ABSTRACT
The study of the mechanisms involved in adaptation remains a timely issue, particularly in the context of global changes. To better understand these mechanisms of rapid adaptation, invasive species are a good model because they are subjected to new and/or different environmental factors. Using different lines of different geographical origin of the invasive pest Drosophila suzukii, we characterized the phenotypic response to oxidative stress. Subsequently, we tested the involvement of the Jheh gene cluster in this response and the possible role of transposable elements. We show that the resistance to oxidative stress of the lines appears to be related to their invasive status and we confirm the role of the Jheh gene cluster in this response. We have not identified any transposable elements in this gene region that could influence the expression of the gene.
Summary statement The responses to oxidative stress of the invasive species, Drosophila suzukii, show variability between genotypes related to their invasion status. The genes of the juvenile hormone epoxide hydrolase cluster are involved in this response.
INTRODUCTION
The rapid spread of invasive species in a huge spectrum of environments relies on multiple factors, from genetics to phenotypic plasticity, probably including fine molecular mechanism such as hormonal production or epigenetic gene regulation (Beldade et al., 2011; Marin et al., 2019; Stapley et al., 2015). Phenotypic plasticity, i.e., the ability of a genotype to express different phenotypes in different environments (Ghalambor et al., 2015) has been proposed as one of the most promising explanations for invasive success, particularly in the case of founder population depleted of genetic variation (Estoup et al., 2016; Marin et al., 2019). Among deleterious environments that can be encountered by invasive species, oxidative stress caused by phytosanitary products is one of them. The invasive pest, Drosophila suzukii, is a good model to investigate the adaptive process during invasion (Gibert et al., 2016). This species which belong to the group of the fruit fly D. melanogaster, originally comes from Asia and was detected simultaneously both in North America (U.S.A) and in Europe in 2008. North America was invaded by native Japan populations derived from Hawaii. In Europe, several introductions were detected from U.S.A and from China (Fraimout et al., 2017). Currently, D. suzukii is present in both North and South America, in Europe from the south (Spain) to the East (Poland, Ukraine) and it has also been observed in Russia (CABI, 2020; Lavrinienko et al., 2017).
Characterization of the phenotypic and molecular responses of D. suzukii to changing environmental conditions may provide information to the mechanisms involved in the ability of invasive species to cope with environmental variation. Paraquat (N,N′-dimethyl-4,4′-bipyridinium dichloride) is one of the most widely used herbicide in the world leading to the production of ROS (reactive oxygen species) (Tsai, 2018). Oxidative stress due to the use of paraquat in the field has also been used in the laboratory as a good proxy for studying stress resistance (Bus J S and Gibson J E, 1984; Rzezniczak et al., 2011). Paraquat was banned since 2007 in Europe but is still used in many other regions like in U.S.A or Japan. Paraquat exposition is known to induce a reduction in the lifespan associated with changes in gene expression (Finkel and Holbrook, 2000; Liguori et al., 2018; Vermeulen et al., 2005). One of the candidate genes involved in paraquat resistance is the cluster of Jheh (Juvenile hormone epoxide hydrolase) genes, which are not only involved in the lifespan but also in response to the oxidative environment (Flatt and Kawecki, 2007; Guio et al., 2014). Moreover in D. melanogaster, an insertion of a transposable element (TE) Bari-Jheh, near the cluster of the Jheh genes has been described as driving an increase of resistance in presence of paraquat (Guio et al., 2014).
Using several strains of D. suzukii, we measured responses to oxidative stress at the phenotypic and molecular level. We made the hypothesis that different genetic backgrounds from native and invasive populations will have different responses to oxidative stress and that the Jheh cluster may be involved on it. Due to the over-representation of TEs in the genome of D. suzukii (33% of the repeated elements, (Sessegolo et al., 2016)), compared to other Drosophila species, we looked for the presence of TEs in this region in the different lines. We monitored lifespan after paraquat exposure and measured the expression of three genes of Jheh cluster Jheh-1, Jheh-2 and Jheh-3 in six isofemale lines, four from the invasive regions, North America (Watsonville and Dayton) and France (Paris and Montpellier) and two from the native area, Japan (Sapporo and Tokyo). We evaluated the genetic diversity within and between lines by sequencing introns of the Jheh genes, searched for TEs and for transcription factor binding sites (TFBS). Our results suggest a strong effect of the genotype on the resistance to stress and changes in Jheh expression levels, with no link with TEs.
