Shaping the environment - Drosophila suzukii larvae construct their own niche

Abstract

In holometabolous insects, the choice of oviposition substrate by the adult needs to be coordinated with the developmental needs of the larva. Drosophila suzukii female flies possess an enlarged serrated ovipositor, which has enabled them to conquer the ripening fruit as an oviposition niche. They insert their eggs through the skin of priced small fruits. However, this specialization seems to clash with the nutritional needs for larval development since ripening fruits have a low protein content and are high in sugars. In this work, we studied how D. suzukii larvae develop in and interact with the blueberry. We show that despite its hardness and composition, D. suzukii's first instar larvae are able to use the ripening fruit by engaging in niche construction. They display unique physical and behavioral characteristics that allow them to process the hard-ripening fruit and provoke an improvement in its composition that better suits larval nutritional needs.

Keywords: Biological sciences; Entomology; Evolutionary biology.

Graphical abstract

Figure 1

Developmental progression of D. suzukii larvae in blueberries and their physical adaptations (A) The procedure developed to study larvae in blueberries. Top, left: Incision into the skin of a blueberry to expose the flesh and place five first instar larvae inside. Bottom, left: Schematics of the procedure. 1. An incision was made to generate and open a flap. 2. Larvae were placed on the exposed fruit tissue. 3. The skin flap was closed. Right: Blueberries in tissue sample bags hanging inside an incubator and the environmental parameters used. (B) Average proportion of five initially placed D. suzukii (top) or D. melanogaster (bottom), larvae in each developmental stage, found in each fruit at one-day intervals after introduction into fruits (days after infestation). n indicates number of fruits for each treatment. Bars show the means, error bars the SEM. Survival of Drosophila larvae of the two species was compared using a binomial generalized linear model (glm), F = 53.059, p < 0.001 (∗∗∗). (C) Description of D. suzukii and D. melanogaster first instar larvae mouth hooks. Top panels: Lateral (left column) and bottom view (mid column) of the mouth hooks of D. melanogaster (top row) and D. suzukii (bottom row). On the right, enlarged isolated tracings of the structures are shown for a clearer comparison (Scale bar: 20 μm). Bottom graphs: Bar graphs show the mean length of the first tooth from the top of the mouth hook (H1), the length from the top of the mouth hook hinge to the bump posterior to the first tooth (H2), length of the H piece (P1), and vertical length from the top of the dorsal posterior process to the joining of the bottom posterior and anterior processes, passing through the end of the H piece (P2). Bars indicate the mean and error bars the SEM. n indicates the number of mouth hooks analyzed. The size of mouth hook parameters between the two Drosophila species was compared with ANOVAs, H1: F = 427.80, H2: F = 1938.00, P1: F = 165.00, P2: F = 32.09, p for all comparison <0.001 (∗∗∗). (D and E) Average proportion of the initially placed larvae of D. suzukii or D. melanogaster in each developmental stage, found in each fruit at one-day intervals after introduction into fruits (days after infestation), (D) D. melanogaster in softened blueberries and (E) Co-infestation experiments. Fruits and larvae were scored independently and discarded at each time point. Bars represent means, error bars SEMs. n indicates the number of larvae containing fruits. (F) Average (mean) cumulative proportion of the initially placed larvae of D. melanogaster that pupate at one-day intervals after introduction into vials with infested blueberry tissue from 4 days D. suzukii infested blueberries. The same vials were scored each day for the presence of pupae. Error bars: SEM. n represents the number of vials scored for pupae.

Figure 2

D. suzukii larval activity modifies the composition of the fruit, elevates its protein content, and increases its physiologically perceived nutritional value (A) Qualitative changes in blueberries at 1-day intervals after the introduction of D. suzukii larvae (top row) or D. melanogaster larvae (bottom row). The red arrowhead marks the introduction point. (B) Percentage of protein in sample tissue from single fruits either displaying no signs of larval activity (virgin fruit tissue – green bars) or showing signs of larval activity (infested tissue – purple bars) at one-day intervals after placement of first instar larvae in fruits (days after infestation). The protein content of fruits was analyzed using generalized least squares models. Common uppercase letters on top of the graph indicate non-significant differences between treatments. Bars represent the mean, and error bars the SEM, and n the number of investigated fruits. (C) Progression of the average larval weight (mg) at 1-day intervals during development in blueberries from recently hatched D. suzukii first instar larvae. Blue circles represent the mean, error bars the SEM, n the number of investigated samples of larvae. The beige curve represents the 3-parameter sigmoidal growth curve with the equation shown in the graph. All three parameters of the equation are statistically significant (a: t = 14.080, b: t = 26.049, c: t = 5.912, p for all parameters <0.001). (D) Volatile compounds emitted by individual blueberry samples on different days after sham-treated (left), D. melanogaster-infested (center), and D. suzukii-infested (right) blueberries. Bars: Presence of a chemical in a sample. Red bars: Chemicals of interest exclusively present in D. suzukii-infested samples. (E) Schematics of expected dILP2 immunofluorescence signal dynamics in IPCs of larvae feeding on poor or rich protein media (left). Normalized immunofluorescence signals of anti-dILP2 in IPCs of larvae feeding on virgin blueberries (VB - green bar) or 4-day D. suzukii-infested blueberries (IB - purple bar) (right) Bars represent the means, error bars the SEM, and n the number of larvae investigated. two-way ANOVA with tissue type and Drosophila species as explanatory variables (Tissue type: F = 17.405, p < 0.001 (∗∗∗), Drosophila species: F = 0.077, p = 0.782, Tissue type: Drosophila species: F = 1.074, p = 0.302).

