Research ArticleGenetic Diversity of Apis spp in Thailand Inferred from 28SrRNA Nuclear and Cytochrome b Mitochondrial Gene Sequences

ThitipanMeemongkolkiat1 Atsalek Rattanawannee2 and Chanpen Chanchao 3

1Program in Biotechnology Faculty of Science Chulalongkorn University 254 Phayathai Road Bangkok 10330 Thailand2Department of Entomology Faculty of Agriculture Kasetsart University 50 NgamWong Wan Rd ChatuchakBangkok 10900 Thailand3Department of Biology Faculty of Science Chulalongkorn University 254 Phayathai Road Bangkok 10330 Thailand

Correspondence should be addressed to Chanpen Chanchao chanpencchulaacth

Received 8 November 2018 Revised 14 January 2019 Accepted 31 January 2019 Published 20 February 2019

Academic Editor David Roubik

Copyright copy 2019 Thitipan Meemongkolkiat et al This is an open access article distributed under the Creative CommonsAttribution License which permits unrestricted use distribution and reproduction in any medium provided the original work isproperly cited

Knowledge of the genetic diversity of Apis spp is important in order to provide a better understanding of breeding strategies thatrelate to the conservation of wild species and colony survival of farmed species Here honeybees of five Apis species were collectedfrom 12 provinces throughoutThailand After DNA extraction 28S rRNA nuclear (710 bp) and cytochrome b (cytb) mitochondrial(520 bp) gene fragments were sequenced Homologous sequences (nucleotide identity of over 95) were obtained from GeneBankusing the BLASTn algorithm aligned and analysed bymaximum likelihood and Bayesian inference phylogenetics For 28S rRNA alow genetic variation was detected within species (haplotype diversity ranging from 0212 to 0394) while 19 polymorphic sites weredetected between species Although the relative haplotype diversity was high a low nucleotide divergence was found (07) withmigratory species For cytb the sequence divergence ranged from 024 to 388 within species and 735 to 1307 between speciesThe divergence of cytb was higher than that of 28S rRNA A cerana showed two distinct clades between SouthernThailand and theother regions Groups of A cerana (Asian cavity-nesting) A mellifera (European cavity-nesting) A dorsata (giant open-nesting)and A florea and A andreniformis (dwarf bees) were defined in the 28S rRNA and cytb tree topologies

1 Introduction

The diversity of honeybees (Apis spp) is important in agri-cultural areas because they play a crucial role in pollinatinga wide variety of crop flowers [1] Large forest areas havebeen increasingly fragmented for agriculture and other landmanagement with an increase in larger monoculture andindustrial areas Furthermore climate change has becomemore serious These changes have likely affected the diversityof Apis spp since each species has different food plant andhabitat preferences

The diversity of Apis spp in Thailand has been reportedon over the past few decades Morphological variation wasreported in A florea [2] A andreniformis [3] and A dorsata[4] as well as the genetic variation of A dorsata and A mel-lifera [5 6] Studies on the genetic variation ofApis have usedmitochondrial DNA gene fragments [7 8] single nucleotide

polymorphisms [9] and allozymes [10] to differentiate thebreeds and provide a more powerful approach to determinethe degree of genetic variation among and between honeybeepopulations

Mitochondrial genes are the most popular molecularmarkers used to determine the degree of genetic diversityin insects [11] since the mitochondrial genome typicallyshows higher average mutation rates (10x) than the nucleargenome [12] due to nucleotide imbalance [13] and the smalleffective population size associated with effective haploidinheritance across eukaryotic taxa [14] For eusocial insectsthe polyandrous mating system seems to insert the highersubstitution rates in mitochondrial genes than in nucleargenes since the mitochondrial DNA is inherited exclusivelythrough the maternal lineage [15] Thus the inheritance ofmitochondrial DNA is transmitted via only one queen whilethe nuclear genome is derived from the parental line This

HindawiPsycheVolume 2019 Article ID 5823219 11 pageshttpsdoiorg10115520195823219

2 Psyche

Myanmar

Laos

CAMBODIA

98 102 106N

W

S

E

18

14

10

6

18

14

10

6

98 Malaysia 106

International boundary

0 100 200 km

Adorsata Aflorea Acerana Aandreniformis Amellifera

Figure 1 Sampling sites of Apis spp in Thailand

leads to a bias between the effective size of the nuclear andmitochondrial DNA pools [16]

However the use of only mitochondrial DNA as a geneticmarker to resolve phylogenetic studies can be insufficientand possibly lead to erroneous results Accordingly the phy-logenetic relationships between nuclear and mitochondrialgenomes have been compared and combined [17] Nucleargene introns appear to be acceptable for this analysis Usingbothmaximum likelihood (ML) and neighbour-joining anal-yses of the mitochondrial NADH dehydrogenase subunit 2(ND2) and nuclear elongation factor 1-120572 (EF1-120572) intron ofApisspp produced the same basic topologies of Apis spp [18]

Although the resolution of such a circumscribed taxo-nomic group as Apis spp has been reported uncertaintyremains concerning important aspects of the phylogeneticrelationships within honeybees and more data are requiredto reconfirm the status of honeybee diversity in ThailandThus this work aimed to investigate the level of geneticvariation and phylogenetic relationships within and amongApis spp populations in Thailand using partial sequencedata from the nuclear 28S rRNA region compared with the

mitochondrial cytochrome b gene (cytb) These data willprovide complementary information and a more satisfactorycomprehension of the phylogenetic variation and the originwithin and among Apis spp on an evolutionary time scale

2 Materials and Methods

21 Sample Collection A total of 76 colonies of honeybeescomprising Apis cerana (26) A mellifera (11) A dorsata(15) A florea (22) and A andreniformis (2) were collectedfrom 12 provinces in eastern central northern western andsouthern Thailand ranging from 8∘ 461015840 04410158401015840 N to 18∘ 45101584034110158401015840 N latitude and 98∘571015840 56310158401015840 E to 102∘ 271015840 09010158401015840 Elongitude during 2016 and 2017 (Figure 1 and Supplement 1)The samples were frozen with dry ice and stored at -20∘C inthe laboratory until used

22 Partial Nucleotide Sequences of the 28S rRNA and CytbGenes Total genomic DNA was extracted from the thoraxusing a QIAamp DNA Mini Kit (catalog 51304 Qiagen)The 28S rRNA and cytb gene fragments were amplified using

Psyche 3

Table 1 Frequency () of transitions and transversions of 28S rRNA and cytb

Species(No of sequences)

Transitions TransversionsAG TC AT AC GT GC

28S rRNAA cerana (26)lowast 0 0 1934 2394 2607 3068A florea (22)lowast 3421 0 1274 1541 1748 2015A dorsata (15)lowast 5279 0 913 1112 1249 1448A mellifera (11)lowast 4272 0 1094 1368 1493 1767All species (75)lowastlowast(includingA andreniformis) 3269 2973 722 885 994 1157

cytbA cerana (10)lowast 865 7941 453 268 329 144A florea (11)lowast 0 0 3830 2358 2642 1170A mellifera (9)lowast 008 9954 014 008 010 004All species (33)lowastlowast(includingA andreniformis and Adorsata)

2598 2856 1728 1035 1239 546

Remark ldquolowastrdquo and ldquolowastlowastrdquo show the nucleotide compositions within and among respectively Apis species from the different regions inThailand

the polymerase chain reaction (PCR) with the specific 28SrRNA (forward 51015840-AGAGAGAGTTCAAGAGTACGTG-31015840and reverse 51015840-TAGTTCACCATCTTTCGGGTCCC-31015840)and cytb (forward 51015840-TGAAATTTTGGATCAATTCTTGG-31015840 and reverse 51015840- TCCAAGAGGATTAGATGATCCAG-31015840) primer pairs [3 19] Each PCR amplification wasperformed in a 25 120583l final reaction volume comprising 125120583l 2X EmeraldAmp PCR master mix (catalog RR300ATakara) 1 120583l of each of the 10 120583M forward and reverse primerat least 30 ng of genomic DNA template and nuclease-freedistilled H

2O The thermocycling condition was 94∘C for

2 min 30 sec followed by 35 cycles of 94∘C for 1 min x ∘Cfor 1 min (x = 65∘C for 28S rRNA and 50∘C for cytb) and72∘C for 1 min and then a final 72∘C for 10 min Afterelectrophoresis on a 2 (wv) agarose gel at 80 V for 30 minPCR products were visualized by UV transillumination afterethidium bromide staining The expected product sizes ofthe amplified 28S rRNA and cytb fragments were 710 and520 bp respectively The amplified DNA fragments werepurified using a QIAquick Gel Purification Kit (catalog 28704 Qiagen) and then commercially direct sequenced atMacrogen Inc (Korea) The obtained sequences were usedto search for homologous sequences in the National Centrefor Biotechnology Information (NCBI) GenBank databaseusing the BLASTn algorithm and aligned

23 Sequences Analysis and Diversity Indices The MEGA7program version 70 [20] was used to generate the multiplesequence alignments and to obtain the number of substi-tutions between sequences nucleotide compositions andthe transition or transversion ratios for the 28S rRNA andcytb sequences The genetic variation within and betweenspecies was determined as the number of haplotypes (No)number of polymorphic sites (s) haplotype diversity (h) [21]average number of nucleotide differences (k) and nucleotidediversity (120587) [22] using the DNAsp V50 program [23]The sequence divergence () with Jukes and Cantor [24]

between and within species and the average number ofnucleotide differences were estimated using theMEGA7 V70and ARLEQUIN ver 3522 software [25]

24 Phylogenetic Analysis The ML and Bayesian inference(BI) methods were used to reconstruct the phylogeneticrelationships in the 28S rRNA and cytb sequencesThe best-fitmodels of nucleotide substitution were determined using theAkaike information criterion (AIC) forML [26] and Bayesianinformation criterion (BIC) for BI [27] as implemented in theKakusan 4 program [28] The ML analysis was performed inTreefinder with 1000 bootstrap replicates to estimate branchconfidence values [29] Tree topologies with bootstrap values(BV) of 70 or greater were regarded as sufficiently resolvedThe BI was performed using the MrBayes V31 program [30]using the selected model with four simultaneous chains runfor at least 1000000 generations starting from a random treewith tree sampling every 100 generations Both ML and BIgave a consensus of the phylogenetic tree

3 Results

31 Molecular Data and Sequence Analysis A total of 106sequences (73 sequences from 28S rRNA and 33 sequencesfrom cytb) were obtained and have been deposited in Gen-Bank with the accession codes as listed in Supplement 1 Theaverage nucleotide composition of the 28S rRNA sequencedid not vary much between the Apis species with an acrossall species average of A (1733) T (2105) G (3184) andC (2978) giving a high GC content of 616 (Supplement2) The average cytb sequence composition varied slightlybetween species with an average between all species of A(342) T (425) G (118) andC (114) with a highA+Tcontent of 772 (Supplements 2 and 3)

For the 28S rRNA 75 individualswere analysed and a totalof 612 positions were analysed in the final dataset Transver-sion sites were less frequent than transitions (Table 1) For

4 Psyche

Table 2 Molecular diversity indices within the 28S rRNA and cytb sequences in each Apis species in Thailand

Species(No of sequences) No S h (plusmn SD) k 120587plusmn SD)

28S rRNAA cerana (26) 2 1 0212 plusmn 0097 0212 000033 plusmn 000041A florea (22) 3 2 0394 plusmn 0119 0541 000085 plusmn 000086A dorsata (15) 4 10 0371 plusmn 0153 1333 000213 plusmn 000491

CytbA cerana (10) 8 16 0933 plusmn 0077 5956 00154 plusmn 00147A florea (11) 3 3 0473 plusmn 0162 1200 000201 plusmn 000250Number of haplotypes (No) number of polymorphic sites (s) haplotype diversity (h) [21] average number of nucleotide differences (k) and nucleotide diversity(120587) [22] with standard deviation (SD)

cytb 33 sequences were analysed including the 1st 2nd and3rd codons and noncoding positions separately giving a totalof 325 positions in the final dataset with 76 polymorphicsites of which 27 sites showed transitions 42 sites showedtransversions and seven sites carried both types of basesubstitution (Table 1)

Based on the sequence alignment within species the28S rRNA revealed low genetic diversity indices and lowgenetic variation among Asian Apis spp (Tables 2 and 3)The sequence divergence (pairwise comparison) over the28S rRNA within A cerana A florea and A dorsata wasderived from 612 638 and 627 bp respectively using 2622 and 15 sequences from across Thailand (Table 3) Thesequence alignment between species revealed 19 polymorphicsites comprising seven AG transitions six TC transitionstwo AT transversions three GT transversions and one GCtransversion The between species transition and transver-sion mutations in the 28S rRNA sequence were 6242and 3758 respectively Our results also demonstrated thatthe intraspecific sequence divergence was lower than theinterspecific sequence divergence (Table 4) For cytb onlyone and two sequences were obtained for A dorsata and Aandreniformis respectively The number of base differencesper sequence from385 positions is shown inTable 4 revealing33 different sequence variants in five species The sequencealignment between species revealed 76 polymorphic sites ofwhich there were 11 AG transitions 20 TC transitions 37AT transversions one GT transition one GC transitionand seven sites that carried both types of base substitutionThe transition and transversion mutations in cytb sequencesbetween species were 5454 and 5546 respectively Themean sequence divergence for all the variants was 89 Ingeneral the sequence divergence was high among speciesranging from 65 to 252 and lower between individuals ofthe same species

Based on the cytb sequence alignment the cytb analysisof A cerana (n = 10) revealed eight haplotypes from 16polymorphic sites that contained four AG transitions eightTC transitions and four AT transversions (Supplement4) The A cerana populations from the south of Thailandshowed a clear sequence divergence from the other regionsof Thailand The sequence divergence () within speciesand the average number of nucleotide differences observedin pairwise comparison among A cerana are shown in

Table 5 The highest sequence divergence was detected whenHAP1ndashHAP6 (including A cerana from east central andwest regions) was compared with HAP7 and HAP8 (fromsouth regions) (Supplement 4) These results suggest that Acerana in southernThailand were from different populationsthan those in the other regions of Thailand

In A florea the pairwise comparison of the cytbsequences suggested thatA florea from two colonies from SA(central regions)were slightly different from the other regions(Table 5) Interestingly the sequence divergence amongA cerana and A mellifera (1022) was lower than thatamong A dorsata A florea and A andreniformis (Table 4)Meanwhile A florea showed less sequence divergence fromA andreniformis (735) than fromA cerana andA mellifera(1219 and 1307 respectively)

32 Phylogenetic Analysis Thebest-fit model of evolution forconstructing the phylogenetic tree by ML analysis was foundby the AIC to be TVM + gamma while that for the BI treeunder the BIC was HKY85 + homogeneous The ML treesconstructed for the 28S rRNA and cytb gene sequences wereessentially the same as the corresponding BI tree and so onlythe ML tree topologies are illustrated here (Figures 2 and 3)

For the 28S rRNA sequences (612ndash638 bp) the phy-logenetic relationships within A cerana A florea and Adorsata from all geographical regions showed no evidenceof geographical clustering in the phylogenetic tree althoughthe nonnative A mellifera showed some divergence withinthe species Overall a low genetic variation within all AsianApis spp was noticed based on the 75 sequences of this studyplus two sequences of A andreniformis from GenBank andusing Trigona collina as an outgroup (Figure 2)The Apis sppwere subdivided into four major clades (BV range from 50to 98 PP range from 060 to 095) strongly supportingthe monophyly of Apis species in Thailand (BV = 100PP = 10) Interestingly all four Apis species were clearlygrouped into their respective species clade except for the twodwarf honeybee speciesA florea andA andreniformis whichwere not completely resolved (Figure 2) This reflects the lowsequence divergence between these dwarf species whichwerebasal in the Apis groups (Figure 2)

For the cytb region (383 bp in length) the ML phyloge-netic dendrograms based on 33 sequences from this studyfrom all Apis species including only one and two sequences

