The APICULTURAL SOCIETY OF KOREA
1

Current Issue

Journal of Apiculture - Vol. 38 , No. 4
[Paper List]

[ Original Article ]
Journal of Apiculture - Vol. 33, No. 4, pp. 227-237
Abbreviation: J. Apic.
ISSN: 1225-0252 (Print)
Print publication date 30 Nov 2018
Received 14 Sep 2018 Revised 16 Oct 2018 Accepted 22 Oct 2018
DOI: https://doi.org/10.17519/apiculture.2018.11.33.4.227

Morphometric and Genetic Variation of Tropilaelaps Mites Infesting Apis dorsata and A. mellifera in Thailand
Tipwan Suppasat1, * ; Siriwat Wongsiri2
1Biology Program, School of Science, University of Phayao, Phayao, 56000, Thailand
2Agricultural Interdisciplinary Program, Graduate School, Maejo University, Chiang Mai, 50290, Thailand

Correspondence to : * E-mail: tsuppasat@yahoo.com

Funding Information ▼
Abstract

The majority parasitic bee mites of Thailand in genus Tropilaelaps are infesting colonies of native bees (Apis dorsata) and introduced bees (A. mellifera). The investigation aims to study morphological and genetic variation of Tropilaelaps mites infected different hosts. Adult mites were collected from honey bee brood throughout Thailand. Traditional and geometrical morphometrics were measured on photograph by using TPS program. Additional, COI gene variations were examined by PCR-RFLP and nucleotides sequencing. Tree of mites relationships were constructed by NJ and MP assumptions. Morphometric results indicated T. mercedesae were major species infesting on A. dorsata and A. mellifera. Mophological variation represented at anal and epigynial plate, which the shape of the anal plate apex margin has been key character to identify between T. mercedesae (bell to blunt shape) and T. koenigerum (pear shape). However, the discriminant analysis suggested that geometric results were potential to classify Thai Tropilaelaps populations from different hosts better than traditional morphometric. Otherwise, PCR-RFLP clearly detected the site of Dra I and Xba I digestion of Thai Tropilaelaps morphotypes. The COI sequences of T. koenigerum were founded infesting only A. dorsata in Thailand and four sequences that related to the Thai T. mercedesae morphotypes. The NJ and MP tree were clearly classified Thai Tropilaelaps species which were suggested both from morphological and molecular analysis. This information might be basically of taxonomic status, but this should have implication for controlling these mites in Thailand and other countries.

Keywords: Apis dorsata, COI, Geometric, Phylogeny, Tropilaelaps, Variation
INTRODUCTION

Honey bee mites belonging to Genus Tropilaelaps are serious pest of beekeeping in the world. They have tiny oval body shape and approximately 1 mm in length. Tropilaelaps mites are ectoparasite on honey bee body; induce of several diseases such as viruses, bacteria or protozoa destroying the bee colonies. This causes Colonies Collapse Disorder (CCD) in Apis mellifera beekeeping (Sammataro et al., 2000). The mites invade native naturally colony of giant honey bee (Apis dorsata or A. larboriosa), those are original in Asia. The first taxonomic data named Tropilaelaps clareae, was found from A. dorsata in Philippine (Delfinado-Baker and Baker, 1961). Recently T. clareae has shifted host to commercial honey bee (A. mellifera) and becomes cosmopolitan pest to destroy beekeeping. Another species was T. koenigerum from A. larboriosa, found distributing in Brunei, India, Nepal, Thailand and Vietnam except in Sri Lanka (Burgett et al., 1983; Tangjingjai et al., 2003). Sometimes T. koenigerum was found in A. dorsata but was not serious in A. mellifera beekeeping (Koeniger et al., 2002; Anderson and Morgan, 2007).

At present taxonomic data of the parasitic bee mites has been revised by using morphological as well as molecular study. To cite an example the mites belonging to the genus Varroa which infests two honey bees species i.e. A. cerana or A. mellifera can be mentioned. There are four species (V. destructor, V. jacobsoni, V. rindereri and V. underwoodi) and the serious pest of A. mellifera beekeeping in the world has been found V. destructor (Anderson and Trueman, 2000). Situation of parasitic bee mites in Thailand is worth mentioning because of their variability of genetically haplotype as genus Varroa infested on native host (A. cerana). There are 5 haplotypes as Vietnam haplotype of V. destructer and 4 haplotypes of V. jacobsoni such as Malaysia, Norththai1, Norththai 2 and Samui haplotypes (Warrit et al., 2006; Warrit and Lekprayoon, 2011). In Thailand, A. mellifera colonies were more infested with Korean haplotype of V. destructor than Japan/Thailand haplotype and were not found V. jacobsoni (Suppasat, 2007). Anderson and Morgan (2007) revised the mite belonging genus Tropilaelaps, naturally infested single nesting bees (A. d. dorsata, A. d. binghami, A. d. breviligula and A. laboriosa). The genus Tropilaelaps has two species, T. mercedesae and T. thaii among which the most predominating damage causing mite in A. mellifera is T. mercedecae. Recently, in China, the tropilaelaps mites infesting A. mellifera was T. mercedesae instead T. clareae, supported by using COI sequences (Luo et al., 2011).

