Fine-Scale Genetic Structure of Curculio chinensis (Coleoptera: Curculionidae) Based on Mitochondrial COI: The Role of Host Specificity and Spatial Distance

Simple Summary Curculio chinensis is a phytophagous pest that feeds on oil-tea Camellia in South and Southwest China. This pest is highly dependent on its hosts and habitats. The genetic basis of this pest in various hosts, which could enrich our understanding of whether host-specificity existed and how the population is structured, is poorly explored. This study aimed to evaluate the genetic diversity, genetic differentiation, and phylogenetic structure of C. chinensis in 2 major host species, Camellia meiocarpa Hu and Camellia oleifera Abel, using 1083 mitochondrial COI. Structural differentiation was observed among populations in monoculture plantations of Camellia meiocarpa and Camellia oleifera. The mean genetic distance between Haplogroup 1 and Haplogroup 2 was significantly lower than that between C. chinensis and its related species. Two haplogroups have recently undergone a demographic expansion, and a significant asymmetrical effective migration was observed from C. chinensis populations in Cam. meiocarpa to populations in Cam. oleifera in comparison to migrating back, which is likely due to the increased cultivation of oil-tea Camellia in Jiangxi. Our findings can serve as a guide for future genomic research to improve prediction and facilitate control strategies for C. chinensis. Abstract The Camellia weevil, Curculio chinensis (Chevrolat, 1978), is a dominant oligophagous pest that bores into the fruit of oil-tea Camellia. Genetic differentiation among populations in various hosts can easily occur, which hinders research on pest management. In this study, the genetic structure, genetic diversity, and phylogenetic structure of local C. chinensis populations were examined using 147 individuals (from 6 localities in Jiangxi), based on 2 mitochondrial COI markers. Results indicated that the C. chinensis population in Jiangxi exhibits a high haplotype diversity, especially for the populations from Cam. meiocarpa plantations. Structural differentiation was observed between Haplogroup 1 (73 individuals from Ganzhou, Jian, and Pingxiang) in the monoculture plantations of Cam. meiocarpa and Haplogroup 2 (75 individuals from Pingxiang and Jiujiang) in Cam. oleifera. Two haplogroups have recently undergone a demographic expansion, and Haplogroup 1 has shown a higher number of effective migrants than Haplogroup 2. This suggests that C. chinensis has been spreading from Cam. meiocarpa plantations to other oil-tea Camellia, such as Cam. oleifera. The increased cultivation of oil-tea Camellia in Jiangxi has contributed to a unique genetic structure within the C. chinensis population. This has, in turn, expanded the distribution of C. chinensis and increased migration between populations.


