Gene Flow Across Host‐Associated Populations of the Rice

Hereward James (Orcid ID: 0000-0002-6468-6342)

Gene flow across host-associated populations of the rice stem borer Chilo suppressalis Walker (Lepidoptera: Crambidae) – implications for Bt resistance management in rice.

Yongmo Wang†, Chen Huang†, Bing Hu†, Yue Liu†, Gimme H. Walter#, James P. Hereward#*

†Hubei Insect Resources Utilization and Sustainable Pest Management Key Laboratory, College

of Plant Science & Technology, Huazhong Agricultural University, 430070 Wuhan, P.R. China

#School of Biological Sciences, The University of Queensland, Brisbane, QLD, Australia, 4068

* corresponding author, email address: [email protected] (J.P. Hereward)

This is the author manuscript accepted for publication and has undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1002/ps.5567

Background:

The rice stem borer, Chilo suppressalis, is a serious pest of rice, but also damages an

aquatic vegetable, water oats (Zizania latifolia Turcz.). The time at which mating occurs is different between populations of rice stem borer associated with rice and those associated with water-oats, which suggests that undetected cryptic species may be associated with these plant hosts. If true, this would have significant management implications. This study is the first empirical test of this idea, using population genetic tools from two sampling cohorts. We genotyped 320 rice stem borer individuals from

2014, collected from rice and water-oats across five locations (where they exist in sympatry), using seven microsatellite loci.

Results:

We found no genetic structuring associated with host plant species. On water oats,

some rice stem borers were found that had a similar mating time to the rice population,

so in 2016, a second cohort of samples was screened by their timing of mating to get

‘pure rice feeders’ and ‘pure water oats feeders’. These samples were genotyped with

microsatellites, mitochondrial DNA (mtDNA – COI & COII), and a nuclear gene

(EF1-α). Our mtDNA data suggest a relatively low amount of population subdivision

associated with plant host, but the microsatellite data revealed no such genetic

Conclusions:

Our results indicate gene flow between rice and water oats populations of rice stem borer, indicating that water oats will likely provide a refuge for resistance management of Bt rice.

Detecting cryptic species is crucial to conducting and interpreting applied research on agricultural pests1 and differences in biological characteristics across populations often signal the presence of undetected cryptic species2,3,4. The rice stem borer Chilo suppressalis (Walker) is one of the most serious insect pests of rice in

Asia5. It feeds on several host plants in Poaceae, but mainly rice, Oryza sativa L., and an aquatic vegetable - water oats, Zizania latifolia Turcz6. Rice stem borers occur on rice and water oats simultaneously when the two crops are grown in the same area.

Even when water oats is not cultivated locally, wild populations are often found, especially in southern China.

Biological differences across rice borer populations that feed upon rice and those that feed on water oats are well documented. In general, the body size of individuals that fed on water oats is significantly bigger than that of rice-fed borers7,8.

Morphometric differences in male genitalia are also evident between the two9. The blend ratios of the three components of their sex pheromones are significantly different from each other10. The overwintering biological traits, such as supercooling points, parasitism rates and diapause intensity are significantly different between overwintering larvae collected from the two host plants11,12.

Several behavioral differences have also been described between the rice stem

borer populations associated with rice and water-oat hosts. After overwintering, spring

flight (following eclosion) is about 15 days earlier in water oats fields than in rice

fields13. Individuals from rice and water oats collected near one another differ in susceptibility to some pesticides14. More importantly, there is a distinct difference in the time at which mating occurs, with the rice feeders mainly mating in the ﬁrst half of the scotophase and the water oats feeders mating in the latter half15. This difference

has been confirmed by several studies in different locations10,16. This suggests that rice

stem borer could comprise cryptic species associated with these different host plants.

However, this has not been sufficiently investigated since the allozyme based study of

Ishiguro et al.17, who found no genetic difference across hosts based on the three of

their markers that were informative.

Understanding the extent of gene flow between these two host-associated

populations of rice stem borer is critical to designing appropriate resistance

management strategies. Several genetically modified rice lines expressing Bt (Bacillus

thuringiensis) toxins have been developed in China and are waiting commercial

release18. In many countries a refuge is mandated as part of the ‘high-dose and refuge’

percentage of a Bt crop is sown with non-transgenic crop. In China, refuge strategies

are not generally employed, but the small size of farms and the mix of crops grown

appears to have provided an unplanned, area-wide, refuge strategy, at least in species that use multiple host plants such as Helicoverpa armigera 19,20. In contrast, the pink

bollworm (Pectinophora gossypiella) only feeds on cotton, and under the ‘no refuge’

strategy resistance to the Cry1Ac toxin has been rising21. If there is gene flow

between these two populations then a ‘no-refuge’ strategy may work well for rice stem borer in Bt rice, because water oats will act as a natural refuge. If, however, there is no geneflow between the host associated populations and they represent cryptic species, then a refuge strategy would be recommended.

