The nematode Caenorhabditis elegans has a cationic amphiphilic drug (CAD) defense system

Levon Tokmakjian; Duhyun Han; Kateryna Sihuta; Aanchal Aggarwal; Somayeh Pirhadi; Yao Wang; Andrew R Burns; Brittany Cooke; Siyue Ren; Marios Gavrielatos; Jiabao Liu; Henry M Krause; Carolyn L Cummins; Justin Nodwell; David R Koes; Peter J Roy

doi:10.1371/journal.pbio.3004006

. 2026 Sep 21;24(9):e3004006. doi: 10.1371/journal.pbio.3004006

The nematode Caenorhabditis elegans has a cationic amphiphilic drug (CAD) defense system

Levon Tokmakjian ^1,², Duhyun Han ^1,², Kateryna Sihuta ^2,³, Aanchal Aggarwal ^2,³, Somayeh Pirhadi ⁴, Yao Wang ⁴, Andrew R Burns ^2,³, Brittany Cooke ^2,³, Siyue Ren ^1,², Marios Gavrielatos ⁴, Jiabao Liu ^2,³, Henry M Krause ^2,³, Carolyn L Cummins ⁵, Justin Nodwell ⁶, David R Koes ⁴, Peter J Roy ^1,^2,^3,^*

Editor: Mark J Alkema⁷

¹ Department of Pharmacology and Toxicology, University of Toronto, Toronto, Ontario, Canada

² The Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, Ontario, Canada

³ Department of Molecular Genetics, University of Toronto, Toronto, Ontario, Canada

⁴ Department of Computational and Systems Biology, University of Pittsburgh, Pittsburgh, Pennsylvania, United States of America

⁵ Department of Pharmaceutical Sciences, Leslie Dan Faculty of Pharmacy, University of Toronto, Toronto, Ontario, Canada

⁶ Department of Biochemistry, University of Toronto, Toronto, Ontario, Canada

⁷UMass Chan Medical School, UNITED STATES OF AMERICA

^✉

* E-mail: peter.roy@utoronto.ca

The authors have declared that no competing interests exist.

Roles

Levon Tokmakjian: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Validation, Visualization, Writing – original draft

Duhyun Han: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Validation, Writing – original draft

Kateryna Sihuta: Data curation, Investigation, Methodology, Validation, Writing – original draft

Aanchal Aggarwal: Investigation, Methodology

Somayeh Pirhadi: Formal analysis, Investigation

Yao Wang: Formal analysis, Investigation

Andrew R Burns: Investigation, Methodology

Brittany Cooke: Investigation, Methodology

Siyue Ren: Investigation

Marios Gavrielatos: Data curation, Formal analysis, Investigation, Methodology, Validation, Visualization, Writing – original draft, Writing – review & editing

Jiabao Liu: Formal analysis, Investigation, Methodology, Validation, Visualization, Writing – original draft, Writing – review & editing

Henry M Krause: Funding acquisition, Resources, Supervision, Writing – review & editing

Carolyn L Cummins: Conceptualization

Justin Nodwell: Resources

David R Koes: Conceptualization, Formal analysis, Methodology, Resources, Supervision, Visualization, Writing – original draft

Peter J Roy: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Resources, Supervision, Visualization, Writing – original draft, Writing – review & editing

Mark J Alkema: Academic Editor

PMCID: PMC13619054 PMID: 42766653

Abstract

Cationic Amphiphilic Drugs (CADs) severely disrupt lysosomal function, which leads to a cellular pathology in humans characterized by excess phospholipids called phospholipidosis. Through a forward genetic screen and mining of published datasets, we discovered that CADs induce the expression of the CYP-35B family of cytochrome P450s and the PGP-13 p-glycoprotein pump via the nuclear receptors NHR-70 and NHR-107 in the nematode C. elegans. A pgp-13 fluorescent reporter revealed hundreds of human drugs that up-regulate the CAD defense system in vivo. Chemoinformatic analyses indicate that the pgp-13 reporter may be useful in identifying CADs that have pathogenic potential in humans. Mutant analyses coupled to metabolomics and structural modeling show that the CYP-35Bs are necessary and sufficient for CAD metabolism, and that CYP-35B2 D311 is key in mediating electrostatic interactions with the positively charged CADs. We also show that CAD metabolites are effluxed via PGP-13 acting partially redundantly with PGP-14 and that an intact defense system is necessary to resist CAD-induced pathology. Finally, we demonstrate that bacteria that likely cohabitate with C. elegans in nature trigger the CAD defense system, providing a plausible explanation for why a pathway that protects against anthropogenic small molecules exists in nematodes.

Cationic amphiphilic drugs (CADs) can collaterally disrupt lysosomal function, inducing pathologies in humans. This study in C. elegans shows that a xenobiotic module couples CYP-35B–mediated CAD catabolism to efflux, likely serving to detoxify natural products generated by microbes in the soil.

Introduction

Cationic Amphiphilic Drugs (CADs) can severely disrupt lysosomal function and lead to a cellular pathology called drug-induced phospholipidosis (PLD) [1]. CADs are small lipophilic molecules that typically have a secondary or tertiary amine with a pKa around 8 [2]. At physiological pH, a good fraction of the CAD molecules is unprotonated and can pass through membranes. In acidic conditions, a greater fraction of the CAD molecules is positively charged and becomes trapped in acidic organelles like the lysosome and accumulate there to high concentrations [3]. The protonated CADs neutralize the negatively charged membranes of the acidic organelles [1]. Consequently, the phospholipid-degrading enzymes that normally interact with the lysosome via its negative membrane charge can no longer do so in the presence of CADs [4]. In this way, CADs drive lipid accumulation (aka PLD) as an unintended side-effect.

CADs are problematic for several reasons. First, CADs are more likely to be associated with drug-induced liver injury (DILI) compared to non-CADs [5]. DILI can lead to liver transplantation and death for many of those afflicted [6,7]. Second, most PLD-inducing CADs have physicochemical properties (i.e., a hydrophobic backbone coupled to a positively charged atom) in common with blockers of the human ether-a-go-go related gene product hERG (KCNH2 or Kv11.1) [8]. Indeed, 77% of PLD-inducing CADs have been shown to block hERG activity [9]. hERG is a voltage-gated potassium channel that is key in maintaining cardiac rhythm and whose disruption can lead to sudden cardiac death [10]. Given their associated DILI and cardiac pathologies, cationic amphiphilic molecules are considered high risk in medicinal chemistry pipelines [10,11]. Despite this, roughly 5% of all small molecule drugs used in the clinic are PLD-inducing CADs [12].

A master drug clearance pathway in mammals consists of the PXR nuclear receptor, which when bound by a variety of compounds, will drive the expression of multiple drug-metabolizing cytochrome P450s and the drug efflux pump ABCB1 (aka MDR1, aka Pgp) [13–16]. However, known drug detoxification pathways are only modestly effective when responding to CADs. For example, although the canonical CAD amiodarone induces ABCB1 expression, the drug inhibits the pump’s efflux activity [17,18]. Similarly, while multiple cytochrome P450s can metabolize amiodarone, the resulting metabolites inhibit these same P450s [19]. Another classic CAD is chlorpromazine, which activates the PXR drug response [20]. While the P450s CYP1A2, CYP2D6, and CYP3A4 metabolize chlorpromazine [21], chlorpromazine itself inhibits ABCB1’s activity [18].

Here, we have discovered a coordinated CAD defense system consisting of the nuclear receptors NHR-70 and NHR-107, the CYP-35B P450 family, and the PGP-13 and PGP-14 efflux pumps using the nematode worm C. elegans. We discovered this pathway serendipitously through our search for small molecules that can circumvent the disruption of a C. elegans homologue of a human disease gene. Through our previous small molecule screens [22], we discovered one class of small molecules (not CADs) that kills the nematode C. elegans by forming crystals exclusively in the pharynx cuticle [23–26]. A forward genetic screen revealed that the p-glycoprotein (PGP) pump PGP-14, which is expressed exclusively in the pharynx myoepithelium, is needed for polar lipid accumulation within the pharynx cuticle, which in turn acts as a sink for hydrophobic small molecules that form lethal crystals [24,25]. Loss-of-function of pgp-14 confers resistance to the lethality associated with crystal-forming compounds.

PGP-14 is most homologous to the human PGPs ABCB4, a pump that exports phospholipids from hepatocytes to form bile [27,28], and the aforementioned ABCB1 multidrug efflux pump [25,29,30]. Without functional ABCB4, phospholipids are not present to promote bile acid entry within the hepatocytic duct, leading to hepatocytic damage that can result in a childhood liver disease called progressive familial intrahepatic cholestasis 3 (PFIC3) [31]. There is a clear analogy to PGP-14’s role in establishing the lipid barrier within the C. elegans pharynx cuticle and ABCB4’s role in phospholipid export from hepatocytes.

Given PGP-14’s analogous role and homology to ABCB4, we were curious to explore the potential utility of the pgp-14(0) null mutant as a model for ABCB4 deficiency in humans. Towards this end, we screened the Spectrum library of 2,560 drugs and natural products for those that suppress pgp-14(0). All resulting suppressors are structurally diverse CADs. Through a forward genetic screen, we discovered that CADs suppress pgp-14(0) because they up-regulate an ABCB1 homolog called PGP-13 that acts redundantly with PGP-14. In addition to PGP-13, CADs up-regulate multiple cytochrome P450s of the CYP-35B family, via the nuclear receptor transcription factors NHR-70 and NHR-107. Whether the CADs act as ligands for these nuclear receptors is an open question. Using mass-spectrometry, we show that the P450s metabolize CADs into multiple products, some of which are effluxed by PGP-13 acting either alone or redundantly with PGP-14. In this way, we have discovered a novel coordinated CAD defense pathway, without which the animals become more sensitive to the pathological effects of CADs.

To better understand the diversity of molecules that engage the CAD defense system we built a fluorescent reporter of pgp-13. We used this reporter to screen the Spectrum small molecule library at a single concentration and identified 293 structurally diverse inducers, 78% of which have either secondary or tertiary amines, indicating that C. elegans has a robust and plastic CAD detection system. A published chemoinformatic predictor of PLD-inducing CADs [11] predicts that 73% of our 293 hits are PLD-inducing CADs. The same chemoinformatic tool predicts the same fraction of PLD-inducing CADs from a set of 112 gold standard PLD-inducing CADs that we assembled from the literature. Hence, our in vivo CAD defence reporter is a good in vivo predictor of CAD-induced pathologies in humans and provides a facile model for testing predictions.

Finally, we asked whether bacteria that likely cohabitate with C. elegans in nature produce products that engage the CAD-defense system. We screened 65 Streptomyces strains, the genus of which are highly abundant within the natural environment of C. elegans [32,33] and found two strains that robustly up-regulate the CAD defense reporter in a nuclear receptor-dependent manner. These observations suggest that the CAD defense system exists in nematodes to defend against natural product CADs that have the potential to harm the worm.

Results

A Screen for small molecules that Rescue pgp-14(0) yields multiple CADs

To determine whether the phenotypes associated with the pgp-14(0) null mutant could be suppressed by small molecules, we screened 2,560 clinically approved drugs and natural products of the Spectrum library (Microsource Inc) for those that can rescue pgp-14(0)’s resistance to the lethal crystal-forming wact-190 (3-(4-(benzyloxy)phenyl)-5-(pyridin-4-yl)-1,2,4-oxadiazole) (Fig 1A) [25]. We focused on 14 reproducible hits (Fig 1B; S1 Data File-tab 1, column ‘AB’). These hits rescued pgp-14(0)’s resistance to both the death/arrest phenotype induced by crystal-forming molecules (Fig 1C) and the ability of the molecules to form crystals (Fig 1D). In the wildtype, crystal-forming molecules form obvious crystals over 24 hours. In the pgp-14(0) mutant, crystals rarely if ever form. pgp-14(0) mutants grown in the rescuing drugs take about two days to form obvious crystals in the arrested animals. Hence, the 14 rescuing compounds have slower temporal dynamics compared to wild type PGP-14.

Fig 1 — A. A schematic illustrating the logic of the drug suppressor screen of *pgp-14(0)*. 30 µM of drug and 15 µM of wact-190 were used in the screen. B. Silicos-IT Log P estimations are indicated in brackets. C. 6-day viability assays illustrating how select CADs, tested at 30 µM, resensitize the *pgp-14(0)* mutant *ok2660* to the lethality of 30 µM wact-190. Either three or six independent repeats were performed (N = 3 or 6); an upper ceiling of 50 animals per sample was counted. An asterisks indicates p < 0.001 in a Student’s 2-tailed T test, calculated only for the *pgp-14(0) +* wact-190 condition relative to the *pgp-14(0) +* solvent-only control (indicated with a blue asterisks). D. 30 µM of wact-190 and 30 µM of the indicated drug were used. At least 3 independent repeats were performed (N ≥ 3), with at least 5 worms counted per sample (n ≥ 5). An asterisks indicates p < 0.05 in a Student’s 1-tailed T test, calculated only for the *pgp-14(0)* background relative to the solvent-only control (indicated with a blue asterisks). For both graphs, standard error of the mean is shown. See S1 Data for the data underlying the graphs.

The 14 rescuing compounds are: i) drugs used in the clinic for diverse indications, ii) structurally diverse, iii) tertiary amine-containing molecules (and therefore positively charged at pH ≤ 7), and iv) hydrophobic molecules as measured by a computationally predicted hydrophobicity score (cLogP) greater than +2 (Fig 1B; S1 Data File-tab 1). Hence, these 14 compounds are considered CADs [2]. The Spectrum library has a total of 336 tertiary amine-containing molecules (S1 Data File-tab 1, column ‘FZ’). A hyper Hence, these 14 compounds are cogeometric p-value calculation indicates that the chances of all 14 hits being tertiary amine-containing molecules by random chance alone is very low (p ≤ 4E-13).

CADs compensate for the loss of PGP-14 through the upregulation of PGP-13

To understand how the CADs rescue pgp-14(0), we carried out a forward genetic screen for mutants that suppress the ability of CADs to rescue pgp-14(0). We grew randomly EMS-mutagenized pgp-14(0) mutants in a solution containing a crystal-forming compound (wact-190) and a CAD (terconazole). Normally, the CAD resensitizes the pgp- 14(0) mutant to the lethality conferred by wact-190 crystal formation. We simply selected for mutants that could live under these conditions (see the bottom line of Fig 1A).

The results of the screen yielded multiple alleles of nhr-70, nhr-107, pgp-13, and tat-3 (Fig 2A; S1 Data File-tab 4). For each of the four genes, there are alleles that are highly likely to represent loss-of-function alleles, yielding either early stops, deletions, or splicing defects (Figs 2A; S1). We do not further discuss tat-3 here. NHR-70 and NHR-107 are distantly related to the human HNF4 family of nuclear receptors and do not have clear one-to-one human orthologs [34]. In humans, the HNF4s are primarily expressed in endodermal epithelia, including the liver and intestine [35,36] and are closely related to the xenobiotic-responsive PXR and CAR nuclear receptors [37]. In blast searches with the PGP-13 query, human ABCB1 is the best human match. Of note, the pgp-13 gene is situated immediately next to pgp-14 in the genome, which is remarkably similar to the physical relationship between ABCB1 and ABCB4 in humans (S2 Fig).

The screen results suggested that CADs directly or indirectly activate the NHR-70 and NHR-107 nuclear receptors to transcriptionally up-regulate PGP-13, the closest paralog of PGP-14 [25,38], to restore the pgp-14(0) mutant’s sensitivity to wact-190-mediated lethality (Fig 2B). See Fig 9 for an expanded model. We tested this model in multiple ways. First, we used qRT-PCR to examine pgp-13 mRNA expression in response to the CAD terconazole in the background of the nuclear receptor mutants. We observed a 47-fold induction of pgp-13 expression in response to the CAD terconazole (p = 1.7E-05) and find that this upregulation is dependent on functional NHR-70 and NHR-107 (p = 4.1E-04 for both) (Fig 2C). Second, we asked whether pgp-13, nhr-70, and nhr-107 were necessary for the CAD’s ability to restore pgp-14(0)’s sensitivity to wact-190 crystal formation and found that they were (S3 Fig). Third, we tested whether expression of ligand-binding domain fusions of NHR-70 and NHR-107 to the GAL4 DNA-binding domain (DBD) could drive reporter expression in HepG2 cells in response to CADs and found that they could not (S4 Fig). It is possible that this heterologous system lacks other C. elegans proteins necessary drive NHR-dependent transcription. Finally, we also performed computational modeling to predict whether the CADs directly interact with each of the NHR’s ligand-binding domains (S5 Fig). The positive controls behaved weaker than expected and the modeling provided no evidence of direct interaction. We consider the question of how the CADs interact with the NHRs an open question.

Fig 9 — CADs may directly interact with the nuclear receptors NHR-70 and/or NHR-107 to drive the expression of gene products involved in CAD metabolism and export. Alternatively, metabolite(s) generated in response to CAD perturbation of the cell may engage NHR-70 and NHR-107 and drive the CAD defense response. Metabolites of the CAD molecules themselves are likely exported by both PGP-13 and PGP-14. The CADs can rescue the *pgp-14(0)* phenotype (i.e., resistance to crystal-induced death that is dependent on the likely export of polar lipids) via the induction of compensatory PGP-13 expression.

We explored the spatiotemporal expression pattern of pgp-13 with a transcriptional YFP reporter, which revealed expression in the pharynx myo-epithelium and the intestine in a CAD and nuclear receptor-dependent manner (Fig 2D–2K). Expression is evident (p < 0.05) within 2 hours of CAD exposure and peaks approximately 24 hours after initial exposure (Fig 2H and 2I; S1 Data 1-tab 9). Intriguingly, one CAD’s ability to induce pgp-13 expression (terconazole) is dependent on both NHR-70 and NHR-107, while another (amiodarone) is dependent on only NHR-70 (Fig 2J and 2K). We expand on these observations below. Upon CAD induction, a functional translational fusion reporter of PGP-13 (S6 Fig) shows subcellular localization to the apical membrane of the pharynx myoepithelium that is highly similar to that of PGP-14 [25] (Fig 2L and 2M). The PGP-13 fusion protein also localizes to the apical membrane of the intestine (Fig 2N and 2O). Given the paralogy of PGP-13 and PGP-14 [25] and the spatial overlap of their expression pattern, we infer that CAD-induced PGP-13 expression fortuitously compensates for the loss of pgp-14 and accounts for why CADs were identified as pgp-14(0) suppressors in our initial screen.

pgp-13 is a robust in vivo marker of diverse biologically-accessible CADs

We were curious to know what other small molecule structures might up-regulate pgp-13 expression in a nuclear receptor-dependent manner. We therefore screened 2,400 compounds from the Spectrum library at a final concentration of 30 µM and identified 293 molecules that induce expression from the pgp-13p::YFP reporter (S1 Data file-tab 1- column ‘J’). With each molecule reduced to its core scaffold (i.e., removing everything but rings, connectors, and carbon chains), there are 195 distinct structural scaffolds, indicating that the hits are structurally diverse (Fig 3A). We evaluated our hits using two approaches. First, we compared our hits to a list of 112 molecules known to induce phospholipidosis in humans that we compiled from the literature and are also present in the Spectrum library [11,39–42] (S1 Data file-tab 1 columns HU-IB). We refer to these molecules as gold standards. The pgp-13-inducers capture 68% of the gold standards, which is 5.6-fold more than expected by random chance (p = 2E-46) (Fig 3B). To understand why 32% of the gold standards failed to induce pgp-13 expression, we compared the physicochemical features of the gold standards that induce pgp-13 expression compared to those that do not. We find that the latter are less likely to cross biological barriers (p < 1E-5); they are significantly more polar (as measured by hydrogen bond donors and acceptors and total polar surface area), hydrophilic (as measured by LogP) and less likely to cross the skin and blood-brain barriers of humans (as measured by LogKp and a published predictive formula [43,44], respectively) (Fig 3C). We tested whether three of the failed gold standards would elicit pgp-13 reporter expression at higher concentrations and found that they did not (S7 Fig). It may be that some gold standards fail to induce pgp-13 expression because of differences in pharmacokinetic properties (i.e., absorption, metabolism, export) compared to those that elicit reporter expression.

Fig 3 — A. A structural similarity map of the 293 hits reveals 195 unique core scaffolds. Each of the nodes represents one of the 293 molecules reduced to its core scaffold (rings connected by chains). If two molecules within the collection of 293 molecules share the same scaffold, they are connected with a line. Gold-colored nodes represent gold standard PLD-inducing molecules. Bold outlines are the 195 molecules further investigated in Fig 4. For clusters containing 3 or more molecules, the core scaffold is shown below, most of which have tertiary amines that are not part of the core scaffold. The identity of each molecule is represented by a number (see S1 Data -tab 1-column J to decode). B. Venn diagram showing the overlap between the *pgp-13*-inducing hits, the gold standards, and the predicted PLD-inducing CADs within the Spectrum library’s 2,560 small molecules. C. Graphs of the select physicochemical properties (from SwissADME) of the molecules within the 112 gold standard (GOLD STD) set and the 288 predicted PLD-inducing CAD set (PRED MODEL) parsed on whether the molecule induces *pgp-13* expression or not. The physicochemical properties of all molecules are presented in S1 Data -tab 1. MLogP was used for LogP; LogKp is a measure of skin permeability (the higher the number the higher the permeability); with blood-brain barrier (BBB) permeability, each molecule is given a digital score (1 = likely permeable; 0 = likely impermeable). A two-tailed Student T test determined significance. The mean and the standard error of the mean are shown.

Second, we compared the pgp-13-inducers to the Spectrum molecules predicted to be PLD-inducing CADs in humans. To identify PLD-inducing CADs, we employed a chemoinformatic formula that predicts whether a molecule is a CAD that will induce PLD in human cells based on the molecule’s lipophilicity and pKa [11] (see Methods; S1 Data file-tab 1, columns GL-GZ)). With this formula, we identified 288 Spectrum molecules that are predicted to be PLD-inducing CADs (Fig 3B; S1 Data file-tab 1, column HA). This predictive formula identifies 73% of the gold standards (Fig 3B), which is 6.1-fold more than expected by random chance (p = 3E-55). The predictive formula also identifies 73% of the pgp-13 inducers (Fig 3B), which is also 6.1-fold more than expected by random chance (p = 6E-169). Like the gold standards, the predicted PLD-inducing CADs that fail to induce pgp-13 expression are significantly more polar and less likely to pass the skin barrier than those molecules that drive pgp-13 expression (Fig 3C; S1 Data file-tab 1). Importantly, the chemoinformatic predictor, which was trained on human data [11], does equally well with predicting which molecules will induce pgp-13 expression in worms. This reinforces the idea that the properties of molecules that drive pgp-13 expression in worms are intimately related to those that make a molecule a PLD-inducing CAD in humans.

Most pgp-13 inducers require NHR-70 and NHR-107

Above, we show that terconazole’s induction of pgp-13 is dependent on both NHR-70 and NHR-107, while amiodarone’s induction of pgp-13 is dependent on only NHR-70. We were curious to know whether most CADs require both nuclear receptors to induce pgp-13 expression. We therefore assembled 195 molecules from our original 293 hits into a new mini-library (based on additional stringency and availability) (see bolded nodes of Fig 3A). We screened this mini-library for the ability of the molecules to induce pgp-13p::YFP expression in an NHR-70 and NHR-107-dependent manner. We found that 149 (76%) of the 195 molecules were dependent on both nuclear receptors to induce robust pgp-13 expression (which we call ‘class I’ molecules) (p < 0.05), 24 molecules (12%) were dependent on only NHR-70 (which we call ‘class II’ molecules) (p < 0.05), and 3 (2%) were not dependent on either nuclear receptor (which we call ‘class 3′ molecules) (p < 0.05) (Fig 4A; S1 Data File-tab 1-columns ID-LP). Another 19 molecules did not fall into any of these bins based on statistical analyses and cutoffs.

