Co-opted genes of algal origin protect C. elegans against cyanogenic toxins

Abstract

Amygdalin is a cyanogenic glycoside enriched in the tissues of many edible plants, including seeds of stone fruits such as cherry (Prunus avium), peach (Prunus persica), and apple (Malus domestica). These plants biosynthesize amygdalin in defense against herbivore animals, as amygdalin generates poisonous cyanide upon plant tissue destruction.^1,2,3,4 Poisonous to many animals, amygdalin-derived cyanide is detoxified by potent enzymes commonly found in bacteria and plants but not most animals.⁵ Here we show that the nematode C. elegans can detoxify amygdalin by a genetic pathway comprising cysl-1, egl-9, hif-1, and cysl-2. A screen of a natural product library for hypoxia-independent regulators of HIF-1 identifies amygdalin as a potent activator of cysl-2, a HIF-1 transcriptional target that encodes a cyanide detoxification enzyme in C. elegans. As a cysl-2 paralog similarly essential for amygdalin resistance, cysl-1 encodes a protein homologous to cysteine biosynthetic enzymes in bacteria and plants but functionally co-opted in C. elegans. We identify exclusively HIF-activating egl-9 mutations in a cysl-1 suppressor screen and show that cysl-1 confers amygdalin resistance by regulating HIF-1-dependent cysl-2 transcription to protect against amygdalin toxicity. Phylogenetic analysis indicates that cysl-1 and cysl-2 were likely acquired from green algae through horizontal gene transfer (HGT) and functionally co-opted in protection against amygdalin. Since acquisition, these two genes evolved division of labor in a cellular circuit to detect and detoxify cyanide. Thus, algae-to-nematode HGT and subsequent gene function co-option events may facilitate host survival and adaptation to adverse environmental stresses and biogenic toxins.

Keywords: C. elegans; EGL-9; HIF-1; amygdalin; cyanide detoxifcation; cysl-1; cysl-2; gene co-option; green algae; horizontal gene transfer.

Figure 1.. A compound screen of natural product library identifies amygdalin as a potent activator of cysl-2p::GFP.

(A), Schematic of compound screens that led to identification of amygdalin. (B), Representative bright field and epifluorescence images showing dose-dependent up-regulation of cysl-2p::GFP by amygdalin. Scale bar: 50 μm. (C), Quantification of percentages of animals with cysl-2p::GFP expression after 48 h of amygdalin treatment with increasing doses. Values are means ± S.D with ***P < 0.001 (one-way ANOVA with post-hoc Tukey HSD, N = 5 independent experiments, n > 50 animals per experiment). (D), Quantification of percentages of animals with cysl-2p::GFP expression after increasing durations of amygdalin treatment at the fixed dose of 2 mg/mL. Values are means ± S.D with ***P < 0.001 (one-way ANOVA with post-hoc Tukey HSD, N = 5 independent experiments, n > 50 animals per experiment). (E), Volcano plot showing significantly (adjusted p value < 0.05) up- (green) or down- (blue) regulated genes in amygdalin-treated animals, with cysl-2 noted (arrow). A list of all genes with expression values and statistics are in Data S1.

Figure 2.. Amygdalin activates cysl-2p::GFP through cyanide, CYSL-1, EGL-9 and HIF-1.

(A), Representative bright field and epifluorescence images showing up-regulation of cysl-2p::GFP by amygdalin (1 mg/mL) in wild-type but not hif-1(ia4), cysl-1(ok762) loss-of-function mutant animals. Scale bar: 50 μm. (B), Schematic illustrating the metabolic pathway of amygdalin, leading to generation of prunasin, glucose, hydrogen cyanide and benzaldehyde. (C), Quantification of percentages of animals with cysl-2p::GFP expression after 48 h of treatment with amygdalin (2 mg/mL), prunacin (2 mg/mL), glucose (2 mg/mL), potassium cyanide (0.2 mg/mL) and benzaldehyde (1 mg/mL). Values are means ± S.D. (D), Schematic showing cysl-1 suppressor screens resulting in 3 mutants, all of which are allelic to egl-9 based on whole-genome sequencing and complementation tests. (E), Table summary of cysl-2p::GFP activation in animals with indicated genotypes and treatment conditions. Penetrance and numbers of animals examined are noted. (F), Model illustrating the dis-inhibitory regulatory pathway that mediates the transcriptional response to amygdalin.