Introduction
Curculio chinensis (Chevrolat, 1978) (Coleoptera: Curculionidae), one of the dangerous forest pests in China, was first reported in the 1960s.It is now distributed in oil-tea Camellia Insects 2024, 15 plantations in South and Southwest China, but it has not been found abroad [1].C. chinensis predominantly damages the fruits and seeds of Camellia species [2].Young tea leaves and shoots can be damaged by adults' piercing-sucking, which can then easily lead to infection by anthracnose.The eggs and larvae were deposited in fruits of Camellia species, leading to spoiled seeds or even complete fruit failure [3][4][5].
China is the origin and distribution center of Camellia spp.The vast majority of oil-tea Camellia, a unique woody-oil species, is located in China [6].C. chinensis is a notorious oligophagous pest with weak flying ability in the primary oil-tea Camellia plantation in China.The investigation of the population showed that the occurrence of C. chinensis is closely related to the host species and the climate of plantations [7,8].C. chinensis and C. spp.have shown significant variation related to host isolation (Cam.oleifera and Cam.sinensis) using mitochondrial COI.For C. chinensis, the tea saponin content varies among host species and affects the composition of gut microbiota.The current study focuses on the adaptation of gut bacteria to the phytochemical resistance of oil-tea Camellia [9,10].For other species of Coleoptera with limited flight ability, natural geographical factors and host plants often play crucial roles in influencing the genetic structure and gene flow among populations.C. camellia, a closely related species to C. chinensis, is influenced not only by geographical isolation but also by host plants [11], and the migration pattern of C. camellia imposed significant effects on the geographic configuration of the coevolution level between C. camellia and its host, Cam.Japonica [12,13].Various levels of geographic differentiation and host specificity have been detected among interspecific and intraspecific species of Camellia using mitochondrial genes [14,15].Therefore, in addition to gut bacteria adaptation, the genetic structure characteristics and gene flow patterns of C. chinensis are speculated to vary due to diverse host species and geographical isolation.This makes C. chinensis an ideal subject for revealing adaptive evolution.
Mitochondrial genes, such as COI, have the advantages of stable composition and conservative gene arrangement and have been extensively utilized in population genetics studies of Curculionidae [11,14,15].Although mitochondrial genes are maternally inherited, their small effective size, interspecific diversity, and relatively high mutation rates make them especially valuable for providing crucial insights into population dynamics, predicting potential migration events, and gaining a better understanding of the evolutionary potential related to hosts and habitats.But shorter mitochondrial gene sequences often lack sufficient variable sites to accurately resolve genetic diversity and population differentiation.For example, the results based on mitochondrial COI (544 bp) revealed a unique haplotype in the Jiangxi population, which was significantly different from the other populations in China [16,17].However, mitochondrial ATP synthase-based results showed higher diversity and more private haplotypes in the Jiangxi population compared to the other populations [18].On the other hand, the aforementioned findings also suggest that there may be local adaptive differentiation in the Jiangxi population.
Jiangxi Province is one of the major cultivation regions for oil-tea Camellia in China.The various Camellia species in Jiangxi provide diverse habitats for the occurrence of C. chinensis, making Jiangxi Province an ideal area to study the local adaptability of this pest.In the present study, our main objective was to test the host specificity of C. chinensis and examine its population dynamics while taking into account the genetic variation of this pest collected from two major species, Camellia meiocarpa Hu and Camellia oleifera Abel, within Jiangxi plantations.First, the population diversity, genetic differentiation, and phylogeographical structure of C. chinensis were assessed using variations in mitochondrial COI sequences.Subsequently, population dynamics were estimated from recurrent migration between genetic haplogroups and the demographic history of populations.The correlation between genetic differentiation and host plants, as well as geographical isolation, was evaluated using an analysis of molecular variance and a matrix correspondence test.This will infer the adaptive differentiation of C. chinensis to hosts and geographical factors and will enable better prediction and management of C. chinensis in Jiangxi.

Sample Collection
During 2022 and 2023, specimens of C. chinensis were selectively collected from six localities in Jiangxi (Figure 1) using a Z-sampling method (six plots from each locality).These localities were infested by this weevil over the years and covered species-distributed regions, including Cam. meiocarpa plantations, Cam.oleifera plantations, and areas with a mix of Cam.meiocarpa/Cam.oleifera plantations (Table 1; Figure 1).A total of 1000 specimens were preserved in absolute alcohol at −20 • C until they were identified and used for DNA extraction.They are now deposited at the Institute of Jiangxi Oil-tea Camellia, Jiujiang University, Jiujiang, China.A total of 1 adult (identified using morphological characters [1] and sequencing firstly in populations) and 21-24 mature larvae were chosen for 6 populations.

Sample Collection
During 2022 and 2023, specimens of C. chinensis were selectively collected from six localities in Jiangxi (Figure 1) using a Z-sampling method (six plots from each locality).These localities were infested by this weevil over the years and covered species-distributed regions, including Cam. meiocarpa plantations, Cam.oleifera plantations, and areas with a mix of Cam.meiocarpa/Cam.oleifera plantations (Table 1; Figure 1).A total of 1000 specimens were preserved in absolute alcohol at −20 °C until they were identified and used for DNA extraction.They are now deposited at the Institute of Jiangxi Oil-tea Camellia, Jiujiang University, Jiujiang, China.A total of 1 adult (identified using morphological characters [1] and sequencing firstly in populations) and 21-24 mature larvae were chosen for 6 populations.