Figure 3

D. suzukii larval behavior is conducive to niche construction (A) Example of annotated tracks of a single digging first instar larva after 75 min. Recently hatched larvae were placed in a hole (black circle) in the gel matrix (1% agarose, 10% blueberry juice); yellow: reused tracks; magenta: backward movement; blue: turns; cyan: branching; green: exploration events; orange: returns to the starting point. Gray marks the area considered to have been influenced by larval activity (0.25mm to each side of the track.). (B) Frequency of trip lengths in intervals of 1 mm. Bars represent the mean, error bars the SEM. n represents individual larvae digging trials. (C) Number of trips per trial. White line in the violin plot represents the median, white broken line the 25^th and 75^th percentiles, black dots the individual values, black line the mean, and error bars the SEM. n represents individual larvae digging trials. (D) Proportion of the total area of the arena that was under the influence of the larva after 75 min in each trial. n represents individual larvae digging trials. Black horizontal lines represent the mean, and error bars the SEM. Boxplots in gray show the 25^th and 75^th percentiles, whiskers extend to the 10^th and 90^th percentiles, white line represents the median, and black dots individual data points. Linear regression: regression line (black) with 95% CI (gray). (E) Average length of tunnels from 10 larvae tested individually (n) after 75 min trials, black horizontal lines represent the means, and error bars the SEM. (F) Average number of each behavioral feature occurring per trip. n represents individual larvae digging trials. White line in the violin plots represents the median, white broken line the 25^th and 75^th percentiles, black dots the individual values, black line the mean, and error bars the SEM. (G) Progression of the complexity of the network of tunnels as quantified by the sum of the number of paths, chambers, and openings to the start point through the 75 min trials. n represents individual larvae digging trials. Black horizontal lines represent the mean, and error bars the SEM. Boxplots in gray show the 25^th and 75^th percentiles, whiskers extend to the 10^th and 90^th percentiles, white line represents the median, and black dots individual data points. Linear regression: regression line (black) with 95% CI (gray). (H) Open-field choice assay allowing a choice between larval-infested blueberry (IB, purple perimeter) and virgin blueberry (VB, green perimeter). Larval choice was defined as larval presence within the area around the target tissue. Larvae outside both areas were ignored. Larvae started grouped at the central gray circle. (I and J) Preference index of first instar D. suzukii larvae in choice assays at 15 min toward agarose-embedded fruit tissue (I) or at 30 min toward volatile chemicals identified in D. suzukii larval modified blueberries (J). N = nothing, VB = virgin blueberry, IB = infested blueberry, VS = virgin side of infested blueberry, IS = infested side of infested blueberry, MO = Mineral oil, IA = Isoamyl alcohol, AA = Acetic acid, Ac = Acetoin, Mix = mixture of the three chemicals, VB + Mix = virgin blueberry additioned with the mix of the three chemicals. White line in the violin plots represents the median, white broken line the 25^th and 75^th percentiles, black dots the individual values, black line the mean, and error bars the SEM. n represents individual assays with groups of 20 larvae. Individual Student’s t tests against a hypothetical mean = 0. (I) N v N, t = 0.08896. N v VB, t = 3.103. N v IB, t = 3.703. VB v IB, t = 3.552. VS v IS, t = 2.862. (J) VB v IB, t = 2.867. MO v IA, t = 1.702. MO v Ac, t = 0.9689. MO v AA, t = 0.6790. MO v Mix, t = 7.486. VB + Mix v VB, t = 2.850. ns: non statistically significant differences, ∗: statistically significant at p < 0.05, ∗∗∗: statistically significant at p < 0.001. (K) Representative image of a choice digging assay. Traces in green, magenta, and cyan depict the paths taken by the three larvae tested during the assay. Broken lines mark the quadrants defined for quantification. Larvae started the assay on quadrant 0. For the statistical analysis, quadrants 0 and 1 were defined as the Start position, and quadrants 4 and 5 as the End position. Shades of green mark the side of the virgin blueberry tissue, and the green circle, its placement (VB). Shades of purple mark the side of the infested blueberry tissue (IB) and a purple circle its placement. (L) Mean proportion of larvae present in each quadrant at 2 min intervals. Shades of green mark the quadrants on the virgin blueberry side (VB) and purple the quadrants on the infested blueberry side (IB). Black outlines mark the time points analyzed in M. (M) Mean proportion of larvae present at the Start position (quadrants 0–1) at the beginning of the assay (2 min) and at the End position (quadrants 4–5) at the end of the assay (58 min). Binomial generalized linear model (glm) with tissue type and time as explanatory variables different letters indicate statistically significant differences. Explanatory variables: Time (Deviance = −2.209, p = 0.137), Side (Deviance = −0.964, p = 0.326), Time:Side (Deviance = −13.516, p < 0.001). White line in the violin plots represents the median, white broken line the 25^th and 75^th percentiles, black dots the individual values, black line the mean, and error bars the SEM. n represents individual assays with groups of 3 larvae.