species identification of aphids (insecta: hemiptera ... · pdf fileet al of aphid, . )), ,...

Molecular Ecology Resources (2008) 8, 1189–1201 doi: 10.1111/j.1755-0998.2008.02297.x

© 2008 The AuthorsJournal compilation © 2008 Blackwell Publishing Ltd

Blackwell Publishing LtdDNA BARCODING

Species identification of aphids (Insecta: Hemiptera: Aphididae) through DNA barcodes

R. G. FOOTTIT,* H. E . L . MAW,* C . D. VON DOHLEN† and P. D . N. HEBERT‡*National Environmental Health Program, Invertebrate Biodiversity, Agriculture and Agri-Food Canada, K. W. Neatby Bldg., 960 Carling Ave., Ottawa, ON, Canada K1A 0C6, †Department of Biology, Utah State University, 5305 Old Main Hill, Logan, UT 84322, USA, ‡Biodiversity Institute of Ontario, Department of Integrative Biology, University of Guelph, Guelph, ON, Canada N1G 2W1

Abstract

A 658-bp fragment of mitochondrial DNA from the 5′ region of the mitochondrial cytochromec oxidase 1 (COI) gene has been adopted as the standard DNA barcode region for animallife. In this study, we test its effectiveness in the discrimination of over 300 species of aphidsfrom more than 130 genera. Most (96%) species were well differentiated, and sequence variationwithin species was low, averaging just 0.2%. Despite the complex life cycles and parthenogeneticreproduction of aphids, DNA barcodes are an effective tool for identification.

Keywords: Aphididae, COI, DNA barcoding, mitochondrial DNA, parthenogenesis, speciesidentification

Received 28 December 2007; revision accepted 3 June 2008

Introduction

The aphids (Insecta: Hemiptera: Aphididae) and relatedfamilies Adelgidae and Phylloxeridae are a group ofapproximately 5000 species of small, soft-bodied insects thatfeed on plant phloem using piercing/sucking mouthparts.Aphids have complex life cycles involving many morph-ologically distinct forms, and parthenogenetic generationsalternating with a sexual generation, and in about 10% ofspecies, this is associated with host alternation (Harrewijn& Minks 1987). An evolutionary tendency towards the lossof taxonomically useful characters, and morphologicalplasticity due to host and environmental factors, complicatesthe recognition of species and the analysis of relationshipsat all levels (Foottit 1997). The presence of different morph-ological forms of a single species on different hosts and atdifferent times of the year makes it particularly difficultto correctly identify routinely collected field samples. Yetaccurate identifications are needed because many speciesof aphids are pests in agriculture, forestry and horticulture.In addition to causing direct feeding damage, they are vectors

of numerous plant diseases (Eastop 1977; Harrewijn &Minks 1987; Blackman & Eastop 2000; Harrington & vanEmden 2007). Aphids are also an important invasive riskbecause their winged forms are easily dispersed by windand because feeding aphids are readily transported withtheir plant hosts. Furthermore, their parthenogeneticmode of reproduction means that solitary individualsor small populations can become established and rapidlyincrease in number. As a result, aphids are significanteconomically important invasive pests throughout the world(for example, Stufkens & Teulon 2002; Foottit et al. 2006;Messing et al. 2007).

Reliable identification of species is essential for theintegrated management of pest aphids and for the earlydetection and risk analysis of newly introduced species(Miller & Foottit 2009). Molecular taxonomic approacheshave provided additional valuable characters for the re-solution of taxonomic problems and the discovery of newspecies within the Aphididae (Foottit 1997). DNA barcodinghas been proposed as a standardized approach to thecharacterization of life forms in numerous groups of livingorganisms (Hajibabaei et al. 2007) including the insects(Floyd et al. 2009). In animals, the selected region is the5′-terminus of the mitochondrial cytochrome c oxidasesubunit 1 (COI) gene (Hebert et al. 2003). In a group withnumerous obstacles to taxonomic resolution, DNA barcodes

Correspondence: Dr Robert G. Foottit, Agriculture and Agri-Food Canada, K. W. Neatby Bldg., 960 Carling Ave., Ottawa,Ontario K1A 0C6, Canada. Fax: (613) 759 1701; E-mail: [email protected]

1190 D N A B A R C O D I N G


could aid in the routine identification of species in appliedsettings, the detection of morphologically cryptic species,the detection of host-specific lineages, and the associationof morphologically distinct life-cycle forms within a species(Foottit & Miller 2009). This study examines the utility ofDNA barcoding in achieving these goals through a prelim-inary analysis of sequence variation in the COI barcoderegion for 335 species of Aphididae, with special emphasison the Aphidinae, the most diverse subfamily of aphids.Subsequent papers will present detailed DNA barcoderesults for other subfamilies of the Aphididae and for therelated family Adelgidae.