Morphological variation of tropilaelaps mites showed the differences in size and shape such as ventral and dorsal plate, epigynial plate, anal plate, setae and stigmata characters. T. clareae, T. koenigerum and T. thaii can be identified based on the anal plate characters while T. mercedecae bears diverse characters (Anderson and Morgan, 2007; Anderson and Robert, 2013; Kavinseksan and Wongsiri, 2016). Morphometric analysis including traditional and geometric morphometric is powerful techniques for determining morphology (Daly, 1985; Rohlf, 1999). Recently, geometric morphometric is popular method to analyzed size, shape and orientation in several insect wings such as mosquitoes (Rohlf, 2002). Likewise, the geometric morphometric was used to investigate the mite Aceria guerreronis Keifer (Acari: Eriophyidae), those populations distributed in America, Africa and Asia. The geometric shape of coxigenital region (anal plate) variation of the mite is effective to show relation with their population distribution (Navia et al., 2006). The eleven species of oppiid mites (Acari: Oribatida) in Turkey were determined 17 measurements for traditional morphometric and were plotted 20 landmarks on ventral surface for geometric morphometric. Both of methods separated in genera of oppiid mites (Baran et al., 2011). Genetic variation of COI gene is popular to apply for species identification of Varroa and Tropilaelaps genera. Moreover, biogeography of mites and their native hosts were represented with the COI sequences which were hypothesized by using phylogenetic relationship. By the way, the tropilaelaps mite, examined highly fashionable identification in morphological and genetic variability of ITS and COI genes of mitochondrial DNA (Anderson and Morgan, 2007; Luo et al., 2011).

In Thailand, the first known species belonging to genus Tropilaelaps was T. clareae, which infested on naturally A. dorsata or A. cerana colonies. Later it also shifted and became serious pest of A. mellifera colonies while T. koenigerum infected only A. dorsata (Burgett et al., 1983; Tangjingjai et al., 2003). In early 2000, the T. clareae and T. koenigerum were investigated by using PCR-RFLP of ITS gene and RAPD methods to develop specific marker for identification (Tangjingjai et al., 2003). At present the most predominating Tropilaelaps mite infesting A. mellifera is T. mercedesae (Anderson and Morgan, 2007). This research aims to examine morphological and genetic variation of the tropilaelaps mites infesting native host (A. dorsata) and exotic host (A. mellifera) in Thailand. The Thai tropilaelaps mite species are identified by morphometric and COI markers.

MATERIALS AND METHODS
Thai tropilaelaps collections

The adult tropilaelaps mites were collected from honey bee brood of native host (A. dorsata) and exotic host (A. mellifera) in Thailand between 2008-2010. The mites (n=307) were collected from infesting colonies of A. dorsata (15 colonies, n=88) and A. mellifera (21 colonies, n=219). The tropilaelaps samples were kept in 70% ethanol for morphometric study and were preserved in absolute ethanol then kept them at -4°C for molecular study.

Morphometric analysis

Collected mites were mounted in Hoyer’s medium (Krantz, 1978) on slide. All samples were photographed using Nikon EClipsee 200 digital microscope, and photographs were used for morphometric analysis with tpsDig 2.04 (Rohlf, 2010a). Firstly, traditional morphometric was measured width (W) and length (L) of ventral shield (Ven), anal shield (Anal) and epigynial shield (Epig) for 6 characters (W-Ven, L-Ven, W-Anal, L-Anal, W-Epig and L-Epig) in totally (Fig. 1a). Then, the statistically significance of 6 character measurements of tropilaelaps mites from different hosts were analyzed by using one-way ANOVA and discriminant analysis. Secondly, geometric morphometric digitized landmark on 152 mite photographs that divided female mite infesting A. dorsata for 60 samples and A. mellifera for 92 samples. All of tropilaelaps images were plotted 30 landmarks on ventral side by using tpsDig 2.04 (Fig. 1b). Geometric analysis calculated variation in each landmark. And, principle component analysis (PCA) were generated partial and relative warp by using tpsRelW 1.49 (Rohlf, 2010b).