Psyche 5

Table3Estim

ated

sequ

ence

divergence

over

the2

8SrRNA

sequ

ence

pairs

with

ineach

speciesSequ

ence

divergence

()w

ithin

speciesisa

bove

thed

iagonalwhilethea

verage

numbero

fnu

cleotided

ifferencesisb

elowthed

iagonal

Acerana

Central(Sam

utsong

kram

)West(Ka

nchanabu

riRa

tchabu

ri)North

(Chiangm

aiN

an)

East(C

hanthabu

ri)South(C

humph

on

Suratth

aniNakho

nsi

tham

marat)

Central

-003

007

004

004

West

025

-006

00

North

045

040

-006

006

East

025

0040

-0

South

025

0040

0-

Aflorea

Central(Sam

utsong

kram

Nakho

nsaw

an)

West(Ka

nchanabu

riRa

tchabu

ri)Ea

st(C

hanthabu

ri)South(C

humph

on)

North

(Pichit)

Central

-008

003

003

001

West

057

-007

007

001

East

020

050

-000

009

South

020

050

000

-009

North

064

077

060

060

-

Adorsata

Central(Sam

utsong

kram

Nakho

nsaw

an)

West(Ka

nchanabu

ri)North

(Nan)

South(Suratthani)

Central

-024

029

024

West

150

-005

000

North

183

033

-005

South

150

000

033

-

6 Psyche

Table 4 Estimated sequence divergence in the 28S rRNA and cytb sequence pairs between species The pairwise comparison uses 612 and383 bp of 28SrRNA and cytb sequence respectively

28S rRNA A cerana A mellifera A dorsata A florea A andreniformisA cerana - 055 061 120 117A mellifera 339 - 081 105 105A dorsata 376 498 - 113 109A florea 729 642 686 - 003A andreniformis 711 636 667 023 -cytb A mellifera A andreniformis A dorsata A cerana A floreaA mellifera - 1148 1058 1022 1307A andreniformis 3461 - 1289 1249 735A dorsata 3400 3850 - 1259 1345A cerana 3250 3740 3830 - 1219A florea 3900 2277 4200 3657 -Sequence divergence () within species is above the diagonal while the average number of nucleotide differences is below the diagonal

Table 5 Estimated sequence divergence over cytb sequence pairs (383 bp) within A cerana (10 samples) and A florea (11 samples)

A cerana Chantaburi (East) Kanchanaburi (West) Samutsongkram (Central) Suratthani (South)Chantaburi (East) - 084 105 388Kanchanaburi (West) 322 - 026 306Samutsongkram(Central) 400 100 - 279

Suratthani (South) 1450 1150 105 -A florea Chantaburi (East) Chumphon (South) Pichit (North) Ratchaburi (West) Samutsongkram (Central)Chantaburi (East) - 0 024 0 074Chumphon (South) 0 - 024 0 074Pichit(North) 100 100 - 024 049

Ratchaburi (West) 0 0 100 - 074Samutsongkram(Central) 300 300 200 300 -

Sequence divergence () with Jukes and Cantor between species is above the diagonal and the average number of nucleotide differences is below the diagonal

of A dorsata and A andreniformis respectively plus fivesequences of Apis spp from GenBank were constructedwith T collina as the outgroup The obtained tree revealedfour major clades (Figure 3) where each Apis species wasclearly separated from the others including A florea and Aandreniformis Although A florea in Thailand was dividedinto small subgroups with three haplotypes in the populationfrom Samutsongkram (central region) there was a low levelof genetic variation in cytb in the A florea populations Thebasal group was A dorsata For A cerana (clade 1) the 10A cerana samples in Thailand of this study plus three Acerana sequences from GenBank were subdivided into thetwo large groups of SouthernThailand and the other regionsplus one subgroup in the Eastern region (Figure 3) The Acerana populations from Suratthani were clearly separatedfrom the other regions (BV = 100 PP = 095) while thesamples from Eastern Thailand showed a slight differencefrom the other regions although this had only weak support(BV = 64) This low support might reflect the nucleotidedifferences among the haplotypes (Supplement 4) that arestill placed in the same clade These results suggested that A

cerana in Southern Thailand are from a different populationto those in the other regions of Thailand

4 Discussion

A total of 73 28S rRNA and 33 cytb sequences were obtainedThe overall sequence diversity of both gene fragments withinand among species was low especially for 28S rRNA thatshowed only a slight sequence divergence for all haplotypes(07)This suggests that there is little variation in 28S rRNAwithinApis spp inThailandThe 28S rRNA gene is often usedas a representative nuclear gene at the family-species leveland above where for example it was reported that the mostconserved 28S-D3 region could be used to define Encarsiaspp [31]

However using 28S rRNA alone may be insufficientespecially at the species level and below In addition to 28SrRNA other nuclear genes have been used like itpr [32]and the EF1-120572 intron [18] in order to resolve the familylevel phylogenies in bees and other insects Neverthelessthe nuclear EF1-120572 intron has been reported to be useful for

Psyche 7

ACCM1

ACCM2ACSA3ACCM3ACCN1ACCN3ACCP1ACCP2ACNS2ACNS3ACRA1ACRA2ACCP3ACKA1

ACN2

ACKA2ACKA3ACSA1ACNS1ACSA2ACSU1ACSU2ACSU3

ADKA1ADKA2ADKA3

ADN2ADN3ADNA1

ADN98081

87078

A2ADN1

ADSA1ADSA2ADSA3 ADSA4ADSU1ADSU2ADSU3

ACCN2 ACN1ACSA47607763072

AA1

AA2

AAKAAFCP2AFKA1

AFSA3AFPJ5

AFPJ4AFRA3AFSA3

AFSA1AFSA2

AFPJ5AFRA2

56067

63065AFCN1AFNA1AFNA2AFCN2AFCN3AFCP1AFCP3AFPJ1AFPJ2AFPJ3AFRA1

50095

Tetragonilla collina (FJ042138)

002

50060

84067

98091

8907070063 AMCM1

AMP2AMCM3

AMP1AMP3

AMCM2

AMN1

AMCM4

AMCM5

AMN2

AMN3

Clade II

Clade III

Clade IV

Clade I

Figure 2 Phylogenetic relationship of honeybee species in Thailand based upon the 28S rRNA sequences by ML analysis Trigonacollina (accession FJ042138) was utilized as the outgroup Node support inferred from Bayesian PP and ML BV is shown Sample(speciesgeographic location) codes are as in Supplement 1

8 Psyche

Apis dorsata (KC294229)

ADKA

Apis andreniformis (EU040176)Apis andreniformis (EU040168)Apis andreniformis (EU040165)AACNAAKA

91082

100097

AFCN1AFCN2AFCP1AFCP2AFCP3AFPJ1AFSA1AFSA2AFPJ2AFRA1AFRA2Apis florea (JX982136)

Apis florea (KC170303)

90092

100100

AMCM1AMN2AMCM2AMCM3AMP3AMP2AMN1AMN3AMP1

100093

ACCN1ACCN3ACCN2ACKA1ACKA2ACKA3ACSA1ACSA2

100064

100095

Apis cerana (1) (AP017941)Apis cerana (2) (AP017941)

Apis cerana (GQ162109)ACSU1

ACSU2

Trigono collino (AY575081)

300

Clade II

Clade IV

Clade IV-1

Clade IV-2

Clade I

Clade III

Clade II-1

Clade II-2

Figure 3 Phylogenetic relationship of honeybee species in Thailand based upon cytb sequences by BI and including Apis spp fromGenBank (accession KC294229 EU040176 EU040168 EU040165 JX982136 KC170303 AP017941 AP017942 and GQ162109) Trigonacollina (accession AY575081) was utilized as the outgroup Node support as Bayesian PP and ML BV is shown Sample (speciesgeographiclocation) codes are as in Supplement 1

classifying honeybees between species but not for resolvingwithin species [18] Likewise it was reported that the 28SrRNA was unlikely to recover deep divergences at the familylevel in bees [33] In contrast in combination with othergenes such as opsin cytb and 16S rRNA 28S rRNA appearedto contribute to resolving basal bee divergence

In this study the number of transversions in cytb washigher than the number of transitions in each Apis species(Table 1) This is in agreement with the transitional biasreported previously [18] Within A cerana the cytb sequencedivergence ranged from 026 to 388 but interestingly washigher in the southern populations than in the other regions

Psyche 9

of Thailand (Table 5) The phylogenetic results are consistentwith the molecular diversity indices of A cerana where Acerana populations in southernThailandwere separated fromthose in the East West and Central regions (Figure 3) Theresult could be explained by the biogeography of A ceranaIn addition the data concur with the previous PCR-RFLPanalysis [34]

For A florea only a very low cytb sequence divergencewas detected (0ndash074) with only one haplotype Onlytwo colonies from Samutsongkram province were slightlydifferent from the others (Figure 3)

Overall the sequence divergence of cytb (89 plusmn 50) washigher than that for 28S rRNA (07 plusmn 05) consistent witha higher evolutionary rate for mitochondrial than nucleargenomes [11] Although the mitochondrial cytb showed ahigher rate of variation it still showed a low variation in theseApis spp among the different geographical regions This mayreflect that Apis spp are highly eusocial and polyandrouswhich reduces the substitution rate and counteracts the effectof genetic drift by increasing the effective population size[35] The ancestral genome in diverging populations will bemaintained and so the genetic distances between populationsare reduced [36] The other possible factor leading to alow genetic variation is migration behaviour The migrationdistance ofA dorsata has been reported to be 50 kmwide andup to 200 km [37] InA cerana although they havemigratoryswarms flying dozens of km [38] they have a strong tendencyto swarm 5ndash6 times a year and are well adapted to new areassince it is more resistant to some diseases and pest problemsin each area [39]

TheMLandBI phylogenetic analyses of the 28S rRNA andcytb sequences showed the same topologies with four majorclades among Apis spp (for ML Figures 2 and 3) Using Tcollina as the outgroup the 28S rRNA tree placed the dwarfbees in a basal position within the genus while A dorsatawas located in the basal position in the cytb tree consistentwith previous analysis [40] Although both the cytb and28S rRNA phylogenetic trees in this study could not clearlyresolve the relationship between these Apis species theycould separate A dorsata from others (Figures 2 and 3) Aclose relationship has been reported between A dorsata andA cerana in the 28S rRNA tree and molecular indices [32]Although A cerana and A mellifera have been reported tobe closely related to each other [18] they have ecological andbehavioural differences [41] In this study the 28S rRNA andcytb topologies clearly separated A cerana and A mellifera inThailand from each other although they were still closer toeach other compared to the other Apis species

Despite the fact that A dorsata and A florea are open-nesting bees A dorsata was closer to A cerana and Amellifera than to A florea and A andreniformis in this studyThis can be explained by the evolution of behaviour in Apiswhere cavity-nesting is secondarily derived [42] MoreoverA dorsata has a buzzing dance in daylight and at night whileA florea has only a silent waggle dance and cavity-nestingbees adapted this buzzing waggle dance in a dark condi-tion Thus a loss of visual cues was compensated for withsound

5 Conclusion

Overall the genetic variation of Apis spp from 64 popula-tions within 12 provinces in Thailand showed low geneticdiversity indices and variation within and between speciesThis may indicate that habitat fragmentation monocropagriculture industrial management and climate changeshave not affected the diversity of Apis spp in Thailand yetbut clearly further work is of vital importance in establishingthe actual genetic diversity and relationship within Apissubspecies including the analysis of other genes for betterresolution

Data Availability

The authors confirm that all data in the manuscript are freelyavailable

Conflicts of Interest

The authors declare that there are no conflicts of interestregarding the publication of this paper

Acknowledgments

This work was financially supported by the Science Achieve-ment Scholarship of Thailand and the 90th Anniversary ofChulalongkorn University Fund (RatchadaphiseksomphotEndowment Fund)

Supplementary Materials

Supplement 1 detail of sample collection andGenBank acces-sion numbers Supplement 2 frequency () of the nucleotidecompositions of 28S rRNA and cytb regions Supplement 3mean frequency () for base compositions at different codonpositions in the cytb gene fragment Supplement 4 the 16polymorphic sites among eight cytb haplotypes of A ceranain Thailand Only positions that are different from haplotypeHAP1 are indicated (Supplementary Materials)

References

[1] R Winfree J R Reilly I Bartomeus D P Cariveau NM Williams and J Gibbs ldquoSpecies turnover promotes theimportance of bee diversity for crop pollination at regionalscalesrdquo Science vol 359 no 6377 pp 791ndash793 2018

[2] S Wongsiri C Lekprayoon R Thapa et al ldquoComparativebiology of Apis andreniformis and Apis florea in Thailandrdquo BeeWorld vol 78 no 1 pp 23ndash35 1997

[3] A Rattanawannee C Chanchao and S Wongsiri ldquoMorpho-metric and genetic variation of small dwarf honeybees Apisandreniformis Smith 1858 in Thailandrdquo Insect Science vol 14no 6 pp 451ndash460 2007

[4] B P Oldroyd K E Osborne and M Mardan ldquoColony related-ness in aggregations of Apis dorsata Fabricius (HymenopteraApidae)rdquo Insectes Sociaux vol 47 no 1 pp 94-95 2000

[5] A Rattanawannee C Chanchao J Lim S Wongsiri and BP Oldroyd ldquoGenetic structure of a giant honey bee (Apis

Psyche 3

Table 1 Frequency () of transitions and transversions of 28S rRNA and cytb

Species(No of sequences)

Transitions TransversionsAG TC AT AC GT GC

28S rRNAA cerana (26)lowast 0 0 1934 2394 2607 3068A florea (22)lowast 3421 0 1274 1541 1748 2015A dorsata (15)lowast 5279 0 913 1112 1249 1448A mellifera (11)lowast 4272 0 1094 1368 1493 1767All species (75)lowastlowast(includingA andreniformis) 3269 2973 722 885 994 1157

cytbA cerana (10)lowast 865 7941 453 268 329 144A florea (11)lowast 0 0 3830 2358 2642 1170A mellifera (9)lowast 008 9954 014 008 010 004All species (33)lowastlowast(includingA andreniformis and Adorsata)

2598 2856 1728 1035 1239 546

Remark ldquolowastrdquo and ldquolowastlowastrdquo show the nucleotide compositions within and among respectively Apis species from the different regions inThailand

the polymerase chain reaction (PCR) with the specific 28SrRNA (forward 51015840-AGAGAGAGTTCAAGAGTACGTG-31015840and reverse 51015840-TAGTTCACCATCTTTCGGGTCCC-31015840)and cytb (forward 51015840-TGAAATTTTGGATCAATTCTTGG-31015840 and reverse 51015840- TCCAAGAGGATTAGATGATCCAG-31015840) primer pairs [3 19] Each PCR amplification wasperformed in a 25 120583l final reaction volume comprising 125120583l 2X EmeraldAmp PCR master mix (catalog RR300ATakara) 1 120583l of each of the 10 120583M forward and reverse primerat least 30 ng of genomic DNA template and nuclease-freedistilled H

2O The thermocycling condition was 94∘C for

2 min 30 sec followed by 35 cycles of 94∘C for 1 min x ∘Cfor 1 min (x = 65∘C for 28S rRNA and 50∘C for cytb) and72∘C for 1 min and then a final 72∘C for 10 min Afterelectrophoresis on a 2 (wv) agarose gel at 80 V for 30 minPCR products were visualized by UV transillumination afterethidium bromide staining The expected product sizes ofthe amplified 28S rRNA and cytb fragments were 710 and520 bp respectively The amplified DNA fragments werepurified using a QIAquick Gel Purification Kit (catalog 28704 Qiagen) and then commercially direct sequenced atMacrogen Inc (Korea) The obtained sequences were usedto search for homologous sequences in the National Centrefor Biotechnology Information (NCBI) GenBank databaseusing the BLASTn algorithm and aligned

23 Sequences Analysis and Diversity Indices The MEGA7program version 70 [20] was used to generate the multiplesequence alignments and to obtain the number of substi-tutions between sequences nucleotide compositions andthe transition or transversion ratios for the 28S rRNA andcytb sequences The genetic variation within and betweenspecies was determined as the number of haplotypes (No)number of polymorphic sites (s) haplotype diversity (h) [21]average number of nucleotide differences (k) and nucleotidediversity (120587) [22] using the DNAsp V50 program [23]The sequence divergence () with Jukes and Cantor [24]

between and within species and the average number ofnucleotide differences were estimated using theMEGA7 V70and ARLEQUIN ver 3522 software [25]

