Strains with variants or expression differences in ben-1 were predicted to be ABZ resistant

To predict ABZ resistant strains in the three Caenorhabditis species, we identified high-impact variants (SNVs, INDELs, or SVs predicted to disrupt beta-tubulin function) in ben-1 (see Materials and Methods). In a set of 611 C. elegans wild strains [35], we identified 65 strains (10.6% of all strains) with 33 unique high-impact variants in ben-1 (S1 Table). Of the 33 variants, 24 were previously phenotyped and 20 were associated with ABZ resistance (S1A Fig) [14,36]. In 641 C. briggsae wild strains, 22 strains (3.43% of all strains) with eight unique high-impact variants in ben-1 were identified (S2 Table). In a set of 518 C. tropicalis wild strains, only two strains (0.39% of all strains) with unique high-impact variants in ben-1 were identified (S3 Table).

Next, because low ben-1 expression was previously correlated with ABZ resistance in C. elegans wild strains [37], we evaluated whether ben-1 expression levels were predictive of ABZ responses in C. elegans. We assessed the relationship between the expression of ben-1 [37] and ABZ responses in 180 C. elegans wild strains [14, 36] (p-value = 5.16e-16, r² = 0.344). We hypothesized that low ben-1 expression could contribute to ABZ resistance in C. briggsae and C. tropicalis, but expression data from wild strains in these species have yet to be collected.

Altogether, we identified nine novel variants in Cel-ben-1 (E3stop, Y50C, P80S, VDN113N, Q131L, frameshift 319, frameshift 368, stop445S, and a duplication) (S1 Table), one strain with low Cel-ben-1 expression, eight variants in Cbr-ben-1 (W21stop, V91I, Q94K, D128E, Q134H, S218L, M299V, and R359H) (S2 Table), and two variants in Ctr-ben-1 (P80T and R121Q) (S3 Table) to test for ABZ resistance. In addition to these selected strains, we also selected C. briggsae and C. tropicalis strains that lacked high-impact variants in beta-tubulin genes but were closely related to strains with variants (S2, S3 Figs, and S4 Table). The strains with no high-impact variants in any beta-tubulin gene might control for differences in wild strain genetic backgrounds.

High-throughput larval development assays (HTLDAs) reveal species-specific associations between ben-1 variation and ABZ resistance

Nematodes grow longer as they progress through development, and BZs slow this progression [14– 16,32,36]. A longer animal length (i.e., larger animals) corresponds to increased ABZ resistance, and a shorter animal length (i.e., smaller animals) corresponds to increased ABZ sensitivity. To evaluate the effects of ABZ on animal length (a proxy for development), we used image-based HTLDAs to expose all C. elegans, C. briggsae, and C. tropicalis wild strains with novel high-impact variants in ben-1 or low ben-1 expression (i.e., predicted resistant strains) to control (DMSO) (S4, S5, and S6 Figs) and drug (ABZ) conditions (S7, S8, and S9 Figs) (see Materials and Methods). The assay included 48 replicates per strain with five to 30 animals per replicate in DMSO or ABZ conditions. In DMSO conditions, strains naturally varied in animal length, but all animal measurements were categorized as the L4 larval stage by our custom CellProfiler models [38,39], indicating all strains underwent normal development in control conditions. In ABZ conditions, the reported nematode length (i.e., normalized animal length (μm)) of each strain is the difference between animal lengths in DMSO and ABZ.

To define the role of ben-1 in BZ response and classify wild strains as BZ resistant, we first obtained or created ben-1 deletions in the reference strains for each of the three Caenorhabditis species. In C. elegans, the strain with a ben-1 LoF variant in the N2 strain background (ECA882) has been used as an ABZ-resistant control previously [14,32,36]. For C. briggsae and C. tropicalis, we used CRISPR-Cas9 genome editing to generate two independent ben-1 deletion alleles per species (see Materials and Methods). In C. briggsae, deletions of Cbr-ben-1 were created in the AF16 reference strain background (ECA3953 and ECA3954) (S10 Fig and S5 Table). In C. tropicalis, deletions of Ctr-ben-1 were created in the NIC58 reference strain background (ECA4247 and ECA4248) (S11 Fig and S5 Table). All deletions of ben-1 in each of the three Caenorhabditis species conferred high levels of ABZ resistance (S12 and S13 Figs). Wild strains were then classified as resistant using a threshold where the median animal lengths after ABZ exposure were no more than two SD below the median animal length of the species-specific ben-1 deletion strain.

In C. elegans, seven of the nine strains with ben-1 variants showed minimal developmental delays after exposure to ABZ, a phenotype similar to loss of ben-1, and were classified as resistant (Fig 1). The two remaining strains with ben-1 variants (P80S and stop445S) and the strain with low ben-1 expression were not strongly resistant to ABZ. The P80S variant might partially alter ben-1 function, causing a moderate resistance phenotype. In stop445S, the normal stop codon was replaced with a serine. This variant likely does not affect ben-1 function because position 445 is at the end of the BEN-1 protein, and only four amino acid residues (NRKL) are added beyond the wild-type stop codon. Finally, the strain with low ben-1 expression and no high-impact variants in ben-1 (JU1581) was sensitive to ABZ, indicating that the selected threshold of ben-1 expression (≥3.75 TPM) still retained strains with adequate ben-1 function. Overall, natural allelic variation in Cel-ben-1 is associated with BZ resistance, recapitulating previous findings that BZ resistance is associated with a diverse set of LoF variants in Cel-ben-1 [14,29,32,36].

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. High-throughput larval development assays in the presence of albendazole for C. elegans strains with unique high-impact variants in BEN-1.

The regressed median animal length values for populations of nematodes grown in 30 μM albendazole (ABZ) are shown on the y-axis. Each point represents the normalized median animal length value (i.e., the difference between animal lengths in DMSO and ABZ for each strain) of a well containing approximately five to 30 animals. Data are shown as Tukey box plots with the median as a solid horizontal line, and the top and bottom of the box representing the 75th and 25th quartiles, respectively. The top whisker is extended to the maximum point that is within the 1.5 interquartile range from the 75th quartile. The bottom whisker is extended to the minimum point that is within the 1.5 interquartile range from the 25th quartile. The gray dashed line marks the C. elegans resistance threshold, defined as two standard deviations below the mean of the ben-1 deletion strain in the N2 reference strain background. Results for (A) the N2 reference strain (orange) and a strain with a ben-1 deletion in the N2 background (red), and (B) all wild C. elegans strains with unique high-impact variants in ben-1 are sorted by their relative resistance to ABZ based on median animal length. Wild C. elegans strains are colored by beta-tubulin variant status. An asterisk indicates a stop gained at the indicated BEN-1 position.

https://doi.org/10.1371/journal.ppat.1014306.g001

In C. briggsae, we performed HTLDAs on the AF16 reference strain (ABZ sensitive), two ben-1 deletion strains (ECA3953 and ECA3954) (ABZ resistant), 11 strains with eight unique variants in ben-1, and 13 strains genetically related to strains with ben-1 variants in DMSO (control) (S5 Fig) and ABZ conditions (Figs 2, S8 and S12). Of the 11 strains with Cbr-ben-1 variants, only two strains with unique variants in BEN-1 (Q134H and W21stop) were ABZ resistant (Fig 2). The Q134H amino acid change has been associated with ABZ resistance in A. caninum and validated in C. elegans [11]. An early stop gain at position 21 is predicted to cause the premature termination of protein synthesis and LoF. Of all the assayed strains with variants in BEN-1, six (V91I, Q94K, D128E, S218L, M299V, and R359H) were not resistant to ABZ. Additionally, eleven C. briggsae strains were exposed to a maximum concentration of 120 μM ABZ, where only two strains with the variants Q134H and W21stop were resistant at these higher concentrations (Fig 3). Overall, because only one of the seven missense variants was associated with ABZ resistance, we could not reliably predict C. briggsae ABZ resistance based on the presence of a missense variant alone. To evaluate the potential fitness consequences of ben-1 LoF alleles, fecundity assays were performed between the AF16 reference strain and the two strains with a loss of ben-1 in the AF16 background (ECA3953 and ECA3954) (Fig 4A and S6 Table). We found no significant differences in fecundity between the three strains, indicating that a loss of ben-1 does not affect fitness in this species. Similar results have been reported in C. elegans, where loss of Cel-ben-1 has no effect on fitness [36]. These results indicate that, although a loss of Cbr-ben-1 confers ABZ resistance with no fitness detriment, ABZ resistance associated with Cbr-ben-1 variants is uncommon.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. High-throughput larval development assays in the presence of albendazole for C. briggsae strains with unique high-impact variants in BEN-1.