We conducted a population genetics analysis of rice stem borer to test for the

presence of cryptic species associated with these two hosts. This study consists of two

phases. In the first phase, we selected five sampling sites where both rice and water

oats were co-planted and collected rice stem borers from the two host plants within about a kilometer of each other. The samples were genotyped using seven

microsatellites. In the second phase, we selected two of the five sampling sites, and

field-collected individuals from rice and water oats were screened in the laboratory

markers, two mtDNA genes (COI and COII), and a nuclear gene (EF1-α). This study

deepens our understanding of the evolutionary status of the host-associated

populations of the rice stem borer and provides helpful insights for developing

effective management strategies for this pest.

Materials and methods

Insect sampling and rearing

In 2014, five sampling sites, Hangzhou, Taizhou, Jinhua, Changde and Wuhan,

were targeted (Fig. 1). These locations are alongside the middle and lower reaches of the Yangtze River where rice is planted as the main food crop. Water oats, as a

traditional aquatic vegetable, is widely planted in Hangzhou, Taizhou and Jinhua

especially. Although water oats is cultivated only sporadically in small acreages in

Changde and Wuhan, wild populations frequently grow in irrigation canals and ponds

near paddy fields.

At each site, more than 400 larvae (2nd to 5th instar) were collected from rice

fields, and at the same time matched numbers of larvae were collected from water

oats fields within about a kilometer of the rice site. A subset of the samples was stored

laboratory using the same host plant that they were collected from at 28±1℃, 80-90%

RH and L16:D8 photoperiod, to observe the time at which they mated. Samples from

the same site and the same host were designated a population.

In 2016, two sampling sites, Wuhan and Hangzhou, were targeted. In 2014 we found that some field-collected populations from water oats exhibited two separate mating peaks during the scotophase, which indicated that there may be some rice-associated stem-borers in water oats fields16. Thus, the rice stem borer samples

were screened by the time that they mated before genotyping them. Samples were

reared on their natal host plant to the adult stage under the same conditions described

above, and then individuals from rice fields that mated three to four hours after the onset of scotophase were selected as genuine rice-associated stem borers, and individuals from water oats fields that mated five to six hour after the onset of scotophase were selected as genuine water oats associated stem borers (Fig. 2).

Microsatellite genotyping and analysis

Genomic DNA was extracted from individuals of the rice stem borer using a universal genomic DNA extraction kit (TaKaRa) following the manufacture’s

(2014 and 2016) were genotyped using seven microsatellite loci (CS1, CS5, CS11,

CS20, Cs248, Cs218, Cs133) that had been isolated from C. suppressalis by Ishiguro

& Tsuchida22 and Liu et al.23. The loci were amplified with fluorescently labeled

(FAM) primers following the PCR conditions described by the authors. PCR products

were separated using an ABI3730 DNA sequencer (Applied Biosystems, USA) with

the GeneScan-500 (LIZ) internal size standard and the peak allele calls were

confirmed manually in Geneious 9.1.3 (Biomatters). At least 32 larvae were

genotyped in each host-associated population included in each analysis.

Analysis of microsatellite data

Basic statistics were calculated in GenAlEx version 6.50124, and included the

number of alleles, and the observed heterozygosity (HO) and expected heterozygosity

(HE) under Hardy-Weinberg equilibrium (HWE) for each locus and each population.

Fixation index FIS values were also calculated using GenAlEx. To test for deviation

from HWE, Hardy-Weinberg exact tests were performed using GENEPOP version

3.425. The significance of these statistics was estimated through 5,000 random

permutations with Bonferroni corrected p values.

Several methods were used to analyze the genetic differentiation and structuring

among populations. Pairwise FST values were calculated using GenAlEx to estimate

genetic differentiation between each population pair. Exact probability tests for allelic

frequencies on each pairwise comparison were made using a Markov chain method

based on 999 permutations. Individual-based assignment to genetic clusters was

performed using the Bayesian model-based clustering method implemented in

STRUCTURE version 2.3.426 under the “no-admixture” model with no use of

population priors. We ran 10 independent simulations each consisting of a burn-in

period of 500,000 chains and a run length of 1,000,000 chains for each of 1 to 10 possible clusters, K. To determine the most likely value of K, these results were uploaded to the web-based software STRUCTURE HARVESTER

(http://taylor0.biology.ucla.edu/structureHarvester/),27 using the delta K method28 to

assess and visualize likelihood values across multiple values of K. The results of 10

runs using the most likely value of K with different starting seeds were then permuted

and combined using CLUMPP29 and plotted using DISTRUCT30. Finally, principal

components analysis (PCA) was performed using the adegenet package31 in R32.

