1. Neuroscience
Download icon

C. elegans avoids toxin-producing Streptomyces using a seven transmembrane domain chemosensory receptor

Short Report
Cited
0
Views
517
Comments
0
Cite as: eLife 2017;6:e23770 doi: 10.7554/eLife.23770

Abstract

Predators and prey co-evolve, each maximizing their own fitness, but the effects of predator–prey interactions on cellular and molecular machinery are poorly understood. Here, we study this process using the predator Caenorhabditis elegans and the bacterial prey Streptomyces, which have evolved a powerful defense: the production of nematicides. We demonstrate that upon exposure to Streptomyces at their head or tail, nematodes display an escape response that is mediated by bacterially produced cues. Avoidance requires a predicted G-protein-coupled receptor, SRB-6, which is expressed in five types of amphid and phasmid chemosensory neurons. We establish that species of Streptomyces secrete dodecanoic acid, which is sensed by SRB-6. This behavioral adaptation represents an important strategy for the nematode, which utilizes specialized sensory organs and a chemoreceptor that is tuned to recognize the bacteria. These findings provide a window into the molecules and organs used in the coevolutionary arms race between predator and potential prey.

https://doi.org/10.7554/eLife.23770.001

Introduction

The evolutionary interactions of predators with their prey have profoundly influenced the history of life on earth (Lotka, 1920). To better understand these interactions and their influences at the molecular and cellular levels, we studied the bacteriovore Caenorhabditis elegans and several species of the Gram-positive bacterium Streptomyces that have evolved strong defenses to avoid predation by nematodes. Streptomyces avermitilis, S. milbemycinicus, and S. costaricanus all produce potent nematicides or antinematodal compounds (Miller et al., 1979; Burg et al., 1979; Egerton et al., 1979; Takiguchi et al., 1980; Dicklow et al., 1993; Esnard et al., 1995) (Figure 1A). Interestingly, these bacteria have been used to produce drugs against nematode infections, such as avermectin and milbemycin, which paralyze nematodes by irreversibly opening invertebrate glutamate-gated chloride channels (Cully et al., 1994; Holden-Dye and Walker, 2014). Despite this usage, little is known about whether or how C. elegans recognizes and responds to Streptomyces in the environment.

Figure 1 with 1 supplement see all
C. elegans nematodes rapidly avoid Streptomyces at their head and tail.

(A) Phylogenetic tree including strains of Streptomyces bacteria and Bacillus subtilis. (B) Schematic diagram of a dry-drop behavioral assay that selectively tests the response of the PHA and PHB phasmid neurons. Animals are placed on a dry plate and induced to move backward with a hair pick. A drop of bacterial extract or control buffer is placed behind the animal and quickly dries, preventing wicking along the animal. As the animal moves into the drop, the time before the animal halts movement is measured. Response times are normalized to a positive control, the average response to sodium dodecyl sulfate (SDS), assayed on the same day. A shorter response time indicates a more rapid avoidance response. (C) Quantification of the rapid avoidance response of C. elegans phasmid neurons to Streptomyces, but not to Escherichia coli or B. subtilis. One-way ANOVA and Tukey’s post-hoc tests were performed to determine significance. (D) Schematic diagram of a dry-drop behavioral assay that selectively tests the response of amphid neurons in the head. Animals are placed on a dry plate and allowed to move forward. A drop of bacterial extract or control buffer is placed in front of the animals and quickly dries. The proportion of the time that animals halt movement into the dry drop is recorded. (E) Quantification of the rapid avoidance response of C. elegans amphid neurons to species of Streptomyces, but not E. coli or B. subtilis. Two independent sample z-tests and the Hochberg multiple comparison adjustment procedure were performed. (C and E) For each experimental sample, n ≥ 40 for the experimental group, n ≥ 40 for the 0.6% SDS positive control, and n ≥ 40 for the saline buffer negative control. ***B, p<0.001; *B, p<0.05; ns B, not significant compared with a control buffer; ***S, p<0.001; NS S, not significant compared with SDS positive control. Exact values and additional pairwise comparisons are included in Figure 1—source data 1 and Figure 1—source data 2.