Fig. 1. 
Traditional morphometric was measured for 6 characters of width and length on ventral plate such as body size (-) represented 1and 2, epigynial plate (--) represented 3 and 4 and anal plate (…) represented 5 and 6 (a). Geometric landmarks were plotted 30 points on ventral side of mite image (b).

Molecular analysis

DNA analysis of adult female tropilaelaps mites were investigated in mitochondrial DNA i.e. COI gene. There were 53 tropilaelaps mites, were form different hosts; A. dorsata (n=23) and A. mellifera (n=30). One mite was extracted DNA using GeneluteTM Mammalian Genomic DNA Kit (Sigma, USA). Polymerase Chain Reaction (PCR) of COI was amplified using pair of primers TCF-1 and TCR-2 following Anderson and Morgan (2007). PCR product was checked on 1.5% agarose gel electrophoresis and was purified with RBC hiyield gel/PCRDNA fragments extraction kit (RBC, Taiwan). Then, PCR-RFLP was examined with 5μl of the PCR products that were digested using Dra I and Xba I (BioLabs, UK). Restriction fragments were run on 2.0% agarose gel electrophoresis and scored restriction patterns. Further, COI sequences were generated nucleotide sequences of Thai tropilaelaps mites. Multiple sequences of Thai tropilaelaps were aligned using ClustalW in Mega 6.0 program (Tamura et al., 2011). All the Thai tropilaelaps haplotypes were aligned using 4 sequences of T. clareae (EF025464, EF025466, EF025467 and EF025468), 3 sequences of T. koenigerum (EF025449, EF025450 and EF025451), 9 sequences of T. mercedesae (EF025429, EF025430, EF025431, EF025432, EF025437, EF025438, EF025445, EF025446 and EF025447) and T. thaii (EF025452).

Tree reconstruction of Thai tropilaelaps mites relationship were reconstructed by using neighbor-joining (NJ) and maximum parsimony (MP) methods, additional statistical with bootstrap 1,000 replications in Mega 6.0. The first method of neighbor-joining was performed by using distance matrices with p-distance and Kimura-2 parameter (K2) models. Also maximum parsimony was generated using both models. The tree reconstruction followed Morgan and Anderson (2007) and added performance of K2 model. All of tree reconstruction methods used condense consensus tree that performed statistical value at 50% cut off following majority’s rule to show relationship between Thai tropilaelaps mites and other from NCBI database.

RESULTS
Morphometric variation of Thai tropilaelaps mites

The traditional morphometric showed that almost all Thai tropilaelaps mites infesting different hosts (A. dorsata (Ad) and A. mellifera (Am)) was T. mercedesae (Tm), but 3 samples infesting A. dorsata were resembles to T. koenigerum (Tk) in their anal plate as pear shape. The Thai tropilaelaps mites infesting different hosts were significantly difference (P<0.05) in their average for 4 morphometric characters (Table 1).

Table 1. 
Morphometric of Thai tropilaelaps mites were showed significantly different (*) of 6 morphometric characters at ventral plate by using Independent T-test (P<0.05, df1=87, df2=218, df3=305), those were 4 significant characters as L-Ven, L-Anal, WAnal and L-Epig
Characters Tropilaelaps size average (Mean±S.D.) mm.
Tropilaelaps infesting
A. dorsata
(n=88, df1=87)		 Tropilaelaps infesting
A. mellifera
(n=219, df2=218 )		 Overall
(n=307, df3=305)
Length of ventral plate (L-Ven)* 1.293±0.081 1.252±0.043 1.263±0.060
Width of ventral plate (W-Ven) 0.706±0.053 0.706±0.027 0.706±0.036
Length of anal plate (L-Anal)* 0.289±0.018 0.279±0.018 0.282±0.018
Width of anal plate (W-Anal)* 0.183±0.016 0.190±0.012 0.188±0.014
Length of epigynial plate (L-Epig)* 0.750±0.104 0.788±0.066 0.777±0.081
Width of epigynial plate (W-Epig) 0.144±0.072 0.133±0.011 0.136±0.040

Traditional morphometric results indicated that Tropilaelaps infesting A. dorsata have longer body size and anal plate than mites infesting A. mellifera. But, the anal plate of mites infesting A. mellifera was wider than the others. In contrast mites from A. mellifera have narrow epigynial plate. Additionally, the mahalanobis distance discriminant supported 4 morphometric characters such as length of ventral plate, length and width of anal plate and length of epigynial plate, those represented the relationship in each group of mites infesting different honey bee hosts. Moreover, the canonical discriminant function showed eigenvalue for 0.484, 100% of cumulative and canonical correlation for 0.571, also test of function showed Wilks’ lambda was 0.674. And cross validation predicted group of Thai tropilaelaps mite correctly classified into each species for 78.8% (Table 2).