COI, COII and EF1-α sequencing and analysis

This article is protected by copyright. All rights reserved. Fragments of mitochondrial genes (Cytochrome C Oxidase Ⅰ and Ⅱ, COⅠ and COⅡ) and a nuclear gene (elongation factor-1α, EF1-α) were amplified in a subset of the

2016 samples. COI was amplified from the rice stem borer samples using the primers

LCO1490 (5'-GGTCAACAAATCATAAAGATATTGG-3') and HCO2198

(5'-TAAACTTCAGGGTGACCAAAAAATCA-3')33; COII was amplified using the primers PIERRE (5'-AGAGCCTCTCCTTTAATAGAACA-3') and E VA

(5'-GAGACCATTACTTGCTTTCAGTCATCT-3')34; EF1-α was amplified using the

primers EF3 (5′-GAACGTGAACGTGGTATCAC-3′) and EF2

(5′-ATGTGAGCAGTGTGGCAATCCAA-3′)35. The PCR amplification was

performed in a 20 µL reaction mixture containing 1 μL template DNA, 2 µL10×

buffer, 0.4 μM each of the primers, 200 μM each dNTP and 1 U Taq polymerase

(TaKaRa). The PCR conditions were as follows: 95 ℃ for 5 min, 35 cycles consisting

of denaturation at 95 ℃ for 30 sec, annealing at 57 ℃ for COⅠ, 55 ℃ for COⅡ and

58 ℃ for EF1-α for 30 sec and extension at 72 ℃ for 45 sec, and a final extension at

72 ℃ for 5 min. PCR products were sequenced bi-directionally after confirmation by

1% agarose gel electrophoresis. Sequences were edited with codoncode aligner 7.1.2

(CodonCode Corporation) and aligned with Geneious 9.1.3 (Biomatters Ltd,

Auckland, New Zealand). Haplotype networks were constructed with PopART 1.7.2

(www.popart.otago.ac.nz) using TCS network criteria based on sequences of COⅠ,

the Geneious software 9.1.3 (Biomatters Ltd., Auckland, New Zealand). Phylogenetic

trees for combination of COI and COII were constructed using Mr Bayes method with

GTR substitution model and the neighbour-joining (NJ) method37.The clusters were assessed by bootstrapping 1000 replicates. The mitochondrial genome of Chilo auricilius was used as the outgroup (GenBank accession number KJ174087).

Results

Genetic variation and structuring of 2014 samples

A total of 320 individuals was genotyped using seven microsatellite loci and each

locus exhibited allelic polymorphism. The percentage of polymorphic loci ranged from 85.7% to 100% in a single population. The number of effective alleles in a single locus ranged from 1.036 (CS20) to 6.833 (CS248) across all populations and ranged from 2.233 (HZ_W) to 3.718 (HZ_R) in a single population across all loci (supporting information, Table S1). The number of effective alleles in rice associated populations

(3.368) was higher than in the water oats associated populations (2.954), but this difference was not significant (t-test, df=8, p= 0.141). The average HO ranged from

0.202 to 0.385 and the average HE ranged from 0.411 to 0.527 in each locus of a

single population. There were two to six loci that deviated significantly from

deviated significantly from HWE in all populations (supporting information, Table

S1).

Genetic differentiation was measured by pairwise FST which varied from 0.012 to

0.089 (Table 1). Almost all (44 out of 45) of the population pairs (HZ_R vs CD_R)

exhibited significant allelic frequency differences (p<0.05), but these were not always

accompanied by substantial FST values (>0.05). The pairwise FST between HZ_W and

other populations were all >0.05, indicating the water oats-associated population from

Hangzhou is the most genetically different (Table 1). The PCA clustered all samples

together and showed no pattern of clustering associated with host plant (Fig. 3). The

deltaK method indicated that K=4 was the most likely number of clusters in the structure analysis, but there was no additional clustering evident beyond K = 2 based on a visual inspection of the plots. Although structure plot indicated that both the rice and water oats populations from Hangzhou were relatively clearly differentiated from all the other populations, no structure was associated with host plant species at this site (Fig. 3).

Genetic variation and structuring of 2016 samples

This article is protected by copyright. All rights reserved. The four populations sampled in 2016 were genotyped with seven microsatellite loci, and the results showed allelic diversity at each of the loci. The percentage of polymorphic loci in each population ranged from 85.71% to 100%. Only one allele was found in HZ-W at CS20. The number of effective alleles ranged from 1.052

(CS20) to 8.878 (CS248) at a single locus across all populations and ranged from

3.488 (WH_W) to 3.515 (WH_R) in a single population across all loci (supporting

information, Table S2). There was no difference in number of effective alleles across the host-associated populations. All four populations deviated from Hardy-Weinberg equilibrium at CS5, but not at CS1 and CS20 (supporting information, Table S2).

Pairwise FST varied from 0.019 to 0.053 (Table 2). Of the six pairwise comparisons, only WH_R vs WH_W did not exhibit significant allelic differentiation

(p=0.059) with a small FST=0.019. The PCA analysis clustered all four populations

similarly, but some individuals from rice in Hangzhou fell outside the main cluster

(Fig. 4). The structure plot indicated a slight difference between the Hangzhou rice

samples and the other three, but did not otherwise show clear structure associated with

host plant species (Fig. 4).

A total of 24 individuals was sequenced successfully for the COI, COII and

insertions or deletions were found in the alignment of the gene regions. The COI

sequences showed polymorphism at 26 nucleotide positions, and 10 different

haplotypes were identiﬁed. The COII sequences also showed polymorphism at 26

nucleotide positions, and 11 different haplotypes were identiﬁed. There was no

variation in the sequences of EF1-α (only one haplotype).

