Global genetic diversity, lineage distribution, and Wolbachia infection of the alfalfa weevil Hypera postica (Coleoptera: Curculionidae)

Abstract The alfalfa weevil (Hypera postica) is a well‐known example of a worldwide‐distributed pest with high genetic variation. Based on the mitochondrial genes, the alfalfa weevil clusters into two main mitochondrial lineages. However, there is no clear picture of the global diversity and distribution of these lineages; neither the drivers of its diversification are known. However, it appears likely that historic demographic events including founder effects played a role. In addition, Wolbachia, a widespread intracellular parasite/symbiont, likely played an important role in the evolution of the species. Wolbachia infection so far was only detected in the Western lineage of H. postica with no information on the infecting strain, its frequency, and its consequences on the genetic diversity of the host. We here used a combination of mitochondrial and nuclear sequences of the host and sequence information on Wolbachia to document the distribution of strains and the degree of infection. The Eastern lineage has a higher genetic diversity and is found in the Mediterranean, the Middle East, Eastern Europe, and eastern America, whereas the less diverse Western lineage is found in Central Europe and the western America. Both lineages are infected with the same common strain of Wolbachia belonging to Supergroup B. Based on neutrality tests, selection tests, and the current distribution and diversification of Wolbachia in H. postica, we suggested the Wolbachia infection did not shape genetic diversity of the host. The introduced populations in the United States are generally genetically less diverse, which is in line with founder effects.

The alfalfa weevil is a widespread common pest of alfalfa Medicago sativa Linnaeus, 1753 (Fabaceae) (Sanaei, Seiedy, & de Castro, 2015;Summers, 1998). A recent study suggested that this species has a Palearctic origin, but has undergone human-mediated translocations to several parts of the world .
The invasion of H. postica to much of the Holarctic region is well documented with the first introductions to the America in 1904 (Titus, 1909), Japan in 1982 (Kuwata, Tokuda, Yamaguchi, & Yukawa, 2005), and Korea in 2002 (Hong & Kim, 2002). Historically, the alfalfa weevil populations in the America were categorized into three lineages which were called Eastern, Egyptian, and Western strains (Bundy, Smith, English, Sutton, & Hanson, 2005). Each strain was introduced to America via an independent route and time (Radcliffe & Flanders, 1998). While there is a lack of distinct morphological characters (Bland, 1984;Pienkowski, Hsieh, & Lecato, 1969;Sanaei, Seiedy, & Momtazi, 2015b), several fluctuating ecological traits were diagnosed to be strain specific including response to parasitoids and location of pupation (Coles & Day, 1977;Dewitt & Armbrust, 1972;Litsinger & Apple, 1973). However, the usefulness of these characters is questionable as strains with certain ecological characters were determined only based on their location and most of the mentioned characters were not applicable or showed huge variation and overlap among strains (Bundy et al., 2005;.
At the end of the 20th century, studies started using molecular markers to quantify strain divergence (Erney, Pruess, Danielson, & Powers, 1996;Hsaio, 1996). The results of mitochondrial analyses (parts of COI and CytB genes with a total length of 1,031 bp) showed that the Eastern and Egyptian strains were similar and together have an approximate nucleotide difference of 5% to the Western strain (Erney et al., 1996). While nuclear genes failed to recover any pattern of diversification (Böttger, Bundy, Oesterle, & Hanson, 2013; for more information on genes please refer to Appendices S1 and S2), several studies confirmed the strong mitochondrial gene divergence between Western and Eastern (including American Eastern and Egyptian strains) lineages (Böttger et al., 2013;Iwase, Nakahira, et al., 2015;Kuwata et al., 2005;. However, the genetic diversity underlying each lineage remains unknown. Several factors may have contributed to the mitochondrial lineage divergence and diversification in H. postica; these may include population history, adaptation to different hosts, or disrupted interpopulation gene flow (Iwase, Nakahira, et al., 2015;Iwase, Tani, et al., 2015;Sanaei, Seiedy, & Momtazi, 2015a;. One additional potentially important agent is the intracellular alpha-proteobacteria Wolbachia (Hertig, 1936;Werren, 1997). Several hundreds of strains of Wolbachia have been diagnosed so far within nematodes and arthropods with various effects on their hosts (Gerth & Bleidorn, 2017;Werren, Windsor, & Guo, 1995;Zug & Hammerstein, 2012). Recently, it has been estimated that Wolbachia infects 38.3% of all beetle species (Kajtoch & Kotásková, 2018). Wolbachia is transmitted to the host progeny via the cytoplasm of the eggs and is therefore maternally transferred to the next generation (Bourtzis, Dobson, Braig, & O'Neill, 1998;Branca, Vavre, Silvain, & Dupas, 2009;Hoffmann, 1988;Telschow, Hammerstein, & Werren, 2002). Wolbachia neutrality tests, selection tests, and the current distribution and diversification of Wolbachia in H. postica, we suggested the Wolbachia infection did not shape genetic diversity of the host. The introduced populations in the United States are generally genetically less diverse, which is in line with founder effects.
The American populations of the alfalfa weevil were one of the first hosts for which CI (Hsaio, 1996;Hsiao & Hsiao, 1985a, 1985b and potential sex ratio disorder (Hsiao & Hsiao, 1985b)

