Three kinds of rice planthoppers (L. striatellus, N. lugens, and S. furcifera) were collected in 2010 from a rice field in Nanjing, China. They were maintained in the laboratory on seedlings of japonica rice variety Wuyujing 3 (Oryza sativa L.) under a 16L:8D photoperiod at 26 ± 2°C with 65 ± 5% relative humidity (RH). The rice was grown in nutrient rich soil in a climate chamber. RWW (L. oryzophilus) were provided by Professor Mingxing Jiang at Zhejiang University and two kinds of mirid bugs (C. livdipennis and T. chinensis) were provided by Professor Zengrong Zhu at Zhejiang University. CLB (Brontispa longissimi) was obtained from Professor Zhengqiang Peng at the Chinese Academy of Tropical Agricultural Science in Hainan province, China. E. sophia was provided by Professor Fanghao Wan at the Chinese Academy of Agricultural Science and M. cingulum was provided by Professor Jian Hu at Sun Yet-Sen University. The D. melanogaster Canton-S strain was used as the external reference (Bennett et al., 2003; Tsoumani and Mathiopoulos, 2012). The fly culture was fed on a cornmeal-agar-molasses medium and maintained under a 12L:12D photoperiod at 25 ± 2°C and 60 ± 5% RH. A. pisum and A. lucorum were maintained in a laboratory at the Chinese Academy of Agricultural Science, Beijing, China.

Samples were prepared using a standard procedure with slight modification (Galbraith et al., 1983; Brown et al., 2005; Dolezel et al., 2007). Insects were anesthetized using carbon dioxide for 20 s and the heads were dissected in an ice-cold plastic Petri dish. The heads were completely homogenized in 500 μL ice-cold Galbraith's Buffer (pH 7.0) containing 45 mM MgCl₂, 20 mM MOPS (3-N-morpholinopropane sulfonic acid), 30 mM sodium citrate, and 0.1% (vol/vol) Triton X-100. The homogenate was washed slowly 2–3 times and filtered into a 1.5-mL Eppendorf tube using 38-μm nylon mesh. RNA was removed by adding RNase A (Takara, Japan) to the homogenate at a final concentration of 20 μg/ml and incubated at 25°C for 10 min. Then the solutions were centrifuged at 1000 g for 5 min. The precipitates were suspended with 400 μL phosphate buffer (pH 7.4) and stained with 50 μg/mL propidium iodide stock solution in darkness at 4°C for 10 min. The suspensions obtained in the final step were analyzed using the MoFlo™ XDP High Speed Cell Sorter and Analyzer (Beckman Coulter, CA, USA). The cell DNA content was measured using the fluorescent intensity of each sample exposed to a laser at 488-nm wavelength. The same parameter settings were applied to D. melanogaster samples. Summit Software (Beckman Coulter, CA, USA) was used to obtain the nuclei peaks (FL3-Log-Height or FL3-A). The genome sizes of the samples were calculated as follows: Sample 1C value=Reference 1C×(sample 2C mean peak positionreference 2C mean peak position), where reference 1C-value is the genome size of D. melanogaster, which is 176.4 megabase pairs (Mb) (1 pg = 978 Mb). For haploid and triploid cells, sample peaks of the 1C and 3C positions were used, and the sample 1C-value was obtained by multiplying by 2 and 2/3, respectively (Dolezel et al., 2007). All experiments were repeated at least for two times.

A paired-end library with an insert size of approximately 350 bp was constructed using the Illumina TruSeq Nano DNA (350) DNA sample preparation kit, following manufacturer instructions, and was sequenced using the Illumina HiSeqX system in Macrogen Inc., Korea. Quality control of raw sequence data was done using FastQC. The reads were filtered before assembly to ensure that a pair of paired-end reads had more than 90% of bases with quality ≥Q20. High-quality cleaned Illumina sequences were subjected to k-mer counting using JELLYFISH (Marçais and Kingsford, 2011) with the k-mer size set to 17. k-mer depth distribution was counted and the peak value of the depth distribution was identified. Since the short reads by the Illumina Solexa sequencing are randomly generated, the depth of the k-mer coverage should follow a Poisson distribution. So, the mean k-mer depth equals the peak value of the k-mer depth distribution. The genome size was calculated using the formula: Genome size=total number of k-mer/peak value of k-merfrequency distribution.

