The marginal cells of the Caenorhabditis elegans pharynx scavenge cholesterol and other hydrophobic small molecules

Abstract

The nematode Caenorhabditis elegans is a bacterivore filter feeder. Through the contraction of the worm's pharynx, a bacterial suspension is sucked into the pharynx's lumen. Excess liquid is then shunted out of the buccal cavity through ancillary channels made by surrounding marginal cells. We find that many worm-bioactive small molecules (a.k.a. wactives) accumulate inside of the marginal cells as crystals or globular spheres. Through screens for mutants that resist the lethality associated with one crystallizing wactive we identify a presumptive sphingomyelin-synthesis pathway that is necessary for crystal and sphere accumulation. We find that expression of sphingomyelin synthase 5 (SMS-5) in the marginal cells is not only sufficient for wactive accumulation but is also important for absorbing exogenous cholesterol, without which C. elegans cannot develop. We conclude that sphingomyelin-rich marginal cells act as a sink to scavenge important nutrients from filtered liquid that might otherwise be shunted back into the environment.

Fig. 1

Wactives can form objects in the marginal cells of the anterior pharynx. a A chart showing our survey of the effects of 238 wactives on C. elegans L1 animals. Each row is a distinct wactive. The first column indicates whether a wactive accumulates as a crystal, sphere, or neither at a concentration of 30 μM. The second column indicates whether the wactive also kills/arrests worm development at a concentration of 30 μM. The percentages in the legend refer to the percentage of animals harboring the indicated objects. See “Methods” for details. b Examples of the accumulation of crystals and spheres in the anterior pharynx of young adults treated for 24 h to allow easy visualization of the channels. The first row shows differential interference contrast images; the second row shows whether the objects are birefringent (and therefore crystalline); and the third row shows whether the object fluoresces. White arrows indicate the channels, blue arrows indicate crystals, and orange arrows indicate spheres. The white asterisk indicates where the micrograph was spliced to ensure channel is in the correct focal plane. The scale is the same for all panels. c A schematic of C. elegans (courtesy of WormAtlas). The anterior pharynx is boxed. d Violin plots of the physical–chemical properties of molecules that crystallize (cr) (n = 38 distinct molecules), form spheres (sp) (n = 33 distinct molecules), or fail to form objects in the animal (no) (n = 152 distinct molecules). The gray shape represents all results and the thickness indicates how common that value is within the dataset; a white circle indicates the median; a red circle indicates the mean; the center thick line represents one standard of deviation; the thinner line represents the second standard of deviation. The number above each plot shows the significance of the difference compared to the “no-object” data using the nonparametric two-tailed Mann–Whitney U test. Source data for a and d are provided in the Source Data file

Fig. 2

Wactive-induced lethality and object formation are correlated. a Crystal and sphere formation in the pharynx was analyzed after a 48-h incubation of L1 larvae in the respective wactive and is indicated in blue. The lethality induced by the respective wactive, assayed after 50 L1s were incubated with the molecules for 6 days, is indicated in red. Animals in each well were counted if not over-grown with worms; otherwise, the well was scored as 100 animals. Low numbers of animals (see wact-190 or wact-128, for example) are invariably arrested L1s. The twofold serial dilution dose of the wactive, in micromolar, is indicated at the bottom of the three columns of graphs. The small molecule structure is illustrated below each graph. b–e An analysis of the ability of Evans Blue dye (EBD) to penetrate the anterior pharynx. L1-stage worms were incubated with control (1% dimethyl sulfoxide (DMSO) only) or 60 µM of the indicated compound for 24 h. Worms were then incubated in EBD for 4 h. b–d We see an enrichment of EBD signal in the anterior pharynx of worms treated with the crystal-forming wact-190 but not with the sphere-forming wact-34 or control. The scale is the same for all panels. For each pair of images, the differential interference contrast image is on the left, and the fluorescent image showing EBD signal is on the right. e Quantification of the EBD signal intensity of wact-190-treated worms relative to the DMSO negative control treated worms. Asterisk indicates a significant difference between the control and the wact-190-treated sample (p < 0.05) using Student’s T test. Means and standard error are derived from n = 3 independent biological trials with at least six animals analyzed per sample per trial. Error bars represent standard error of the mean (SEM) for all graphs in this figure. Source data for a and e are provided in the Source Data file

