Intra-species variation of ovipositor morphology and oviposition preference in Drosophila suzukii (Diptera: Drosophilidae)

Females of Drosophila suzukii have an enlarged and sclerotized ovipositor (oviscapt, hypogynium), which is a key evolutionary trait that enabled D. suzukii to lay eggs through the skin of ripening fruits. We investigated the intra-species variation of D. suzukii in ovipositor morphology and oviposition preference for substrate hardness among 27 strains recently established from local populations in Japan. A significant difference between strains was observed in the length of the ovipositor as well as in the number of bristles on it. Further analysis using representative strains revealed that the ovipositor size variation among strains was independent of the body size variation, suggesting that the proportion between the ovipositor and other body parts was different among the strains. In the two-choice oviposition assay using 2% and 7% agar substrates, four strains preferred the soft substrates, and one strain preferred the hard substrates. Competitive cross-modal assay between substrate hardness and surface curvature revealed that the preference for curvature overrode the preference for hardness in all the tested strains.


Introduction
In arthropods, the size of genitalia is known to be less sensitive to the nutritional condition than other body parts, resulting in low allometric slope-even individuals with relatively large or small body sizes have genitalia of "standard" size (Eberhard 2009;Emlen et al. 2012).A proposed explanation for this trend is called the "one-size-fits-all" hypothesis, where an individual with a standard size of genitalia is expected to be mechanically compatible with the largest number of mates that are also assumed to have a variable size of genitalia (Eberhard et al. 1998).Although this hypothesis is about the plastic response of body size to the nutritional condition, the same stabilizing mechanism may apply to the genetic variation within a species as well-sexual selection should stabilize the genitalia size reducing the variation among strains compared to the other body parts.
Drosophila suzukii (Matsumura) is an invasive pest originating from eastern Asia, expanding its range worldwide (Little et al. 2020;Tait et al. 2021).One of the key evolutionary traits that make D. suzukii an agricultural pest is its ovipositor morphology.Unlike other Drosophila species, D. suzukii has a large, sclerotized ovipositor with thickened bristles on it, being capable of penetrating the skin of fresh fruits on trees (Hamby et al. 2016;Walsh et al. 2011).In this sense, the evolution of the ovipositor morphology is ascribed to natural selection favoring D. suzukii to exploit a new niche (Atallah et al. 2014).If a larger ovipositor is more adaptive to utilize a wider range of fruits, such as those with thicker and harder skin, there may be intra-species variation in the ovipositor morphology among populations of D. suzukii depending on the local availability of fruit species.
At the same time, the ovipositor is a part of female genitalia, and the reproduction of D. suzukii suffered from the ovipositor evolution-the mating posture of D. suzukii diverged from the sibling species, illustrating the interaction between sexual and natural selection on the ovipositor morphology (Muto et al. 2018).Although D. suzukii has acquired altered genital coupling mechanics circumventing the interference of the enlarged ovipositor, an extremely larger ovipositor may still be maladaptive destabilizing the copulation.In this case, sexual selection should suppress the further evolution of the ovipositor size, and the observed size variation of ovipositor among the populations is expected to be smaller than that of the other body parts as in the allometric relationship observed among individuals with different nutritional conditions.
In this study, we explored the variation in the ovipositor morphology and oviposition behavior among populations of D. suzukii in its native range.Using 27 strains recently established from individuals collected from various parts of Japan, we measured the length of the ovipositor and the number of bristles on it.We also tested the oviposition preference of these strains for the hardness of substrates.

D. suzukii strains
The 27 D. suzukii strains used in this study have been described in our previous study (Akutsu and Matsuo 2022).They originated from wild populations collected at various locations in Japan during 2019-2021 (Table S1).
The strains have been maintained on a standard Drosophila medium made of corn meal, glucose, and dry yeast at 20 °C with the 16L:8D light cycle.In this condition, one generation took 3 weeks (17 generations/year).The individuals used in the size measurement and the oviposition assay were reared at 25 °C from the larval stage with the same light cycle.

Body size measurement
Ovipositor morphology was measured for each of the 27 strains.Ovipositor was dissected from female adults and immersed in 10% KOH solution for 24 h before being mounted on a slide glass using Hoyer's mountant.Pictures of specimens were taken under a phase contrast microscope using a digital camera (FLEXACAM C1, Leica Microsystems, Wetzlar, Germany), and the length of the ovipositor was measured using ImageJ v1.53 (Fig. 1; Schneider et al. 2012).The number of all bristles on one side of the ovipositor was counted as well.For each strain, 20 specimens were analyzed.
For the selected four strains, the measurement of the head width and the thorax length, as well as the ovipositor morphology, was carried out using the individuals independent from those used for the measurement described above.Pictures of the head and the thorax were taken using a digital camera (DP21, Olympus, Tokyo, Japan) mounted on a dissection microscope (SZ61, Olympus), and the head width and the thorax length were measured using ImageJ (Fig. 1).For each strain, 20 females were analyzed.