DNA Extraction, Optimization of PCR Amplification
Genomic DNA was extracted from a leg of an adult or muscle tissue of larvae using the EsayPure ® Genomic DNA Kit (TransGen Biotech, Beijing, China).The remaining specimens were deposited.DNA concentration was measured using an ND-1000 spectrophotometer (Bio-Rad, Hercules, CA, USA) and diluted to 20 ng/µL with ddH 2 O.

Sequencing and Population Genetic Structure Analysis 2.3.1. Genetic Diversity Analyses
DNA sequences were derived from 147 individuals selected from 6 locations representing 6 populations.A total of 2 fragments of 544 bp and 825 bp were obtained from each sample using 2 pairs of primers.The raw sequences were then proofread using Geneious Primer v. 11.0.14.1 [19].Sequences of target genes were confirmed by aligning the resulting sequences to the mitochondrial genome MZ417388 [20] in the National Center for Biotechnology Information (NCBI) database using the BLAST tool.The aligned sequences were then joined together using MEGA v. 7.0 [21].Mitochondrial COI sequences from 147 samples, each 1083 bp long, were used for genetic diversity analyses.The number of haplotypes (Nh), haplotype diversity (Hd), nucleotide diversity (π), the number of polymorphic sites (S), and the mean number of nucleotide difference (k) were estimated by DnaSP v. 6.12.03 [22].

Phylogenetic Analyses and Genetic Structure
Phylogenetic trees were reconstructed using 147 mitochondrial COI sequences and 5 sequences from GenBank (MF409663, MF409669, MF409675, MF409681, and MF409682).Maximum likelihood (ML) and Bayesian inference (BI) were performed using IQ-TREE and MrBayes in PhyloSuite v.1.2.3 [23,24] after finding the GTR+ F +G4 model based on the Akaike information criterion (AIC) in the ModelFinder.Two sequences (OR976214, OR976215) of C. davidi were set as the outgroup.A total of 5000 ultrafast bootstraps and 1000 replicates for the SH-aLRT branch test were run in ML [25].MrBayes uses Markov chain Monte Carlo (MCMC) to perform Bayesian inference of phylogeny [26].A total of 2 separate analyses, 4 MCMC chains, 2 × 10 6 generations (the removal of the first 25% of samples), and 1000 sampling statistics were set for running until the average standard deviations of split frequencies were below 0.01 and the effective sample size (ESS) was above 100 [27].The tree was visualized and edited with an online tool called the Interactive Tree of Life (iTOL) v. 5 [28].
Haplotype networks were analyzed and edited using Network v. 10 to infer relationships between haplotypes from 147 samples.An admixture model was chosen in the Bayesian inference-based software STRUCTURE v. 2.3.4 [29,30] to detect clusters of multi-site haplotypes in populations using correlated allele frequencies.Inferred clusters (K) 1 to 6 were set with 20 independent runs of each and 1 × 10 6 (MCMC) repetitions (a 100,000 repetition burn-in period in each run).CLUMPP v1.1.2[31] and DISTRUCT v. 1.1 [32] were used to permute the cluster labels across runs and display the genetic structure results after determining the most likely number of genetic clusters (K) based on the result of the ad hoc statistic (∆K) using STRUCTURE HARVESTER [33,34].Sequence divergences among haplogroups and other COI sequences from 11 mitogenomes (MT560591, MK654677, KX087269, MG728095, NC045101, NC027577, KX087330, MT232762, MW023069, NC022680, NC051548) in the same family downloaded from GenBank were calculated with MEGA v. 7.0 [21] using Kimura two-parameter distances.