The regressed median animal length values for populations of nematodes grown in 30 μM albendazole (ABZ) are shown on the y-axis. Each point represents the normalized median animal length value of a well containing approximately five to 30 animals. Data are shown as Tukey box plots with the median as a solid horizontal line, and the top and bottom of the box representing the 75th and 25th quartiles, respectively. The top whisker is extended to the maximum point that is within the 1.5 interquartile range from the 75th quartile. The bottom whisker is extended to the minimum point that is within the 1.5 interquartile range from the 25th quartile. The gray dashed line marks the C. briggsae resistance threshold, defined as two standard deviations below the mean of the ben-1 deletion strain in the AF16 reference strain background. Results for (A) the AF16 reference strain (green) and a strain with a ben-1 deletion in the AF16 background (red), and (B) all wild C. briggsae strains with unique high-impact variants in ben-1 are sorted by their relative resistance to ABZ based on median animal length. No variant (N. V.) strains (gray) paired with strains that have a high-impact variant in ben-1 that pass the resistance threshold are shown alongside each corresponding strain with a high-impact variant in ben-1. Wild C. briggsae strains are colored by beta-tubulin variant status.

https://doi.org/10.1371/journal.ppat.1014306.g002

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Dose-response curves for C. briggsae strains in albendazole.

Normalized animal lengths (y-axis) are plotted for each strain as a function of the dose of albendazole (ABZ) in the high-throughput larval development assay (HTLDA) (x-axis). Strains are denoted by color. Lines extending vertically from points represent the standard deviation from the mean response. Statistical normalization of animal lengths is described in Materials and Methods. C. elegans strains N2 and ECA882 were added for ABZ-susceptible and ABZ-resistant controls, respectively.

https://doi.org/10.1371/journal.ppat.1014306.g003

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Variation in lifetime fecundity between the C. briggsae (AF16) and C. tropicalis (NIC58) reference strains and strains with a loss of ben-1 in the AF16 and NIC58 reference strain backgrounds.

Bar plots for the lifetime fecundity, y-axis, for each strain on the x-axis are shown. Error bars show the standard deviation of lifetime fecundity among six to nine replicates. (A) A comparison of lifetime fecundities between the C. briggsae laboratory reference strain AF16 (green) and the two independently edited C. briggsae AF16 strains with a loss of ben-1 (red). (B) A comparison between the C. tropicalis laboratory reference strain NIC58 (blue) and two independently edited C. tropicalis NIC58 strains with a loss of ben-1 (red). Statistical significance was determined using Tukey HSD. Significance of each comparison is shown above each comparison pair (p > 0.05 = ns, p < 0.05 = *, p < 0.01 = **, Tukey HSD).

https://doi.org/10.1371/journal.ppat.1014306.g004

In C. tropicalis, we performed HTLDAs on the reference strain NIC58 (ABZ sensitive), two independent deletions of Ctr-ben-1 (ECA4247 and ECA4248) (ABZ resistant), two C. tropicalis strains with variants in ben-1 (P80T and R121Q), and two strains genetically related to ben-1 variant strains in DMSO (S6 Fig) and ABZ conditions (Figs 5 and S9). We found that all C. tropicalis wild strains displayed ABZ sensitivity, indicating that P80T and R121Q are not associated with ABZ resistance (Fig 5). Additionally, all tested wild strains were exposed to a maximum concentration of 120 μM ABZ, and none of the strains displayed resistance at these higher concentrations (Fig 6). Finally, to evaluate the potential fitness consequences caused by a loss of ben-1, fecundity assays were performed between the NIC58 reference strain and the two strains with a loss of ben-1 in the NIC58 background (ECA4247 and ECA4248) (Fig 4B and S7 Table). We found that a deletion of ben-1 caused a significant reduction in fecundity in C. tropicalis (Fig 4B). The fitness defect caused by a loss of ben-1 could explain the absence of naturally occurring ben-1 variants in wild C. tropicalis strains, suggesting that C. tropicalis is unlikely to acquire natural ben-1 variants that confer ABZ resistance without strong selection. If ABZ resistance is present in C. tropicalis wild strains, its genetic basis remains unknown and involves factors distinct from ben-1. Lastly, we only tested the effects of ABZ, testing another BZ analog would be necessary to determine whether the observed patterns of resistance are consistent throughout the BZ drug class.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. High-throughput larval development assays in the presence of albendazole for C. tropicalis strains with unique high-impact variants in BEN-1.

The regressed median animal length values for populations of nematodes grown in 30 μM albendazole (ABZ) are shown on the y-axis. Each point represents the normalized median animal length value of a well containing approximately five to 30 animals. Data are shown as Tukey box plots with the median as a solid horizontal line, and the top and bottom of the box representing the 75th and 25th quartiles, respectively. The top whisker is extended to the maximum point that is within the 1.5 interquartile range from the 75th quartile. The bottom whisker is extended to the minimum point that is within the 1.5 interquartile range from the 25th quartile. The gray dashed line marks the C. tropicalis resistance threshold, defined as two standard deviations below the mean of the ben-1 deletion strain in the NIC58 reference strain background. Results for (A) the NIC58 reference strain (blue) and a strain with a ben-1 deletion in the NIC58 background (red), and (B) all wild C. tropicalis strains with unique high-impact variants in ben-1 are sorted by their relative resistance to ABZ based on median animal length. Wild C. tropicalis strains are colored by beta-tubulin variant status.

https://doi.org/10.1371/journal.ppat.1014306.g005

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. Dose-response curves for C. tropicalis strains in albendazole.

Normalized animal lengths (y-axis) are plotted for each strain as a function of the dose of albendazole (ABZ) in the high-throughput larval development assay (HTLDA) (x-axis). Strains are denoted by color. Lines extending vertically from points represent the standard deviation from the mean response. Statistical normalization of animal lengths is described in Materials and Methods.

https://doi.org/10.1371/journal.ppat.1014306.g006

Resistance-associated ben-1 variants are located within the protein core near the predicted BZ-binding site

To evaluate whether the predicted molecular consequence and severity of ben-1 variants were predictive of BZ resistance, we characterized the identities and predicted functional impacts of each amino acid altering variant. We distinguished between alleles that disrupt multiple amino acids (i.e., frameshifts, stop/start altering variants, and SVs) and alleles that affect a single amino acid (i.e., missense alleles). Multi-site alleles were excluded from further analysis because they are predicted to severely disrupt or terminate BEN-1 function, whereas missense alleles might preserve an intact protein and allow interpretation of residue-level effects on BZ interaction.