Two-choice oviposition assay for the substrate hardness
Oviposition substrate was made of agar solution at two concentrations (2 and 7%; Seakem® LE Agarose, Lonza, Basel, Switzerland) cast in silicone molds for UV resincrafting hobbies, which have hemispherical holes of various sizes.In this study, we used holes with two sizes (3.8 and 9.6 mm radius) to prepare oviposition substrates with a part of a spherical surface, which have been shown to promote oviposition of D. suzukii and other Drosophila species (Akutsu and Matsuo 2022;Akutsu and Matsuo 2023).The volume of each substrate was fixed to 150 µl regardless of its radius.
Ten females at the age of 7-9 days after eclosion were introduced into a Petri dish (9 cm diameter), in which 10 oviposition substrates were placed on a wet cotton pad (3 × 6 cm).For testing the preference for substrate hardness, 5 soft (2%) and 5 hard (7%) substrates with a 3.8 mm radius were used.For testing the cross-modal priority between the hardness and curvature, the radius of the substrates with the preferred hardness (depended on strain; see Results) was changed to 9.6 mm, allowing the females to choose between the preferred hardness and the higher curvature (3.8 mm, preferred to 9.6 mm; Akutsu and Matsuo 2022).Because there were two types of strains (preferred harder or softer substrates), there were two types of analyses.
The assay started 6 h before the transition to the dark phase.After 24 h of oviposition at 25 °C, the number of eggs laid on each substrate was counted.Each fly was used only once.Ten replications were made for each strain.When the total number of eggs in a dish was less than 10, the corresponding data were excluded from the analysis, and additional replications were made.The preference index was calculated as (N hard − N soft )/N hard + N soft ), where N hard and N soft are the number of eggs laid on the hard and soft substrates, respectively.

Statistical analysis
The difference in the morphological traits among the strains was examined using the Kruskal-Wallis rank sum test.The results of oviposition preference were tested using the Wilcoxon signed rank test for the deviation from a null hypothesis (no preference for substrate hardness) in each strain.The relationship between body size and ovipositor morphology was examined using the standardized major axis (SMA) provided by the SMATR package for R (Taskinen and Warton 2013) including 'strain' as a group parameter.All the analyses were conducted using R version 4.2.0 (R core team 2022).

Effect of body size on ovipositor morphology
The observed variation in ovipositor morphology may reflect the body size variation among strains through allometry.To test this possibility, the influence of body size on ovipositor morphology was examined in four strains that showed large deviation from the overall mean value in the ovipositor length or the bristle number (Fig. 3).Thorax length was significantly correlated with head width in two strains (Fig. 3A: lines #14 and #24).The regression lines were not significantly different among strains (p = 0.225), showing that the strains shared the same allometric relationship between the two body parts.In contrast, the regression lines for ovipositor length were significantly different among the strains (p < 0.001), suggesting that the allometric relationship between the sizes of these two parts was variable among the strains (Fig. 3B).Bristle number was different among the strains (p < 0.001), but the correlation with head  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 A B C Fig. 2 Variation in ovipositor morphology and oviposition preference among the strains.A Ovipositor length.The horizontal line represents the mean of the whole dataset.Shaded boxes indicate the strains whose range (including outliers) did not overlap with it.B The number of bristles on an ovipositor.The horizontal line represents the mean of the whole dataset.Shaded boxes indicate the strains whose range (including outliers) did not overlap with it.C Oviposition preference for the hardness of oviposition substrates.The horizontal line shows that the level of the preference index at zero.The preference index was calculated as (N hard − N soft )/(N hard + N soft ), where N hard and N soft are the number of eggs laid on the hard and soft substrates, respectively.The shaded boxes indicate the strains for which the preference index was significantly different from zero at a p < 0.01 level with the Wilcoxon signed-rank test width was not significant in all the strains, showing that the bristle number variation among strains was independent of the body size variation (Fig. 3C).These results proved that the variation in ovipositor morphology among strains was accompanied by changes in the proportion between the size of ovipositor and other body parts.