Fig. 3

A microscopic analysis of wact-190 crystal formation. a A schematic of the pharynx with the three channels indicated with white arrows; the central lumen and grinder are in black. The procorpus of the anterior pharynx, which is shown in b, is boxed in orange. b Qualitative analysis of wact-190 crystal formation over time. Synchronized fourth-stage (L4) animals were grown in 60 μM of wact-190 in liquid culture for the time indicated at the top of each column. The top images show differential interference contrast micrographs; the corresponding bottom images are taken with polarized filters to visualize birefringence. Each pair of images shows the procorpus (anterior quarter) of the pharynx. The white arrows indicate channels; green arrows indicate select crystals. The scale is the same for all panels. c Quantitative analysis of wact-190 crystal formation over time. Synchronized first-stage (L1) worms of the indicated genotype were incubated in liquid with 60 μM wact-190 for the indicated time period and crystals were identified by their birefringence. Error bars represent standard error of the mean (SEM) and n = 3 independent biological samples. Source data are provided in the Source Data file. d–f Transmission electron micrographs (TEMs) of pharynx cross-sections of the indicated genotype and wactive treatment. The channels are indicated with white arrows; presumptive crystals with green arrows; marginal cell plasma membranes with fuchsia arrows. See Supplementary Fig. 5 for transverse sections. g A schematic of the pharynx illustrating the approximate area of the cross-sections imaged by TEM. The marginal cells are indicated in fuchsia. Image adapted from WormAtlas with permission

Fig. 4

Mutant genes that resist the lethality induced by wact-190. Mutants listed in rows 1–17 were isolated in our forward genetic screen for those that resist the lethal/arrest phenotype of wact-190. In addition to these 17, we identified 29 other strains with a mutant allele of pgp-14 that will be described elsewhere. Additional mutants beyond row 17 are deletion alleles obtained from either the C. elegans Genetics Centre or from Shohei Mitani. While all of the mutants isolated in our screen resist wact-190 on solid substrate, only some also resist wact-190 effects in liquid culture. F53B1.2 is a paralog of the sms genes, and F33D4.4 is a paralog of ttm-5. Viability assays were done with at least four replicates (see “Methods” for details). Residue numbers for both isoforms of sptl-2 are shown. Protein changes noted in red indicate a presumptive null allele; “X” indicates an early non-sense codon. For five wact-190-resistant strains (rows 13–17), the mutant gene responsible for wact-190 resistance has not been identified and are not discussed further. Source data and homozygous mutation information for the relevant strains are provided in the Source Data file