In C. elegans, seven unique missense substitutions in BEN-1 were identified, of which five (Y50C, Q131L, S145F, A185P, M275I) were associated with BZ resistance (S7 Fig) [14]. In C. briggsae, seven unique missense substitutions in Cbr-BEN-1 were identified, but only Q134H was associated with BZ resistance (S8 Fig). In C. tropicalis, both missense substitutions (P80T and R121Q) were not associated with BZ resistance (S9 Fig). To evaluate if missense variants with larger predicted functional impacts were associated with BZ resistance, we quantified the severity of each substitution using BLOSUM62 and Grantham scores [40,41]. BLOSUM scores measure the evolutionary likelihood of observing a specific missense substitution, and Grantham scores measure the physicochemical severity of a substitution. We reasoned that missense substitutions that confer BZ resistance would either disrupt BEN-1 function or affect BZ binding. In both cases, substitutions predicted to be more structurally disruptive or evolutionarily less tolerated should be enriched among resistance-associated alleles. Because C. elegans had the most missense variants associated with BZ resistance, we performed a regression analysis of each strain’s response to ABZ by the BLOSUM or Grantham scores of the strain’s beta-tubulin variant. For C. elegans, neither BLOSUM nor Grantham scores showed a significant correlation with ABZ response (S14 Fig). By contrast, too few missense variants were associated with BZ resistance among C. briggsae and C. tropicalis strains to meaningfully interpret BLOSUM or Grantham scores for these species (S8 Table). The limited number of missense substitutions among wild strains and the even smaller subset associated with BZ resistance constrained statistical power to evaluate structure-function relationships between naturally occurring beta-tubulin variants and BZ response.

To complement the sequence-based analysis, we evaluated the structural position of each variant relative to the predicted BZ-binding site. We mapped missense variants onto an AlphaFold-predicted BEN-1 protein structure that showed that the five C. elegans missense substitutions associated with resistance are located within the core, whereas the two substitutions not associated with resistance (P80S and D404N) are on the protein surface (Fig 7A). By contrast, all C. briggsae (except Q134H) and all C. tropicalis missense substitutions are on the protein surface and are not associated with resistance (Fig 7B and 7C). For structural comparison, we modeled H. contortus tbb-isotype-1 beta-tubulin bound to ABZ, and highlighted the canonical missense variants at positions 167, 198, and 200 (Fig 7D), all of which are located within the core of the structure, near the predicted BZ-binding pocket and are associated with resistance [18–22,42]. These results suggest that BZ resistance is associated with missense substitutions occurring within the protein core likely affecting the theoretical BZ-binding pocket.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 7. Structure of beta-tubulin orthologs bound to albendazole.

AlphaFold3 models of BEN-1 orthologs bound to albendazole (ABZ) are shown for (A) Cel-BEN-1, (B) Cbr-BEN-1, (C) Ctr-BEN-1, and (D) Hcon-TBB-ISO-1. ABZ is colored yellow. Residues impacted by amino acid substitutions are colored by resistance (red) or sensitivity (blue).

https://doi.org/10.1371/journal.ppat.1014306.g007

Natural variants in tbb-1, tbb-2, mec-7, and tbb-4 are not associated with ABZ resistance across the three free-living Caenorhabditis species

Although it has been established that ben-1 is the primary gene involved in ABZ resistance in C. elegans [14,32,36], we know less about the role each beta-tubulin gene plays in C. briggsae and C. tropicalis ABZ resistance. Therefore, we identified variants in the other four conserved beta-tubulin genes across the three Caenorhabditis species. We found no variants in Cel-tbb-1 or Cel-tbb-2 in any C. elegans wild strains (CaeNDR Release ID: C. elegans - 20231213) (Fig 8). However, we did identify strains with one missense variant (A9T) in Cel-mec-7, a splice donor variant in Cel-mec-7, or one missense variant in Cel-tbb-4 (Q8H) (S1 Table). The strain with a splice donor variant in Cel-mec-7 also carried a high-impact variant (*445S) in Cel-ben-1. We also assessed the relationship between the expression of Cel-tbb-1, Cel-tbb-2, Cel-mec-7, or Cel-tbb-4 [37] and ABZ responses in 180 wild strains (S15 Fig) [14,36]. In C. briggsae, we identified strains with missense variants in Cbr-tbb-1 (T35A, V64I, A275T, and L377I), Cbr-tbb-2 (E441A), Cbr-tbb-4 (A271V), and Cbr-mec-7 (T136I, N165S, and S338C) (CaeNDR Release ID: C. briggsae - 20240129) (S2 Table). In C. tropicalis, we identified no high-impact variants in Ctr-tbb-1, one missense variant in Ctr-tbb-2 (N89S), two missense variants in Ctr-mec-7 (P80F, D433E), and one missense variant in Ctr-tbb-4 (A78V) (CaeNDR Release ID: C. tropicalis - 20231201) (S3 Table). Because expression data from C. briggsae and C. tropicalis wild strains have yet to be collected, we could not test correlations of beta-tubulin gene expression with resistance. We performed HTLDAs under control (DMSO) (S16, S17, and S18 Figs) and ABZ conditions (S19, S20 and S21 Figs) on strains in the three Caenorhabditis species carrying variants in the conserved beta-tubulin genes tbb-1, tbb-2, mec-7, and tbb-4, along with strains genetically related to strains with variants in those beta-tubulin genes. No variants outside of ben-1 conferred ABZ resistance in any species, indicating that natural missense variants in tbb-1, tbb-2, mec-7, and tbb-4 are not associated with BZ resistance in the three selfing Caenorhabditis species.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 8. The frequency of predicted high-impact consequences in the five conserved beta-tubulin genes present in natural populations of C. elegans, C. briggsae, and C. tropicalis.

The frequency of single nucleotide variants (SNVs) and structural variants (SVs) present in natural populations of C. elegans (n = 611), C. briggsae (n = 641), and C. tropicalis (n = 518) (y-axis) are shown by their predicted consequence in each beta-tubulin gene (x-axis). The frequency of each consequence was calculated as the number of strains carrying a particular predicted consequence divided by the total number of strains sampled for each species. The total number of isotype reference strains with a given predicted consequence are displayed on top of each bar plot.

https://doi.org/10.1371/journal.ppat.1014306.g008

Differences in tissue- and cell-specific expression patterns of beta-tubulins might influence BZ susceptibility

Tissue-specific differences in beta-tubulin expression among the three Caenorhabditis species could influence BZ resistance. To compare beta-tubulin expression patterns, we analyzed two whole-animal single-cell transcriptomic datasets that quantify gene expression patterns across C. elegans, C. briggsae, and C. tropicalis [44,45]. A comparison of embryonic gene expression levels in C. elegans and C. briggsae revealed that beta-tubulin gene expression during embryogenesis is highly conserved (S9 and S10 Tables). On average, the five beta-tubulin gene distances, which quantify expression conservation across homologous cell types, were small (mean beta-tubulin JSDgene = 0.29) (S9 Table) [43] and expression breadth across all cell types (i.e., cell-type specificity) was conserved (S10 Table). Among beta-tubulin genes, tbb-2 displayed the greatest divergence in embryonic gene expression breadth, the largest difference across species in cell-type specificity (S10 Table). Overall, embryonic beta-tubulin expression patterns are conserved between C. elegans and C. briggsae, although tbb-2 expression breadth is an exception (S10 Table).