Table 2. 
Discriminant analysis of Thai tropilaelaps mites performed with significant characters (Variables) and compared between traditional and geometric methods. Results were represented statistical discriminant value from each morphometric methods of the tropilaelaps mite with canonical correlation, Wilks’ Lambda and predicted group of infesting their hosts using cross validation
Morphometric
Methods 		Variables Standardized
canonical
discriminant
function
coefficients 		Canonical
correlation 		Eigenvalue Wilks’
Lambda 		Cross validation
Predicted group of mites
Infesting
A.dorsata		 Infesting
A.mellifera
Traditioal
morphometric 		L-Ven 0.559 0.571 0.484 0.674 76.7% 23.3%
L-Anal 0.426 15.9% 84.1%
W-Anal -0.415 Cross-validated group
L-Epig -0.714 78.8%
Geometric
morphometric 		warp 1y -0.398 0.769 1.448 0.409 86.7% 13.3%
warp7y 0.987 10.9% 89.1%
warp8x -0.325 Cross-validated group
warp11y\ 0.327 88.2%
warp21x 0.356
warp21y 0.607
unix 0.437

In addition to results of geometric morphometric represented superimposition of 30 landmarks belonging to 152 female tropilaelaps mites, showed the deviation in shape of the mites infesting A. dorsata and A. mellifera (Fig. 2a). The landmark digitization of Thai tropilaelaps mites were shown consensus configuration presented by the point of the origin vectors. The variation of mites shape was shown highly variances (S2=0.00006 to 0.00012) at ventral plate shape at the landmark (LM) 1 to 12. While, the potentially LM was shown at anal plate, were highly relative contribution of each landmark (the lowest at LM 6, SS=0.02253 to the highest at LM 20, SS=0.21008). Focus on the analysis demonstrated to describe the morphological variation between Thai tropilaelaps mites infesting different hosts. Furthermore, the relative warp results showed centroid size of Thai tropilaelaps mites, were performed 56 relative warp deformations for the populations. In addition to mahalanobis distance indicated 7 relative warps indicated variability shape of the Thai tropilaelaps mite populations from their infesting different honey bee hosts. Another results of PCA plotting graph showed distribution from coordination of Thai tropilaelaps mite populations of PC1 on X and PC2 on Y axes (Fig. 2c).


Fig. 2. 
The Tropilaelaps consensus configuration showed vector variation in 30 landmarks (a). The anal plate shape of Thai tropilaelaps infesting A. dorsata (1, 2) and A. mellifera (3, 4) were as same as T. mercedesae following Anderson and Morgan (2007). There were 2 types of anal plate shape, represented with “a” and “b” (b). The PCA of Thai tropilaelaps samples distributed from coordinate axis of X and Y (c). The scatter plots were shown using relative warp analysis of the mite infesting A. dorsata (●) and A. mellifera (○).

All of morphological analysis indicated that Thai tropilaelaps mites were as same as species of T. mercedaecae. They showed variation of the mites infesting different hosts in their morphological size and shape. Moreover, the discriminant analysis showed values of the canonical discriminant function were eigenvalue for 1.448, 100% of cumulative and canonical correlation for 0.769. Also, test of function supported with Wilks’ Lambda for 0.409. And, cross validation in the geometric shape were correctly classified for 88.2% (Table 2) that higher than the morphometric analysis. However, Thai tropilaelaps mites were similar to T. mecedecae, showed variation in their anal shield in both methods.