Fig 4 — A. A heatmap showing the ability of 195 cherry-picked small molecules from the Spectrum library to induce *pgp-13p*::YFP expression in either a *wild type* (RP3507), *nhr-70(tm1697)* (RP3518) or *nhr-107(tm4512)* (RP3516) background at a concentration of 30 µM. 149 small molecules induced lower *pgp-13* reporter expression in the *nhr-70(0)* and *nhr-107(0)* mutants compared to control (p < 0.05). Twenty-five small molecules induced lower expression of the *pgp-13* reporter in only the *nhr-70(0)* mutant compared to control (p < 0.05), and another 3 molecules did not have significantly lower expression in either mutant (p > 0.05). See S1 Data -tab 1- columns ID-LP for details. B. Violin plots of the indicated physicochemical properties for the class I and II molecules (indicated in A) relative to 2024 true negatives within the Spectrum library. The median is indicated. A two-tailed Student T test determined significance. For clarity, 21, 1, 13, and 63 data points from the true negative class are left off from the molecular weight, cLogP, rotatable bonds, and van der Walls volume graphs, respectively (but are accounted for in the medians and statistical analyses).

Physicochemical analyses show that class II molecules are enriched for molecules that are longer and less bulky than class I molecules; class II molecules are significantly (p < 0.001) more hydrophobic (cLogP), more flexible (rotatable bonds), and have a larger volume (van der Waals volume) without having a greater molecular weight compared to class I (Fig 4B; S1 Data File-tab 12). Perhaps relatedly, class II molecules are more likely (p < 0.001) to penetrate human skin than class I molecules (LogKp) (Fig 4B). These results provide a physicochemical basis for why different CADs may rely on a different combination of nuclear receptors to drive a CAD-response. Furthermore, there are a small number of CADs that drive pgp-13 reporter expression independent of either nuclear receptor, suggesting that additional CAD-detection components beyond NHR-70 and NHR-107 may exist.

Metabolic analyses reveals a CAD detoxification system

The above results suggest that the CAD-induced expression of PGP-13 is part of a drug-response pathway and that there may be other genes like pgp-13 whose expression is induced by the CADs in a nuclear receptor-dependent manner. To identify such genes, we examined published global expression analyses in response to the CADs mianserin and tamoxifen [45,46]. Twenty-one genes, including pgp-13, were significantly and coincidentally up-regulated in both studies. We cross-referenced this gene list with a meta-analysis of 2243 genome-wide expression experiments that identified 11 genes that are significantly co-regulated with pgp-13 [47]. In total, five genes were common among the three datasets; pgp-13, three cytochrome P450s (aka P450s), including cyp-35B1, cyp-35B2, and cyp-35B3, and the predicted oxidoreductase cror-1 (Fig 2C). CYP-35B1 is most similar to CYP-35B2 at the amino acid level (75%), while CYP-35B3 is least like the other two paralogs at 73% and 72% identity with CYP-35B1 and -35B2, respectively. The CYP-35B family members belong to P450 clan 2 and do not have a clear one-to-one orthology with homologs of human clan 2 P450s [48], although human clan 2 member CYP2D6 is known to interact with positively charged molecules and can metabolize CADs [49,50]. We verified by qRT-PCR that these genes are indeed upregulated by multiple CADs in a manner dependent on either NHR-70 or NHR-70 and NHR-107 (Figs 2C and S8). We hypothesize that these five enzymes are part of a CAD defense system (Fig 2B). cror-1 will not be further discussed here.

We investigated whether the components that are upregulated in response to CADs are involved in CAD detoxification. We first tested the hypothesis that NHR-70 is key in turning on machinery that metabolizes the CADs. To test this, we compared the ability of wild-type worms to generate oxidized products of the CAD to that of the nhr-70(0) null mutant using liquid chromatography-coupled mass spectrometry (LCMS). For LCMS analyses, we focused on CADs that: i) induce robust expression from the pgp-13p::YFP reporter; ii) are structurally distinct; iii) have a tertiary amine and/or are predicted to be a PLD-inducing CAD; and iv) have a chloro group, which is not naturally found in animal metabolomes and exist as two highly abundant isotopes that are easily identified in LCMS analyses [51]. The chosen CADs are shown in Fig 5. Of the six CADs examined in detail, worms generated NHR-70-dependent oxidized metabolites above background levels for all molecules except terconazole (p ≤ 0.05)(Fig 5a). This indicates that not all CADs that up-regulate the CAD defense system are robustly metabolized by NHR-70-driven products.

Fig 5 — A. The abundance of mono or di oxidized products of the indicated parent molecule in the lysates of wild-type worms relative to their abundance in the indicated mutants as measured by LCMS. The strains used are wild type (N2), TM1697 *nhr-70(tm1697),* RB2216 *cyp-35B1(ok2998)*, TM7435 *cyp-35B2(tm7435)*, and TM12255 *cyp-35B3(tm12255).* B. Similar to A, the abundance of metabolites that accumulate in the lysate of the indicated PGP mutant relative to the wild type. The strains used are RB894 *pgp-13(ok747)*, RP2960 *pgp-14(ok2660)*, and RP3495 *pgp-13(tr696) pgp-14(ok2660).* In both A and B, counts are normalized to the abundance of the parental molecule in each strain, and the exact masses of each metabolite and the p values (one tailed Student T test relative to the normalized parent molecule counts) are found in S1 Data. N ≥ 3 biological trials, n ≥ 2 technical repeats. C. The abundance of the parent and mono and di oxidized metabolites in yeast (lysate and buffer combined) expressing the indicated P450 relative to (-) control yeast that do not express heterologous P450s. Chlorprothix., chlorprothixene. Statistical tests (one tailed Student T test) are done relative to parent mass, or when indicated with a horizontal line, relative to another P450-expressing strain. All p-values and exact masses are reported in S1 Data, tabs 14 and 15. No blinding was used in the experimental design. In all graphs, the standard error is shown.

Next, we asked whether any of the three cytochrome P450s, whose expression is induced by CADs in an NHR-70-dependent manner, are necessary for CAD metabolism (oxidation). We again used LCMS to measure the accumulation of CAD metabolites in wild-type worms compared to loss-of-function mutants of cyp-35b1, 35b2, and 35b3. We found that the loss of CYP-35B2 reduces the oxidation of all five CADs tested while the loss of CYP-35B1 reduces the oxidation of only three (p ≤ 0.05) (Fig 5a). The loss of CYP-35B3 shows no decrease in oxidized CADs, consistent with the observation that the expression of CYP-35B3 is not induced by CADs as much as CYP-35B1 and B2 (Fig 2c). This data demonstrates that CYP-35B1 and -35B2 are necessary for robust CAD metabolism.

We next investigated the role of PGP-13 in CAD detoxification by asking whether CAD metabolites accumulate more in the pgp-13(0) null mutant relative to wild type. Accumulation of CAD metabolites in the pgp-13(0) mutant would suggest a role for the PGP in drug efflux. Of the six CADs examined, we found multiple metabolites of four CADs that accumulate in the mutant (Fig 5b). Given that PGP-13 acts at least partially redundantly with PGP-14 in crystal formation (see above), we asked whether PGP-13’s role in CAD efflux for some drugs might be obscured by the activity of PGP-14. Indeed, the two CADs for which we failed to see CAD metabolite accumulation in the pgp-13(0) single mutant accumulate in the pgp-13(0) pgp-14(0) double mutant (Fig 5b). Hence, PGP-13 plays a role in CAD metabolite efflux both independently and redundantly with PGP-14, depending on the CAD in question.

We also investigated whether any of CYP-35B1, B2, or B3 are sufficient for CAD metabolism. To do so, we exploited our previously established methodology to heterologously express worm P450s in the yeast S. cerevisiae to determine their metabolic capability [52–54]. Because the yeast experiments are less labor intensive compared to the worm metabolic analysis, we were able to examine more CADs. Of the 11 CADs examined, 9 were oxidized by at least one of the three P450s with CYP-35B2 metabolizing 9, CYP-35B1 metabolizing 8, and CYP-35B3 metabolizing 6 (p ≤ 0.05) (Fig 5c). We were curious to determine whether the P450s inability to metabolize fipexide and terconazole in yeast might be related to their potential endogenous detoxification and/or a failure to robustly accumulate in yeast. We therefore compared the ratio of the accumulation of the CAD in the yeast cell lysate relative to what is found in the buffer for each of the 11 CADs. The only CAD that has a distinct accumulation pattern is clomiphene (p < 0.05) (S9 Fig). Hence, the failure of fipexide and terconazole to be metabolized by yeast-expressed CYP-35B family members is not due to yeast detoxification.

CYP-35B1 is better able to metabolize chlorprothixene and eticlopride compared to CYP-35B2, which is seemingly contradictory to the worm mutational analysis shown in Fig 5a. However, cyp-35b2 is expressed ~32 fold more than cyp-35b1 in worms (Fig 2c). In yeast, all 11 P450s are expressed using the same galactose-inducible promoter and are likely expressed at roughly equivalently levels. Hence, while 35B1 may be able to metabolize select CAD structures better than 35B2 in yeast, its lower expression in worms has less of an impact on CAD metabolism. Regardless, our data demonstrates that CYP-35B family members are both necessary and sufficient to metabolize CADs in C. elegans.

Computational modeling reveals CYP-35B2 D311 as key to in CAD metabolism

Given the positive charge of CADs in cells, we were curious to know whether electrostatic interactions play a role in the ability of the CYP-35B family to metabolize CADs. Using CYP-35B2 and chlorprothixene and clomipramine as exemplar interactions, we employed structural modeling and molecular dynamics to model how charged and uncharged CADs interact with the P450. Protein structures, with the HEME and CAD, were predicted using Boltz-1 [55] and then simulated with Amber 22 [56]. We found that the carboxylate of residue D311 is frequently within hydrogen bonding distance (<3 angstroms) of the charged moiety of the CADs, but is never within this distance when the uncharged molecules are modeled (Figs 6a, 6b and S10).

Fig 6 — A, B. Computational modeling and simulation reveals that charged (blue) chlorprothixene (A) and clomipramine (B) frequently hydrogen bond with D311 while uncharged (yellow) versions do not consistently interact with D311. Representative conformations from molecular dynamics simulations are shown inset. C. The abundance of the indicated presumptive compound as identified by exact mass spectrometry as produced in yeast expressing either the wild type, D311N or D311G CYP-35B2. The abundance found in lysate and buffer was combined. Y axis is log₁₀. N = 3 biological trials, n = 2 technical repeats. A one-tailed Student T test was used to measure differences. A black asterisk indicates p < 0.05 when compared to the wild type; a purple asterisks indicates p < 0.05 when comparing the two mutants. OH, hydroxylated metabolite; GSH, glutathione conjugate; GSH-red, reduced glutathione conjugate; dealkyl; dealkylated metabolite; AB1515, 5-(furan-2-yl)-3-(m-tolyl)-1,2,4-oxadiazole; CLPTX, chlorprothixene; CLPRM, clomipramine. No blinding was used in the experimental design. The standard error of the mean is shown. See S1 Data for the data underlying the graphs.

To test whether the anionic D311 residue of CYP-35B2 plays a role in cationic amphiphilic drug metabolism, we substituted D311 for the more conservative but uncharged asparagine residue, and for the less conservative glycine residue. We then used LCMS to ask what impact these substitutions have on the ability of CYP-35B2 to metabolize chlorprothixene and clomipramine in our heterologous yeast assay. We focused on exact masses consistent with the hydroxylation and/or dealkylation of the molecules. For most metabolites, both D311N and D311G produced less than wild type (p < 0.05) (Fig 6c). Furthermore, D311G disrupts metabolism more than D311N for most metabolites (p < 0.05) (Fig 6c), which is consistent with the divergence in structures of the two substitutions. To control for the possibility that the mutations simply make the enzyme less efficient, we compared the ability of the two mutant P450s to generate metabolites from a non-CAD tool compound that we discovered that we call AB1515 (5-(furan-2-yl)-3-(m-tolyl)-1,2,4-oxadiazole) relative to the wild type. The two mutants generated more AB1515 metabolites than the wild type (p < 0.05)(Fig 6c). Hence, D311 is likely key in mediating the interaction with CADs via electrostatic interactions.

The CAD detoxification pathway protects against pathogenesis

The results presented above suggest that a CAD-responsive pathway exists to protect worms from the potentially pathogenic effects of CADs. We tested this idea in two ways. First, we examined the viability of wild-type worms grown in the presence of various CADs compared to relevant nuclear receptor and pgp mutants grown in the same compound using our standard 6-day viability assay [22]. In the wild type, the CADs have different lethal potencies, depending on their structure (Fig 7A). While pgp-13 and pgp-14 single mutants were hypersensitive to some CADs, the pgp-13 pgp-14 double mutant showed more hypersensitivity to most CADs (Fig 7A). Consistent with our earlier results that show that PGP-13 expression can compensate for the loss of PGP-14 (Figs 5B and S6), these data suggest that PGP-13 and PGP-14 play redundant roles in the export CADs and their metabolites. Mutants of nhr-70 and nhr-107 are similarly hypersensitive to most CADs, but especially those where CAD metabolite production is particularly dependent on NHR-70, namely azelastine and clomipramine (Fig 7A).

Fig 7 — A. Drugs assayed are indicated on the top, genotype on the bottom. Concentrations are the same for all drugs. Note the ~4-fold hypersensitivity of azelastine and clomipramine in the nuclear receptor mutants (white arrows). Note the apparent redundancy of PGP-13 with PGP-14 in their presumptive role in drug efflux (pink arrows) (i.e., without the two pumps, the animals are more sensitive to the CADs, compared to either single mutant). N = 3 biological trials and n = 2 wells with 20 animals per well. B. Examples of cDCF signal in the indicated genotypes following 24-hour incubation with solvent or the indicated CAD. Note the diffuse fluorescence throughout the intestine (arrows) in the wild type animals treated with CADs at the higher concentrations or in the mutants treated with CADs at the lower concentrations. C. Quantification of puncta for the animals assayed in differing concentrations of the CADs. D. Quantification of puncta for the animals assayed in different mutant backgrounds. For both C and D, the legend is on the right, N = 3 biological trials with n = 10 worms/trial. Statistical significance was determined using a two-way ANOVA followed by Dunnett’s multiple comparisons test, where **** indicates p < 0.0001 and * indicates p < 0.05. Standard error of the mean is shown. See S1 Data for the data underlying the graphs.

Second, we investigated whether C. elegans lysosome-related organelles are disrupted by CADs. We focused on the gut granules of the worm because they are well-characterized large acidic organelles that are terminal compartments for degradation and recycling endocytic pathways [57–59]. Previous work has employed carboxy-2′,7′-dichlorofluoresceine diacetate (cDCFDA) to study the pH of the gut granules [60]. cDCFDA is converted to the fluorescent cDCF (carboxy-2′,7′-dichlorofluoresceine) by esterases, which are robustly expressed in the gut [61]. Because cDCF has a pKa of 4.2 with multiple ionizable groups, a significant fraction of cDCF remains deprotonated even in acidic compartments like the gut granules and consequently accumulates within the organelle because of its charge [62].

CADs are well-known to neutralize acidic compartments and consequently disrupt their function [63]. Because cDCF accumulation in acidic organelles is dependent on a pH difference within the organelle relative to the cytosol, we used cDCFDA to track gut granule neutralization. We performed a small dose-response analysis of two CADs (chlorprothixene and clomipramine) using the cDCFDA gut granule acidity marker. We found that 60 µM chlorprothixene and 30 µM clomipramine had an unaltered pattern of concentrated cDCF accumulation in gut granules like the solvent-only negative controls. However, when incubated in twice as much CAD, cDCF localization became less concentrated with the gut granules (Fig 7B, 7C). These observations are consistent with the CAD’s ability to neutralize acidic organelles.

Given that NHR-70 and NHR-107 are key in the metabolism of chlorprothixene and clomipramine (Fig 5), we reasoned the corresponding loss-of-function mutants would increase the internal concentration of the parent CAD molecules and enhance the loss of cCDF signal within the gut granules. Indeed, we observed the same diffuse cDCF pattern in the gut with the 60 µM chlorprothixene and 30 µM clomipramine treatments in the nhr-70(0) and nhr-107(0) mutant backgrounds that were observed only at the higher concentrations in the wildtype background (Fig 7C, 7D). Of note, we also tested the pHTomato reporter [64], but it failed to generate signal with our positive control (10 mM chloroquine) (S11 Fig). These results further validate our model that the CAD-response machinery helps defend against CAD-related pathologies. Whether CADs induce phospholipidosis in C. elegans is an open question.

Soil microbes elicit the CAD defence response

Above, we show that hundreds of diverse compounds trigger a robust CAD defense response in C. elegans, prompting the question of why the pathway exists at all. To address this, we asked whether bacteria in the natural environment of the worm might also trigger the expression of the CAD defense reporter. Actinobacteria, of which Streptomyces constitute the largest genus, are common within the natural environment of C. elegans [32,33,65], and produce an abundance of bioactive natural products that include molecules that we employ as antibiotics, anticancer therapeutics, immune suppressants, antifungals, and other drugs [66,67]. We screened 65 diverse Actinobacteria strains assembled from diverse locations around North America by scientists at Cubist Pharmaceuticals Inc (later sold to Merck). Many of these strains produce obvious secondary metabolites, most of which are secreted into the media and/or deposited onto the surface of sporulating colonies (e.g., S12 Fig). From this collection, we identified 2 strains (1G9 and 1E6) capable of eliciting obvious pgp-13p::YFP expression (Fig 8). We tested whether this increase in signal was dependent on NHR-70 and NHR-107 and found that it was (p < 0.005), indicating that the increase in pgp-13 expression is likely elicited by a CAD-like natural product made by the bacteria (Fig 8). Together, these observations suggest that a CAD defense pathway may help C. elegans survive in the presence of CAD-producing microbes.

Fig 8 — A. Examples of one non-hit (5E1) and two hits (1E6 and 1G9) from our screen for *Streptomyces* strains that can elicit YFP expression from the RP3507 *trIs114* [*pgp-13p::YFP; p382(myo-2p::mCherry*)] CAD reporter strain after 24 hours of co-incubation. The top panels are brightfield images of the bacterial colony, the middle panels show the YFP signal (white arrows show one example) or lack thereof (red arrow), and the bottom panels show the constitutive expression of mCherry in the pharynx that marks the presence of the worms. B. Representative high-magnification images of individual worms exposed to the three bacterial strains. The white arrowheads indicate the pharynx; the black arrowheads indicate the intestine. C. Quantification of the YFP signal in the indicated genetic background grown in the indicated *Streptomyces* strain for 24 hours. N ≥ 3 biological trials with n = 10 worms/trial. A two way ANOVA was used to determine statistical differences. Red asterisks indicate p < 0.0001 relative the corresponding genotype grown on 5E1; black asterisks indicate p < 0.005 relative to wild type within the same growth conditions. See S1 Data for the data underlying the graphs.

Discussion

Here, we demonstrated that C. elegans harbors a pathway that defends against cationic amphiphilic small molecules (aka CADs). The pathway is adept at responding to a wide variety of CAD structures. CAD exposure upregulates the expression of CAD-metabolizing cytochrome P450s and PGP-13, which acts partially redundantly with its paralog PGP-14 to export CAD metabolites (Fig 9). Without an intact CAD defense pathway, the animal is less able to protect itself against the effects of CAD accumulation. Given the unclear orthology between components of this pathway and human homologs, how these insights will translate to CAD pharmacokinetics in humans remains to be determined. However, the findings indicate that animals can indeed mount a robust and coordinated defense against CADs that have phospholipidosis-inducing potential.

Evidently, PGP-13 acts as a multidrug CAD metabolite efflux pump but can also weakly compensate for the loss of PGP-14, which likely exports phospholipids [25] (Fig 9). The relationship between PGP-13 and PGP-14 is remarkably similar to that of human MDR1 (aka ABCB1) and ABCB4; the MDR1 multidrug export pump is also capable of exporting phospholipids but less effectively than ABCB4 [68,69]. Furthermore, PGP-13 and PGP-14 directly neighbor one another in the genome, as do MDR1 and ABCB4 (S2 Fig). Whether the conservation of the physical relationship between the paralogous pairs is coincidental or functional remains to be determined.

One goal of this study was to determine whether small molecules could be used to suppress the phenotypes associated with the loss-of-function of pgp-14, which has the potential to serve as a model for ABCB4 deficiency in humans. The only drugs recovered from our screen for pgp-14(0) suppressors were the CADs, which as described above, can rescue the loss of PGP-14 by upregulating the expression of its paralog PGP-13. Whether CAD treatment in humans might induce the expression of the ABCB4 paralog ABCB1 or some other phospholipid efflux pump and provide some measure of relief for PFIC3 patients with ABCB4 hypomorphic mutations is an open question. That ABCB1 is drug-inducible in hepatocytes [70,71] provides some encouragement for the hypothesis, but the complications surrounding the use of phospholipidosis-inducing drugs are a potential confounding issue. Investigating whether CADs can suppress the phenotypes of ABCB4 deficiency in mdr2 mutant mice [72] would be one step towards testing this hypothesis.

Many more Spectrum library small molecules induce PGP-13 expression than rescue the pgp-14(0) null mutant phenotype. Three reasons may explain this discrepancy. First, the temporal kinetics of CAD rescue of pgp-14(0) is slow. We observe only small crystals after multiple days of coincubation of the mutant with the CADs plus a crystalizing molecule. By contrast, we see crystals within minutes of coincubation of a crystalizing molecule with wild type. Hence, a CAD that may be a slightly weaker PGP-13 inducer might be missed in the pgp-14(0) suppressor screen. Second, some CADs that induce PGP-13 expression are toxic and confound pgp-14(0) suppression. Finally, many CADs that we identified as pgp-13 inducers elicit expression more robustly in the intestine as opposed to the pharynx, which will reduce their ability to suppress the pgp-14(0) phenotype, which is restricted to the pharynx.

It remains an open question as to whether CADs interact directly with the NHR-70 and NHR-107 nuclear receptors to induce a transcriptional response. Given the extensive diversity of CADs that induce the pathway, this model would require that the ligand binding domains of NHR-70 and NHR-107 be equivalently plastic in their interaction with ligands as the P450s that they regulate. Alternatively, the nuclear receptors may be responding to an endogenous metabolite generated as a secondary consequence of CAD accumulation. Using our LCMS protocol that we optimized for small molecule detection, we did not observe a mass that reproducibly accumulates in worms treated with different CADs relative to controls. However, it remains possible and perhaps even likely that endogenous metabolites that accumulate in CAD-treated worms are not captured by our current methodology. Future studies may shed light on this issue.

We demonstrated that the expression of CYP-35B family members 35B1, 35B2 and 35B3 is induced by CADs. Ethanol, caffeine, ethidium bromide, PCB52, polycyclic aromatic hydrocarbons, and several pesticides, none of which are CADs, are also known to induce CYP-35B expression [48,73]. Five of these molecules (caffeine, chlorpyrifos, dichlorvos, rotenone, and thiabendazole) are included in the Spectrum small molecule library and were all negative for the induction of pgp-13 expression in our screen. Hence, the CYP-35B family members must have multiple distinct regulatory inputs in response to different xenobiotics.

We have also demonstrated that CYP-35B1 and 35B2 are necessary for CAD metabolism in worms and that all three P450s are sufficient for this activity upon being expressed in yeast. Metabolic studies of C. elegans P450s are sparse [48], and we could not find evidence within the literature of previously characterized metabolic reactions involving CYP-35B family members. However, our finding that our tool molecule AB1515 is metabolized by CYP-35B2 indicates that at least CYP-35B2’s substrate selectivity extends beyond CADs. Our results also suggest that CYP-35B2’s D311 contributes to a filtering mechanism, likely by increasing the affinity to CADs via electrostatic interactions whilst at the same time reducing interactions to non-CAD molecules like AB1515.

