The fall armyworm, Spodoptera frugiperda (Smith, 1797), is one of the most important noctuid moth pests in the world and be known to damage economically important cultivated rice, maize, sorghum, cabbage, beet, peanut, soybean, alfalfa, onion, cotton, pasture grasses, millet, tomato and potato (Chapman et al., 2000; Montezano et al., 2018). Until now, it has a broad host range and attacks more than 350 species of plants (Montezano et al., 2018), but it prefers maize sometimes resulting in yield losses (>70%) when it outbreaks (Johnson, 1987).

This species is native to tropical and subtropical regions of the Americas (Goergen et al., 2016). Since S. frugiperda was first described in 1797, its outbreaks have occurred irregularly in USA with severe damages in 1870, 1912 to 1920, and 1975 to 1977 (Sparks, 1986). After it invaded into Nigeria in West Africa in 2016, it covered through about 40 countries in sub- Saharan Africa during one year (Nagoshi et al., 2018). In July 2018, it was firstly found in India, and then has spread rapidly to other Asian countries including Bangladesh, China, Laos, Myanmar, Sri Lanka, Thailand, and Vietnam (Ma et al., 2019).

In Korea, S. frugiperda was firstly found on 13 June 2019 at a maize field of Gujwa, Jeju-si, Jeju Island by experts of Jeju Agricultural Extension and Service Center. The life stages of samples were 2^nd to 3^rd instars and crop damage rates were measured as about 5% (per 100 plants). This moth was further found at three maize fields adjacent to the first founding location of Jeju Island (Fig. 1). The subsequent surveys by Rural Development Administration (RDA) and the provincial Agricultural Research & Extension Services showed that many maize fields had been infested by this moth and damage rates of the maize fields were less than 1% on many counties of Jella-do and Gyeongsang-do in June and July 2019.

In this study, we analyzed a mitochondrial cytochrome c oxidase subunit I (COI) gene about the first invaded populations of S. frugiperda, and the larval samples collected at several counties in Korea to confirm genetically as one species S. frugiperda and found several populations collected in Korea were separated into two clades (A and B) based on COI sequences.

Materials and Methods

Sample Collection

Sampling was conducted from June to July 2019 throughout four provinces of Korea: Jeollanam-do (JN), Jeollabuk-do (JB), Gyeongsangnam-do (GN), and Jeju-do (JJ). After crop damage was observed by naked eyes, larvae (Fig. 2) were collected from maize fields using larval tweezer. Collection details, geographical locations, host plants, and dates of collection are summarized in Table 1.

A total of 31 larvae were collected, and individual samples were preserved in 99% ethanol. Voucher specimens were deposited in the insect collection of the National Institute of Agricultural Sciences, Korea.

DNA Extraction, Amplification, and Sequencing

Genomic DNA extraction was performed using DNeasy® Blood & Tissue Kit (QIAGEN Inc., Dusseldorf, Germany) according to the manufacturer's protocol. Samples for extraction consisted of a single individual from the same colony. PCR amplification was conducted with one primer set, LCO1490 (5′ -GGTCAACAAATCATAAAGATATTGG-3′) and HCO2198 (5′-TAAACTTCAGGGTGACCAAAAAATCA-3′) (Folmer et al., 1994), using AccuPower® PCR PreMix (Bioneer, Seoul, Korea) with the following thermal cycle parameters for 20 amplification reactions: initial denaturation for 5 min at 94°C, followed by 34 cycles of 1 min each at 94°C, 1 min at 45.2°C, and 1 min at 72°C, with a final extension for 5 min at 72°C. PCR products were visualized on agarose gels after electrophoresis. Single bands were purified using a QIAquick PCR purification kit (QIAGEN, Dusseldorf, Germany). PCR products were sequenced in both directions by ABI 3730xl sequencer (Applied Biosystems). The resulting chromatograms were evaluated for miscalls and ambiguities and assembled into contigs in SeqManTMPro (version 7.1.0, 2006; DNAStar, Inc., Madison, Wisconsin, USA). The sequences were visually checked individually for protein coding frame-shifts to avoid pseudogenes (Zhang and Hewitt, 1996). Consensus files were aligned using Clustal X 1.83 (Thompson et al., 1997). All sequences are deposited in the GenBank.

Data Analysis

For identifying 31 moth samples, a neighbor-joining (NJ) tree was constructed based on 31 new COI sequences analyzed in this study, together with 27 COI sequences of S. frugiperda from the GenBank (http://www.ncbi.nlm.nih.gov/genbank/). Alignments of nucleotide sequences were performed using CLUSTALX with default conditions. A NJ analysis was conducted for the combined data set, in MEGA 5.0 (Tamura et al., 2011). Pairwise sequence divergences between the 58 COI sequences for each were calculated using a Kimura’s 2-parameter (K2P) distance model (Kimura, 1980) in MEGA 5.0 (Tamura et al., 2011). Descriptive statistics (number of variable sites and number of haplotypes,) were calculated using DNASP ver. 5.0.