Figure 3.. Amygdalin resistance requires CYSL-1, HIF-1 and CYSL-2 acting in a regulatory pathway.

(A), Quantification of percentages of post-L4 stage wild-type animals survived after 24, 48 or 72 h of treatment with amygdalin (10 mg/mL). (B), Quantification of percentages of post-L4 stage cysl-1(ok762) loss-of-function mutant animals survived after 24, 48 or 72 h of treatment with amygdalin (10 mg/mL). (C), Quantification of percentages of post-L4 stage cysl-2(ok3516) loss-of-function mutant animals survived after 24, 48 or 72 h of treatment with amygdalin (10 mg/mL). (D), Quantification of percentages of post-L4 stage cysl-1(ok762); egl-9(sa307) double loss-of-function mutant animals survived after 24, 48 or 72 h of treatment with amygdalin (10 mg/mL). (E), Quantification of percentages of post-L4 stage hif-1(ia04); egl-9(sa307) double loss-of-function mutant animals survived after 24, 48 or 72 h of treatment with amygdalin (10 mg/mL). (F), Quantification of percentages of post-L4 stage cysl-2(ok3516); egl-9(sa307) double loss-of-function mutant animals survived after 24, 48 or 72 h of treatment with amygdalin (10 mg/mL). (G), Quantification of percentages of post-L4 stage cysl-2(ok3516); cysl-1(ok762) double loss-of-function mutant animals survived after 24, 48 or 72 h of treatment with amygdalin (10 mg/mL). (H), Quantification of percentages of post-L4 stage hif-1(ia04); cysl-1(ok762) double loss-of-function mutant animals survived after 24, 48 or 72 h of treatment with amygdalin (10 mg/mL). (I), Quantification of percentages of post-L4 stage rhy-1(n5500) loss-of-function mutant animals survived after 24, 48 or 72 h of treatment with amygdalin (10 mg/mL). All strains carry nIs470 [cysl-2p::GFP + myo-2p::mCherry]. Values are means ± S.D. with ***P < 0.001 (one-way ANOVA with post-hoc Tukey HSD, N = 5 independent experiments, n > 50 animals per experiment). n.s., non-significant.

Figure 4.. Nematode cysl genes were likely acquired from green algae by horizontal gene transfer.

(A), IQ-TREE generated maximum likelihood (ML) phylogenetic tree of homologs of nematode CYSL proteins across a broad range of species. Species groups are shaded. Clades of similar proteins are collapsed and shown as triangles. Bootstrap support values for relevant branches are shown. Bolded values indicate the support for a clade that only contains CYSL-like proteins from nematodes (100% bootstrap support) and a well-supported clade that contains only CYSL-like proteins from nematodes and a group of green algae in the Chlorophyta lineage (93% bootstrap support). Within the nematodes, two clear, well-supported clades of CYSL proteins exist, one that contains C. elegans CYSL-1 and one that contains C. elegans CYSL-2. (B), ML phylogenetic analyses were performed with multiple programs and multiple amino acid substitution models as indicated. Support values (bootstrap and SH-aLRT when available) are shown for a clade containing only nematode proteins and a clade containing nematode and chlorophyte green algae proteins as are bolded in part A. Asterisks next to model names indicate that best fitting model as described in Methods. (C), Expanded phylogenetic tree of CYSL proteins from nematodes and their nearest homologs in green algae. Font colors denote major groups of nematode species Bootstrap support values are shown at relevant nodes. For panels A-C, a complete list of all sequences used is in Data S2 and all trees, including support values, are found in Data S3. CYSL homologs in nematode species from the WGS (sequences from all genome sequencing projects) and TSA (sequences from transcriptome sequencing projects) databases using tBLASTn searches except those shown in Figure 4C are listed in Data S4.