Intersecting experimental evolution and CRISPR screens to identify novel toxin resistance loci

Michele Marconcini
Steeve Cruchet
Srishti Goswami
Raghuvir Viswanatha
Matthew Butnaru
Joydeep De
Camilla Roselli
Dafni Hadjieconomou
Norbert Perrimon
Stephanie E Mohr
Richard Benton author has email address

Center for Integrative Genomics, Faculty of Biology and Medicine, University of Lausanne, Lausanne, Switzerland
Department of Genetics, Blavatnik Institute, Harvard Medical School, Boston, United States
Howard Hughes Medical Institute, Boston, United States
Institut du Cerveau-Paris Brain Institute, Sorbonne Université, Inserm, CNRS, Hôpital Pitié-Salpêtrière, Paris, France

https://doi.org/10.7554/eLife.111773.2

Open access
Copyright information

Figures and data

Experimental evolution of D. simulans with increased OA resistance.
A Phylogeny of the drosophilid trio studied in this work. Mya: million years ago. B Schematic of octanoic acid (OA) resistance tube assay (see Materials and Methods) and selection procedure. Created with BioRender.com. C Species differences in resistance of D. simulans 04 and D. sechellia 07 strains to 3 μl OA. D Schematic illustrating the generation of the base populations for Evolve-and-Resequence (see Materials and Methods). Created with BioRender.com. E Changes in survivability to 3 μl OA of the ten D. simulans populations over generations 0-25. Box plots represent the interquartile ranges of the survivability for all the replicates; blue lines shows the generalized additive model (GAM) regression; grey shading represents the 95% confidence interval for the mean responses at each time point. F As in E for the 10 populations combined. G OA resistance for the combined 10 populations with increasing OA volumes during generation 25-31. H As in E, with 9 μl OA during generations 31-50. I As in H for the 10 populations combined. J Comparison of OA resistance of unevolved D. simulans (i.e., the original 10 outcrossed populations maintained on standard fly food; see Materials and Methods), evolved (generation 50 (G50)) D. simulans, and D. sechellia. K-N Phenotypic comparison of unevolved and evolved (G50+) D. simulans populations 1-3 for K noni resistance (G54), L fecundity (G54), M developmental viability (G57), and N lifespan (G57). For K-M, significance was assessed using the unpaired t-test correcting for multiple testing. NS p>0.05, ***p<0.001. For L-N, tests were run on standard media. No significant differences between unevolved and evolved flies were detected using the Cox regression model. See Materials and Methods for details and Data S1 for raw data.

Dynamic directional selection of allele frequencies during experimental evolution.
A Schematic of the analysis. B Identity-by-state analysis. The heatmap depicts the proportion of shared alleles between populations, with values ranging from 0.8 (blue, lowest shared proportion) to 1 (red, identical alleles). C,D Manhattan plot showing the probability of allele frequency changes between C generations 0 and 25, and D generations 0 and 50. Top panel shows the results of the CMH test; orange dots indicate significant variants at a threshold of-log₁₀(p) ≥40, a conservative approximation of the top 0.000005% simulated SNPs under neutrality (Fig. S1B). Bottom panel shows the absolute log-Bayes factors estimated by Bait-ER; orange dots indicate significant variants at the conservative threshold log(0.99/0.01). Raw data in Data S2. E Circular plots illustrating overlap of predicted genomic regions under directional selection between the two statistical approaches at generation 25 (left) and generation 50 (middle), or between generation 25 and 50 (right). Raw data in Data S3.

A genome-wide pooled CRISPR screen for OA susceptibility and resistance genes in D. melanogaster cultured cells.
A Left: images of S2R+ cells cultured in the absence or presence of OA. Scale bar, 50 μm. Right: dose-dependent effects of OA on S2R+ cell viability, through measurement of ATP levels with a CellTiter-Glo assay (see Materials and Methods). B Schematic of CRISPR screen design. C Top 5% “susceptibility” genes (see Materials and Methods and Data S4). Box plots represent the interquartile ranges for all the replicates of the Z-scores obtained through maximum likelihood estimation to test for selection. The inset (bottom-right) displays the gene-ontology analysis of these genes. D As in C, for the top 5% “resistance” genes. E Venn diagram showing the overlap between all the genes identified in the experimental evolution at G25 and G50, and genes associated with susceptibility and resistance in the CRISPR screen. F Subportion of the top panel (CMH test) of the Manhattan plot from Fig. 2C for chromosome arms 2L and 3L, with the red arrows highlighting the regions magnified on the right showing the significant variants and linkage disequilibrium blocks (triangles) for kraken and Alkbh7 (CG14130).

