Genetic Diversity and Connectivity of Ocypode ceratophthalmus in the East and South China Seas and Its Implications for Conservation

Simple Summary This study investigated the genetic diversity and connectivity of 15 Ocypode ceratophthalmus populations in the East and South China Seas based on two genetic markers. The results showed that O. ceratophthalmus had a high genetic diversity among all collected populations, and an insignificant population structure was observed by a hierarchical analysis of molecular variance and fixation index. Additionally, Migrate-n revealed high historical gene flow and migration rates among populations. The results of this study could inform the construction and management of marine protected areas in the East and South China Seas. Abstract The East and South China Seas are rich in marine resources, but they are also under great pressure from climate change and human activities. Maintaining diversity and connectivity between communities is thought to be effective in mitigating these pressures. To assess the diversity and connectivity among the populations of Ocypode ceratophthalmus in the East and South China Seas, 15 populations from or near 15 marine protected areas in the two seas were studied using COI and D-Loop as genetic markers. The results showed that O. ceratophthalmus populations had high diversity, and the results of a hierarchical analysis of molecular variance and fixation index found that there were no significant genetic structures among these populations. High historical gene flow and high migration rates were further observed among populations by Migrate-n. Furthermore, the COI sequences further showed the asymmetric migration rate with a higher migration rate from south to north than from north to south. This information could provide recommendations for the management of marine protected areas in the East and South China Seas.


Introduction
The East and South China Seas are important areas for the utilization of marine resources and marine conservation. Currently, 12,933 tropical and subtropical species in the East China Sea (ECS) have been recorded, of which half (48%) were endemic [1], and the ECS was also an important fishery for the world, contributing about 6 million tons of catches annually [2]. The South China Sea (SCS), adjacent to the ECS, is regarded as the hotspot of tropical biodiversity in shallow seas, with 450 species of coral (about 7% of the world's total coral reef areas), 2 million hectares of mangrove (12% of the world's total mangrove areas), 1027 species of fish, 91 species of shrimp and 73 species of cephalopods [3]. Furthermore, the SCS provides 6 billion tons of catch annually, accounting for 10% of the world's total catch [3]. Nevertheless, with the rapid economic development in the East and South China Seas, the regions were also facing threats, such as overfishing, biodiversity decline and habitat degradation, which urgently need to be strengthened for conservation [3]. The marine protected area (MPA) was considered the core initiative of marine biodiversity conservation [4], and MPAs were established to protect the ECS and SCS 20 years ago by * In this column, the contents in parentheses indicate that the sampling site is in or near the marine protected areas, "in" indicates that the sampling site is in the reserve, and "near" indicates that the sampling site is near the marine protected areas.
The obtained samples were morphologically identified according to the previous study [25], samples were transported back to the laboratory, measurements of carapace width, carapace length, abdomen width, abdomen length, weight and gender were conducted [26]. Then, samples were stored in 95% ethanol at −20 • C until DNA extraction. About 20 mg of muscle tissue from the ambulatory leg was obtained to extract the total DNA by using MolPure ® Cell/Tissue DNA Kit (Yeason, Shanghai, Chian), and the DNA was stored at −20 • C and used as a template in the PCR reactions. PCR was implemented to amplify COI and D-Loop using the following primers, LCO1490l (5 -GGTCAACAAATCATAAAGATATTGG-3 ) and HCO2198 (5 -TAAACTTCAGGGTGAC CAAAAAATCA-3 ) for COI, 13,323F (5 -GCGAATGCTGGCACAAACAT-3 ) and 14,378R (5 -AGGGAGTGGTGCAATTCCAT-3 ) for D-loop. The PCR amplification reaction included 25 µL 2× Hieff ® PCR Master Mix, 2 µL of upstream and downstream primers (10 µM), about 55 ng of template DNA, and double-distilled water was added to the total volume of 50 µL. Then, the PCR thermal cycling program was set as follows: denaturation at 94 • C for 5 min, followed by 35 cycles of denaturation for 94 • C 30 s, annealing at 55 • C for 30 s, elongation at 72 • C for 1 min; and the final extension step for 10 min at 72 • C. All PCR products were detected by 1.5% agarose gel electrophoresis. Finally, the PCR products were purified and sequenced by Tsingke Biotech Company (Shanghai, China). and downstream primers (10 μM), about 55 ng of template DNA, and double-distilled water was added to the total volume of 50 μL. Then, the PCR thermal cycling program was set as follows: denaturation at 94 °C for 5 min, followed by 35 cycles of denaturation for 94 °C 30 s, annealing at 55 °C for 30 s, elongation at 72 °C for 1 min; and the final extension step for 10 min at 72 °C. All PCR products were detected by 1.5% agarose gel electrophoresis. Finally, the PCR products were purified and sequenced by Tsingke Biotech Company (Shanghai, China).