The genome sizes of the nine insect species, estimated by flow cytometry, are presented in Table 1. D. melanogaster was the external reference and N. lugens and A. pisum, with published genomes, were positive controls. All experiments were repeated at least two times and the results indicated that all replicates had good reproducibility. The estimated genome sizes of N. lugens were 1130 Mb for male and 1110 Mb for female. A. pisum was estimated to be 460 Mb, which was consistent with genome assembly results (Richards et al., 2010; Xue et al., 2014). Among the nine insects, L. oryzophilus had the largest genome size (981 Mb, 1.003 pg), which was approximately two-fold that of other Coleoptera B. longissimi (554 Mb for male and 443 Mb for female). The genome sizes of the other eight insects (omitting L. oryzophilus) were <1 Gb (the nucleic content ≤1 pg) and the average genome size was 504 Mb. These data indicate that the insects studied had a moderate genome size. Both S. furcifera and L. striatellus had a smaller genome size than N. lugens (Figure 1). The mirids C. livdipennis and T. chinensis had a genome size around 400 Mb but A. lucorum had a genome size of 878 Mb. Hymenoptera tend to have smaller genome sizes than other orders. The genome sizes of wasps in this study were <200 Mb with the exception of E. sophia, which had a larger genome size of 372 Mb.

Males and females of five species, B. longissima, L. striatellus, S. furcifera, C. liydipennis, and M. cingulum, were used to estimate genome size. Genome sizes were different in the males and females of four species except C. liydipennis. The sex differences of genome size were 111, 97, 77, and 4 Mb in B. longissima, L. striatellus, S. furcifera, and M. cingulum, respectively (Table 1, Figure 1). The largest difference between males and females was in B. longissima, while the smallest difference was in M. cingulum. The sex difference of genome size may have been due to variation in sex chromosome numbers. Male L. striatellus and S. furcifera have 29 chromosomes whereas female have 30 chromosomes (Kobayashi and Noda, 2007; Noda, 2009). The female hemipterans, L. striatellus, S. furcifera, and C. liydipennis, had larger genomes than the males. However, the genome sizes of male B. longissima and M. cingulum were larger than females. The sex related differences of genome size is a complex and unanswered scientific question.

Non-native, invasive species often cause damage to the environment, agriculture, economy, and humans. Among the nine insects studied, L. oryzophilus and B. longissimi are invasive insects. Both of these species cause significant economic losses in China. We investigated the relationship between species invasiveness and genome size. We collected genome size information on 18 invasive insects from the Animal Genome Size Database (http://www.genomesize.com). The genome sizes of invasive insects were compared with the average genome size of non-invasive species in the corresponding family. There was no distribution bias in the genome sizes of invasive insects (Table 2), suggesting that the invasiveness of an organism is unrelated to genome size.

The k-mer analysis was used to estimate the genome sizes of L. striatellus and M. cingulum using Illumina Solexa sequencing data. The 17-mer depth distributions showed a single peak, indicating that both insects had very low heterogeneity. Based on k-mer analysis, the genome size of L. striatellus was 657 Mb and that of M. cingulum was 136 Mb (Figure 2). These values were slightly different from estimates made using flow cytometry. This difference might be due to the variability of the two methods. Another possible reason was that different samples were used for the k-mer analysis. If this is the case, the genome size differences between different samples require further investigation. Regardless, it still can be concluded that L. striatellus had a moderate-size genome of 500–700 Mb while M. cingulum had a small-size genome of ~150 Mb. The exact genome size requires confirmation by additional genome-sequencing.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.