Morphological identification of Amphitetranychus species (Acari: Tetranychidae) with crossbreeding, esterase zymograms and DNA barcode data

The genus Amphitetranychus Oudemans (Tetranychidae) consists of only three species, A. quercivorus (Ehara & Gotoh), A. savenkoae (Reck) and A. viennensis (Zacher). The original description of A. savenkoae was extremely simple and had no drawing of the aedeagus; however, a subsequent study described only the aedeagus. The present study investigated all three species in detail using a combination of morphological traits, crossbreeding experiments, esterase zymograms and the mitochondrial cytochrome c oxidase subunit I (COI) gene. Morphological differences in the peritremes and male aedeagi were observed among the three species. Complete reproductive isolation was confirmed in the reciprocal crosses between the morphologically similar A. savenkoae and A. quercivorus (no female offspring were produced). Esterase zymograms differed interspecifically, but not intraspecifically (among individuals in a given species). All three species formed clearly separate clades with 100% bootstrap values in the COI tree, and A. savenkoae was more closely related to A. quercivorus than to A. viennensis, which corresponded to the morphological similarity of their aedeagi and setal counts on tarsi IV. A key to Amphitetranychus species is provided.

Conventionally, researchers use morphological characters to separate species within a genus [8,12,15,17,18]. Among the strains of Amphitetranychus species collected in the present study, A. viennensis was easily identified because of having a distinct aedeagal shape, whereas A. quercivorus and A. savenkoae were quite difficult to separate based only on the aedeagus. Because few morphological features are known to vary among Amphitetranychus species, they may be easier to be distinguished using genetic, physiological and biological methods [19,20]. Therefore, we used a multidisciplinary approach to elucidate the taxonomy of these species based on: (i) morphological redescriptions with a key to species, (ii) crossbreeding experiments to assess the degree of reproductive isolation between A. quercivorus and A. savenkoae strains, (iii) esterase zymogram analysis, and (iv) sequencing of the partial mitochondrial cytochrome c oxidase subunit I (COI) gene and construction of a phylogenetic tree.

Mite samples
A list of species and strains used in the present study is provided in Table 1. The species A. savenkoae was imported to Japan with the authorization of the Ministry of Agriculture, Forestry and Fisheries of Japan (no. 25-Y-532) on 29 Aug. 2013. Laboratory stocks were separately reared on leaf discs (ca. 16 cm 2 ) of cherry, Prunus × yedoensis Matsum., for A. viennensis; deciduous oak, Quercus mongolica Fisch. ex Ledeb., for A. quercivorus and deciduous oak, Q. robur L., for A. savenkoae. The leaf discs were placed on water-saturated polyurethane mats in plastic dishes (90-mm diameter, 20-mm depth) at 25±1˚C under a 16:8 h light:dark photoperiod. In winter, diapause females of A. quercivorus and A. viennensis were produced by the eggs, which were reared at 15˚C under a 10:14 h light:dark photoperiod in the laboratory. When they reached adulthood, they were put in dark boxes at ca. 75% RH using saturated aqueous sodium chloride solution. The boxes were kept in a refrigerator at 5˚C from December to April until 2013 after collection, the period when cherry and deciduous oak leaves were unavailable.

Morphological analyses
Adult males and females for each species were mounted on permanent slides using Hoyer's medium. The specimens were examined using an Olympus 1 BX53 differential interference contrast microscope equipped with an Olympus 1 DP72 digital camera and drawn by a camera lucida (U-DA, Olympus 1 ) attached to the microscope. Illustrations were done with Adobe Illustrator (Adobe Systems Incorporated, USA). Measurements were performed using the imaging software Sensiv Measure 1 ver. 2.6.0. All measurements are given in micrometers (μm) and correspond to mean values followed by the standard error (SE) and the minimum and maximum values (range) of specimens examined. Body length measurements represent the distance between tip of gnathosoma to end of idiosoma. Setae were measured from the center of the setal base to the end of the tip. Terminology and abbreviations used in the description follow Lindquist [21]. Leg setal counts are given in the order: coxa, trochanter, femur, genu, tibia and tarsus. Leg tactile and eupathidial setal numbers are provided first followed by solenidia in parentheses. Some of the voucher specimens of the redescribed species JP17F17397 (Grant-in-Aid for JSPS Research Fellow). The funder (Japan Society for the Promotion of Science KAKENHI) had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Nihon BioData Corporation provided support in the form of salaries for Tomoko Matsuda, but did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. The specific roles of these authors are articulated in the 'author contributions' section.
will be deposited in the collections of the Institute of Zoology, Ilia State University, Georgia, and others were deposited in the Laboratory of Applied Entomology and Zoology, Ibaraki University (AEZIU), Japan, under the serial voucher specimen numbers.

