Comparative genetic diversity and structure of the Rhus gall aphid Schlechtendalia chinensis and its host-plant Rhus chinensis

Studying the population genetic structure of both parasites and their host-plants is expected to yield new valuable insights into their coevolution. In this study, we examined and compared the population genetic diversity and structure of 12 populations of the Rhus gall aphid, Schlechtendalia chinensis , and its host-plant, Rhus chinensis , using amplified fragment length polymorphism (AFLP) markers. AMOVA analysis showed that the genetic variance of the aphid and its host-plant were both higher within populations compared to that among them, suggesting that a co-evolutionary history has yielded similar patterns of population genetic structure. We did not detect significant correlation between genetic and geographic distance for either the aphid or host-plant populations, therefore rejecting an isolation by distance model for the demographic histories of the two species. However, our results appeared to suggest that genetically diverse host -plant Rhus populations correlated to similarly genetically diverse populations of gall aphid parasites.


23
The Rhus gall aphids (Hemiptera: Aphididae: Eriosomatinae) have complex life cycles with alternating sexual and parthenogenetic generations, and they are unique in alternating between Rhus (Anacardiaceae) as their primary host in summer and mosses (Bryophyta) as their secondary hosts in winter to complete their life cycle 1 . Rhus gall aphids include six genera and 12 species, which live on primary host Rhus species to form galls with rich tannins to be applied as a raw material in different fields, e.g., medicine, food, dye, chemical and military industry 2 .
Among these aphids, Schlechtendalia chinensis is widely distributed in East Asia, and Rhus chinensis is its unique primary host-plant 2 .
Diversification in parasitic species is known to be tightly linked to diversification in their hosts, so that even small evolutionary or demographic changes in the host may profoundly impact population structures or adaptive change in the parasite 3 . These linked, or coevolutionary relationships have been studied in numerous parasitic species and their plant or animal hosts 4,5 . For example, Jobet et al. examined the population genetic diversity and structure of the urban cockroach and its haplodiploid parasite, an oxyuroid nematode, using RAPD markers and found that the genetic diversity within populations was higher than that between populations for both the oxyuroid nematode and the host-plant 4 . Similarly, Jerome & Ford revealed that the gene flow among populations of the dwarf mistletoe, Arceuthobium americanum Nutt. ex Engelm. (Viscaceae) was similar to its host-plant comprising of Pinus species, and they considered that the population structures were influenced by geographic isolation among populations of Pinus and different environmental conditions 5 .
In a previous study of Schlechtendalia chinensis from China, random amplified polymorphic DNA (RAPD) revealed high genetic variation that the authors attributed to geographic isolation among populations 6 . Subsequently, Ren et al. compared the population structures of S. chinensis and Rhus chinensis in eight populations from Guizhou Province in southwestern China using inter-simple sequence repeat (ISSR) markers 7 . Their results showed 24 that the population genetic structure of S. chinensis was similar to its host-plant, but there was no significant correlation between geographic and genetic distances for either the aphid or its host-plant. A more recent study discussed the origin and genetic divergence of Melaphidina aphids between East Asia and North America to suggest that the distribution of the aphids was influenced by the both the host-plant and the environment, with evidence for the latter being that the aphids do not occur throughout the full range of the hosts 8 .
By comparison to parasites, the demography of the host plants might be influenced by many factors such as rates of seed and pollen transmission and, consequently, gene flow among populations and species as well as historical factors of climate, geology and regional biota 9,10 .
The genus Rhus (family Anacardiaceae) contains a number of wide spread species, which migrated from North America into Asia during the late Eocene (33.8 ± 3.1 million years ago) by the Bering land bridge 11 . The species of R. chinensis exhibited a demographic structure in the temperate and subtropical zones in China, which has been impacted by the uplift of the Qinghai-Tibet Plateau (QTP) 12 .
In this study, we investigated the population genetic structures of Schlechtendalia chinensis and its unique host-plant Rhus chinensis from 12 corresponding populations from six provinces, and tested the correlations between their population structure, genetic diversity, and gene flow using AFLP markers. We expect that our results will provide a framework for coevolution between insects and host plants, and also for the further genetic studies and conservative action.

Results
We found that eight of 64 pairs of AFLP primers tested had high levels of polymorphism and were, therefore, useful for investigating the genetic structure of Schlechtendalia chinensis and Rhus chinensis (Table 1). These primers produced 269 polymorphic bands for the 12 aphid populations, and the rate of polymorphism was 75.6%. For the host-plant, 333 specific bands were produced, and the rate of polymorphism was 81.5% (Table 2).  distances (FST/1-FST) between the aphid and its host-plant populations indicated that there was no significant correlation between the two species (P > 0.05) (Fig. 1A). There was also no  Based on the genetic distances, the NJ tree of the host-plant showed that the populations were clustered into three groups (I, II and III), which were shown in Fig. 3B

Discussion
In the present study, we found that the genetic structure of the aphid, Schlechtendalia chinensis, was partially similar to that of the host plant; or, stated another way, similar by some measures. The level of genetic diversity of the host-plant was higher than other woody plants in the region, such as Phellodendron amurense Rupr. [16][17][18] . High levels of genetic diversity were often observed in plants with long life spans, such as trees, wide geographic distributions, wind pollination, and an abundant fruit yield 19 . Rhus chinensis is a widespread species, which is a perennial small tree or shrub, with long generation cycle, with pollen dissemination by wind [19][20] . Therefore, these functional traits of the plant likely interacted with the environment to 211 strongly influence the genetic diversity of the species 9, 20-21 . The

Materials and Methods
Sample collection. We collected aphid galls and host-plant leaf samples from 12 locations in China (Fig. 3A). When the galls were mature, we cut them open and collected the aphids in absolute ethanol. We stored leaves in plastic zipper bags with silica gel prior to DNA isolation.
We deposited all the voucher specimens at the School of Life Science in Shanxi University,

China.
DNA extraction and AFLP analysis. We immersed the aphid samples in sterile water for 24 hours with the water being changed every eight hours. Thereafter, we extracted genomic DNA of aphid individuals using the conventional phenol/chloroform method 30  Nei's genetic diversity index (H) and Shannon's information index (I) 32 , and performed a mantel test using NTSYSpc 0/1-version 2.11 software based on the binary matrix to analyze 213 the relationships among genetic and geographic distances between the 12 populations 33-34 .
AMOVA was conducted using ARLEQUIN 3.01 35 . We constructed neighbor-joining (NJ) trees of both aphid and its host-plant based on genetic distances using Mega 4.0 software 36 . To compare the most likely number of population genetic clusters (K) in the AFLP datasets for the aphid and host plants (independently), we compared values of K value from two to nine, with ten replicates performed for each K, a burn-in of 1 × 10 5 iterations, and 1 × 10 5 Monte Carlo and Markov Chain (MCMC) steps. We determined the best-fit of clusters using ΔKin STRUCTURE HARVESTER 37 .