Methods and materials

Specimen collection and taxon sampling

Taxon assignment follows the current world catalogue ofaphids (Remaudière & Remaudière 1997) with updates tosubfamily names according to Nieto Nafría et al. (1998).Species authorship and date of publication may be foundin Remaudière & Remaudière (1997).

Between 1991 and 2006, aphid samples were collectedinto liquid nitrogen or 95% ethanol for subsequent use inmolecular systematics studies. Voucher specimens from eachcollection were mounted on microscope slides and depositedin the Canadian National Collection of Insects (Agricultureand Agri-Food Canada, Ottawa). Voucher specimens foradditional samples contributed by Dr K. Pike are depositedat the Irrigated Agriculture Research and Extension Center,Washington State University, Prosser. Samples of aphidDNA (particularly Hormaphidinae: Cerataphidini) wereprovided by D. Stern (vouchers in Natural History Museum,London).

Samples were selected to ensure coverage of mostsubfamilies of the Aphididae and a wide range of generawithin the subfamily Aphidinae. Meyer & Paulay (2005)concluded that the lack of broad geographical samplingfor single species was likely to have resulted in a seriousunderestimation of within-species variation, and that thefailure to survey closely related species would overestimatesequence divergences between congeneric taxa. To addressthe issue of intraspecific variation, we studied samples ofselected species (Aphis fabae, Aphis gossypii, Aphis spiraecola)from widely separated locales, while for four genera(Aphis, Ericaphis, Macrosiphum, Uroleucon), we analysed asmany species as available. We also tested the effects ofcomprehensive sampling on levels of sequence diver-gence among species in the genus Aphis. The completedata set includes 690 samples, covering 335 species,134 genera and 16 subfamilies (Table 1). Associatedspecimen information is available in the ‘Barcoding theAphididae’ project on Barcode of Life Data Systems (BOLD;www.barcodinglife.org).

DNA extraction, amplification and sequencing

Single aphid specimens, stored in 95% ethanol, weretransferred to coded tubes in a Matrix box (TrakMatesmicroplate system; Matrix Technologies), and sent to theBiodiversity Institute of Ontario (BIO) for DNA extractionand sequencing. Alternatively, DNA was extracted, vacuumdried and sent to BIO in 96-well plates. Standard protocols(Hajibabaei et al. 2005) were employed for DNA extractionand amplification, sequencing of the COI barcode region,sequence editing and alignment. Both sequence informationand collection/taxonomic information for each specimenwere entered in BOLD (Hebert & Ratnasingham 2007). TotalDNA was extracted from individual specimens and theprimer pair LepF (ATTCAACCAATCATAAAGATATTGG)and LepR (TAAACTTCTGGATGTCCAAAAAATCA) wasused to amplify an approximately 700 bp DNA fragmentof mitochondrial CO1 which was subsequently sequencedin both directions using the same primers. All sequencesobtained in this study have been deposited in GenBank(accession nos EU701270–EU701959) and are also accessibleon BOLD (www.barcodinglife.org, ‘Barcoding the Aphididae’project).

Data analysis

Electropherograms for the CO1 gene were edited and alignedwith Sequencher (version 4.5; Gene Codes Corporation).Pairwise nucleotide sequence divergences were calculatedusing the Kimura 2-parameter model of base substitution(Kimura 1980), the best model for species level analysiswith low distances (Hebert et al. 2003), and neighbour-joining(NJ) analysis (Saitou & Nei 1987) was used to examinerelationships among taxa and population samples. NJ treeswere produced using the taxon ID tree function on BOLD.

Results

The results of the overall NJ analysis of distances amongthe samples of 335 species are summarized in Fig. 1. Itshould be noted that the tree presented here is intended asa representation of the distance matrix only, and should notbe interpreted as a phylogenetic hypothesis. Most of thenominal species showed very limited intraspecific variation(mean 0.201%, SE 0.004), while sequence divergences amongcongeneric taxa averaged 7.25% (Table 2, Fig. 2; range0.46–13.1%, mean 7.25%, SE 0.013% for all Aphididae tested;range 0.46–11.3%, mean 7.22%, SE 0.013% for Aphidinaeonly). The exceptional cases of low distance betweencongeners (Table 3) are representatives of groups with knowntaxonomic problems with clusters of morphologically verysimilar species, as in certain Aphis and Illinoia. Among thelimited number of species examined, the high values seemto occur in biologically diverse genera lacking obvious

www.barcodinglife.org


D N A B A R C O D I N G 1191


apomorphic morphological characters, such as Acyrthosiphon,or those with several described subgenera, such as Myzusand Nasonovia.