Crossbreeding experiments
To determine the reproductive compatibility between A. savenkoae (Kherson, voucher specimen no. 676) and A. quercivorus (Tsukuba, no. 610), reciprocal crosses were carried out. Individual females in the teleiochrysalis stage (the final immature resting stage) that were obtained from each stock culture were transferred onto a small leaf disc (ca. 4 cm 2 ) with two adult males obtained from the stock cultures of either the same or another species. Five days after the start of oviposition on each disc, the adults (female and males) were removed. Eggs on each leaf disc were checked daily to determine hatchability, survival rate of immature stages and ratio of female offspring. All experiments were carried out at 25˚C and under a 16:8 h light:dark photoperiod.

Electrophoresis
Electrophoresis (native polyacrylamide gel electrophoresis) was carried out to examine whether the three Amphitetranychus species could be discriminated by esterase zymograms. Live females were individually put into a 1.5-ml Eppendorf tube and then homogenized in 10 μl of 32% (w/v) sucrose with 0.1% Triton X-100 and 0.002% bromophenol blue by a pipette tip. The gels were 1-mm thick, 90-mm wide and 83-mm high, and contained Triton X-100 (concentration 0.05% in the separating gels and 0.1% in the stacking gels). The acrylamide concentration was 7.5% in the separating gels and 2.5% in the stacking gels. Electrophoresis was carried out at a constant current of 20 mA/gel at 5˚C for 2 h. The electrode buffer was 25 mM Tris/192 mM glycine, pH 8.6. To detect non-specific esterase, the gels were incubated for 40 min at 30˚C in 0.02% of α-naphthyl acetate in 0.1 M phosphate buffer (pH 6.5) which contained 1% acetone; then, the gels were soaked for 1 h in 0.2% Fast Blue BB salt solution. Ten individuals of each species were used for esterase zymogram analysis.

Molecular analyses
We examined whether the three Amphitetranychus species could be distinguished by the mitochondrial COI gene sequences. Mite species were separately reared on leaf discs of each original host plant on a water-saturated polyurethane mat in a plastic dish (90-mm diameter, 20-mm depth) at 25ºC under a 16:8 h light:dark photoperiod, until analysis. Adult females from each species were arbitrarily selected and used for molecular analyses. Total DNA was extracted from the whole body of each female using PrepMan Ultra Reagent (Applied Biosystems, Foster City, CA). The primers C1-J-1718 [22] (5 0 -GGAGGATTTGGAAATTGATTAGT TCC-3 0 ) and COI REVA [23] (5 0 -GATAAAACGTAATGAAAATGAGCTAC-3 0 ) were used for polymerase chain reaction (PCR  [24,25]. Obtained sequences were aligned using CLUSTAL W in MEGA7 [26]. For the maximum likelihood (ML) analysis, we used the best-fit model (GTR+G model) chosen by MEGA7. ML trees were constructed with MEGA7. Branch robustness was tested by bootstrap analysis with 100 replications.

Data analyses
One-way analysis of variance (ANOVA) was used to compare the influence of crosses on the number of eggs laid per female during the first five days of the oviposition period, egg hatchability, survival rate of immature stages and offspring sex ratio (i.e. the fraction of daughters). Means were compared among cross combinations using Tukey's HSD test [27]. To normalize the data, log transformation was used for the number of eggs laid, and logistic transformation was used for egg hatchability, survival rate of immature stages and offspring sex ratio data [28].  [12]; Ehara [31]. Type-species: Tetranychus viennensis Zacher [5] Diagnosis. Amphitetranychus has no medio-dorsal spur on all empodia of the legs of both sexes; dorsocentral area of opisthosoma more or less with transverse striae; peritreme anastomosed distally; empodium usually without spur, splitting distally into three pairs of hairs; two sets of duplex setae of tarsus I well separated; with only one pair of para-anal setae; and tibia II with six tactile setae.
Male (n = 6). Body pale greenish-yellow. Dorsum-Dorsal body setae simple, acicular, longer than distances between bases of consecutive setae. Gnathosoma-Palptarsus as in Fig 3A; peritreme as in female. Legs-Empodia as in Fig 3B. Leg setal counts as follows (Fig 3C-3F): (1) Tarsus I with four tactile setae and three solenidia proximal to proximal set of duplex seta, one solenidion (φ♂) near duplex seta (Fig 3C). Aedeagus-Shaft of aedeagus with dorsal margin upturned distally forming a neck which ends with a terminal knob, constructing an approximately 45-degree angle between the axis of the knob and dorsal margin of shaft ( Fig  3G).
Male (n = 10). Body pale greenish-yellow. Dorsum-Dorsal setae slender, longer than distances between bases of consecutive setae. Gnathosoma-Palptarsus as in Fig 6A; peritreme as in female. Legs-Empodia as in Fig 6B. Leg setal counts as in A. quercivorus (Fig 6C-6F). Tarsus I with four tactile setae and two solenidia proximal to proximal set of duplex seta (Fig 6C). Tarsus II with five tactile setae and one solenidion proximal to duplex seta, two tactile setae at the same level of duplex seta (Fig 6D). Aedeagus-Similar in shape to A. quercivorus; however, aedeagal knob markedly wider than neck (Fig 6G).