The only CADs that the CYP-35B family failed to metabolize are fipexide and terconazole, despite these molecules being able to robustly induce the marker of CAD defense (see Fig 4a and S1 Data file-tab 1). There may be multiple reasons why these CADs are not metabolized. First, azoles such as terconazole are known to inhibit P450s [74,75], making it possible that terconazole disrupts CYP35B function. Second, fipexide has the lowest pKa (6.09) of the 11 molecules tested and barely makes the CAD cutoff score (see S1 Data file-tab 1). Hence, fipexide is likely a weaker interactor with the CYP-35B family members given that the positive charge of CADs is an important determinant of their metabolism by these P450s. Finally, the shape of fipexide and terconazole sets them apart from the remaining CADs tested in yeast; they harbor the longest atom chains among the 11 molecules. Regardless, it is clear that there is not a one-to-one correspondence in being able to engage the CAD detection machinery and being subject to its defenses.

Obviously, the CAD detoxification pathway in C. elegans did not evolve to defend against anthropogenic CAD molecules that have existed for fewer than 100 years. Indeed, our small-scale screen of 65 Streptomyces revealed multiple strains that induced the CAD defense reporter, suggesting that natural products engage the CAD defense pathway. Furthermore, we know that individual Streptomyces strains generate different types of secondary metabolites depending on their growth conditions [76,77], suggesting that we have only skimmed the surface in identifying strains capable of engaging the C. elegans CAD defense pathway. At the very least, our observations provide a framework for rationalizing the existence of a pathway that was discovered using anthropocentric-derived molecules.

Methods

C. elegans husbandry, screening and microscopy

Unless otherwise noted, the following canonical alleles were used in genetic analyses: The N2 wild type Bristol strain, RB2216 cyp-35B1(ok2998), TM7435 cyp-35B2(tm7435), TM12255 cyp-35B3(tm12255), TM1697 nhr-70(tm1697), TM4512 nhr-107(tm4512), RB894 pgp-13(ok747), RP2960 pgp-14(ok2660), and the RP3495 pgp-13(tr696) pgp-14(ok2660), RB894 pgp-13(ok747), RP2960 pgp-14(ok2660), RP3522 nhr-70(tm1697); pgp-14(ok2660), and RP3523 nhr-107(tm4512); pgp-14(ok2660) double mutants. All strains were cultured using standard techniques [78] at 20 ^oC. The six-day liquid viability assays were performed as described previously [22]. Briefly, E. coli (HB101) was cultured in Lysogeny broth (LB) media at 37 ^oC overnight to make a saturated culture. The cells were collected by centrifugation (3,200 rpm for 20 min) and 2-fold concentrated by resuspending them in 0.5 volume of nematode growth medium (NGM; 3 mg/mL NaCl, 2.5 mg/ml peptone, 5 mg/mL cholesterol, 1 mM CaCl₂, 1 mM MgSO₄ and 25 mM KH₂PO₄). 80 µL of HB101/NGM mix was dispensed into each well of a 96-well plate. Then compounds (typically 300 nL of 10 mM stock) or solvent-only (DMSO) control is added to the 96-well plate using a 96 pinner pinning tool according to the manufacturer’s instruction (V&P Scientific). Approximately 20 first larval stage (L1) animals in 20 µL of M9 buffer (6 mg/mL Na₂HPO₄, 3 mg/mL KH₂PO₄, 5 mg/ml NaCl, 2mM MgSO₄) were then dispensed to each well of the 96-well plate. 96-well plates are incubated in a 20 ^oC shaker at 200 rpm for 6 days. Unless otherwise indicated, all samples are assayed in technical duplicates for every run.

The screen and initial characterization of pgp-14(0) small molecule suppressors

The objective of this screen was to identify small molecule suppressors of the pgp-14(0) mutant phenotype. Normally, wild-type animals succumb to the lethal effects of crystal formation of the wact-190 small molecule within the pharynx cuticle; pgp-14(0) mutants resists these effects [24,25]. Hence, we endeavored to identify small molecules that resensitize pgp-14(0) mutants to the lethality and crystal formation conferred by wact-190, we grew N2 wild-type animals and pgp-14 mutants tr458 and tr477 (both loss-of-function mutations [25]) in liquid format in 96-well plates in duplicate (see the description of the 6-day viability assay above). With a subset of the library, mutants with the pgp-14(ok2660) deletion allele were also screened. N2 was treated with 30 µM of the screening compounds from Microsource’s Spectrum Library (S1 Data File-tab 1), while the pgp-14 mutants were treated with 30 µM of the screening compounds and 15 µM of wact-190 (Chembridge Inc). Our 2015 Spectrum Library contains 2,560 clinically approved drugs and natural products (S1 Data File-tab 1). For the screen, the number of adults (A) and larvae (L) was manually counted in each well for up to 50 animals. Beyond that, wells were given a score of ‘overgrown’ or OG (see OG, A and L scores in S1 Data File). A series of retests, where a ceiling of 50 animals were counted, revealed 14 Spectrum molecules that induced lethality in the pgp-14 null mutant at 30 µM when co-incubated with 15 µM or 30 µM of wact-190 (see column ‘Z’ in S1 Data File-tab 1). None of the Spectrum hits induced obvious phenotype in the absence of wact-190. We also tested whether candidate suppressors could restore wac-190 crystal formation in the pgp-14(0) (but not induce crystal formation in wild type animals) using previously-described methodology [24]. Briefly, 50 synchronized L1 animals were incubated in 30 µM of the Spectrum small molecule hits and either 30 µM of wact-190 or solvent-only control (1% DMSO) for 48 hours. Then, animals were transferred to a microcentrifuge tube and spun down at 2,800 rpm. After aspiration, 2 μL of 50 mM levamisole was added to 10 μL worm pellet. The worm pellets were placed on 3% agarose pad on glass slides and cover-slips were applied. A Leica DMRA 2 compound microscope was used at 40X magnification to investigate the presence of crystals in the anterior pharynxes of the worms. Openlab software was used to take Differential interference contrast (DIC) and birefringence images.

A forward genetic screen for mutants that suppress the ability of CADs to restore pgp-14(0) sensitivity to the lethal effects of wact-190 crystal formation

The forward genetic screen mutagenesis was conducted as described in previous work [79]. Briefly, RP3412 pgp-14(ok2660); trIs104[p1138(pgp-14p::SMS-5B(genomic)::FLAG::mCherry); p1131(pgp-14p::YFP)] parental (P0) worms were mutagenized in ethyl methanesulfonate at a concentration of 50 mM for 4 hours. We included extra copies of functional SMS-5 in the genetic background (i.e. trIs104) because we knew that mutations in sms-5 would suppress the ability of the CADs to resensitize pgp-14(0) to wact-190 crystal formation because SMS-5 is necessary for crystal formation [24]. 100 synchronized and mutagenized F2 (second-generation) L1 worms were dispensed into each well of a 96-well plate, with a total volume of 50 µL containing 30 µM terconazole (CAS# 67915-31-5; Sigma-Aldrich) and 30 µM wact-190 suspended in liquid NGM with HB101 E. coli cell as a food source. The plates incubated over 14 days at 20 °C in a shaking (200 rpm) incubator. Plates were then examined for any wells containing adults and progeny, which was indicative of the presence of a mutant that resists the activity of the CAD’s (terconazole) (i.e., resists the CAD’s ability to resensitize the pgp-14(0) mutant to the lethal effects of wact-190). These mutants were then isolated (ultimately only one line per resistant well) and cloned to enable further characterization.To identify the mutant genes that were likely responsible for CAD-resistance, genomic DNA was prepared, and sequences were analyzed as previously described [80]. In brief, 100 µL of worms were collected in a 15 mL conical tube and washed three times with M9 buffer, followed by a single wash using phosphate-buffered saline (PBS). Excess media was aspirated without perturbing the worms, followed by flash freezing via liquid nitrogen and processing using the QIAGEN DNeasy Blood and Tissue Kit (Cat # 69504). DNA library construction and sequencing were conducted via the services of the Centre for Applied Genomics (Toronto) services. In brief, the 150-bp paired-end sequencing reads were examined for quality via FastQC [81], and Trim Galore! was used to cut off low-quality reads as well as adapter sequences. The processed reads were then aligned via the BWA-mem tool to the C. elegans N2 reference genome – release WS282 [82]. Sequence alignments were sorted by coordinate order, and duplicate reads were marked and removed using Picard tools (https://broadinstitute.github.io/picard/). Variants were called and consolidated utilizing GATK’s HaplotypeCaller and CombineGVCFs [83]. Variants were then filtered on the following criteria to ensure high-confidence variant calls: Fisher Strand (FS) > 60.0, Mapping Quality (MQ) < 40.0, Mapping Quality Rank Sum (MQRankSum) <−12.5, Qual By Depth (QD) < 2.0, Quality Score (QUAL) < 30.0, Read Position Rank Sum (ReadPosRankSum) <−20.0 or <−8.0, and Symmetric Odds Ratio (SOR) > 3.0. Variants were subsequently annotated using Annovar to generate a list of exonic variants, from which candidate genes were identified [84].

RNA extraction, cDNA synthesis and quantitative real-time PCR

Synchronized L1 worms were treated with either vehicle (1% DMSO) or small molecule in 1% DMSO for 96 hours in liquid NGM with HB101 E. coli as a food source at a final concentration of two worms per microlitre in a total of 8 mL. RNA extraction was performed as described in [85]. RNA within the aqueous portion of TRIzol (Invitrogen) was purified utilizing an Aurum Total RNA Mini Kit (Bio-Rad), and treated with TURBO DNase (Thermo Fisher) to remove genomic DNA. cDNA synthesis was conducted from purified RNA using the Applied Biosystems High-Capacity cDNA Reverse Transcription Kit (cat#4368814). Primers for targets were designed to span at least 1 exon-exon junction using NCBI Primer Blast [86]. SsoAdvanced Universal SYBR Green Supermix (Bio-Rad) was utilized with 5 ng of cDNA and primers at 500 nM of each primer to conduct q-RTPCR on the QuantStudio 6 Flex Real-Time PCR System. The data was analyzed utilizing the comparative C_T method normalized to the common standard rpl-33 [87,88]. RPL-33: F 5′-GGCCCACAACAAGACTTTGAAAACC-‘3; R 5′-ATGGGTAGAGAAGCACGCGG-3’PGP-14: F 5′-CACGACCAAATTGAACGACAGTATG-3′; R 5′-TGTAGTTGAAGTCATTTGTCTGGC-3’PGP-13: F 5′-TTTCGCTACTGCAATTGTTCCGC-3′; R 5′-ACACATTTGTCATTCGACCCGAG-3′

CYP-35B1: F: 5′-GCTGAACACGAGATGTGCCG-3′; R: 5′-ACGTTTTCCGACGAGCAGAGA– 3′

CYP-35B2: 2. F: 5′-GGGCCAGTATCTCTTCCACTT-3′; R: 5′-GTGAAAACATCTCCATAAATCTGTC-3′

CYP-35B3: F: 5′-GGATGAGCTGAATGCAAGATGTGC-3′; R: 5′-CACTGCCAACAGTTAGGTCAAAG-3′

CROR-1: F: 5′-GTGGGCAACATGCGCAAC-3′; R: 5′-TTGTAGATGTTTTCCACCGACTC-3′

CROR-2: F: 5′-GGGCGTTTGAGGAGATGATG-3′; R: 5′-CTTACGATCAGCCAAGAACGC-3′

Construction and analyses of pgp-13 transcriptional and translational reporters

The pPRLT1187 pgp-13p::YFP transcriptional reporter was constructed by PCR amplifying 1.5 kb of the pgp-13 sequence 5′ to the ATG using the 5′ primer 5′-aagtcTCTAGAattatgaaagaaatccgagaggtctg-3′ and the 3′ primer 5′-gtactgACCGGTcagcggttcgttttatacaattgtg-3′ from wild type (N2) genomic DNA, and then cloning the PCR product into a modified construct from the Andy Fire vector collection kit (L4666). Animals were made transgenic for this construct through microinjection into wild type (N2) animals and then chromosomally integrated using the ultraviolet radiation technique [89], resulting in trIs114 [pgp-13p::YFP; p382(myo-2p::mCherry)].

The mNeonGreen::PGP-13 N-terminal fusion translational reporter was built with the help of In vivo Biosystems Inc by using CRISPR to insert mNeonGreen coding sequence, containing three synthetic introns, in-frame with a 30 bp linker (GGTTCGGGCTCGGGTTCTGGAAGTGGTAGT) immediately 5′ to the initiator methionine codon of the pgp-13 gene, resulting in the integrated transgene trIs115. The sgRNA guides included the 5′ sgRNA GATGCCACCACAAGAAAACA and the 3′ sgRNA TTTTCTTGTGGTGGCATCAG. Importantly, the trIs115 mNeonGreen::PGP-13 fusion gene was created in the background of the pgp-14(ok2660) deletion allele. We used this background to determine whether the resulting fusion protein was functional; the only known phenotype at the time of reporter construction was that pgp-13’s loss-of-function is capable of suppressing the ability of CADs to resensitize pgp-14(0)’s resistance to the lethal effects of wact-190 crystals. Furthermore, pgp-13 and pgp-14 are neighboring genes and recovering a pgp-13 knock-in recombinant with a neighboring pgp-14(0) mutant would be highly unlikely. Hence the need for the creation of the mNeonGreen knock-in into pgp-13 in the background of a pgp-14(0) mutant. The key result that demonstrated that mNeonGreen::PGP-13 is functional is that the RP3561 strain (which encodes the mNeonGreen::PGP-13 fusion in the background of the pgp- 14(ok2660) null) yields crystals in the background of the CAD terconazole (which suppresses pgp-14(0) because it upregulates pgp-13 expression). This outcome is identical to that of wild-type pgp-13 in the background of pgp-14(0). If pgp-13 had lost function through the fusion of mNeonGreen, terconazole would not have been able to resensitize pgp-14(0) to wact-190 crystals. See S6 Fig for details.

Analyses of pgp-13 temporal expression and dose-response

The temporal and dose-response analyses of pgp-13p::YFP reporter expression were done in 96-well plate format with either a wild-type background (strain RP3507) and/or the backgrounds of the nhr-70(tm1697) and nhr-107(tm4512) null mutants (strains RP3518 and RP3516, respectively). For each experiment, 3 independent biological repeats were performed with ~150 worms per well. For the dose-response analyses, at least 12 technical repeats for each data point were prepared for each trial. For the temporal analysis at least 36 technical repeats were prepared for each trial. For each, ~150 synchronized L1 worms were added to each well. Compound was added to the wells using a 96 pin pinner to a final concentration of 0, 15 or 30 µM, with a final DMSO concentration of 1%. Plates were incubated at 20 ^oC for 48 hours on a shaking incubator. YFP and RFP signals were measured with a CLARIOstar fluorimeter (BMG Labtech Inc). Only YFP signal was considered in the analyses because fluctuations in basal RFP expression threw YFP::RFP ratio wildly, while YFP-only signal yielded reproducible signal patterns.

Analyses of pgp-13 Localization

Characterization of the expression pattern of the mNeonGreen::PGP-13 fusion reporter (RP3561 trIs115 [mNeonGreen::PGP-13; pgp-14(ok2660)]) in the head of the animal was conducted by incubating ~100 synchronized L1 worms in either solvent control or 30 µM of terconazole for 48 hours in liquid NGM with HB101 E. coli cells (optical density (OD600) ≥ 1.2) as a food source. Following the incubation on a tube rotator, worms were given the chitin stain calcofluor white at a final concentration of 104 µM for 1 hour and then imaged. Intestinal expression was investigated by incubating ~100 synchronized L1 RP3561 animals in liquid NGM with HB101 E. coli cell as a food source with solvent or 30 µM of terconazole for 72 hours. Following the incubation, worms were incubated in 1 mg/mL Texas Red-dextran (40,000 MW) for 1.5 hours. Imaging was performed using a Leica DM6B microscope equipped with 63x and 20x objectives.

Cell-based assay testing CAD interaction with NHRs

We built the plasmids expressing chimeric GAL4 DBD fusions pPRLT1204 (FLAG-6His-GAL4 DBD–NHR-70 LBD (Ligand Binding Domain)) and pPRLT1205 (FLAG-6His-GAL4 DBD–NHR-107 LBD) using the pcDNA3.1 FSH GAL4-hPPARa backbone (a gift from the Carolyn Cummins Lab). The parental vector was digested with NheI and PmeI to excise the hPPARa coding region, and the remaining ~6.0 kb backbone fragment containing the CMV promoter and the N-terminal tag cassette (3xFLAG, Strep-Tag II, 6xHIS) was fused in-frame to the yeast GAL4 DBD. Codon-optimized sequences encoding the hinge and ligand-binding domains but lacking the N-terminal DBD of NHR-70 and NHR-107 flanked by terminal 5′ NheI and 3′ PmeI restriction sites were synthesized by Twist Bioscience. Synthetic fragments were digested with NheI and PmeI to yield a 941 bp insert for nhr-70 and an 866 bp insert for nhr-107. Each digested insert was ligated into the linearized pcDNA3.1 FSH GAL4 backbone. Construct integrity and in-frame fusion of the resulting open reading frames (encoding a 488-amino-acid fusion protein for pPRLT1204 and a 463-amino-acid fusion protein for pPRLT1205) were confirmed by sequencing.

HepG2 cells were maintained in RPMI 1640 media (Gibco, 11870593) containing 10% fetal bovine serum (FBS) and 1% Penicillin/Streptomycin (P/S) at 37 °C in a humidified atmosphere of 5% CO₂. For transfection assays, HepG2 cells were cultured into a clear bottom white 96-well plate in RPMI 1640 media (Gibco, 11870593) containing 10% FBS. After 24 hrs incubation, cells were transfected with the plasmid DNA using Lipomaster 3000 (Vazyme, TL301−01). The plasmid DNA contained 50 ng of UAS-luciferase reporter, 15 ng of GAL4-NHR70, or GAL4-NHR107, or GAL4-hPXR, 20 ng of the Renilla reniformis luciferase transfection control plasmid, and 65 ng of pGEM filler vector. Six hours after transfection, cells were treated with vehicle (DMSO) or various concentrations of compounds. The following day (16 hours after ligand addition), luminescence activity in the cell lysate was measured using Dual-Luciferase Reporter Assay (Promega, Madison, WI, USA).

To assay for the expression of the NHR(LBD)::GAL4(DBD) fusions, total protein was extracted using Passive lysis buffer (E1941, Promega) containing protease inhibitor cocktail (Thermo Fisher Scientific, 78430) for 30 min on ice. Protein was quantified by BCA, then 2 × loading buffer was added, incubated at 95 °C for 5 min and separated on 12% SDS gels. Samples were wet-transferred onto nitrocellulose membrane (GE Healthcare, Chicago, Illinois, USA) at 200 mA for 60 min. Blocking was performed for 15 min at room temperature in blocking solution (5% skim milk in PBS with 0.05% Tween-20 (PBST)), incubations with the primary and secondary antibodies were performed as suggested by the manufacturer. Amersham ECL Western Blotting Detection Reagent (Cytiva, RPN2109) was used for signal detection. Blots were exposed to an X-ray film (Thermo Fisher Scientific, 34091) for 30–60 s. Uncropped blots are available in the provided S1 Data file. Monoclonal Anti-FLAG M2 antibody (F3165, 1:1000) was purchased from Sigma. Polyclonal β-Actin antibody (4967, 1:1000) was purchased from Cell Signaling Technology.

Computational modeling of CAD interactions with NHR-70 and NHR-107

The C. elegans nuclear receptors NHR-70 and NHR-107 were each co-folded with four CADs: terconazole, amiodarone, chlorprothixene, and clomipramine. Both charged and uncharged protonation states were modeled for every drug. Structure prediction used AlphaFold3 (v3.0.0, with model parameters AlphaFold-beta-20231127) [90] and Boltz-2 [91].

In these docking experiments, we also included the human PXR nuclear receptor, which is well characterized to bind the PXR agonists SR12813 and hyperforin as positive controls. Of note, there is some weak evidence in the literature that PXR mediates a response to some CADs, but direct physical interaction remains an open question. We also included the human VDR nuclear receptor as a negative control because it has not been reported to interact with the SR12813 and hyperforin ligands.

For all predictions, the LBD sequence was used in addition to the N-terminal linker region as a buffer region. The constructs corresponded to residues 82–389 of NHR-70, 75–357 of NHR-107, 137–430 of PXR, and 121–423 of VDR, taken from UniProt entries G5EEL3, G5EEH2, O75469, and P11473, respectively. Each construct begins immediately after the DBD; the NHR-70 and NHR-107 constructs extend to the native C terminus, while the PXR and VDR constructs stop four residues short of it. Charged and uncharged forms of each CAD were supplied as explicit SMILES. SR12813 and hyperforin were modeled in a single neutral form. Each receptor was co-folded with all 10 ligand forms, giving 40 runs per tool and 80 predicted complexes in total.

Each AF3 run used five model seeds and generated five diffusion samples per seed, giving 25 structures. Alignments came from the built-in AF3 data pipeline, and template search was enabled. AF3 identified four templates for each receptor. Each Boltz-2 run generated five diffusion samples over five recycling steps. The Boltz-2 binding-affinity module was enabled for every run. The top-ranked structure was selected from each run, using the ranking score for AF3 and the iPTM score for Boltz-2.

GNINA v1.3.2 was then used to re-dock each ligand into its predicted receptor structure, generating Vina and CNN-based affinity and pose scores across all receptor–ligand–protonation combinations [92,93]. The receptor was held rigid throughout. The ligand pose from the co-folded complex was not carried into docking; a fresh three-dimensional conformer was generated from the canonical SMILES using Open Babel 3.1.1 [94]. The co-folded pose was used only to define the search box, with 4 Å of padding. Docking used an exhaustiveness of 16 and nine poses were retained. The top-ranked pose was used for all reported scores.

GNINA CNN affinity approximates −log10(Kd), where a score of 5 corresponds to ~10 µM, 6 to ~1 µM, 7 to ~100 nM, and 8 to ~10 nM predicted binding affinity. Each unit increase reflects a 10-fold improvement. CNN pose score reflects the predicted probability that a docked pose is within 2 Å root mean square deviation (RMSD) of the true binding mode, where higher scores indicate greater geometric confidence. We interpret these scores as relative rankings across receptors and ligands rather than absolute affinity predictions and compare them within each tool to assess whether NHR-70 and NHR-107 score similarly to PXR or VDR.

To assess the validity of the structures, RMSD was calculated between AF3 and Boltz-2 predictions aligned on the LBD Cα atoms only, excluding the flexible N-terminal linker. The LBD was taken to begin at residues 131 (NHR-70), 107 (NHR-107), 144 (PXR), and 125 (VDR) in UniProt numbering. Superposition was performed by least squares using Biopython [95]. For PXR and VDR the comparison was restricted further to those LBD residues also resolved in the corresponding crystal structure, which were identified by pairwise sequence alignment. This prevented flexible C-terminal extensions absent from the experimental models from contributing. The crystal structures used were 2O9I for PXR [96] and 1DB1 for VDR [97]. The same procedure was used to compare each prediction with its crystal structure, and the resulting values are reported in S5 Fig.

Computational modeling of CAD interactions with cytochrome P450s

As experimentally resolved structures are not available for the complexes of interest, Boltz-1 [55], an open source implementation of AlphaFold 3 [90], was used to co-fold the enzyme, HEME, and CAD. Thirty structures were generated and the top ranked (by confidence score) model was selected for molecular dynamics simulation. All confidence, pTM, ligand ipTM, and complex pLDDT scores were better than 0.92. For simulation with AMBER [98], HEME parameters for a ferric high-spin state were adapted from [99], and a covalent linkage was made between HEME and its coordinating cysteine, C444. CADs were parameterized using antechamber and GAFF [100]. The uncharged version of each CAD was setup using the identical pose as the charged version, but with a different protonation state. The systems were then solvated and neutralized with sodium ions using tleap. Initial equilibration consisted of two minimization steps (the first with protein restraints), followed by 1 ns of restrained NVT dynamics (heating from 0 to 300 K), and 1 ns of unrestrained NPT dynamics at 300 K and 1 atm. All simulations used the AMBER ff15ipq force field [101] and TIP3P water model, selected for its consistency with NMR-derived structural data [102]. Three independent 100-ns production runs were performed for each system. Analysis was performed using MDAnalysis [103]. Representative ligand poses for each simulation (i.e., Fig 6) were selected by identifying the medoid of the largest cluster, where clusters were determined using RMSD as a distance metric for the fcluster method of scipy with a radius of 2.0 angstroms and average linkage.