COI sequences variation of Thai tropilaelaps mites

From morphological data of Thai tropilaelaps mites were identified into T. mercedesae (“Tm”) based on morphological of anal shield shape following Anderson and Morgan (2007). The COI gene of Thai tropilaelaps mites showed PCR products size approximately 535-538 bp. as same as previously study of genus Tropilaelaps. Likewise, restriction pattern of Dra I and Xba I only digested the PCR products of T. mecedesae but was not digest the T. koenigerum (Fig. 3). However, RFLP patterns of Thai tropilaelaps mites, T. mercedesae, were not different between mites infesting different hosts or the variation of anal shield morphotypes. However, there were 3 samples of Thai tropilaelaps mites infesting A. dorsata, identified into T. koenigerum (“Tk”). The COI gene variation represented for 5 difference haplotypes of COI sequences within Thai tropilaelaps mites. The comparison of COI sequences of Thai tropilaelaps mites consisted of 3 sequences of T. koenigerum infesting A. dorsata represented “AdTk” haplotype. The 23 sequences of T. mercedesae infesting A. dorsata showed 2 difference haplotypes; “AdTma” and “AdTmb” and the 30 sequences of mites infesting A. mellifera represented “AmTma” and “AmTmb” haplotypes. The multiple indicated that 4 nucleotides sequences of AdTma, AdTmb, AmTma and AmTmb were closely resemble to the references sequences (EF025445, “TmThailand1EF025445”). Especially, the nucleotide sequences of AmTma and AmTmb showed 100% similarity with the references sequence. But the sequences showed transition substitution at 513 site of T. Otherwise, the sequences of T. mercedecae infesting A. dorsata, AdTma, showed transition substitution at 200 sites of T and AdTmb showed at 457 sites. Both of them were resemble to T. koenigerum sequences. Moreover, the sequences of AdTmb showed transversional substitution at 467 sites of C that was closely related with T. clareae (TcMindanao1EF025464) and T. thaii (TtVeitnam1EF025452) (Fig. 3 and 4). However, the sequences of AdTk showed 100% similarity with T. koenigerum.


Fig. 3. 
Multiple alignments of COI gene sequences were compared by 18 Tropilaelaps sequences from GenBank database and 5 sequences of Thai tropilaelaps mites (AdTk, AdTma, AdTmb, AmTma and AmTmb) using Mega version 6.0. The partial sequences were shown for their differences. The restriction site of Dra I and Xba I presented in the box. Also, variable sequences of Thai tropilaelaps mites represented substitution base such as transition showed with bold capital letter and transversion showed with underline bold capital letter.


Fig. 4. 
Phylogenetic tree reconstructed based on 538 bp. COI sequences of five Thai tropilaelaps and GenBank sequences. NJ and MP methods were performed with bootstrap 1,000 replications and calculated 50% majority cut-off tree branching by using Mega 6.0 program. The tree results of NJ (a) and MP (b) indicated that 4 Thai tropilaelaps sequences (AdTma, AdTmb, AmTma and AmTmb) clustered in T. mercedesae clade and one sequences (AdTk) was in T. koenigerum clade. In detail of NJ tree (a) reconstructed by p-distance and Kimura-2-Parameter (K2P), they were similarity in their relationship (SBL=0.375. And, MP tree (b) were random addition of 10 of no. of initial trees, CI=0.705, RI=0.883, RCI=0.623.

Phylogenetic relationship of COI sequences represented similar results which performed under all three criteria as neighbor-joining (NJ) and maximum parsimony (MP), additional statistical tested with bootstrap 1,000 replications. Almost all Thai tropilaelaps mites infesting A. dorsata and A. mellifera were same as T. mecedesae sequences from database and found a small proportion T. koenigerum sequences. But the investigation was not found T. clareae or T. thaii sequences. However, the tree indicated that 4 sequences of Thai tropilaelaps mites grouped into the T. mercedesae clade were mostly resemble to sequences accession number EF025445. While, the mites infesting A. dorsata from Lampang province of Thailand “AdTk” sequences clustered into T. koenigerum clade. The phylogenetic of neighbor-joining were performed with p-distance and K2P. Unweighted NJ tree were generated bootstraps for 1,000 replications and 50% majority rule consensus and condense tree showed SBL for 0.342 (Fig. 4a). In the maximum parsimony was shown most parsimonious trees length (207). The consistency index (CI) was 0.670 the retention index (RI) was 0.883 and the composite index (RCI) was 0.623 (Fig. 4b).