TCS networks inferred from COI, COII and EF1-α sequences were constructed

(Fig. 5). There were five COI haplotypes shared by more than one individual. Of

these five, three consisted of three individuals each that were only from rice or only from water oats, and the other two consisted of five individuals from both rice and water oats (Fig. 5). There were six shared COII haplotypes. Three of these comprised individuals only from rice, and two comprised individuals only from water oats, and the other one consisted of seven individuals from both rice and water oats. Some haplotypes were shared by only rice or water oats individuals from different sampling sites. A phylogenetic tree of rice and water oats feeders was constructed from

MrBayes runs based on concatenated COI and COII sequences (Fig. 5). There were several clades or sub-clades consisting of individuals only from rice or water oats.

These results, based on mtDNA, indicated some differentiation between the rice and

Discussion

The extent of the biological differences between the rice and water oats

populations of C. suppresalis are suggestive of the presence of sibling species, but we

find no strong evidence for a lack of gene-flow across these populations in this study.

The microsatellite data do not provide evidence of host-associated differentiation, and

the mitochondrial DNA data, despite some indication of host association, does not

clearly delineate samples from the two hosts. The nuclear gene EF1-α is identical in

all host-associated samples and was therefore uninformative with respect to resolving

potential population sub-structuring. This lack of differentiation is evident despite the

separation of individuals based on the time at which they mated in the 2016 phase of

the study.

How do we reconcile these patterns of genetic differentiation with the clear

differences in mating time10,16,38,39 and morphometric traits, such as body size and

male genital structure40,41. The divergence between these two host-associated lineages may be relatively recent and the nuclear gene EF1-α might still the same across them

haplotypes across hosts, despite some signal of host association in the mtDNA, could also represent incomplete lineage sorting (where gene lineages have not had sufficient time to accumulate fixed differences), or female specific behaviour (which is more evident in mitochondrial data as it is maternally inherited).

Can we infer anything about the evolutionary history of the association of rice stem borer with these plants from their domestication as crops? Rice has been domesticated several times42. In Asia, Oryza sativa was domesticated at least 7,700 years ago43, but maybe as long as 12,000 years ago44. Water oats was domesticated at

least 2,000 years ago45, based on written records. Rice stem borer could, therefore,

have been present on these two domesticated crops for a long time, and is likely to

have been present on the wild ancestors of both crops for even longer, but a recent

host-shift from one to the other cannot be excluded. There is insufficient evidence at

present to discriminate among these scenarios.

Ishiguro et al. studied the genetic divergence of rice and water oats populations

of C. suppressalis using allozyme polymorphism22. They tested 13 loci but only 3 loci

displayed useful polymorphism, and their results indicated that there was no evidence

This article is protected by copyright. All rights reserved. of gene flow between the two host-associated populations. But they did recommend that genetic markers with higher polymorphism should be used to repeat their results.

We used microsatellites developed for this species17,23 and despite recovering a high number of alleles in total, for several markers most individuals only had a few alleles.

Nonetheless, in this study these microsatellite markers showed no evidence of host-associated differentiation in C. suppressalis.

Cross-mating can occur between rice and water oats C. suppressalis in the

laboratory10,16,39,40. In these crosses, the rate of mating was significantly different in

crosses between females from rice and males from water oats compared to the

opposite direction. This might indicate that cross-mating under field conditions would be more likely in one direction rather than the other.

Perhaps research on another lepidopteran Spodoptera frugiperda (Lepidoptera:

Noctuidae) might be instructive here. Spodoptera frugiperda is also an important agricultural pest, and widely distributed in the Western hemisphere46,47. There are two

morphologically identical strains associated with corn and rice48. The corn ‘strain’ has been reported to infest mainly corn, sorghum and cotton, while the rice ‘strain’ has been found mostly in rice and wild grasses such as Johnson grass (Sorghum halepense)

As with C. suppressalis, the corn and rice populations have different mating

times, corn populations mainly mate in the first two thirds of the night and rice

populations mostly mate in the last third50,51. Mitochondrial markers seem to be

reliable at differentiating the two ‘strains’52. Ten polymorphic sites in the sex-linked

triose phosphate isomerase gene (Tpi) also appear reliable for distinguishing them53,54.

Conversely, data from nuclear markers such as AFLP’s have found less evidence for

host associated genetic differentiation55. In the laboratory, crosses between the corn

and rice S. frugiperda showed a similar asymmetry to that in C. suppressalis56, and

the inheritance of the time of females calling, male calling and copulation, exhibited strong maternal effects51. Putative hybrids were identified with mtDNA from one

strain and nuclear DNA from another in wild-collected samples48, but this was based

on AFLP data.