caused by
Wolbachia were diagnosed. The unsuccessful cross mating between uninfected Eastern and infected Western lineage individuals indicated a reproduction barrier caused by bidirectional CI (Hsiao & Hsiao, 1985b). A study of American populations further suggested that the Western lineage is naturally infected, whereas the Eastern lineage is naturally resistant to Wolbachia (Bundy et al., 2005;Hsaio, 1996). However, in Japan, an Eastern lineage sample was found to be infected with Wolbachia; yet, here none of the examined Western lineage samples were infected (Iwase, Tani, et al., 2015). Recently, it has also been shown that in the absence of Wolbachia infection in Japan, both lineages produce viable eggs (Iwase & Tani, 2016) providing further evidence that CI induced by Wolbachia did not result in complete reproductive isolation and hence speciation. However, we still know little on the frequency, type, and diversity of Wolbachia in the worldwide populations of H. postica. Further, its effect on host genetic diversification remains unclear. While the genetic composition of invasive populations may be the result of historical demographic events (Estoup & Guillemaud, 2010;Szűcs, Melbourne, Tuff, Weiss-Lehman, & Hufbauer, 2017), the potential role of Wolbachia in the reduction of genetic diversity of alfalfa weevil populations remains unknown.
In order to fill this knowledge gap, we generated a large molecular dataset (mitochondrial and nuclear DNA and molecular information on Wolbachia) of several H. postica populations from different parts of the world. For the first time, we mapped the current distribution of the two lineages of alfalfa weevil in its native and invasive distribution range covering North America, Europe, the Middle East, and Eastern Asia. We further report the degrees of intra-and interpopulation divergence and conducted neutrality analyses to test for signs of selection potentially caused by Wolbachia. Finally, we report the degree of infection for all populations and discuss its potential role for the genetic diversification of the host species. These were also tested for Wolbachia to investigate whether closely related species are infected with the same Wolbachia strains.

| Molecular analyses
DNA was extracted from hind legs using the Wizard™ Genomic DNA Extraction Kit (Promega). Two mitochondrial (COI and CytB) and two nuclear genes (EF1a and CAD) were amplified with specifically designed primers (for information on amplicon size, primer design, detailed PCR, and sequencing conditions, please refer to Appendix S2). Sequences were checked and aligned using ChromasPro v.

| Basic statistics and tests for selection
Polish populations and few Western lineage samples from Iran and Bulgaria, we excluded these from AMOVA. To assess whether the examined genes evolved randomly or not, Tajima's D (Tajima, 1989), Fu's D, and Li's D test (Fu & Li, 1993) were performed in DnaSP.
Based on the flight ability (Prokopy & Gyrisco, 1965) and human-mediated translocations of H. postica , we assumed our populations as panmictic (nonstructured  (Yang, 1997) of PamlX (Xu & Yang, 2013) following the guidelines to test for natural selection effects on a protein-coding mitochondrial gene (Jeffares, Tomiczek, Sojo, & Reis, 2015). Therefore, in order to estimate the dN/dS ratio for all branches, COI and CytB sequences and the maximum-likelihood phylogenetic trees constructed by RAxML v. 8 (Stamatakis, 2014) were used.