A screen of the spectrum library for small molecules that induce pgp-13 expression

The Spectrum Screen for Small Molecule Inducers of pgp-13 Expression: To identify small molecule inducers of pgp-13 expression, we screened the RP3507 trIs114 [pgp-13p::YFP; p382(myo-2p::mCherry)] animals against 2,400 molecules of the Spectrum Library (Microsource Inc) (S1 Data File-tab 1). We deposited synchronized ~150 L1s into the wells of a 96-well plate along with 30 µM (final concentration) of compound in 1% DMSO and E. coli (an OD600 of ~1.5) as a food source and incubated at 20 °C for 48 hours on a shaking incubator. Duplicate plates were set up and analyzed independently. The fluorescence signal from the pharynx and intestines of the worms was scored independently and manually binned into one of seven categories: 6, saturated; 5, very bright; 4, bright; 3 medium bright; 2 obvious signal, but not bright; 1 signal just above background; 0 no obvious signal. Given that the correlation between the two manual counts was +0.96 (S1 Data File), we were confident in this scoring system. For the purposes of this analysis, we considered any molecule that resulted in an average score in the pharynx and/or the intestine of 2.5 or more to be a consistent inducer of pgp-13 reporter expression.

Generating a structural similarity map

Molecule scaffolds were generated from Chemical SMILES of the Spectrum library were provided by the vendor (Microsource Inc). SMILES were edited to remove salts, and then reduced to their core scaffolds using the Strip-it command line tool version 1.0.2 (RINGS_WITH_LINKERS_1 scaffold definition). The resulting scaffold molecules were analyzed for pairwise structural similarity by calculating the Tanimoto coefficient of shared FP2 fingerprints using OpenBabel (http://openbabel.org) and a structural similarlity score of 1 (identical) was used to generate the structural similarity map, which was visualized using Cytoscape version 3.10.3 [104].

Physicochemical properties of molecules were downloaded from SwissADME [43] and investigated for differences among different groups of molecules.

Establishing a predictive model of phospholipidosis(PLD)-inducing CADs

We constructed several predictive models of PLD-inducing CAD molecules based on the published model [11], each using a different method that predicts hydrophobicity of a small molecule (i.e., cLogP). We estimated the pKa of the most basic group of each molecule by using the MolGpKa API [105]. The original MolGpKa API python script was modified to allow iteration through lists of smiles, and Microsoft Excel (v 16.43) was then used to deconvolute the output. We evaluated the different models against a suit of 112 gold standard CAD-inducing molecules identified from the literature (see S1 Data File-tab 1), and then evaluated each based on their capture rate of the gold standards using the hypergeometric p value calculation. The ‘PLD-inducing-CAD’ predictive model that captured the highest fraction of gold standards as a factor of the number of predicted PLD-inducing CADs (6.1 fold more than expected by chance alone; hypergeometric p-value E-55) was produced using the MLogP model (SwissADME) (S1 Data File-tab 1). We therefore proceeded with our analyses using this model.

Spectrum MiniLibrary analyses

From the 293 Spectrum Library hits that induce pgp-13p::YFP expression, we selected those that induced robust expression in the pharynx (an average score of 3 or more) and were plentiful within our cherry-picking stock plates), leaving a total of 195 hits to analyze in more detail (bolded nodes in Fig 3a). These 195 molecules were re-arrayed into new 96 well plates and screened with the strains RP3507 pgp-13p::YFP, RP3516 pgp-13p::YFP; nhr-107(0), and RP3518 pgp-13p::YFP; nhr-70(0) as described above with synchronized L1s grown for 48 hours at 20 ^oC in a shaking incubator. The wild-type (RP3507) strain was screened two independent times (N = 2) with two technical repeats each (n = 2); the two mutant strains were each screened four independent times (N = 4) with two technical repeats each (n = 2). Although we collected and reported data for expression in both the pharynx and intestine, we focused only on the pharynx because of the low background signal.

Sample preparation of C. elegans for mass spectrometry analyses

30,000 synchronized L1 C. elegans were placed into 1.5 mL microcentrifuge tubes in 495 µL M9 buffer per sample. Compounds were added to achieve the desired final concentration, where DMSO final concentration is at 1% (v/v). Samples were incubated on a tube rotator to ensure proper at room temperature for 24 hours. Following the incubation, worms were transferred to 96-well filter plates (Pall P/N 8684) and the buffer was removed via vacuum filtration. Worms were then washed with 600 µL of M9 buffer and then resuspended in 50 µL of water and immediately frozen at −80 °C.

Frozen worm pellets were thawed and resuspended in 400 µL of butanol (BuOH) with the addition of 100 µL of 0.5 mm glass beads. Samples were lysed using a bead-beating device (Bead Ruptor 24, Omni International, Kennesaw, GA, USA), followed by centrifugation at 14,000 rpm for 10 min. 350 µL of the BuOH extract was then transferred to a new Eppendorf Safe-lock tube (cat. #022363204), and 350 µL of methanol (MeOH) is added to the original vial which was then bead-beaten a second time. After centrifugation at 14,000 rpm for 10 min, 350 µL of MeOH extract was combined with the BuOH extract in the new tube. The combined extracts were then dried at 60°C utilizing an Eppendorf Vacufuge (V-AL a pre-set mode for drying alcohol-based solutions). Dried samples were then stored at −20 °C until further processing.

The dried extracts of worm lysate and buffer (and yeast lysate and buffer-see below) were resuspended using 100 µL of a 50:50 mixture of LCMS grade acetonitrile and water. Samples were then vortexed for at least 30 s, followed by sonication for 15 min and centrifugation at 14,000 rpm for 15 min. For LCMS analysis, 25 µL of the supernatant was transferred into an amber vial insert and the remainder was stored at −20 °C.

Expression of wild-type and mutant C. elegans CYP-35B P450s in the yeast S. cerevisiae

The C. elegans cyp-35b1, cyp-35b2, and cyp-35b3 cDNA sequences were obtained from WormBase (version WS296) and were codon-optimized for yeast expression. The codon-optimized ORFs were synthesized by Twist Biosciences and placed into the SpeI and HindIII restriction sites of the ATCC p416 GAL1 expression vector (URA3 selection marker; CEN6/ARSH4 origin of replication) [106]. The S. cerevisiae stain RP4400 (S228c background) MATa his3Δ leu2Δ ura3Δ lys2Δ pdr5Δ snq2Δ yor1Δ can1::pGAL1-Ce emb-8 was transformed using the lithium acetate/single stranded carrier DNA/PEG method [107] with the cyp-35B1, cyp-35B2, cyp-35B3, and empty vector negative control plasmids and were selected for on SD (-)URA medium. For the mutants, codon-optimized DNA fragments encoding cyp-35B2 variants carrying the D311N or D311G substitutions were synthesized by Twist Bioscience with 5′ and 3′ adaptors: 5′- CAAAAAATTGTTAATATACCTCTATACTTTAACGTCAAGGAGAAAAAACCCCGGATTCTAGAACTAGT-3′, and 5′- AAGCTTATCGATACCGTCGACCTCGAGTCATGTAATTAGTTATGTCACGCTTACATTCACGCCCTCCCCCCACATC-3′, respectively, whose sequence corresponds to the multi-cloning site of the aforementioned ATCC p416 GAL1 expression vector. The fragments were co-transformed along with linearized p416 at the multicloning site into the S. cerevisiae BY4741 (MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0) strain and selected for on SD (-) URA media. The plasmids were then extracted and transformed en mass into E. coli DH5α competent cells and plated on LB plates with 100 µg/ml ampicillin. E.coli colonies were miniprepped and sequenced (Plasmidsaurus Inc) to determine which were correct. The resulting correct plasmids were individually transformed into aforementioned S. cerevisiae strain RP4400 as described above.

Sample preparation of S. cerevisiae for mass spectrometry analyses

Yeast strains were grown overnight at 30 °C in 5 mL of SD-URA medium with 2% galactose. The following morning, cultures were pelleted by centrifugation for 3 min at 3,000 rpm. The pellet was then washed with 12 mL of autoclaved Milli-Q water, followed by centrifugation under the aforementioned conditions. The cell pellets were then resuspended with 5 mL of autoclaved Milli-Q water, and the OD600 of each culture was calculated using a 1:50 culture dilution in a cuvette. The volume is then adjusted to reach a final OD600 of 5. Following this, 5 µl of drug in solvent or solvent-only is added to 495 µL of the culture for each sample. Samples were then incubated at 30 °C for 4 hours in a rotary shaker. Following incubation, samples were pulse-spun and transferred to 96-well filter plates (Pall P/N 8684) and the buffer was removed via vacuum filtration into a collection plate. Incubation media collection plates were then spun at 1,000 rpm for 1 min to remove residual droplets on the side walls. Cells were then resuspended in 50 µL of autoclaved Milli-Q water and immediately frozen at −80 °C in Safe-Lock Eppendorf tubes, whereas the incubation buffer was transferred to Safe-lock Eppendorf tubes and frozen.

Frozen yeast pellets were thawed and resuspended in 800 µL of a 50:50 mixture of BuOH and MeOH alongside 100 µL of 0.5 mm glass beads. Samples were lysed using a bead-beating device as described above, following by centrifugation at 14,000 rpm for 10 min. From each sample, 800 µL of the BuOH:MeOH extract was transferred into an Eppendorf Safe-Lock microtube for subsequent drying.

To dry the samples, the incubation buffer is thawed, and tubes are placed inside an Eppendorf vacufuge ensuring balance after the device has been preheated for 15 min at 60 °C. Samples are dried for ~2–3 hours at 60 °C (V-AL setting). When samples are dried, samples are stored at −20 °C until sample resuspension. Samples are resuspended as described above for the worm samples.

Liquid chromatography–quadrupole time‑of‑flight mass spectrometry

Agilent 1260 Infinity II with 6545 LC-MS/QTOF mass spectrometer was used to analyze samples in positive ionization mode with Dual AJS electrospray ionization (ESI) equipped with Agilent ZORBAX Eclipse Plus C18 column (2.1 x 50 mm, 1.8-μm particles) and ZORBAX Eclipse Plus C18 guard column (2.1 x 5 mm, 1.8-μm particles). LC parameters used were: injection volume 5 μL with 5 μL needle wash with sample prior to injection. The autosampler chamber temperature is maintained at 4 °C and the column oven temperature is maintained at 40 °C. Mass spectrometry parameters were: gas temperature 320 °C, drying gas flow 8 L/min, nebulizer 35 psi, sheath gas 350 °C at 11 liters per minute, VCap 3,500 V, Nozzle voltage 1,000 V, fragmentor 125 V, skimmer 65 V. The solvent gradient with the flow of 0.5 ml per minute started with 95% mobile phase A (Optima LC/MS H2O+0.1% Formic Acid, Fisher Chemical P/N LS118-4) and 5% mobile phase B (Optima LC/MS Acetonitrile+0.1% Formic Acid, Fisher Chemical P/N LS120-4), increased linearly to 50% B at 10 min, followed by an increase to 100%B at 27.50 min. The post-run time was 2 min (instrument conditioning at 95% mobile phase A). The raw data were analyzed using Agilent MassHunter Qualitative Analysis 12.0 and MassHunter Profinder 10.0 to find differences in samples of different strains and conditions.

CAD sensitivity analysis

Worms were incubated with CADs in NGM liquid media supplemented with E. coli HB101 (OD600 ≥ 1.2) in a shaking incubator (200 rpm) at 20 °C for 3 days prior to quantification. Twenty worms were deposited per well (two wells per condition and genotype) in a total volume of 50 µL, where live moving worms were quantified up to a ceiling of 50 after a 6 day incubation period. Data were normalized to each strain’s respective solvent control.

Analysis of gut granules using cDCFDA

Synchronized L1 animals were grown in NGM liquid media supplemented with E. coli HB101 (OD600 ≥ 1.2) for 24 hours at 20 °C on a tube rotator prior to addition of CADs, which were then added at the indicated final concentrations and incubated for another 24 hours. Following this, 5-(and-6)-carboxy-2′,7′-dichlorofluorescein diacetate (cDCFDA) (Cayman Chemicals Item #34692) was added to a final concentration of 50 µM and incubated for 2.5 hours. After the incubation, worms were washed seven times with M9 media, anesthetized with 250 mM sodium azide, and mounted on agarose pads. Imaging was performed using a Leica DM6B microscope with a 20x objective. Puncta were quantified using ImageJ. The GFP channel images capturing cDCF signal were processed with a Difference of Gaussians filter (σ₁ = 1.0, σ₂ = 1.5) to isolate punctate structures. Images were then thresholded using the RenyiEntropy method and segmented using the Watershed algorithm. Particles were then subsequently quantified in the intestine using size (0.11–40 µM²) and circularity (0.45–1.00) parameters.

Use of the pHTomato reporter

Gravid XW19180 adult animals were bleached to isolate embryos, which were subsequently incubated in M9 buffer for 24 hours to generate a synchronized population of L1 animals. The synchronized L1 larvae were transferred into liquid NGM containing Escherichia coli HB101, as described above, and cultured at 20 °C. Upon reaching the L4 stage, animals expressing PhsNUC-1::pHTomato were subjected to a single heat shock treatment at 33 °C for 30 min [64]. Following heat shock, drug treatments were initiated by adding drugs to the liquid culture at the indicated concentrations, and the animals were incubated at 20 °C for 24 in the indicated treatment groups. Following the 24-hour incubation, adults (1 day post-L4) were anesthetized with 250 mM sodium azide, and mounted on agarose pads. For each treatment group, Z-stacks of five worms were captured across three independent biological replicates, and the resulting Z-series were collapsed using a sum of slices tool in ImageJ/Fiji software. To quantify the average pHTomato fluorescence intensity per lysosome, individual worms were isolated on the maximum intensity projections. Lysosomal puncta boundaries were segmented using an Otsu Dark threshold and segmented with a watershed algorithm. A particle filter restricted data collection to objects displaying a circularity index between 0.40 and 1.00.

The screening and characterization of actinomyces bacterial strains for CAD reporter induction

Sixty-five Cubist Actinomyces strains were plated in the wells of 12 well plate with solid MYM agar at 30 °C for 7 days in duplicate. ~1,000 synchronized L1 RP3507 trIs114 [pgp-13p::YFP; p382(myo-2p::mCherry)] CAD reporter worms were deposited into each well. Twenty-four hours later wells were inspected for worms expressing YFP using an MZFLIII Leica epifluorescent dissection microscope. Subsequent detailed analysis was done by growing ‘hit’ strains plus a nonhit control on solid 6 cm MYM agar plates at 30°C for three days. 500 synchronized L1 RP3507 animals were then transferred to the bacterial lawns and incubated for 24 hours. Following co-incubation, worms were manually picked, anesthetized using 250 mM sodium azide, and mounted on agarose pads. Imaging of reporter signals in the head of animals was performed on a Leica DM6B microscope with a 63x objective, and quantified using ImageJ.

Supporting information

S1 Fig. A schematic of mutations recovered in our forward genetic screen. A.

A topology schematic of PGP-13 and mutations identified in our screen with missense in green and nonsense in pink. Previously recovered PGP-14 missense (blue) and nonsense (red) mutations are also shown. B. Topology schematic of NHR-70 nuclear hormone receptor and associated mutations. C. Topology schematic of NHR-107 nuclear hormone receptor and associated mutations. In B and C, green stars indicate a missense and pink star indicate nonsense.

(TIFF)

pbio.3004006.s001.tiff^{(15.3MB, tiff)}

S2 Fig. A comparison of the relative positions of relevant PGPs in C. elegans and humans.

The screenshot of the genomic position of pgp-13 and pgp-14 on chromosome X is from WormBase (version WS297) [108]. The screenshot of the genomic position of ABCB4 and ABCB1 on chromosome 7 is from the UCSC Genome Browser (accessed Nov 12, 2025) [109].

(TIFF)

pbio.3004006.s002.tiff^{(15.3MB, tiff)}

S3 Fig. PGP-13, NHR-70, and NHR-107 are necessary for CAD resensitization of pgp-14(0) to wact-190 crystal formation. A.

Images of the heads of animals of the indicated genotype (to the left) and treatment (at the top of the columns). Treatment time was 48 hours. The imaging channel for each animal is indicated at the bottom of the columns; DIC, differential interference contrast; BIRE, birefringence. Anterior is up. The scale in the top left applies to all images. The relevant strains/genotypes are N2 wild type; RP2960 pgp-14(ok2660); RB894 pgp-13(ok747); RP3495 pgp-13(tr696) pgp-14(ok2660); nhr-70(tm1697), RP3522 nhr-107(tm1697); pgp-14(ok2660); nhr-107(tm4512); and RP3523 nhr-107(tm4512); pgp-14(ok2660). wact-190 crystals are indicated with a pink arrow. The key result that demonstrates each of PGP-13, NHR-70, and NHR-107 are necessary for the CAD (terconazozle) to resensitize the pgp-14(0) null to wact-190 crystal formation are indicated with green box. B. Quantification of the number of animals within the population that exhibited wact-190 crystals in the experiment shown in A. N = 3 independent trials with at least 24 animals counted per sample (n ≥ 24). A two-tailed Student T test was used to determine p values, except where there are absolute differences (all trials for a sample yields zero or 100 percent). See S1 Data for the data underlying the graphs.

(TIFF)

pbio.3004006.s003.tiff^{(15.3MB, tiff)}

S4 Fig. Screening of CADs for NHR-70 and NHR-107 activation.

A-C HepG2 transfected with UAS-luciferase reporter constructs and expression plasmids for (A) Gal4-NHR-70, (B) Gal4-NHR-107, or (C) Gal4-human PXR were treated with medium (10 μM for CADs, 1 μM for SR12813, purple bars) and low (1 μM for CADs, 0.1 μM for SR12813, pink bars) concentrations of various indicated drugs (e.g., Nicardipine, chlorpromazine). Data are presented as fold change in normalized luciferase activity relative to vehicle control. Data are expressed as means ± SD (n = 3 technical replicates). D. Western blot showing the protein expression of NHR-70 and NHR-107 in the transfected cell lines. Cell lysates from NHR-70 and NHR-107-expressing cells were probed with anti-FLAG antibody. Green and orange boxes highlight the NHR-70 and NHR-107 protein bands, respectively, at 56.3 kDa (predicted) and 54.5 kDa (predicted). β-actin is used as a loading control, with molecular weight markers of 58 kD and 50 kD indicated. See S1 Data for the data underlying the graphs.

(TIFF)

pbio.3004006.s004.tiff^{(15.3MB, tiff)}

S5 Fig. Computational modeling of CAD interaction with NHR ligand binding domains.

A-B. Structural validation of AF3 and Boltz-2 predicted receptor structures. A. Mean LBD Cα RMSD of AF3 (solid) and Boltz-2 (hatched) predictions versus PDB crystal structures for PXR (2O9I) and VDR (1DB1). Error bars show the min–max range across the eight CAD–protonation combinations (four CADs × two protonation states); the control ligands are not included. B. Inter-tool LBD Cα RMSD between AF3 and Boltz-2 for all four receptors (n = 8 CAD–protonation combinations per receptor, control ligands excluded); horizontal bars indicate the mean. C. Structural overlay of AF3-predicted (green) and PDB crystal (orange) LBDs for PXR (left) and VDR (right), with SR12813 shown as sticks. D-K. AlphaFold3 co-folding and GNINA re-docking scores across all receptor–ligand combinations. D-F. GNINA CNN affinity (↑ higher = stronger predicted binder). G-I. CNN pose scores (↑ higher = more realistic binding geometry). The two categories are split into charged (D, G), uncharged (E, H), and PXR control ligands SR12813 and hyperforin (F, I). Note that protonation states only differ for the four CADs; SR12813 and hyperforin were modeled in a single neutral form. J and K. AF3 internal confidence metrics including chain pair iPTM (J; ↑ higher = more confident ligand placement) and interface PAE min (K; ↓ lower = less positional uncertainty between receptor and ligand chains), each displaying the best-scoring protonation state per cell (C = charged, U = uncharged). L-S. Boltz-2 co-folding and GNINA re-docking scores across all receptor–ligand combinations. L-N. GNINA CNN affinity (↑ higher = stronger predicted binder). O-Q. CNN pose score (↑ higher = more realistic binding geometry). The two categories are split into charged (L, O), uncharged (M, P), and PXR control ligands SR12813 and hyperforin (N, Q). Note that protonation states only differ for the four CADs; SR12813 and hyperforin were modeled in a single neutral form. R and S. Boltz-2 internal confidence metrics: chain pair iPTM (R; ↑ higher = more confident ligand placement) and interface PAE min (S; ↓ lower = less positional uncertainty between receptor and ligand chains), each displaying the best-scoring protonation state per cell (C = charged, U = uncharged). To assess the validity of the structures, RMSD was calculated between AF3 and Boltz-2 predictions aligned on the LBD Cα atoms only, excluding the flexible N-terminal linker. Both tools accurately reproduced the PXR (2O9I) and VDR (1DB1) crystal structures at sub-angstrom to low-angstrom precision (AF3: 0.76–0.79 Å for PXR, 1.18–1.39 Å for VDR; Boltz-2: 0.83–0.94 Å for PXR, 0.48–0.59 Å for VDR), validating the predicted folds (A). Inter-tool LBD agreement was strong across all receptors, with 30 of 32 receptor-ligand-protonation combinations at or below 2 Å; the two exceptions were NHR-107 with uncharged chlorprothixene (2.13 Å) and uncharged amiodarone (2.12 Å) (B; mean LBD Cα RMSD: NHR-70,0.80 Å, NHR-107,1.65 Å, PXR 0.45 Å, VDR 1.03 Å). For PXR and VDR, the inter-tool RMSD in (B) was computed over residues covered by the respective crystal structure, to exclude flexible C-terminal extensions not resolved in the PDB; for NHR-70 and NHR-107, for which no experimental structure exists, all LBD Cα atoms were used. We had three benchmarks to gauge whether the docking experiments were yielding useful predictions. First, if the CADs physically interact with NHR-70 and NHR-107 AND the computational docking experiments were predictive, we expected that the charged versions of CADs to yield greater affinity for the C. elegans receptors relative to the uncharged versions because the CADs would be largely charged within the cell. In AF3-based docking, the uncharged forms scored nearly always higher (i.e., a predicted greater affinity) than charged forms with NHR-70 and NHR-107 (D and E). Similar results are found using Boltz-2 (L and M). Hence, the computational predictions failed to meet this benchmark. Second, experiments using nhr-70(0) and nhr-107(0) mutants indicated that terconazole, chlorprothixene, and clomipramine depend on both NHR-70 and NHR-107 to drive the CAD defense response. By contrast, amiodarone is much more dependent on NHR-70 than on NHR-107. Hence, if the CADs physically interact with NHR-70 and NHR-107 AND the computational docking experiments were insightful, we would expect that amiodarone to have a higher predicted affinity for the NHR-70 ligand compared to NHR-107. In the charged state, AF3 CNN affinity (D) was comparable between the two receptors, while pose score (G) clearly favored NHR-70 (0.75 versus 0.62). In the uncharged state, the trends reversed (E and H). With Boltz-2, there was little difference in predicted affinity with charged amiodarone (L), and the pose score favored NHR-107, not NHR-70 (O), with similar results for uncharged amiodarone (M and P). Hence, this benchmark is not clearly met. Finally, if the computational docking experiments were indeed predictive, we expected PXR to show higher predicted affinity for SR12813 and hyperforin than VDR. Here, the trends for affinity are all consistent for the positive control ligands (F and N), and the pose scores are largely in agreement with expectations (I and Q). Hence, the computational predictions are in reasonable agreement with expectations with the positive controls, but the magnitude of the differences between the positive control pairs and the negative control pairs is expected to be larger. See S1 Data for the data underlying the graphs.