24 Phylogenetic Analysis The ML and Bayesian inference(BI) methods were used to reconstruct the phylogeneticrelationships in the 28S rRNA and cytb sequencesThe best-fitmodels of nucleotide substitution were determined using theAkaike information criterion (AIC) forML [26] and Bayesianinformation criterion (BIC) for BI [27] as implemented in theKakusan 4 program [28] The ML analysis was performed inTreefinder with 1000 bootstrap replicates to estimate branchconfidence values [29] Tree topologies with bootstrap values(BV) of 70 or greater were regarded as sufficiently resolvedThe BI was performed using the MrBayes V31 program [30]using the selected model with four simultaneous chains runfor at least 1000000 generations starting from a random treewith tree sampling every 100 generations Both ML and BIgave a consensus of the phylogenetic tree

3 Results

31 Molecular Data and Sequence Analysis A total of 106sequences (73 sequences from 28S rRNA and 33 sequencesfrom cytb) were obtained and have been deposited in Gen-Bank with the accession codes as listed in Supplement 1 Theaverage nucleotide composition of the 28S rRNA sequencedid not vary much between the Apis species with an acrossall species average of A (1733) T (2105) G (3184) andC (2978) giving a high GC content of 616 (Supplement2) The average cytb sequence composition varied slightlybetween species with an average between all species of A(342) T (425) G (118) andC (114) with a highA+Tcontent of 772 (Supplements 2 and 3)

For the 28S rRNA 75 individualswere analysed and a totalof 612 positions were analysed in the final dataset Transver-sion sites were less frequent than transitions (Table 1) For

4 Psyche

Table 2 Molecular diversity indices within the 28S rRNA and cytb sequences in each Apis species in Thailand

Species(No of sequences) No S h (plusmn SD) k 120587plusmn SD)

28S rRNAA cerana (26) 2 1 0212 plusmn 0097 0212 000033 plusmn 000041A florea (22) 3 2 0394 plusmn 0119 0541 000085 plusmn 000086A dorsata (15) 4 10 0371 plusmn 0153 1333 000213 plusmn 000491

CytbA cerana (10) 8 16 0933 plusmn 0077 5956 00154 plusmn 00147A florea (11) 3 3 0473 plusmn 0162 1200 000201 plusmn 000250Number of haplotypes (No) number of polymorphic sites (s) haplotype diversity (h) [21] average number of nucleotide differences (k) and nucleotide diversity(120587) [22] with standard deviation (SD)

cytb 33 sequences were analysed including the 1st 2nd and3rd codons and noncoding positions separately giving a totalof 325 positions in the final dataset with 76 polymorphicsites of which 27 sites showed transitions 42 sites showedtransversions and seven sites carried both types of basesubstitution (Table 1)

Based on the sequence alignment within species the28S rRNA revealed low genetic diversity indices and lowgenetic variation among Asian Apis spp (Tables 2 and 3)The sequence divergence (pairwise comparison) over the28S rRNA within A cerana A florea and A dorsata wasderived from 612 638 and 627 bp respectively using 2622 and 15 sequences from across Thailand (Table 3) Thesequence alignment between species revealed 19 polymorphicsites comprising seven AG transitions six TC transitionstwo AT transversions three GT transversions and one GCtransversion The between species transition and transver-sion mutations in the 28S rRNA sequence were 6242and 3758 respectively Our results also demonstrated thatthe intraspecific sequence divergence was lower than theinterspecific sequence divergence (Table 4) For cytb onlyone and two sequences were obtained for A dorsata and Aandreniformis respectively The number of base differencesper sequence from385 positions is shown inTable 4 revealing33 different sequence variants in five species The sequencealignment between species revealed 76 polymorphic sites ofwhich there were 11 AG transitions 20 TC transitions 37AT transversions one GT transition one GC transitionand seven sites that carried both types of base substitutionThe transition and transversion mutations in cytb sequencesbetween species were 5454 and 5546 respectively Themean sequence divergence for all the variants was 89 Ingeneral the sequence divergence was high among speciesranging from 65 to 252 and lower between individuals ofthe same species

Based on the cytb sequence alignment the cytb analysisof A cerana (n = 10) revealed eight haplotypes from 16polymorphic sites that contained four AG transitions eightTC transitions and four AT transversions (Supplement4) The A cerana populations from the south of Thailandshowed a clear sequence divergence from the other regionsof Thailand The sequence divergence () within speciesand the average number of nucleotide differences observedin pairwise comparison among A cerana are shown in

Table 5 The highest sequence divergence was detected whenHAP1ndashHAP6 (including A cerana from east central andwest regions) was compared with HAP7 and HAP8 (fromsouth regions) (Supplement 4) These results suggest that Acerana in southernThailand were from different populationsthan those in the other regions of Thailand

In A florea the pairwise comparison of the cytbsequences suggested thatA florea from two colonies from SA(central regions)were slightly different from the other regions(Table 5) Interestingly the sequence divergence amongA cerana and A mellifera (1022) was lower than thatamong A dorsata A florea and A andreniformis (Table 4)Meanwhile A florea showed less sequence divergence fromA andreniformis (735) than fromA cerana andA mellifera(1219 and 1307 respectively)

32 Phylogenetic Analysis Thebest-fit model of evolution forconstructing the phylogenetic tree by ML analysis was foundby the AIC to be TVM + gamma while that for the BI treeunder the BIC was HKY85 + homogeneous The ML treesconstructed for the 28S rRNA and cytb gene sequences wereessentially the same as the corresponding BI tree and so onlythe ML tree topologies are illustrated here (Figures 2 and 3)

For the 28S rRNA sequences (612ndash638 bp) the phy-logenetic relationships within A cerana A florea and Adorsata from all geographical regions showed no evidenceof geographical clustering in the phylogenetic tree althoughthe nonnative A mellifera showed some divergence withinthe species Overall a low genetic variation within all AsianApis spp was noticed based on the 75 sequences of this studyplus two sequences of A andreniformis from GenBank andusing Trigona collina as an outgroup (Figure 2)The Apis sppwere subdivided into four major clades (BV range from 50to 98 PP range from 060 to 095) strongly supportingthe monophyly of Apis species in Thailand (BV = 100PP = 10) Interestingly all four Apis species were clearlygrouped into their respective species clade except for the twodwarf honeybee speciesA florea andA andreniformis whichwere not completely resolved (Figure 2) This reflects the lowsequence divergence between these dwarf species whichwerebasal in the Apis groups (Figure 2)

For the cytb region (383 bp in length) the ML phyloge-netic dendrograms based on 33 sequences from this studyfrom all Apis species including only one and two sequences

Psyche 5

Table3Estim

ated

sequ

ence

divergence

over

the2

8SrRNA

sequ

ence

pairs

with

ineach

speciesSequ

ence

divergence

()w

ithin

speciesisa

bove

thed

iagonalwhilethea

verage

numbero

fnu

cleotided

ifferencesisb

elowthed

iagonal

Acerana

Central(Sam

utsong

kram

)West(Ka

nchanabu

riRa

tchabu

ri)North

(Chiangm

aiN

an)

East(C

hanthabu

ri)South(C

humph

on

Suratth

aniNakho

nsi

tham

marat)

Central

-003

007

004

004

West

025

-006

00

North

045

040

-006

006

East

025

0040

-0

South

025

0040

0-

Aflorea

Central(Sam

utsong

kram

Nakho

nsaw

an)

West(Ka

nchanabu

riRa

tchabu

ri)Ea

st(C

hanthabu

ri)South(C

humph

on)

North

(Pichit)

Central

-008

003

003

001

West

057

-007

007

001

East

020

050

-000

009

South

020

050

000

-009

North

064

077

060

060

-

Adorsata

Central(Sam

utsong

kram

Nakho

nsaw

an)

West(Ka

nchanabu

ri)North

(Nan)

South(Suratthani)

Central

-024

029

024

West

150

-005

000

North

183

033

-005

South

150

000

033

-

6 Psyche

Table 4 Estimated sequence divergence in the 28S rRNA and cytb sequence pairs between species The pairwise comparison uses 612 and383 bp of 28SrRNA and cytb sequence respectively

28S rRNA A cerana A mellifera A dorsata A florea A andreniformisA cerana - 055 061 120 117A mellifera 339 - 081 105 105A dorsata 376 498 - 113 109A florea 729 642 686 - 003A andreniformis 711 636 667 023 -cytb A mellifera A andreniformis A dorsata A cerana A floreaA mellifera - 1148 1058 1022 1307A andreniformis 3461 - 1289 1249 735A dorsata 3400 3850 - 1259 1345A cerana 3250 3740 3830 - 1219A florea 3900 2277 4200 3657 -Sequence divergence () within species is above the diagonal while the average number of nucleotide differences is below the diagonal

Table 5 Estimated sequence divergence over cytb sequence pairs (383 bp) within A cerana (10 samples) and A florea (11 samples)

A cerana Chantaburi (East) Kanchanaburi (West) Samutsongkram (Central) Suratthani (South)Chantaburi (East) - 084 105 388Kanchanaburi (West) 322 - 026 306Samutsongkram(Central) 400 100 - 279

Suratthani (South) 1450 1150 105 -A florea Chantaburi (East) Chumphon (South) Pichit (North) Ratchaburi (West) Samutsongkram (Central)Chantaburi (East) - 0 024 0 074Chumphon (South) 0 - 024 0 074Pichit(North) 100 100 - 024 049

Ratchaburi (West) 0 0 100 - 074Samutsongkram(Central) 300 300 200 300 -

Sequence divergence () with Jukes and Cantor between species is above the diagonal and the average number of nucleotide differences is below the diagonal

of A dorsata and A andreniformis respectively plus fivesequences of Apis spp from GenBank were constructedwith T collina as the outgroup The obtained tree revealedfour major clades (Figure 3) where each Apis species wasclearly separated from the others including A florea and Aandreniformis Although A florea in Thailand was dividedinto small subgroups with three haplotypes in the populationfrom Samutsongkram (central region) there was a low levelof genetic variation in cytb in the A florea populations Thebasal group was A dorsata For A cerana (clade 1) the 10A cerana samples in Thailand of this study plus three Acerana sequences from GenBank were subdivided into thetwo large groups of SouthernThailand and the other regionsplus one subgroup in the Eastern region (Figure 3) The Acerana populations from Suratthani were clearly separatedfrom the other regions (BV = 100 PP = 095) while thesamples from Eastern Thailand showed a slight differencefrom the other regions although this had only weak support(BV = 64) This low support might reflect the nucleotidedifferences among the haplotypes (Supplement 4) that arestill placed in the same clade These results suggested that A

cerana in Southern Thailand are from a different populationto those in the other regions of Thailand

4 Discussion

A total of 73 28S rRNA and 33 cytb sequences were obtainedThe overall sequence diversity of both gene fragments withinand among species was low especially for 28S rRNA thatshowed only a slight sequence divergence for all haplotypes(07)This suggests that there is little variation in 28S rRNAwithinApis spp inThailandThe 28S rRNA gene is often usedas a representative nuclear gene at the family-species leveland above where for example it was reported that the mostconserved 28S-D3 region could be used to define Encarsiaspp [31]

However using 28S rRNA alone may be insufficientespecially at the species level and below In addition to 28SrRNA other nuclear genes have been used like itpr [32]and the EF1-120572 intron [18] in order to resolve the familylevel phylogenies in bees and other insects Neverthelessthe nuclear EF1-120572 intron has been reported to be useful for

Psyche 7

ACCM1

ACCM2ACSA3ACCM3ACCN1ACCN3ACCP1ACCP2ACNS2ACNS3ACRA1ACRA2ACCP3ACKA1

ACN2

ACKA2ACKA3ACSA1ACNS1ACSA2ACSU1ACSU2ACSU3

ADKA1ADKA2ADKA3

ADN2ADN3ADNA1

ADN98081

87078

A2ADN1

ADSA1ADSA2ADSA3 ADSA4ADSU1ADSU2ADSU3

ACCN2 ACN1ACSA47607763072

AA1

AA2

AAKAAFCP2AFKA1

AFSA3AFPJ5

AFPJ4AFRA3AFSA3

AFSA1AFSA2

AFPJ5AFRA2

56067

63065AFCN1AFNA1AFNA2AFCN2AFCN3AFCP1AFCP3AFPJ1AFPJ2AFPJ3AFRA1

50095

Tetragonilla collina (FJ042138)

002

50060

84067

98091

8907070063 AMCM1

AMP2AMCM3

AMP1AMP3

AMCM2

AMN1

AMCM4

AMCM5

AMN2

AMN3

Clade II

Clade III

Clade IV

Clade I

Figure 2 Phylogenetic relationship of honeybee species in Thailand based upon the 28S rRNA sequences by ML analysis Trigonacollina (accession FJ042138) was utilized as the outgroup Node support inferred from Bayesian PP and ML BV is shown Sample(speciesgeographic location) codes are as in Supplement 1

8 Psyche

Apis dorsata (KC294229)

ADKA

Apis andreniformis (EU040176)Apis andreniformis (EU040168)Apis andreniformis (EU040165)AACNAAKA

91082

100097

AFCN1AFCN2AFCP1AFCP2AFCP3AFPJ1AFSA1AFSA2AFPJ2AFRA1AFRA2Apis florea (JX982136)

Apis florea (KC170303)

90092

100100

AMCM1AMN2AMCM2AMCM3AMP3AMP2AMN1AMN3AMP1

100093

ACCN1ACCN3ACCN2ACKA1ACKA2ACKA3ACSA1ACSA2

100064

100095

Apis cerana (1) (AP017941)Apis cerana (2) (AP017941)

Apis cerana (GQ162109)ACSU1

ACSU2

Trigono collino (AY575081)

300

Clade II

Clade IV

Clade IV-1

Clade IV-2

Clade I

Clade III

Clade II-1

Clade II-2

Figure 3 Phylogenetic relationship of honeybee species in Thailand based upon cytb sequences by BI and including Apis spp fromGenBank (accession KC294229 EU040176 EU040168 EU040165 JX982136 KC170303 AP017941 AP017942 and GQ162109) Trigonacollina (accession AY575081) was utilized as the outgroup Node support as Bayesian PP and ML BV is shown Sample (speciesgeographiclocation) codes are as in Supplement 1

classifying honeybees between species but not for resolvingwithin species [18] Likewise it was reported that the 28SrRNA was unlikely to recover deep divergences at the familylevel in bees [33] In contrast in combination with othergenes such as opsin cytb and 16S rRNA 28S rRNA appearedto contribute to resolving basal bee divergence

In this study the number of transversions in cytb washigher than the number of transitions in each Apis species(Table 1) This is in agreement with the transitional biasreported previously [18] Within A cerana the cytb sequencedivergence ranged from 026 to 388 but interestingly washigher in the southern populations than in the other regions

Psyche 9

of Thailand (Table 5) The phylogenetic results are consistentwith the molecular diversity indices of A cerana where Acerana populations in southernThailandwere separated fromthose in the East West and Central regions (Figure 3) Theresult could be explained by the biogeography of A ceranaIn addition the data concur with the previous PCR-RFLPanalysis [34]

For A florea only a very low cytb sequence divergencewas detected (0ndash074) with only one haplotype Onlytwo colonies from Samutsongkram province were slightlydifferent from the others (Figure 3)

Overall the sequence divergence of cytb (89 plusmn 50) washigher than that for 28S rRNA (07 plusmn 05) consistent witha higher evolutionary rate for mitochondrial than nucleargenomes [11] Although the mitochondrial cytb showed ahigher rate of variation it still showed a low variation in theseApis spp among the different geographical regions This mayreflect that Apis spp are highly eusocial and polyandrouswhich reduces the substitution rate and counteracts the effectof genetic drift by increasing the effective population size[35] The ancestral genome in diverging populations will bemaintained and so the genetic distances between populationsare reduced [36] The other possible factor leading to alow genetic variation is migration behaviour The migrationdistance ofA dorsata has been reported to be 50 kmwide andup to 200 km [37] InA cerana although they havemigratoryswarms flying dozens of km [38] they have a strong tendencyto swarm 5ndash6 times a year and are well adapted to new areassince it is more resistant to some diseases and pest problemsin each area [39]