Priority between two oviposition cues: hardness and curvature
The five strains that showed specific oviposition preference for substrate hardness allowed us to examine the interaction between multiple oviposition cues based on different sensory modalities.We previously found that D. suzukii and other Drosophila species prefer oviposition substrates with smaller radii (higher curvature) to larger radii (Akutsu and Matsuo 2022;Akutsu and Matsuo 2023).In this study, we tested the cross-modal priority between hardness and curvature as oviposition cues.The result showed that the preference for higher curvature overrode the preference for hardness in all the tested strains (Fig. 4).

Discussion
In this study, we observed a significant variation in ovipositor size among strains, whereas the variation in body size was not significant (Figs. 2, 3).Our results are consistent with the previous studies comparing three D. suzukii strains originated from the wild populations in France, the USA, and Japan, which showed that the size of ovipositor was significantly different among the three strains (Varón-González et al. 2020), whereas the size of wings was not (Fraimout et al. 2018).These results collectively show that the ovipositor of D. suzukii does not follow the general rule of arthropod genitalia, supporting the view that it is a site of conflict between sexual selection and natural selection.There are two possible explanations for this observation, which are not exclusive to each other.

Natural selection
First, natural selection on the ovipositor size may exist and be variable among local populations.The most likely advantage of having a larger ovipositor is the capacity to penetrate the harder fruit skin (Atallah et al. 2014).It should be noted, however, that the preference for substrate hardness was overridden by the preference for surface curvature in all the tested strains (Fig. 4).Oviposition preference for curved surfaces with a smaller radius has been shown in D. suzukii, as well as in D. takahashii and D. simulans (Akutsu and Matsuo 2022;Akutsu and Matsuo 2023), suggesting that this trait is ancestral and shared by multiple Drosophila species.In contrast, substrate hardness may be relatively less important as a cue of oviposition site selection even in D. suzukii.In such a case, natural selection on the ovipositor size would not be sufficient for explaining the observed variation among strains.

Sexual selection
Second, sexual selection on the ovipositor size may be relaxed in D. suzukii.Evolutionary change in the genital coupling mechanics may have freed the ovipositor for further morphological change, allowing different populations to have different ovipositor morphology.In this regard, it would subpulchrella has an enlarged, serrated ovipositor as well, it shares the "conventional" mating posture with sibling species that have the "normal" ovipositor morphology (Muto et al. 2018).Therefore, the sexual selection on the ovipositor size should be more stringent in D. subpulchrella than in D. suzukii.Consequently, a smaller variation in ovipositor size is expected to be observed in this species.Another useful experiment would be the mating assay between D. suzukii strains with different ovipositor sizes.If sexual selection is relaxed in D. suzukii, inter-strain mating should not be disturbed by larger ovipositor size.Nevertheless, it is possible that the ovipositor size and male genitalia coevolve in each population, maintaining strong sexual selection within each local population.In this case, copulation between the strains with different ovipositor sizes is expected to be disrupted, and a compensating variation in male genital morphology should be observed among strains.
In conclusion, our study revealed that substantial intraspecies variation exists in the ovipositor morphology and oviposition behavior of D. suzukii.Since the ovipositor of D. suzukii is a site of interaction between natural and sexual selection, the intra-species variation provides an exceptional opportunity to analyze the interaction experimentally.
Fig. 4 Competitive cross-modal assay between substrate hardness and surface curvature.A The strains that preferred the softer substrates with the same curvature.B The strain that preferred the harder substrates with the same curvature.The tested combinations of hardness (agar concentration) and curvature (substrate radius) are shown above and below the plots.Note that the results for substrates with the same curvature (the left box in each panel) are the same as those shown in Fig. 2C

Fig. 1
Fig.11Measurements of body parts.A Ovipositor length was measured along the longitudinal axis of the sclerotized part of an ovipositor plate.The number of all the bristles on an ovipositor plate was counted regardless of their thickness, because the distinction between thin and thick bristles was often unclear.B Head width was measured as the widest distance between the compound eye surfaces.C Thorax length was measured as the distance from the socket of the upper humeral bristle to the posterior tip of the scutellum not supported in a prior test.p values indicate the significance against the null hypothesis that the regression lines have the same elevation for all the strains.Dashed regression lines were not supported at a p < 0.05 level for R 2 ≠ 0. A Head width vs thorax length.B Head width vs ovipositor length.C Head width vs bristle number be worth examining if the ovipositor morphology of D. subpulchrella is more constrained than D. suzukii.Although D.
. Asterisks indicate the results of Wilcoxon signed rank test (null hypothesis: no preference; **: p < 0.01, *: p < 0.05).N = 10 for each combination of substrates, and 10 females were used in each replicate