A beta-glucosidase of an insect herbivore determines both toxicity and deterrence of a dandelion defense metabolite

Abstract

Gut enzymes can metabolize plant defense compounds and thereby affect the growth and fitness of insect herbivores. Whether these enzymes also influence feeding preference is largely unknown. We studied the metabolization of taraxinic acid β-D-glucopyranosyl ester (TA-G), a sesquiterpene lactone of the common dandelion (Taraxacum officinale) that deters its major root herbivore, the common cockchafer larva (Melolontha melolontha). We have demonstrated that TA-G is rapidly deglucosylated and conjugated to glutathione in the insect gut. A broad-spectrum M. melolontha β-glucosidase, Mm_bGlc17, is sufficient and necessary for TA-G deglucosylation. Using cross-species RNA interference, we have shown that Mm_bGlc17 reduces TA-G toxicity. Furthermore, Mm_bGlc17 is required for the preference of M. melolontha larvae for TA-G-deficient plants. Thus, herbivore metabolism modulates both the toxicity and deterrence of a plant defense compound. Our work illustrates the multifaceted roles of insect digestive enzymes as mediators of plant-herbivore interactions.

Keywords: Melolontha melolontha; Taraxacum officinale; ecology; plant defense; root herbivore; sesquiterpene lactone; β-glucosidase.

Figure 1.. Taraxinic acid β-D-glucopyranosyl ester is rapidly deglucosylated and conjugated to glutathione upon ingestion by Melolontha melolontha.

(A) Relative abundance of taraxinic acid β-D-glucopyranosyl ester (TA-G) and its aglycone taraxinic acid (TA) in diets enriched with Taraxacum officinale latex and in Melolontha melolontha larval gut, hemolymph, and frass after feeding on latex-containing and control diets. Ant = anterior; post = posterior. N = 5. For relative quantification of TA-glutathione (TA-GSH) conjugates in M. melolontha, refer to Figure 1—figure supplement 1. (B) High-pressure liquid chromatography-mass spectrometry (HPLC-MS) full scan (positive mode) of the anterior midgut of M. melolontha larvae fed with latex-containing and control diets. (C) Schematic illustration of proposed TA-G metabolism in M. melolontha. For nuclear magnetic resonance (NMR) analysis of TA-GSH conjugates, refer to Figure 1—figure supplements 2–6. Distribution of the total deglucosylated and conjugated metabolites of TA-G in M. melolontha larvae that consumed 100 µg TA-G within 24 hr. The size of the circles is relative to the total amount of conjugates. Values denote the percentage of metabolites found in each body part and are the mean of eight replicates. For long-term distribution of TA-Cys in M. melolontha, refer to Figure 1—figure supplement 7. (E) Relative proportions of TA-G metabolites in quantities from panel (D). Values denote the mean of eight replicates. Raw data are available in Figure 1—source data 1.

Figure 1—figure supplement 1.. Relative quantification of TA-glutathione conjugates in Melolontha melolontha larvae feeding on diets with and without Taraxacum officinale latex.

N = 5. TA = taraxinic acid; GSH = glutathione; Cys = cysteine; Gly = glycine; Glu = glutamate.

Figure 1—figure supplement 2.. 500 MHz ¹H-¹³C HSQC NMR spectrum of the partially purified Melolontha melolontha midgut extract.

Intensity level was adjusted to suppress noise. Gray rectangles mask impurities for clarity. Resonances above noise level are highlighted by pink circles. Missing resonances in the noise level are not considered in the picture. Sample was measured in MeOH-d₄.

Figure 1—figure supplement 3.. Structure of synthesized TA-G-GSH with chemical shifts (500 MHz, in MeOD-d₄).

Figure 1—figure supplement 4.. Structure of synthesized TA-G-Cys with chemical shifts (500 MHz, in MeOD-d₄).

Figure 1—figure supplement 5.. Structure of synthesized TA-GSH with chemical shifts (500 MHz, in MeOD-d₄).

Figure 1—figure supplement 6.. Structure of synthesized TA-Cys with chemical shifts (500 MHz, in MeOD-d₄).

Figure 1—figure supplement 7.. Accumulation of TA-Cys in different body parts of Melolontha melolontha upon feeding for 1 month on Taraxacum officinale plants.

The mean of four to six replicates is displayed.

Figure 2.. Insect rather than plant enzymes deglucosylate TA-G.