Results

Amplification Result and Sequence Information

A total of 31 COI sequences were successfully amplified from the 31 specimens and obtained bands of approximately 650 bp on the gel. We aligned the sequence once the PCR products were sequenced. Several base pairs were removed because of ambiguous alignment, which resulted in a final count of 546 bp. From the combined dataset (new 31 COI sequences + 27 COI sequences of the GenBank), we determined 29 variable sites at nucleotide positions 16, 51, 84, 133, 138, 140, 172, 176, 185, 247, 286, 288, 294, 316, 324, 369, 406, 420, 444, 450, 468, 486, 504, 513, 514, 543, and 546.

Genetic Diversity and Distances

Totally, 27 haplotypes of S. frugiperda were detected from the 58 COI sequences in the world. Among the 27 haplotypes, most of samples collected in Korea (29 out of the 31 COI sequences) belonged to the hap-1; whereas, the hap-2 was detected from samples which were collected in JN province. The genetic distances among the 27 haplotypes were ranged from 0.20% to 2.20%, and the hap-1 and hap-2 populations detected in Korea revealed a 1.90% genetic distance.

Phylogenetic Analysis of S. frugiperda

The NJ tree was generated based on the aligned dataset of 58 COI sequences of S. frugiperda and one COI sequence of Spodoptera litura as an outgroup (Fig. 3). The NJ tree revealed two distinct clades: clade (A) consisted of 19 haplotypes (hap-1, 3, 8, 9, 10, 11, 12, 13, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, and 27) and 48 COI sequences (including 25 COI sequences of the JB, JN, JJ, GN populations), and clade (B) consisted of seven haplotypes (hap-2, 4, 5, 6, 7, 14, and 15) and 9 COI sequences (including two COI sequences of the JN populations). In the clade (A), no genetic differences among the JB, JN, JJ, and GN populations were observed (Fig. 3).

Discussion

Until early April 2019, S. frugiperda has been known to be distributed in the America, sub-Saharan Africa, and Asia (including Bangladesh, China, India, Myanmar, Sri Lanka, Thailand, and Yemen) (CABI, 2019). In this study, we firstly found the occurrences of S. frugiperda on maize fields in Korea (Fig. 2) and examined their 31 COI sequences together with the 27 COI sequences from the Genbank. From the 58 COI sequences, the first invaded Korean populations were comprised of two haplotypes, hap-1 and hap-2, and the phylogenetic tree revealed that S. frugiperda was separated into two clades (A and B) with the hap-1 and hap-2 of the Korean populations were included in each clade.

In this study, we checked distribution countries of the 27 haplotypes based on each of haplotype sequence data from GenBank (Table 2). In the phylogenic tree, 29 COI sequences from Korea (hap-1) were identical to native ones from Brazil, Canada, Costa Rica, Dominica, and USA, as well as invasive ones from Ghana, Kenya, Nigeria, South Africa, Uganda, China, India, and Vietnam; whereas, two COI sequences from Korea (hap-2) were identical to native ones from Brazil, Canada, Mexico, Puert Rico, and USA, as well as invasive ones from Ghana, Kenya, Sao Tome, Uganda, and India. It means that two haplotypes (hap-1 and hap-2) among the 27 haplotypes have been involved mostly in intra- and intercontinental dispersal, including Africa and Asia.

According to Nagoshi et al. (2019), the expansion of S. frugiperda in Africa can be explained by a single introduction, showing low numbers of haplotypes, regional similarities in haplotype composition, and regional differences in haplotype frequencies. If the long-distance migration of S frugiperda is one of reasons for the rapid invasion into Korea, the migration source, as estimated by Ma et al. (2019), might be the southern or middle regions in China. Zhang et al. (2019) recently reported that two haplotypes, hap-1 and hap-2, were found in the southern regions in China, which were the same haplotypes in Korea, until early June 2019, and the hap-1 was predominant in number of sequences (> 96%) in China, It was similar that the hap-1 was also predominant in Korea, accounting for 93.5%.

S. frugiperda has a remarkable dispersal capacity and this feature is understood to have evolved as part of its life history strategy (Jonhson, 1987). Considering its high spreading performance, large reproductive capacity (Murúa and Virla, 2004), absence of diapause (Jonhson, 1987), and wide host plant range, it is likely that the pest will be able to become one of important migratory insect pests in most of Korea. So, there is an urgent need for developing pest control methods and detection tools to mitigate the impact of the pest in Korea. In addition, further studies about migration behavior using combined molecular markers should be conducted to estimate the source areas or migration times.