Table 4 summarizes within-species divergence for repli-cated species. Only a few of these (Myzus cerasi, Macro-siphum euphorbiae, Aphis coreopsidis, Neomyzus circumflexus,Sitiobion avenae) had within-species divergences thatexceeded 1%.

Some subfamilies form more discrete clusters on thetaxon ID tree than others (Fig. 1). The Aphidinae, withthe embedded Pterocommatinae, form a single clusteras do the Lachninae. Other diverse groups, such as theEriosomatinae and Calaphidinae are not as cohesivealthough certain recognized tribes, subtribes and othergeneric groupings within these subfamilies form distinctclusters. If nearest neighbours are used to classify unknownspecimens, about 15% of these were placed in a wrongsubfamily, although all Aphidinae and Lachninae werecorrectly classified. This was calculated by removing allspecies in a genus from those subfamilies with morethan one genus represented in the barcode data set and

determining the nearest neighbour within the remainingdata set.

The node in Fig. 1 consisting of the Aphidinae andPterocommatinae, is expanded in Fig. 3 and subgroups ofthe Aphidinae are expanded in Figs 4–8 (Tribe Macrosiphini)and in Figs 9–12 (Tribe Aphidini). In the subfamily Aphidinae,species of most genera form distinct clusters. The exceptionsare a few genera (Acyrthosiphon, Ericaphis, Nasonovia, Myzus)which may be polyphyletic. For example, most Myzusspecies cluster together (Fig. 6a), but M. varians and M.(Sciamyzus) ascalonicus are well removed from the mainMyzus cluster (see Fig. 3).

Pairwise sequence divergences among samples withinreplicated species are given in Table 4. Figure 13 shows theNJ analysis of 84 samples of the completely parthenoge-netic species, Aphis gossypii, collected from Germany,Canada, continental USA, Hawaii and Micronesia. Themajority of samples (n = 42), representing the entire geo-graphical range that was represented, possess identicalbarcodes. The maximum sequence divergence among thesamples of this species is 0.62%.

Table 1 Summary of taxonomic distribution material sampled relative to known world fauna. A full list of material and associated data isavailable in the ‘Barcoding the Aphididae’ file on BOLD (www.barcodinglife.org). Classification follows Remaudière & Remaudière (1997)and Nieto Nafría et al. (1998)

Subfamily

No. of taxa sampled No. of taxa in world

Genera Species Recognized genera Described species & subspecies

Anoeciinae 1 1 1 24Aphidinae 68 218 337 2860Calaphidinae 18 22 91 356Chaitophorinae 2 9 11 178Drepanosiphinae 2 4 5 37Eriosomatinae 16 35 60 369Greenideinae 1 2 16 173Hormaphidinae 7 7 41 197Lachninae 8 18 18 397Lizeriinae 1 1 3 34Mindarinae 1 2 1 9Pterocommatinae 2 5 5 57Neophyllaphidinae 1 1 1 12Phyllaphidinae 2 3 2 14Saltusaphidinae 3 3 12 71Tamaliinae 1 4 1 4

Subfamilies not sampledAiceoninae – – 1 17Israelaphidinae – – 1 4Macropodaphidinae – – 1 10Parachaitophorinae – – 1 3Parastheniinae – – 2 6Phloeomyzinae – – 1 2Spicaphidinae – – 2 13Taiwanaphidinae – – 2 13Thelaxinae – – 4 19

Totals 134 335 620 4879




Fig. 1 Basal nodes of neighbour-joining tree based on distances from Kimura 2-parameter model. Species belonging to miscellaneous taxaand species not clustering with other members of the higher taxon to which they belong are shown individually, while clusterscorresponding to major recognized taxa are shown as boxes attached to the basal node of the cluster. The node representing subfamiliesAphidinae and Pterocommatinae is expanded in Figs 3–12. Nodes corresponding to other recognized taxa indicate included number of taxa,and will be dealt with in future publications. Terminal species clusters with identical sequence are collapsed into a single terminal node withindication of the number of identical samples given in brackets.



Discussion

For the most part, species of the Aphidinae correspond toa cluster of similar barcode sequences and these clustersare distinct from neighbouring barcode clusters. Although

96% of sequence divergences among species pairs wasgreater than 3%, there were cases where morphologicallyand biologically well-delineated species, such as species inthe genera Aphis and Illinoia, exhibited divergences of lessthan 1%. For example, Illinoia morrisoni which feeds onconifers in western North America and Illinoia spiraecolawhich feeds on Spiraea in eastern North America differ byonly 0.61% sequence divergence (Table 3). In most caseswhere we included replicated samples of species, intraspecificdivergence was usually less than 0.4% (Table 4). Weanalysed sequence variation in a large number of Aphisgossypii, a widespread obligatorily parthenogenetic pest,from sites in Micronesia, North America and Europe. Thisspecies is known to have several host-associated genotypiclineages (Guldemond et al. 1994; Chavigny & Vanlerberghe-Masutti 1998) and although some clusters were detected byDNA barcoding, the maximum divergence within thespecies was still less than 0.62% (Fig. 13). By contrast, speciesof Aphis that are morphologically very similar to A. gossypii(the species in Fig. 12c bounded by A. gossypii and soybeanaphid, Aphis glycines) showed mean divergences fromA. gossypii of 1.6% to 4.3%. The clear delineation of A. gossypii