| Phylogenetic analyses
To reconstruct a phylogenetic tree for the sampled alfalfa weevil populations, we used MrBayes v. 3.2.1 (Ronquist & Huelsenbeck, 2003). In a first step, the best substitution model was estimated using jModelTest v. 2.1.10 (Darriba, Taboada, Doallo, & Posada, 2012). All four loci were concatenated using SequenceMatrix v. 1.8 (Vaidya, Lohman, & Meier, 2011). PAUP was used to perform a partition-homogeneity test with 10,000 replicates, which suggested congruence of all partitions (p value >.05). We ran MrBayes for 10 million generations, sampling every 1,000 generations yielding a total of 10,000 trees; the first 25% of trees were discarded as burnin. Convergence was checked using the average standard deviations of split frequencies, which were below 0.005 and effective sample sizes (above 200). Trees were visualized and edited with FigTree v.

| Wolbachia detection and strain determination
To detect Wolbachia and to determine the strain, we used five primers of the Multilocus Sequence Typing System (MLST, https ://pubml st.org/Wolba chia) and the wsp locus (Baldo, Hotopp, et al., 2006). In addition, we adopted the ARM  (Iwase, Tani, et al., 2015). PCR conditions for each primer combination followed the original MLST publications (Baldo, Hotopp, et al., 2006). PCR products were purified and sequenced as described above.
Based on the MLST protocol (https://pubml st.org/Wolba chia/), the Wolbachia strain type (profile) of each positively infected weevil was determined. We obtained an additional 26 Wolbachia sequences for phylogenetic comparison from GenBank and the MLST database in order to confirm the Supergroup of the alfalfa weevil Wolbachia strains. (Table S2). Additional sequences were chosen based on the degree of relatedness (most close matches) to our sequences and were further supplemented by several random samples from both Supergroups (Table S2). We then constructed a phylogenetic tree based on a concatenated alignment of all MLST genes (excluding the highly variable wsp gene) using the same methods and settings described above. As Wolbachia has a high mutation and recombination rate, we additionally used the ClonalFrameML approach (Didelot & Wilson, 2015), which corrects the branch length and position of taxa to account for possible recombination, in R (v. 3.2.2), to further validate the Supergroup position of the detected strain. The infection frequency for each population was calculated, infected populations were plotted on a map, and infected haplotypes were marked on the haplotype network (Figures 1 and 2). (1,776 bp, COI and CytB combined) revealed a deep divergence between the two lineages (average of 90 nucleotide differences, 6.15%;

| Lineage determination and global distribution of the alfalfa weevil
Appendix S3, Table S3, Figure S1). The phylogenetic tree ( Figure S2) and the haplotype network ( Figure 2)  Japan, both lineages had the same frequency ( Figure 1, Table 2).

| Genetic and haplotype diversity
Regardless of the region, populations belonging to the Eastern lineage comprised higher levels of mitochondrial diversity compared to the Western lineage (  Table 2A). Based on the haplotype network ( Figure 2), two haplotypes of the Eastern lineage (HPIRAN93) were globally common and were found in Iran, Bulgaria, and Missouri.
In addition, there was one dominant haplotype detected in Japan, Korea, and Iran. The dominant haplotype of the Western lineage was HPUSAMO19 ( Figure 2). This haplotype was observed in all Western lineage populations except that from the Czech Republic.