https://doi.org/10.7554/eLife.23770.002

Results

To determine whether C. elegans responds to Streptomyces, we used dry-drop avoidance assays that test the response of either the phasmid sensory neurons in the tail or the amphid sensory neurons in the head (Park et al., 2011; Hilliard et al., 2002) to Streptomyces (Figure 1B,D). First, we tested the response to exposure at the tail. Interestingly, animals rapidly halted backwards movement when exposed to Streptomyces strains but not Escherichia coli, a Gram-negative species used to feed worms in the laboratory, or Bacillus subtilis, a Gram-positive species found in soil and vegetation (WilpatWipat and Harwood, 1999) (Figure 1C). These data indicate that C. elegans have the ability to recognize and specifically avoid Streptomyces. Few compounds have been fully characterized with respect to these neurons, but dilute concentrations of sodium dodecyl sulfate (SDS) have been shown to elicit a rapid and robust avoidance response in C. elegans hermaphrodites (Hilliard et al., 2002), the primary sex found in nature (Brenner, 1974). The avoidance response to Streptomyces is comparable to the response to SDS (Figure 1C). We hypothesize that the tail-mediated response to Streptomyces is due to a secretion product, as animals responded similarly to cell-free supernatants (Figure 1C).

We wondered whether C. elegans would demonstrate a similar avoidance response upon encountering Streptomyces at their head. Thus, we tested the response to exposure at the head. Animals exposed to a control buffer, E. coli or B. subtilis at the head usually continued moving forward, but animals exposed to S. avermitilis, S. milbemycinicus or S. costaricanus usually halted movement, avoiding Streptomyces (Figure 1D,E). These data indicate that C. elegans also recognize Streptomyces at the head. C elegans did not demonstrate avoidance in chemotaxis assays designed to test responses to volatile odorants, suggesting that direct contact is required (Figure 1—figure supplement 1).

SDS is a laboratory detergent, but the very similar responses of C. elegans to Streptomyces and SDS led us to hypothesize that Streptomyces might secrete a small molecule similar to SDS in activity or structure. The avoidance could be due to a non-specific interaction that is based on the compound’s surfactant properties, or it could be due to specific binding and recognition of SDS. We found that the surfactant activity of SDS was not required for the avoidance response, as concentrations ten-fold below the critical micelle concentration (Rahman and Brown, 1983) were still avoided (Figure 2A, Figure 2—figure supplement 1, Figure 2—figure supplement 2). Similarly, a surfactant (Triton X-100) at its critical micelle concentration (Tiller et al., 1984) did not elicit a strong avoidance response (Figure 2A). These findings indicate that SDS is unlikely to elicit a response through a non-specific detergent effect. Therefore, we looked to define the structural elements of SDS that are responsible for the avoidance response.

Figure 2 with 2 supplements see all
C. elegans avoids the SDS analogs dodecanoate and decanoate.

(A) Quantification of the similar phasmid responses to 3.4 mM and 0.6 mM SDS, which is ten-fold below the critical micelle concentration of SDS. All other assays in this study were performed with 0.6 mM SDS to preclude the possibility of surfactant activity. The surfactant Triton X-100 at its critical micelle concentration is not strongly sensed by the phasmid neurons. (B) Quantification of the reduced responses to molecules with shorter tail length or a less polar head group than SDS. SDS and sodium hexyl sulfate were tested at 0.6 mM and 1-dodecanol was tested at 1 mM. (C) Quantification of the similar phasmid responses to SDS, dodecanoate, decanoate and nonanoate. All compounds except SDS were tested at 1 mM. Note that compounds are represented as the predominant form in a pH 7 solution, the carboxylates. (D) Quantification of the similar amphid responses to SDS, dodecanoate, decanoate and undecanoate, and slower responses to nonanoate and 1-decanol. All compounds except SDS were tested at 1 mM. Note that carboxylate forms are again represented. (A–C) For phasmid assays, one-way ANOVA and Tukey’s post-hoc test were performed. (D) For amphid assays, two-sample z-tests and the Hochberg multiple comparison adjustment procedure were performed. (A–D) For all experimental samples, n ≥ 40 for the experimental group, n ≥ 40 for the 0.6% SDS positive control, and n ≥ 40 for the saline buffer negative control. ***B, p<0.001; *B, p<0.05; ns B, not significant compared with a control buffer; ***S, p<0.001; ns S, not significant compared with SDS positive control. Exact values and additional pairwise comparisons are included in Figure 2—source data 1 and Figure 2—source data 2.