Table 2 Summary of pairwise sequence divergences among congeneric species of Aphidinae

GenusNo. of species: examined (in world)

Distance (range, percentage)

Acyrthosiphon 6 (84) 4.88 to 8.61Amphorophora 4 (23) 0.93 to 6.12Aphis 53 (604) 0.46 to 11.05Aulacorthum 3 (41) 3.59 to 7.91 (excluding Neomyzus)Brachycaudus 3 (50) 3.92 to 5.59Capitophorus 4 (31) 6.05 to 8.61Carolinaia 2 (18) 5.42Cavariella 3 (40) 5.93 to 7.91Cedoaphis 2* (2) 4.24 to 5.06Ericaphis 5 (11) 1.24 to 6.41Hyperomyzus 4 (19) 1.54 to 6.47Illinoia 6 (41) 0.77 to 4.65 (incl. subgenera Illinoia and Amphorophorina)Macrosiphoniella 4 (143) 5.06 to 6.54Macrosiphum 15 (139) 1.08 to 6.97Muscaphis 4 (8) 4.09 to 5.94Myzus 6 (66) 5.55 to 11.3 (incl. subgenera Myzus, Nectarosiphon, Sciamyzus)Nasonovia 5 (47) 4.73 to 6.05 (incl. subgenera Ranakimia and Kakimia)Nearctaphis 3 (10) 3.28 to 4.57Paradoxaphis 2 (2) 4.77Pleotrichophorus 2 (56) 5.71Pseudoepameibaphis 2 (4) 4.41Rhopalosiphum 7 (19) 0.96 to 8.27Sitobion 2 (82) 1.39 to 1.66Toxoptera 2 (5) 7.14 to 7.91Uroleucon 7 (215) 2.66 to 3.01 (incl. subgenera Uroleucon, Uromelan, Lambersius)Utamphorophora 2 (7) 7.16

*examined material includes at least one undescribed species.

Fig. 2 Frequency distribution of pairwise nearest-neighbourdistances among congeneric species in subfamily Aphidinae.



indicates that continuous parthenogenetic reproduction byitself does not negatively influence the utility of barcodesfor identifying species.

Some species, such as Neomyzus circumflexus, exhibitrelatively large intraspecific variation in sequence diver-gence (3.14%, Table 4). This indicates the possibility of thepresence of multiple cryptic taxa within the current speciesdefinition.

In general, congeneric species of Aphidinae formedcohesive barcode assemblages of well-delineated species,so that DNA barcodes are useful in many cases for genus-level identifications. However, we intentionally includedrepresentative genera that are known to be taxonomically

unresolved, and the above-mentioned correspondence ofspecies and genera to discrete clusters of barcodes wasviolated in these areas of known taxonomic instability. Forexample, genus Macrosiphum is likely paraphyletic withrespect to the genus Illinoia, and Illinoia (sensu lato) itself issuspected of being polyphyletic, a conclusion supported

Fig. 3 Expansion of node from neighbour-joining analysis ofFig. 1 comprising Aphidinae and Pterocommatinae. The majorityof the members of tribe Macrosiphini of subfamily Aphidinaeform a single cluster expanded in Fig. 4. The majority of membersof subtribe Rhopalosiphina of tribe Aphidini form a single clusterexpanded in Fig. 9 (but Melanaphis appears in this figure,Hyalopterus in Fig. 10). All members of subtribe Aphidina oftribe Aphidini, plus one Macrosiphonine (Lipaphis) and oneRhopalosiphonine (Hyalopterus) form a single cluster expanded inFig. 10. The remaining genera except the Pterocommatinae(Pterocomma and Plocamaphis) belong to tribe Macrosiphini.Terminal species clusters with identical sequence are collapsedinto a single terminal node with indication of the number ofidentical samples.