Crossbreeding experiments
In the intra-specific crosses, total egg production, hatchability, survival rate of immature stages and offspring sex ratio (% females) were not significantly different in both A. quercivorus and A. savenkoae ( Table 3). The reciprocal crosses between A. quercivorus and A. savenkoae produced no female offspring (Table 3), indicating that these two species are reproductively isolated. In arrhenotokous species such as spider mites, unfertilized eggs develop into haploid males, whereas fertilized eggs develop into diploid females. Reduction in number of eggs laid and lower egg hatchability and survival rate of immature stages were observed in the inter-specific crosses compared with the intra-specific crosses.

Esterase zymograms
There was no intra-specific variation in esterase zymograms for the three Amphitetranychus species. Esterase band patterns were species-specific and the three Amphitetranychus species could be electrophoretically discriminated (Fig 10).

Molecular analyses
After alignment, the COI fragment had 618 nucleotide sites, of which 113 were parsimonyinformative and contained no insertions or deletions. In the COI tree (Fig 11), all three Amphitetranychus species formed clearly separate clades with 100% bootstrap values. The genetic distances between species were 7-8% between A. quercivorus and A. savenkoae, 9-10% between A. quercivorus and A. viennensis, and 11-12% between A. savenkoae and A. viennensis. These results indicated that the COI sequences are effective for identifying Amphitetranychus species. Additionally, although the associated bootstrap value was relatively low (59%), A. savenkoae was more closely related to A. quercivorus than A. viennensis.

Discussion
The present work provides morphological, crossbreeding and molecular evidence that the genus Amphitetranychus currently includes three taxonomically distinct species: A. quercivorus, A. savenkoae and A. viennensis. Here, we discuss each piece of evidence based on different geographic strains of these species. Comparative examination of morphological features for the five examined strains (one Ukrainian and four Japanese) showed a clear separation between the taxonomic identities of the three species based on female peritremes and male aedeagi. This separation was consistent and similar to that observed in other strains from Korea [16], except for A. savenkoae, which is redescribed in detail for the first time here. Amphitetranychus savenkoae was first described by Reck [8] from specimens collected on Quercus sp. (Fagaceae) in Georgia and was subsequently recorded from Ukraine [12,13]. Reck's [8] original description of A. savenkoae was not sufficiently detailed and lacked   drawings and information on the male aedeagus. Additionally, although Mitrofanov et al. [12] provided a drawing of the male aedeagus, they did not provide any descriptions. Moreover, Reck [8] described female tarsi I and II with 18 and 15 setae, respectively; however, the present specimens had tarsi I and II with only 16 and 14 setae, respectively. This variation in setal counts may be because the old Russian acarologists did not separate in their descriptions between tactile setae and solenidia. We conclude that the three species have strong morphological similarities in measurements and leg chaetotaxy, but the most distinguishable trait was the shape of the peritreme. The crossbreeding experiments confirmed that the studied strains of A. quercivorus and A. savenkoae were reproductively isolated and represent different biological species. Concerning A. viennensis, Gotoh and Takayama [32] reported its reproductive incompatibility with A. quercivorus by crossing populations collected from rosaceous and fagaceous trees, respectively. Therefore, the cross experiments herein were conducted only between the two species, A. quercivorus and A. savenkoae. Furthermore, A. viennensis can be easily separated from the other two species by its distinct shape of aedeagus.
Furthermore, the band patterns of the esterase zymograms were clearly different among the three Amphitetranychus species. Enohara and Amano [33] and Gotoh et al. [34] also applied the same technique to separate various Tetranychus species, and considered enzyme zymograms to be a powerful tool for species separation.
The COI nucleotide sequences also showed that the three Amphitetranychus species are different species because they formed separate clades. Khaing et al. [16] reported 11% variation in COI sequences between the Korean strains of A. quercivorus and A. viennensis. The COI variation (9-10%) between the Japanese strains of A. quercivorus and A. viennensis in our experiments was also similar to the COI variation between the Korean strains.
We conducted various analyses that all produced the same results, which clarified the taxonomic status of the three Amphitetranychus species. The aedeagal shapes indicate the existence of two different morphological groups, which is consistent with the higher similarity between A. quercivorus and A. savenkoae. This close relationship is also supported by the COI analysis, which showed that A. quercivorus is more closely related to A. savenkoae than to A. viennensis. Recently, several case studies have used integrative taxonomy and employed the same methods to separate cryptic species [23,35,36]. We conclude that integration of different approaches is required to distinguish different and closely related taxa and, therefore, to delimit species.