More recently, genomes have been developed for both corn and rice S.

frugiperda57,58 and resequencing has been conducted on nine corn strain individuals and nine rice strain individuals (based on their mtDNA designations). The genome-wide data clearly separates the two strains, regardless of which reference

how we interpret these situations in which host-associated populations appear to represent independent gene pools, but cannot be readily differentiated genetically. Is the recorded discrepancy between mitochondrial and nuclear DNA results due to sex-biased mating (in alignment with the laboratory cross-mating studies) or does it simply reflect the inadequacies of approaches like AFLPs to resolve nuclear genetic differences in closely related lineages, ones that might only be resolved by analysing genome-wide data? Mitochondrial DNA is generally inherited through the maternal line (and can reveal sex-biased population processes) but it also generally has a higher mutation rate than nuclear DNA (especially inn insects) because of the biochemical conditions within the mitochondria.

The results of the present study indicate that gene flow takes place in the field between rice and water oats associated populations of C. suppressalis. The key question now is how much gene flow occurs between populations on these two hosts under natural conditions. Relatively infrequent (and directional) mating between the two hosts might be sufficient for the microsatellite data to indicate a lack of host association, but also allow the two populations to remain differentiated. Parentage analysis between samples from the same two hosts would allow this to be tested.

host-associated populations than the microsatellite data does, as in S. frugiperda58.

Understanding the geographic context of the interaction between rice and water

oats populations is also critical to interpreting their evolutionary relationships relative

to each other. This type of scenario is often interpreted as providing evidence of

ongoing, sympatric speciation as was the case with the apple maggot fly, Rhagoletis

pomonella59. This pest was believed to be undergoing speciation in sympatry

following a host switch from hawthorn to apple following the introduction of apples

to the USA by European settlers60. Subsequent studies showed, however, that the

divergence between the two host-associated populations happened on different

mountain ranges in Mexico61. This highlights how data can be misinterpreted if not

considered at the appropriate geographic scale. Studies using mtDNA have indicated

that there is geographic structure across C. suppressalis populations in China, but

these studies have only sampled this insect from rice62,63. Further genetic studies

between both host plant species and across the distribution of the rice stem borer would be very helpful in interpreting the evolution of this system.

Our results indicate gene flow between the water-oats and rice populations (as

This article is protected by copyright. All rights reserved. defined by the time at which they mate), and that the ‘rice’ population can be found on both rice and water oats (Fig. 2). This has implications for resistance management following the deployment of Bt rice. Water oats is widely cultivated from northern

Beijing to southern Guangdong province and Taiwan, and from eastern Sichuan

province to Shanghai45. Both rice and water oats are often planted next to each other

as crops and given that water oats is larger than rice, less than 10% land area would

likely provide a 10% refuge by volume. In parts of china where water oats is not

cultivated, the wild relative of water oats is often found in irrigation canals and ponds close to cultivated rice64. Our results show that water oats growing near rice would likely provide suitable refuges for Bt resistance management of this serious rice pest,

and structured refuges might be unnecessary. Future studies on the dispersal of C.

suppressalis would clarify the geographic scale at which wild populations of Z.

latifolia would serve as unstructured refuges.

Acknowledgments

This project was supported by the National Key R&D Program of China

(2017YFD0200602) and the Fundamental Research Funds for the Central Universities of China (2662017JC006).

References

1. Hereward J, Hutchinson JA, McCulloch GA and Walter GH, Divergence among generalist herbivores: the Frankliniella schultzei species complex in Australia (Thysanoptera: Thripidae). Arthropod-Plant Interactions, 1-13(2017). 2. Paterson, HEH, The recognition of cryptic species among economically important insects. In: Heliothis: Research Methods and Prospects (ed. MP Zalucki), pp. 1-10. Springer-Verlag, New York, USA (1991). 3. Walter GH, Insect pest management and ecological research. Cambridge University Press, Cambridge, UK. (2003). 4. Bickford D, Lohman DJ, Sodhi NS, Ng PKL, Meier R, et al., Cryptic species as a window on diversity and conservation. Trends Ecol Evol 22:148–55 (2007). 5. Zhang Q, Strategies for developing green super rice. Proc Natl Acad Sci USA 104:16402–16409 (2007). 6. Jiang WH, Li HD, Cheng XF, Ye JR, Feng YB and Han ZJ, Study on host plants for reproduction of Chilo Suppressalis. J Asia-Pac Entomol 18 (3): 591–595 (2015). 7. Takasaki T, Noda M and Murata Z, Ecology of rice stem borer moth on Zizania latifolia. I. Occurrence of rice stem borer on Zizania latifolia. Proc Assoc Plant Prot Kyushu 15:118–121 (1969). (in Japanese) 8. Chen JM, Yu XP, Zheng XS, Xu HX, Lv ZX, Zhang JF, Biological performance of the striped stem borer, Chilo suppressalis Walker fed on Jiaobai, Zizania caduciflora, and rice plants. Acta Agric Zhejiangensis 15: 139–143 (2003). (in Chinese)