(A) Left and right panels: plant-mediated enzymatic deglucosylation of TA-G at pH 3–8. Taraxacum officinale latex was collected from wounded roots and incubated in buffers adjusted to different pH values. N = 3. (B) Deglucosylation activity of untreated and boiled extracts of Melolontha melolontha gut content and gut tissue incubated at pH 8.0 with boiled latex extracts. The p-values of a two-way analysis of variance (ANOVA) are shown. N = 6. Error bars = SEM. The intestinal pH of M. melolontha is shown for comparative purposes (data from Egert et al., 2005). For in vitro M. melolontha glucosidase inhibition assays, refer to Figure 2—figure supplement 1 - 2. (C) Relative abundance of TA-G and its metabolites in the diet and regurgitant of larvae fed with carrot slices coated with either intact (+) or heat-deactivated (-) T. officinale latex. Heat deactivation of latex did not significantly affect the deglucosylation of TA-G in M. melolontha. p-values refer to two-way ANOVAs. N = 4. TA-G: taraxinic acid β-D-glucopyranosyl ester; TA = taraxinic acid; GSH = glutathione; Cys = cysteine. Peak area was normalized across all treatments based on the maximal value of each metabolite. Raw data are available in Figure 2—source data 1.

Figure 2—figure supplement 1.. TA-G deglucosylation activity (TA/(TA + TA-G)) of Melolontha melolontha anterior midgut samples in the presence of either the carboxylesterase inhibitor bis(p-nitrophenyl)phosphate or the α- and β-glucosidase inhibitor castanospermin.

Only the glucosidase inhibitor reduced deglucosylation of taraxinic acid β-D-glucopyranosyl ester (TA-G). N = 6. Error bars denote SEM.

Figure 2—figure supplement 2.. TA-G deglucosylation activity (TA/(TA + TA G)) of Melolontha melolontha anterior midgut samples in the presence of acarbose, an α-glucosidase-specific inhibitor, or castanospermine, an α- and β-glucosidase inhibitor.

Only castanospermine reduced deglucosylation of TA-G. TA-G = taraxinic acid β-D-glucopyranosyl ester; TA = taraxinic acid. N = 3. Standard errors denote SEM.

Figure 3.. Melolontha melolontha midgut β-glucosidases hydrolyze TA-G and other plant defensive glycosides.

(A) Phylogeny of newly identified Melolontha melolontha β-glucosidases and previously reported β-glucosidases of Tenebrio molitor (Tm bGlc, AF312017.1) and Chrysomela populi (Cp bGlc, KP068701.1), and myrosinases (thioglucosidases) of Phyllotreta striolata (Ps myrosinase, KF377833.1) and Brevicoryne brassicae (Bb myrosinase, AF203780.1) based on amino acid similarities using maximum likelihood method. Bootstrap values (N = 1000) are shown next to each node. Amino acid sequence alignments of the β-glucosidases are shown in Figure 3—figure supplement 1. (B) Heat map of average (n = 3) gene expression levels of M. melolontha β-glucosidases in the anterior and posterior midgut of larvae feeding on diets supplemented with water, taraxinic acid β-D-glucopyranosyl ester (TA-G), or Taraxacum officinale latex-containing diet. FPKM = fragments per kilobase of transcript per million mapped reads. (C) Activity of heterologously expressed M. melolontha β-glucosidases with TA-G, a mixture of maize benzoxazinoids, the salicinoid salicin, 4-methylsulfinylbutyl glucosinolate (4-MSOB), cellobiose, and the fluorogenic substrate 4-methylumbelliferyl-β-D-glucopyranoside (Glc-MU). Glucosidase activities of three consecutive assays with excreted proteins from insect High Five cells were measured. Negative controls (buffer, non-transfected wild-type cells, and cells transfected with green fluorescent protein) did not hydrolyze any defense metabolite. Results from the individual assays are shown in Figure 3—figure supplement 2. For deglycosylation of these compounds by M. melolontha gut protein crude extracts, refer to Figure 3—figure supplement 3. Deglycosylation assays with recombinant Mm_bGlc17 yielded highest aglycone formation; Figure 3—figure supplement 4. Raw data are available in Figure 3—source data 1.

Figure 3—figure supplement 1.. Amino acid sequence alignment of β-glucosidases of Melolontha melolontha and other insect species.

The catalytic glutamates of the glucosidase-characteristic NEP and ITENG motifs are indicated with red triangles.

Figure 3—figure supplement 2.. Activity of heterologously expressed Melolontha melolontha β-glucosidases and negative controls (GFP = green fluorescent protein; WT = non-transfected wild type; buffer) toward plant defensive glycosides, cellobiose, and the standard substrate 4-methylumbelliferyl-β-D-glucopyranoside (Glc-MU), which fluoresces upon deglucosylation.

The glucosidase activity of three deglucosylation assays was categorized according to relative activity. Both total protein levels and catalytic activity may contribute to the overall activity. Na = not available.

Figure 3—figure supplement 3.. Deglucosylation of defensive glycosides by boiled and non-boiled Melolontha melolontha anterior midgut extracts in vitro.

Aglycone formation was normalized to the maximal peak area of all samples. p-values of two-way analyses of variance (ANOVAs) are shown for each metabolite separately. N = 10. Error bars denote SEM. TA-G = taraxinic acid β-D-glucopyranosyl ester; BXDs = benzoxazinoids; 4-MSOB = 4-methylsulfinylbutyl glucosinolate.