Sequence and expression analyses of kraken and Alkbh7.
A AlphaFold protein models for D. melanogaster Kraken (AF-O18391-F1-model_v4) and Alkbh7 (AF-Q9VTP1-F1-model_v4) (Jumper, et al. 2021; Varadi, et al. 2022), highlighting a conserved putative catalytic serine in the active site in Kraken, and the mitochondrial targeting sequence in Alkbh7 (predicted by MitoFates (Fukasawa, et al. 2015)). See also Fig. S2A,B. B Tissue-specific expression of D. melanogaster kraken and Alkbh7 from bulk RNA-sequencing data from the Fly Atlas 2.0 (Krause, et al. 2022). C tSNE plots illustrating D. melanogaster kraken expression in single-cell transcriptomes of the Malpighian tubules and gut from the Fly Cell Atlas (Stringent 10x datasets) (Li, et al. 2022). D Left: RNAscope detection of kraken (green) transcripts in the gut and Malpighian tubules, with nuclei counterstained with DAPI (blue). Right: higher-magnification images showing kraken transcript expression in the indicated tissues. Scale bars, 200 μm. This expression pattern was observed in tissues from >20 individuals. E Expression levels of the kraken and Alkbh7 in the indicated laboratory strains (left panel) and wild caught strains (right panel) of adult female D. sechellia and D. simulans measured by qPCR (Data S6). Expression is represented as calibrated and normalized relative quantities (CNRQ). Significance was assessed using the unpaired t-test correcting for multiple testing and represented using letter codes. F Expression levels of kraken and Alkbh7 in the indicated strains and species at larval or adult stages as indicated from published RNA-seq data. (References: 1, (Watanabe, et al. 2019), 2 (Ma, et al. 2018), 3 (Kalra, et al. 2024)). Significance was assessed using the unpaired t-test correcting for multiple testing and represented using letter codes when multiple replicates were present in the original data. G Expression levels of the candidate genes at generation 0, 25 and 50 of the experimentally-evolved D. simulans populations measured by qPCR (Data S6). Significance was assessed using the paired t-test correcting for multiple testing and represented using letter codes.

Functional validation of the contribution of kraken and Alkbh7 to OA resistance.
A-C Survival curves of adult flies of the indicated genotypes tested in the plate assay with 2 μl OA for kraken and Alkbh7 RNAi (A), mutants (B) and overexpression (C). Each replicate consisted of 10 2-7-day-old female flies. N replicates ≥6, n flies ≥60. Shading indicates 95% confidence intervals. Genotypes: A act-Gal4/+;UAS-kraken^RNAi/+, UAS-kraken^RNAi/+, act-Gal4/+, act-Gal4/+;UAS-Alkbh7^RNAi/+, UAS-Alkbh7^RNAi/+, act-Gal4/+ B kraken¹, Alkbh7¹, Canton-S (wild-type genetic background control for both mutants), C act-Gal4/+;UAS-kraken^o/e/+, UAS-kraken^o/e/+, act-Gal4/+, act-Gal4/+;UAS-Alkbh7^o/e/+, UAS-Alkbh7^o/e/+, act-Gal4/+. Significance is based on the resulting mixed effect Cox regression model. NS p>0.05, *p<0.05, **p<0.01, ***p<0.001. The same data for D. sechellia are shown in both plots in C. D Survival curves of wild-type (Dsec07), and kraken (Dsec\kraken^RFP) and Alkbh7 (Dsec\Alkbh7^RFP) mutant D. sechellia in the plate assay with 30 μl OA. Statistics as in A. E Long-term survivability of the same genotypes as in D (n = 60 per condition) in control conditions (no OA) (left) and food supplemented with 0.9% OA (right). Statistics as in A. Raw data in Data S7.

Drosophila strains.

Effective population size estimates from the experimentally-evolved populations.

qPCR primer efficiency.

Linkage disequilibrium decay, and comparison of simulated and observed p-value distributions in the experimentally evolved populations.
A Linkage disequilibrium decay at generations 25 (left) and 50 (right). Black curves represent the trajectory of r2 values (y-axis left) as genomic distances increase (x-axis). Red curves represent the same but for D’ values (y-axis right). B Violin plots showing the distribution of p-values from the CMH test, comparing allele frequency changes between generation 25 and 0 (left), and generation 50 and 0 (right). Blue violins represent p-values derived from forward evolution simulations, while orange violins correspond to p-values from experimentally evolved populations. Red dashed lines mark the top 0.000005% of simulated p-values, serving as the significance threshold for the experimental data.