TheMLandBI phylogenetic analyses of the 28S rRNA andcytb sequences showed the same topologies with four majorclades among Apis spp (for ML Figures 2 and 3) Using Tcollina as the outgroup the 28S rRNA tree placed the dwarfbees in a basal position within the genus while A dorsatawas located in the basal position in the cytb tree consistentwith previous analysis [40] Although both the cytb and28S rRNA phylogenetic trees in this study could not clearlyresolve the relationship between these Apis species theycould separate A dorsata from others (Figures 2 and 3) Aclose relationship has been reported between A dorsata andA cerana in the 28S rRNA tree and molecular indices [32]Although A cerana and A mellifera have been reported tobe closely related to each other [18] they have ecological andbehavioural differences [41] In this study the 28S rRNA andcytb topologies clearly separated A cerana and A mellifera inThailand from each other although they were still closer toeach other compared to the other Apis species

Despite the fact that A dorsata and A florea are open-nesting bees A dorsata was closer to A cerana and Amellifera than to A florea and A andreniformis in this studyThis can be explained by the evolution of behaviour in Apiswhere cavity-nesting is secondarily derived [42] MoreoverA dorsata has a buzzing dance in daylight and at night whileA florea has only a silent waggle dance and cavity-nestingbees adapted this buzzing waggle dance in a dark condi-tion Thus a loss of visual cues was compensated for withsound

5 Conclusion

Overall the genetic variation of Apis spp from 64 popula-tions within 12 provinces in Thailand showed low geneticdiversity indices and variation within and between speciesThis may indicate that habitat fragmentation monocropagriculture industrial management and climate changeshave not affected the diversity of Apis spp in Thailand yetbut clearly further work is of vital importance in establishingthe actual genetic diversity and relationship within Apissubspecies including the analysis of other genes for betterresolution

Data Availability

The authors confirm that all data in the manuscript are freelyavailable

Conflicts of Interest

The authors declare that there are no conflicts of interestregarding the publication of this paper

Acknowledgments

This work was financially supported by the Science Achieve-ment Scholarship of Thailand and the 90th Anniversary ofChulalongkorn University Fund (RatchadaphiseksomphotEndowment Fund)

Supplementary Materials

Supplement 1 detail of sample collection andGenBank acces-sion numbers Supplement 2 frequency () of the nucleotidecompositions of 28S rRNA and cytb regions Supplement 3mean frequency () for base compositions at different codonpositions in the cytb gene fragment Supplement 4 the 16polymorphic sites among eight cytb haplotypes of A ceranain Thailand Only positions that are different from haplotypeHAP1 are indicated (Supplementary Materials)

References

[1] R Winfree J R Reilly I Bartomeus D P Cariveau NM Williams and J Gibbs ldquoSpecies turnover promotes theimportance of bee diversity for crop pollination at regionalscalesrdquo Science vol 359 no 6377 pp 791ndash793 2018

[2] S Wongsiri C Lekprayoon R Thapa et al ldquoComparativebiology of Apis andreniformis and Apis florea in Thailandrdquo BeeWorld vol 78 no 1 pp 23ndash35 1997

[3] A Rattanawannee C Chanchao and S Wongsiri ldquoMorpho-metric and genetic variation of small dwarf honeybees Apisandreniformis Smith 1858 in Thailandrdquo Insect Science vol 14no 6 pp 451ndash460 2007

[4] B P Oldroyd K E Osborne and M Mardan ldquoColony related-ness in aggregations of Apis dorsata Fabricius (HymenopteraApidae)rdquo Insectes Sociaux vol 47 no 1 pp 94-95 2000

[5] A Rattanawannee C Chanchao J Lim S Wongsiri and BP Oldroyd ldquoGenetic structure of a giant honey bee (Apis

4 Psyche

Table 2 Molecular diversity indices within the 28S rRNA and cytb sequences in each Apis species in Thailand

Species(No of sequences) No S h (plusmn SD) k 120587plusmn SD)

28S rRNAA cerana (26) 2 1 0212 plusmn 0097 0212 000033 plusmn 000041A florea (22) 3 2 0394 plusmn 0119 0541 000085 plusmn 000086A dorsata (15) 4 10 0371 plusmn 0153 1333 000213 plusmn 000491

CytbA cerana (10) 8 16 0933 plusmn 0077 5956 00154 plusmn 00147A florea (11) 3 3 0473 plusmn 0162 1200 000201 plusmn 000250Number of haplotypes (No) number of polymorphic sites (s) haplotype diversity (h) [21] average number of nucleotide differences (k) and nucleotide diversity(120587) [22] with standard deviation (SD)

cytb 33 sequences were analysed including the 1st 2nd and3rd codons and noncoding positions separately giving a totalof 325 positions in the final dataset with 76 polymorphicsites of which 27 sites showed transitions 42 sites showedtransversions and seven sites carried both types of basesubstitution (Table 1)

Based on the sequence alignment within species the28S rRNA revealed low genetic diversity indices and lowgenetic variation among Asian Apis spp (Tables 2 and 3)The sequence divergence (pairwise comparison) over the28S rRNA within A cerana A florea and A dorsata wasderived from 612 638 and 627 bp respectively using 2622 and 15 sequences from across Thailand (Table 3) Thesequence alignment between species revealed 19 polymorphicsites comprising seven AG transitions six TC transitionstwo AT transversions three GT transversions and one GCtransversion The between species transition and transver-sion mutations in the 28S rRNA sequence were 6242and 3758 respectively Our results also demonstrated thatthe intraspecific sequence divergence was lower than theinterspecific sequence divergence (Table 4) For cytb onlyone and two sequences were obtained for A dorsata and Aandreniformis respectively The number of base differencesper sequence from385 positions is shown inTable 4 revealing33 different sequence variants in five species The sequencealignment between species revealed 76 polymorphic sites ofwhich there were 11 AG transitions 20 TC transitions 37AT transversions one GT transition one GC transitionand seven sites that carried both types of base substitutionThe transition and transversion mutations in cytb sequencesbetween species were 5454 and 5546 respectively Themean sequence divergence for all the variants was 89 Ingeneral the sequence divergence was high among speciesranging from 65 to 252 and lower between individuals ofthe same species

Based on the cytb sequence alignment the cytb analysisof A cerana (n = 10) revealed eight haplotypes from 16polymorphic sites that contained four AG transitions eightTC transitions and four AT transversions (Supplement4) The A cerana populations from the south of Thailandshowed a clear sequence divergence from the other regionsof Thailand The sequence divergence () within speciesand the average number of nucleotide differences observedin pairwise comparison among A cerana are shown in

Table 5 The highest sequence divergence was detected whenHAP1ndashHAP6 (including A cerana from east central andwest regions) was compared with HAP7 and HAP8 (fromsouth regions) (Supplement 4) These results suggest that Acerana in southernThailand were from different populationsthan those in the other regions of Thailand

In A florea the pairwise comparison of the cytbsequences suggested thatA florea from two colonies from SA(central regions)were slightly different from the other regions(Table 5) Interestingly the sequence divergence amongA cerana and A mellifera (1022) was lower than thatamong A dorsata A florea and A andreniformis (Table 4)Meanwhile A florea showed less sequence divergence fromA andreniformis (735) than fromA cerana andA mellifera(1219 and 1307 respectively)

32 Phylogenetic Analysis Thebest-fit model of evolution forconstructing the phylogenetic tree by ML analysis was foundby the AIC to be TVM + gamma while that for the BI treeunder the BIC was HKY85 + homogeneous The ML treesconstructed for the 28S rRNA and cytb gene sequences wereessentially the same as the corresponding BI tree and so onlythe ML tree topologies are illustrated here (Figures 2 and 3)

For the 28S rRNA sequences (612ndash638 bp) the phy-logenetic relationships within A cerana A florea and Adorsata from all geographical regions showed no evidenceof geographical clustering in the phylogenetic tree althoughthe nonnative A mellifera showed some divergence withinthe species Overall a low genetic variation within all AsianApis spp was noticed based on the 75 sequences of this studyplus two sequences of A andreniformis from GenBank andusing Trigona collina as an outgroup (Figure 2)The Apis sppwere subdivided into four major clades (BV range from 50to 98 PP range from 060 to 095) strongly supportingthe monophyly of Apis species in Thailand (BV = 100PP = 10) Interestingly all four Apis species were clearlygrouped into their respective species clade except for the twodwarf honeybee speciesA florea andA andreniformis whichwere not completely resolved (Figure 2) This reflects the lowsequence divergence between these dwarf species whichwerebasal in the Apis groups (Figure 2)

For the cytb region (383 bp in length) the ML phyloge-netic dendrograms based on 33 sequences from this studyfrom all Apis species including only one and two sequences

Psyche 5

Table3Estim

ated

sequ

ence

divergence

over

the2

8SrRNA

sequ

ence

pairs

with

ineach

speciesSequ

ence

divergence

()w

ithin

speciesisa

bove

thed

iagonalwhilethea

verage

numbero

fnu

cleotided

ifferencesisb

elowthed

iagonal

Acerana

Central(Sam

utsong

kram

)West(Ka

nchanabu

riRa

tchabu

ri)North

(Chiangm

aiN

an)

East(C

hanthabu

ri)South(C

humph

on

Suratth

aniNakho

nsi

tham

marat)

Central

-003

007

004

004

West

025

-006

00

North

045

040

-006

006

East

025

0040

-0

South

025

0040

0-

Aflorea

Central(Sam

utsong

kram

Nakho

nsaw

an)

West(Ka

nchanabu

riRa

tchabu

ri)Ea

st(C

hanthabu

ri)South(C

humph

on)

North

(Pichit)

Central

-008

003

003

001

West

057

-007

007

001

East

020

050

-000

009

South

020

050

000

-009

North

064

077

060

060

-

Adorsata

Central(Sam

utsong

kram

Nakho

nsaw

an)

West(Ka

nchanabu

ri)North

(Nan)

South(Suratthani)

Central

-024

029

024

West

150

-005

000

North

183

033

-005

South

150

000

033

-

6 Psyche

Table 4 Estimated sequence divergence in the 28S rRNA and cytb sequence pairs between species The pairwise comparison uses 612 and383 bp of 28SrRNA and cytb sequence respectively

28S rRNA A cerana A mellifera A dorsata A florea A andreniformisA cerana - 055 061 120 117A mellifera 339 - 081 105 105A dorsata 376 498 - 113 109A florea 729 642 686 - 003A andreniformis 711 636 667 023 -cytb A mellifera A andreniformis A dorsata A cerana A floreaA mellifera - 1148 1058 1022 1307A andreniformis 3461 - 1289 1249 735A dorsata 3400 3850 - 1259 1345A cerana 3250 3740 3830 - 1219A florea 3900 2277 4200 3657 -Sequence divergence () within species is above the diagonal while the average number of nucleotide differences is below the diagonal

Table 5 Estimated sequence divergence over cytb sequence pairs (383 bp) within A cerana (10 samples) and A florea (11 samples)

A cerana Chantaburi (East) Kanchanaburi (West) Samutsongkram (Central) Suratthani (South)Chantaburi (East) - 084 105 388Kanchanaburi (West) 322 - 026 306Samutsongkram(Central) 400 100 - 279

Suratthani (South) 1450 1150 105 -A florea Chantaburi (East) Chumphon (South) Pichit (North) Ratchaburi (West) Samutsongkram (Central)Chantaburi (East) - 0 024 0 074Chumphon (South) 0 - 024 0 074Pichit(North) 100 100 - 024 049

Ratchaburi (West) 0 0 100 - 074Samutsongkram(Central) 300 300 200 300 -

Sequence divergence () with Jukes and Cantor between species is above the diagonal and the average number of nucleotide differences is below the diagonal

of A dorsata and A andreniformis respectively plus fivesequences of Apis spp from GenBank were constructedwith T collina as the outgroup The obtained tree revealedfour major clades (Figure 3) where each Apis species wasclearly separated from the others including A florea and Aandreniformis Although A florea in Thailand was dividedinto small subgroups with three haplotypes in the populationfrom Samutsongkram (central region) there was a low levelof genetic variation in cytb in the A florea populations Thebasal group was A dorsata For A cerana (clade 1) the 10A cerana samples in Thailand of this study plus three Acerana sequences from GenBank were subdivided into thetwo large groups of SouthernThailand and the other regionsplus one subgroup in the Eastern region (Figure 3) The Acerana populations from Suratthani were clearly separatedfrom the other regions (BV = 100 PP = 095) while thesamples from Eastern Thailand showed a slight differencefrom the other regions although this had only weak support(BV = 64) This low support might reflect the nucleotidedifferences among the haplotypes (Supplement 4) that arestill placed in the same clade These results suggested that A

cerana in Southern Thailand are from a different populationto those in the other regions of Thailand

4 Discussion

A total of 73 28S rRNA and 33 cytb sequences were obtainedThe overall sequence diversity of both gene fragments withinand among species was low especially for 28S rRNA thatshowed only a slight sequence divergence for all haplotypes(07)This suggests that there is little variation in 28S rRNAwithinApis spp inThailandThe 28S rRNA gene is often usedas a representative nuclear gene at the family-species leveland above where for example it was reported that the mostconserved 28S-D3 region could be used to define Encarsiaspp [31]

However using 28S rRNA alone may be insufficientespecially at the species level and below In addition to 28SrRNA other nuclear genes have been used like itpr [32]and the EF1-120572 intron [18] in order to resolve the familylevel phylogenies in bees and other insects Neverthelessthe nuclear EF1-120572 intron has been reported to be useful for

Psyche 7

ACCM1

ACCM2ACSA3ACCM3ACCN1ACCN3ACCP1ACCP2ACNS2ACNS3ACRA1ACRA2ACCP3ACKA1

ACN2

ACKA2ACKA3ACSA1ACNS1ACSA2ACSU1ACSU2ACSU3

ADKA1ADKA2ADKA3

ADN2ADN3ADNA1

ADN98081

87078

A2ADN1

ADSA1ADSA2ADSA3 ADSA4ADSU1ADSU2ADSU3

ACCN2 ACN1ACSA47607763072

AA1

AA2

AAKAAFCP2AFKA1

AFSA3AFPJ5

AFPJ4AFRA3AFSA3

AFSA1AFSA2

AFPJ5AFRA2

56067

63065AFCN1AFNA1AFNA2AFCN2AFCN3AFCP1AFCP3AFPJ1AFPJ2AFPJ3AFRA1

50095

Tetragonilla collina (FJ042138)

002

50060

84067

98091

8907070063 AMCM1

AMP2AMCM3

AMP1AMP3

AMCM2

AMN1

AMCM4

AMCM5

AMN2

AMN3

Clade II

Clade III

Clade IV

Clade I

Figure 2 Phylogenetic relationship of honeybee species in Thailand based upon the 28S rRNA sequences by ML analysis Trigonacollina (accession FJ042138) was utilized as the outgroup Node support inferred from Bayesian PP and ML BV is shown Sample(speciesgeographic location) codes are as in Supplement 1

8 Psyche

Apis dorsata (KC294229)

ADKA

Apis andreniformis (EU040176)Apis andreniformis (EU040168)Apis andreniformis (EU040165)AACNAAKA

91082

100097

AFCN1AFCN2AFCP1AFCP2AFCP3AFPJ1AFSA1AFSA2AFPJ2AFRA1AFRA2Apis florea (JX982136)

Apis florea (KC170303)

90092

100100

AMCM1AMN2AMCM2AMCM3AMP3AMP2AMN1AMN3AMP1

100093

ACCN1ACCN3ACCN2ACKA1ACKA2ACKA3ACSA1ACSA2

100064

100095

Apis cerana (1) (AP017941)Apis cerana (2) (AP017941)

Apis cerana (GQ162109)ACSU1

ACSU2

Trigono collino (AY575081)