(TIFF)

pbio.3004006.s005.tiff^{(15.3MB, tiff)}

S6 Fig. The mNeonGreen::PGP-13 translational fusion is functional.

A. Images of the heads of animals of the indicated genotype (to the left) and treatment (at the top of the columns). Treatment time was 48 hours. The imaging channel for each animal is indicated at the bottom of the columns; DIC, differential interference contrast; BIRE, birefringence; GFP, green fluorescent protein. Anterior is up. The scale in the top left applies to all images. The relevant strains/genotypes are N2 wild type; RB894 pgp-13(ok747); RP2960 pgp-14(ok2660); RP3495 pgp-13(tr696) pgp-14(ok2660); RP3561 trIs115 [mNeonGreen::PGP-13] pgp-14(ok2660). wact-190 crystals are indicated with a pink arrow; mNeonGreen::PGP-13 signal is indicated with white arrows. The key result that demonstrates mNeonGreen::PGP-13 functionality is that the strain containing the mNeonGreen insertion into PGP-13 in the background of the pgp-14(ok2660) null yields crystals (indicated by a green box) in the background of the CAD terconazole (which suppresses pgp-14(0) because it upregulates pgp-13 expression), just like wild-type pgp-13 in the background of pgp-14(0) (yellow box). If pgp-13 loses its function, terconazole cannot resensitize pgp-14(0) to wact-190 crystals (red box). Note that the controls for this experiment are the same as that shown in S3 Fig (the experiments were done at the same time). B. Quantification of the number of animals within the population that exhibited wact-190 crystals in the experiment shown in A. N = 3 independent trials with at least 24 animals counted per sample (n ≥ 24). A two-tailed Student T test was used to determine p values, except where there are absolute differences (all trials for a sample yields zero or 100 percent). See S1 Data for the data underlying the graphs.

(TIFF)

pbio.3004006.s006.tiff^{(15.3MB, tiff)}

S7 Fig. Gold standard CADs fail to induce pgp-13 reporter expression at higher concentrations.

Compounds were tested for their ability to induce expression of the pgp-13 reporter in living worms in the same manner as described in the manuscript. Two different-day independent trials were performed and 10 worms were analyzed for each condition in each trial. One way ANOVA was used to analyze differences. Two stars indicates p = 0.006; ns = not significant. See S1 Data for the data underlying the graphs.

(TIFF)

pbio.3004006.s007.tiff^{(15.3MB, tiff)}

S8 Fig. The qRT-PCR analysis of pgp-13, pgp-14, cyp-35b1, cyp-35b2, and cyp-35b3 transcript abundance in wild type, nhr-70(tm1697), and nhr-107(tm4512) animals.

The worms were treated for 96 hours with either solvent control, terconazole (30 µM), clomipramine (30 µM) clomiphene (20 µM), azelastine (20 µM), or clemizole (20 µM). Data shown represents mean ± S.E.M.; N = 3 biological trials with n = 2 technical repeats each. Statistical significance was determined via two-way ANOVA followed by Fisher’s LSD test comparing each treatment to the solvent control within each respective genotype. Black asterisks indicate statistical significance of p < 0.05 relative to the solvent-only control. Red asterisks indicate significance of p < 0.05 between the mutant strains compared to wild type within each treatment group. See S1 Data for the data underlying the graphs.

(TIFF)

pbio.3004006.s008.tiff^{(15.3MB, tiff)}

S9 Fig. The accumulation of CAD parent molecules in yeast cells. A.

For every independent repeat of every experiment shown in Fig 5c (N ≥ 12 independent repeats), we calculated the ratio of the parent (i.e., nonmetabolized) CAD molecule that is present in the cell lysate versus that found in the buffer. With the assembled dataset, we performed a one-way ANOVA analyses revealing p < 0.001. Standard error of the mean is shown. B. An all-against-all statistical comparison was performed on the data shown in A using the Tukey’s multiple comparisons test. Only the accumulation of clomiphene in yeast is significantly different than other CADs (p < 0.05). See S1 Data for the data underlying the graphs.

(TIFF)

pbio.3004006.s009.tiff^{(15.3MB, tiff)}

S10 Fig. A survey of the potential CYP35-B2 binding site residue contacts with exemplar CADs.

A-B. Each heatmap for the indicated charged and uncharged CAD shows the percent of time during 3 x 100 ns simulation that each atom of the CAD ligand was in contact (<4 Å) with any atom of a given CYP-35B2 binding site residue. The difference in charge is due to different protonation of the nitrogen (bottom row of all the heatmaps). The atom numbers on the left of the charts correspond to those indicated on the molecules to the right of the charts. The legend below B is for all charts. C. Illustration of the solvent accessible surface area of selected C. elegans P450s showing that even though D311 is conserved in sequence, it is not conserved in a structural role. For a selection of CYPs the apo structure was predicted with Boltz-1 and simulated for 3 x 100 ns with Amber, and the solvent accessible surface area of D311 or the equivalent residues was measured. Only in CAD metabolizing P450s of the CYP-35B family was this residue highly accessible and therefore available for interacting with substrates.

(TIFF)

pbio.3004006.s010.tiff^{(15.3MB, tiff)}

S11 Fig. Testing the utility of the PhsNUC-1::pHTomato reporter.

CADs, including the positive control 10mM chloroquine, were tested for their ability to increase the intensity of the pH sensitive fluorophore pHTomato. A. Representative images of the XW19180 PhsNUC-1::pHTomato reporter strain 24 hours after heat shock. B. Quantification of the pHTomato fluorescent signal in the indicated conditions. Three different-day independent trials were performed and 5 worms were analyzed for each condition in each trial (N = 3, n = 3). One way ANOVA was used to analyze differences. ns = not significant.

(TIFF)

pbio.3004006.s011.tiff^{(15.3MB, tiff)}

S12 Fig. Examples of the varied colony morphologies of different Streptomyces strains, including the hits (1E6 and 1G9) and exemplar non-hit (5E1).

Strains were grown in 12-well plates over the course of 7 days. In addition to the hits and nonhit, additional strains are magnified to highlight features of interest or the obvious secreted secondary metabolites (black arrows) that some strains make under these culture conditions. Also note the differences in the color of the agar base for some strains, which also reflects the secretion of different secondary metabolites.

(TIFF)

pbio.3004006.s012.tiff^{(15.3MB, tiff)}

S1 Data. Source data file.

The Source Data File contains the data associated with the display items. The first tab provides a navigational index.

(XLSX)

pbio.3004006.s013.xlsx^{(39MB, xlsx)}

S1 Raw Image. The raw western blot validation of FLAG-tagged protein expression that is cropped for S4 Fig.

D. NHR-70 and NHR-107 were detected using an anti-FLAG antibody. Human PXR (hPXR), mouse PXR (mPXR), and wild-type HEK293T cell lysates were included as controls. Samples were collected in triplicate. β-Actin was used as a loading control, and molecular weight markers are indicated on the left.

(PDF)

pbio.3004006.s014.pdf^{(123.9KB, pdf)}

Acknowledgments

We thank Igor Stagljar and Jamie Snider for training and use of their CLARIOstar fluorimeter. We thank June Tan, Michael Schertzberg, and Andrew Fraser for performing Linux-based calculations for the structural similarity map and guidance on sequence data analyses. We thank Muntasir Kamal for mentorship. We thank Jessica Lacoste, Jessica Knox, and Doreen Fang for technical assistance, and the Charlie Boone and Brenda Andrews labs for the gift of the Y15771 S. cerevisiae yeast strain. We thank Tim Hughes and Ali Fathi for preliminary computational analyses. Some C. elegans strains were provided by the CGC, which is funded by NIH Office of Research Infrastructure Programs (P40 OD010440). We thank Shohei Mitani and the National Bioresource Project of Japan for multiple C. elegans mutant strains.

Abbreviations

BBB: blood-brain barrier
CADs: cationic amphiphilic drugs
DBD: DNA-binding domain
DIC: differential interference contrast
DILI: drug-induced liver injury
ESI: electrospray ionization
FBS: fetal bovine serum
FS: Fisher Strand
LB: lysogeny broth
LBD: ligand binding domain
LCMS: liquid chromatography-coupled mass spectrometry
MQ: mapping quality
NGM: nematode growth medium
OD600: optical density
P0: parental
PBS: phosphate-buffered saline
PFIC3: progressive familial intrahepatic cholestasis 3
Pgp: p-glycoprotein
QD: qual by depth
RMSD: root mean square deviation
SOR: symmetric odds ratio

Data Availability

The data are in the figures, manuscript, and Source Data File.

Funding Statement

Funders include the Canadian Institutes of Health Research Project Grants 173448 and 186156 (PJR), Canadian Institutes of Health Research Project Grant PJT-186117 (HMK), a Canada Research Chair (Tier 1) grant in Chemical Biology (PJR), an Ontario Graduate Scholarship to LT, a grant from the National Institute of General Medical Sciences R35GM140753 (DK), and a NSERC grant RGPIN-2020-07212 (CC). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