This article is protected by copyright. All rights reserved. 9. Matsukura K, Hoshizaki S, Ishikawa Y and Tatsuki S, Morphometric differences between rice and water-oats population of the striped stem borer moth, Chilo suppressalis (Lepidoptera: Crambidae). Appl Entomol Zool 41:529–535 (2006). 10. Samudra IM , Emura K , Hoshizaki S, et al., Temporal differences in mating behavior between rice- and water-oats-populations of the striped stem borer, Chilo suppressalis (Walker)(Lepidoptera: Crambidae). Appl Entomol Zool 37(2):257-262(2002). 11. Hou M, Lin W and Han Y, Seasonal change in supercooling points and glycerol content in overwintering larvae of the Asiatic rice borer from rice and water-oat plants. Environ Entomol 38:1182–1188 (2009). 12. Ding N, Dalin P, Zhu QH, et al., A comparison of the larval overwintering biology of the striped stem borer, Chilo suppressalis (Lepidoptera: Crambidae), in rice and water-oat fields. Appl Entomol Zool 48:147-153 (2013). 13. Matsukura K, Hoshizaki S, Ishikawa Y and Tatsuki S, Differences in timing of the emergence of the overwintering generation between rice and water-oats populations of the striped stem borer moth, Chilo suppressalis (lepidoptera: crambidae). Appl Entomol Zool, 44:485-489 (2009). 14. Konno Y and Tanaka F, Aliesterase isozymes and insecticide susceptibility in rice-feeding and water-oat-feeding strains of the rice stem borer, Chilo suppressalis Walker (Lepidoptera: Pyralidae). Appl Entomol Zool 31:326–329 (1996a). 15. Konno Y and Tanaka F, Mating time of the rice-feeding and water-oats-feeding strains of the rice stem borer, Chilo suppressalis (Walker) (Lepidoptera: Pyralidae). Jpn J Appl Entomol Zool 40: 245–247 (1996b). (in Japanese with English summary)

This article is protected by copyright. All rights reserved. 16. Huang C, Hu B, Li J, Wang Y, Water-oats harbors two strains of the striped stem borer Chilo suppressalis (Lepidoptera: Crambidae) with temporal divergence in mating behavior. Appl Entomol Zool 51:457-463 (2016). 17. Ishiguro N, Yoshida K and Tsuchida K, Genetic differences between rice and water-oat feeders in the rice stem borer, Chilo suppressalis (Walker) (Lepidoptera: Crambidae). Appl Entomol Zool 41:585–593 (2006). 18. Liu Q, Hallerman E, Peng Y, Li Y, Development of Bt rice and Bt maize in China and their efficacy in target pest control. Int J Mol Sci 17(10):1561 (2016). 19. Qiao F, Huang J, Rozelle S and Wilen J, Natural refuge crops, buildup of resistance, and zero-refugestrategy for Bt cotton in China. Science China Life Sciences 53 (10): 1227–38(2010). 20. Jin L, Zhang H, Lu Y , Yang Y, Wu K and Tabashnik BE, et al., Large-scale test of the natural refuge strategy for delaying insect resistance to transgenic Bt crops. Nat Biotech, 33(2): 169-174 (2015). 21. Wan P, Huang Y, Wu H, Huang M, Cong S, Tabashnik BE and Wu K, Increased frequency of pink bollworm resistance to Bt toxin Cry1Ac in China. PloS One 7 (1): e29975 (2012). 22. Ishiguro N and Tsuchida K, Polymorphic microsatellite loci for the rice stem borer, Chilo suppressalis (Walker) (Lepidoptera: Crambidae). Appl Entomol Zool 41: 565-568 (2006). 23. Liu YD, Hou ML and Hao LX, Twelve nuclear microsatellite loci for rice stem borer (Chilo suppressalis W.). Mol Ecol Resour 9: 180-182 (2009). 24. Peakall R and Smouse PE, GenAlEx 6: genetic analysis in Excel. Population genetic software for teaching and research. Mol Ecol Notes 6:288-295 (2006). 25. Raymond M and Rousset F, GENEPOP (Version 1.2): Population genetics software for exact tests and ecumenicism. J Hered 86 :248-249 (1995).

This article is protected by copyright. All rights reserved. 26. Pritchard JK, Stephens MJ and Donnelly PJ, Inference of population structure using multilocus genotype data. Genetics 155: 945-959 (2000). 27. Earl DA and von Holdt BM, STRUCTURE HARVESTER: a website and program for visualizing STRUCTURE output and implementing the Evanno method. Conserv Genet Resour 4: 359-361 (2012). 28. Evanno GS, Regnaut SJ and Goudet J, Detecting the number of clusters of individuals using the software STRUCTURE: A simulation study. Mol Ecol 14: 2611-2620 (2005). 29. Jakobsson M, Rosenberg NA (2007) CLUMPP: a cluster matching and permutation program for dealing with label switching and multimodality in analysis of population structure. Bioinformatics 23:1801–1806. 30. Rosenberg NA, DISTRUCT: a program for the graphical display of population structure. Mol Ecol Notes 4:137–138 (2004). 31. Jombart T, adegenet: a R package for the multivariate analysis of genetic markers. Bioinformatics, 24(11):1403-1405 (2008). 32. R Core Team, R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/. (2018). 33. Hebert PD, Cywinska A, Ball SL, de Waard JR, Biological identifications through DNA barcodes. Pro R Soc Lond B. 270: 313-321 (2003) 34. Caterino MS and Sperling FA, Papilio phylogeny based on mitochondrial cytochrome oxidase I and II genes. Mol Phylogenet Evol 11: 122-137 (1999). 35. von Dohlen CD and Teulon DAJ. Phylogeny and historical biogeography of New Zealand indigenous Aphidini aphids (Hemiptera, Aphididae): an hypothesis. Ann Entomol Soc Am 96: 107–116 (2003).