300

Clade II

Clade IV

Clade IV-1

Clade IV-2

Clade I

Clade III

Clade II-1

Clade II-2

Figure 3 Phylogenetic relationship of honeybee species in Thailand based upon cytb sequences by BI and including Apis spp fromGenBank (accession KC294229 EU040176 EU040168 EU040165 JX982136 KC170303 AP017941 AP017942 and GQ162109) Trigonacollina (accession AY575081) was utilized as the outgroup Node support as Bayesian PP and ML BV is shown Sample (speciesgeographiclocation) codes are as in Supplement 1

classifying honeybees between species but not for resolvingwithin species [18] Likewise it was reported that the 28SrRNA was unlikely to recover deep divergences at the familylevel in bees [33] In contrast in combination with othergenes such as opsin cytb and 16S rRNA 28S rRNA appearedto contribute to resolving basal bee divergence

In this study the number of transversions in cytb washigher than the number of transitions in each Apis species(Table 1) This is in agreement with the transitional biasreported previously [18] Within A cerana the cytb sequencedivergence ranged from 026 to 388 but interestingly washigher in the southern populations than in the other regions

Psyche 9

of Thailand (Table 5) The phylogenetic results are consistentwith the molecular diversity indices of A cerana where Acerana populations in southernThailandwere separated fromthose in the East West and Central regions (Figure 3) Theresult could be explained by the biogeography of A ceranaIn addition the data concur with the previous PCR-RFLPanalysis [34]

For A florea only a very low cytb sequence divergencewas detected (0ndash074) with only one haplotype Onlytwo colonies from Samutsongkram province were slightlydifferent from the others (Figure 3)

Overall the sequence divergence of cytb (89 plusmn 50) washigher than that for 28S rRNA (07 plusmn 05) consistent witha higher evolutionary rate for mitochondrial than nucleargenomes [11] Although the mitochondrial cytb showed ahigher rate of variation it still showed a low variation in theseApis spp among the different geographical regions This mayreflect that Apis spp are highly eusocial and polyandrouswhich reduces the substitution rate and counteracts the effectof genetic drift by increasing the effective population size[35] The ancestral genome in diverging populations will bemaintained and so the genetic distances between populationsare reduced [36] The other possible factor leading to alow genetic variation is migration behaviour The migrationdistance ofA dorsata has been reported to be 50 kmwide andup to 200 km [37] InA cerana although they havemigratoryswarms flying dozens of km [38] they have a strong tendencyto swarm 5ndash6 times a year and are well adapted to new areassince it is more resistant to some diseases and pest problemsin each area [39]

TheMLandBI phylogenetic analyses of the 28S rRNA andcytb sequences showed the same topologies with four majorclades among Apis spp (for ML Figures 2 and 3) Using Tcollina as the outgroup the 28S rRNA tree placed the dwarfbees in a basal position within the genus while A dorsatawas located in the basal position in the cytb tree consistentwith previous analysis [40] Although both the cytb and28S rRNA phylogenetic trees in this study could not clearlyresolve the relationship between these Apis species theycould separate A dorsata from others (Figures 2 and 3) Aclose relationship has been reported between A dorsata andA cerana in the 28S rRNA tree and molecular indices [32]Although A cerana and A mellifera have been reported tobe closely related to each other [18] they have ecological andbehavioural differences [41] In this study the 28S rRNA andcytb topologies clearly separated A cerana and A mellifera inThailand from each other although they were still closer toeach other compared to the other Apis species

Despite the fact that A dorsata and A florea are open-nesting bees A dorsata was closer to A cerana and Amellifera than to A florea and A andreniformis in this studyThis can be explained by the evolution of behaviour in Apiswhere cavity-nesting is secondarily derived [42] MoreoverA dorsata has a buzzing dance in daylight and at night whileA florea has only a silent waggle dance and cavity-nestingbees adapted this buzzing waggle dance in a dark condi-tion Thus a loss of visual cues was compensated for withsound

5 Conclusion

Overall the genetic variation of Apis spp from 64 popula-tions within 12 provinces in Thailand showed low geneticdiversity indices and variation within and between speciesThis may indicate that habitat fragmentation monocropagriculture industrial management and climate changeshave not affected the diversity of Apis spp in Thailand yetbut clearly further work is of vital importance in establishingthe actual genetic diversity and relationship within Apissubspecies including the analysis of other genes for betterresolution

Data Availability

The authors confirm that all data in the manuscript are freelyavailable

Conflicts of Interest

The authors declare that there are no conflicts of interestregarding the publication of this paper

Acknowledgments

This work was financially supported by the Science Achieve-ment Scholarship of Thailand and the 90th Anniversary ofChulalongkorn University Fund (RatchadaphiseksomphotEndowment Fund)

Supplementary Materials

Supplement 1 detail of sample collection andGenBank acces-sion numbers Supplement 2 frequency () of the nucleotidecompositions of 28S rRNA and cytb regions Supplement 3mean frequency () for base compositions at different codonpositions in the cytb gene fragment Supplement 4 the 16polymorphic sites among eight cytb haplotypes of A ceranain Thailand Only positions that are different from haplotypeHAP1 are indicated (Supplementary Materials)

References

[1] R Winfree J R Reilly I Bartomeus D P Cariveau NM Williams and J Gibbs ldquoSpecies turnover promotes theimportance of bee diversity for crop pollination at regionalscalesrdquo Science vol 359 no 6377 pp 791ndash793 2018

[2] S Wongsiri C Lekprayoon R Thapa et al ldquoComparativebiology of Apis andreniformis and Apis florea in Thailandrdquo BeeWorld vol 78 no 1 pp 23ndash35 1997

[3] A Rattanawannee C Chanchao and S Wongsiri ldquoMorpho-metric and genetic variation of small dwarf honeybees Apisandreniformis Smith 1858 in Thailandrdquo Insect Science vol 14no 6 pp 451ndash460 2007

[4] B P Oldroyd K E Osborne and M Mardan ldquoColony related-ness in aggregations of Apis dorsata Fabricius (HymenopteraApidae)rdquo Insectes Sociaux vol 47 no 1 pp 94-95 2000

[5] A Rattanawannee C Chanchao J Lim S Wongsiri and BP Oldroyd ldquoGenetic structure of a giant honey bee (Apis

Psyche 5

Table3Estim

ated

sequ

ence

divergence

over

the2

8SrRNA

sequ

ence

pairs

with

ineach

speciesSequ

ence

divergence

()w

ithin

speciesisa

bove

thed

iagonalwhilethea

verage

numbero

fnu

cleotided

ifferencesisb

elowthed

iagonal

Acerana

Central(Sam

utsong

kram

)West(Ka

nchanabu

riRa

tchabu

ri)North

(Chiangm

aiN

an)

East(C

hanthabu

ri)South(C

humph

on

Suratth

aniNakho

nsi

tham

marat)

Central

-003

007

004

004

West

025

-006

00

North

045

040

-006

006

East

025

0040

-0

South

025

0040

0-

Aflorea

Central(Sam

utsong

kram

Nakho

nsaw

an)

West(Ka

nchanabu

riRa

tchabu

ri)Ea

st(C

hanthabu

ri)South(C

humph

on)

North

(Pichit)

Central

-008

003

003

001

West

057

-007

007

001

East

020

050

-000

009

South

020

050

000

-009

North

064

077

060

060

-

Adorsata

Central(Sam

utsong

kram

Nakho

nsaw

an)

West(Ka

nchanabu

ri)North

(Nan)

South(Suratthani)

Central

-024

029

024

West

150

-005

000

North

183

033

-005

South

150

000

033

-

6 Psyche

Table 4 Estimated sequence divergence in the 28S rRNA and cytb sequence pairs between species The pairwise comparison uses 612 and383 bp of 28SrRNA and cytb sequence respectively

28S rRNA A cerana A mellifera A dorsata A florea A andreniformisA cerana - 055 061 120 117A mellifera 339 - 081 105 105A dorsata 376 498 - 113 109A florea 729 642 686 - 003A andreniformis 711 636 667 023 -cytb A mellifera A andreniformis A dorsata A cerana A floreaA mellifera - 1148 1058 1022 1307A andreniformis 3461 - 1289 1249 735A dorsata 3400 3850 - 1259 1345A cerana 3250 3740 3830 - 1219A florea 3900 2277 4200 3657 -Sequence divergence () within species is above the diagonal while the average number of nucleotide differences is below the diagonal

Table 5 Estimated sequence divergence over cytb sequence pairs (383 bp) within A cerana (10 samples) and A florea (11 samples)

A cerana Chantaburi (East) Kanchanaburi (West) Samutsongkram (Central) Suratthani (South)Chantaburi (East) - 084 105 388Kanchanaburi (West) 322 - 026 306Samutsongkram(Central) 400 100 - 279

Suratthani (South) 1450 1150 105 -A florea Chantaburi (East) Chumphon (South) Pichit (North) Ratchaburi (West) Samutsongkram (Central)Chantaburi (East) - 0 024 0 074Chumphon (South) 0 - 024 0 074Pichit(North) 100 100 - 024 049

Ratchaburi (West) 0 0 100 - 074Samutsongkram(Central) 300 300 200 300 -

Sequence divergence () with Jukes and Cantor between species is above the diagonal and the average number of nucleotide differences is below the diagonal

of A dorsata and A andreniformis respectively plus fivesequences of Apis spp from GenBank were constructedwith T collina as the outgroup The obtained tree revealedfour major clades (Figure 3) where each Apis species wasclearly separated from the others including A florea and Aandreniformis Although A florea in Thailand was dividedinto small subgroups with three haplotypes in the populationfrom Samutsongkram (central region) there was a low levelof genetic variation in cytb in the A florea populations Thebasal group was A dorsata For A cerana (clade 1) the 10A cerana samples in Thailand of this study plus three Acerana sequences from GenBank were subdivided into thetwo large groups of SouthernThailand and the other regionsplus one subgroup in the Eastern region (Figure 3) The Acerana populations from Suratthani were clearly separatedfrom the other regions (BV = 100 PP = 095) while thesamples from Eastern Thailand showed a slight differencefrom the other regions although this had only weak support(BV = 64) This low support might reflect the nucleotidedifferences among the haplotypes (Supplement 4) that arestill placed in the same clade These results suggested that A

cerana in Southern Thailand are from a different populationto those in the other regions of Thailand

4 Discussion

A total of 73 28S rRNA and 33 cytb sequences were obtainedThe overall sequence diversity of both gene fragments withinand among species was low especially for 28S rRNA thatshowed only a slight sequence divergence for all haplotypes(07)This suggests that there is little variation in 28S rRNAwithinApis spp inThailandThe 28S rRNA gene is often usedas a representative nuclear gene at the family-species leveland above where for example it was reported that the mostconserved 28S-D3 region could be used to define Encarsiaspp [31]

However using 28S rRNA alone may be insufficientespecially at the species level and below In addition to 28SrRNA other nuclear genes have been used like itpr [32]and the EF1-120572 intron [18] in order to resolve the familylevel phylogenies in bees and other insects Neverthelessthe nuclear EF1-120572 intron has been reported to be useful for

Psyche 7

ACCM1

ACCM2ACSA3ACCM3ACCN1ACCN3ACCP1ACCP2ACNS2ACNS3ACRA1ACRA2ACCP3ACKA1

ACN2

ACKA2ACKA3ACSA1ACNS1ACSA2ACSU1ACSU2ACSU3

ADKA1ADKA2ADKA3

ADN2ADN3ADNA1

ADN98081

87078

A2ADN1

ADSA1ADSA2ADSA3 ADSA4ADSU1ADSU2ADSU3

ACCN2 ACN1ACSA47607763072

AA1

AA2

AAKAAFCP2AFKA1

AFSA3AFPJ5

AFPJ4AFRA3AFSA3

AFSA1AFSA2

AFPJ5AFRA2

56067

63065AFCN1AFNA1AFNA2AFCN2AFCN3AFCP1AFCP3AFPJ1AFPJ2AFPJ3AFRA1

50095

Tetragonilla collina (FJ042138)

002

50060

84067

98091

8907070063 AMCM1

AMP2AMCM3

AMP1AMP3

AMCM2

AMN1

AMCM4

AMCM5

AMN2

AMN3

Clade II

Clade III

Clade IV

Clade I

Figure 2 Phylogenetic relationship of honeybee species in Thailand based upon the 28S rRNA sequences by ML analysis Trigonacollina (accession FJ042138) was utilized as the outgroup Node support inferred from Bayesian PP and ML BV is shown Sample(speciesgeographic location) codes are as in Supplement 1

8 Psyche

Apis dorsata (KC294229)

ADKA

Apis andreniformis (EU040176)Apis andreniformis (EU040168)Apis andreniformis (EU040165)AACNAAKA

91082

100097

AFCN1AFCN2AFCP1AFCP2AFCP3AFPJ1AFSA1AFSA2AFPJ2AFRA1AFRA2Apis florea (JX982136)

Apis florea (KC170303)

90092

100100

AMCM1AMN2AMCM2AMCM3AMP3AMP2AMN1AMN3AMP1

100093

ACCN1ACCN3ACCN2ACKA1ACKA2ACKA3ACSA1ACSA2

100064

100095

Apis cerana (1) (AP017941)Apis cerana (2) (AP017941)

Apis cerana (GQ162109)ACSU1

ACSU2

Trigono collino (AY575081)

300

Clade II

Clade IV

Clade IV-1

Clade IV-2

Clade I

Clade III

Clade II-1

Clade II-2

Figure 3 Phylogenetic relationship of honeybee species in Thailand based upon cytb sequences by BI and including Apis spp fromGenBank (accession KC294229 EU040176 EU040168 EU040165 JX982136 KC170303 AP017941 AP017942 and GQ162109) Trigonacollina (accession AY575081) was utilized as the outgroup Node support as Bayesian PP and ML BV is shown Sample (speciesgeographiclocation) codes are as in Supplement 1

classifying honeybees between species but not for resolvingwithin species [18] Likewise it was reported that the 28SrRNA was unlikely to recover deep divergences at the familylevel in bees [33] In contrast in combination with othergenes such as opsin cytb and 16S rRNA 28S rRNA appearedto contribute to resolving basal bee divergence

In this study the number of transversions in cytb washigher than the number of transitions in each Apis species(Table 1) This is in agreement with the transitional biasreported previously [18] Within A cerana the cytb sequencedivergence ranged from 026 to 388 but interestingly washigher in the southern populations than in the other regions

Psyche 9

of Thailand (Table 5) The phylogenetic results are consistentwith the molecular diversity indices of A cerana where Acerana populations in southernThailandwere separated fromthose in the East West and Central regions (Figure 3) Theresult could be explained by the biogeography of A ceranaIn addition the data concur with the previous PCR-RFLPanalysis [34]

For A florea only a very low cytb sequence divergencewas detected (0ndash074) with only one haplotype Onlytwo colonies from Samutsongkram province were slightlydifferent from the others (Figure 3)

Overall the sequence divergence of cytb (89 plusmn 50) washigher than that for 28S rRNA (07 plusmn 05) consistent witha higher evolutionary rate for mitochondrial than nucleargenomes [11] Although the mitochondrial cytb showed ahigher rate of variation it still showed a low variation in theseApis spp among the different geographical regions This mayreflect that Apis spp are highly eusocial and polyandrouswhich reduces the substitution rate and counteracts the effectof genetic drift by increasing the effective population size[35] The ancestral genome in diverging populations will bemaintained and so the genetic distances between populationsare reduced [36] The other possible factor leading to alow genetic variation is migration behaviour The migrationdistance ofA dorsata has been reported to be 50 kmwide andup to 200 km [37] InA cerana although they havemigratoryswarms flying dozens of km [38] they have a strong tendencyto swarm 5ndash6 times a year and are well adapted to new areassince it is more resistant to some diseases and pest problemsin each area [39]