Isolation by Hosts and Distance Analysis
The analysis of molecular variance (AMOVA) was performed to partition genetic variations among haplogroups and within haplogroups, which were divided based on the results of haplotype relationship analyses in Arlequin version 3.5.2[35] with 10,000 permutations, as well as the fixation indices.The genetic differentiation between populations of different hosts in the same sample site and AMOVA among haplogroups in different sample sites were also conducted to infer the effects of the host.
Within Jiangxi plantations, the anticipated high levels of gene flow at a small spatial scale should somewhat restrict the geographic differentiation of the weevil trait.We thus predicted that the difference in genetic distances of populations would not be corrected with increasing geographic distance between weevil populations if the extent of gene flow affected the degrees of local adaptation.Mantel tests, in which the effects of geographic distance between localities were controlled, were conducted to examine the correlation between geographic distance and interpopulation genetic distance using IBDWS v. 3.23 [36,37] The pairs of geographic distance/genetic distance between populations were generated using Geographic Distance Matrix Generator v. 1.2.3 and MEGA v. 7.0.

Population Dynamics
Neutrality tests (Tajima's D and Fu's Fs statistics) of 6 populations were calculated to test for evidence of recent population expansion using DnaSP v. 6.12.03 [38,39].To infer the gene flow between haplogroups, Bayesian inference of population genetic parameters was conducted using the program Migrate-n 3.1.6.The DNA sequence model and the full model were used to estimate the migration rate (M) and the mutation-scaled population size (θ) [40].After the initial run with F ST , θ and M were used for the remaining three runs.One hundred million MCMC steps were taken, with the first 2 × 10 6 steps discarded, and static heating schemes with 4 chains were sampled to estimate θ and M in the Bayesian search strategy.The effective number of migrants of each population per generation (Nem) can be calculated as 4θM, and the effective population size (N) of each population can be calibrated by the mutation rate (N = θ/2µ; µ, mutation rate per site per generation is assumed to be 1.8 × 10 −8 [13]).N and Nem can reveal the effective population size and population interactions of C. chinensis, which can help us recognize the focal population and immigration of this hidden pest for monitoring and control.

Genetic Diversity
A total of 36 haplotypes were identified from 147 COI sequences in six populations of C. chinensis (OR976178-OR976213).High genetic diversity was detected in Jiangxi populations.The number of haplotypes in various populations ranged from 4 to 9, with haplotype diversity ranging from 0.545 to 0.777 and nucleotide diversity ranging from 0.00075 to 0.02496.The number of polymorphic sites (S) and the mean number of nucleotide difference (k) were 0.81 to 27.033 and 3 to 75, respectively (Table 1).Five common haplotypes were found.Populations in Cam.meiocarpa plantations and Cam.oleifera plantations have the most common haplotypes, H8 and H15, respectively (Table A1).There were clear differences in the number of polymorphic sites (S) and the mean number of nucleotide differences (k) between C. chinensis populations from Cam. oleifera plantations and Cam.meiocarpa plantations.For populations in different plantations, the highest S (18) and k (1.8) were found in PS populations, while the lowest S (3) and k (0.633) were discovered in the JX population (Table 1).

Haplotype Relationship and Genetic Differentiation Analyses
Thirty-six haplotypes were used for the phylogenetic analysis.ML and Bayesian results showed that haplotypes of C. chinensis populations were separated into two haplogroups with high support values (Figure 2).Haplogroup 1 mainly includes 23 haplotypes from GX, JS, and PS populations; the other haplotypes from JD, PL, and JX populations clustered into Haplogroup 2. Haplogroup 1 clustered with MF409663 in Clade 1, while Haplogroup 2 clustered with MF409682 in Clade 2 based on 544 bp in a previous study [17] (Figure 2A).The haplotypes of populations in Cam.oleifera plantations clustered closely, as well as the haplotypes of populations in Cam.meiocarpa plantations.The common haplotypes H8 and H15 were detected in the PL population (Figure 2B).populations.The number of haplotypes in various populations ranged from 4 to 9, with haplotype diversity ranging from 0.545 to 0.777 and nucleotide diversity ranging from 0.00075 to 0.02496.The number of polymorphic sites (S) and the mean number of nucleotide difference (k) were 0.81 to 27.033 and 3 to 75, respectively (Table 1).Five common haplotypes were found.Populations in Cam.meiocarpa plantations and Cam.oleifera plantations have the most common haplotypes, H8 and H15, respectively (Table A1).There were clear differences in the number of polymorphic sites (S) and the mean number of nucleotide differences (k) between C. chinensis populations from Cam. oleifera plantations and Cam.meiocarpa plantations.For populations in different plantations, the highest S (18) and k (1.8) were found in PS populations, while the lowest S (3) and k (0.633) were discovered in the JX population (Table 1).