This article is protected by copyright. All rights reserved. 36. Bandelt HJ, Forster P and Rohl A, Median-joining networks for inferring intraspecific phylogenies. Mol Biol Evol, 16: 37-48 (1999). 37. Saitou N and Nei M, The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol Biol Evol 4: 406– 425 (1987). 38. Zheng X, Yu X, Lu Z, Chen J, Xu H, Zhang J, Ye G, Occurrence pattern of the striped stem borer, Chilo suppressalis Walker, in water-oat, Zizania caduciflora L. fields. Acta Agric Zhejiangensis 15:133-138 (2003). (in Chinese) 39. Quan WL, Liu W, Zhou RQ, Chen R, Ma WH, Lei CL, et al., Difference in diel mating time contributes to assortative mating between host plant-associated populations of Chilo suppressalis. Sci Rep, 7:45265 (2017). 40. Hu B, Mating behavior and genetic structure of rice and water-oats associated population of the Chilo supperessalis (Walker)[D]. Wuhan, Huazhong Agricultural University, 2013. 41. Yu X, Xu H, Lv Z, Chen J, Zheng X, Tao L, Differentiation of striped stem borer (SSB), Chilo suppressalis Walker from rice and Zizania caduciflora habitats. Acta Ecologica Sinica 22:341-345 (2002). (in Chinese) 42. Hilbert L, Góes Neves E, Pugliese F, Whitney BS, Shock M et al., Evidence for mid-Holocene rice domestication in the Americas. Nat Ecol Evolut 1(11): 1693– 1698 (2017). 43. Zong Y, Chen Z, Innes JB, Chen C, Wang Z, Wang H, Fire and flood management of coastal swamp enabled first rice paddy cultivation in east China. Nature 449 (7161): 459–462 (2007). 44. Sweeney M and McCouch S, The complex history of the domestication of tice. Ann Botany 100 (5): 951–957 (2007). 45. Guo HB, Li SM, Peng J, Ke WD, Zizania latifolia Turcz. cultivated in China. Genet Resour Crop Ev 54:1211–1217 (2007).

This article is protected by copyright. All rights reserved. 46. Meagher RL, Nagoshi RN, Stuhl C, Mitchell ER, Larval development of fall armyworm (Lepidoptera: Noctuidae) on different cover crop plants. Fla Entomol 87:454–460 (2004). 47. Barros EM, Torres JB, Ruberson JR, Oliveira MD, Development of Spodoptera frugiperda on different hosts and damage to reproductive structures in cotton. Entomol Exp Appl 137:237–245 (2010). 48. Prowell DP, McMichael M and Silvain JF, Multilocus genetic analysis of host use, introgression, and speciation in host strains of fall armyworm (Lepidoptera: Noctuidae). Ann Entomol Soc Am 97:1034–1044(2004). 49. Pashley DP, Host-associated genetic differentiation in fall armyworm (Lepidoptera: Noctuidae): a sibling species complex? Entomol Soc Am 79:898– 904 (1986). 50. Pashley DP, Hammond AM, Hardy TN, Reproductive isolating mechanisms in fall armyworm host strains (Lepidoptera: Noctuidae). Ann Entomol Soc Am 85:400–405 (1992). 51. Schöfl G, Heckel DG and Groot AT, Time-shifted reproductive behaviours among fall armyworm (Noctuidae: Spodoptera frugiperda) host strains: evidence for differing modes of inheritance. J Evol Biol, 22: 13 (2009). 52. Nagoshi RN, Meagher RL, Adamczyk JJJ, Braman SK, Brandenburg RL, Nuessly G, New restriction fragment length polymorphisms in the cytochrome oxidase i gene facilitate host strain identification of fall armyworm (lepidoptera: noctuidae) populations in the southeastern united states. J Econ Entomol 99: 671 (2006). 53. Nagoshi RN, The fall armyworm Triose Phosphate Isomerase (Tpi) gene as a marker of strain identity and interstrain mating. Ann Entomol Soc Am 103: 283– 292 (2010).