Figure 1.
Sampling locations and ocean currents schematic in the studied area. TWC, Taiwan Warm Current; SCSWC, South China Sea Warm Current; MZCC, the coastal currents of Zhejiang and Fujian; GDCC, Guangdong coastal current. The direction of MZCC and GDCC varies seasonally, from south to north in summer and vice versa in winter [27], which is indicated by the solid line with bidirectional arrows in the figure.

Statistical Analyses of Genetic Data
All sequences were checked and edited using Geneious. Multiple sequences were aligned using the Geneious Alignment (Identity 1.0/0.0), and all alignments were checked visually. MEGA11 was used to evaluate the composition of the nucleotides. DnaSP v.4.0 was used to calculate the genetic diversity indices such as the number of haplotypes (h), haplotype diversity (Hd) and nucleotide diversity (π).
A map of haplotypes distribution was constructed using ArcGIS, R 4.2 and Arlequin 3.5, and minimum spanning networks were constructed using PopART 1.7. Hierarchical analysis of molecular variance (AMOVA) of the different levels was used to further examine the population structure. The AMOVA based on three levels included population level (  GDCC, Guangdong coastal current. The direction of MZCC and GDCC varies seasonally, from south to north in summer and vice versa in winter [27], which is indicated by the solid line with bidirectional arrows in the figure.

Statistical Analyses of Genetic Data
All sequences were checked and edited using Geneious. Multiple sequences were aligned using the Geneious Alignment (Identity 1.0/0.0), and all alignments were checked visually. MEGA11 was used to evaluate the composition of the nucleotides. DnaSP v.4.0 was used to calculate the genetic diversity indices such as the number of haplotypes (h), haplotype diversity (Hd) and nucleotide diversity (π).
A map of haplotypes distribution was constructed using ArcGIS, R 4.2 and Arlequin 3.5, and minimum spanning networks were constructed using PopART 1.7. Hierarchical analysis of molecular variance (AMOVA) of the different levels was used to further examine the population structure. The AMOVA based on three levels included population level . The significance of AMOVA was analyzed by 10,000 permutations. The fixation index (F ST ) was employed to assess the pairwise genetic divergence between different populations and the significance was obtained by 10,000 permutations.
Migrate-n 3.6 was conducted to estimate the mutation-scaled migration rate M (M = m / µ, where m was the historical migration rate among populations and µ was mutation per generation) and mutation-scaled population size θ (θ = Ne × µ, where Ne was historical effective population size). The runs were recorded every 100 steps for a total of 1,000,000 long-chain Markov chain Monte Carlo (MCMC) steps and a burn-in of 1000. To improve the efficiency of the MCMC search, a static heating scheme with four different temperatures (1.0, 1.5, 3.0 and 1,000,000.0) was used. We inspected histograms of estimated θ and M posterior values to assess convergence. We calculated historical gene flow (Nm) by using the equation Nm = θ × M (when using mitochondrial gene).

Population Genetic Diversity
A total of 253 COI sequences (538 bp, OP989704-956) and 95 D-Loop sequences (611 bp-612 bp, ON504951-995) were obtained from 15 populations of horn-eyed ghost crab. For COI sequences, the average contents of A, T, G and C were 26.6%, 34.6%, 16.8% and 22.0%, respectively. A total of 50 polymorphic sites were detected, including 24 singleton variable sites and 26 parsimony informative sites. Haplotype diversity ranged from 0.80 to 0.98, and nucleotide diversity ranged from 0.0022 to 0.0054 (Table 2). When the 15 populations were considered as a metapopulation, the haplotype diversity was 0.87, and the nucleotide diversity was 0.0036 (Table 2). The average contents of A, T, G and C were 42.7%, 32.6%, 9.6% and 15.0%, respectively, according to the results based on D-Loop sequences. A total of 116 polymorphic sites were found, including 89 parsimony informative sites and 27 singleton variable sites. The genetic diversity parameters showed that nucleotide diversity ranged from 0.0192 to 0.0379, and haplotype diversity ranged from 0.90 to 1.00 (Table 2). When the 15 populations were considered as a whole, the nucleotide diversity was 0.0279, and the haplotype diversity was 1.00. In conclusion, both genetic markers demonstrated high genetic diversity in horn-eyed ghost crab populations.