Figure 3—figure supplement 4.. Taraxinic acid aglycone formation of the heterologously expressed Melolontha melolontha β-glucosidases and negative controls (GFP = green fluorescent protein; WT = non-transfected wild type; buffer) of two deglucosylation assays.

Both total protein levels and catalytic activity may contribute to the aglycone formation. Na = not available.

Figure 4.. Silencing of Mm_bGlc17 reduces TA-G deglucosylation and modifies the impact of TA-G on larval growth and host plant choice.

(A) Taraxinic acid β-D-glucopyranosyl ester (TA-G) deglucosylation activity (TA/(TA + TA G)) of gut extracts from Melolontha melolontha larvae in which different β-glucosidases were silenced through RNA interference (RNAi), resulting in stable and specific silencing of the individual glucosidases, Figure 4—figure supplements 2–3. Silencing of Mm_bGlc17 significantly reduced hydrolysis of TA-G by gut extracts. A green fluorescent protein-derived double-stranded RNA (GFP dsRNA) construct was used as a negative control. Mm_bGlc16&17-treated larvae received a 50:50 (v/v) mixture of both dsRNA species. Deglucosylation activity was normalized to that of boiled control samples to correct for the background of non-enzymatic hydrolysis. N = 9–10. p-value of a one-way analysis of variance (ANOVA) is shown. Different letters indicate a significant difference according to Tukey’s honest significance test. Error bars = SEM. (B) Weight gain of Mm_bGlc17-silenced and GFP-control M. melolontha larvae growing on transgenic TA-G-deficient or control Taraxacum officinale lines. N = 11–15. p-values refer to a two-way ANOVA and Student’s t-tests. Error bars = SEM. For comparing growth of GFP- and Mm_bGlc17-silenced larvae between TA-G-deficient and control lines, refer to Figure 4—figure supplement 4. The experiment was repeated once with similar results; Figure 4—figure supplement 5. (C) Gene expression (relative to actin) of Mm_bGlc17-silenced and GFP-control M. melolontha larvae feeding on transgenic TA-G-deficient or control T. officinale lines. N = 12–14. p-values refer to a two-way ANOVA (log-transformed data) and Kruskal-Wallis rank sum tests (non-transformed values). (D) Choice of Mm_bGlc17-silenced and GFP-control larvae between transgenic TA-G-deficient and control T. officinale lines. Silencing of Mm_bGlc17 abolished the choice of control larvae for TA-G-deficient lines. p-values refer to binomial tests. Choice was stable over time; see Figure 4—figure supplement 6. Raw data are available in Figure 4—source data 1.

Figure 4—figure supplement 1.. Injection of dsRNA by a sterile syringe between the second and third segment of Melolontha melolontha.

Figure 4—figure supplement 2.. Tubulin mRNA expression in Melolontha melolontha larvae treated with 2.5 µg or 0.25 µg GFP or tubulin dsRNA per g larval mass 2, 5, and 10 days after injection (N = 3).

Error bars denote SEM. Asterisks indicate significant differences in relative tubulin expression between green fluorescent protein (GFP) and tubulin double-stranded RNA (dsRNA)-treated larvae according to the two-tailed Student’s t-test (*p<0.5; **p<0.01; ***p<0.001).

Figure 4—figure supplement 3.. Gene expression (relative to actin) of Mm_bGlc16, Mm_bGlc17, and Mm_bGlc18 in Mm_bGlc17 dsRNA-injected larvae 2 days after dsRNA application.

Asterisks indicate significant differences in the relative gene expression between non-injected and Mm_bGlc17-injected, as well as between green fluorescent protein (GFP)-injected and Mm_bGlc17-injected larvae according to two-tailed Student’s t-tests (*p<0.5). Error bars denote SEM. N = 6–8.

Figure 4—figure supplement 4.. Weight gain of Mm_bGlc17-silenced and GFP-control Melolontha melolontha larvae growing on transgenic TA-G-deficient or control Taraxacum officinale lines (N = 11–15).

p-values refer to Student’s t-tests. Error bars = SEM.

Figure 4—figure supplement 5.. Repetition of the experiment on weight gain of Mm_bGlc17-silenced and GFP-control Melolontha melolontha larvae growing on transgenic TA-G-deficient or control Taraxacum officinale lines.

(A) within the first 3 weeks, (B) 2-8 weeks, (C) 8 weeks (N=12-20) p-values refer to two-way analyses of variance (ANOVAs) and Student’s t-tests. Error bars = SEM.

Figure 4—figure supplement 6.. Choice of Melolontha melolontha larvae that were treated with either GFP (A) or Mm_bGlc17 (B) dsRNA between TA-G-deficient transgenic and wild-type Taraxacum officinale plants 1–4 hr after the start of the experiment.

p-values refer to binomial tests. Larvae that were not chosen were excluded from the analysis. GFP = green fluorescent protein; TA-G = taraxinic acid β-D-glucopyranosyl ester; dsRNA = double-stranded RNA.