MATERIAL AND METHODS
Drosophila suzukii lines and rearing conditions
D. suzukii lines were sampled in 2014 from one native country (Japan: Sapporo and Tokyo) and 2 invaded areas (USA: Watsonville and Dayton and France: Montpellier and Paris). Field-inseminated females were isolated to establish half-sib families called isofemale lines commonly used to investigate Drosophila natural populations (David et al., 2005). Flies were reared in modified medium (drosophila agar type, ref.66-103, Apex™,9 g.L−1; cornmeal 33 g.L−1; yeast, dried yeast, ref.75570, LYNSIDEⓇ 17 g.L−1; industrial sugar 50 g.L−1; nipagin, Tegosept, ref.20-258, Apex™ 4 g.L−1; 96% ethanol 40 ml.L−1; distilled water 1 L) from Dalton et al., (2011), in a humidified, temperature-controlled incubator at 22.5°C, 70 % of relative humidity and a 16:8 LD cycle. The recipe of the modified medium was to bring to boil agar, cornmeal, yeast extract and sugar in distilled water. Then wait out of the fire about 10 minutes until the mixture cooled to 53°C before adding diluted nipagin in 96% ethanol. Medium is then poured in vials and cooled at room temperature before to be stored at 4°C. All the experiments were made with 4 to 7 days old flies.
Oxidative stress resistance experiments
We used paraquat (methyl viologen dichloride hydrate, ref. 75365-73-0, Sigma-AldrichⓇ) to mimic oxidative stress. Oxidative stress was assessed by adding paraquat directly in the medium (10 mM) before the cooling step and below 53°C. The control experiment was made with the same medium but without paraquat. We used one isofemale line per locality (total of six) named Montpellier (France), Paris (France), Sapporo (Japan), Tokyo (Japan), Dayton (U.S.A.) and Watsonville (U.S.A.). We made three replicates per line and per sex, with ten flies per replicate. Survival was monitored every 24h. Flies were transferred into new vials every three to four days to limit microbial development.
RT-qPCR analysis of Jheh genes
We quantified the expression of the three Jheh genes (Jheh-1, Jheh-2 and Jheh-3) by RT-qPCR after induction of oxidative stress and in control condition. Adult males and females were exposed during 24h to medium culture with 20 mM of paraquat. After 24h the flies were immediately dissected in PBS 1X solution (Gibco Thermo-Fisher) in order to extract carcasses for both sexes and eliminate germline tissues. We made three replicates per sex and treatment and used four flies per replicate.
RNA extraction was made using Direct-zol™-96 RNA Kits (Zymo Research), following the manufacturer recommendations and RNA was treated with DNAse. cDNA were obtained from 0.5 μg of RNA using SuperScript™ IV VILO™ Master Mix (Invitrogen). RT negative control was made with RNA but without the reverse transcriptase to control for genomic DNA contamination. cDNA were stored at −80°C before the quantification step. Gene expression was then quantified by quantitative PCR and Rp49 was used as housekeeping gene. Primers were designed using the D. suzukii referenced genome (Table S1, (Chiu et al., 2013)). Their efficiency was between 91.1% to 97.2% (RP49: 91.6 %, Jheh-1: 97.2%, Jheh-2: 95.2%, Jheh-3: 91.1%). 2 μl of the cDNA sample were supplemented with 5 μL of SsoADV Universal SYBR Green Supermix (BioRad) mix 2X, 0.3 μl of each primer (10μM) and 2.4 μl of pure water. We made technical duplicates for each sample. PCR reactions were made in a BioRAd CFX-96 with a program consisted of an initial activation of 95°C for 10 minutes and then 40 cycles each comprising 15 seconds at 95°C, 10 seconds at 60°C and 72°C.