Next, because neuronal ben-1 expression restores BZ susceptibility in C. elegans [31], we focused on the divergence of beta-tubulin expression across homologous neuronal cell classes at the L2 larval stage [43]. For each gene, we quantified neuron-class-specific expression divergence across the three Caenorhabditis species using the proportion of neuronal classes where gene expression differs (S11 Table) [43]. At least one species expressed ben-1 in 98 of 118 neuronal cell classes. However, all three Caenorhabditis species expressed ben-1 in only 35 of these classes (Jaccard distance = 0.64) (S11 Table). By contrast, tbb-1 and tbb-2 had highly conserved expression across the three Caenorhabditis species (98 of 118 neuronal cell classes expressed tbb-1 or tbb-2 in all three species) (Jaccard distance = 0) (S11 Table) [43]. The broad and conserved neuronal expression of tbb-1 and tbb-2 contrasts with ben-1, which showed narrower and species-specific expression patterns. Next, because ben-1 expression in cholinergic neurons restores BZ susceptibility in C. elegans [43,44], we compared cholinergic neuronal expression and identified several cell classes with species-specific patterns. Generally, beta-tubulin cell specificity is conserved across the three Caenorhabditis species (S9, S10, and S11 Tables). However, examples of expression divergence among C. elegans, C. briggsae, and C. tropicalis were identified for ben-1 and tbb-2. The divergence in ben-1 expression across cholinergic cell classes indicates that the neuronal sites where ben-1 function causes BZ susceptibility could differ among species. Future work should test this hypothesis using transgenic strains that express ben-1 in different tissues and neuronal cells.

The most diverse high-impact variants are found in Cel-ben-1

To better understand the evolution of predicted BZ resistance alleles in the three Caenorhabditis species, we assessed the population-wide frequencies of each beta-tubulin variant. First, to determine the prevalence of high-impact variants in the five conserved beta-tubulin genes (tbb-1, tbb-2, mec-7, tbb-4, and ben-1) across Caenorhabditis species, we quantified the frequency of each consequence (deletion, duplication, frameshift, in-frame deletion, inversion, missense, splice donor, and start/stop altering) in each species. With global sampling of the three Caenorhabditis species (C. elegans: 611 strains, C. briggsae: 641 strains, C. tropicalis: 518 strains), we found that C. elegans variants predicted to cause deleterious functional effects were present in 1% of strains. By contrast, variants predicted to cause deleterious functional effects in C. briggsae and C. tropicalis were rare (< 0.05% of all strains in either species) (Fig 8). The most variation in beta-tubulin genes was identified in Cel-ben-1, where deletions, frameshifts, in-frame deletions, inversions, missense, stop/start altering variants, and a duplication were found. Next, we found one missense variant in Cel-tbb-4 and several missense variants and a splice donor in Cel-mec-7. Overall, C. elegans has acquired the most diverse set of high-impact variants and predicted functional effects on ben-1. In C. briggsae, we found 21 missense amino acid substitutions and a single start/stop altering consequence in Cbr-ben-1. In both Cbr-tbb-1 and Cbr-tbb-2, we identified rare missense consequences. Additionally, we found nine strains with splice variants in Cbr-tbb-4 and one strain with a missense variant. In C. tropicalis, we found missense consequences in all beta-tubulin genes, except tbb-1. Additionally, we found six strains with splice variants in Ctr-tbb-4. These findings highlight that Cel-ben-1 has the most diverse set of variants, reinforcing its role in BZ resistance.

Next, we examined the geographic distribution of strains carrying high-impact variants in tbb-1, tbb-2, mec-7, tbb-4, and ben-1 to determine if beta-tubulin variants were associated with natural sampling location. In C. elegans, variants in ben-1 were found globally with no discernible geographic pattern but were concentrated in clades that have experienced recent selective sweeps [45,46] (Fig 9B). Cel-ben-1 variants in swept clades suggest that these mutations arose as relatively recent evolutionary events in response to BZ-like compounds in the natural niche. By contrast, in C. briggsae, variants in ben-1 were distributed throughout the species tree and found on more ancestral branches (i.e., earlier diverged lineages in the species) (Fig 9C). For C. tropicalis, the limited number of variants in ben-1 precludes any definitive conclusions regarding their evolutionary patterns (Fig 9D). Because few variants are found in tbb-1 (S22 Fig), tbb-2 (S23 Fig), mec-7 (S24 Fig), and tbb-4 (S25 Fig), we cannot identify the evolutionary patterns of BZ resistance in these genes for any of the three species.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 9. The global distribution of Caenorhabditis strains that contain predicted high-impact variation in BEN-1.

(A) Each point corresponds to the sampling location of an individual C. elegans (orange), C. briggsae (green), or C. tropicalis (blue) isotype reference strain with a predicted high-impact consequence in BEN-1. A genome-wide phylogeny of (B) 611 C. elegans, (C) 641 C. briggsae, and (D) 518 C. tropicalis isotype reference strains, where each point denotes an isotype reference strain with a predicted high-impact consequence in BEN-1 is shown. The base layer of the map was obtained from the Natural Earth world countries shape file accessed via the R package rnaturalearth and function ne_countries(). The scale was set to return a medium-scale base map with the scale = medium parameter. The direct link to the base layer of the map can be accessed here: https://www.naturalearthdata.com/.

https://doi.org/10.1371/journal.ppat.1014306.g009

Finally, because substrates harbor distinct microbial communities that can influence the evolution of BZ resistance alleles, we determined if strains carrying a high-impact variant in a beta-tubulin gene were associated with specific substrates. Substrate categories were obtained from CaeNDR collection metadata and were classified into 12 major substrates (i.e., arthropod, bait, compost, fungus, mollusk, moss, rotting flower, rotting nut, rotting stem, rotting wood, soil, or vegetal litter) [35,47,48]. A substrate enrichment analysis was performed to assess correlations between 12 substrates and all strains in the three Caenorhabditis species (S12 Table). However, no significant enrichment was observed between a high-impact variant in a beta-tubulin gene and any given substrate (Fisher’s Exact Test, p = 1) (S26Fig). Because no geographic or substrate enrichment was observed, evolutionary pressures driving beta-tubulin variation are likely not strongly tied to substrate. However, our broad substrate categories and small number of BZ resistant C. briggsae and C. tropicalis strains might obscure finer-scale ecological patterns. Future studies characterizing microbial communities associated with each substrate might clarify the selective pressures on Caenorhabditis nematodes.

Beta-tubulin-mediated BZ resistance varies across natural populations of Caenorhabditis nematodes

This study provides new insights into beta-tubulin-mediated ABZ resistance across three Caenorhabditis species. Each Caenorhabditis species harbored a unique set of predicted high-impact beta-tubulin alleles, but only variants in ben-1 conferred resistance. With additional wild strains since our first study, we identified more Cel-ben-1 LoF alleles associated with BZ resistance, which confirmed that a diverse collection of predicted LoF variants in ben-1 are associated with ABZ resistance in C. elegans [14, 16]. In C. briggsae, strains harboring only two of the eight unique Cbr-ben-1 variants were resistant to ABZ. One Cbr-ben-1 variant (W21stop) causes early protein termination, and the other (Q134H) alters a residue within the Cbr-BEN-1 protein that likely affects ABZ binding and is associated with resistance in A. caninum [11]. A CRISPR-Cas9-generated deletion of Cbr-ben-1 conferred resistance similar to that displayed by wild strains with these alleles and did not impact fitness. To date, resistance has yet to be identified in C. tropicalis wild strains. However, a CRISPR-Cas9-generated deletion of Ctr-ben-1 conferred resistance and caused a significant reduction in fitness (fecundity), which likely explains why just two high-impact variants were identified among wild strains and neither caused loss of ben-1 function. Finally, because we hypothesized that similar selective pressures would cause the three Caenorhabditis species to evolve BZ resistance by mutation in a beta-tubulin gene, we assessed niche overlap among these species. C. elegans occupies a niche distinct from C. briggsae and C. tropicalis, with only occasional overlap with C. briggsae [49, 50]. Consistent with this ecological separation, ben-1 resistance alleles do not colocalize between species. The two C. briggsae ben-1 resistance alleles were identified in strains collected in Aberdeen, Scotland (PE887) and Puerto Aventuras, Mexico (NIC1052), neither is a region where C. elegans has been sampled. C. tropicalis is found in warmer climates with no overlaps with C. elegans and rare overlaps with C. briggsae. Therefore, unique selection pressures potentially add to the independent evolution of BZ resistance driven by ben-1 mutations. Overall, our results highlight the complexity of BZ resistance. Accurate prediction of BZ resistance across nematode species requires a clear understanding of the contributions of both beta-tubulin dependent and beta-tubulin independent mechanisms.