TheMLandBI phylogenetic analyses of the 28S rRNA andcytb sequences showed the same topologies with four majorclades among Apis spp (for ML Figures 2 and 3) Using Tcollina as the outgroup the 28S rRNA tree placed the dwarfbees in a basal position within the genus while A dorsatawas located in the basal position in the cytb tree consistentwith previous analysis [40] Although both the cytb and28S rRNA phylogenetic trees in this study could not clearlyresolve the relationship between these Apis species theycould separate A dorsata from others (Figures 2 and 3) Aclose relationship has been reported between A dorsata andA cerana in the 28S rRNA tree and molecular indices [32]Although A cerana and A mellifera have been reported tobe closely related to each other [18] they have ecological andbehavioural differences [41] In this study the 28S rRNA andcytb topologies clearly separated A cerana and A mellifera inThailand from each other although they were still closer toeach other compared to the other Apis species

Despite the fact that A dorsata and A florea are open-nesting bees A dorsata was closer to A cerana and Amellifera than to A florea and A andreniformis in this studyThis can be explained by the evolution of behaviour in Apiswhere cavity-nesting is secondarily derived [42] MoreoverA dorsata has a buzzing dance in daylight and at night whileA florea has only a silent waggle dance and cavity-nestingbees adapted this buzzing waggle dance in a dark condi-tion Thus a loss of visual cues was compensated for withsound

5 Conclusion

Overall the genetic variation of Apis spp from 64 popula-tions within 12 provinces in Thailand showed low geneticdiversity indices and variation within and between speciesThis may indicate that habitat fragmentation monocropagriculture industrial management and climate changeshave not affected the diversity of Apis spp in Thailand yetbut clearly further work is of vital importance in establishingthe actual genetic diversity and relationship within Apissubspecies including the analysis of other genes for betterresolution

Data Availability

The authors confirm that all data in the manuscript are freelyavailable

Conflicts of Interest

The authors declare that there are no conflicts of interestregarding the publication of this paper

Acknowledgments

This work was financially supported by the Science Achieve-ment Scholarship of Thailand and the 90th Anniversary ofChulalongkorn University Fund (RatchadaphiseksomphotEndowment Fund)

Supplementary Materials

Supplement 1 detail of sample collection andGenBank acces-sion numbers Supplement 2 frequency () of the nucleotidecompositions of 28S rRNA and cytb regions Supplement 3mean frequency () for base compositions at different codonpositions in the cytb gene fragment Supplement 4 the 16polymorphic sites among eight cytb haplotypes of A ceranain Thailand Only positions that are different from haplotypeHAP1 are indicated (Supplementary Materials)

References

[1] R Winfree J R Reilly I Bartomeus D P Cariveau NM Williams and J Gibbs ldquoSpecies turnover promotes theimportance of bee diversity for crop pollination at regionalscalesrdquo Science vol 359 no 6377 pp 791ndash793 2018

[2] S Wongsiri C Lekprayoon R Thapa et al ldquoComparativebiology of Apis andreniformis and Apis florea in Thailandrdquo BeeWorld vol 78 no 1 pp 23ndash35 1997

[3] A Rattanawannee C Chanchao and S Wongsiri ldquoMorpho-metric and genetic variation of small dwarf honeybees Apisandreniformis Smith 1858 in Thailandrdquo Insect Science vol 14no 6 pp 451ndash460 2007

[4] B P Oldroyd K E Osborne and M Mardan ldquoColony related-ness in aggregations of Apis dorsata Fabricius (HymenopteraApidae)rdquo Insectes Sociaux vol 47 no 1 pp 94-95 2000

[5] A Rattanawannee C Chanchao J Lim S Wongsiri and BP Oldroyd ldquoGenetic structure of a giant honey bee (Apis

6 Psyche

Table 4 Estimated sequence divergence in the 28S rRNA and cytb sequence pairs between species The pairwise comparison uses 612 and383 bp of 28SrRNA and cytb sequence respectively

28S rRNA A cerana A mellifera A dorsata A florea A andreniformisA cerana - 055 061 120 117A mellifera 339 - 081 105 105A dorsata 376 498 - 113 109A florea 729 642 686 - 003A andreniformis 711 636 667 023 -cytb A mellifera A andreniformis A dorsata A cerana A floreaA mellifera - 1148 1058 1022 1307A andreniformis 3461 - 1289 1249 735A dorsata 3400 3850 - 1259 1345A cerana 3250 3740 3830 - 1219A florea 3900 2277 4200 3657 -Sequence divergence () within species is above the diagonal while the average number of nucleotide differences is below the diagonal

Table 5 Estimated sequence divergence over cytb sequence pairs (383 bp) within A cerana (10 samples) and A florea (11 samples)

A cerana Chantaburi (East) Kanchanaburi (West) Samutsongkram (Central) Suratthani (South)Chantaburi (East) - 084 105 388Kanchanaburi (West) 322 - 026 306Samutsongkram(Central) 400 100 - 279

Suratthani (South) 1450 1150 105 -A florea Chantaburi (East) Chumphon (South) Pichit (North) Ratchaburi (West) Samutsongkram (Central)Chantaburi (East) - 0 024 0 074Chumphon (South) 0 - 024 0 074Pichit(North) 100 100 - 024 049

Ratchaburi (West) 0 0 100 - 074Samutsongkram(Central) 300 300 200 300 -

Sequence divergence () with Jukes and Cantor between species is above the diagonal and the average number of nucleotide differences is below the diagonal

of A dorsata and A andreniformis respectively plus fivesequences of Apis spp from GenBank were constructedwith T collina as the outgroup The obtained tree revealedfour major clades (Figure 3) where each Apis species wasclearly separated from the others including A florea and Aandreniformis Although A florea in Thailand was dividedinto small subgroups with three haplotypes in the populationfrom Samutsongkram (central region) there was a low levelof genetic variation in cytb in the A florea populations Thebasal group was A dorsata For A cerana (clade 1) the 10A cerana samples in Thailand of this study plus three Acerana sequences from GenBank were subdivided into thetwo large groups of SouthernThailand and the other regionsplus one subgroup in the Eastern region (Figure 3) The Acerana populations from Suratthani were clearly separatedfrom the other regions (BV = 100 PP = 095) while thesamples from Eastern Thailand showed a slight differencefrom the other regions although this had only weak support(BV = 64) This low support might reflect the nucleotidedifferences among the haplotypes (Supplement 4) that arestill placed in the same clade These results suggested that A

cerana in Southern Thailand are from a different populationto those in the other regions of Thailand

4 Discussion

A total of 73 28S rRNA and 33 cytb sequences were obtainedThe overall sequence diversity of both gene fragments withinand among species was low especially for 28S rRNA thatshowed only a slight sequence divergence for all haplotypes(07)This suggests that there is little variation in 28S rRNAwithinApis spp inThailandThe 28S rRNA gene is often usedas a representative nuclear gene at the family-species leveland above where for example it was reported that the mostconserved 28S-D3 region could be used to define Encarsiaspp [31]

However using 28S rRNA alone may be insufficientespecially at the species level and below In addition to 28SrRNA other nuclear genes have been used like itpr [32]and the EF1-120572 intron [18] in order to resolve the familylevel phylogenies in bees and other insects Neverthelessthe nuclear EF1-120572 intron has been reported to be useful for

Psyche 7

ACCM1

ACCM2ACSA3ACCM3ACCN1ACCN3ACCP1ACCP2ACNS2ACNS3ACRA1ACRA2ACCP3ACKA1

ACN2

ACKA2ACKA3ACSA1ACNS1ACSA2ACSU1ACSU2ACSU3

ADKA1ADKA2ADKA3

ADN2ADN3ADNA1

ADN98081

87078

A2ADN1

ADSA1ADSA2ADSA3 ADSA4ADSU1ADSU2ADSU3

ACCN2 ACN1ACSA47607763072

AA1

AA2

AAKAAFCP2AFKA1

AFSA3AFPJ5

AFPJ4AFRA3AFSA3

AFSA1AFSA2

AFPJ5AFRA2

56067

63065AFCN1AFNA1AFNA2AFCN2AFCN3AFCP1AFCP3AFPJ1AFPJ2AFPJ3AFRA1

50095

Tetragonilla collina (FJ042138)

002

50060

84067

98091

8907070063 AMCM1

AMP2AMCM3

AMP1AMP3

AMCM2

AMN1

AMCM4

AMCM5

AMN2

AMN3

Clade II

Clade III

Clade IV

Clade I

Figure 2 Phylogenetic relationship of honeybee species in Thailand based upon the 28S rRNA sequences by ML analysis Trigonacollina (accession FJ042138) was utilized as the outgroup Node support inferred from Bayesian PP and ML BV is shown Sample(speciesgeographic location) codes are as in Supplement 1

8 Psyche

Apis dorsata (KC294229)

ADKA

Apis andreniformis (EU040176)Apis andreniformis (EU040168)Apis andreniformis (EU040165)AACNAAKA

91082

100097

AFCN1AFCN2AFCP1AFCP2AFCP3AFPJ1AFSA1AFSA2AFPJ2AFRA1AFRA2Apis florea (JX982136)

Apis florea (KC170303)

90092

100100

AMCM1AMN2AMCM2AMCM3AMP3AMP2AMN1AMN3AMP1

100093

ACCN1ACCN3ACCN2ACKA1ACKA2ACKA3ACSA1ACSA2

100064

100095

Apis cerana (1) (AP017941)Apis cerana (2) (AP017941)

Apis cerana (GQ162109)ACSU1

ACSU2

Trigono collino (AY575081)

300

Clade II

Clade IV

Clade IV-1

Clade IV-2

Clade I

Clade III

Clade II-1

Clade II-2

Figure 3 Phylogenetic relationship of honeybee species in Thailand based upon cytb sequences by BI and including Apis spp fromGenBank (accession KC294229 EU040176 EU040168 EU040165 JX982136 KC170303 AP017941 AP017942 and GQ162109) Trigonacollina (accession AY575081) was utilized as the outgroup Node support as Bayesian PP and ML BV is shown Sample (speciesgeographiclocation) codes are as in Supplement 1

classifying honeybees between species but not for resolvingwithin species [18] Likewise it was reported that the 28SrRNA was unlikely to recover deep divergences at the familylevel in bees [33] In contrast in combination with othergenes such as opsin cytb and 16S rRNA 28S rRNA appearedto contribute to resolving basal bee divergence

In this study the number of transversions in cytb washigher than the number of transitions in each Apis species(Table 1) This is in agreement with the transitional biasreported previously [18] Within A cerana the cytb sequencedivergence ranged from 026 to 388 but interestingly washigher in the southern populations than in the other regions

Psyche 9

of Thailand (Table 5) The phylogenetic results are consistentwith the molecular diversity indices of A cerana where Acerana populations in southernThailandwere separated fromthose in the East West and Central regions (Figure 3) Theresult could be explained by the biogeography of A ceranaIn addition the data concur with the previous PCR-RFLPanalysis [34]

For A florea only a very low cytb sequence divergencewas detected (0ndash074) with only one haplotype Onlytwo colonies from Samutsongkram province were slightlydifferent from the others (Figure 3)

Overall the sequence divergence of cytb (89 plusmn 50) washigher than that for 28S rRNA (07 plusmn 05) consistent witha higher evolutionary rate for mitochondrial than nucleargenomes [11] Although the mitochondrial cytb showed ahigher rate of variation it still showed a low variation in theseApis spp among the different geographical regions This mayreflect that Apis spp are highly eusocial and polyandrouswhich reduces the substitution rate and counteracts the effectof genetic drift by increasing the effective population size[35] The ancestral genome in diverging populations will bemaintained and so the genetic distances between populationsare reduced [36] The other possible factor leading to alow genetic variation is migration behaviour The migrationdistance ofA dorsata has been reported to be 50 kmwide andup to 200 km [37] InA cerana although they havemigratoryswarms flying dozens of km [38] they have a strong tendencyto swarm 5ndash6 times a year and are well adapted to new areassince it is more resistant to some diseases and pest problemsin each area [39]

TheMLandBI phylogenetic analyses of the 28S rRNA andcytb sequences showed the same topologies with four majorclades among Apis spp (for ML Figures 2 and 3) Using Tcollina as the outgroup the 28S rRNA tree placed the dwarfbees in a basal position within the genus while A dorsatawas located in the basal position in the cytb tree consistentwith previous analysis [40] Although both the cytb and28S rRNA phylogenetic trees in this study could not clearlyresolve the relationship between these Apis species theycould separate A dorsata from others (Figures 2 and 3) Aclose relationship has been reported between A dorsata andA cerana in the 28S rRNA tree and molecular indices [32]Although A cerana and A mellifera have been reported tobe closely related to each other [18] they have ecological andbehavioural differences [41] In this study the 28S rRNA andcytb topologies clearly separated A cerana and A mellifera inThailand from each other although they were still closer toeach other compared to the other Apis species

Despite the fact that A dorsata and A florea are open-nesting bees A dorsata was closer to A cerana and Amellifera than to A florea and A andreniformis in this studyThis can be explained by the evolution of behaviour in Apiswhere cavity-nesting is secondarily derived [42] MoreoverA dorsata has a buzzing dance in daylight and at night whileA florea has only a silent waggle dance and cavity-nestingbees adapted this buzzing waggle dance in a dark condi-tion Thus a loss of visual cues was compensated for withsound

5 Conclusion

Overall the genetic variation of Apis spp from 64 popula-tions within 12 provinces in Thailand showed low geneticdiversity indices and variation within and between speciesThis may indicate that habitat fragmentation monocropagriculture industrial management and climate changeshave not affected the diversity of Apis spp in Thailand yetbut clearly further work is of vital importance in establishingthe actual genetic diversity and relationship within Apissubspecies including the analysis of other genes for betterresolution

Data Availability

The authors confirm that all data in the manuscript are freelyavailable

Conflicts of Interest

The authors declare that there are no conflicts of interestregarding the publication of this paper

Acknowledgments

This work was financially supported by the Science Achieve-ment Scholarship of Thailand and the 90th Anniversary ofChulalongkorn University Fund (RatchadaphiseksomphotEndowment Fund)

Supplementary Materials

Supplement 1 detail of sample collection andGenBank acces-sion numbers Supplement 2 frequency () of the nucleotidecompositions of 28S rRNA and cytb regions Supplement 3mean frequency () for base compositions at different codonpositions in the cytb gene fragment Supplement 4 the 16polymorphic sites among eight cytb haplotypes of A ceranain Thailand Only positions that are different from haplotypeHAP1 are indicated (Supplementary Materials)

References

[1] R Winfree J R Reilly I Bartomeus D P Cariveau NM Williams and J Gibbs ldquoSpecies turnover promotes theimportance of bee diversity for crop pollination at regionalscalesrdquo Science vol 359 no 6377 pp 791ndash793 2018

[2] S Wongsiri C Lekprayoon R Thapa et al ldquoComparativebiology of Apis andreniformis and Apis florea in Thailandrdquo BeeWorld vol 78 no 1 pp 23ndash35 1997

[3] A Rattanawannee C Chanchao and S Wongsiri ldquoMorpho-metric and genetic variation of small dwarf honeybees Apisandreniformis Smith 1858 in Thailandrdquo Insect Science vol 14no 6 pp 451ndash460 2007

[4] B P Oldroyd K E Osborne and M Mardan ldquoColony related-ness in aggregations of Apis dorsata Fabricius (HymenopteraApidae)rdquo Insectes Sociaux vol 47 no 1 pp 94-95 2000

[5] A Rattanawannee C Chanchao J Lim S Wongsiri and BP Oldroyd ldquoGenetic structure of a giant honey bee (Apis

Psyche 7

ACCM1

ACCM2ACSA3ACCM3ACCN1ACCN3ACCP1ACCP2ACNS2ACNS3ACRA1ACRA2ACCP3ACKA1

ACN2

ACKA2ACKA3ACSA1ACNS1ACSA2ACSU1ACSU2ACSU3

ADKA1ADKA2ADKA3

ADN2ADN3ADNA1

ADN98081

87078

A2ADN1

ADSA1ADSA2ADSA3 ADSA4ADSU1ADSU2ADSU3

ACCN2 ACN1ACSA47607763072

AA1

AA2

AAKAAFCP2AFKA1

AFSA3AFPJ5

AFPJ4AFRA3AFSA3

AFSA1AFSA2

AFPJ5AFRA2

56067

63065AFCN1AFNA1AFNA2AFCN2AFCN3AFCP1AFCP3AFPJ1AFPJ2AFPJ3AFRA1

50095

Tetragonilla collina (FJ042138)