Genetic diversity of isofemale lines
We sequenced intronic regions of Jheh gene cluster of the six lines used in this study. DNA was extracted individually from 10 females per line with the 96-Well Plate Animal Genomic DNA Miniprep kit (ref. BS437, Biobasic) following the manufacturer instructions. Primers were designed to flank the intronic regions for the three Jheh genes (Table S1) and Phusion high fidelity DNA Polymerase (2 U/μL) (F-530XL Thermofisher Scientific) was used to amplify sequences. The same PCR program was used for all primers pairs: 98°C for 10 minutes, followed by 40 cycles composed of 30 seconds at 98°C, 1 minute at 56°C and 20 seconds at 72°C and a final elongation step for 1 minute at 72°C. The sequencing of the two strands was done directly from the PCR product by BIOFIDAL sequencing company (Vaulx en Velin, France). Sequences were manually curated with CLC Main Workbench 8 software (Qiagen) before being aligneed with the Muscle program implemented in the workbench to generate haplotypes by line for each intron. MEGA X software was used to calculate pairwise comparison and nucleotide diversity using p-method option (Table S2–S3) (Kumar et al., 2018).
Detection of Transposable elements and transcription factor binding sites
We sequenced the intergenic regions of the Jheh gene cluster, plus the 5’ and 3’ regions of the cluster (Table S1). DNA was extracted from one female per population as described above. Classical PCR method was used with the following program, 10 minutes at 95°C followed by several cycles composed of 30 seconds at 95°C, 30 seconds at 63°C, 3 minutes at 72°C and a final elongation of 15 minutes at 72°C. The number of cycles was25 for the region before Jheh-1 and between Jheh-1 and Jheh-2, 35 cycles for the region between Jheh-2 and Jheh-3 and 30 cycles after Jheh-3. We identified TEs in the intergenic regions by a blast against a homemade data base of the TE sequences from the D. suzukii reference genome (Paris et al., 2020, Mérel et al., in prep.).
For TFBS (Table 1), we used conSite website to screen all TFBS from insect in our sequences (Sandelin et al., 2004). To complete our analysis, we used the TFBS obtained from Villanueva-Cañas et al. (2019) and we extracted PFM (position frequency matrix) of the 14 TFBS from the JASPAR2018 database (v.1.1.1) (Parcy et al., 2017). Then, we used TFBSTools (v.1.22.0) package from R software (v. 3.6.0) to convert in PWM (position weight matrix), and then search on the 6 lines and the reference genome of D. suzukii (Paris et al., 2020; R Core Team, 2019; Tan and Lenhard, 2016).
Statistical analysis
Survival monitoring
Survival data were analyzed using a linear mixed model with lmer function from lme4 provided on R (v. 3.6.0) after a log transformation, confirmation of normality and homoscedasticity (Bates et al., 2015). This model was chosen after log-likelihood comparison between models (linear model with raw or log transformed data, survival model with a Weibull distribution).
We analyzed sexes separately to limit interaction terms, and focused on the effect of the treatment, the lines and their interaction. Biological replicates were added as random effect and we plotted exponential of the values and associated confidence interval on the Fig. S1. Those effects can be interpreted as multiplicative effect on the mean lifespan compared to the reference chosen here as the non-exposed group from Sapporo (e.g. the Sapporo reference is centered on 1 and the effect of paraquat 0.18 involves a survival time under paraquat for Sapporo of 0.18 or 18% of the survival time without paraquat).
qPCR analysis
RT-qPCR raw data were analyzed using R and EasyqpcR library (1.21.0) for the quantification and normalization with RP49 (Sylvain, 2012). Data were analyzed separately for the three genes (Jheh-1 -2 and -3) and sex using a linear model (ANOVA2, Table S4) after log transformation to validate homoscedasticity and normality. Pairwise comparisons were made using a Tuckey test.
RESULTS
D. suzukii wild type lines have significant differences in life span
To investigate the influence of the genotypes from different geographical origins on the lifespan, we compared the invasive and native D. suzukii lines in control condition (Fig. 1A, Fig. S1 & Table 2). The lifespan ranges from 31 to 55 days for females and from 25 to 45 days for males. For females we observed a strong genotype effect related to geographic location: the genotypes that lived the longest were those of Dayton and Paris (about 1.88-1.96 times more than Sapporo, Fig. S1). Sapporo, Tokyo, and Watsonville were not significantly different and with the lowest lifespan. For males, the four invasive genotypes from Europe and U.S.A had a higher lifespan than Sapporo. Tokyo was similar to Sapporo.
