Chaoboridae, a family of Diptera, is commonly known as phantom midges. These are common midges with cosmopoli-tan distribution. Aquatic larvae of Chaoborus, a common genus of the family, are widely distributed in lakes, ponds, and reservoirs. Omnivorous Chaoborus larvae are crucial predators in structuring zooplankton communities, especially for the small-sized prey such as Daphnia, water flea. Larvae of Chaoborus are commonly known to produce predator-induced polyphenism in Daphnia sp., which is a morphologi-cal defense for planktonic crustaceans by adaptive develop-mental plasticity (Tollrian and Dodson, 1999; Simon et al., 2011).

Despite the important ecological role of Chaoborus, their taxonomy and phylogenetic relationships remain unresolved. According to Dupuis et al. (2008), the monophyletic relation-ships of species in Chaoborus were highly questionable, and several cryptic species have been suggested. Especially, 2 cryptic species in Chaoborus flavicans were indicated accord-ing to its habitats, morphological characters, and mitochon-drial cytochrome oxidase I (COI) sequences. Based on the morphological characters and 18S and 5.8S ribosomal DNA sequences, it has been suggested that Chaoboridae is more closely related to the Culicidae than Ceratopogonidae in the culicomorphan Diptera (Miller et al., 1997; Sæther, 2000). In other molecular phylogenetic studies (Friedrich and Tautz, 1997; Cameron et al., 2007) conducted with the purpose of testing the traditional hypotheses on relationships between families of the dipterans, the species in Chaoboridae were not included in analyses.

Species in Chaoboridae have never known to inhabit Korea until recently, and even the family name did not appear in the Checklist of Insects from Korea (The Entomological Society of Korea and Korean Society of Applied Entomology, 1994). Recently, Jeong (2010) reported Chaoborus flavicans from the Sangchun reservoir in Gyeonggi-do, based mainly on mandible morphology and partial COI sequences. Therefore, their taxonomy and distribution are very poorly understood in Korea. In this study, molecular identification and phylo-genetic relationships of Chaoborus were analyzed as a fun-damental study for understanding the role of Chaoborus in predator-prey interactions in a freshwater ecosystem.

Larvae of Chaoborus were collected by a Bongo net of 60 cm mouth diameter and 300 ㎛ mesh size in the Sangchun reservoir in Gyeonggi-do and the Ildae reservoir in Jeolla-nam-do. Collected specimens were preserved in 100% ethyl alcohol before storing at -20℃ until DNA extraction. DNA was extracted by using an AccuPrep Genomic DNA extrac-tion kit (Bioneer, Korea).

For molecular identification, the partial mitochondrial COI region was PCR amplified by employing primers COIF and COIR (Table 1), in a total volume of 20 μL consisting of 2× TOPsimple DyeMIX-Tenuto (Enzynomics, Korea), ~100 ng template DNA, 2 pmol dNTPs, and 5 pmol of each primer. The PCR protocol consisted of initial denaturation at 94℃ for 3 min, 35 cycles of 94℃ for 30 sec, 45℃ for 30 sec, and 72℃ for 1 min, followed by final extension, 72℃ for 7 min. PCR products were purified using the AccuPrep PCR Purifi-cation kit (Bioneer). Sequencing reactions were performed using the BigDye Terminators kit 3.1, and run on an ABI 3730 Automated Sequencer (Applied Biosystems, USA).Using Blastn search, COI sequences similar to the present sequences were retrieved (Table 2) and multiply aligned by CLUSTAL W (Larkin et al., 2007) in the Geneious Pro 5.4.6 program (Biomatters, New Zealand). Neighbor-Joining tree

[Table 1.] List of PCR primers used in this study

based on the Kimura two-parameter model was inferred by PAUP4.0b10* (Swofford, 2003). Tree robustness was exam-ined by bootstrap analysis using 1,000 replicates.

For phylogenetic analysis, more than 4 kb DNA including mitochondrial coding genes COI, COII, ATP6, ATP8, and COIII was amplified from a individual collected from the Sangchun reservoir by using PCR primers, CI-J-1632 and C3-J-5460 (Table 1) by using the same PCR reaction compo-sition for the molecular identification and the following PCR protocol: 92℃ for 2 min and 40 cycles of 92℃ for 30 sec, 45℃ for 30 sec, 68℃ for 12 min, followed by final exten-sion, 68℃ for 20 min. For PCR amplification of genes inside the long PCR fragment, several internal PCR primers were designed (Table 1) and separate PCR reactions were execut-ed using these primers. ND3 was amplified by the PCR pri-mers, C3-I-F and N4-I-R. PCR products were purified using the AccuPrep PCR Purification kit. Sequencing reactions were performed using BigDye Terminators kit 3.1, and run on an ABI 3730 Automated Sequencer.

For comparison, another 9 dipteran species and Locusta migratoria were used as an outgroup from which complete mitochondrial DNA sequences are known were retrieved from the GenBank (Table 3). Based on the conserved (C)/vari-able(V) site ratios and percentage of gaps and invariable sites as an estimation of reliability of alignment (Lee et al., 2006), ATP8 did not show reliability (data not shown). ATP8 gene was commonly excluded from phylogenetic studies using mitochondrial genome. Except ATP8 gene 5 genes were used for further phylogenetic analyses. After translating the DNA sequences, reading frames of DNA sequences were confirm-ed by concatenating and aligning 5 mitochondrial coding genes by using CLUSTAL W in the Geneious Pro 5.4.6 pro-gram.