Haplotype Relationship and Genetic Differentiation Analyses
Thirty-six haplotypes were used for the phylogenetic analysis.ML and Bayesian results showed that haplotypes of C. chinensis populations were separated into two haplogroups with high support values (Figure 2).Haplogroup 1 mainly includes 23 haplotypes from GX, JS, and PS populations; the other haplotypes from JD, PL, and JX populations clustered into Haplogroup 2. Haplogroup 1 clustered with MF409663 in Clade 1, while Haplogroup 2 clustered with MF409682 in Clade 2 based on 544 bp in a previous study [17] (Figure 2A).The haplotypes of populations in Cam.oleifera plantations clustered closely, as well as the haplotypes of populations in Cam.meiocarpa plantations.The common haplotypes H8 and H15 were detected in the PL population (Figure 2B).The results of haplotype networks and STRUCTURE corresponded to two haplogroups in the phylogenetic tree (Figure 3; Figure A1).H8 and H15 were common haplotypes shared by 49 and 39 individuals in different plantations.The others were private in one population, except for H7 in GX and JS, and H19 in JD, JX, and PL.The GX, JS, and PS populations have a similar genetic composition in Haplogroup 1, while the genetic composition of JD is similar to JX in Haplogroup 2. PL has the most complex genetic composition in both Haplogroup 1 and Haplogroup 2.
(1083 bp) are haplotypes from Jiangxi populations.The bootstrap values of ML and the posterior probability of BI are given (>90/0.9).
The results of haplotype networks and STRUCTURE corresponded to two haplogroups in the phylogenetic tree (Figure 3; Figure A1).H8 and H15 were common haplotypes shared by 49 and 39 individuals in different plantations.The others were private in one population, except for H7 in GX and JS, and H19 in JD, JX, and PL.The GX, JS, and PS populations have a similar genetic composition in Haplogroup 1, while the genetic composition of JD is similar to JX in Haplogroup 2. PL has the most complex genetic composition in both Haplogroup 1 and Haplogroup 2.

Isolation by Hosts and Distance Analysis
The genetic distances within the haplogroups of C. chinensis ranged from 0.0011 to 0.018, with a mean of 0.0013 for Haplogroup 1 and 0.0122 for Haplogroup 2. The genetic distances between Haplogroup 1 and Haplogroup 2 ranged from 0.0415 to 0.0593, and those between species within Curculio ranged from 0.1279 to 0.1992.The genetic distances between C. chinensis and related species within Curculionidae ranged from 0.1710 to 0.2392, with a mean of 0.2045 (Table A2).

Isolation by Hosts and Distance Analysis
The genetic distances within the haplogroups of C. chinensis ranged from 0.0011 to 0.018, with a mean of 0.0013 for Haplogroup 1 and 0.0122 for Haplogroup 2. The genetic distances between Haplogroup 1 and Haplogroup 2 ranged from 0.0415 to 0.0593, and those between species within Curculio ranged from 0.1279 to 0.1992.The genetic distances between C. chinensis and related species within Curculionidae ranged from 0.1710 to 0.2392, with a mean of 0.2045 (Table A2).

Population Dynamics
The results of Tajima's D and Fu's Fs indicate a recent expansion for JS, PS, JD, and JX populations (Table 1).Significant population expansion was supported by both Tajima's D and Fu's Fs in the JS and PS populations (Tajima's D < 0, p < 0.05; Fu's Fs < 0, p < 0.05).The JD and JX populations have also undergone a population expansion, but the expansion was not significant (Tajima's D < 0, P > 0.05; Fu's Fs < 0, p > 0.05).The significantly negative values of Tajima's D and Fu's Fs suggest a recent demographic expansion for Haplogroup 1 (Tajima's D = −2.417,p < 0.01; Fu's Fs = −22.941,p < 0.01).Populations in Haplogroup 2 showed evidence of a recent demographic expansion, as indicated by the negative value of Fu's Fs (Fu's Fs: −0.950, p > 0.100).