As expected, exposure to paraquat reduced life span on average from 82 to 77% for females and males (Fig. 1A, Fig. S1). The two lines with the best paraquat resistance in absolute value were still Dayton and Paris in both sexes (Fig. 1A and Table 2). We then wanted to have an estimate of paraquat sensitivity (i.e., the slope difference Fig.1 B) taking into account the longevity of each line by estimating the value of the interaction coefficients (i.e., the slope difference compared to Sapporo) in Table 2 and statistically tested in Fig. S1. Again, the effect was not similar between genotypes and sexes. For females, Paris was the line presenting significantly the highest sensitivity (−0.87, Table 2) with a reduction of 28% of the life span compared to Sapporo (Fig. S1). For males, the reduction in life span was significantly the highest for Montpellier and Watsonville (−0.84 and −0.83) with a reduction from 34 and 32% by comparison with Sapporo. These results reveal a strong genotype-by-environment interaction in the response to oxidative stress and also a sex effect. It is interesting to note that despite the shorter life span of Japanese genotypes, and in particular of Sapporo in the absence of treatment, these genotypes were the most resistant to paraquat exposure, as shown by the lowest ratio of paraquat lifetime to control lifetime (Table 2).
Jheh genes expression changes with the paraquat treatment
To investigate the effect of paraquat-mediated oxidative stress on the gene expression level, we focused on Jheh gene cluster described as potentially involved in stress response in insects and mammals (Guio et al., 2014; Oesch et al., 2000). We quantified the level of expression of the Jheh genes (Jheh-1, Jheh-2 and Jheh-3) in adult males and females flies for the six genotypes described above (Fig. 2).
We observed strong differences between males and females. For males, gene expression was not significantly different between control and paraquat treatment for the six genotypes and for the three gene. In females, the effect of paraquat was different according to the gene and the genotype (Table S4). For Jheh-1 and Jheh-2, oxidative stress resulted in a significant increase of gene expression for the two French genotypes and the Tokyo genotype. On the contrary, the Sapporo genotype exhibited a significant reduction of Jheh-1 expression in presence of paraquat. For Jheh-3, we observed a downregulation of the gene expression only for the Sapporo genotype.
Low Genetic diversity of lines in Jheh cluster
To assess the levels of neutral genetic diversity within and between lines, we sequenced intronic regions for Jheh genes for each genotype (Fig. 3). As expected, the within-line polymorphism was very low (Table 3, Fig. S2), with the exception of Watsonville with 0.0792 for the first intron of Jheh-1. The number of haplotypes was also low (Table S2). The first intron of Jheh-2 presents the highest levels of diversity, contrasting with the other introns. This corresponds to a residual polymorphism that is still present in the lines despite the laboratory rearing.
Depending on the intronic regions we found between two to four haplotypes per genotype (Table S2). We computed the diversity between genotypes (global intronic nucleotide diversity π) using the most common haplotype for each intron, showing that on average these values are very small, with the highest value for Jheh-2.1 as mentioned above (Table S3).
Jheh harbour transcription factor binding sites (TFBS)
Transcription factor binding site (TFBS) are cis regulatory sequences that are recognized by transcription factors and modify gene expression. Several TFBS are known to be involved during oxidative response. We detected 9 of the 14 previous identified TFBS in the intergenic regions: HIF1A, br, cad, Cf2, Deaf1, CnC, dl, hb and Ubx (Fig. 3 & 4, Table S5). Comparison of the number of TFBS between genotypes (Fig. 4) revealed several differences but not clear link with the changes in expression observed for the Jheh genes. For example, the Sapporo genotype which consistently showed a decrease in the expression of all three genes, did not appear to have a different specific TFBS. The two French genotypes which exhibited systematically an increase of expression after paraquat treatment appeared to have an increased number of putative TFBS. For example, the two French genotypes showed two Deaf1 motives when compared to the other genotypes upstream of TSS of the Jheh-1 gene. In the case of the Jheh-2 gene, the French genotypes presented a significant number of TFBS, with for example six TFBS for the Montpellier genotype (2Ubx, 2hb, CnC and cad). In this region, no genotype showed the same pattern of TFBS and it was similar for Jheh-3.
Transposable elements do not affect Jheh gene expression
The presence of TE in the vicinity or within the Jheh cluster could impact the gene expression during oxidative stress because they could bring Antioxidant Response Element for transcription factors or by modifying chromatin state (Guio and González, 2015; Guio et al., 2014). Surprisingly, and even if D. suzukii harbors more than 30% of TEs, no full insertion was observed in the Jheh cluster indicating that we are probably in regions of high recombination. However, we did identify small pieces of TE that are quite conserved between the genotypes but no TFBS were detected in these sequences (Fig. 3, Table 4 & S5). No obvious link seems to exist between gene expression and the presence of TE in the Jheh cluster.