002

50060

84067

98091

8907070063 AMCM1

AMP2AMCM3

AMP1AMP3

AMCM2

AMN1

AMCM4

AMCM5

AMN2

AMN3

Clade II

Clade III

Clade IV

Clade I

Figure 2 Phylogenetic relationship of honeybee species in Thailand based upon the 28S rRNA sequences by ML analysis Trigonacollina (accession FJ042138) was utilized as the outgroup Node support inferred from Bayesian PP and ML BV is shown Sample(speciesgeographic location) codes are as in Supplement 1

8 Psyche

Apis dorsata (KC294229)

ADKA

Apis andreniformis (EU040176)Apis andreniformis (EU040168)Apis andreniformis (EU040165)AACNAAKA

91082

100097

AFCN1AFCN2AFCP1AFCP2AFCP3AFPJ1AFSA1AFSA2AFPJ2AFRA1AFRA2Apis florea (JX982136)

Apis florea (KC170303)

90092

100100

AMCM1AMN2AMCM2AMCM3AMP3AMP2AMN1AMN3AMP1

100093

ACCN1ACCN3ACCN2ACKA1ACKA2ACKA3ACSA1ACSA2

100064

100095

Apis cerana (1) (AP017941)Apis cerana (2) (AP017941)

Apis cerana (GQ162109)ACSU1

ACSU2

Trigono collino (AY575081)

300

Clade II

Clade IV

Clade IV-1

Clade IV-2

Clade I

Clade III

Clade II-1

Clade II-2

Figure 3 Phylogenetic relationship of honeybee species in Thailand based upon cytb sequences by BI and including Apis spp fromGenBank (accession KC294229 EU040176 EU040168 EU040165 JX982136 KC170303 AP017941 AP017942 and GQ162109) Trigonacollina (accession AY575081) was utilized as the outgroup Node support as Bayesian PP and ML BV is shown Sample (speciesgeographiclocation) codes are as in Supplement 1

classifying honeybees between species but not for resolvingwithin species [18] Likewise it was reported that the 28SrRNA was unlikely to recover deep divergences at the familylevel in bees [33] In contrast in combination with othergenes such as opsin cytb and 16S rRNA 28S rRNA appearedto contribute to resolving basal bee divergence

In this study the number of transversions in cytb washigher than the number of transitions in each Apis species(Table 1) This is in agreement with the transitional biasreported previously [18] Within A cerana the cytb sequencedivergence ranged from 026 to 388 but interestingly washigher in the southern populations than in the other regions

Psyche 9

of Thailand (Table 5) The phylogenetic results are consistentwith the molecular diversity indices of A cerana where Acerana populations in southernThailandwere separated fromthose in the East West and Central regions (Figure 3) Theresult could be explained by the biogeography of A ceranaIn addition the data concur with the previous PCR-RFLPanalysis [34]

For A florea only a very low cytb sequence divergencewas detected (0ndash074) with only one haplotype Onlytwo colonies from Samutsongkram province were slightlydifferent from the others (Figure 3)

Overall the sequence divergence of cytb (89 plusmn 50) washigher than that for 28S rRNA (07 plusmn 05) consistent witha higher evolutionary rate for mitochondrial than nucleargenomes [11] Although the mitochondrial cytb showed ahigher rate of variation it still showed a low variation in theseApis spp among the different geographical regions This mayreflect that Apis spp are highly eusocial and polyandrouswhich reduces the substitution rate and counteracts the effectof genetic drift by increasing the effective population size[35] The ancestral genome in diverging populations will bemaintained and so the genetic distances between populationsare reduced [36] The other possible factor leading to alow genetic variation is migration behaviour The migrationdistance ofA dorsata has been reported to be 50 kmwide andup to 200 km [37] InA cerana although they havemigratoryswarms flying dozens of km [38] they have a strong tendencyto swarm 5ndash6 times a year and are well adapted to new areassince it is more resistant to some diseases and pest problemsin each area [39]

TheMLandBI phylogenetic analyses of the 28S rRNA andcytb sequences showed the same topologies with four majorclades among Apis spp (for ML Figures 2 and 3) Using Tcollina as the outgroup the 28S rRNA tree placed the dwarfbees in a basal position within the genus while A dorsatawas located in the basal position in the cytb tree consistentwith previous analysis [40] Although both the cytb and28S rRNA phylogenetic trees in this study could not clearlyresolve the relationship between these Apis species theycould separate A dorsata from others (Figures 2 and 3) Aclose relationship has been reported between A dorsata andA cerana in the 28S rRNA tree and molecular indices [32]Although A cerana and A mellifera have been reported tobe closely related to each other [18] they have ecological andbehavioural differences [41] In this study the 28S rRNA andcytb topologies clearly separated A cerana and A mellifera inThailand from each other although they were still closer toeach other compared to the other Apis species

Despite the fact that A dorsata and A florea are open-nesting bees A dorsata was closer to A cerana and Amellifera than to A florea and A andreniformis in this studyThis can be explained by the evolution of behaviour in Apiswhere cavity-nesting is secondarily derived [42] MoreoverA dorsata has a buzzing dance in daylight and at night whileA florea has only a silent waggle dance and cavity-nestingbees adapted this buzzing waggle dance in a dark condi-tion Thus a loss of visual cues was compensated for withsound

5 Conclusion

Overall the genetic variation of Apis spp from 64 popula-tions within 12 provinces in Thailand showed low geneticdiversity indices and variation within and between speciesThis may indicate that habitat fragmentation monocropagriculture industrial management and climate changeshave not affected the diversity of Apis spp in Thailand yetbut clearly further work is of vital importance in establishingthe actual genetic diversity and relationship within Apissubspecies including the analysis of other genes for betterresolution

Data Availability

The authors confirm that all data in the manuscript are freelyavailable

Conflicts of Interest

The authors declare that there are no conflicts of interestregarding the publication of this paper

Acknowledgments

This work was financially supported by the Science Achieve-ment Scholarship of Thailand and the 90th Anniversary ofChulalongkorn University Fund (RatchadaphiseksomphotEndowment Fund)

Supplementary Materials

Supplement 1 detail of sample collection andGenBank acces-sion numbers Supplement 2 frequency () of the nucleotidecompositions of 28S rRNA and cytb regions Supplement 3mean frequency () for base compositions at different codonpositions in the cytb gene fragment Supplement 4 the 16polymorphic sites among eight cytb haplotypes of A ceranain Thailand Only positions that are different from haplotypeHAP1 are indicated (Supplementary Materials)

References

[1] R Winfree J R Reilly I Bartomeus D P Cariveau NM Williams and J Gibbs ldquoSpecies turnover promotes theimportance of bee diversity for crop pollination at regionalscalesrdquo Science vol 359 no 6377 pp 791ndash793 2018

[2] S Wongsiri C Lekprayoon R Thapa et al ldquoComparativebiology of Apis andreniformis and Apis florea in Thailandrdquo BeeWorld vol 78 no 1 pp 23ndash35 1997

[3] A Rattanawannee C Chanchao and S Wongsiri ldquoMorpho-metric and genetic variation of small dwarf honeybees Apisandreniformis Smith 1858 in Thailandrdquo Insect Science vol 14no 6 pp 451ndash460 2007

[4] B P Oldroyd K E Osborne and M Mardan ldquoColony related-ness in aggregations of Apis dorsata Fabricius (HymenopteraApidae)rdquo Insectes Sociaux vol 47 no 1 pp 94-95 2000

[5] A Rattanawannee C Chanchao J Lim S Wongsiri and BP Oldroyd ldquoGenetic structure of a giant honey bee (Apis

8 Psyche

Apis dorsata (KC294229)

ADKA

Apis andreniformis (EU040176)Apis andreniformis (EU040168)Apis andreniformis (EU040165)AACNAAKA

91082

100097

AFCN1AFCN2AFCP1AFCP2AFCP3AFPJ1AFSA1AFSA2AFPJ2AFRA1AFRA2Apis florea (JX982136)

Apis florea (KC170303)

90092

100100

AMCM1AMN2AMCM2AMCM3AMP3AMP2AMN1AMN3AMP1

100093

ACCN1ACCN3ACCN2ACKA1ACKA2ACKA3ACSA1ACSA2

100064

100095

Apis cerana (1) (AP017941)Apis cerana (2) (AP017941)

Apis cerana (GQ162109)ACSU1

ACSU2

Trigono collino (AY575081)

300

Clade II

Clade IV

Clade IV-1

Clade IV-2

Clade I

Clade III

Clade II-1

Clade II-2

Figure 3 Phylogenetic relationship of honeybee species in Thailand based upon cytb sequences by BI and including Apis spp fromGenBank (accession KC294229 EU040176 EU040168 EU040165 JX982136 KC170303 AP017941 AP017942 and GQ162109) Trigonacollina (accession AY575081) was utilized as the outgroup Node support as Bayesian PP and ML BV is shown Sample (speciesgeographiclocation) codes are as in Supplement 1

classifying honeybees between species but not for resolvingwithin species [18] Likewise it was reported that the 28SrRNA was unlikely to recover deep divergences at the familylevel in bees [33] In contrast in combination with othergenes such as opsin cytb and 16S rRNA 28S rRNA appearedto contribute to resolving basal bee divergence

In this study the number of transversions in cytb washigher than the number of transitions in each Apis species(Table 1) This is in agreement with the transitional biasreported previously [18] Within A cerana the cytb sequencedivergence ranged from 026 to 388 but interestingly washigher in the southern populations than in the other regions

Psyche 9

of Thailand (Table 5) The phylogenetic results are consistentwith the molecular diversity indices of A cerana where Acerana populations in southernThailandwere separated fromthose in the East West and Central regions (Figure 3) Theresult could be explained by the biogeography of A ceranaIn addition the data concur with the previous PCR-RFLPanalysis [34]

For A florea only a very low cytb sequence divergencewas detected (0ndash074) with only one haplotype Onlytwo colonies from Samutsongkram province were slightlydifferent from the others (Figure 3)

Overall the sequence divergence of cytb (89 plusmn 50) washigher than that for 28S rRNA (07 plusmn 05) consistent witha higher evolutionary rate for mitochondrial than nucleargenomes [11] Although the mitochondrial cytb showed ahigher rate of variation it still showed a low variation in theseApis spp among the different geographical regions This mayreflect that Apis spp are highly eusocial and polyandrouswhich reduces the substitution rate and counteracts the effectof genetic drift by increasing the effective population size[35] The ancestral genome in diverging populations will bemaintained and so the genetic distances between populationsare reduced [36] The other possible factor leading to alow genetic variation is migration behaviour The migrationdistance ofA dorsata has been reported to be 50 kmwide andup to 200 km [37] InA cerana although they havemigratoryswarms flying dozens of km [38] they have a strong tendencyto swarm 5ndash6 times a year and are well adapted to new areassince it is more resistant to some diseases and pest problemsin each area [39]

TheMLandBI phylogenetic analyses of the 28S rRNA andcytb sequences showed the same topologies with four majorclades among Apis spp (for ML Figures 2 and 3) Using Tcollina as the outgroup the 28S rRNA tree placed the dwarfbees in a basal position within the genus while A dorsatawas located in the basal position in the cytb tree consistentwith previous analysis [40] Although both the cytb and28S rRNA phylogenetic trees in this study could not clearlyresolve the relationship between these Apis species theycould separate A dorsata from others (Figures 2 and 3) Aclose relationship has been reported between A dorsata andA cerana in the 28S rRNA tree and molecular indices [32]Although A cerana and A mellifera have been reported tobe closely related to each other [18] they have ecological andbehavioural differences [41] In this study the 28S rRNA andcytb topologies clearly separated A cerana and A mellifera inThailand from each other although they were still closer toeach other compared to the other Apis species

Despite the fact that A dorsata and A florea are open-nesting bees A dorsata was closer to A cerana and Amellifera than to A florea and A andreniformis in this studyThis can be explained by the evolution of behaviour in Apiswhere cavity-nesting is secondarily derived [42] MoreoverA dorsata has a buzzing dance in daylight and at night whileA florea has only a silent waggle dance and cavity-nestingbees adapted this buzzing waggle dance in a dark condi-tion Thus a loss of visual cues was compensated for withsound

5 Conclusion

Overall the genetic variation of Apis spp from 64 popula-tions within 12 provinces in Thailand showed low geneticdiversity indices and variation within and between speciesThis may indicate that habitat fragmentation monocropagriculture industrial management and climate changeshave not affected the diversity of Apis spp in Thailand yetbut clearly further work is of vital importance in establishingthe actual genetic diversity and relationship within Apissubspecies including the analysis of other genes for betterresolution

Data Availability

The authors confirm that all data in the manuscript are freelyavailable

Conflicts of Interest

The authors declare that there are no conflicts of interestregarding the publication of this paper

Acknowledgments

This work was financially supported by the Science Achieve-ment Scholarship of Thailand and the 90th Anniversary ofChulalongkorn University Fund (RatchadaphiseksomphotEndowment Fund)

Supplementary Materials

Supplement 1 detail of sample collection andGenBank acces-sion numbers Supplement 2 frequency () of the nucleotidecompositions of 28S rRNA and cytb regions Supplement 3mean frequency () for base compositions at different codonpositions in the cytb gene fragment Supplement 4 the 16polymorphic sites among eight cytb haplotypes of A ceranain Thailand Only positions that are different from haplotypeHAP1 are indicated (Supplementary Materials)

References

[1] R Winfree J R Reilly I Bartomeus D P Cariveau NM Williams and J Gibbs ldquoSpecies turnover promotes theimportance of bee diversity for crop pollination at regionalscalesrdquo Science vol 359 no 6377 pp 791ndash793 2018

[2] S Wongsiri C Lekprayoon R Thapa et al ldquoComparativebiology of Apis andreniformis and Apis florea in Thailandrdquo BeeWorld vol 78 no 1 pp 23ndash35 1997

[3] A Rattanawannee C Chanchao and S Wongsiri ldquoMorpho-metric and genetic variation of small dwarf honeybees Apisandreniformis Smith 1858 in Thailandrdquo Insect Science vol 14no 6 pp 451ndash460 2007

[4] B P Oldroyd K E Osborne and M Mardan ldquoColony related-ness in aggregations of Apis dorsata Fabricius (HymenopteraApidae)rdquo Insectes Sociaux vol 47 no 1 pp 94-95 2000

[5] A Rattanawannee C Chanchao J Lim S Wongsiri and BP Oldroyd ldquoGenetic structure of a giant honey bee (Apis

Psyche 9

of Thailand (Table 5) The phylogenetic results are consistentwith the molecular diversity indices of A cerana where Acerana populations in southernThailandwere separated fromthose in the East West and Central regions (Figure 3) Theresult could be explained by the biogeography of A ceranaIn addition the data concur with the previous PCR-RFLPanalysis [34]

For A florea only a very low cytb sequence divergencewas detected (0ndash074) with only one haplotype Onlytwo colonies from Samutsongkram province were slightlydifferent from the others (Figure 3)

Overall the sequence divergence of cytb (89 plusmn 50) washigher than that for 28S rRNA (07 plusmn 05) consistent witha higher evolutionary rate for mitochondrial than nucleargenomes [11] Although the mitochondrial cytb showed ahigher rate of variation it still showed a low variation in theseApis spp among the different geographical regions This mayreflect that Apis spp are highly eusocial and polyandrouswhich reduces the substitution rate and counteracts the effectof genetic drift by increasing the effective population size[35] The ancestral genome in diverging populations will bemaintained and so the genetic distances between populationsare reduced [36] The other possible factor leading to alow genetic variation is migration behaviour The migrationdistance ofA dorsata has been reported to be 50 kmwide andup to 200 km [37] InA cerana although they havemigratoryswarms flying dozens of km [38] they have a strong tendencyto swarm 5ndash6 times a year and are well adapted to new areassince it is more resistant to some diseases and pest problemsin each area [39]