Genetic Diversity of C. chinensis in Different Hosts
C. chinensis populations in Jiangxi have substantially high genetic diversity.The results were different from those reported in the Jiangxi population [17] but similar to the results in C. camellia [11].This can be attributed to multiple mutation sites from 6 populations in different hosts based on 1083 bp mitochondrial sequences (70% complete mitochondrial COI [20]) in this study, while 1 conserved site from only 1 population in a natural and isolated Cam.Oleifera plantation near the Wuyi Mountains based on 544 bp mitochondrial sequences in the previous study [17].The genetic diversity (S and k) of GX, JS, and PS populations in Cam.oleifera plantations was lower than that of JX and JD populations in Cam.meiocarpa plantations.Additionally, significant differences in S and k were detected in PL and PS populations, which have similar geographical locations (near the Luoxiao Mountains) but different hosts.Unique private haplotypes were found in populations collected from Cam. oleifera plantations (11,30.5% of all haplotypes) and Cam.meiocarpa plantations (20, 55.5% of all haplotypes) (Table A1; Table 1).This suggests significant genetic variation among samples from different hosts.

Population Genetic Structure and Nucleotide Divergences
Two haplogroups were detected in samples from Jiangxi, which were from Cam. meiocarpa plantations and Cam.oleifera plantations, respectively.The results of haplotype phylogenetic relationships, networks, and STRUCTURE can elucidate the relationship between two haplogroups.In this study, 36 haplotypes were observed, but only H8 (PL population) was shared by two haplogroups.The PL population from the plantation includes Cam.Oleifera, and Cam.meiocarpa also has the same genetic components as the two haplogroups from Cam. oleifera and Cam.meiocarpa plantations, respectively (Figure 2B).The AMOVA analysis revealed a high level of genetic differentiation among the haplogroups with two hosts and populations in different haplogroups.
The mean genetic distances within haplogroups did not exhibit any significant difference (A/B: t-test: t = 1.911, d.f.= 4, p = 0.196), which was notably lower than the distances between haplogroups (with a mean of 0.0529) (C/A: t-test: t = 18.823, d.f.= 10, p < 0.001; C/B: t-test: t = 7.137, d.f.= 10, p < 0.001).The mean genetic distance between Haplogroup 1 and Haplogroup 2 was significantly less than that between C. chinensis and its related species (C/D: t-test: t = −9.387,d.f.= 12, p < 0.001; C/E: t-test: t = −36.402,d.f.= 25, p < 0.001) (Figure 4) [41].The genetic distance between the two haplogroups of C. chinensis was greater than the intraspecific distance of other species in Jiangxi.This is similar to the intraspecific genetic variation found in one-quarter of the species from BOLD (>0.03) [42,43].The above situation suggests a clear differentiation between the two haplogroups, which may be related to host shifts.
AMOVA analysis revealed a high level of genetic differentiation among the haplogroups with two hosts and populations in different haplogroups.
The mean genetic distances within haplogroups did not exhibit any significant difference (A/B: t-test: t = 1.911, d.f.= 4, p = 0.196), which was notably lower than the distances between haplogroups (with a mean of 0.0529) (C/A: t-test: t = 18.823, d.f.= 10, p < 0.001; C/B: t-test: t = 7.137, d.f.= 10, p < 0.001).The mean genetic distance between Haplogroup 1 and Haplogroup 2 was significantly less than that between C. chinensis and its related species (C/D: t-test: t = −9.387,d.f.= 12, p < 0.001; C/E: t-test: t = −36.402,d.f.= 25, p < 0.001) (Figure 4) [41].The genetic distance between the two haplogroups of C. chinensis was greater than the intraspecific distance of other species in Jiangxi.This is similar to the intraspecific genetic variation found in one-quarter of the species from BOLD (>0.03) [42,43].The above situation suggests a clear differentiation between the two haplogroups, which may be related to host shifts.A2. *** and ns denote significant difference and no significant difference, respectively.