References

1.Breiden B, Sandhoff K. Emerging mechanisms of drug-induced phospholipidosis. Biol Chem. 2019;401(1):31–46. doi: 10.1515/hsz-2019-0270 [DOI] [PubMed] [Google Scholar]
2.Halliwell WH. Cationic amphiphilic drug-induced phospholipidosis. Toxicol Pathol. 1997;25(1):53–60. doi: 10.1177/019262339702500111 [DOI] [PubMed] [Google Scholar]
3.Tummino TA, Rezelj VV, Fischer B, Fischer A, O’Meara MJ, Monel B, et al. Drug-induced phospholipidosis confounds drug repurposing for SARS-CoV-2. Science. 2021;373(6554):541–7. doi: 10.1126/science.abi4708 [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Abe A, Shayman JA. The role of negatively charged lipids in lysosomal phospholipase A2 function. J Lipid Res 2009;50:2027–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Chen M, Borlak J, Tong W. High lipophilicity and high daily dose of oral medications are associated with significant risk for drug-induced liver injury. Hepatology. 2013;58(1):388–96. doi: 10.1002/hep.26208 [DOI] [PubMed] [Google Scholar]
6.Chalasani N, Bonkovsky HL, Fontana R, Lee W, Stolz A, Talwalkar J, et al. Features and outcomes of 899 patients with drug-induced liver injury: the DILIN prospective study. Gastroenterology. 2015;148(7):1340-52.e7. doi: 10.1053/j.gastro.2015.03.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Dirven H, Vist GE, Bandhakavi S, Mehta J, Fitch SE, Pound P, et al. Performance of preclinical models in predicting drug-induced liver injury in humans: a systematic review. Sci Rep. 2021;11(1):6403. doi: 10.1038/s41598-021-85708-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Creanza TM, Delre P, Ancona N, Lentini G, Saviano M, Mangiatordi GF. Structure-based prediction of hERG-related cardiotoxicity: a benchmark study. J Chem Inf Model. 2021;61(9):4758–70. doi: 10.1021/acs.jcim.1c00744 [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Sun H, Xia M, Shahane SA, Jadhav A, Austin CP, et al. Are hERG channel blockers also phospholipidosis inducers? Bioorg Med Chem Lett 2013;23:4587–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Vandenberg JI, Perry MD, Perrin MJ, Mann SA, Ke Y, Hill AP. hERG K(+) channels: structure, function, and clinical significance. Physiol Rev. 2012;92(3):1393–478. doi: 10.1152/physrev.00036.2011 [DOI] [PubMed] [Google Scholar]
11.Pelletier DJ, Gehlhaar D, Tilloy-Ellul A, Johnson TO, Greene N. Evaluation of a published in silico model and construction of a novel Bayesian model for predicting phospholipidosis inducing potential. J Chem Inf Model. 2007;47(3):1196–205. doi: 10.1021/ci6004542 [DOI] [PubMed] [Google Scholar]
12.Schieferdecker S, Eberlein A, Vock E, Beilmann M. Development of an in silico consensus model for the prediction of the phospholipigenic potential of small molecules. Comput Toxicol. 2022;22:100226. doi: 10.1016/j.comtox.2022.100226 [DOI] [Google Scholar]
13.Blumberg B, Sabbagh W Jr, Juguilon H, Bolado J Jr, van Meter CM, Ong ES, et al. SXR, a novel steroid and xenobiotic-sensing nuclear receptor. Genes Dev. 1998;12(20):3195–205. doi: 10.1101/gad.12.20.3195 [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Kliewer SA, Moore JT, Wade L, Staudinger JL, Watson MA, Jones SA, et al. An orphan nuclear receptor activated by pregnanes defines a novel steroid signaling pathway. Cell. 1998;92(1):73–82. doi: 10.1016/s0092-8674(00)80900-9 [DOI] [PubMed] [Google Scholar]
15.Geick A, Eichelbaum M, Burk O. Nuclear receptor response elements mediate induction of intestinal MDR1 by rifampin. J Biol Chem. 2001;276(18):14581–7. doi: 10.1074/jbc.M010173200 [DOI] [PubMed] [Google Scholar]
16.Takeshita A, Taguchi M, Koibuchi N, Ozawa Y. Putative role of the orphan nuclear receptor SXR (steroid and xenobiotic receptor) in the mechanism of CYP3A4 inhibition by xenobiotics. J Biol Chem. 2002;277(36):32453–8. doi: 10.1074/jbc.M111245200 [DOI] [PubMed] [Google Scholar]
17.Cermanova J, Fuksa L, Brcakova E, Hroch M, Kucera O, Kolouchova G, et al. Up-regulation of renal Mdr1 and Mrp2 transporters during amiodarone pretreatment in rats. Pharmacol Res. 2010;61(2):129–35. doi: 10.1016/j.phrs.2009.08.005 [DOI] [PubMed] [Google Scholar]
18.Hodges LM, Markova SM, Chinn LW, Gow JM, Kroetz DL, Klein TE, et al. Very important pharmacogene summary: ABCB1 (MDR1, P-glycoprotein). Pharmacogenet Genomics. 2011;21(3):152–61. doi: 10.1097/FPC.0b013e3283385a1c [DOI] [PMC free article] [PubMed] [Google Scholar]
19.McDonald MG, Au NT, Rettie AE. P450-based drug-drug interactions of amiodarone and its metabolites: diversity of inhibitory mechanisms. Drug Metab Dispos 2015;43:1661–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Faucette SR, Zhang T-C, Moore R, Sueyoshi T, Omiecinski CJ, LeCluyse EL, et al. Relative activation of human pregnane X receptor versus constitutive androstane receptor defines distinct classes of CYP2B6 and CYP3A4 inducers. J Pharmacol Exp Ther. 2007;320(1):72–80. doi: 10.1124/jpet.106.112136 [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Boyd-Kimball D, Gonczy K, Lewis B, Mason T, Siliko N, et al. Classics in chemical neuroscience: chlorpromazine. ACS Chem Neurosci 2019;10:79–88. [DOI] [PubMed] [Google Scholar]
22.Burns AR, Luciani GM, Musso G, Bagg R, Yeo M, Zhang Y, et al. Caenorhabditis elegans is a useful model for anthelmintic discovery. Nat Commun. 2015;6:7485. doi: 10.1038/ncomms8485 [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Kamal M, Knox J, Horne RI, Tiwari OS, Burns AR, Han D, et al. A rapid in vivo pipeline to identify small molecule inhibitors of amyloid aggregation. Nat Commun. 2024;15(1):8311. doi: 10.1038/s41467-024-52480-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Kamal M, Moshiri H, Magomedova L, Han D, Nguyen KCQ, Yeo M, et al. The marginal cells of the Caenorhabditis elegans pharynx scavenge cholesterol and other hydrophobic small molecules. Nat Commun. 2019;10(1):3938. doi: 10.1038/s41467-019-11908-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Kamal M, Tokmakjian L, Knox J, Han D, Moshiri H, Magomedova L, et al. PGP-14 establishes a polar lipid permeability barrier within the C. elegans pharyngeal cuticle. PLoS Genet. 2023;19(11):e1011008. doi: 10.1371/journal.pgen.1011008 [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Kamal M, Tokmakjian L, Knox J, Mastrangelo P, Ji J, Cai H, et al. A spatiotemporal reconstruction of the C. elegans pharyngeal cuticle reveals a structure rich in phase-separating proteins. Elife. 2022;11:e79396. doi: 10.7554/eLife.79396 [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Smit JJ, Schinkel AH, Oude Elferink RP, Groen AK, Wagenaar E, van Deemter L, et al. Homozygous disruption of the murine mdr2 P-glycoprotein gene leads to a complete absence of phospholipid from bile and to liver disease. Cell. 1993;75(3):451–62. doi: 10.1016/0092-8674(93)90380-9 [DOI] [PubMed] [Google Scholar]
28.Prescher M, Kroll T, Schmitt L. ABCB4/MDR3 in health and disease - at the crossroads of biochemistry and medicine. Biol Chem. 2019;400(10):1245–59. doi: 10.1515/hsz-2018-0441 [DOI] [PubMed] [Google Scholar]
29.Alam A, Kowal J, Broude E, Roninson I, Locher KP. Structural insight into substrate and inhibitor discrimination by human P-glycoprotein. Science. 2019;363(6428):753–6. doi: 10.1126/science.aav7102 [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Nosol K, Romane K, Irobalieva RN, Alam A, Kowal J, Fujita N, et al. Cryo-EM structures reveal distinct mechanisms of inhibition of the human multidrug transporter ABCB1. Proc Natl Acad Sci U S A. 2020;117(42):26245–53. doi: 10.1073/pnas.2010264117 [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Park HJ, Kim TH, Kim SW, Noh SH, Cho KJ, Choi C, et al. Functional characterization of ABCB4 mutations found in progressive familial intrahepatic cholestasis type 3. Sci Rep. 2016;6:26872. doi: 10.1038/srep26872 [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Kampfer P. The Family Streptomycetaceae, Part I: Taxonomy in The Prokaryotes., edited by M. Dworkin, Falkow, S., Rosenberg, E., Schleifer, KH., Stackebrandt, E. Springer, New York, NY; 2006. [Google Scholar]
33.Samuel BS, Rowedder H, Braendle C, Félix M-A, Ruvkun G. Caenorhabditis elegans responses to bacteria from its natural habitats. Proc Natl Acad Sci U S A. 2016;113(27):E3941-9. doi: 10.1073/pnas.1607183113 [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Robinson-Rechavi M, Maina CV, Gissendanner CR, Laudet V, Sluder A. Explosive lineage-specific expansion of the orphan nuclear receptor HNF4 in nematodes. J Mol Evol. 2005;60(5):577–86. doi: 10.1007/s00239-004-0175-8 [DOI] [PubMed] [Google Scholar]
35.Sasaki S, Urabe M, Maeda T, Suzuki J, Irie R, Suzuki M, et al. Induction of hepatic metabolic functions by a novel variant of hepatocyte nuclear factor 4γ. Mol Cell Biol. 2018;38(24):e00213-18. doi: 10.1128/MCB.00213-18 [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Vemuri K, Radi SH, Sladek FM, Verzi MP. Multiple roles and regulatory mechanisms of the transcription factor HNF4 in the intestine. Front Endocrinol (Lausanne). 2023;14:1232569. doi: 10.3389/fendo.2023.1232569 [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Bookout AL, Jeong Y, Downes M, Yu RT, Evans RM, Mangelsdorf DJ. Anatomical profiling of nuclear receptor expression reveals a hierarchical transcriptional network. Cell. 2006;126(4):789–99. doi: 10.1016/j.cell.2006.06.049 [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Sheps JA, Ralph S, Zhao Z, Baillie DL, Ling V. The ABC transporter gene family of Caenorhabditis elegans has implications for the evolutionary dynamics of multidrug resistance in eukaryotes. Genome Biol. 2004;5(3):R15. doi: 10.1186/gb-2004-5-3-r15 [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Morelli JK, Buehrle M, Pognan F, Barone LR, Fieles W, Ciaccio PJ. Validation of an in vitro screen for phospholipidosis using a high-content biology platform. Cell Biol Toxicol. 2006;22(1):15–27. doi: 10.1007/s10565-006-0176-z [DOI] [PubMed] [Google Scholar]
40.Shahane SA, Huang R, Gerhold D, Baxa U, Austin CP, Xia M. Detection of phospholipidosis induction: a cell-based assay in high-throughput and high-content format. J Biomol Screen. 2014;19(1):66–76. doi: 10.1177/1087057113502851 [DOI] [PMC free article] [PubMed] [Google Scholar]
41.Bauch C, Bevan S, Woodhouse H, Dilworth C, Walker P. Predicting in vivo phospholipidosis-inducing potential of drugs by a combined high content screening and in silico modelling approach. Toxicol In Vitro. 2015;29(3):621–30. doi: 10.1016/j.tiv.2015.01.014 [DOI] [PubMed] [Google Scholar]
42.Hinkovska-Galcheva V, Treadwell T, Shillingford JM, Lee A, Abe A, Tesmer JJG, et al. Inhibition of lysosomal phospholipase A2 predicts drug-induced phospholipidosis. J Lipid Res. 2021;62:100089. doi: 10.1016/j.jlr.2021.100089 [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Daina A, Michielin O, Zoete V. SwissADME: a free web tool to evaluate pharmacokinetics, drug-likeness and medicinal chemistry friendliness of small molecules. Sci Rep. 2017;7:42717. doi: 10.1038/srep42717 [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Shaker B, Lee J, Lee Y, Yu M-S, Lee H-M, Lee E, et al. A machine learning-based quantitative model (LogBB_Pred) to predict the blood-brain barrier permeability (logBB value) of drug compounds. Bioinformatics. 2023;39(10):btad577. doi: 10.1093/bioinformatics/btad577 [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Rangaraju S, Solis GM, Thompson RC, Gomez-Amaro RL, Kurian L, Encalada SE, et al. Suppression of transcriptional drift extends C. elegans lifespan by postponing the onset of mortality. Elife. 2015;4:e08833. doi: 10.7554/eLife.08833 [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Diot C, García-González AP, Vieira AF, Walker M, Honeywell M, Doyle H, et al. Bacterial diet modulates tamoxifen-induced death via host fatty acid metabolism. Nat Commun. 2022;13(1):5595. doi: 10.1038/s41467-022-33299-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Schmauder L, Richter K. hsp-90 and unc-45 depletion induce characteristic transcriptional signatures in coexpression cliques of C. elegans. Sci Rep. 2021;11(1):12852. doi: 10.1038/s41598-021-91690-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Larigot L, Mansuy D, Borowski I, Coumoul X, Dairou J. Cytochromes P450 of Caenorhabditis elegans: implication in biological functions and metabolism of xenobiotics. Biomolecules. 2022;12(3):342. doi: 10.3390/biom12030342 [DOI] [PMC free article] [PubMed] [Google Scholar]
49.de Groot MJ, Ackland MJ, Horne VA, Alex AA, Jones BC. Novel approach to predicting P450-mediated drug metabolism: development of a combined protein and pharmacophore model for CYP2D6. J Med Chem. 1999;42(9):1515–24. doi: 10.1021/jm981118h [DOI] [PubMed] [Google Scholar]
50.Wang A, Stout CD, Zhang Q, Johnson EF. Contributions of ionic interactions and protein dynamics to cytochrome P450 2D6 (CYP2D6) substrate and inhibitor binding. J Biol Chem. 2015;290(8):5092–104. doi: 10.1074/jbc.M114.627661 [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Zhu M, Zhang H, Humphreys WG. Drug metabolite profiling and identification by high-resolution mass spectrometry. J Biol Chem. 2011;286(29):25419–25. doi: 10.1074/jbc.R110.200055 [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Burns AR, Baker RJ, Kitner M, Knox J, Cooke B, Volpatti JR, et al. Selective control of parasitic nematodes using bioactivated nematicides. Nature. 2023;618(7963):102–9. doi: 10.1038/s41586-023-06105-5 [DOI] [PubMed] [Google Scholar]
53.Knox J, Burns AR, Cooke B, Cammalleri SR, Kitner M, Ching J, et al. Cyprocide selectively kills nematodes via cytochrome P450 bioactivation. Nat Commun. 2024;15(1):5529. doi: 10.1038/s41467-024-49738-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Collins JB, Dilks CM, Hahnel SR, Rodriguez B, Fox BW, Redman E, et al. Naturally occurring variation in a cytochrome P450 modifies thiabendazole responses independently of beta-tubulin. PLoS Pathog. 2025;21(1):e1012602. doi: 10.1371/journal.ppat.1012602 [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Wohlwend J, Corso G, Passaro S, Getz N, Reveiz M, et al. Boltz-1 Democratizing Biomolecular Interaction Modeling. bioRxiv. 2025.
56.Case DA, Cerutti DS, Cruzeiro VWD, Darden TA, Duke RE, et al. Recent Developments in Amber Biomolecular Simulations. J Chem Inf Model 2025;65:7835–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
57.Clokey GV, Jacobson LA. The autofluorescent “lipofuscin granules” in the intestinal cells of Caenorhabditis elegans are secondary lysosomes. Mech Ageing Dev. 1986;35(1):79–94. doi: 10.1016/0047-6374(86)90068-0 [DOI] [PubMed] [Google Scholar]
58.Bossinger O, Schierenberg E. The use of fluorescent marker dyes for studying intercellular communication in nematode embryos. Int J Dev Biol. 1996;40(1):431–9. doi: 10.1387/ijdb.8735958 [DOI] [PubMed] [Google Scholar]
59.Hermann GJ, Schroeder LK, Hieb CA, Kershner AM, Rabbitts BM, Fonarev P, et al. Genetic analysis of lysosomal trafficking in Caenorhabditis elegans. Mol Biol Cell. 2005;16(7):3273–88. doi: 10.1091/mbc.e05-01-0060 [DOI] [PMC free article] [PubMed] [Google Scholar]
60.Baxi K, Ghavidel A, Waddell B, Harkness TA, de Carvalho CE. Regulation of lysosomal function by the DAF-16 forkhead transcription factor couples reproduction to aging in Caenorhabditis elegans. Genetics. 2017;207(1):83–101. doi: 10.1534/genetics.117.204222 [DOI] [PMC free article] [PubMed] [Google Scholar]
61.Edgar LG, McGhee JD. Embryonic expression of a gut-specific esterase in Caenorhabditis elegans. Dev Biol 1986;114:109–18. [DOI] [PubMed] [Google Scholar]
62.Nedergaard M, Desai S, Pulsinelli W. Dicarboxy-dichlorofluorescein: a new fluorescent probe for measuring acidic intracellular pH. Anal Biochem. 1990;187(1):109–14. doi: 10.1016/0003-2697(90)90425-9 [DOI] [PubMed] [Google Scholar]
63.Nielsen IØ, Clemmensen KKB, Fogde DL, Dietrich TN, Giacobini JD, Bilgin M, et al. Cationic amphiphilic drugs induce accumulation of cytolytic lysoglycerophospholipids in the lysosomes of cancer cells and block their recycling into common membrane glycerophospholipids. Mol Biol Cell. 2024;35(3):ar25. doi: 10.1091/mbc.E23-06-0263 [DOI] [PMC free article] [PubMed] [Google Scholar]
64.Li X, Sun Y, Wang X. Probing lysosomal activity in vivo. Biophys Rep. 2021;7(1):1–7. doi: 10.52601/bpr.2021.200047 [DOI] [PMC free article] [PubMed] [Google Scholar]
65.Dirksen P, Marsh SA, Braker I, Heitland N, Wagner S, Nakad R, et al. The native microbiome of the nematode Caenorhabditis elegans: gateway to a new host-microbiome model. BMC Biol. 2016;14:38. doi: 10.1186/s12915-016-0258-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
66.Ho LK, Nodwell JR. David and Goliath: chemical perturbation of eukaryotes by bacteria. J Ind Microbiol Biotechnol. 2016;43(2–3):233–48. doi: 10.1007/s10295-015-1686-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
67.Daniel-Ivad M, Pimentel-Elardo S, Nodwell JR. Control of Specialized metabolism by signaling and transcriptional regulation: opportunities for new platforms for drug discovery? Annu Rev Microbiol. 2018;72:25–48. doi: 10.1146/annurev-micro-022618-042458 [DOI] [PubMed] [Google Scholar]
68.van Helvoort A, Smith AJ, Sprong H, Fritzsche I, Schinkel AH, Borst P, et al. MDR1 P-glycoprotein is a lipid translocase of broad specificity, while MDR3 P-glycoprotein specifically translocates phosphatidylcholine. Cell. 1996;87(3):507–17. doi: 10.1016/s0092-8674(00)81370-7 [DOI] [PubMed] [Google Scholar]
69.Raggers RJ, Vogels I, van Meer G. Multidrug-resistance P-glycoprotein (MDR1) secretes platelet-activating factor. Biochem J. 2001;357(Pt 3):859–65. doi: 10.1042/0264-6021:3570859 [DOI] [PMC free article] [PubMed] [Google Scholar]
70.Williamson B, Dooley KE, Zhang Y, Back DJ, Owen A. Induction of influx and efflux transporters and cytochrome P450 3A4 in primary human hepatocytes by rifampin, rifabutin, and rifapentine. Antimicrob Agents Chemother. 2013;57(12):6366–9. doi: 10.1128/AAC.01124-13 [DOI] [PMC free article] [PubMed] [Google Scholar]
71.Järvinen E, Hammer HS, Pötz O, Ingelman-Sundberg M, Stage TB. 3D spheroid primary human hepatocytes for prediction of cytochrome P450 and drug transporter induction. Clin Pharmacol Ther. 2023;113(6):1284–94. doi: 10.1002/cpt.2887 [DOI] [PubMed] [Google Scholar]
72.Weber ND, Odriozola L, Martínez-García J, Ferrer V, Douar A, Bénichou B, et al. Gene therapy for progressive familial intrahepatic cholestasis type 3 in a clinically relevant mouse model. Nat Commun. 2019;10(1):5694. doi: 10.1038/s41467-019-13614-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
73.Menzel R, Bogaert T, Achazi R. A systematic gene expression screen of Caenorhabditis elegans cytochrome P450 genes reveals CYP35 as strongly xenobiotic inducible. Arch Biochem Biophys. 2001;395(2):158–68. doi: 10.1006/abbi.2001.2568 [DOI] [PubMed] [Google Scholar]
74.Aoyama Y, Yoshida Y, Sato R. Yeast cytochrome P-450 catalyzing lanosterol 14 alpha-demethylation. II. Lanosterol metabolism by purified P-450(14)DM and by intact microsomes. J Biol Chem 1984;259: 661–6. [PubMed] [Google Scholar]
75.Yoshida Y, Aoyama Y. Yeast cytochrome P-450 catalyzing lanosterol 14 alpha-demethylation. I. purification and spectral properties. J Biol Chem. 1984;259(3):1655–60. doi: 10.1016/s0021-9258(17)43458-2 [DOI] [PubMed] [Google Scholar]
76.Yoon V, Nodwell JR. Activating secondary metabolism with stress and chemicals. J Ind Microbiol Biotechnol. 2014;41(2):415–24. doi: 10.1007/s10295-013-1387-y [DOI] [PubMed] [Google Scholar]
77.Saito S, Arai MA. Methodology for awakening the potential secondary metabolic capacity in actinomycetes. Beilstein J Org Chem. 2024;20:753–66. doi: 10.3762/bjoc.20.69 [DOI] [PMC free article] [PubMed] [Google Scholar]
78.Lewis JA, Fleming JT. Basic culture methods. Methods Cell Biol 1995;48:3–29. [PubMed] [Google Scholar]
79.Burns AR, Kwok TCY, Howard A, Houston E, Johanson K, Chan A, et al. High-throughput screening of small molecules for bioactivity and target identification in Caenorhabditis elegans. Nat Protoc. 2006;1(4):1906–14. doi: 10.1038/nprot.2006.283 [DOI] [PubMed] [Google Scholar]
80.Kwok TCY, Ricker N, Fraser R, Chan AW, Burns A, Stanley EF, et al. A small-molecule screen in C. elegans yields a new calcium channel antagonist. Nature. 2006;441(7089):91–5. doi: 10.1038/nature04657 [DOI] [PubMed] [Google Scholar]
81.Andrews S. FastQC: a quality control tool for high throughput sequence data. 2010. pp.
82.Li H, Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics 2009;25:1754–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
83.McKenna A, Hanna M, Banks E, Sivachenko A, Cibulskis K, Kernytsky A, et al. The genome analysis toolkit: a mapreduce framework for analyzing next-generation DNA sequencing data. Genome Res. 2010;20(9):1297–303. doi: 10.1101/gr.107524.110 [DOI] [PMC free article] [PubMed] [Google Scholar]
84.Wang K, Li M, Hakonarson H. ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data. Nucleic Acids Res. 2010;38(16):e164. doi: 10.1093/nar/gkq603 [DOI] [PMC free article] [PubMed] [Google Scholar]
85.Riccio C. Extracting total RNA from Caenorhabditis elegans using phase-lock gel separation tubes. 2019.
86.Ye J, Coulouris G, Zaretskaya I, Cutcutache I, Rozen S, Madden TL. Primer-BLAST: a tool to design target-specific primers for polymerase chain reaction. BMC Bioinformatics. 2012;13:134. doi: 10.1186/1471-2105-13-134 [DOI] [PMC free article] [PubMed] [Google Scholar]
87.Schmittgen TD, Livak KJ. Analyzing real-time PCR data by the comparative C(T) method. Nat Protoc. 2008;3(6):1101–8. doi: 10.1038/nprot.2008.73 [DOI] [PubMed] [Google Scholar]
88.Tao J, Hao Y, Li X, Yin H, Xie N, et al. Systematic Identification of housekeeping genes possibly used as references in Caenorhabditis elegans by large-scale data integration. Cells 2020;9. [DOI] [PMC free article] [PubMed]
89.Mello C, Fire A. DNA transformation. Methods Cell Biol 1995;48:451–82. [PubMed] [Google Scholar]
90.Abramson J, Adler J, Dunger J, Evans R, Green T, Pritzel A, et al. Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature. 2024;630(8016):493–500. doi: 10.1038/s41586-024-07487-w [DOI] [PMC free article] [PubMed] [Google Scholar]
91.Passaro S, Corso G, Wohlwend J, Reveiz M, Thaler S, et al. Boltz-2: towards accurate and efficient binding affinity prediction. bioRxiv. 2025.
92.McNutt AT, Francoeur P, Aggarwal R, Masuda T, Meli R, et al., GNINA 1.0: molecular docking with deep learning. J Cheminform 2021;13:43. [DOI] [PMC free article] [PubMed] [Google Scholar]
93.McNutt AT, Li Y, Meli R, Aggarwal R, Koes DR. GNINA 1.3: the next increment in molecular docking with deep learning. J Cheminform 2025;17:28. [DOI] [PMC free article] [PubMed] [Google Scholar]
94.O'Boyle NM, Banck M, James CA, Morley C, Vandermeersch T, et al. Open Babel: An open chemical toolbox. J Cheminform 2011;3:33. [DOI] [PMC free article] [PubMed] [Google Scholar]
95.Cock PJA, Antao T, Chang JT, Chapman BA, Cox CJ, Dalke A, et al. Biopython: freely available Python tools for computational molecular biology and bioinformatics. Bioinformatics. 2009;25(11):1422–3. doi: 10.1093/bioinformatics/btp163 [DOI] [PMC free article] [PubMed] [Google Scholar]
96.Xue Y, Chao E, Zuercher WJ, Willson TM, Collins JL, Redinbo MR. Crystal structure of the PXR-T1317 complex provides a scaffold to examine the potential for receptor antagonism. Bioorg Med Chem. 2007;15(5):2156–66. doi: 10.1016/j.bmc.2006.12.026 [DOI] [PMC free article] [PubMed] [Google Scholar]
97.Rochel N, Wurtz JM, Mitschler A, Klaholz B, Moras D. The crystal structure of the nuclear receptor for vitamin D bound to its natural ligand. Mol Cell. 2000;5(1):173–9. doi: 10.1016/s1097-2765(00)80413-x [DOI] [PubMed] [Google Scholar]
98.Pearlman DA, Case DA, Caldwell JW, Ross WS, Cheatham TE III, DeBolt S, et al. AMBER, a package of computer programs for applying molecular mechanics, normal mode analysis, molecular dynamics and free energy calculations to simulate the structural and energetic properties of molecules. Computer Phy Commun. 1995;91(1–3):1–41. doi: 10.1016/0010-4655(95)00041-d [DOI] [Google Scholar]
99.Shahrokh K, Orendt A, Yost GS, Cheatham TE 3rd. Quantum mechanically derived AMBER-compatible heme parameters for various states of the cytochrome P450 catalytic cycle. J Comput Chem. 2012;33(2):119–33. doi: 10.1002/jcc.21922 [DOI] [PMC free article] [PubMed] [Google Scholar]
100.Wang J, Wolf RM, Caldwell JW, Kollman PA, Case DA. Development and testing of a general amber force field. J Comput Chem. 2004;25(9):1157–74. doi: 10.1002/jcc.20035 [DOI] [PubMed] [Google Scholar]
101.Debiec KT, Cerutti DS, Baker LR, Gronenborn AM, Case DA, Chong LT. Further along the Road Less Traveled: AMBER ff15ipq, an original protein force field built on a self-consistent physical model. J Chem Theory Comput. 2016;12(8):3926–47. doi: 10.1021/acs.jctc.6b00567 [DOI] [PMC free article] [PubMed] [Google Scholar]
102.Koes DR, Vries JK. Evaluating amber force fields using computed NMR chemical shifts. Proteins. 2017;85(10):1944–56. doi: 10.1002/prot.25350 [DOI] [PMC free article] [PubMed] [Google Scholar]
103.Michaud-Agrawal N, Denning EJ, Woolf TB, Beckstein O. MDAnalysis: a toolkit for the analysis of molecular dynamics simulations. J Comput Chem. 2011;32(10):2319–27. doi: 10.1002/jcc.21787 [DOI] [PMC free article] [PubMed] [Google Scholar]
104.Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 2003;13(11):2498–504. doi: 10.1101/gr.1239303 [DOI] [PMC free article] [PubMed] [Google Scholar]
105.Pan X, Wang H, Li C, Zhang JZH, Ji C. MolGpka: a web server for small molecule pKa prediction using a graph-convolutional neural network. J Chem Inf Model. 2021;61(7):3159–65. doi: 10.1021/acs.jcim.1c00075 [DOI] [PubMed] [Google Scholar]
106.Montalibet J, Kennedy BP. Using yeast to screen for inhibitors of protein tyrosine phosphatase 1B. Biochem Pharmacol. 2004;68(9):1807–14. doi: 10.1016/j.bcp.2004.06.024 [DOI] [PubMed] [Google Scholar]
107.Gietz RD, Schiestl RH. High-efficiency yeast transformation using the LiAc/SS carrier DNA/PEG method. Nat Protoc. 2007;2(1):31–4. doi: 10.1038/nprot.2007.13 [DOI] [PubMed] [Google Scholar]
108.Harris TW, Arnaboldi V, Cain S, Chan J, Chen WJ, Cho J, et al. WormBase: a modern model organism information resource. Nucleic Acids Res. 2020;48(D1):D762–7. doi: 10.1093/nar/gkz920 [DOI] [PMC free article] [PubMed] [Google Scholar]
109.Kent WJ, Sugnet CW, Furey TS, Roskin KM, Pringle TH, Zahler AM, et al. The human genome browser at UCSC. Genome Res. 2002;12(6):996–1006. doi: 10.1101/gr.229102 [DOI] [PMC free article] [PubMed] [Google Scholar]

PLoS Biol. doi: 10.1371/journal.pbio.3004006.r001

Decision Letter 0

Ines Alvarez-Garcia

2 Feb 2026

Dear Dr Roy,

Thank you for submitting your manuscript entitled "A Cationic Amphiphilic Drug (CAD) Defense System in the Nematode Caenorhabditis elegans" for consideration as a Research Article by PLOS Biology.

Your manuscript has now been evaluated by the PLOS Biology editorial staff as well as by an academic editor with relevant expertise and I am writing to let you know that we would like to send your submission out for external peer review.

However, before we can send your manuscript to reviewers, we need you to complete your submission by providing the metadata that is required for full assessment. To this end, please login to Editorial Manager where you will find the paper in the 'Submissions Needing Revisions' folder on your homepage. Please click 'Revise Submission' from the Action Links and complete all additional questions in the submission questionnaire.

Once your full submission is complete, your paper will undergo a series of checks in preparation for peer review. After your manuscript has passed the checks it will be sent out for review. To provide the metadata for your submission, please Login to Editorial Manager (https://www.editorialmanager.com/pbiology) within two working days, i.e. by Feb 04 2026 11:59PM.

Feel free to email us at plosbiology@plos.org if you have any queries relating to your submission.

Kind regards,

Ines

Ines Alvarez-Garcia, PhD

Senior Editor

PLOS Biology

ialvarez-garcia@plos.org

PLoS Biol. doi: 10.1371/journal.pbio.3004006.r002

Decision Letter 1

Ines Alvarez-Garcia

3 Apr 2026

Dear Dr Roy,

Thank you for your patience while your manuscript entitled "A Cationic Amphiphilic Drug (CAD) Defense System in the Nematode Caenorhabditis elegans" was peer-reviewed at PLOS Biology. Please also accept my apologies for the delay in sending you our decision. The manuscript has now been evaluated by the PLOS Biology editors, an Academic Editor with relevant expertise, and by three independent reviewers.

The reviews are attached below. As you will see, the reviewers find the conclusions novel and interesting, but they also raise several issues that will need to be addressed before we can consider the manuscript for publication. Reviewer 1 thinks that a ligand binding essay should be performed to confirm CAD binding to NHR-70 and NHR-107 to strengthen the study. This reviewer also requests an additional assay to confirm the experiments with lysosomal acidity using cDCFDA, among other issues. Reviewer 2 is mostly satisfied and only raises minor points, whereas Reviewer 3 notes that parts of the study are a bit hard to follow, thus the paper would benefit from some rewriting to make it more accessible for a wide audience. In addition, the reviewer suggests that a cartoon could be used to explain the model better.

In light of the reviews, which you will find at the end of this email, we would like to invite you to revise the work to thoroughly address the reviewers' reports. Given the extent of revision needed, we cannot make a decision about publication until we have seen the revised manuscript and your response to the reviewers' comments. Your revised manuscript is likely to be sent for further evaluation by all or a subset of the reviewers.

In addition to these revisions, you will need to complete some formatting changes, which you will receive in a follow up email. A member of our team will be in touch with a set of requests shortly.

We expect to receive your revised manuscript within 3 months. Please email us (plosbiology@plos.org) if you have any questions or concerns, or would like to request an extension.

At this stage, your manuscript remains formally under active consideration at our journal; please notify us by email if you do not intend to submit a revision so that we may withdraw it.

**IMPORTANT - SUBMITTING YOUR REVISION**

Your revisions should address the specific points made by each reviewer. Please submit the following files along with your revised manuscript:

1. A 'Response to Reviewers' file - this should detail your responses to the editorial requests, present a point-by-point response to all of the reviewers' comments, and indicate the changes made to the manuscript.

*NOTE: In your point-by-point response to the reviewers, please provide the full context of each review. Do not selectively quote paragraphs or sentences to reply to. The entire set of reviewer comments should be present in full and each specific point should be responded to individually, point by point.

You should also cite any additional relevant literature that has been published since the original submission and mention any additional citations in your response.

2. In addition to a clean copy of the manuscript, please also upload a 'track-changes' version of your manuscript that specifies the edits made. This should be uploaded as a "Revised Article with Changes Highlighted" file type.

3. Resubmission Checklist

When you are ready to resubmit your revised manuscript, please refer to this resubmission checklist: https://plos.io/Biology_Checklist

To submit a revised version of your manuscript, please go to https://www.editorialmanager.com/pbiology/ and log in as an Author. Click the link labelled 'Submissions Needing Revision' where you will find your submission record.

Please make sure to read the following important policies and guidelines while preparing your revision and fulfil the editorial requests:

a) *PLOS Data Policy*

Please note that as a condition of publication PLOS' data policy (http://journals.plos.org/plosbiology/s/data-availability) requires that you make available all data used to draw the conclusions arrived at in your manuscript. If you have not already done so, you must include any data used in your manuscript either in appropriate repositories, within the body of the manuscript, or as supporting information (N.B. this includes any numerical values that were used to generate graphs, histograms etc.). Please also indicate in each figure legend where the data can be found. For an example see here: http://www.plosbiology.org/article/info%3Adoi%2F10.1371%2Fjournal.pbio.1001908#s5

b) *Published Peer Review*

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out. Please see here for more details:

https://blogs.plos.org/plos/2019/05/plos-journals-now-open-for-published-peer-review/

c) *Blot and Gel Data Policy*

Please provide the original, uncropped and minimally adjusted images supporting all blot and gel results reported in an article's figures or Supporting Information files. We will require these files before a manuscript can be accepted so please prepare them now, if you have not already uploaded them. Please carefully read our guidelines for how to prepare and upload this data: https://journals.plos.org/plosbiology/s/figures#loc-blot-and-gel-reporting-requirements

d) *Protocols deposition*

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosbiology/s/submission-guidelines#loc-materials-and-methods

Thank you again for your submission to our journal. We hope that our editorial process has been constructive thus far, and we welcome your feedback at any time. Please don't hesitate to contact us if you have any questions or comments.

Sincerely,

Ines

Ines Alvarez-Garcia, PhD

Senior Editor

PLOS Biology

ialvarez-garcia@plos.org

------------------------------------

Reviewers' comments

Rev. 1:

This study identifies a defence system in Caenorhabditis elegans that protects against cationic amphiphilic drugs (CADs), which can cause lysosomal dysfunction. Using a combination of genetics, transcriptomics, metabolomics, chemoinformatics, and structural modeling, the authors provide compelling evidence that nuclear receptors NHR 70 and NHR 107 regulate CYP 35B enzymes and PGP 13/PGP 14 transporters to detoxify CADs. The authors provide evidence for CYP 35Bs metabolizing CADs and PGP 13/14 exporting metabolites, both critical for reducing drug-induced toxicity. A pgp 13 fluorescent reporter enabled them to screen a compound library, identifying CAD-like compounds consistent with computational predictions. They show that the defence system is also activated by certain Streptomyces strains, hinting at its evolutionary purpose. While some mechanisms remain unclear, in general, the study is well designed and executed, have several elements of novelty and will be of broad interest to the scientific community.

Comments:

1. One of the main drawbacks of the study is the lack of direct evidence of CAD binding to NHR-70 and NHR-107. Direct evidence in form of a ligand binding assay would strengthen their study.

2. Does NHR-70 and NHR-107 directly regulate the expression of the pgp genes?

3. 76% of compounds require both NHR-70 and NHR-107 while 12% require only NHR-70. This suggests a potential combinatorial regulation. Do NHR-70 and NHR-107 form heterodimers?

4. The relevance of the last section of the manuscript that deals with checking the response to Actinobacteria strains is quite vague. Out of the 65 strains, only two showed obvious increase in the pgp-13::gfp expression while all these strains are claimed to produce secondary metabolites. What are the metabolites produced by the strains 1G9 and 1E6 that induce the GFP?

5. The experiments with lysosomal acidity using cDCFDA should be supported by another assay (may refer https://doi.org/10.52601/bpr.2021.200047 or similar studies).

6. The authors claim, "Consistent with this interpretation, many of the gold standards that failed to induce pgp-13 expression have EC50 concentrations for PLD-induction in human cells that far exceed the concentration at which we screened in worms (30 μM) (e.g., amantadine, memantine, spectinomycin (MORELLI et al. 2006)". They need to show in the main text that higher concentrations would induce pgp-13:gfp.