TheMLandBI phylogenetic analyses of the 28S rRNA andcytb sequences showed the same topologies with four majorclades among Apis spp (for ML Figures 2 and 3) Using Tcollina as the outgroup the 28S rRNA tree placed the dwarfbees in a basal position within the genus while A dorsatawas located in the basal position in the cytb tree consistentwith previous analysis [40] Although both the cytb and28S rRNA phylogenetic trees in this study could not clearlyresolve the relationship between these Apis species theycould separate A dorsata from others (Figures 2 and 3) Aclose relationship has been reported between A dorsata andA cerana in the 28S rRNA tree and molecular indices [32]Although A cerana and A mellifera have been reported tobe closely related to each other [18] they have ecological andbehavioural differences [41] In this study the 28S rRNA andcytb topologies clearly separated A cerana and A mellifera inThailand from each other although they were still closer toeach other compared to the other Apis species

Despite the fact that A dorsata and A florea are open-nesting bees A dorsata was closer to A cerana and Amellifera than to A florea and A andreniformis in this studyThis can be explained by the evolution of behaviour in Apiswhere cavity-nesting is secondarily derived [42] MoreoverA dorsata has a buzzing dance in daylight and at night whileA florea has only a silent waggle dance and cavity-nestingbees adapted this buzzing waggle dance in a dark condi-tion Thus a loss of visual cues was compensated for withsound

5 Conclusion

Overall the genetic variation of Apis spp from 64 popula-tions within 12 provinces in Thailand showed low geneticdiversity indices and variation within and between speciesThis may indicate that habitat fragmentation monocropagriculture industrial management and climate changeshave not affected the diversity of Apis spp in Thailand yetbut clearly further work is of vital importance in establishingthe actual genetic diversity and relationship within Apissubspecies including the analysis of other genes for betterresolution

Data Availability

The authors confirm that all data in the manuscript are freelyavailable

Conflicts of Interest

The authors declare that there are no conflicts of interestregarding the publication of this paper

Acknowledgments

This work was financially supported by the Science Achieve-ment Scholarship of Thailand and the 90th Anniversary ofChulalongkorn University Fund (RatchadaphiseksomphotEndowment Fund)

Supplementary Materials

Supplement 1 detail of sample collection andGenBank acces-sion numbers Supplement 2 frequency () of the nucleotidecompositions of 28S rRNA and cytb regions Supplement 3mean frequency () for base compositions at different codonpositions in the cytb gene fragment Supplement 4 the 16polymorphic sites among eight cytb haplotypes of A ceranain Thailand Only positions that are different from haplotypeHAP1 are indicated (Supplementary Materials)

References

[1] R Winfree J R Reilly I Bartomeus D P Cariveau NM Williams and J Gibbs ldquoSpecies turnover promotes theimportance of bee diversity for crop pollination at regionalscalesrdquo Science vol 359 no 6377 pp 791ndash793 2018

[2] S Wongsiri C Lekprayoon R Thapa et al ldquoComparativebiology of Apis andreniformis and Apis florea in Thailandrdquo BeeWorld vol 78 no 1 pp 23ndash35 1997

[3] A Rattanawannee C Chanchao and S Wongsiri ldquoMorpho-metric and genetic variation of small dwarf honeybees Apisandreniformis Smith 1858 in Thailandrdquo Insect Science vol 14no 6 pp 451ndash460 2007

[4] B P Oldroyd K E Osborne and M Mardan ldquoColony related-ness in aggregations of Apis dorsata Fabricius (HymenopteraApidae)rdquo Insectes Sociaux vol 47 no 1 pp 94-95 2000

[5] A Rattanawannee C Chanchao J Lim S Wongsiri and BP Oldroyd ldquoGenetic structure of a giant honey bee (Apis

10 Psyche

dorsata) population in northern Thailand implications forconservationrdquo Insect Conservation and Diversity vol 6 no 1pp 38ndash44 2013

[6] O Songram S Sittipraneed and S Klinbunga ldquoMitochondrialDNAdiversity and genetic differentiation of the honeybee (Apiscerana) in Thailandrdquo Biochemical Genetics vol 44 no 5-6 pp256ndash269 2006

[7] B Branchiccela C Aguirre G Parra P Estay P Zunino andK Antunez ldquoGenetic changes in Apis mellifera after 40 years ofAfricanizationrdquo Apidologie vol 45 no 6 pp 752ndash756 2014

[8] N V Ostroverkhova O L Konusova A N Kucher T NKireeva A A Vorotov and E A Belykh ldquoGenetic diversity ofthe locus COI-COII of mitochondrial DNA in honeybee popu-lations (Apis mellifera L) from the Tomsk regionrdquoGenetika vol51 no 1 pp 89ndash100 2015

[9] N C Chapman B A Harpur J Lim et al ldquoHybrid origins ofAustralian honeybees (Apis mellifera)rdquo Apidologie vol 47 no 1pp 26ndash34 2016

[10] D R Smith and T C Glenn ldquoAllozyme polymorphisms inSpanish honeybees (Apis melliferaiberica)rdquo Journal of Heredityvol 86 no 1 pp 12ndash16 1995

[11] M Bouga V Evangelou D Lykoudis I Cakmak and FHatjina ldquoGenetic structure ofMarchalina hellenica (HemipteraMargarodidae) populations from Turkey preliminary mtDNAsequencing datardquo Biochemical Genetics vol 49 no 11-12 pp683ndash694 2011

[12] J W O Ballard and M C Whitlock ldquoThe incomplete naturalhistory of mitochondriardquo Molecular Ecology vol 13 no 4 pp729ndash744 2004

[13] S Song L J Wheeler and C K Mathews ldquoDeoxyribonu-cleotide pool imbalance stimulates deletions in HeLa cellmitochondrial DNArdquo The Journal of Biological Chemistry vol278 no 45 pp 43893ndash43896 2003

[14] MNeiman andDR Taylor ldquoThe causes ofmutation accumula-tion inmitochondrial genomesrdquo Proceedings of the Royal SocietyB Biological Science vol 276 no 1660 pp 1201ndash1209 2009

[15] M S Meusel and R F A Moritz ldquoTransfer of paternalmitochondrial DNA during fertilization of honeybee (Apismellifera L) eggsrdquo Current Genetics vol 24 no 6 pp 539ndash5431993

[16] J Schmitz and R F A Moritz ldquoSociality and the rate of rDNAsequence evolution in wasps (Vespidae) and honeybees (Apis)rdquoJournal of Molecular Evolution vol 47 no 5 pp 606ndash612 1998

[17] KWongsa O Duangphakdee andA Rattanawannee ldquoGeneticstructure of the Aphis craccivora (Hemiptera Aphididae) fromThailand inferred from mitochondrial COI gene sequencerdquoJournal of Insect Science vol 17 no 4 p 84 2017

[18] M C Arias and W S Sheppard ldquoPhylogenetic relationships ofhoney bees (HymenopteraApinaeApini) inferred fromnuclearand mitochondrial DNA sequence datardquoMolecular Phylogenet-ics and Evolution vol 37 no 1 pp 25ndash35 2005

[19] B N Danforth J Fang and S Sipes ldquoAnalysis of family-level relationships in bees (Hymenoptera Apiformes) using 28Sand two previously unexplored nuclear genes CAD and RNApolymerase IIrdquo Molecular Phylogenetics and Evolution vol 39no 2 pp 358ndash372 2006

[20] S Kumar G Stecher and K Tamura ldquoMEGA7 molecularevolutionary genetics analysis version 70 for bigger datasetsrdquoMolecular Biology and Evolution vol 33 no 7 pp 1870ndash18742016

[21] M Nei Molecular Evolutionary Genetics Columbia UniversityPress New York NY USA 1987

[22] M Nei andW H Li ldquoMathematical model for studying geneticvariation in terms of restriction endonucleasesrdquo Proceedings ofthe National Acadamy of Sciences of the United States of Americavol 76 no 10 pp 5269ndash5273 1979

[23] P Librado and J Rozas ldquoDnaSP v5 a software for comprehen-sive analysis of DNA polymorphism datardquo Bioinformatics vol25 no 11 pp 1451-1452 2009

[24] T H Jukes and C R Cantor ldquoEvolution of protein moleculesrdquoinMammalian Protein Metabolism Academic Press New YorkNY USA 1969

[25] L Excoffier and H E L Lischer ldquoArlequin suite ver 35 anew series of programs to perform population genetics analysesunder Linux and Windowsrdquo Molecular Ecology Resources vol10 no 3 pp 564ndash567 2010

[26] H Akaike ldquoA new look at the statistical model identificationrdquoIEEE Transactions on Automatic Control vol 19 pp 716ndash7231974

[27] G Schwarz ldquoEstimating the dimension of a modelrdquoTheAnnalsof Statistics vol 6 no 2 pp 461ndash464 1978

[28] A S Tanabe ldquoKakusan A computer program to automate theselection of a nucleotide substitution model and the configura-tion of a mixed model on multilocus datardquo Molecular EcologyResources vol 7 no 6 pp 962ndash964 2007

[29] G Jobb A Von Haeseler and K Strimmer ldquoTREEFINDER apowerful graphical analysis environment for molecular phylo-genetics - art no 18 (Retracted article See vol 15 243 2015)rdquoBMC Evolutionary Biology vol 4 p 18 2004

[30] J P Huelsenbeck and F Ronquist ldquoMrBayes bayesian inferenceof phylogenetic treesrdquo Bioinformatics vol 17 no 8 pp 754-7552001

[31] S Schmidt F Driver and P De Barro ldquoThe phylogenetic char-acteristics of three different 28S rRNA gene regions in Encarsia(Insecta Hymenoptera Aphelinidae)rdquo Organisms Diversity ampEvolution vol 6 no 2 pp 127ndash139 2006

[32] N Lo R S Gloag D L Anderson and B P Oldroyd ldquoAmolecular phylogeny of the genus Apis suggests that the gianthoney bee of the Philippines A breviligulaMaa and the plainshoney bee of southern India A indica Fabricius are validspeciesrdquo Systematic Entomology vol 35 no 2 pp 226ndash233 2010

[33] S A Cameron and P Mardulyn ldquoMultiple molecular data setssuggest independent origins of highly eusocial behavior in bees(HymenopteraApinae)rdquo Systematic Biology vol 50 no 2 pp194ndash214 2001

[34] D Sihanuntavong S Sittipraneed and S Klinbunga ldquoMito-chondrial DNA diversity and population structure of the honeybee Apis cerana in Thailandrdquo Journal of Apicultural Researchvol 38 no 3-4 pp 211ndash219 1999

[35] B J Cole ldquoMultiple mating and the evolution of social behaviorin the hymenopterardquo Behavioral Ecology and Sociobiology vol12 no 3 pp 191ndash201 1983

[36] B P Oldroyd A J Smolenski J-M Cornuet et al ldquoLevelsof polyandry and intracolonial genetic relationships in Apisdorsata (Hymenoptera Apidae)rdquo Annals of the EntomologicalSociety of America vol 89 no 2 pp 276ndash283 1996

[37] N Koeniger and G Koeniger ldquoObservations and experimentson migration and dance communication of Apis dorsata in SriLankardquo Journal of Apicultural Research vol 19 no 1 pp 21ndash341980

[38] W S Robinson ldquoApis ceranaswarms abscond to battle andelude hornets (Vespaspp) in northern Thailandrdquo Journal ofApicultural Research vol 52 no 3 pp 160ndash172 2013

Psyche 11

[39] T Atmowidi P Rianti and A Sutrisna ldquoPollination effec-tiveness of Apis cerana Fabricus and Apis mellifera Linnaeus(Hymenoptera Apidae) in Jatropha curcas L (Euphorbiaceae)rdquoBiotropia vol 15 no 2 pp 129ndash134 2008

[40] R Raffiudin and R H Crozier ldquoPhylogenetic analysis ofhoney bee behavioral evolutionrdquo Molecular Phylogenetics andEvolution vol 43 no 2 pp 543ndash552 2007

[41] H K Sharma J K Gupta and B S Rana ldquoResource parti-tioning among Apismellifera and Apis cerana under mid-hillconditions ofHimachal Pradeshrdquo inProceedings of the 4th AsianApiculture Association International Conference pp 213ndash215Kathmandu Nepal 2000

[42] M S Engel and T R Schultz ldquoPhylogeny and behavior inhoney bees (hymenoptera apidae)rdquoAnnals of the EntomologicalSociety of America vol 90 no 1 pp 43ndash53 1997

Hindawiwwwhindawicom

International Journal of

Volume 2018

Zoology

Hindawiwwwhindawicom Volume 2018

Anatomy Research International

PeptidesInternational Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Journal of Parasitology Research

GenomicsInternational Journal of

Hindawiwwwhindawicom Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Hindawiwwwhindawicom Volume 2018

BioinformaticsAdvances in

Marine BiologyJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Neuroscience Journal

Hindawiwwwhindawicom Volume 2018

BioMed Research International

Cell BiologyInternational Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Biochemistry Research International

ArchaeaHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Genetics Research International

Hindawiwwwhindawicom Volume 2018

Advances in

Virolog y Stem Cells International

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Enzyme Research

Hindawiwwwhindawicom Volume 2018

International Journal of

MicrobiologyHindawiwwwhindawicom

Nucleic AcidsJournal of

Volume 2018

Submit your manuscripts atwwwhindawicom

Psyche 11

[39] T Atmowidi P Rianti and A Sutrisna ldquoPollination effec-tiveness of Apis cerana Fabricus and Apis mellifera Linnaeus(Hymenoptera Apidae) in Jatropha curcas L (Euphorbiaceae)rdquoBiotropia vol 15 no 2 pp 129ndash134 2008

[40] R Raffiudin and R H Crozier ldquoPhylogenetic analysis ofhoney bee behavioral evolutionrdquo Molecular Phylogenetics andEvolution vol 43 no 2 pp 543ndash552 2007

[41] H K Sharma J K Gupta and B S Rana ldquoResource parti-tioning among Apismellifera and Apis cerana under mid-hillconditions ofHimachal Pradeshrdquo inProceedings of the 4th AsianApiculture Association International Conference pp 213ndash215Kathmandu Nepal 2000

[42] M S Engel and T R Schultz ldquoPhylogeny and behavior inhoney bees (hymenoptera apidae)rdquoAnnals of the EntomologicalSociety of America vol 90 no 1 pp 43ndash53 1997

Hindawiwwwhindawicom

International Journal of

Volume 2018

Zoology

Hindawiwwwhindawicom Volume 2018

Anatomy Research International

PeptidesInternational Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Journal of Parasitology Research

GenomicsInternational Journal of

Hindawiwwwhindawicom Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Hindawiwwwhindawicom Volume 2018

BioinformaticsAdvances in

Marine BiologyJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Neuroscience Journal

Hindawiwwwhindawicom Volume 2018

BioMed Research International

Cell BiologyInternational Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Biochemistry Research International

ArchaeaHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Genetics Research International

Hindawiwwwhindawicom Volume 2018

Advances in

Virolog y Stem Cells International

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Enzyme Research

Hindawiwwwhindawicom Volume 2018

International Journal of

MicrobiologyHindawiwwwhindawicom

Nucleic AcidsJournal of

Volume 2018

Submit your manuscripts atwwwhindawicom

Hindawiwwwhindawicom

International Journal of

Volume 2018

Zoology

Hindawiwwwhindawicom Volume 2018

Anatomy Research International

PeptidesInternational Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Journal of Parasitology Research

GenomicsInternational Journal of

Hindawiwwwhindawicom Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Hindawiwwwhindawicom Volume 2018

BioinformaticsAdvances in

Marine BiologyJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Neuroscience Journal

Hindawiwwwhindawicom Volume 2018

BioMed Research International

Cell BiologyInternational Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Biochemistry Research International

ArchaeaHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Genetics Research International

Hindawiwwwhindawicom Volume 2018

Advances in

Virolog y Stem Cells International

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Enzyme Research

Hindawiwwwhindawicom Volume 2018

International Journal of

MicrobiologyHindawiwwwhindawicom

Nucleic AcidsJournal of

Volume 2018

Submit your manuscripts atwwwhindawicom