Nematode survival under BZ exposure depends on beta-tubulin dosage and drug-binding ability

BZs bind to beta-tubulins and inhibit the polymerization of microtubules [2–4]. Therefore, despite the presence of beta-tubulin independent resistance mechanisms [51,52], beta-tubulins have a large impact on BZ efficacy as an anthelmintic treatment. Two factors determine how beta-tubulin impacts susceptibility to BZs: (1) the beta-tubulin’s ability to bind to BZs and (2) the dosage of the beta-tubulin protein, which can be modified by tissue-specific expression. Cell-specific expression of drug transporters or efflux pumps might also influence the intracellular BZ concentration. Nematode species have distinct beta-tubulin gene complements with divergence in number of genes, BZ binding affinities, expression levels and cell types, and redundancy. Ultimately, this divergence in beta-tubulin complement shapes the type(s) of resistance-associated alleles that arise in each nematode species. In species where the primary BZ-binding beta-tubulin is redundant with other isoforms, as in selfing Caenorhabditis species, loss of a single copy can confer resistance without disrupting essential microtubule functions. In species without redundancy, such as H. contortus, beta-tubulin variants that alter the BZ binding affinity are the only available beta-tubulin dependent route to BZ resistance.

In C. elegans, ben-1 is the primary target of BZs but is redundant with tbb-1 and tbb-2 [28–30]. LoF alleles in ben-1 reduce the total amount of BZ-binding protein while preserving essential microtubule functions, which permits nematode development during BZ exposure. A similar pattern likely occurs in C. briggsae, where rare high-impact variants in ben-1 confer resistance, and other beta-tubulin genes maintain essential microtubule functions. Although natural Ctr-ben-1 resistance alleles were not identified, a deletion of ben-1 conferred BZ resistance in C. tropicalis but imposed a significant fitness cost, which could explain the few high-impact alleles among wild strains. Fewer C. tropicalis wild strains were surveyed (518) compared with C. elegans (611) and C. briggsae (641), which could partly explain why fewer high-impact variants were identified. Reduced variation in Ctr-ben-1 might also partly reflect lower genome-wide C. tropicalis genetic diversity compared to C. elegans and C. briggsae [53–55]. By contrast, H. contortus relies on the essential beta-tubulin tbb-isotype-1 and cannot tolerate loss without severe fitness consequences [18]. The differences between the Caenorhabditis species and H. contortus demonstrate that the prediction of beta-tubulin dependent resistance requires the identification of the total number and expression levels of beta-tubulin genes and determination of which encoded beta-tubulins can bind to BZ. Altogether, BZ resistance depends not on the copy number of beta-tubulins alone but on the availability of functional beta-tubulins capable of interacting with BZs.

How can we accurately predict BZ resistance across nematode species?

To accurately predict BZ resistance across nematode species, we need (1) tractable nematode models with beta-tubulin gene complements that resemble that of parasite species, (2) improved parasitic nematode genomes that enable the comprehensive identification of beta-tubulin genes, (3) functional tests that define how parasite beta-tubulin alleles contribute to BZ response, and (4) to define the contribution of non-beta-tubulin genes to overall BZ response. To date, no experimentally tractable model nematode species exist with beta-tubulin genes that have both limited redundancy and BZ binding properties similar to those of parasitic nematodes. Pristionchus pacificus, a free-living clade V nematode has three beta-tubulin genes, comprising two ben-1 orthologs (Ppa-ben-1.1 and Ppa-ben-1.2) and an ortholog of Cel-mec-7 (Ppa-mec-7) [56]. All three beta-tubulin genes in Pristionchus pacificus are predicted to have high ABZ binding affinity (S27 Fig). Using CRISPR-Cas9 genome editing, we found that homozygous LoF alleles for either Ppa-ben-1.1 or Ppa-ben-1.2 could not be recovered in P. pacificus, suggesting that both genes are essential (S13 Table). The fact that P. pacificus has fewer beta-tubulin genes than C. elegans likely contributes to the lack of redundancy among beta-tubulins. The number, essentiality, and BZ binding affinity of the P. pacificus beta-tubulins more closely resembles that of H. contortus, which positions P. pacificus as a valuable free-living model to study essential beta-tubulin function, beta-tubulin dosage effects, and BZ resistance relevant to parasitic nematode species.

Second, to understand the role beta-tubulins play in BZ resistance across parasitic nematode species, we must significantly improve genomes and gene models. Most parasitic nematode genomes remain incomplete or are poorly annotated, which obscures beta-tubulin copy number and gene identity. Recent efforts have produced higher quality genomes for some parasitic nematodes [57–59], but we must still define beta-tubulin copy number across diverse nematode species. For example, improved reference genomes and gene models for C. briggsae [60,61] and C. tropicalis [53,55] enabled accurate beta-tubulin gene identification for these species. Until genome assemblies, technologies, and analytical techniques improve, we will be unable to accurately predict the number of beta-tubulin genes in a given species and their respective BZ binding affinities.

Third, functional validation of parasitic resistance alleles presents additional challenges. Although tools such as RNAi can work in parasitic nematodes [62], delivery challenges and difficulty isolating edited individuals limit its use. An alternative strategy is to introduce predicted parasite resistance alleles into C. elegans to test effects on fitness and BZ response, as shown previously [14–16]. Finally, some parasitic nematodes, such as ascarid species, exhibit BZ resistance without known beta-tubulin resistance alleles [63–65]. This pattern suggests that resistance arises independently of detected beta-tubulin sequence changes or reflects incomplete identification of beta-tubulin genes caused by poor genome assemblies. To date, the underlying cause of BZ resistance in ascarid species remains unclear. Reduced functional beta-tubulin dosage, altered BZ binding, or both could contribute to BZ resistance in ascarids [66]. Improved genome assemblies and annotation are required to identify all beta-tubulin genes and drug-binding sites to define the mechanisms of BZ resistance in ascarid species. In the future, the introduction of newly identified ascarid alleles into free-living nematode models can directly test resistance and fitness, and clarify the mechanisms that drive BZ resistance in clade III nematodes. Additionally, we must define how beta-tubulin redundancy, essentiality, dosage, and physical interactions with BZs shape the evolution of resistance to improve the prediction of BZ resistance across the huge diversity of parasitic nematode species.