Fig. 4 Expansion of node in Fig. 3 containing taxa belonging tosubtribe Macrosiphini. A subset of Macrosiphini is expanded inFig. 5. Two other clusters corresponding to groups of biologicallysimilar genera are expanded in Fig. 6. Terminal species clusterswith identical sequence are collapsed into a single terminal nodewith indication of the number of identical samples.



by the barcode results. For example, species now placed inthe Illinoia subgenus Amphorophorina cluster with certainspecies of Macrosiphum rather than with other Illinoiaspecies (Fig. 7). This is also the situation within the generaAulacorthum and Ericaphis. Ericaphis gentneri (Fig. 5) isdistant from the remainder of the genus Ericaphis (Fig. 8).Conversely, indigenous North American species currentlyplaced in Aulacorthum (A. dorsatum and A. pterinigrum), andassociated with Ericaceae, have barcodes similar to thoseof the main group of Ericaphis species, most of which feedon Ericaceae and Rosaceae (Fig. 8), but unlike that of thecosmopolitan polyphagous Aulacorthum solani. Interestingly,this alliance is also reflected in morphology. In fact, specimensof A. dorsatum have the same barcode sequence as somespecimens identified as Ericaphis wakibae. The need for furthertaxonomic analysis of these genera is thus indicated. Thegenus Aphis is extremely diverse, with over 400 describedspecies and attempts have been made to designate subgenerawithin Aphis or to subdivide it into several genera. Althoughcertain groups, such as Braggia (Fig. 11c) and Zyxaphis (seeFig. 10), are well defined with respect to barcode sequence,

they are not well separated from Aphis as a whole. Similarly,species allied to Aphis species, but with morphologicalsingularities, that are currently placed in distinct genera(Siphonatrophia, Sanbornia and Brachyunguis), join to taxacurrently belonging to Aphis sensu lato. The two species

Table 3 Range of pairwise interspecific distance in taxa for whichpairwise sequence divergence is low (less than 2%)

Genus Species 1 Species 2Percentage divergence

Aphis manitobensis vs. varians 0.46 to 0.77gossypii vs. oestlundi 1.54 to 1.86gossypii vs. ichigo 1.76 to 1.93ichigo vs. idaei 1.60ichigo vs. oestlundi 1.60decepta vs. clydesmithi 1.80decepta vs. helianthi 1.86decepta vs. ceanothi 1.86

Illinoia azaleae vs. kalmiaflora 0.77 to 0.92liriodendri vs. kalmiaflora 1.70liriodendri vs. azaleae 1.33 to 1.39liriodendri vs. spiraecola 1.23liriodendri vs. morrisoni 1.55morrisoni vs. kalmiaflora 1.70morrisoni vs. spiraecola 0.61spiraecola vs. azaleae 1.39 to 1.70spiraecola vs. kalmiaflora 1.88

Amphorophora agathonica vs. rubicumberlandi 0.93Muscaphis stroyani vs. musci 0.93Rhopalosiphum insertum vs. enigmae 0.96 to 1.15

insertum vs. enigmae 1.15Macrosiphum albifrons vs. daphnidis 1.55

euphorbiae vs. daphnidis 1.08euphorbiae vs. albifrons 1.54gaurae vs. albifrons 1.38

Ericaphis fimbriata vs. scammelli 1.24 to 1.70Sitobion avenae vs. phyllanthi 1.39 to 1.66

Fig. 5 Expansion of Macrosiphine node in Fig. 4. A clustercontaining four genera with distinctly reticulate siphunculi (plussome members of two other genera) is expanded in Fig. 7(Macrosiphum group). A group of species associated with Ericaceae(Ericaphis group) is expanded in Fig. 8. Terminal species clusterswith identical sequence are collapsed into a single terminal nodewith indication of the number of identical samples.



Table 4 Observed range of pairwise within-species sample divergences for species of Aphidinae replicated over broad geographical range,with number and geographical origin of samples

SpeciesDistance (range, percentage) n Geographical origin of samples

A. Myzus cerasi 0 to 1.27 7 Canada (ON, BC), USA (WA)Macrosiphum euphorbiae 0 to 1.24 4 USA (WA, HI)(excluding one Hawaiian spm) 0 to 0.31 3 USA (WA, HI)Aphis coreopsidis 1.23 2 USA (NC, HI)Neomyzus circumflexus 3.14 2 New Zealand, Columbia

(border interceptions in USA)Sitobion avenae 1.17 2 Canada (ON), USA (SC)Aphis (Protaphis) middletonii 0 to 0.79 9 Canada (ON, QC, AB), USA (NC, UT)Aphis varians 0.15 to 0.77 3 Canada (NB, AB, BC)Illinoia azaleae 0 to 0.92 5 Canada (NL, QC, ON), USA (HI)Ericaphis scammelli 0.75 3 Canada (BC), USA (WA)

B. Acyrthosiphon macrosiphum 0 2 Canada (BC), USA (CA)Acyrthosiphon pisum 0 8 Canada (NB, QC, ON), USA (WA)Aphis craccivora 0 to 0.32 6 Canada (NB), USA (CO, HI),