This article is protected by copyright. All rights reserved. 54. Nagoshi RN, Meagher RL, Hay-Roe M, Inferring the annual migration patterns of fall armyworm (Lepidoptera: Noctuidae) in the United States from mitochondrial haplotypes. Ecol Evol 2:1458–1467 (2012). 55. Juárez ML, Schöfl G, Vera MT, Vilardi JC, Murúa MG, Willink E, et al., Population structure of Spodoptera Frugiperda maize and rice host forms in South America: Are They Host Strains? Entomol Exp Appl 152 (3): 182–199 (2014). 56. Groot AT, Marr M, Heckel DG, Schofl G, The roles and interactions of reproductive isolation mechanisms in fall armyworm (Lepidoptera: Noctuidae) host strains. Ecol Entomol 35:105–118 (2010). 57. Gouin A, Bretaudeau A, Nam K, et al., Two genomes of highly polyphagous lepidopteran pests (Spodoptera frugiperda, Noctuidae) with different host-plant ranges. Sci Rep 7: 11816(2017). 58. Kiwoong NAM, Nhim S, Robin S, Bretaudeau A, Negre N and d’Alencon E, Divergent selection causes whole genome differentiation without physical linkage among the targets in Spodoptera Frugiperda (Noctuidae). bioRxiv, https://doi.org/10.1101/452870 (2018). 59. Bush GL, Host race formation and sympatric speciation in Rhagoletis fruit ﬂies (Diptera: Tephritidae). Psyche 99:335–357(1992). 60. Berlocher SH and Mcpheron BA, Population structure of Rhagoletis pomonella, the apple maggot fly. Heredity 77(1):83-99 (1996). 61. Xie X, Rull J, Michel AP, Velez S, Forbes AA and Lobo NF, et al., Hawthorn-infesting populations of Rhagoletis pomonella in mexico and speciation mode plurality. Evolution, 61(5): 1091-1105 (2007). 62. Meng XF, Shi Mi and Chen XX, Population genetic structure of Chilo Suppressalis (Walker) (Lepidoptera: Crambidae): strong subdivision in China

This article is protected by copyright. All rights reserved. inferred from microsatellite markers and mtDNA gene sequences. Mol Ecol 17 (12): 2880–2897 (2008). 63. Tang, Xiao-Tian, Fu-Shan Zheng, Ming-Xing Lu, and Yu-Zhou Du. New ideas about genetic differentiation of Chilo Suppressalis (Lepidoptera: Pyralidae) populations in China based on the mtDNA Cytochrome B gene. Mitochondrial DNA 27 (2): 1567–1573 (2016). 64. Yuanyuan Chen, Yang Liu, Xiangrong Fan, Wei Li and Yanling Liu. Landscape-Scale Genetic Structure of Wild Rice Zizania latifolia: The Roles of Rivers, Mountains and Fragmentation. Frontiers in Ecology and Evolution. (5):17 (2017).

Figure legends.

Fig. 1. Rice stem borer (bottom right) was sampled from rice (top left) and water oats

(bottom left) from the five localities shown in the map.

Fig. 2. Selection of rice and water oats populations based on the time that they mated

(data from Huang et al. 2016).

Fig. 3. PCA of the microsatellite data for all individuals genotyped in the first phase of the study in 2014 (top) and structure plot of the same data for K=2 (bottom).

Fig. 4. PCA of the microsatellite data for all individuals genotyped in the second

phase of the study in 2016 (top) and structure plot of the same data for K=2 (bottom).

Fig. 5. Phylogenetic tree of rice and water oats populations of C. suppressalis from

MrBayes runs based on combined COI and COII sequences (left), with Chilo

auricilius KJ174087 as the outgroup. TCS network inferred from EF1-αsequences

and COI and COII mitochondrial sequences are also presented. The samples were

collected from Hangzhou and Wuhan in 2016 and were screened by timing of mating

(see Methods) before being sequenced.

Table 1 Pairwise FST values (below diagonal) and statistical significance for each

comparison (above diagonal) to indicate allelic differentiation across 10 C.

suppressalis populations collected from five sites and two host plant species, rice (R)

and water oats (W), in 2014

Pop. HZ_R HZ_W TZ_R TZ_W JH_R JH_W CD_R CD_W WH_R WH_W

HZ_R 0.001 0.001 0.018 0.001 0.001 0.265 0.002 0.017 0.013

HZ_W 0.062 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001

TZ_W 0.019 0.074 0.019 0.003 0.003 0.009 0.001 0.004 0.002

JH_R 0.032 0.089 0.030 0.029 0.001 0.033 0.001 0.001 0.001

JH_W 0.042 0.080 0.017 0.022 0.052 0.001 0.001 0.001 0.001

CD_R 0.012 0.067 0.023 0.022 0.020 0.036 0.001 0.038 0.001

CD_W 0.029 0.061 0.035 0.033 0.053 0.039 0.032 0.001 0.001

WH_R 0.021 0.058 0.025 0.025 0.038 0.038 0.018 0.029 0.001

WH_W 0.022 0.068 0.049 0.032 0.052 0.056 0.035 0.041 0.032

HZ, Hangzhou; TZ, Taizhou; JH, Jinhua; CD, Changde; WH, Wuhan

comparison (above diagonal) of allelic differentiation for four C. suppressalis

populations collected from two sites and two host plants, rice (_R) and water-oats

(_W), in 2016. Samples were screened by timing of mating (see methods) before being genotyped; HZ, Hangzhou; WH, Wuhan

Pop. WH-R WH-W HZ-R HZ-W

WH-R 0.059 0.001 0.001

WH-W 0.019 0.001 0.023

HZ-R 0.051 0.053 0.001

HZ-W 0.029 0.021 0.052