Identification of beta-tubulin loci

Amino acid sequences for all six C. elegans beta-tubulin proteins were obtained from WormBase (WS283) [67] and used as queries in a BLASTp search (Version 2.12.0) [68] against protein sequence databases constructed using gene models for C. briggsae [61] and C. tropicalis [53]. To construct the protein sequence database, we extracted gene model transcript features from the gene feature file with gffread (Version 0.9.11) [69] and processed them using the makeblastdb function from BLAST (Version 2.12.0). From the BLASTp search, we identified C. briggsae and C. tropicalis protein sequences with the highest percent identity (PID) to each C. elegans beta-tubulin protein. Only protein sequences with the highest PID in both searches were considered orthologs (S14 Table). For some C. elegans beta-tubulin orthologs, multiple C. briggsae or C. tropicalis gene models contained multiple splice isoforms. All gene models for all beta-tubulin transcripts were manually inspected, and isoforms that were not fully supported by short-read RNA sequencing data were removed.

Single nucleotide variant (SNV) and indel calling and annotation

To identify single nucleotide variants (SNVs) or indels (insertions and deletions) in the beta-tubulin genes across the selfing Caenorhabditis species, we used the Variant Annotation Tool from the Caenorhabditis Natural Diversity Resource (CaeNDR) (Release IDs: C. elegans - 20231213, C. briggsae - 20240129, C. tropicalis - 20231201) [35]. The identified SNVs and indels included small insertions and deletions, frameshifts, altered stop and start codons, nonsynonymous changes, and splice variants (S1, S2 and S3 Tables).

Structural variant (SV) calling and annotation

Structural variant (SV) calling was performed using DELLY (Version 0.8.3), a SV caller optimized to detect large insertions, deletions, and other complex structural variants such as inversions, translocations, and duplications in paired-end short-read alignments [70] and shown to perform well on C. elegans short-read sequence data [71]. SVs that overlapped with beta-tubulin genes were extracted using bcftools (Version 1.10.1) [72]. Insertions, deletions, inversions, and duplications that passed the DELLY (Version 0.8.3) default quality threshold (greater than three supporting read pairs with a median MAPQ > 20), filtered to high-quality genotypes (genotype quality > 15), and had at least one alternative allele were retained. For complex variants (inversions and duplications), the identification of at least one split-read pair was required (variants flagged as a precise SV by DELLY). To validate SVs that passed quality filtering, each SV was manually inspected for breakpoints in the raw-read alignments (Wally, Version 0.5.8) and for impacts on the beta-tubulin coding sequence (CaeNDR Genome Browser) [35] (S15 Table). We retained SVs where raw read alignments suggested that the SV impacted the beta-tubulin coding sequence. We compared the Cel-ben-1 SVs called by DELLY to those SVs identified previously [71]. DELLY successfully recalled structural variants in several strains, including deletions in JU751, JU830, JU1395, JU2582, JU2587, JU2593, JU2829, and QX1233, as well as an inversion in MY518. However, DELLY did not detect a previously reported transposon insertion in strain JU3125. To assess if other SVs could have been missed by DELLY, we manually inspected the read alignments for all strains that had not been previously phenotyped to check if any other SVs were not detected by DELLY. We confirmed the presence of novel Cel-ben-1 SVs in multiple strains, including putative deletions in ECA706 and NIC1832, and a duplication in NIC1107. Additionally, we identified a previously undetected putative deletion in JU4287. We also examined the amino acids at position 200 in TBB-1, TBB-2, MEC-7, TBB-4, and BEN-1 orthologs and found that all MEC-7, TBB-4, and BEN-1 orthologs contained phenylalanine, and that all TBB-1 and TBB-2 orthologs contained tyrosine at position 200 (S27 Fig).

Association of Cel-ben-1 expression with ABZ response

Two previous assays measured developmental responses of wild C. elegans strains after ABZ exposure [14, 36]. For 180 of these wild C. elegans strains, whole-animal expression levels (transcripts per million estimates [TPM]) were collected from untreated young-adult animals [37]. We identified strains with low Cel-ben-1 expression by selecting strains with TPM values more than one standard deviation (SD) below the mean expression level across all 207 wild strains with expression data. Of the 180 wild C. elegans strains, 105 strains were measured for both ABZ response and gene expression. A linear model was built using the lm function in R to account for assay effects. Subsequently, the residuals of the linear model were used to normalize previous measures of ABZ response. We evaluated the linear fit between each strain’s expression of ben-1 (S1 Fig), tbb-1, tbb-2, mec-7, or tbb-4 (S15 Fig) and the developmental delay following ABZ exposure.

Phylogenetic analysis

We characterized the relatedness of isotype reference strains (genetically unique strains) with beta-tubulin variants using species trees downloaded from CaeNDR and generated by the ‘post-gatk-nf’ pipeline (https://github.com/AndersenLab/post-gatk-nf) (Release IDs: C. elegans - 20231213, C. briggsae - 20240129, C. tropicalis - 20231201) [35]. Briefly, the trees were generated using high-quality SNVs in isotype reference strains retained in the hard-filtered variant call format (VCF) file. vcf2phylip [73] and the bioconvert [74] function phylip2stockholm were used to prepare inputs for quicktree, which was used to construct a tree using a neighbor-joining algorithm [75]. All versions of these software can be accessed from the ‘post-gatk’ docker container (https://hub.docker.com/r/andersenlab/tree), used by the ‘post-gatk-nf’ pipeline. We visualized the trees for each species using the ggtree function from the ggtree (v3.6.2) R package [76].

Strain selection and maintenance

Eighteen C. elegans strains, 45 C. briggsae strains, and 15 C. tropicalis strains from the CaeNDR [35] were used in this study (S1, S2, and S3 Table). Isolation details for each strain are included in CaeNDR. For each species, we selected strains that had variants (SNV or SV) with unique high-impact consequences in tbb-1, tbb-2, mec-7, tbb-4, or ben-1 that had not been previously phenotyped. High-impact consequences included changes to amino acids, start and stop codon positions, or splice variants predicted to disrupt beta-tubulin function [35]. Strains with high-impact consequences in a beta-tubulin gene are herein referred to as “predicted resistant” strains. Strains that were closely related to predicted resistant strains with no high-impact consequences in beta-tubulin genes were also included for C. briggsae and C. tropicalis and herein classified as “predicted susceptible” strains (S2, S3 Figs and S4 Table). Although these predicted susceptible strains are closely related to the predicted resistant strains, non-beta-tubulin variants in the genetic background of these strains could impact BZ susceptibility. The reference strains for all three species were included.

Before measuring ABZ responses, C. elegans and C. briggsae animals were maintained at 20ºC and C. tropicalis animals were maintained at 25ºC. All animals were maintained on 6 cm plates with modified nematode growth medium (NGMA), which contains 1% agar and 0.7% agarose to prevent animals from burrowing [77]. The NGMA plates were seeded with the Escherichia coli strain OP50 as a nematode food source. All strains were grown for three generations without starvation on NGMA plates before anthelmintic exposure to reduce the transgenerational effects of starvation stress [78].

CRISPR-Cas9 genome editing

To validate the role that ben-1 plays in ABZ resistance in the three Caenorhabditis species, we used CRISPR-Cas9 to create ben-1 deletions. For C. elegans, we used a previously generated strain (ECA882) with a ben-1 deletion in the N2 background [14,15,32,36]. The ben-1 deletions were generated in the AF16 background for C. briggsae and the NIC58 background for C. tropicalis. Injections were performed by InVivo Biosystems (Eugene, OR), and deletions of Cbr-ben-1 and Ctr-ben-1 were confirmed using PCR (S10 and S11 Figs). Briefly, two primer pairs were designed for the deletion alleles for each species, with each pair designed to bind to a region external or internal to both of the deletions. Confirmation of deletion was performed by performing two amplification reactions for each sample: (1) the use of both external primers, and (2) the use of an internal and external pair (S5 Table). The parental strain was used as a control in each PCR. Deletions were confirmed by a reduction in the size of the external-external amplicon in the edited strains compared to the unedited parental control strain. Homozygosity was confirmed by the loss of a band amplified from the external-internal primer pair. Edited strains underwent two generations of PCR confirmation for homozygosity. Two independent edits of each allele in each species were generated to control for any potential off-target effects caused by CRISPR-Cas9 genome editing (S12, S13 Figs and S5 Table).