7. In Figure 7B and C, single and double of pgp-13 and pgp-14 should be added. Are they redundant?

Rev. 2:

Overall comments:

This study is a very nice combination of genetic, computational, biochemical, and ecological/evolutionary approaches. It addresses an interesting question from a drug discovery and safety perspective, and contributes impactfully to the field's very early-stage understanding of xenobiotic metabolism pathways in C. elegans. It is quite a lot of rigorously-done work, and clearly and fairly interpreted and discussed. It is clearly written and well-organized. I am not sure I have ever had so little to say in >25 years of reviewing manuscripts. Thank you!

Specific comments:

Figure 1 legend: "Three independent repeats were performed (N>3); an upper ceiling of 50 animals per sample were counted"—I think this should be "N = 3"? and please indicate also the lower range of animals per sample counted.

I recommend including a metadata file in the first tab of the Supplementary excel file.

Page 17 "dependent" typo: "We tested whether this increase in signal was dependant on NHR-70 and NHR-107 and found that it was (p<0.005)"

Rev. 3:

This is an interesting piece of work that uses an extensive range of experimental approaches in C elegans to probe mechanisms of xenobiotic detoxification. The take home message seems to be that the nematodes can mount a defence against CADs - this is a novel finding as I understand that such a mechanism has not been identified in other organisms. The most exciting prospect would be that these systems do exist in human but are as yet undiscovered with the possibility that this work might provide a route in.

The starting point for the study is a bit hard to follow. Is it really about PGP-14 as a model for human ABC4 deficiency? The paper writes about the pgp-14 phenotype - is this constrained to the resistance to wact-190? Or is it more extensive? If it's really about ABC4 and pgp-14 why not conduct a complementation experiment? i.e. can ABCB4 rescue pgp-14?

But rather I think that their screen has provided an interesting way to reveal the interaction between PGP-13, PGP-14, upstream NHR regulation and CAD metabolism.

If I understand correctly, the authors propose that in the pgp-14 ko CADS sensitize the worm to wact-190 because they upregulate pgp-13, which can then perform the same function in terms of promoting crystal formation. Furthermore, they reveal upstream regulation of this phenomenon. They then show that pgp-13 ko, like pgp-14 ko, can shift the dose-response for CAD lethality. I would like to suggest the authors rewrite some of the framing of this work to make it more 'digestible'. A cartoon would really help with the message.

Some more specific comments:

Why no further comment on tat-3?

Could the authors comment more explicitly on the baseline effect of the CADs in the study on wild-type? And how the dosing was chosen for the pgp-14 ko screen in this regard?

Fig 7. How are the worms identified as dead? Is it possible they are just immobilised? What was the stage of the worm? Adult? What was the exposure time?

Typos; page 9, 17 and possibly elsewhere- 'dependant' should be 'dependent'

Fig 1 legend D. should this really be 30mM?

P 23 typo 'Hence, we endeavored To'

PLoS Biol. 2026 Sep 21;24(9):e3004006. doi: 10.1371/journal.pbio.3004006.r003

Author response to Decision Letter 2

4 Jul 2026

Attachment

Submitted filename: CADs-Response to Reviews & Appendices-Final.pdf

pbio.3004006.s017.pdf^{(3.3MB, pdf)}

PLoS Biol. doi: 10.1371/journal.pbio.3004006.r004

Decision Letter 2

Ines Alvarez-Garcia

21 Aug 2026

Dear Dr Roy,

Thank you for your patience while we considered your revised manuscript entitled "A Cationic Amphiphilic Drug (CAD) Defense System in the Nematode Caenorhabditis elegans" for publication as a Research Article at PLOS Biology. This revised version of your manuscript has been evaluated by the PLOS Biology editors, the Academic Editor and one of the original reviewers.

Based on the comments (attached below), we are likely to accept this manuscript for publication, provided you satisfactorily address the remaining points raised by Reviewer 1. After consulting with the Academic Editor, we agree that all the negative results should be reported in the manuscript, but that the statement that "CADs may directly interact with the nuclear receptors" is fine. In point 4, you could replace "at least as good as" with "a good predictor" as 78% vs 73% is in the margin of error. You could also say that "C. elegans may increase its survival rate in the presence of CAD-producing microbes." Please also make sure to address the policy-related requests stated below my signature.

In addition, we would like you to consider a suggestion to improve the title:

"C. elegans have a cationic amphiphilic drug (CAD) defense system”

As you address these items, please take this last chance to review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the cover letter that accompanies your revised manuscript.

In addition to these revisions, you may need to complete some formatting changes, which you will receive in a follow up email. A member of our team will be in touch with a set of requests shortly. If you do not receive a separate email within a few days, please assume that checks have been completed, and no additional changes are required.

We expect to receive your revised manuscript within two weeks.

To submit your revision, please go to https://www.editorialmanager.com/pbiology/ and log in as an Author. Click the link labelled 'Submissions Needing Revision' to find your submission record. Your revised submission must include the following:

- a cover letter that should detail your responses to any editorial requests, if applicable, and whether changes have been made to the reference list

- a Response to Reviewers file that provides a detailed response to the reviewers' comments (if applicable, if not applicable please do not delete your existing 'Response to Reviewers' file.)

- a track-changes file indicating any changes that you have made to the manuscript.

NOTE: If Supporting Information files are included with your article, note that these are not copyedited and will be published as they are submitted. Please ensure that these files are legible and of high quality (at least 300 dpi) in an easily accessible file format. For this reason, please be aware that any references listed in an SI file will not be indexed. For more information, see our Supporting Information guidelines:

https://journals.plos.org/plosbiology/s/supporting-information

*Published Peer Review History*

Please note that you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out. Please see here for more details:

https://plos.org/published-peer-review-history/

*Press*

Should you, your institution's press office or the journal office choose to press release your paper, please ensure you have opted out of Early Article Posting on the submission form. We ask that you notify us as soon as possible if you or your institution is planning to press release the article.

*Protocols deposition*

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Please do not hesitate to contact me should you have any questions.

Sincerely,

Ines

Ines Alvarez-Garcia, PhD

Senior Editor

PLOS Biology

ialvarez-garcia@plos.org

------------------------------------------------------------------------

DATA POLICY:

Thank you for providing the data underlying all the graphs shown in the figures. Please ensure that the data file is named using the following format verbatim: S1 Data, and that is saved using exactly the following convention: S1_Data.xlsx (using an underscore).

In addition, please make sure that this file is mentioned in the figure legends of all the corresponding figures to state where the data can be found.

------------------------------------------------------------------------

CODE POLICY

Per journal policy, if you have generated any custom code during the course of this investigation, please make it available without restrictions. Please ensure that the code is sufficiently well documented and reusable, and that your Data Statement in the Editorial Manager submission system accurately describes where your code can be found. More information on our Code Policy, what and how to share can be found here: https://journals.plos.org/plosbiology/s/code-availability

Please note that we cannot accept sole deposition of code in GitHub, as this could be changed after publication. However, you can archive this version of your publicly available GitHub code to Zenodo. Once you do this, it will generate a DOI number, which you will need to provide in the Data Accessibility Statement (you are welcome to also provide the GitHub access information). See the process for doing this here: https://docs.github.com/en/repositories/archiving-a-github-repository/referencing-and-citing-content

------------------------------------------------------------------------

Reviewers' comments

Rev. 1:

I thank the authors for the extensive new work and for a rebuttal that is candid about experiments that did not succeed. My concern is narrow: four experiments were done at my request, all returned negative results, and none appear in the manuscript.

1. Three approaches — GAL4-LBD/luciferase in HepG2, AlphaFold3/Boltz-2 docking with PXR and VDR benchmarks, and LBD pull-downs with LC-MS, all returned negative results but are not reported in the manuscript; a reader would never know about these. "Remains an open question" no longer fairly describes what you know. Please add Appendices A and B as supplemental figures with brief text, including your caveats about absent worm-specific cofactors. New Figure 9 opens with "CADs may directly interact with the nuclear receptors." Please give the two arms of the model equal weight and cite the negative data in the legend.

2. pHTomato showed no change in response to any CAD or to 10 mM chloroquine. The failed positive control indicates that the assay was uninformative for intestinal LROs; this is worth publishing, as it will spare others the effort. Please report Figure R1. cDCFDA therefore remains the sole pathology readout, and phospholipidosis itself is still never measured. Please state in the Discussion that the relationship to PLD is inferred.

3. Amantadine, memantine and spectinomycin were tested at higher concentrations and none induced (Figure R2). The revision replaces the original claim with an uncheckable restatement of the same permeability hypothesis while the disconfirming data stay unpublished. Please include Figure R2 and state in the Results that the concentration explanation does not hold.

4. I accept the deferrals on ChIP-seq, motif analysis, heterodimerization, and identifying the Streptomyces metabolites. The claims should track them: the bacterial section supports NR-dependent induction by environmental bacteria, not yet an explanation for why the pathway evolved; please note that NR regulation of the targets is not shown to be direct; and please reframe "at least as good as a computational predictor," since the reporter recovers 68% of gold standards to the model's 73% on the same library the hits came from.

5. The one-tailed t-tests in Figures 5 and 6 need justification and multiple-comparison correction; please state whether hit scoring was blinded and why the myo-2p::mCherry control was dropped; and deposit the screen WGS and raw LCMS data.

In summary, the genetics is carefully conducted, the necessity/sufficiency case for the CYP-35Bs is convincing, and the reporter plus annotated compound set is a resource the field will use. But as it stands, the sensing step remains unresolved and now has three negative datasets against it; the pathology in the title is measured only by proxy; and the reporter's translational claim rests on a comparison whose main caveat has been experimentally excluded.

Without the additional validation described in points 1-3, I do not think the work supports the breadth of claim expected for this venue. It would be better suited to a specialist journal, such as PLOS Genetics or Genetics, where the negative results and limitations could be reported fully and would make a solid, useful contribution. I leave the broader question of general-interest significance to the editors, who are better placed to weigh it alongside the other reviewers' assessments.

PLoS Biol. 2026 Sep 21;24(9):e3004006. doi: 10.1371/journal.pbio.3004006.r005

Author response to Decision Letter 3

1 Sep 2026

Attachment

Submitted filename: CAD-response to second round reviews-1f.pdf

pbio.3004006.s018.pdf^{(232.4KB, pdf)}

PLoS Biol. doi: 10.1371/journal.pbio.3004006.r006

Decision Letter 3

Ines Alvarez-Garcia

4 Sep 2026

Dear Dr Roy,

Thank you for the submission of your revised Research Article entitled "The Nematode Caenorhabditis elegans has a Cationic Amphiphilic Drug (CAD) Defense System" for publication in PLOS Biology. On behalf of my colleagues and the Academic Editor, Mark Alkema, I am delighted to let you know that we can in principle accept your manuscript for publication, provided you address any remaining formatting and reporting issues. These will be detailed in an email you should receive within 2-3 business days from our colleagues in the journal operations team; no action is required from you until then. Please note that we will not be able to formally accept your manuscript and schedule it for publication until you have completed any requested changes.

Please take a minute to log into Editorial Manager at http://www.editorialmanager.com/pbiology/, click the "Update My Information" link at the top of the page, and update your user information to ensure an efficient production process.

PRESS

We frequently collaborate with press offices. If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximise its impact. If the press office is planning to promote your findings, we would be grateful if they could coordinate with biologypress@plos.org. If you have previously opted in to the early version process, we ask that you notify us immediately of any press plans so that we may opt out on your behalf.

We also ask that you take this opportunity to read our Embargo Policy regarding the discussion, promotion and media coverage of work that is yet to be published by PLOS. As your manuscript is not yet published, it is bound by the conditions of our Embargo Policy. Please be aware that this policy is in place both to ensure that any press coverage of your article is fully substantiated and to provide a direct link between such coverage and the published work. For full details of our Embargo Policy, please visit http://www.plos.org/about/media-inquiries/embargo-policy/.

Many congratulations and thanks again for choosing PLOS Biology for publication and supporting Open Access publishing. We look forward to publishing your study.

Sincerely,

Ines

Ines Alvarez-Garcia, PhD

Senior Editor

PLOS Biology

ialvarez-garcia@plos.org

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

S1 Fig. A schematic of mutations recovered in our forward genetic screen. A.

(TIFF)

pbio.3004006.s001.tiff^{(15.3MB, tiff)}

S2 Fig. A comparison of the relative positions of relevant PGPs in C. elegans and humans.

(TIFF)

pbio.3004006.s002.tiff^{(15.3MB, tiff)}

S3 Fig. PGP-13, NHR-70, and NHR-107 are necessary for CAD resensitization of pgp-14(0) to wact-190 crystal formation. A.

(TIFF)

pbio.3004006.s003.tiff^{(15.3MB, tiff)}

S4 Fig. Screening of CADs for NHR-70 and NHR-107 activation.

(TIFF)

pbio.3004006.s004.tiff^{(15.3MB, tiff)}

S5 Fig. Computational modeling of CAD interaction with NHR ligand binding domains.

(TIFF)

pbio.3004006.s005.tiff^{(15.3MB, tiff)}

S6 Fig. The mNeonGreen::PGP-13 translational fusion is functional.

(TIFF)

pbio.3004006.s006.tiff^{(15.3MB, tiff)}

S7 Fig. Gold standard CADs fail to induce pgp-13 reporter expression at higher concentrations.

(TIFF)

pbio.3004006.s007.tiff^{(15.3MB, tiff)}

S8 Fig. The qRT-PCR analysis of pgp-13, pgp-14, cyp-35b1, cyp-35b2, and cyp-35b3 transcript abundance in wild type, nhr-70(tm1697), and nhr-107(tm4512) animals.

(TIFF)

pbio.3004006.s008.tiff^{(15.3MB, tiff)}

S9 Fig. The accumulation of CAD parent molecules in yeast cells. A.

(TIFF)

pbio.3004006.s009.tiff^{(15.3MB, tiff)}

S10 Fig. A survey of the potential CYP35-B2 binding site residue contacts with exemplar CADs.

(TIFF)

pbio.3004006.s010.tiff^{(15.3MB, tiff)}

S11 Fig. Testing the utility of the PhsNUC-1::pHTomato reporter.

(TIFF)

pbio.3004006.s011.tiff^{(15.3MB, tiff)}

S12 Fig. Examples of the varied colony morphologies of different Streptomyces strains, including the hits (1E6 and 1G9) and exemplar non-hit (5E1).

(TIFF)

pbio.3004006.s012.tiff^{(15.3MB, tiff)}

S1 Data. Source data file.

The Source Data File contains the data associated with the display items. The first tab provides a navigational index.

(XLSX)

pbio.3004006.s013.xlsx^{(39MB, xlsx)}

S1 Raw Image. The raw western blot validation of FLAG-tagged protein expression that is cropped for S4 Fig.

(PDF)

pbio.3004006.s014.pdf^{(123.9KB, pdf)}

Attachment

Submitted filename: CADs-Response to Reviews & Appendices-Final.pdf

pbio.3004006.s017.pdf^{(3.3MB, pdf)}

Attachment

Submitted filename: CAD-response to second round reviews-1f.pdf

pbio.3004006.s018.pdf^{(232.4KB, pdf)}

Data Availability Statement

The data are in the figures, manuscript, and Source Data File.

[pbio.3004006.ref001] 1.Breiden B, Sandhoff K. Emerging mechanisms of drug-induced phospholipidosis. Biol Chem. 2019;401(1):31–46. doi: 10.1515/hsz-2019-0270 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref002] 2.Halliwell WH. Cationic amphiphilic drug-induced phospholipidosis. Toxicol Pathol. 1997;25(1):53–60. doi: 10.1177/019262339702500111 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref003] 3.Tummino TA, Rezelj VV, Fischer B, Fischer A, O’Meara MJ, Monel B, et al. Drug-induced phospholipidosis confounds drug repurposing for SARS-CoV-2. Science. 2021;373(6554):541–7. doi: 10.1126/science.abi4708 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref004] 4.Abe A, Shayman JA. The role of negatively charged lipids in lysosomal phospholipase A2 function. J Lipid Res 2009;50:2027–35. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref005] 5.Chen M, Borlak J, Tong W. High lipophilicity and high daily dose of oral medications are associated with significant risk for drug-induced liver injury. Hepatology. 2013;58(1):388–96. doi: 10.1002/hep.26208 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref006] 6.Chalasani N, Bonkovsky HL, Fontana R, Lee W, Stolz A, Talwalkar J, et al. Features and outcomes of 899 patients with drug-induced liver injury: the DILIN prospective study. Gastroenterology. 2015;148(7):1340-52.e7. doi: 10.1053/j.gastro.2015.03.006 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref007] 7.Dirven H, Vist GE, Bandhakavi S, Mehta J, Fitch SE, Pound P, et al. Performance of preclinical models in predicting drug-induced liver injury in humans: a systematic review. Sci Rep. 2021;11(1):6403. doi: 10.1038/s41598-021-85708-2 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref008] 8.Creanza TM, Delre P, Ancona N, Lentini G, Saviano M, Mangiatordi GF. Structure-based prediction of hERG-related cardiotoxicity: a benchmark study. J Chem Inf Model. 2021;61(9):4758–70. doi: 10.1021/acs.jcim.1c00744 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref009] 9.Sun H, Xia M, Shahane SA, Jadhav A, Austin CP, et al. Are hERG channel blockers also phospholipidosis inducers? Bioorg Med Chem Lett 2013;23:4587–90. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref010] 10.Vandenberg JI, Perry MD, Perrin MJ, Mann SA, Ke Y, Hill AP. hERG K(+) channels: structure, function, and clinical significance. Physiol Rev. 2012;92(3):1393–478. doi: 10.1152/physrev.00036.2011 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref011] 11.Pelletier DJ, Gehlhaar D, Tilloy-Ellul A, Johnson TO, Greene N. Evaluation of a published in silico model and construction of a novel Bayesian model for predicting phospholipidosis inducing potential. J Chem Inf Model. 2007;47(3):1196–205. doi: 10.1021/ci6004542 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref012] 12.Schieferdecker S, Eberlein A, Vock E, Beilmann M. Development of an in silico consensus model for the prediction of the phospholipigenic potential of small molecules. Comput Toxicol. 2022;22:100226. doi: 10.1016/j.comtox.2022.100226 [DOI] [Google Scholar]

[pbio.3004006.ref013] 13.Blumberg B, Sabbagh W Jr, Juguilon H, Bolado J Jr, van Meter CM, Ong ES, et al. SXR, a novel steroid and xenobiotic-sensing nuclear receptor. Genes Dev. 1998;12(20):3195–205. doi: 10.1101/gad.12.20.3195 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref014] 14.Kliewer SA, Moore JT, Wade L, Staudinger JL, Watson MA, Jones SA, et al. An orphan nuclear receptor activated by pregnanes defines a novel steroid signaling pathway. Cell. 1998;92(1):73–82. doi: 10.1016/s0092-8674(00)80900-9 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref015] 15.Geick A, Eichelbaum M, Burk O. Nuclear receptor response elements mediate induction of intestinal MDR1 by rifampin. J Biol Chem. 2001;276(18):14581–7. doi: 10.1074/jbc.M010173200 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref016] 16.Takeshita A, Taguchi M, Koibuchi N, Ozawa Y. Putative role of the orphan nuclear receptor SXR (steroid and xenobiotic receptor) in the mechanism of CYP3A4 inhibition by xenobiotics. J Biol Chem. 2002;277(36):32453–8. doi: 10.1074/jbc.M111245200 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref017] 17.Cermanova J, Fuksa L, Brcakova E, Hroch M, Kucera O, Kolouchova G, et al. Up-regulation of renal Mdr1 and Mrp2 transporters during amiodarone pretreatment in rats. Pharmacol Res. 2010;61(2):129–35. doi: 10.1016/j.phrs.2009.08.005 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref018] 18.Hodges LM, Markova SM, Chinn LW, Gow JM, Kroetz DL, Klein TE, et al. Very important pharmacogene summary: ABCB1 (MDR1, P-glycoprotein). Pharmacogenet Genomics. 2011;21(3):152–61. doi: 10.1097/FPC.0b013e3283385a1c [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref019] 19.McDonald MG, Au NT, Rettie AE. P450-based drug-drug interactions of amiodarone and its metabolites: diversity of inhibitory mechanisms. Drug Metab Dispos 2015;43:1661–9. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref020] 20.Faucette SR, Zhang T-C, Moore R, Sueyoshi T, Omiecinski CJ, LeCluyse EL, et al. Relative activation of human pregnane X receptor versus constitutive androstane receptor defines distinct classes of CYP2B6 and CYP3A4 inducers. J Pharmacol Exp Ther. 2007;320(1):72–80. doi: 10.1124/jpet.106.112136 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref021] 21.Boyd-Kimball D, Gonczy K, Lewis B, Mason T, Siliko N, et al. Classics in chemical neuroscience: chlorpromazine. ACS Chem Neurosci 2019;10:79–88. [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref022] 22.Burns AR, Luciani GM, Musso G, Bagg R, Yeo M, Zhang Y, et al. Caenorhabditis elegans is a useful model for anthelmintic discovery. Nat Commun. 2015;6:7485. doi: 10.1038/ncomms8485 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref023] 23.Kamal M, Knox J, Horne RI, Tiwari OS, Burns AR, Han D, et al. A rapid in vivo pipeline to identify small molecule inhibitors of amyloid aggregation. Nat Commun. 2024;15(1):8311. doi: 10.1038/s41467-024-52480-6 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref024] 24.Kamal M, Moshiri H, Magomedova L, Han D, Nguyen KCQ, Yeo M, et al. The marginal cells of the Caenorhabditis elegans pharynx scavenge cholesterol and other hydrophobic small molecules. Nat Commun. 2019;10(1):3938. doi: 10.1038/s41467-019-11908-0 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref025] 25.Kamal M, Tokmakjian L, Knox J, Han D, Moshiri H, Magomedova L, et al. PGP-14 establishes a polar lipid permeability barrier within the C. elegans pharyngeal cuticle. PLoS Genet. 2023;19(11):e1011008. doi: 10.1371/journal.pgen.1011008 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref026] 26.Kamal M, Tokmakjian L, Knox J, Mastrangelo P, Ji J, Cai H, et al. A spatiotemporal reconstruction of the C. elegans pharyngeal cuticle reveals a structure rich in phase-separating proteins. Elife. 2022;11:e79396. doi: 10.7554/eLife.79396 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref027] 27.Smit JJ, Schinkel AH, Oude Elferink RP, Groen AK, Wagenaar E, van Deemter L, et al. Homozygous disruption of the murine mdr2 P-glycoprotein gene leads to a complete absence of phospholipid from bile and to liver disease. Cell. 1993;75(3):451–62. doi: 10.1016/0092-8674(93)90380-9 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref028] 28.Prescher M, Kroll T, Schmitt L. ABCB4/MDR3 in health and disease - at the crossroads of biochemistry and medicine. Biol Chem. 2019;400(10):1245–59. doi: 10.1515/hsz-2018-0441 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref029] 29.Alam A, Kowal J, Broude E, Roninson I, Locher KP. Structural insight into substrate and inhibitor discrimination by human P-glycoprotein. Science. 2019;363(6428):753–6. doi: 10.1126/science.aav7102 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref030] 30.Nosol K, Romane K, Irobalieva RN, Alam A, Kowal J, Fujita N, et al. Cryo-EM structures reveal distinct mechanisms of inhibition of the human multidrug transporter ABCB1. Proc Natl Acad Sci U S A. 2020;117(42):26245–53. doi: 10.1073/pnas.2010264117 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref031] 31.Park HJ, Kim TH, Kim SW, Noh SH, Cho KJ, Choi C, et al. Functional characterization of ABCB4 mutations found in progressive familial intrahepatic cholestasis type 3. Sci Rep. 2016;6:26872. doi: 10.1038/srep26872 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref032] 32.Kampfer P. The Family Streptomycetaceae, Part I: Taxonomy in The Prokaryotes., edited by M. Dworkin, Falkow, S., Rosenberg, E., Schleifer, KH., Stackebrandt, E. Springer, New York, NY; 2006. [Google Scholar]