https://doi.org/10.7554/eLife.23770.006

To determine which structural elements of SDS are required for avoidance, we tested whether the hydrocarbon chain or the polar head group in SDS was required to induce avoidance. Sodium hexyl sulfate, which has a carbon chain half the length of SDS, elicited reduced chemosensation (Figure 2B). 1-Dodecanol, which has a smaller, less polar head group but the same carbon chain length, was not sensed by the phasmid neurons (Figure 2B). These findings indicate that both the tail length and the nature of the polar head group are important for the response.

Although SDS does not occur in nature, carboxylic acids with similar amphipathic structures do. Both SDS and these carboxylic acids have long hydrophobic alkyl tails and negatively charged head groups at pH 7. Interestingly, the NP10 strain of Streptomyces (Figure 1A) secretes carboxylic acids that have these characteristics (Ilic-Tomic et al., 2015). We therefore used the tail drop assay to test a series of carboxylic acids of varying chain lengths, with and without cis double-bonds (Figure 2C and Figure 2—figure supplement 1), to determine which, if any, C. elegans avoid. Dodecanoic acid and decanoic acid induced a robust avoidance response in the tail (Figure 2C, Figure 2—figure supplement 2). Additionally, these two compounds induce a robust avoidance response in the head (Figure 2D). Thus, sensation of these molecules in particular could contribute to the recognition and avoidance of Streptomyces bacteria by C. elegans.

To determine whether strains of Streptomyces that secrete toxins also secrete dodecanoic acid or decanoic acid, we extracted and labeled the fatty acids secreted from Streptomyces strains and B. subtilis. We used high-resolution mass spectrometry to determine relative production levels in cell-free supernatants (Figure 3, Figure 3—figure supplement 1). We found that S. avermitilis, S. costaricanus and S. milbemycinicus did indeed secrete dodecanoic acid, whereas dodecanoic acid secretion from B. subtilis was not detected above background. A lower signal for decanoic acid was also detected. Although additional cues from Streptomyces probably contribute to the C. elegans response, our observation of dodecanoic acid supports the hypothesis that this molecule is a cue that is produced by Streptomyces.

Figure 3 with 1 supplement see all
Detection of carboxylates secreted by S.avermitilis, S. costaricanus, and S. milbemycinicus.

Carboxylates were extracted from species of Streptomyces and labeled with a pyridinium moiety for easier detection by mass spectrometry. For example, to detect dodecanoate after Streptomyces cell-free supernatants were extracted and concentrated, dodecanoate and the other carboxylates in the extract were converted to pyridinium esters by treatment with carbonyldiimidazole (CDI) in acetonitrile (ACN).

https://doi.org/10.7554/eLife.23770.011

Given that the cue from Streptomyces is secreted, we hypothesized that a chemosensory receptor might mediate the avoidance response. There are two pairs of chemosensory neurons in the C. elegans tail that have exposed dendrites: the PHA and PHB chemosensory neurons (White et al., 1986; Hedgecock et al., 1985). These neurons mediate the rapid avoidance of SDS and other chemicals (Hilliard et al., 2002). In the head, additional neurons play roles in chemical avoidance, including the polymodal nociceptor ASH (Bargmann, 2006). Therefore, to identify the cognate receptor for dodecanoic acid, we performed an expression screen using existing data that are available on Wormbase.org (WormBase web site, 2017) for chemoreceptor genes that are expressed in the PHA and PHB neurons and in at least one set of head or amphid neurons. We performed RNA interference (RNAi) (Fraser et al., 2000; Kamath and Ahringer, 2003) to knock down the function of these genes and screened for defects in the phasmid response to dodecanoic acid. We ranked the chemoreceptor genes by the magnitude of the defect, and focused our work on the gene whose knock-down resulted in the most severe phenotype, that encoding the SRB-6 G protein-coupled receptor (GPCR) (Figure 4—figure supplement 1). srb-6 knock-down also resulted in the most severe defect among knockdowns of srb family members (Figure 4—figure supplement 1). In addition to srb-6’s expression in PHA and PHB neurons in the tail, this gene is expressed in only three sets of sensory neurons in the head; srb-6 is highly expressed in the ASH and ADL neurons, and weakly expressed in the ADF neurons (Troemel et al., 1995).