Fine-Scale Population Dynamics and Potential Effects on Genetic Differentiation
Host shifts can lead to directional gene flow and result in genetic differentiation or even speciation.So, host plant specialization is a critical mechanism for the diversification of phytophagous insects.Insect species with poor migration ability often exhibit genetic differentiation among populations that feed on different host species.This is exemplified by the genetic divergence observed between C. chinensis populations in this study.Haplotype phylogenetic relationships based on 1083 bp COI showed that populations collected from the same host were clustered together.This may be associated with a potential microbial contribution to the chemical adaptability of tea saponin of Camellia species [9,10].Some previous studies have also indicated host-associated divergence in Curculio species [13,15,44].

Fine-Scale Population Dynamics and Potential Effects on Genetic Differentiation
Host shifts can lead to directional gene flow and result in genetic differentiation or even speciation.So, host plant specialization is a critical mechanism for the diversification of phytophagous insects.Insect species with poor migration ability often exhibit genetic differentiation among populations that feed on different host species.This is exemplified by the genetic divergence observed between C. chinensis populations in this study.Haplotype phylogenetic relationships based on 1083 bp COI showed that populations collected from the same host were clustered together.This may be associated with a potential microbial contribution to the chemical adaptability of tea saponin of Camellia species [9,10].Some previous studies have also indicated host-associated divergence in Curculio species [13,15,44].
Our Bayesian estimation of population genetic parameters revealed that GX, JS, and PS C. chinensis populations (N = 5.18 × 10 5 , Nem1→2 = 4.96), which were clustered in Haplogroup 1, were potential source populations, On the other hand, JD, PL, and JX C. chinensis populations in Haplogroup 2 (N = 3.21 × 10 5 , Nem2→1 = 1.00) were identified as sink populations (Table A3).Significant asymmetrical effective migrants (Nem) between Haplogroup 1 and Haplogroup 2 were discovered through non-overlapping 95% confidence intervals.These close phylogenetic relationships and asymmetrical effective migration suggest that C. chinensis populations might have shifted their range from Cam. meiocarpa plantations to Cam. oleifera plantations.No significant geographical differentiation was found between six populations with a similar subtropical monsoon climate (no significant difference in annual mean temperature (19.17-18.35)),but a complex genetic component in the PL population from Cam. oleifera/Cam.meiocarpa plantations and Bayesian estimation of population-effective migrants suggest that genetic differentiation between local C. chinensis populations has been caused by artificial cultivation, transformation, and host adaptability rather than physical barriers.
Genetic migration of many parasitic pests was affected by human plant breeding and commerce.Previous research has reported the effects of human activities on population divergence between Curculio beetles [45][46][47].As the cultivation area of Cam.Oleifera is increasing, C. chinensis is occurring across wider areas.Cam.meiocarpa and Cam.oleifera, the two most widely distributed species in Jiangxi, were promoted for planting from the 1950s and the 1990s, respectively.Many plantations that included these two species were retained during the low-yield transformation of oil-tea Camellia.C. chinensis can be transferred by the seeds during the larval stage and dispersed across different trees in the same plantations at the adult stage.So, the recent demographic expansion was found in JS, PS, JD, and JX populations.Our results also showed that there were more migrants from Haplogroup 1 to Haplogroup 2 than from Haplogroup 2 to Haplogroup 1, and Haplogroup 1 has a closer relationship with MF409663 in Clade 1 from five other provinces in China than Haplogroup 2 [17].Additionally, the PL population was found to have two genetic components and haplotypes from two haplogroups.In this population, 25 individuals were collected from Cam. meiocarpa (8)/Cam.Oleifera (17) plantations and their haplotypes corresponded to H8/H15, H19, and H22-H27.A probable cause is that Cam.oleifera planting areas have been expanding to new plantations, and the low-yield transformation for several decades has led to an increased distribution of hosts and an increase in C. chinensis migrants between Cam. meiocarpa and Cam.oleifera plantations.However, our analyses failed to find key haplotypes or subgroups that play a significant role in connecting different haplogroups, which was caused by the limited number of samples from Cam. meiocarpa/Cam.oleifera plantations or haploid genetic markers.The high diversity and recent demographic expansion suggest that C. chinensis populations will continue to expand across Jiangxi plantations.Overwintering in the soil and the development of eggs/larvae within oil-tea Camellia fruit both contribute to avoiding insecticides and improving survival rates in new plantations.Given the genetic diversity and population dynamics of C. chinensis, it is crucial to prioritize population monitoring (surveys or molecular markerbased population genetic analysis) and implement control measures (such as Beauveria bassiana powder) for C. chinensis populations in Cam.meiocarpa plantations and low-yield transformation plantations.