Table 3. 
The genetic distance estimation between groups was performed using p-distance method in 23 COI sequences using Mega version 6.0. There were 6 groups such as T. clareae, T. mercedesae, T. koenigerum and T. thaii followed COI sequences database (Anderson and Morgan, 2007; Luo et al., 2011). And, there were Thai T. mercedaecae sequences (AdTma, AdTmb, AmTma and AmTmb) and Thai T. koenigerum (AdTk). Overall mean distance estimation for substitution (d) was 2.147. The percentages of distance estimation overall (%) and between groups represented lower than 0.05 in the value (bold number*) supported in their closely related of the COI sequences in each Tropilaelaps groups. The pairwise distance of sequences divergence showed overall estimate distance for 8.3%.
Mites T. mercedecae Thai
T. mercedecae		 Thai
T. koenigerum		 T. koenigerum T. clareae T. thaii
T. mercedecae 0.000
Thai T. mercedecae 0.007* (0.7) 0.000
Thai T. koenigerum 0.146 (14.6) 0.142 (14.2) 0.000
T. koenigerum 0.147 (14.7) 0.143 (14.3) 0.008* (0.8) 0.000
T. clareae 0.125 (12.5) 0.124 (12.4) 0.141 (14.1) 0.144 (14.4) 0.000
T. thaii 0.119 (11.9) 0.114 (11.4) 0.133 (13.3) 0.136 (13.6) 0.110(11.0) 0.000

DISCUSSION

The major of Thai tropilaelaps mites infesting A. dosata were identified into T. mecedesae. We did not detect T. clareae and T. thaii. But, we found 3 mites infesting A. dorsata were small in body size and pear shape anal plate similar to the findings of other reports in context to T. koenigerum morphology (Anderson and Morgan, 2007; Kavinseksan and Wongsiri, 2016). Mophological variation of the female Tropilaelaps was found only T. mercedesae infesting different hosts of A. dorsata and A. mellifera. There was significantly morphological variation in the anal and epigynial plate (Anderson and Morgan, 2007). The measurements of traditional and geometric morphometric on ventral plate image by using tpsDig 2.04 are effective identification Thai tropilaelaps mites. We suggested that the Thai tropilaelaps mites, T. mercedesae, infesting different hosts (A. dorsata and A. mellifera), showed different size and shape of anal and epigynial plate. There were two forms (“a” and “b”) at the apex of anal plate in each mite infesting differently hosts. The “a” form showed as bell shape (infesting A. dorsata), slightly blunt shape (infesting A. mellifera) and “b” showed wide bell shape (infesting A. dorsata) and wide blunt shape (infesting A. mellifera) (Fig. 2b). Almost of T. mercedesae were large size and varies shape from blunt to sharp at the anterior margin of anal and the apex of the epigynial plate. Those showed overlapping of the both plates in adult female mites that found to be artefact caused by mounting on glass slide.

Recently, the geometric morphometric were clearly indicated variation, relationship and their biogeography of the mite populations such as coconut mites, Aceria guerreronis (Navia et al., 2005; 2006) and oppiid mites from Turkey (Baran et al., 2011). The geometric method is powerful to determining shape variation of T. mercedesaae infesting different hosts that was firstly utilized in the female Thai tropilaelaps mites. The result suggested that 30 landmarks on ventral plate showed relative deformation of body shape, anal and epigynial plate variation. Likewise the deformation of coxigenital and ventral regions variation of Aceria guerreronis populations suggests morphometric geometric method using for mite identification (Navia et al., 2006). Furthermore, traditional and geometric methods separated the oppiid mites in genera taxa quite well. But the results of geometric morphometric considered these mites as subgenera such as the genus Oppiella (Baran et al., 2011). These results considered Tropilaelaps mites populations infesting different hosts, A. dorsata and A. mellifera, they were overlapping populations of T. mercedesae (Fig. 2c). However, discriminant analysis indicated that they may be grouped their morphological related to their hosts. And, geometric method was higher percentage of cross-validation to classification than traditional method (Table 2). Although, the analyses conducted in this investigation suggest that geometric methods may be useful in understanding morphological variability of Tropilaelaps mites.

The COI sequences of Thai tropilaelaps mites were reported for 5 haplotypes, especially one of T. koenigerum infesting A. dorsata from Thailand which differed from previous investigation (Anderson and Morgan, 2007; Luo et al., 2011.). In Thailand, T. koenigerum was reported RFLP markers of ITS gene in mitochondrial DNA to identify mites. The ITS polymorphism was not appropriate for genetic diversity of Tropilaelaps spp. at the intraspecific level (Tangjingjai et al., 2003). But the result of COI sequences variation of Thai tropilaelaps mites indicated clearly of two distinct species of mites infesting A. dorsata such as T. koenigerum and T. mercedesae. While, the Thai tropilaelaps mites infesting A. mellifera indicated T. mecedesae. The pairwise distance of COI sequences were estimated evolutionary divergence (Tamura et al., 2011) within and between groups. These results of demonstrated percentage of genetic distance within and between groups were range from 1.2 to 4% and 0.11.1 to 14.7%. Similar to Anderson and Morgan (2007) reports pairwise distances of Tropilaelaps was within 1.2~3.5% and between are 11.1 to14.7%. Moreover, genetic distance of COI gene in the other Acari have shown 2.5% and 16% sequences divergence within and between species of genus Stratiolaelaps as same family as Tropilaelaps. And percentages of sequences divergence were 0~2.1% and 6.1~9.1% within and between species of Varroa as same order as Tropilaelaps (Anderson and Trueman, 2000). We concluded that the divergence distance within and between COI sequences haplotypes of these Thai tropilaelaps mites indicated 2 species such as Thai T. mercedesae and Thai T. koenigerum. Almost of Tropilaelaps mites considered species of T. mercedesae which have common found infesting naturally A. dorsata and domestically A. mellifera colonies of Thailand. Moreover, phylogenetic relationship of NJ and MP tree with bootstrap analysis supports this conclusion.