[pbio.3004006.ref033] 33.Samuel BS, Rowedder H, Braendle C, Félix M-A, Ruvkun G. Caenorhabditis elegans responses to bacteria from its natural habitats. Proc Natl Acad Sci U S A. 2016;113(27):E3941-9. doi: 10.1073/pnas.1607183113 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref034] 34.Robinson-Rechavi M, Maina CV, Gissendanner CR, Laudet V, Sluder A. Explosive lineage-specific expansion of the orphan nuclear receptor HNF4 in nematodes. J Mol Evol. 2005;60(5):577–86. doi: 10.1007/s00239-004-0175-8 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref035] 35.Sasaki S, Urabe M, Maeda T, Suzuki J, Irie R, Suzuki M, et al. Induction of hepatic metabolic functions by a novel variant of hepatocyte nuclear factor 4γ. Mol Cell Biol. 2018;38(24):e00213-18. doi: 10.1128/MCB.00213-18 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref036] 36.Vemuri K, Radi SH, Sladek FM, Verzi MP. Multiple roles and regulatory mechanisms of the transcription factor HNF4 in the intestine. Front Endocrinol (Lausanne). 2023;14:1232569. doi: 10.3389/fendo.2023.1232569 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref037] 37.Bookout AL, Jeong Y, Downes M, Yu RT, Evans RM, Mangelsdorf DJ. Anatomical profiling of nuclear receptor expression reveals a hierarchical transcriptional network. Cell. 2006;126(4):789–99. doi: 10.1016/j.cell.2006.06.049 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref038] 38.Sheps JA, Ralph S, Zhao Z, Baillie DL, Ling V. The ABC transporter gene family of Caenorhabditis elegans has implications for the evolutionary dynamics of multidrug resistance in eukaryotes. Genome Biol. 2004;5(3):R15. doi: 10.1186/gb-2004-5-3-r15 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref039] 39.Morelli JK, Buehrle M, Pognan F, Barone LR, Fieles W, Ciaccio PJ. Validation of an in vitro screen for phospholipidosis using a high-content biology platform. Cell Biol Toxicol. 2006;22(1):15–27. doi: 10.1007/s10565-006-0176-z [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref040] 40.Shahane SA, Huang R, Gerhold D, Baxa U, Austin CP, Xia M. Detection of phospholipidosis induction: a cell-based assay in high-throughput and high-content format. J Biomol Screen. 2014;19(1):66–76. doi: 10.1177/1087057113502851 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref041] 41.Bauch C, Bevan S, Woodhouse H, Dilworth C, Walker P. Predicting in vivo phospholipidosis-inducing potential of drugs by a combined high content screening and in silico modelling approach. Toxicol In Vitro. 2015;29(3):621–30. doi: 10.1016/j.tiv.2015.01.014 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref042] 42.Hinkovska-Galcheva V, Treadwell T, Shillingford JM, Lee A, Abe A, Tesmer JJG, et al. Inhibition of lysosomal phospholipase A2 predicts drug-induced phospholipidosis. J Lipid Res. 2021;62:100089. doi: 10.1016/j.jlr.2021.100089 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref043] 43.Daina A, Michielin O, Zoete V. SwissADME: a free web tool to evaluate pharmacokinetics, drug-likeness and medicinal chemistry friendliness of small molecules. Sci Rep. 2017;7:42717. doi: 10.1038/srep42717 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref044] 44.Shaker B, Lee J, Lee Y, Yu M-S, Lee H-M, Lee E, et al. A machine learning-based quantitative model (LogBB_Pred) to predict the blood-brain barrier permeability (logBB value) of drug compounds. Bioinformatics. 2023;39(10):btad577. doi: 10.1093/bioinformatics/btad577 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref045] 45.Rangaraju S, Solis GM, Thompson RC, Gomez-Amaro RL, Kurian L, Encalada SE, et al. Suppression of transcriptional drift extends C. elegans lifespan by postponing the onset of mortality. Elife. 2015;4:e08833. doi: 10.7554/eLife.08833 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref046] 46.Diot C, García-González AP, Vieira AF, Walker M, Honeywell M, Doyle H, et al. Bacterial diet modulates tamoxifen-induced death via host fatty acid metabolism. Nat Commun. 2022;13(1):5595. doi: 10.1038/s41467-022-33299-5 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref047] 47.Schmauder L, Richter K. hsp-90 and unc-45 depletion induce characteristic transcriptional signatures in coexpression cliques of C. elegans. Sci Rep. 2021;11(1):12852. doi: 10.1038/s41598-021-91690-6 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref048] 48.Larigot L, Mansuy D, Borowski I, Coumoul X, Dairou J. Cytochromes P450 of Caenorhabditis elegans: implication in biological functions and metabolism of xenobiotics. Biomolecules. 2022;12(3):342. doi: 10.3390/biom12030342 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref049] 49.de Groot MJ, Ackland MJ, Horne VA, Alex AA, Jones BC. Novel approach to predicting P450-mediated drug metabolism: development of a combined protein and pharmacophore model for CYP2D6. J Med Chem. 1999;42(9):1515–24. doi: 10.1021/jm981118h [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref050] 50.Wang A, Stout CD, Zhang Q, Johnson EF. Contributions of ionic interactions and protein dynamics to cytochrome P450 2D6 (CYP2D6) substrate and inhibitor binding. J Biol Chem. 2015;290(8):5092–104. doi: 10.1074/jbc.M114.627661 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref051] 51.Zhu M, Zhang H, Humphreys WG. Drug metabolite profiling and identification by high-resolution mass spectrometry. J Biol Chem. 2011;286(29):25419–25. doi: 10.1074/jbc.R110.200055 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref052] 52.Burns AR, Baker RJ, Kitner M, Knox J, Cooke B, Volpatti JR, et al. Selective control of parasitic nematodes using bioactivated nematicides. Nature. 2023;618(7963):102–9. doi: 10.1038/s41586-023-06105-5 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref053] 53.Knox J, Burns AR, Cooke B, Cammalleri SR, Kitner M, Ching J, et al. Cyprocide selectively kills nematodes via cytochrome P450 bioactivation. Nat Commun. 2024;15(1):5529. doi: 10.1038/s41467-024-49738-4 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref054] 54.Collins JB, Dilks CM, Hahnel SR, Rodriguez B, Fox BW, Redman E, et al. Naturally occurring variation in a cytochrome P450 modifies thiabendazole responses independently of beta-tubulin. PLoS Pathog. 2025;21(1):e1012602. doi: 10.1371/journal.ppat.1012602 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref055] 55.Wohlwend J, Corso G, Passaro S, Getz N, Reveiz M, et al. Boltz-1 Democratizing Biomolecular Interaction Modeling. bioRxiv. 2025.

[pbio.3004006.ref056] 56.Case DA, Cerutti DS, Cruzeiro VWD, Darden TA, Duke RE, et al. Recent Developments in Amber Biomolecular Simulations. J Chem Inf Model 2025;65:7835–43. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref057] 57.Clokey GV, Jacobson LA. The autofluorescent “lipofuscin granules” in the intestinal cells of Caenorhabditis elegans are secondary lysosomes. Mech Ageing Dev. 1986;35(1):79–94. doi: 10.1016/0047-6374(86)90068-0 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref058] 58.Bossinger O, Schierenberg E. The use of fluorescent marker dyes for studying intercellular communication in nematode embryos. Int J Dev Biol. 1996;40(1):431–9. doi: 10.1387/ijdb.8735958 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref059] 59.Hermann GJ, Schroeder LK, Hieb CA, Kershner AM, Rabbitts BM, Fonarev P, et al. Genetic analysis of lysosomal trafficking in Caenorhabditis elegans. Mol Biol Cell. 2005;16(7):3273–88. doi: 10.1091/mbc.e05-01-0060 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref060] 60.Baxi K, Ghavidel A, Waddell B, Harkness TA, de Carvalho CE. Regulation of lysosomal function by the DAF-16 forkhead transcription factor couples reproduction to aging in Caenorhabditis elegans. Genetics. 2017;207(1):83–101. doi: 10.1534/genetics.117.204222 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref061] 61.Edgar LG, McGhee JD. Embryonic expression of a gut-specific esterase in Caenorhabditis elegans. Dev Biol 1986;114:109–18. [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref062] 62.Nedergaard M, Desai S, Pulsinelli W. Dicarboxy-dichlorofluorescein: a new fluorescent probe for measuring acidic intracellular pH. Anal Biochem. 1990;187(1):109–14. doi: 10.1016/0003-2697(90)90425-9 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref063] 63.Nielsen IØ, Clemmensen KKB, Fogde DL, Dietrich TN, Giacobini JD, Bilgin M, et al. Cationic amphiphilic drugs induce accumulation of cytolytic lysoglycerophospholipids in the lysosomes of cancer cells and block their recycling into common membrane glycerophospholipids. Mol Biol Cell. 2024;35(3):ar25. doi: 10.1091/mbc.E23-06-0263 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref064] 64.Li X, Sun Y, Wang X. Probing lysosomal activity in vivo. Biophys Rep. 2021;7(1):1–7. doi: 10.52601/bpr.2021.200047 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref065] 65.Dirksen P, Marsh SA, Braker I, Heitland N, Wagner S, Nakad R, et al. The native microbiome of the nematode Caenorhabditis elegans: gateway to a new host-microbiome model. BMC Biol. 2016;14:38. doi: 10.1186/s12915-016-0258-1 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref066] 66.Ho LK, Nodwell JR. David and Goliath: chemical perturbation of eukaryotes by bacteria. J Ind Microbiol Biotechnol. 2016;43(2–3):233–48. doi: 10.1007/s10295-015-1686-6 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref067] 67.Daniel-Ivad M, Pimentel-Elardo S, Nodwell JR. Control of Specialized metabolism by signaling and transcriptional regulation: opportunities for new platforms for drug discovery? Annu Rev Microbiol. 2018;72:25–48. doi: 10.1146/annurev-micro-022618-042458 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref068] 68.van Helvoort A, Smith AJ, Sprong H, Fritzsche I, Schinkel AH, Borst P, et al. MDR1 P-glycoprotein is a lipid translocase of broad specificity, while MDR3 P-glycoprotein specifically translocates phosphatidylcholine. Cell. 1996;87(3):507–17. doi: 10.1016/s0092-8674(00)81370-7 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref069] 69.Raggers RJ, Vogels I, van Meer G. Multidrug-resistance P-glycoprotein (MDR1) secretes platelet-activating factor. Biochem J. 2001;357(Pt 3):859–65. doi: 10.1042/0264-6021:3570859 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref070] 70.Williamson B, Dooley KE, Zhang Y, Back DJ, Owen A. Induction of influx and efflux transporters and cytochrome P450 3A4 in primary human hepatocytes by rifampin, rifabutin, and rifapentine. Antimicrob Agents Chemother. 2013;57(12):6366–9. doi: 10.1128/AAC.01124-13 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref071] 71.Järvinen E, Hammer HS, Pötz O, Ingelman-Sundberg M, Stage TB. 3D spheroid primary human hepatocytes for prediction of cytochrome P450 and drug transporter induction. Clin Pharmacol Ther. 2023;113(6):1284–94. doi: 10.1002/cpt.2887 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref072] 72.Weber ND, Odriozola L, Martínez-García J, Ferrer V, Douar A, Bénichou B, et al. Gene therapy for progressive familial intrahepatic cholestasis type 3 in a clinically relevant mouse model. Nat Commun. 2019;10(1):5694. doi: 10.1038/s41467-019-13614-3 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref073] 73.Menzel R, Bogaert T, Achazi R. A systematic gene expression screen of Caenorhabditis elegans cytochrome P450 genes reveals CYP35 as strongly xenobiotic inducible. Arch Biochem Biophys. 2001;395(2):158–68. doi: 10.1006/abbi.2001.2568 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref074] 74.Aoyama Y, Yoshida Y, Sato R. Yeast cytochrome P-450 catalyzing lanosterol 14 alpha-demethylation. II. Lanosterol metabolism by purified P-450(14)DM and by intact microsomes. J Biol Chem 1984;259: 661–6. [PubMed] [Google Scholar]

[pbio.3004006.ref075] 75.Yoshida Y, Aoyama Y. Yeast cytochrome P-450 catalyzing lanosterol 14 alpha-demethylation. I. purification and spectral properties. J Biol Chem. 1984;259(3):1655–60. doi: 10.1016/s0021-9258(17)43458-2 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref076] 76.Yoon V, Nodwell JR. Activating secondary metabolism with stress and chemicals. J Ind Microbiol Biotechnol. 2014;41(2):415–24. doi: 10.1007/s10295-013-1387-y [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref077] 77.Saito S, Arai MA. Methodology for awakening the potential secondary metabolic capacity in actinomycetes. Beilstein J Org Chem. 2024;20:753–66. doi: 10.3762/bjoc.20.69 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref078] 78.Lewis JA, Fleming JT. Basic culture methods. Methods Cell Biol 1995;48:3–29. [PubMed] [Google Scholar]

[pbio.3004006.ref079] 79.Burns AR, Kwok TCY, Howard A, Houston E, Johanson K, Chan A, et al. High-throughput screening of small molecules for bioactivity and target identification in Caenorhabditis elegans. Nat Protoc. 2006;1(4):1906–14. doi: 10.1038/nprot.2006.283 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref080] 80.Kwok TCY, Ricker N, Fraser R, Chan AW, Burns A, Stanley EF, et al. A small-molecule screen in C. elegans yields a new calcium channel antagonist. Nature. 2006;441(7089):91–5. doi: 10.1038/nature04657 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref081] 81.Andrews S. FastQC: a quality control tool for high throughput sequence data. 2010. pp.

[pbio.3004006.ref082] 82.Li H, Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics 2009;25:1754–60. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref083] 83.McKenna A, Hanna M, Banks E, Sivachenko A, Cibulskis K, Kernytsky A, et al. The genome analysis toolkit: a mapreduce framework for analyzing next-generation DNA sequencing data. Genome Res. 2010;20(9):1297–303. doi: 10.1101/gr.107524.110 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref084] 84.Wang K, Li M, Hakonarson H. ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data. Nucleic Acids Res. 2010;38(16):e164. doi: 10.1093/nar/gkq603 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref085] 85.Riccio C. Extracting total RNA from Caenorhabditis elegans using phase-lock gel separation tubes. 2019.

[pbio.3004006.ref086] 86.Ye J, Coulouris G, Zaretskaya I, Cutcutache I, Rozen S, Madden TL. Primer-BLAST: a tool to design target-specific primers for polymerase chain reaction. BMC Bioinformatics. 2012;13:134. doi: 10.1186/1471-2105-13-134 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref087] 87.Schmittgen TD, Livak KJ. Analyzing real-time PCR data by the comparative C(T) method. Nat Protoc. 2008;3(6):1101–8. doi: 10.1038/nprot.2008.73 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref088] 88.Tao J, Hao Y, Li X, Yin H, Xie N, et al. Systematic Identification of housekeeping genes possibly used as references in Caenorhabditis elegans by large-scale data integration. Cells 2020;9. [DOI] [PMC free article] [PubMed]

[pbio.3004006.ref089] 89.Mello C, Fire A. DNA transformation. Methods Cell Biol 1995;48:451–82. [PubMed] [Google Scholar]

[pbio.3004006.ref090] 90.Abramson J, Adler J, Dunger J, Evans R, Green T, Pritzel A, et al. Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature. 2024;630(8016):493–500. doi: 10.1038/s41586-024-07487-w [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref091] 91.Passaro S, Corso G, Wohlwend J, Reveiz M, Thaler S, et al. Boltz-2: towards accurate and efficient binding affinity prediction. bioRxiv. 2025.

[pbio.3004006.ref092] 92.McNutt AT, Francoeur P, Aggarwal R, Masuda T, Meli R, et al., GNINA 1.0: molecular docking with deep learning. J Cheminform 2021;13:43. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref093] 93.McNutt AT, Li Y, Meli R, Aggarwal R, Koes DR. GNINA 1.3: the next increment in molecular docking with deep learning. J Cheminform 2025;17:28. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref094] 94.O'Boyle NM, Banck M, James CA, Morley C, Vandermeersch T, et al. Open Babel: An open chemical toolbox. J Cheminform 2011;3:33. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref095] 95.Cock PJA, Antao T, Chang JT, Chapman BA, Cox CJ, Dalke A, et al. Biopython: freely available Python tools for computational molecular biology and bioinformatics. Bioinformatics. 2009;25(11):1422–3. doi: 10.1093/bioinformatics/btp163 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref096] 96.Xue Y, Chao E, Zuercher WJ, Willson TM, Collins JL, Redinbo MR. Crystal structure of the PXR-T1317 complex provides a scaffold to examine the potential for receptor antagonism. Bioorg Med Chem. 2007;15(5):2156–66. doi: 10.1016/j.bmc.2006.12.026 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref097] 97.Rochel N, Wurtz JM, Mitschler A, Klaholz B, Moras D. The crystal structure of the nuclear receptor for vitamin D bound to its natural ligand. Mol Cell. 2000;5(1):173–9. doi: 10.1016/s1097-2765(00)80413-x [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref098] 98.Pearlman DA, Case DA, Caldwell JW, Ross WS, Cheatham TE III, DeBolt S, et al. AMBER, a package of computer programs for applying molecular mechanics, normal mode analysis, molecular dynamics and free energy calculations to simulate the structural and energetic properties of molecules. Computer Phy Commun. 1995;91(1–3):1–41. doi: 10.1016/0010-4655(95)00041-d [DOI] [Google Scholar]

[pbio.3004006.ref099] 99.Shahrokh K, Orendt A, Yost GS, Cheatham TE 3rd. Quantum mechanically derived AMBER-compatible heme parameters for various states of the cytochrome P450 catalytic cycle. J Comput Chem. 2012;33(2):119–33. doi: 10.1002/jcc.21922 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref100] 100.Wang J, Wolf RM, Caldwell JW, Kollman PA, Case DA. Development and testing of a general amber force field. J Comput Chem. 2004;25(9):1157–74. doi: 10.1002/jcc.20035 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref101] 101.Debiec KT, Cerutti DS, Baker LR, Gronenborn AM, Case DA, Chong LT. Further along the Road Less Traveled: AMBER ff15ipq, an original protein force field built on a self-consistent physical model. J Chem Theory Comput. 2016;12(8):3926–47. doi: 10.1021/acs.jctc.6b00567 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref102] 102.Koes DR, Vries JK. Evaluating amber force fields using computed NMR chemical shifts. Proteins. 2017;85(10):1944–56. doi: 10.1002/prot.25350 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref103] 103.Michaud-Agrawal N, Denning EJ, Woolf TB, Beckstein O. MDAnalysis: a toolkit for the analysis of molecular dynamics simulations. J Comput Chem. 2011;32(10):2319–27. doi: 10.1002/jcc.21787 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref104] 104.Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 2003;13(11):2498–504. doi: 10.1101/gr.1239303 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref105] 105.Pan X, Wang H, Li C, Zhang JZH, Ji C. MolGpka: a web server for small molecule pKa prediction using a graph-convolutional neural network. J Chem Inf Model. 2021;61(7):3159–65. doi: 10.1021/acs.jcim.1c00075 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref106] 106.Montalibet J, Kennedy BP. Using yeast to screen for inhibitors of protein tyrosine phosphatase 1B. Biochem Pharmacol. 2004;68(9):1807–14. doi: 10.1016/j.bcp.2004.06.024 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref107] 107.Gietz RD, Schiestl RH. High-efficiency yeast transformation using the LiAc/SS carrier DNA/PEG method. Nat Protoc. 2007;2(1):31–4. doi: 10.1038/nprot.2007.13 [DOI] [PubMed] [Google Scholar]

[pbio.3004006.ref108] 108.Harris TW, Arnaboldi V, Cain S, Chan J, Chen WJ, Cho J, et al. WormBase: a modern model organism information resource. Nucleic Acids Res. 2020;48(D1):D762–7. doi: 10.1093/nar/gkz920 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3004006.ref109] 109.Kent WJ, Sugnet CW, Furey TS, Roskin KM, Pringle TH, Zahler AM, et al. The human genome browser at UCSC. Genome Res. 2002;12(6):996–1006. doi: 10.1101/gr.229102 [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

The nematode Caenorhabditis elegans has a cationic amphiphilic drug (CAD) defense system

Levon Tokmakjian

Duhyun Han

Kateryna Sihuta

Aanchal Aggarwal

Somayeh Pirhadi

Yao Wang

Andrew R Burns

Brittany Cooke

Siyue Ren

Marios Gavrielatos

Jiabao Liu

Henry M Krause

Carolyn L Cummins

Justin Nodwell

David R Koes

Peter J Roy

Roles

Abstract

Introduction

Results

A Screen for small molecules that Rescue pgp-14(0) yields multiple CADs

Fig 1. Drug suppressors of pgp-14(0).

CADs compensate for the loss of PGP-14 through the upregulation of PGP-13

Fig 2. Components of a putative CAD-defense system.

Fig 9. A Model of the CAD Defense Pathway.

Fig 3. A Screen of 2,400 spectrum small molecules reveals 293 inducers of pgp-13 expression.

Most pgp-13 inducers require NHR-70 and NHR-107

Fig 4. A Screen of 195 small molecules for their ability to induce pgp-13 expression in an NHR-70 and NHR-107-dependent manner.

Metabolic analyses reveals a CAD detoxification system

Fig 5. Metabolic analyses of CADs.

Computational modeling reveals CYP-35B2 D311 as key to in CAD metabolism

Fig 6. CYP-35B2 D311 is important for CAD metabolism.

The CAD detoxification pathway protects against pathogenesis

Fig 7. CADs induce pathology in C. elegans.

Soil microbes elicit the CAD defence response

Fig 8. Induction of the CAD defense system by Streptomyces bacterial strains.

Discussion

Methods

C. elegans husbandry, screening and microscopy

The screen and initial characterization of pgp-14(0) small molecule suppressors

A forward genetic screen for mutants that suppress the ability of CADs to restore pgp-14(0) sensitivity to the lethal effects of wact-190 crystal formation

RNA extraction, cDNA synthesis and quantitative real-time PCR

Construction and analyses of pgp-13 transcriptional and translational reporters

Analyses of pgp-13 temporal expression and dose-response

Analyses of pgp-13 Localization

Cell-based assay testing CAD interaction with NHRs

Computational modeling of CAD interactions with NHR-70 and NHR-107

Computational modeling of CAD interactions with cytochrome P450s

A screen of the spectrum library for small molecules that induce pgp-13 expression

Generating a structural similarity map

Establishing a predictive model of phospholipidosis(PLD)-inducing CADs

Spectrum MiniLibrary analyses

Sample preparation of C. elegans for mass spectrometry analyses

Expression of wild-type and mutant C. elegans CYP-35B P450s in the yeast S. cerevisiae

Sample preparation of S. cerevisiae for mass spectrometry analyses

Liquid chromatography–quadrupole time‑of‑flight mass spectrometry

CAD sensitivity analysis

Analysis of gut granules using cDCFDA

Use of the pHTomato reporter

The screening and characterization of actinomyces bacterial strains for CAD reporter induction

Supporting information

Acknowledgments

Abbreviations

Data Availability

Funding Statement

References

Decision Letter 0

Ines Alvarez-Garcia

Roles

Decision Letter 1

Ines Alvarez-Garcia

Roles

Author response to Decision Letter 2

Decision Letter 2

Ines Alvarez-Garcia

Roles

Author response to Decision Letter 3

Decision Letter 3