Fig. 5

Mutations in a predicted sphingomyelin synthesis pathway confer resistance to wact-190. a A predicted sphingomyelin synthesis pathway. Whole-genome sequence analysis indicated that mutations in three genes predicted to play a role in sphingomyelin synthesis (sptl-2, ttm-5, and sms-5) are sufficient to confer resistance to wact-190 (see Fig. 4). Independently derived deletion alleles were tested for each of these genes (in green), together with their respective paralogs (in black), and only the deletions of sptl-2, ttm-5, and sms-5 confer resistance (see Fig. 4). b, c 60 μM wact-190 forms crystals in wild-type worms but not in the sms-5(ok2498) mutant. Animals were incubated in the small molecule for 24 h as L4-staged animals. Anterior is up; white and green arrows indicate channels and crystals, respectively. d Quantification of wact-190’s effects on viability and crystal formation. The sms-5-containing fosmid (WRM0626dC03) is harbored as an extra-chromosomal array. The construct expressing SMS-5::FLAG::mCherry from an anterior marginal cell-specific promoter (SMS-5 (MC)) (see Supplementary Fig. 7) is harbored by the transgenic array (trIs104). Viability is quantified with n = 8 independent experiments; crystal counts were done with n = 3 independent experiments. In samples where viability is indicated as <40, the respective wells have only the original young larvae that are arrested or dead. e Accumulation of the indicated small molecule in sms-5 mutants relative to wild-type animals as measured by mass spectrometry. n = 3 independent experiments. Asterisk (*) indicates a significant difference (p < 0.001) using a Student’s T-test. f, g Two focal planes of an adult mosaic for the extra-chromosomal (Ex) array harboring pPRHM1051 (a construct with the SMS-5 fosmid WRM0626dC03 with yellow fluorescent protein (YFP) coding sequence inserted in frame at the C-terminus of SMS-5). The YFP-fusion protein is localized to the anterior marginal cells and is enriched basally at the borders with adjacent muscle cells (white arrows). Supplementary Fig. 6 shows the non-mosaic expression pattern of SMS-5. For each pair of images, fluorescent image is on the left and the differential interference contrast image is on the right. The scale is the same for all corresponding panels. Error bars represent standard error of the mean (SEM) for all graphs in this figure. Source data for d and e are provided in the Source Data file

Fig. 6

sms-5 mutants have less sphingomyelin in the anterior pharynx. a, b The sms-5 deletion allele is hypersensitive to the lethality induced by detergents triton-X 100 and IGEPAL CA-630, shown as a v/v percentage of the liquid media. In both cases, differences are significant (p < 0.05) at a detergent concentration of 0.8% (w/v) using a Student’s T test. SMS-5(MC) represents marginal-cell-specific expression of SMS-5::FLAG::mCherry from the trIs104 integrated transgenic array. Three biological repeats were done with four technical repeats done during each biological repeat. c, d Wild-type and sms-5(ok2498) mutant animals are fixed and stained with the GFP-NT-Lysenin protein probe that recognizes clustered sphingomyelin. White arrows indicate buccal cavity and white arrowheads indicate anterior channels. For each pair of images, the differential interference contrast image is on the left, and the fluorescent image is on the right. The scale is the same for all panels. e Quantification of the relative green fluorescent protein (GFP) signal from more than five biological replicates of GFP-NT-Lysenin staining (see “Methods” for details). Asterisk (*) indicates a significant difference (p = 0.024) using a Student’s T test. n = 7 independent trials with at least 10 worms analyzed during each trial. Error bars represent standard error of the mean (SEM) for all graphs in this figure. Source data for a, b, and e are provided in the Source Data file

Fig. 7

sms-5 mutants have altered sensitivity to small molecules. a L1 viability assays of animals of the indicated genotype (top) grown in liquid culture in quadruplicate in 508 different wactive small molecules at 7.5, 30, or 60 μM. Wells were inspected after 6 days of growth; wells that had ≥50 animals are shown in green, wells that have between 11 and 50 animals are shown in yellow, and wells with ≤11 animals are shown in red. The resulting population growth of each of the four replicates is shown in each column. Each row corresponds to a distinct wactive molecule. The data are clustered along the y axis. The 24 wactives with reduced potency in the mutants are referred to as resistant wactives, and the 169 wactives with increased potency in the mutants are referred to as hypersensitive wactives. b–e Dose–response analyses of C. elegans population growth with two resistant molecules (wact-190 and wact-416) and two hypersensitive molecules (wact-455 and wact-572). The sms-5(ok2498) deletion mutant (“sms-5(0)”) was used in this analysis, along with the rescuing sms-5 transgene whereby SMS-5::FLAG::mCherry (SMS-5(MC)) is expressed exclusively in the anterior marginal cells. Resulting wells with <50 animals in the wact-190 and wact-416 conditions were invariably arrested or dead L1s. Viability counts were done with n = 8 independent experiments; crystal counts were done with n = 3 independent experiments. f Accumulation of the indicated small molecule in the sms-5(ok2498) mutant relative to wild-type L1 animals as measured by mass spectrometry. Standard error of the mean is shown for graphs b–f. n = 6 independent experiments for wact-190, wact-171, wact-498, wact-406, and wact-519; n = 4 for wact-455. g Violin plots comparing the physical–chemical properties of molecules that are classified as lethal in all conditions (dead), resistant, hypersensitive, or not obviously bioactive (alive) at 30 μM. See the legend of Fig. 1 for the description of the violin plots. One, two, and three asterisks indicate a significant difference (p < 0.05, p < 0.01, and p < 0.001, respectively) between the indicated datasets using Student’s T test. Source data are provided in the Source Data file