CNMI (Saipan), Marshal Is. (Majuro)Aphis fabae 0.16 10 USA (FL, WA), Canada (ON, BC),

France, SpainAphis farinosa 0.15 to 0.31 3 Canada (QC, BC), USA (WA)Aphis glycines 0 to 0.18 6 Australia(NSW), China, Japan,

Canada (QC, ON), USA (KY)Aphis gossypii 0.62 82 Denmark, Germany, Canada (ON, MB),

USA (CA, HI), Marshal Is. (Majuro), CNMI (Saipan, Rota), Guam, FSM (Kosrae), Palau, South Korea

Aphis helianthi 0 to 0.31 10 Canada (ON, AB, BC), USA (UT, WA)Aphis lugentis 0.31 2 Canada (BC), USA (NC)Aphis maculatae 0 2 Canada (NB, ON)Aphis neilliae 0.62 3 Canada (ON), USA (NC)Aphis neogillettei 0 2 Canada (NB, ON)Aphis nerii 0 2 Canada (ON), CNMI (Saipan)Aphis pomi 0 4 Canada (NS, ON, BC), USA (NY)Aphis spiraecola 0 14 USA (NC, HI), Palau, Marshal Is.

(Majuro), GuamAphis spiraephila 0.15 2 Canada (NB, ON)Aphthargelia symphoricarpi 0.15 2 Canada (SK), USA (UT)Aulacorthum solani 0 to 0.15 4 Canada (ON), USA (HI), Costa RicaBrachycaudus helichrysi 0.31 3 USA (WA, HI)Braggia eriogoni 0 to 0.31 4 Canada (BC), USA (CA, CO)Hayhurstia atriplicis 0 3 Mexico, Canada (AB, ON)Hyalopterus pruni 0 to 0.31 5 Canada (BC, ON)Lipaphis pseudobrassicae 0.15 to 0.31 3 Canada (ON), USA (WA, HI)Myzus lythri 0 3 Canada (ON), USA (WA)Myzus persicae 0 10 Canada (ON, BC), USA (WA)Rhopalosiphum cerasifoliae 0 6 Canada (NB, ON, SK, BC)Rhopalosiphum padi 0 3 Canada (ON), USA (WA)Toxoptera aurantii 0 6 USA (HI), CNMI (Rota)Toxoptera citricidus 0 4 New Zealand, Guam, USA (FL,HI)Wahlgreniella nervata 0 to 0.32 4 Canada (BC), USA (CA, OR)

(A) Species for which intraspecific distances exceed 0.7%. (B) Species with low intraspecific pairwise distances (< 0.7%) with geographically widely separated samples (note: standard postal abbreviations used for Canadian provinces and US states. CNMI, Commonwealth of Northern Mariana Islands; FSM, Federated States of Micronesia; AB, Alberta; BC, British Columbia; MB, Manitoba; NB, New Brunswick; NL, Newfoundland and Labrador; NS, Nova Scotia; ON, Ontario; QC, Quebec; SK, Saskatchewan; CA, California; CO, Colorado; FL, Florida; HI, Hawaii; KY, Kentucky; NC, North Carolina; OR, Oregon; SC, South Carolina; UT, Utah; WA, Washington; NSW, New South Wales).



placed in the endemic New Zealand genus Paradoxaphis(Fig. 11b) do not form a group distinct from two otherendemic New Zealand species currently placed withinthe genus Aphis, although these species together form asingle cluster. The fact that these known areas of taxonomicuncertainty are highlighted by DNA barcodes emphasizesthe utility of this methodology in revealing generic placementsthat should be reconsidered, and in highlighting inadequategeneric definitions.

Although we found that DNA barcodes are likely to bevery useful in identifying a newly encountered aphidspecimen to species or, with caution, to genus, we note thatidentification cannot reliably be extended to deeper levels(such as tribe or subfamily). If nearest neighbours are used

to classify unknown specimens, about 15% of aphid speciesare placed in the wrong subfamily. This is a particular problemwithin the Eriosomatinae and Calaphidinae, whose membertaxa are positioned in various sections of the NJ tree (Fig. 1).This is most likely a joint result of the low taxonomiccoverage and of sequence convergence due to saturationof mutations at third base positions.

Fig. 6 Expansion of nodes from Fig. 4. (a) Myzus group. Manymembers of these genera are host alternating with primary host inthe Rosaceae (Prunus for Brachycaudus and Myzus, Maloidea forDysaphis). Myzus varians and Myzus (Sciamyzus) ascalonicus do notfall within this cluster (see Fig. 3). (b) Uroleucon group. Members ofthis group have a nonalternating life cycle, with hosts usually inAsteraceae (usually Astereae for Uroleucon, Anthemidae forMacrosiphoniella and Metopeurum). Terminal species clusters withidentical sequence are collapsed into a single terminal node withindication of the number of identical samples.