Sequence and molecular evolutionary analyses of kraken and Alkbh7.
A,B Protein sequence alignment and secondary structure of Kraken (A) and Alkbh7 (B) from the reference sequence of D. melanogaster (NP_477265.1 and NP_648511.2) and the consensus sequences of D. simulans and D. sechellia. Sequence data from PRJNA215932 and PRJNA395473 (see Materials and Methods). Black triangles indicate the residues where signs of selection were detected by the FUBAR analysis (see D). C Results from the McDonald-Kreitman test. NI = neutrality index; alpha = proportion of substitutions driven by positive selection; p-value = p-value resulting from a two-tailed Fisher’s exact test. D FUBAR analysis to detect pervasive selection as implemented in HyPhy. Position = codon position; alpha = mean posterior synonymous substitution rate; beta = mean posterior nonsynonymous substitution rate; posterior prob = posterior probability of positive selection; codons = alternative codons, Freq Dsec = frequency of the alternative codons in D. sechellia in the same order as codons; Freq Dsim = frequency of the alternative codons in D. simulans in the same order as codons. E Survival curves of D. melanogaster overexpressing Dmel\kraken and Dsec\kraken exposed to 2 μl OA. Each replicate consisted of 10 2-7-day-old female flies. Genotypes: act-Gal4/+;UAS-Dmel\kraken/+, act-Gal4/+;UAS-Dsec\kraken/+. N replicates ≥6, n flies ≥60. Shading indicates 95% confidence intervals. Significance is based on the resulting mixed effect Cox regression model. NS p>0.05. Raw data in Data S5.

Additional expression analyses of kraken and Alkbh7.
A Bar plots showing the expression levels of kraken and Alkbh7 across various adult Drosophila tissues and species. Data from (Bontonou, et al. 2024). B Bar plots showing the expression levels of kraken and Alkbh7 across various larval Drosophila tissues and species. Data from (Watanabe, et al. 2019) (“medium diet” condition). C Bar plots showing the effect of exposure to 9 μl OA (in the tube assay) on kraken and Alkbh7 expression in D. sechellia and three evolved (G50) D. simulans populations. D Left: RNAscope detection of kraken (green) and Rpl32 (magenta) transcripts in the gut and Malpighian tubules. Right: higher-magnification showing kraken transcript expression in the indicated tissues. Scale bars, 200 μm. This expression pattern was observed in tissues from >10 individuals.

Validation of RNAi and overexpression lines and generation of mutants for kraken and Alkbh7.
A Quantitative RT-PCR analysis of RNAi-mediated knockdown of kraken and Alkbh7 in D. melanogaster. Genotypes as in Fig. 5A. B Schematics of the CRISPR/Cas9 design to generate D. melanogaster and D. sechellia mutants for kraken and Alkbh7. Validation was performed by PCR and Sanger sequencing. The Dsec\kraken^RFP mutant comprised a 1431 bp deletion (-44 to +1,387 nt relative to the ATG), replaced with the 3xP3-RFP cassette. The Dsec\Alkbh7^RFP mutant was a 1023 bp deletion (-84 nt to +939 nt relative to the ATG), replaced with the 3xP3-RFP cassette. C Quantitative RT-PCR analysis of Gal4/UAS-mediated overexpression of D. melanogaster kraken and Alkbh7. Genotypes as in Fig. 5C.

Control survival curves in the absence of OA.
A Survival curves of adult flies of the indicated genotypes tested in the plate assay without OA. Each replicate consisted of 10 2-7-day-old female flies. N replicates ≥5, n flies ≥50. All genotypes display full viability, rendering datapoints invisible beneath the topmost displayed genotype (UAS-kraken^RNAi). Genotypes as in Fig. 5A-B. B Survival curves of wild-type (Dsec07), and kraken (Dsec\kraken^RFP) and Alkbh7 (Dsec\Alkbh7^RFP) mutant D. sechellia on homogenized noni pulp. N = 50 per strain. Significance is based on the resulting mixed effect Cox regression model. NS p>0.05. C Survival curves of kraken (Dsec\kraken^RFP, top) and Alkbh7 (Dsec\Alkbh7^RFP, bottom) mutant D. sechellia compared to wild-type (Dsec07) flies at 30, 33, 36 degrees Celsius. N ≥ 50 per strain/condition. Statistics as in B.

Sign up for email alerts