Nematode food preparation for NGMA 6 cm plates

The OP50 E. coli strain was used as a nematode food source for NGMA plates. A frozen stock of OP50 was streaked onto a 10 cm Luria-Bertani (LB) agar plate and incubated overnight at 37ºC. The following day, a single bacterial colony was transferred into each of two culture tubes that contained 5 mL of 1x LB. The starter cultures and two negative controls (1X LB without E. coli) were incubated for 18 hours at 37ºC shaking at 210 rpm. The OD₆₀₀ value of the starter cultures were measured using a spectrophotometer (BioRad, SmartSpec Plus) to calculate how much starter culture was needed to inoculate a 1 L culture at an OD₆₀₀ value of 0.005. For each assay, one culture containing 1 L of pre-warmed 1X LB inoculated with the starter culture grew for approximately 4 - 4.5 hours at 37ºC at 210 rpm to an OD₆₀₀ value between 0.45 and 0.6. Cultures were transferred to 4ºC to slow growth. OP50 was spotted on NGMA test plates (two per culture) and grown at 37ºC overnight to assay for contamination.

Nematode food preparation for high-throughput larval development assays (HTLDAs)

One batch of HB101 E. coli was used as a nematode food source for all HTLDAs in this study. A frozen stock of HB101 E. coli was streaked onto a 10 cm LB agar plate and incubated overnight at 37ºC. The following day, a single bacterial colony was transferred into three culture tubes that contained 5 mL of 1x Horvitz Super Broth (HSB). The starter cultures and two negative controls (1X HSB without E. coli) were incubated for 18 hours at 37ºC shaking at 180 rpm. The OD₆₀₀ value of the starter cultures were measured using a spectrophotometer (BioRad, SmartSpec Plus) to calculate how much starter culture was needed to inoculate a 1 L culture at an OD₆₀₀ value of 0.001. A total of four cultures each containing 1 L of pre-warmed 1X HSB inoculated with the starter culture grew for 15 hours at 37ºC while shaking at 180 rpm. After 15 hours, flasks were removed from the incubator and transferred to 4ºC to slow growth. The 1X HSB was removed from the cultures by performing three rounds of centrifugation, where the supernatant was removed, and the bacterial cells were pelleted. Bacterial cells were washed with K medium, resuspended in K medium, pooled, and transferred to a 2 L glass beaker. The OD₆₀₀ value of the bacterial suspension was measured and diluted to a final concentration of OD₆₀₀100 with K medium, aliquoted to 15 mL conical tubes, and stored at -80ºC for use in the HTLDAs.

ABZ dose-response assays for C. briggsae and C. tropicalis

Because ABZ response had been minimally characterized in C. briggsae [79] and has not yet been described in C. tropicalis, we first measured dose-response curves for both species after exposure to ABZ to assess developmental delay. Before performing HTLDAs, ABZ (Sigma-Aldrich, Catalog # A4673-10G) stock solutions were prepared in dimethyl sulfoxide (DMSO) (Fisher Scientific, Catalog # D1281), aliquoted, and stored at -20ºC for use in the assays. For the dose-response assays, animals were exposed to ABZ at the following concentrations (μM): 0 (0.3% DMSO), 0.12, 0.23, 0.47, 0.94, 1.88, 3.75, 7.5, 15, 30, 60, and 120. Animals developed in the presence of ABZ as described in HTLDAs to assess nematode development.

Dose-response model estimation and statistics were performed as described previously [80,81]. Briefly, a four-parameter log-logistic dose-response curve was fit independently for a genetically diverse set of 11 C. briggsae strains (Fig 3) and seven C. tropicalis strains (Fig 6), where normalized median animal length was used as a metric for phenotypic response (see HTLDA data collection and data cleaning). For each strain-specific dose-response model, slope (b) and concentration (e) were estimated with strain as a covariate. We calculated EC₁₀ as we have previously found EC₁₀ response to be more heritable than half maximal effective concentration (EC₅₀) estimates and were used in our analysis [80,81]. A dosage of 30 μM ABZ was closest to the EC₁₀ for C. briggsae and C. tropicalis, consistent with ABZ concentrations used in past C. elegans assays [15,16,36] and in all HTLDAs in this study.

HTLDAs to assess nematode development

Populations of each strain were amplified and bleach-synchronized in three independent assays. Independent bleach synchronizations controlled for variation in embryo survival and subsequent effects on developmental rates. After bleach synchronization, approximately 30 embryos were dispensed into each well of a 96-well microplate in 50 μL of K medium. Each strain had sixteen wells per condition (DMSO or ABZ) in each assay. Three independent assays yielded a total of forty-eight wells per condition per strain. Each 96-well microplate was prepared, labeled, and sealed using gas-permeable sealing films (Fisher Scientific, Catalog # 14-222-043). Plates were placed in humidity chambers to incubate for 24 hours at 20°C for C. elegans and C. briggsae, and 25°C for C. tropicalis while shaking at 170 rpm (INFORS HT Multitron shaker). After 24 hours, every plate was inspected to ensure that all embryos hatched and animals were developmentally arrested at the first larval (L1) stage so all strains started each assay at the same developmental stage. Next, food was prepared to feed the developmentally arrested L1 animals using the required number of OD₆₀₀100 HB101 aliquots (see Nematode food preparation for HTLDAs). The HB101 aliquots were thawed at room temperature, combined into a single conical tube, and diluted to an OD₆₀₀30 with K medium. To inhibit further bacterial growth and prevent contamination, 150 μL of kanamycin was added to the HB101. An aliquot of 100 μM ABZ stock solution was thawed at room temperature and added to an aliquot of OD₆₀₀30 K medium at a 3% volume/volume ratio. Next, 25 μL of the food and ABZ mixture was transferred into the appropriate wells of the 96-well microplates to feed the arrested L1 animals at a final HB101 concentration of OD₆₀₀10 and expose L1 animals to ABZ. Immediately afterward, the 96-well microplates were sealed using a new gas permeable sealing film, returned to the humidity chambers, and incubated for 48 hours at 20°C (C. elegans and C. briggsae) or 42 hours at 25°C (C. tropicalis) while shaking at 170 rpm. After 48 hours (C. elegans and C. briggsae) or 42 hours (C. tropicalis) of incubation and shaking in the presence of food and either 0.3% DMSO or 30 μM ABZ, the 96-well microplates were removed from the incubator and treated with 50 mM sodium azide in M9 for 10 minutes to paralyze and straighten nematodes. After 10 minutes, images of nematodes in the microplates were immediately captured using the Molecular Devices ImageXpress Nano microscope (Molecular Devices, San Jose, CA) using a 2X objective. The ImageXpress Nano microscope acquires brightfield images using a 4.7 megapixel CMOS camera and stores images in a 16-bit TIFF format. The images were used to quantify the development of nematodes in the presence of DMSO or ABZ as described below (see HTLDA data collection and data cleaning). A full step-by-step protocol‌‌ for the HTLDA has been deposited on protocols.io [82].