To further test the role of srb-6 in mediating the rapid avoidance response, we performed behavioral assays with srb-6 presumptive null mutants (Thompson et al., 2013). srb-6 mutants were defective in the tail (Figure 4A) and head (Figure 4B) avoidance response to species of Streptomyces and to dodecanoic acid. These data are consistent with the requirement of this GPCR in the avoidance of both Streptomyces and the fatty acid secreted by these bacteria. In the tail, the limited expression of the srb-6 GPCR to PHA and PHB neurons indicates that srb-6 might function in these neurons. To test this, we cloned the srb-6 cDNA and expressed it under the direction of the ocr-2 promoter. Within the tail region, this promoter selectively drives expression in the PHA and PHB neurons (Tobin et al., 2002). Expression of srb-6 in PHA and PHB significantly restored the srb-6 mutants’ ability to sense both S. avermitilis and dodecanoic acid at the tail (Figure 4C). Taken together, the endogenous expression of srb-6 in PHA and PHB neurons, the ability of srb-6 expression in PHA and PHB to rescue srb-6 behavioral defects, and the role of PHA and PHB as the sole chemosensory neurons in the tail, indicate that the srb-6 GPCR probably functions in one or both of these cells.

Figure 4 with 2 supplements see all
The GPCR SRB-6 is required for the response to Streptomyces, dodecanoate and decanoate.

(A) Quantification of the defect in the phasmid response to SDS, Streptomyces, dodecanoate and decanoate in srb-6 null mutants. (B) Quantification of the defect in the amphid response to Streptomyces, dodecanoate and decanoate in srb-6 mutants. (A–D) For each Streptomyces species in each panel, cell-free supernatants were used for wild-type and srb-6 mutants. (A–B) Decanoate and dodecanoate were tested at 1 mM. Note that carboxylate forms are represented. For experimental samples, n ≥ 40 for the experimental group, n ≥ 40 for the 0.6% SDS positive control, and n ≥ 40 for the saline buffer negative control. ***, p<0.001; NS, not significant. (C) Quantification of the significantly rescued response of srb-6 mutant animals expressing the srb-6 cDNA under the direction of the ocr-2 promoter (pocr-2::srb-6) to S. avermitilis and dodecanoate compared with the response of srb-6 mutant animals without the transgene. In the tail, the ocr-2 promoter drives srb-6 cDNA expression in PHA and PHB neurons. (D) Quantification of the significantly rescued response of srb-6 mutant animals carrying the pocr-2::srb-6 transgene to S. avermitilis and dodecanoate compared with that of srb-6 mutant animals without the transgene. In the head, the ocr-2 promoter drives expression in ASH, ADL, ADF and one other type of amphid neuron (AWA). (C–D) Decanoate and dodecanoate were tested at 1 mM. Note that carboxylate forms are represented. Two similar transgenic lines were examined. For experimental samples, n ≥ 20 for the experimental group, n ≥ 20 for the 0.6% SDS positive control, and n ≥ 20 for the saline buffer negative control. ***, p<0.001; **, p<0.01; *, p<0.05; ns, not significant. (A, C) For phasmid assays, one-factor ANOVA analysis was performed followed by two-sample t-tests and the Hochberg procedure for multiple comparisons. (B, D) For amphid assays, two-sample z-tests and the Hochberg multiple comparison adjustment procedure were performed. Exact values are included in Figure 4—source data 1. (E) Calcium imaging of ASH sensory neurons in living wild-type and srb-6 mutant animals expressing GCaMP6 (Chen et al., 2013) under the direction of the srb-6 promoter upon exposure to a cell-free supernatant of S. avermitilis. The bar graph shows the average areas under the curve (AUC) for the period of S. avermitilis stimulation and error bars are standard error of the mean (SEM). (F) Calcium imaging of ASH neurons in living wild-type and srb-6 mutant animals expressing GCaMP6 under the direction of the srb-6 promoter upon exposure to 1 mm dodecanoate. The bar graph represents the average AUC for the period of dodecanoate stimulation, and error bars are SEM. (E–F) Experimental animals were exposed to a control buffer for the first 15 s, then to stimulus for 60 s, followed by control buffer. Traces show the average percent change in GCaMP6 fluorescence (ΔF/F0) and shading indicates SEM. For experimental samples, n ≥ 15 for each genotype and stimulus. *, p<0.05 in a t-test. See Statistical Analysis in the Materials and methods section for additional details.

