertaken successfully in the field to restore natural resources. Because the S. siliquosa population in Gwanmaedo, Korea, is characterized by the highest genetic diversity (h=0.571±0.014; π=0.062±0.000; Table 1), seedlings can be harvested in the region and transplanted to other regions after artificial incubation. The optimum time for seedling collection is between early October and late November each year when the SST drops to less than 18.4 ℃, and it can be cultured in still water and natural light after zygote formation for better adhesion to the glass or other substrates (Li et al., 2007; Huang et al., 2008). Additionally, the optimum salinity for culturing sporophytes is 18 g/L. The rhizoid hairs that are conducive to attachment develop abundantly, and the young sporophytes grow well under these conditions (Huang et al., 2008). A study indicates that the transplants developed using polyethylene ropes grow and mature well in new environments, which suggests that population restoration of S. siliquosa may be achieved through polyethylene rope transplantation (Gao et al., 2017). However, the growth of S. siliquosa populations may be hindered due to competition and grazing (Gao et al., 2017). In addition to preserving the regions with high genetic diversity, other unique populations or regions should receive special attention. For example, we only found one small patch of S. siliquosa in Rushan, Weihai. Thus, the ex situ conservation may sustain survival of S. siliquosa. Future studies are required to better understand the effect of environmental factors on survival and recruitment of S. siliquosa after the native reciprocal transplantation.

5 CONCLUSION

The present study reveals a strikingly low population genetic diversity of the endangered S. siliquosa in East Asia, with a high risk of extinction. In view of the scenario of diversity loss and habitat fragmentation, our work provides preliminary insights into diversity conservation and population restoration. Based on the degree of genetic differentiation, population connectivity, and geographical distance between China and Korea, we propose that four independent MUs can be designated for the management and protection of S. siliquosa in East Asia. Artificial transplantation may be a feasible step for sustaining manageable resources and functions of seaweeds. To maximize the efficacy of conservation, future research should link the genetic information to the transplantation and restoration of outdoor S. siliquosa resources.

6 DATA AVAILABILITY STATEMENT

The datasets of sequences analyzed in this study can be found in the GenBank (cox1: MW093753–MW093890; mtIGS: MW093891–MW094035; ITS2: MW082642–MW082784).

7 ACKNOWLEDGMENT

We thank Drs. Zhongmin SUN and Jing WANG for providing assistance during field collection and specimens' preservation in Korea.

Electronic supplementary material

Supplementary material (Supplementary Tables S1 & S2) is available in the online version of this article at https://doi.org/10.1007/s00343-021-0400-y.