Fig. 8

A comparison of the lethality and object-forming capability of wactives in animals lacking SMS-5. a The behavior of molecules that form objects in at least 25% of the population are analyzed with respect to their lethal potency in wild-type and the sms-5(ok2498) deletion mutant. The wactive number is shown on the right and black arrowheads indicate the wactives analyzed in b. b. Dose–response analyses for select crystal and sphere-forming wactives in the indicated genetic background (indicated at the top of the two columns of graphs). Each of the wactives is analyzed over a twofold dilution series of concentrations (indicated at the bottom of the three columns of graphs). Wactive molecules are indicated on the left. The standard error of the mean is shown. Source data are provided in the Source Data file

Fig. 9

SMS-5 is required for development in cholesterol-limited conditions. a The growth rate of animals of the indicated genotype grown on cholesterol (xol)-limited plates (containing 50 ng/mL xol) is reported relative to the same strains grown on plates with standard concentrations of cholesterol (5000 ng/mL xol) (y axis). See “Methods” for additional details. Asterisks indicate significance (p < 0.001) relative to the wild-type (wt) control; the red asterisks indicate significance (p < 0.05) relative to the sms-5(ok2498) deletion mutant. The double mutant has a more severe growth defect relative to either single mutant (p = 0.06). n = 10 independent biological trials for the control (wt); n = 3 independent biological trials for all experimental genotypes. b Accumulation of NBD-cholesterol signal of the indicated mutants relative to wild-type controls. Black asterisks indicate significance (p < 0.01) relative to the wild-type control; red asterisk indicates significance (p < 0.05) relative to the sms-5(ok2498) deletion mutant. n = 6 independent biological trials for wt and the sms-5 mutant; n = 3 independent biological trials for the other two strains. In both a and b, standard error of the mean is shown and significance is measured using Student’s T test. Source data are provided in the Source Data file. c–f Images of multiple animals after the indicated strains were incubated in NBD-cholesterol for 6 days. The scale is the same for all images. Green lines exemplify the area used in each animal to calculate signal in b. Red arrow indicates the expression of yellow fluorescent protein that is restricted to the anterior pharynx that is used as a marker to follow the rescuing transgene trIs104. In a, b, and f, SMS-5(MC) represents marginal-cell-specific expression of SMS-5 from the trIs104 integrated transgenic array. Supplementary Fig. 9 shows that, in the absence of NBD-cholesterol, each strain auto-fluoresces in the green fluorescent protein channel to the same extent

Fig. 10

A model of SMS-5’s role in small molecule accumulation. a SMS-5 is key in synthesizing sphingomyelin (SM) on the outer leaflet of the plasma membrane (PM) of the marginal cells of the anterior pharynx. The SM-rich membrane acts as a sink for relatively larger hydrophobic crystallizing wactives but a barrier to comparatively smaller molecules. b Over time, the hydrophobic wactives precipitate out of solution within the SM-rich PM of the wild-type marginal cells of the anterior pharynx. c The absence of SMS-5 results in less SM in the PM, which in turn reduces the accumulation of relatively large hydrophobic molecules. Because of decreased molecular density of a SM-poor PM, relatively smaller molecules can now penetrate the PM barrier