DISCUSSION
A growing body of literature suggests that responses to oxidative stress in Drosophila may be mediated by insertions of TEs, that in some cases could affect gene expression or the chromatin structure (Guio et al., 2014). In D. melanogaster, the Jheh gene cluster has been shown to be involved in the response to paraquat treatment and associated with local adaptation. Guio et al. (2014) compared D. melanogaster genotypes with and without Bari-Jheh TE insertion, and showed that, (i) TE insertion had a cost in the absence of stress, (ii) TE insertion confer increased survival in the presence of oxidative stress, (iii) TE insertion provides antioxidant response elements (AREs) that contribute to altered gene expression (Guio and González, 2015; Guio et al., 2014). In this study, we analyzed the expression of the Jheh gene cluster in several genotypes of D. suzukii to test whether the Jheh gene cluster is involved in the oxidative stress response and whether TEs could also be associated with alterations in gene expression.
We first measured the life span of flies without treatment. We showed that flies of the Japanese genotypes exhibited the shortest lifespan in both males and females. Surprisingly, these lines showed an increased resistance to oxidative stress. The French lines were more sensitive to paraquat than the American ones, although notable differences were observed between lines from the same continent. The negative association we observed between longevity and paraquat resistance had not been observed in previous work with D. melanogaster, in which the opposite association was observed (Finkel and Holbrook, 2000; Liguori et al., 2018). It could be argued that the use of paraquat in Europe has been banned since 2007, which could lead to a loosening of selection on genes related to paraquat resistance, as observed in other organisms (Campos et al., 2014; Shaw, 2000).
We then measured the expression of the Jheh genes previously reported to be involved in the oxidative response. Consistent with the literature of D. melanogaster, we found sex-specific responses to oxidative stress (Guio et al., 2014; Weber et al., 2012). For Jheh-1 and Jheh-2, we observed a significant effect of genotype and treatment, but only for females, contrary to what was reported in D. melanogaster. For Jheh-3, treatment and genotype effect were significant for both males and females. These differences in gene expression could not be associated with the presence of TEs insertions, since only partial sequences were present in the intergenic regions. The presence of various TFBS could contribute to the observed differences. We also quantified the polymorphism in our lines, which could be associated with differences in gene expression. We did not observe total homozygosity in the lines but genetic diversity was much lower than what is observed in natural populations of D. melanogaster. Lack et al. (2016) studied populations from several continents and measured values of nucleotide diversity of up to 0.401 within the population. For inbred DGRP (Drosophila Genetic Reference Panel) lines, the mean intronic diversity was 0.0076 ±0.008, which is close to the values we observed (MacKay et al., 2012). It is therefore unlikely that the residual polymorphism in the Jheh gene region can explain the differences in gene expression.
The striking result in our analysis is the similar pattern of changes in the expression of Jheh-1 and Jheh-2 in females of European genotypes, with an increase in expression, which was associated with lower resistance to oxidative stress, since these are the most sensitive genotypes. On the contrary, the Sapporo genotype systematically showed a reduction in the expression levels of the three genes, which could also be associated with increased resistance in the presence of paraquat, but this was not observed for the Tokyo genotype.
Conclusion
In conclusion, our work shows for the first time how various genotypes of D. suzukii respond to oxidative stress and suggests that populations found in invaded areas are more sensitive than Japanese populations, specially the French ones. We have also confirmed that the Jheh gene cluster is involved in the response to oxidative stress also in D. suzukii, independently of the presence of TE in intergenic regions. This work also suggests that the genetic background and probably trans regulatory sequences are involved in gene expression and stress response. Further phenotypic and genomic studies on natural populations are needed to better understand the success of invasive species such as D. suzukii.
Funding
Experimental procedures were supported by the ANR (grant SWING ANR to P.G. and C.V. grant ExHyb ANR to C.V.) and the Rovaltain foundation (EpiRip project).
Competing interests
The authors declare that they have no competing interests.
List of supplementary tables
Acknowledgements
We thanks to Vincent Mérel who produced the TEs annotation. P. M. produced data, conceived and wrote the manuscript draft. C. V. & P.G. designed experiments, edited the manuscript. A. J. calibrated experimental design. H. H. helped to q-PCR method and analysis.