https://doi.org/10.7554/eLife.23770.013

In the head, the ocr-2 promoter drives expression in ASH, ADL and ADF gustatory neurons in addition to AWA olfactory neurons (Tobin et al., 2002). Using this promoter, we found that expression in this subset of neurons significantly rescued the ability of srb-6 mutants to respond to S. avermitilis and dodecanoic acid exposure at the head (Figure 4D). Although we cannot rule out a function in AWA, the endogenous expression of srb-6 in ASH, ADL and ADF (and not AWA) suggests a role in one or more of these cells.

To further test whether srb-6 could act as a receptor for both dodecanoic acid and cell-free supernatants from S. avermitilis, we visualized calcium changes in ASH neurons in wild-type and srb-6 mutant animals (Figure 4E,F). We found that calcium influx in response to both dodecanoic acid and S. avermitilis was severely reduced in srb-6 mutants, consistent with this GPCR having the ability to detect these stimuli.

The GPCR required for this avoidance response provides C. elegans with the ability to avoid potentially deadly prey. In order to understand the broader evolutionary context for this gene, we performed phylogenetic analysis of the srb-6 receptor. The srb family of chemoreceptors is nematode-specific (Troemel et al., 1995; Chen et al., 2005), consistent with a role for srb-6 in mediating nematode avoidance of Streptomyces. Phylogenetic analysis of the srb-6 receptor indicates that its closest homologs are within the Caenhorabditis genus. Although the sequences for most nematodes are not complete, the next monophyletic group contains several other species of nematodes with life cycles that include free-living soil-dwelling stages (Figure 4—figure supplement 2), suggesting possible conservation of this mechanism.

Discussion

Our results demonstrate that the bacteriovore C. elegans detects and escapes from Streptomyces species that secrete powerful toxins. C. elegans rapidly avoids Streptomyces secretions, which include dodecanoic acid. Interestingly, dodecanoic acid itself has been shown to have nematicidal activity (Tarjan and Cheo, 1956; Abdel-Rahman et al., 2008; Gu et al., 2005; Dong et al., 2014; Zhang et al., 2012). The SRB-6 GPCR is essential for the rapid avoidance of Streptomyces and dodecanoic acid. SRB-6 is expressed in the only two phasmid chemosensory neurons, PHA and PHB, as well as in a subset of chemosensory neurons in the amphids ASH, ADL and ADF (Troemel et al., 1995). Rescue of gene expression in these cells and in one additional amphid olfactory neuron largely restores the response of srb-6 mutants to Streptomyces and dodecanoic acid exposure at the tail or head, consistent with a role for srb-6 in those, or a subset of those, cells. PHA and PHB are the only two chemosensory neurons in the hermaphrodite tail, and they mediate a similar response to dilute SDS. ASH and ADL also mediate similar chemical avoidance responses, and ADF has been shown to play a role in salt chemotaxis (Hilliard et al., 2002; Bargmann, 2006; Jang et al., 2012). Our calcium-imaging experiments directly support a role for ASH in the SRB-6-mediated avoidance response.