HTLDA data collection and data cleaning

CellProfiler (Version 24.10.1) was used to characterize and quantify biological data from the image-based assays. Custom software packages designed to extract animal measurements from images collected on the Molecular Devices ImageXpress Nano microscope were previously described [39]. CellProfiler modules and WormToolbox were developed to extract morphological features of individual animals from images from the HTLDA [82]. Worm model estimations and custom CellProfiler pipelines were written using the WormToolbox in the GUI-based instance of CellProfiler [39]. Next, a Nextflow pipeline (Version 24) was written to run command-line instances of CellProfiler in parallel on the Rockfish High-Performance Computing Cluster (Johns Hopkins University). The CellProfiler workflow can be found at https://github.com/AndersenLab/cellprofiler-nf. The custom CellProfiler pipeline generates animal measurements by using four worm models: three worm models tailored to capture animals at the L4 larval stage, in the L2 and L3 larval stages, and the L1 larval stage, as well as a “multi-drug high dose” (MDHD) model, to capture animals with more abnormal body sizes caused by extreme anthelmintic responses. These measurements comprised our raw dataset. Two C. briggsae strains (NIC1052 and VX34) were not fully paralyzed and straightened at the time of imaging, which created some misclassification of animal measurements. Thus, the animal lengths for strains NIC1052 and VX34 measured by CellProfiler are shorter than the actual animal lengths. However, the difference in animal lengths does not affect the classification of a strain as resistant or sensitive to ABZ. Data cleaning and analysis steps were performed using a custom R package, easyXpress (Version 2.0) [39] and followed methods previously reported [36]. Briefly, using easyXpress, we removed statistical outlier measurements for each strain by condition to reduce the likelihood that statistical outliers influence anthelmintic responses. Finally, we normalized the data by (1) regressing variation attributable to assay and technical replicate effects and (2) normalizing these extracted residual values to the average control phenotype. These normalized length measurements (i.e., normalized animal length (μm)) have the helpful property of being centered on zero in control conditions for each strain, and therefore, control for the differences in the average lengths of the strains. All analyses were performed using the R statistical environment (Version 4.2.1) unless stated otherwise.

C. briggsae and C. tropicalis fecundity assays

To define the fitness costs associated with a loss of ben-1 in C. briggsae and C. tropicalis, we performed fecundity assays. For C. briggsae, we used the two strains with independent edits of ben-1 in the AF16 background (ECA3953 and ECA3954) and the AF16 reference strain. For C. tropicalis, we used the two strains with independent edits of ben-1 in the NIC58 background (ECA4247 and ECA4248) and the NIC58 reference strain. To perform fecundity assays, we placed a single L4 larval stage hermaphrodite from each strain onto a 6 cm NGMA plate that was spotted with E. coli OP50. C. briggsae assay plates were maintained at 20°C, and C. tropicalis assay plates were maintained at 25°C. For each assay plate, the original hermaphrodite parent was transferred to a fresh 6 cm NGMA plate every 24 hours for 192 hours. Ten technical replicates were prepared for each strain. The Basic Imaging Platform from Tau Scientific was used to collect images for each of the assay plates (0, 24, 48, 72, 96, and 120-hours) at either 48 or 72 hours for C. briggsae and C. tropicalis animals, respectively, after the removal of the parent from each NGMA plate. The total offspring was counted from each image by visual inspection using the Multi-point tool in ImageJ (Version 1.54g). The original hermaphrodite parents were excluded from the counts. Replicates where the original hermaphrodite parent died were excluded from the analysis. Only biological replicates with data from at least six assay plates were used to calculate total fecundity (Fig 4, S6 and S7 Tables).

Tissue-specific beta-tubulin gene expression conservation in C. elegans, C. briggsae, and C. tropicalis

To assess beta-tubulin expression divergence across neuronal cell classes, we used whole-animal single-cell transcriptomes of C. elegans, C. briggsae, and C. tropicalis [43]. We quantified neuronal cell expression divergence across species using precomputed Jaccard distances calculated from the neuronal cell classes in at least one species and the number of neuron classes in which a gene is expressed in all three species [43]. We also downloaded gene expression summary data for each homologous cell class from CaenoGen and analyzed expression patterns in cell classes that used acetylcholine (ACh) as a neurotransmitter in C. elegans [83]. We identified cell classes in which ben-1 orthologs showed species-specific presence or absence of expression.

We also analyzed embryonic single-cell transcriptomes from the C. briggsae and C. elegans reference strains to examine the conservation of beta-tubulin gene expression in homologous cell types [44]. We used pre-computed summary statistics to quantify two components of expression divergence among beta-tubulin genes: gene distance (reported as Jensen-Shannon Distances for each gene) and expression breadth (reported as Tau metrics for each species). Beta-tubulin gene distances and expression breadth metrics are available in the gene data summary table file (S9 and S10 Tables). Estimates for three cell populations are distinguished by cell-type assignment: progenitor, terminal, and joint (combined progenitor and terminal cells).

Protein structure and visualization of ben-1 and tbb-isotype-1 variants

For C. elegans, the BEN-1 amino acid sequence was obtained from WormBase (WS283) [67]. For C. briggsae and C. tropicalis, BEN-1 amino acid sequences were the best hits to the C. elegans BEN-1 query from the reciprocal BLASTp search used to identify beta-tubulin orthologs (see Identification of beta-tubulin loci) (S16 Table). The H. contortus tbb-isotype-1 amino acid sequence was obtained from WormBase Parasite (Version: WBPS19) [42]. Protein structures were predicted using AlphaFold3 [84]. All BEN-1 variant data and associated benzimidazole-response phenotypes were compiled from this study and previously published data [14,36]. A custom Python script was used to generate a PyMOL script (Version 3.1.6.1) for structural visualization. Each predicted beta-tubulin structure was aligned to Cel-BEN-1 using the PyMOL `align` command.

Mutation of ben-1 orthologs in Pristionchus pacificus

To test the function of ben-1 in Pristionchus pacificus, we used CRISPR-Cas9 genome editing to delete the two orthologs of Cel-ben-1. Designed crRNAs and tracrRNA were synthesized by Integrated DNA Technologies (IDT). For the creation of the guide RNAs, 3 μL of 100 μM tracrRNA (IDT) was combined with 3 μL of 100 μM crRNA (IDT) and incubated at 95ºC for five minutes, followed by five minutes at room temperature for annealing. The ribonucleoprotein (RNP) complex was then created by combining 0.61 μL of the guide-RNA mixture with 0.25 μL of Cas9 protein (IDT), followed by incubation at 37ºC for 10 minutes. The final microinjection mix was then created by combining the 0.86 μL RNP complex with 9.14 μL of a TE buffer mixture containing a (55 ng/μL) plasmid containing a P. pacificus codon-optimized egl-20p::TurboRFP::rpl-23UTR construct [85] used as a co-injection marker for visual identification of successful injections by the presence of fluorescent F1 individuals (S13 Table). All F1 individuals displaying fluorescence were then isolated and allowed to self-fertilize. After successful F2 embryo hatching, F1 individuals were genotyped for mutations using a heteroduplex mobility assay (HMA). Individuals from F2 broods that were determined to contain heterozygous mutations in F1 mothers were then isolated, allowed to self-fertilize, and genotyped by HMA after F3 embryos had hatched to identify homozygous mutant lines, followed by Sanger sequencing to determine types of mutations induced. Additionally, for mutations that were homozygous lethal, the F3 from heterozygous F2 individuals were likewise genotyped by HMA to quantify survivability based on deviations from expected Mendelian ratios.