| Genetic diversification based on nuclear genes
Due to either the lack of PCR products or only weak amplification success, we were not able to amplify and sequence nuclear genes for all samples. EF1a was only successfully sequenced for 78 samples (Appendix S1). The pairwise distance among haplotypes indicated 1%-4.2% nucleotide differences in the intron and 0.1%-0.5% differences for the exon; the complete sequence showed 0.1%-1.1% differences. However, based on the phylogenetic tree ( Figure S3) and haplotype network, no clear geographic pattern of divergence or any level of lineage classification was observed. For the CAD locus, 111 sequences were analyzed (Appendix S1) with 0.07%-5.9% pairwise nucleotide differences. Similar to EF1a, no geographic/lineage pattern of divergence was observed ( Figure S4). Patterns of the distribution of genetic diversity supported the results of mitochondrial genes, but at a lower degree (Table 2B). American populations showed lower diversity compared to the other populations. Similar to mitochondrial genes, Eastern lineage populations showed higher genetic diversity compared to the Western lineage.
Similarly, dN/dS ratio tests found no significant adaptation or selection on mitochondrial genes in the Western and Eastern lineages (ω = .0423, p value: .439, and ω = .0622, p value: .236, respectively).

| Phylogenetic analysis
Nuclear genes were compatible with mitochondrial genes (partitionhomogeneity test p-value = .17) and were used jointly to construct a phylogenetic tree. In addition to the high support of each lineage, Western lineage sample of Lozitsa (Bulgaria) formed a single supportive clade against other Western lineage samples (Figure 3). This can be an indication of intralineage differentiation.
The ABGD analysis detected a maximum four mOTUs. The best supported groups with sufficient prior interspecific divergence (P) (p > .0129) were the Western and Eastern lineages with a deep gap (Figure 4a,b). However, when decreasing the P threshold, more groups were detected (Figure 4c). The Bulgarian Italian samples represented the first splitting from the Eastern lineage with an average of 3.1% nucleotide differences from the other Eastern lineage members (HPItaly88, HPItaly90, and HPBulPL15). In the Western lineage, two distinct singleton mOTUs were observed: (a) a single Iranian sample (HPIRANTA97) and (b) a sample from Bulgaria-Lozitsa (HPBULLO91; Figure 2).

| Wolbachia detection and strain determination
Except for samples of populations from Iran, Italy and Missouri and H. meles, Wolbachia infection was detected in all other populations F I G U R E 3 Concatenated Bayesian tree based on mitochondrial genes (CytB and COI) and nuclear genes (Ef1a and CAD). Each node number represents the posterior probability value. The blue line is indicator for Eastern lineage, red line for Western lineage, green line for out-groups, and black circle is the posterior probability value above 95% F I G U R E 4 ABGD for mitochondrial genes. Maximum 5 molecular taxonomic units inferred with ABGD are presented. (a) The frequency of divergence classes across all samples (without out-groups) a clear divergence is observed. (b) The distance value is ranked and showed the gap from 2% to 6% distance. (c) The numbers of groups detected by ABGD depending on the prior intraspecific threshold, the number of groups started from two major groups (Western-Eastern) at 0.0129 P, so more than this value we expect only two groups until 0.215 P. The third group (Italian Bulgarian) appeared at p = .007 threshold, fourth group (Iranian Western) at p = .0046, and finally the last group is detected at p = .0028 and less than that. Red spot: recursive partition, yellow spot: initial partition and out-group species with variable infection rates (  Figure 1). The infection rate varied from 66% in a Montana population to 6% in Knezha (Bulgaria) (Figure 1). HPUSAMO19 was also a commonly infected haplotype from Japanese to American populations ( Figure 2). In the Eastern lineage, the most commonly infected haplotype was found in Japan and Korea. However, in Bulgaria, every single infection appeared with a unique host haplotype.    or natural migration [Prokopy & Gyrisco, 1965]), may explain their rich gene pool. However, the high genetic diversity typical for of the Eastern lineage was not observed in some of the introduced populations (e.g., Missouri; Table 1). Such reduced genetic diversity is typical for leading edge and introduced populations due to founder effects.

| A more comprehensive picture of genetic diversity and lineage distribution of alfalfa weevil
The Western lineage has a much lower genetic diversity in all populations including introduced and native distributions ( America, and Eastern Asia ( Figure 2, Table 3). This suggests some local differentiation, but also shows that these were not the source of introduction in the non-native regions.