Conclusions
High diversity and recent demographic expansion of C. chinensis were discovered in Jiangxi Cam.oleifera and Cam.meiocarpa plantations.Two haplogroups with significant genetic divergence were detected through haplotype phylogenetic relationships and networks.Haplogroup 1, collected from Cam. meiocarpa plantations, had more effective migrants compared to Haplogroup 2, which mainly originated from Cam. oleifera plantations.The PL population from Cam. oleifera/Cam.meiocarpa plantations included two genetic components and haplotypes from two haplogroups.These results can improve our understanding of the dispersal of C. chinensis across different host plants and reveal the effect of low-yield transformation on the genetic patterns of this pest.We should pay attention to monitoring and controlling C. chinensis populations in Cam.meiocarpa plantations and low-yield transformation plantations.Our future research would focus on evaluating the influence of hosts on the adaptability of C. chinensis with more sampling from Cam. meiocarpa/Cam.oleifera plantations based on genomic data, including differential expression of their key detoxification genes and the biological characteristics (such as generation time and feeding preference) of this pest on different hosts.

Figure 1 .
Figure 1.The distribution of six Curculio chinensis populations in Jiangxi, China.Population codes are listed in Table1.SimpleMappr was used to produce a distribution map based on the geographical coordinates in Table1.URL: http://www.simplemappr.net/#tabs=0(accessed on 25 December 2023).

Figure 3 .
Figure 3. Haplotype network and STRUCTURE analysis of C. chinensis populations based on concatenated mitochondrial COI.(A).Each circle represents a haplotype, and the sizes indicate the number of individuals.(B).The proportion of populations from two clusters inferred by STRUC-TURE analysis.An individual is represented by a vertical bar.

Figure 3 .
Figure 3. Haplotype network and STRUCTURE analysis of C. chinensis populations based on concatenated mitochondrial COI.(A).Each circle represents a haplotype, and the sizes indicate the number of individuals.(B).The proportion of populations from two clusters inferred by STRUCTURE analysis.An individual is represented by a vertical bar.

Figure 4 .
Figure 4. Comparison of genetic distances between different groupings.A: the mean genetic distances of GX, JS, and PS C. chinensis populations; B: the mean genetic distances of JD, PL, and JX C. chinensis populations; C: the mean genetic distances between Haplogroup 1 and Haplogroup 2; D: the mean genetic distances within Curculio; E: the mean genetic distances between C. chinensis and its related species; and F: the mean genetic distances between species within Curculionidae, except for C. chinensis.Genetic distances are shown in TableA2.*** and ns denote significant difference and no significant difference, respectively.

Figure 4 .
Figure 4. Comparison of genetic distances between different groupings.A: the mean genetic distances of GX, JS, and PS C. chinensis populations; B: the mean genetic distances of JD, PL, and JX C. chinensis populations; C: the mean genetic distances between Haplogroup 1 and Haplogroup 2; D: the mean genetic distances within Curculio; E: the mean genetic distances between C. chinensis and its related species; and F: the mean genetic distances between species within Curculionidae, except for C. chinensis.Genetic distances are shown in TableA2.*** and ns denote significant difference and no significant difference, respectively.

Figure A2 .
Figure A2.Scatter plots of the genetic distance isolation by geographical distance among six Curculio chinensis populations in Jiangxi based on mitochondrial COI.

Table 2 .
AMOVA results of six C. chinensis populations between two haplogroups inferred from haplotype relationship.