Haplotype Analysis
The COI sequences identified 62 haplotypes (Hap 1-62), including 24 shared haplotypes and 38 exclusive haplotypes. The frequencies of 62 haplotypes varied widely; the combined frequencies of Hap 1 and Hap 2 were as high as 49.40%, with the highest frequency of Hap 2 (26.48%), followed by Hap 1 (22.92%), while the remaining haplotypes were less frequent. Two haplotypes (Hap 1-2) were shared by 15 populations (Figure 2), while common haplotypes (Hap 1-2, Hap 7, Hap 10 and Hap 13) also existed in populations that were far apart (up to 1000 km), according to the geographic distribution of haplotypes ( Figure 2). The haplotype network also showed that Hap 1 and Hap 2 were shared by 15 populations, Hap 13 was shared by 11 populations. In general, the haplotype network showed a shallow double star-shaped structure ( Figure S1). sive haplotypes. The frequencies of Hap 25, Hap 29, Hap 47 and Hap 72 were all 2.11%, and the rest of haplotypes were found only once. According to the geographic distribution of haplotypes, Hap 29 was shared by two populations (XM and DZG). There were few shared haplotypes between diverse populations and no dominant haplotypes ( Figure 2). The haplotype network also showed that only Hap 29 was shared by two populations. Additionally, the haplotype network presented a bush-like shape ( Figure S2).

Population Genetic Structure
Based on COI sequences, at the region level, −0.22% of the variation was found between regions, with −0.18% of the variance found within regions among populations and 100.40% within populations. At the ecoregion level, most of the total variation (100.04%) was accounted for by differentiation within populations, with a further −0.80% accounting for variation within ecoregions among populations, and the remainder (0.76%) partitioned among ecoregions (Table 3). At the population level, −0.29% of the total variance  (Figure 2). The haplotype network also showed that only Hap 29 was shared by two populations. Additionally, the haplotype network presented a bush-like shape ( Figure S2).

Population Genetic Structure
Based on COI sequences, at the region level, −0.22% of the variation was found between regions, with −0.18% of the variance found within regions among populations and 100.40% within populations. At the ecoregion level, most of the total variation (100.04%) was accounted for by differentiation within populations, with a further −0.80% accounting for variation within ecoregions among populations, and the remainder (0.76%) partitioned among ecoregions (Table 3). At the population level, −0.29% of the total variance was found among populations, whilst 100.29% of the variance was found within populations. The AMOVA statistical tests were not significant in all three levels (p > 0.05). The pairwise population F ST showed that the genetic differences between populations ranged from −0.0456 to 0.0602. All pairwise F ST values were not statistically significant (p > 0.05; Table S3).
Based on the D-Loop sequences, at the region level, −0.86% of the variation was found between regions, with −0.44% of the variance found within regions among populations and 101.29% within populations. At the ecoregion level, most of the total variation (101.27%) was accounted for by differentiation within populations, with a further 0.04% accounting for variation within ecoregions among populations, and the remainder (−1.31%) partitioned among ecoregions. At the population level, −0.82% of the total variance was found among populations, whilst 100.82% of the variance was found within populations (Table 3). AMOVA statistical tests in all three levels were not significant (Table 3). Furthermore, the F ST values between populations were distributed between -0.1128 and 0.2675, the ge-netic differentiation between the remaining populations showed insignificant differences (p > 0.05), except for the significant differences between MZD and FYLD, and between MZD and HA (p < 0.05; Table S4). Two genetic markers revealed a high level of genetic homogeneity among the 15 populations of horn-eyed ghost crab and there was no significant genetic structure.

Migration and Connectivity
Based on COI sequences, the estimated historical gene flow ranged Based on COI sequences, the estimated migration rate ranged from 134.7 (QA-FYLD) to 390.1 (MZD-QA), while the estimated migration rate varied from 159.3 (PT-SYH) to 284.1 (SYH-NJLD) based on D-Loop sequences. The analysis consequences of two markers exhibited the high migration rates among populations. The COI sequences further showed the asymmetric migration rate with the higher migration rate from south to north than from north to south. Meanwhile, the migration rate from SYH to most other populations was higher than the migration rate from that of most other populations to SYH based on two markers (Figure 3). In addition, Migrate-n also revealed that the effective population sizes of each population based on both markers were relatively close (Table S5).
284.1 (SYH-NJLD) based on D-Loop sequences. The analysis consequences of two markers exhibited the high migration rates among populations. The COI sequences further showed the asymmetric migration rate with the higher migration rate from south to north than from north to south. Meanwhile, the migration rate from SYH to most other populations was higher than the migration rate from that of most other populations to SYH based on two markers (Figure 3). In addition, Migrate-n also revealed that the effective population sizes of each population based on both markers were relatively close (Table S5).