| Type and degree of Wolbachia infection in the alfalfa weevil
Past studies suggested that resistance to Wolbachia evolved in the Eastern lineage of H. postica (Hsaio, 1996). However, according to the best of our knowledge, there is less evidence to support the existence of resistance to all Wolbachia strains in a particular arthropod species (Weinert, Araujo-Jnr, Ahmed, & Welch, 2015). In support of recent findings (Iwase, Tani, et al., 2015), our results refute the clas-  (Iwase & Tani, 2016;Iwase, Tani, et al., 2015). We also found the wHypera1 strain in H. viciae, a closely related species of the alfalfa weevil.
The "phylogenetic distance effect" is a hypothesis predicting a decline in transmission chance of pathogens/symbionts with increasing genetic distance between donor and recipient (Charleston & Robertson, 2002;Engelstädter & Hurst, 2006

| Effect of Wolbachia on the alfalfa weevil genetic diversity
The low genetic diversity of the Western lineage can be an intrinsic character of this lineage. The recent evolved clades usually are observed with lower genetic diversity compare to the elder (ancestor) clade (Tilmon, 2008). Although our data (especially nuclear genes) and analyses cannot confirm or reject the position of the Eastern lineage as the ancestral, the reason of the low genetic diversity of the Western lineage may be found in the young age of this lineage (or other historical elements) rather than a recent genetic reduction.
In the other plausible scenario, genetic diversity reduction might be caused by Wolbachia drives or demographic events. where Wolbachia actually drives the decrease of mitochondrial genetic diversity via sweeping effects (Chen et al., 2016).
The question remains whether Wolbachia has caused the reduced mitochondrial diversity in the Western lineage. We cannot clearly answer this question with our data, but suggest that demographic effects rather than Wobachia have driven this reduced diversity. This is supported by several observations: Firstly, we observed that populations with low genetic diversity (e.g., Nebraska), but also, such with high genetic diversity (e.g., Bulgaria) may be infected with Wolbachia. Further, we did not detect any signs of selection on mitochondrial DNA in most populations regardless of their genetic diversity and lineage (Table 5). Finally, no signs of selection was detected in neither Western nor Eastern lineages. Therefore, there appears to be no clear correlation between Wolbachia infection and the genetic diversity of the alfalfa weevil at most of the locations we studied.
Hence, we suggest that the low genetic diversity noted for some H. postica populations is more likely the result of historical demographic events (i.e., founder effects and bottlenecks) rather than that of Wolbachia infection.
However, in a few populations, our data suggest that Wolbachia may have had a negative effect on genetic diversity. Among populations with sufficient individuals, the two Montana populations have one of the highest infection rates (Table 1, Figure 1). Neutrality test based on mitochondrial genes was rejected for these populations without any trace in the nuclear genes (Table 5)

| CON CLUS ION
The comprehensive molecular data, which is generated in this study, provided a better picture of the global distribution of both alfalfa weevil and its Wolbachia endosymbiont diversity. In addition, we found that regardless of lineage and population, most of the infection of alfalfa weevils are caused by a common strain of Wolbachia named wHypera1. However, high mitochondrial genetic diversity of the Eastern lineage and the low diversity in the Western lineage may be explained better by demographic events rather than Wolbachia infection. The data generated in this study may provide a useful basis for future studies on alfalfa weevil evolution, but may also provide some information for pest management.

ACK N OWLED G M ENT
The authors are indebted to J. Skuhrovec, M. A. Mazur, T. Rand, B.
Puttler, and L. Godfrey for providing part of specimens. We appreciate Jong Seok Kim for his technical aids, Jan Engelstädter for reviewing a draft of manuscript, and three anonymous reviewers for their work. This study was supported by BK21 grant and Chonnam National University.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no competing interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
The datasets supporting the conclusions of this article are included within the article and its additional supporting file (Appendix S1) where you can find the accession numbers of the GenBank repository.