Recently, the beekeeping of A. mellifera in Thailand has found Tropilaelaps infections as same as other countries which wondered the problem of colony collapse disorder (CCD) such as in North America (vanEngelsdorp et al., 2009). But, there were no reports of CCD in Thailand. We are wondering about the CCD problem which has a little understanding of the causes. Parasitic bee mites, Varroa and Tropilaelaps, infection can lead to weakening of the colony. Such colonies have diminishing amounts of brood and adult bees causing collapses of the colonies and T. mercedesae can kill untreated colonies of A. mellifera (Kavinseksan and Wongsiri, 2016).

In conclusion, these results may be shown detail of methods to identify Thai tropilaelaps species using both of morphological and molecular analysis. Which have shown two mites species (T. koenigerum and T. mercedesae) infesting on native host (A. dorsata) and introduced host (A. mellifera). A little number of T. koenigerum, infected only A. dorsata, which do not find these mites in A. mellifera colonies. These species suggests clearly host-parasite relationship. While, T. mercedesae can be found widely infesting on honey bees host. They can great colonize A. mellifera colonies which may be destroyed honey bee development, adult disorder or cause of pathogen transmission. These results suggest that T. mercedesae has been variability in their morphology such as anal plate and COI sequences. The relationship of T. mercedesae infesting difference bee hosts has shown high variation in both data. Reasonable, the T. mercedesae shifted host to A. mellifera, they may be well adaptation and living with the newly environment. So T. mercedesae has become major pest in A. mellifera beekeeping in Thailand and the world. However, the successful morphological marker (traditional and geometric methods) and molecular markers (COI sequences) demonstrates the powerful techniques to identify Thai tropilaelaps mites. These markers are useful for tracing the colonization of newly bee hosts and are benefit for beekeeping of Thailand or other countries.

Acknowledgments

We are grateful for financial support from the Thailand Research Fund (MRG5280153). We also thank Thai beekeepers, the staff of the Center of Excellence: Bee Biology Biodiversity of Insect and Mites of Chulalongkorn University, for their valuable help. Thanks also go to Assist. Prof. Dr. Sureerat Deowanish for her kind support molecular equipment and Miss Kesorn Panakwa for her independent study (IS) works in the undergraduate in the University of Phayao.