For reconstucting the phylogenetic trees based on 5 mito-chondrial coding genes from dipteran species, a substitution model was chosen using MrModeltest version 2.02 (Nylander,

[Table 2.] List of COI nucleotide sequences of Chaoborus speciesusing for phylogenetic analysis

[Table 3.] List of species compared with Chaoborus for phylogeny reconstruction

2004) under Akaike’s information criterion. The GTR+I+G model was used to generate a Bayesian inference and a max-imum likelihood (ML) tree. The Bayesian tree was obtained with MRBAYES version 3.1.2 (Ronquist and Huelsenbeck, 2003) with default options for the prior distribution in the Bayesian inferences. Metropolis-coupled Markov chain Monte Carlo (MCMCMC) analyses were run with one cold chain and 3 heated chains for 1,000,000 generations from 5 mitochondrial coding genes, and sampled every 100 genera-tions. Two independent MCMCMC runs were performed and 2,500 trees were discarded as burn-in from the 5 mito-chondrial coding genes. The final trees near the optimum likelihood score were retained using the appropriate burn-in criterion. Trees from 5 mitochondrial coding genes were re-tained and used for calculation of posterior probabilities. A ML tree was constructed using PAUP4.0b10* by heuristic search with a truncated balanced realization algorithm. The tree stability was examined by bootstrap analysis with 100 replicates.

The sequence alignment is available upon request from the corresponding author.

For molecular identification, 6 individual specimens from Sangchun reservoir in Gyeonggi-do and 3 from Ildae reservoir in Jeollanam-do were used for generation of COI barcodes. In addition, 2 specimens from the National Institute for En-vironmental Studies of Japan were also examined for com-parison. Sequences were deposited in GenBank (accession nos: JQ277990-JQ278000). Based on the sequence compari-son of the COI gene, the specimens from the 2 Korean reser-voirs were almost identical (0.06%), while the specimens from Korean and Japanese populations showed 18% differences in genetic distance which are close to values for putative cryptic species suggested by Dupuis et al. (2008). According to the

[Fig. 1.] Neighbor-joining tree for Chaoborus species based on partial mitochondrial cytochrome oxidase I (COI) gene sequences. Values above the branches indicate ¤50% bootstrap support. Boxes include specimens examined in the study.

neighbor-joining tree (Fig. 1), all the specimens examined in this study belonged to the Chaoborus flavicans group. Two subgroups were recognized in the C. flavicans group 2. Each of the Korean and Japanese populations was assigned to a separate subgroup. The 2 subgroups corresponded to the 2 cryptic species suggested by Dupuis et al. (2008). According to Dupuis et al. (2008), 2 cryptic species have been recogniz-ed by having different habitats (lake-dwelling and pond-dwel-ling) and morphological characteristics, especially the man-dible. Korean populations were included in the lake-dwelling group and were well discriminated from the pond-dwell-ing group, including the Japanese population. Populations from the Palearctic and Nearctic were well recognized in the lake-dwelling group, as indicated by Dupuis et al. (2008).

Sequences of the partial mitochondrial genome were depo-sited in GenBank (accession no: JQ235548). The partial mito-chondrial genome of C. flavicans including COI, tRNA-Leu, COII, tRNA-Lys, tRNA-Asp, ATP8, ATP6, COIII, tRNAGly, ND3, tRNA-Ala. was 4446 bp in length (Table 4). The overall AT content for six coding genes was 70.1% (T=39.0%, C=17.7%, A=31.1%, G=12.3%) similar to that of dipter-an species.

The order of protein coding genes and tRNA was identical to that reported from many of insect species. The initiation and termination codons of the genes examined here were identified using the open reading frame finder and by com-parison with mitochondrial gene sequences of other dipteran species. ATG and ATT were used as initiation codons (Table 4). In case of COI the initiation codon could not found be-cause of the truncation of start region of the gene. The usual TAA termination codon found for all genes examined.

The codon usage of C. flavicans for six mitochondrial pro-tein coding genes and the relative synonymous codon usage values are given in Table 5. Most of values differed from the

[Table 4.] Annotation and gene organization of Chaoborus flavicans partial mitochondrial genome

[Fig. 2.] Maximum likelihood (ML) tree for the selected dipteran species based on 3840 bp of five concatenated mitochondrial genesequences. Values above and below the branches indicate ML bootstrap values and BI posterior probabilities respectively.

equilibrium frequency and the use of synonymous codons was distorted. CUG (Leu), UCG (Ser), UAG (Termination), AAG (Lys), UGG (Trp), AGG (Ser) were not used in C. fla-vicans. Since mtDNA of insects show a high bias against G and C, this could explain the lack of these codons.

The topology of resulting phylogeny based on ML and Bayesian interferences analyses recognized two clusters of the Diptera consisting of Brachycera and Nematocera with high boostrap supports. In the Nematocera, Chaoboridae was more closely related to Ceratopogonidae than to Culicidae in both analyses employed (Fig. 2). However, the present phy-logeny suggested different relationships between families in Nematocera from previous analyses (Miller et al., 1997;Sæther, 2000). These prior studies suggested that Ceratopo-gonidae was diverged early and Chaoboridae and Culicidae were more related. In this study, Ceratopogonidae and Chao-boridae were more related to each other, however, this rela-tionship received little statistical support.

In this study, Chaoborus species from 2 reservoirs in Korea were identified as Chaoborus flavicans by mitochondrial COI gene sequences. Phylogenetic trees based on 5 mitochondrial coding genes by ML and Bayesian inferences showed Chao-borus was more closely related to the Ceratopogonidae than to Culicidae, however, this relationship received little statis-tical support. Therefore, further analyses based on complete mitochondrial DNA sequences and nuclear gene sequences are needed for a more robust validation of the phylogenetic relationship of Chaoborus within dipteran lineages.

[Table 5.] Chaoborus flavicans codon usage of six protein coding region of partial mitochondrial genome