Discussion
In this study, 15 populations of horn-eyed ghost crab were collected in or near 15 MPAs in the East and South China Seas to analyze connectivity among populations based on COI and D-Loop. The results revealed that the high genetic diversity and connectivity among populations, and a high historical gene flow and migration rate among 15 populations were supported by two markers. Additionally, the COI sequences further showed the asymmetric migration rate with a higher migration rate from south to north than from

Discussion
In this study, 15 populations of horn-eyed ghost crab were collected in or near 15 MPAs in the East and South China Seas to analyze connectivity among populations based on COI and D-Loop. The results revealed that the high genetic diversity and connectivity among populations, and a high historical gene flow and migration rate among 15 populations were supported by two markers. Additionally, the COI sequences further showed the asymmetric migration rate with a higher migration rate from south to north than from north to south. The results of population structure and connectivity could reflect the connectivities between the marine protected areas in the two seas. These results could offer information for the management of the marine protected areas in the study areas.

Differences of Genetic Diversity in Populations
High haplotype diversity (Hd > 0.5) and low nucleotide diversity (π < 0.005) were observed based on COI sequences when the 15 populations were considered as a whole. This low nucleotide diversity and high haplotype diversity were in line with the genetic diversity of COI sequences from other marine organisms, particularly marine crustacean species such as Pachygrapsus crassipes [28] and Portunus trituberculatus [29]. D-Loop, on the other hand, showed high haplotype diversity (Hd > 0.5) and high nucleotide diversity (π > 0.005). In contrast, the higher genetic diversity was revealed by D-Loop marker, which may be related to the mutation rate of the fragment [30]. D-Loop is the high mutation region, located on either side of the central conserved region, which evolves two to five times faster than mitochondrial protein-coding genes [31], and thus may be more sensitive in genetic diversity analysis than COI marker [32]. This disagreement was often seen in studies of multiple molecular markers [31].
Because the sampling sites of horn-eyed ghost crab were located in or near MPAs, the results of genetic diversity may provide some reference for the management effectiveness of MPAs. Species and populations are adversely affected by anthropogenic activities, such as habitat modification, which are a global phenomenon [33]. These negative impacts eventually would reduce the genetic diversity [34]. The management effectiveness of a protected area is generally considered to be closely related to its conservation effectiveness, and a good protection management is also expected to have a higher level of biodiversity [35]. In this study, we also found that the marine protected areas with high genetic diversity also had a higher management effectiveness. For example, the high genetic diversity in NJLD and MA was observed, for which a high management effectiveness in both protected areas was reported [36]. FYLD had the low genetic diversity based on two markers, and it also had a lower score of management effectiveness in this protected area [37]. Therefore, it might support that the level of population genetic diversity is related to the management effectiveness of the marine protected areas. However, more tests are required to explore the relationship between genetic diversity and MPA management effectiveness.

Connectivity Differences among Populations
The spatial distribution of haplotypes and the haplotype network based on COI sequences showed that the most common Hap 1 and Hap 2, accounting for almost 49.40% of samples, were found in all 15 populations. Shared haplotypes may be related with the large population sizes and high rates of migration [38]. Among the COI haplotypes, shared haplotypes existed between distant populations (up to 1000 km), which indicated that dispersal of this species may occur over long distances. In contrast, the existence of lower-frequency haplotypes was related with large population sizes, and this is because these haplotypes were not eliminated by the selection when in a large stable population [39]. Whereas the spatial distribution of haplotypes and the haplotype network from D-Loop revealed few shared haplotypes among populations, it may not represent low connectivity among populations, but be due to DNA superdiversity [40,41]. This result may be explained by the combination of small samples and high mutation rates in D-loop sequences [41]. Therefore, results from COI might provide more reliable results on the spatial distribution and connectivity.
In this research, the high level of connectivity among 15 populations, between ECS populations and SCS populations and among three ecoregion populations of horn-eyed ghost crab was observed based on the results of AMOVA and F ST for two markers, which was also consistent with the findings of Collichthys lucidus [42], Larimichthys crocea [43], Pampus chinensis [44] and Thamnaconus hypargyreus [17]. The high connectivity among populations may be due to the long planktonic larval period of 34-42 days for horn-eyed ghost crab [45], which could allow distant geographical distances between populations to be overcome and produce genetic homogeneity [46]. On the other hand, ocean currents may promote connectivity among populations [42]. During the planktonic larval stage of horn-eyed ghost crab in summer, the current from the surface to the bottom is almost consistent with the northeast flow parallel to the coast in the coastal area from Xiangshan to Pearl River Estuary [47], which may result in high connectivity among populations. Meanwhile, this may account for the higher migration rate from south to north revealed by the COI marker in this study. It was also in line with the results of an analysis of seven mangrove species on the southern coast of China [48]. Nevertheless, the directional difference was not observed in the migration rate results based on D-Loop sequences, which may be due to the fact that the higher mutation rate of the D-Loop marker was not conducive to the prediction of gene flow or less number of D-Loop sequences [49]. In addition, Migrate-n also revealed a large population size in each population of horn-eyed ghost crab. Such population may be insensitive to the loss and reorganization of variation by genetic drift, and could maintain the ancestral genetic information [50], leading to high genetic connectivity among populations.