Fig. 7 Expansion of node from Fig. 5. Members of this cluster havesimilar habitus characterized by elongate bodies with attenuateappendages. Most (genera Macrosiphum, Illinoia, Corylobium andCatamergus) have siphunculi with apical reticulate sculpture. Thesubcluster at (a) comprises species from both Macrosiphum andsubgenus Amphorophorina of Illinoia associated with rosaceousshrubs and Caprifoliaceae (sensu lato). The nominate subgenus ofIllinoia forms a discrete cluster at (b). Terminal species clusterswith identical sequence are collapsed into a single terminal nodewith indication of the number of identical samples.



In summary, the present study has shown that DNAbarcodes will be valuable in routine identifications ofunidentified aphid specimens, a capacity that will be par-ticularly important in the detection and management ofinvasive and pest aphid species. Additionally, DNA barcodeswill aid the discovery of morphologically cryptic aphid taxa(Miller & Foottit 2009). However, there is a requirement for

Fig. 10 Expansion of Aphidina node of Fig. 3. Several nodes areexpanded in Figs 11 and 12 as indicated. Terminal species clusterswith identical sequence are collapsed into a single terminal nodewith indication of the number of identical samples. Lipaphis andHyalopterus fall within this cluster but are not members ofAphidina.

Fig. 8 Expansion of Ericaphis group node from Fig. 5, consistingof primarily western North American species associated withEricaceae or Rosaceae (except Ericaphis lilii on lilies). Included arethe North American species currently assigned to Aulacorthum,but with morphological and biological similarities to Ericaphis.Excluded is Ericaphis gentneri from Crataegus (see Fig. 5). Ter-minal species clusters with identical sequence are collapsed into asingle terminal node with indication of the number of identicalsamples.

Fig. 9 Expansion of Rhopalosiphina node of Fig. 3. Members ofthis subtribe are compact-bodied, with short appendages, mostlyassociated with graminoid monocots. Host-alternating specieshave primary hosts in Rosaceae (Amydaloidea and Maloidea).Terminal species clusters with identical sequence are collapsedinto a single terminal node with indication of the number ofidentical samples.



thorough taxonomic analysis of each major group of aphidsbefore DNA barcoding can be routinely utilized, becausemany of the invasive and pest aphid species belong totaxonomic groups which require further taxonomic resolu-tion. The end result of an integrative taxonomic approach,with DNA barcodes playing an important role, will be theestablishment of a stable taxonomy for the Aphididaeand the future development of reliable morphological andmolecular catalogues for this diverse and important group.The insights into problems of species delineation suggestedby this analysis will be expanded into a series of paperstreating the specific taxonomic issues involved.

Fig. 12 Expansion of nodes from Fig. 10. (a) the so-called ‘black’and ‘black-backed’ Aphis species, at one time placed in genusPergandeida, plus Aphis thalictri. The latter is morphologicallyand biologically distinct from the other species in the cluster.(b) Species currently placed in subgenus Bursaphis, associatedwith Ribes (winter host) and Onagraceae (summer hosts). Aphisequiseticola differs in biology but is morphologically similar. (c)Species with morphological similarities to Aphis gossypii; three areassociated with Rubus (Aphis rubifolii, A. ichigo and A. idaei), threewith Lamiaceae, and three use Rhamnus species as winter hosts(as do the closest host-alternating relatives of A. gossypii).

Fig. 11 Expansion of nodes from Fig. 10. (a) Aphis subgenusProtaphis. Several North American species assigned to thissubgenus have recently been synonymized (Eastop & Blackman2005) under the name Aphis middletoni. (b) Four species endemic toNew Zealand currently placed in different genera. One otherspecies endemic to New Zealand, Aphis coprosmae, appears inFig. 10. (c) Genus Braggia, a group of poorly known speciesassociated with Eriogonum in western North America. (d) A groupof mainly eastern North American species mostly associated withshrubs or with Apiaceae (or with alternation between the two). (e)Aphis pomi and related species. Terminal species clusters withidentical sequence are collapsed into a single terminal node withindication of the number of identical samples.



Fig. 13 Tree showing neighbour-joining analysis tree based on Kimura 2-parameter distance for 84 samples of Aphis gossypii. Terminalnodes labelled with country, region of origin and host plant.