References
1. Anderson, D. L., and J. W. H. Trueman, (2000), Varroa jacobsoni (Acari: Varroidae) is more than one species, J. Exp and App. Acarol, 24, p165-189.
2. Anderson, D. L., and M. J. Morgan, (2007), Genetic and morphological variation of bee-parasitic Tropilaelaps mites (Acari: Laelapidae): new and re-defined species, J. Exp and App. Acarol, 43, p1-24.
3. Anderson, D. L., and J. M. K. Roberts, (2013), Standard methods for Tropilaelaps mites research, J. Apic. Res, 52, p1-16.
4. Baran, S., A. Altun, N. Ayyildiz, and A. Kence, (2011), Morphometric analysis of oppiid mites (Acari, Oribatida) collected from Turkey, J. Exp and App. Acarol, 54, p411-420.
5. Burgett, M., P. Akratanakul, and R. A. Morse, (1983), Tropilaelaps clareae: a parasite of honeybees in south-east Asia, Bee World, 64, p25-28.
6. Daly, H. V., (1985), Insect morphometrics, Ann.Rev. Entomol, 30, p415-438.
7. Delfinado-Baker, M., and E. W. Baker, (1961), Tropilaelaps, a new species of mite from Philippines (Laelapidae: Acari), Field. Zool, 44, p53-56.
8. Kavinseksan, B., and S. Wongsiri, (2016), Life history and control of the parasitic bee mite, Tropilaelaps mercedesae Anderson and Morgan (Acari: Laelapidae): A Review, Basic Res. J. Agric. Sci. Rev, 5, p56-71.
9. Krantz, G. W., (1978), Collection, rearing, and preparation for study, In: A Manual of Acarology, Oregon State University Bookstore, p77-98.
10. Koeniger, G., N. Koeniger, D. L. Snderson, C. Lekprayoon, and S. Tingek, (2002), Mites from debris and sealed brood cells of Apis dorsata colonies in Sabah (Barnea) Malaysia, including a new haplotype of Varroa jacobsoni, Apidologie, 33, p15-25.
11. Luo, Q. H., T. Zhou, Q. Wang, P. L. Dai, Y. Y. Wu, and H. L. Song, (2011), Identification of Tropilaelaps mites (Acari, Laelapidae) infesting Apis mellifera in China, Apidologie, 42, p485-498.
12. Navia, D., G. J. de Moraes, G. Roderick, and M. Najavas, (2005), The invasive coconut mites Aceria guerreronis (Acari: Eriophyidae): origin and invasion sources inferred from mitochondrial (16S) and nuclear (ITS) sequences, Bull. Entomol. Res, 95, p505-516.
13. Navia, D., G. J. de Moraes, and R. B. Querino, (2006), Geographic variation in the coconut mite, Aceria guerreronis Keifer (Acari: Eriophyidae): a geometric morphometric analysis, Inter. J. Acarol, 32, p301-314.
14. Rohlf, F. J., (1999), Shape statistics: Procrustes superimpositions and tangent spaces, J. of Class, 16, p197-223.
15. Rohlf, F. J., (2002), Geometric morphometric and phylogeny, p175-193, In, Macleod, N, Forey, PL (eds.), Morphology Shape and Phylogeny, Taylor & Francis Groups.
16. Rohlf, F. J., (2010a), tpsDig. Version 2.04. Morphometric at Stony Brook software, Ecology and Evolution, State University of New York at Stony Brook, New-York, USA.
17. Rohlf, F. J., (2010b), tpsRelw. Version 1.49. Morphometric at Stony Brook software, Ecology and Evolution, State University of New York at Stony Brook, New-York, USA.
18. Sammataro, D., U. Gerson, and G. Needham, (2000), Parasitic mites of honey bees: life history, implications and impact, Ann. Rev. Entomol, 45, p519-548.
19. Suppasat, T., (2007), Genetic relationships between two honey bees (Apis mellifera Linnaeus, 1758 and Apis cerana Fabricius, 1753) and varroa mites in Thailand, PhD. Thesis, Faculty of Science, Chulalongkorn University.
20. Tamura, K., D. Peterson, N. Peterson, M. Nei, and S. Kumar, (2011), MEGA6: Molecular genetics analysis using maximum likelihood, evolutionary distance and maximum parsimony methods, Molec. Bio. Evol, 28, p2731-2739.
21. Tangjingjai, W., P. Verakalasa, S. Sittipraneed, S. Klinbunga, and C. Lekprayoon, (2003), Genetic differences between Tropilaelaps clareae and Tropilaelaps koenigerum in Thailand based on ITS and RAPD analysis, Apidologie, 34, p513-524.
22. vanEngelsdorp, D., J. D. Evans, C. Saegerman, C. Mullin, E. Haubruge, B. K. Nguyen, M. Frazier, J. Frazier, D. Cox-Foster, Y. Chen, R. Underwood, D. R. Tarpy, and J. S. Petti, (2009), Colony collapse disorder: a descriptive study, PLoS One, 4, e6481.
23. Warrit, N., D. R. Smith, and C. Lekprayoon, (2006), Genetic subpopulations of Varroa mites and their Apis cerana hosts in Thailand, Apidologie, 37, p19-30.
24. Warrit, N., and J. Lekprayoon, (2011), Asian Honeybee Mites, In, Hepburn, H. R., and S. E. Radloff (eds.), Honeybees of Asia, Springer-Verlag, Berlin Heidelberg, p347-368.
 
Copyright statement. The Apicultural Society of Korea
In Department of Agricultural Biology, National Institute of Agricultural Sciences, RDA, Wanju 55365, Republic of Korea
Contact: Prof. Hye Kyung Kim(Tel. +82-63-238-9316, E-mail. hyegyeong@korea.kr)
Printed by Junghaengsa, Inc.
(Tel. +82-2-2232-3281, Fax. +82-2-2232-5874, E-mail. jhpub3213@naver.com)