Conservation and Management Implications for MPA
Although it is not clear whether marine protected areas in the East and South China Seas could form a network, the results of this study may provide some suggestions for the construction of such a network due to the sampling sites of horn-eyed ghost crab being located in or near protected areas. The higher migration rate from south to north related to the current in summer was revealed based on COI sequences in this research. Consequently, the hydrodynamic characteristics of the East and South China Seas should be considered in constructing a MPA network. In the near-shore areas of the East and South China Seas, the Guangdong coastal current (GDCC) and the coastal currents of Zhejiang and Fujian (MZCC) are the main currents [27]. In summer, the direction of MZCC and GDCC is northward along the parallel shoreline, but in winter, the MZCC and GDCC flow in the southwesterly direction [47]. These season-changed currents should be included in the design of the MPA network, because organisms that reproduce in different seasons may have completely different connectivity patterns [51], especially those that reproduce in only one season.
Many protected species such as Acropora solitarvensis, Epinephelus akaara and Epinephelus bruneus spawn during summer [52]. The summer fishing moratorium implemented by the Chinese government in the East and South China Seas [53,54] contributes to the protection of larvae and may complement the MPA network [55]. This has also been shown in previous studies. A study of fish resources in Daya Bay showed that the stock density, species number, biodiversity and evenness index increased after the summer fishing moratorium period, indicating that the structure of the ecological community improved [56]. Additionally, in a study in the coastal ocean of Ningbo, the SCS showed that species richness, abundance and biomass of macrobenthos communities increased significantly during the summer fishing moratorium period, which implied this policy could facilitate the resilience of microbenthic communities [57]. However, focusing solely on the spawning patterns of these species for reserve design will decrease protection of fall-winter spawners when the direction of the gyres and the location of upstream larval sources reverses [51]. For example, at least some important species such as Anguilla japonica and Penaeus japonicus also spawn exclusively during fall-winter [52]. So, flexible management strategies could be carried out during the winter for the conservation of these species. For example, seasonal closures or fishing intensity restrictions could be assigned to their spawning sites and core areas of activity [58]. Moreover, seasonal protection corridors could be set up in areas with high connectivity during the planktonic larval period, and these areas should be a priority for new protected areas in the future.

Conclusions
This study investigated the genetic connectivity among the populations of Ocypode ceratophthalmus in the East and South China Seas based on wide geographic sampling and two genetic markers. The results showed that the horn-eyed ghost crab had high genetic diversity. The AMOVA analysis revealed no significant genetic structure and high genetic connectivity among the populations. All the F ST values obtained from COI sequences were inapparent. The F ST values based on D-Loop sequences were only significant between MZD and FYLD as well as between MZD and HA. Migrate-n revealed the high historical gene flow and connectivity among 15 populations based on two markers. The COI sequences further showed asymmetric connectivity with higher connectivity from south to north than from north to south. The outcomes from this study provide suggestions for the construction of marine protected areas.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/biology12030437/s1, Table S1: Current status of genetic connectivity between populations in the East China Sea and populations in the South China Sea; Table S2: The information of carapace width, carapace length, abdomen width, abdomen length, weight and gender for the horn-eyed ghost crab; Table S3: Pairwise F ST values between populations based on COI sequences; Table S4: Pairwise F ST values between populations based on D-Loop sequences; Table S5: The effective population sizes for each population based on COI and D-Loop sequences; Figure S1: The minimum spanning network based on COI sequences; Figure S2: The minimum spanning network based on D-Loop sequences.
Author Contributions: Conceptualization, C.Z. and F.Z.; methodology, C.Z., F.Z. and J.L.; data analysis, F.Z., Y.L. and Z.W. writing, F.Z. and C.Z.; funding acquisition, C.Z. and L.C. All authors have read and agreed to the published version of the manuscript.