Acknowledgements

This work was funded by Agriculture and Agri-Food Canada andthrough grants from NSERC, Genome Canada, the Gordon andBetty Moore Foundation to P.D.N.H. We thank Keith Pike (IrrigatedAgriculture Research and Extension Center, Washington StateUniversity, Prosser), Ross Miller (College of Natural and AppliedSciences, University of Guam, Mangilao), and David Stern (PrincetonUniversity) for providing aphid specimens or DNA. We thankPhilana Dollin, Kendra Duffin, Sophie Berolo and Bryan Brunetfor extracting DNA and Beverly Hymes for the preparation ofslide mounts of voucher specimens. We thank personnel at theBiodiversity Institute of Ontario, particularly, Jeremy deWaard,Robin Floyd, Rob Dooh, Alex Borisenko and Claudia Kleint-Steinkefor processing samples and data management.

References

Blackman RL, Eastop VF (2000) Aphids on the World’s Crops: anIdentification and Information Guide, 2nd edn. John Wiley & SonsLtd., England.

Eastop VF (1977) Worldwide importance of aphids as virus vectors.In: Aphids as Virus Vector (eds Harris KF, Maramorosh K), pp. 4–62.Academic Press, New York. 559 pp

Eastop VF, Blackman RL (2005) Some new synonyms in Aphididae(Hemiptera: Sternorrhyncha). Zootaxa, 1089, 1–36.

van Emden HF, Harrington R (2007) Aphids as Crop Pests. CABInternational, Oxford, UK.

Floyd RM, Wilson JJ, Hebert PDN (2009) DNA barcodes and insectbiodiversity. In: Insect Biodiversity: Science and Society (eds FoottitRG, Adler PH), Blackwell Publishing. In press.

Foottit RG (1997) Recognition of parthenogenetic insect species.In: Species. The Units of Biodiversity (eds Claridge MF, DawahHA, Wilson MR), pp. 291–307. Chapman & Hall, London.

Foottit RG, Halbert SE, Miller GL, Maw E, Russell LM (2006)Adventive aphids (Hemiptera: Aphididae) of America North ofMexico. Proceedings of the Entomological Society of Washington,108, 583–610.

Guldemond JA, Tigges WT, DeVrijer PWF (1994) Host races ofAphis gossypii (Homoptera: Aphididae) on cucumber andchrysanthemum. Environmental Entomology, 23, 1235–1240.

Hajibabaei M, deWaard JR, Ivanova NV et al. (2005) Critical factorsfor assembling a high Volume of DNA barcodes. PhilosophicalTransactions of the Royal Society of London. Series B: BiologicalSciences, 360, 1959–1967.

Hajibabaei M, Singer GAC, Hebert PDN, Hickey DA (2007)DNA barcoding: how it complements taxonomy, molecularphylogenetics and population genetics. Trends in Genetics, 23,167–172.

Hebert PDN, Cywinska A, Ball SL, DeWaard JR (2003) Biologicalidentifications through DNA barcodes. Proceedings of the RoyalSociety B: Biological Sciences, 270, 313–322.

Kimura M (1980) A simple method for estimating evolutionaryrate of base substitutions through comparative studies of nucleotidesequences. Journal of Molecular Evolution, 16, 111–120.

Messing RH, Tremblay MN, Mondor EB, Foottit RG, Pike KS (2007)Invasive aphids attack native Hawaiian plants. Biological Invasions,9, 601–607.

Meyer CP, Paulay G (2005) DNA barcoding: error rates based oncomprehensive sampling. Public Library of Science Biology, 3,2229–2238.

Miller GL, Foottit RG (2009) The taxonomy of crop pests: Theaphids. In: Insect Biodiversity: Science and Society (eds Foottit RG,Adler PH). Blackwell Publishing. In press.

Minks AK, Harrewijn P (eds) (1987) Aphids: Their Biology, NaturalEnemies and Control. Elsevier, New York.

Nieto Nafría JM, Mier Durante MP, Remaudière G (1998) Les nomsdes taxa du groupe-famille chez les Aphididae (Hemiptera).Revue Française d’Entomologie (N. S.), [1997], 19, 77–92.

Ratnasingham S, Hebert PDN (2007) BOLD: The Barcode of LifeData System (www.barcodinglife.org). Molecular Ecology Notes,7, 355–364.

Remaudière G, Remaudière M (1997) Catalogue des Aphididae dumonde/ Catalogue of the World’s Aphididae. Homoptera Aphidoidea.INRA Editions, Versailles, France.

Saitou N, Nei M (1987) The neighbour-joining method: a newmethod for reconstructing phylogenetic trees. Molecular Biologyand Evolution, 4, 406–425.

Teulon DAJ, Stufkens MAW (2002) Biosecurity and aphids in NewZealand. New Zealand Plant Protection, 55, 12–17.

Vanlerberghe-Masutti F, Chavigny P (1998) Host-based geneticdifferentiation in the aphid Aphis gossypii Glover, evidencedfrom RAPD fingerprints. Molecular Ecology, 7, 905–914.


species identification of aphids (insecta: hemiptera ... · pdf fileet al of aphid, . )), ,...

Documents