The detection and escape from Streptomyces bacteria and their secretion, dodecanoic acid, described here are the first environmentally relevant chemosensory functions identified for the hermaphrodite phasmid circuit that inform our understanding of interactions with bacteria. Interestingly, a critical function of PHB phasmid sensory neurons in males has recently been discovered: sensing hermaphrodite-derived mating cues (Oren-Suissa et al., 2016). Synapses between PHBs and their normal partners in hermaphrodites are eliminated early in development, and PHBs form male-specific circuits that regulates this distinct function (Oren-Suissa et al., 2016). The hermaphrodite phasmid circuit regulates all chemosensation in the tail, suggesting that these sensilla may have evolved to allow the nematodes to avoid Streptomyces and other environmental threats. Interestingly, dodecanoic acid is secreted by crown daisy plants (Chrysanthemum coronarium L.), which are intercropped with tomatoes (Solanum lycopersicum) to prevent parasitism by root-knot nematodes (Meloidogyne incognita), and root-knot nematodes are repelled by 4 mM dodecanoic acid (but not by lower concentrations) (Dong et al., 2014). This, in addition to our phylogenetic analyses, suggests conservation among nematodes of this mechanism of escaping toxin-producing organisms that would otherwise be a good food source.

In addition to the detection of dodecanoic acid, we also discovered that decanoic acid elicits a strong escape response in C. elegans that is also mediated by the SRB-6 GPCR. However, high levels of decanoic acid are not secreted by Streptomyces. Decanoic acid has also been shown to have nematicidal properties (Tarjan and Cheo, 1956; Abdel-Rahman et al., 2008; Zhang et al., 2012), and decanoic acid, dodecanoic acid, and fatty acids of similar lengths are present in many plant seeds (Zhang et al., 2012), indicating that avoiding decanoic acid in the environment may also be important to nematodes, outside the context of bacterial avoidance.

The more well-studied interaction between C. elegans and bacterial species is the host–pathogen interaction, in which bacteria infect C. elegans. Caenorhabditis elegans is frequently attracted to pathogens initially, facilitating colonization of the host (Schulenburg et al., 2007; Meisel and Kim, 2014). In the predator–prey interaction disclosed in the present work, however, we demonstrate that C. elegans has evolved a mechanism to recognize and escape from the powerful defenses presented by Streptomyces. This avoidance mechanism probably arose through selection pressure on GPCRs, and by employment of reflex-like neural circuits in the head and tail sensory organs, a mechanism that may be common to other predator–prey interactions. Thus, this study deepens our understanding of the molecules, neural circuitry, and behavior used in the arms race between predators and their prey.

Materials and methods

Caenorhabditis elegans strains

Wild-type C. elegans strains were the Bristol strain N2 or wyIs157 IV ox9 (10). Strains were maintained using standard methods (Brenner, 1974). Worms were grown at 20°C and on NGM plates seeded with OP50 E. coli (RRID:WB-STRAIN:OP50). Mutations used in this study include eri-1(mg366) (Kennedy et al., 2004), and srb-6(gk925263) II (24), which was outcrossed four times by pushing out the linked marker unc-4(e120) II; wyIs157 IV ox4(10). srb-6(gk925263) II is likely to be null, as it has an early stop codon in srb-6: Y186Amber. The outcrossed srb-6(gk925263) II; wyIs157 IV ox4 strain was sequenced to verify the substitution mutation. Transgenes used for rescue of srb-6 and maintained as extrachromosomal arrays are iyEx348 and iyEx350 (30–50 ng/ul pocr-2::srb-6, 40 ng/ul podr-1::GFP [L'Etoile and Bargmann, 2000]), and ‘no transgene’ srb-6 animals were odr-1::GFP negative animals from iyEx348; srb-6 and iyEx350; srb-6. ocr-2 is expressed in PHA and PHB neurons in the tail, and in ASH, ADL, ADF and AWA neurons in the head (Tobin et al., 2002). The transgene used for calcium imaging is iyEx358 (30 ng/ul psrb-6::GCaMP6 and 20 ng/ul punc-22::DsRed).

Bacterial strains and media

For solid cultures, Streptomyces was grown on mannitol-soy agar at 28°C. Bacillus subtilis was grown on mannitol-soy agar or trypticase soy agar (TSA) at 37°C. Escherichia coli for drop tests was grown on TSA at 22°C. For liquid cultures, B. subtilis was grown at 37°C in yeast extract-malt extract (YEME) medium overnight. Streptomyces strains were grown at 28°C in YEME medium for three days. Aliquots of the bacterial culture were pelleted, and the resulting supernatants were filtered through a 0.22 µm filter. Higher-density inocula required dilution. For example, after a 3 day incubation of S. avermitilis, the resulting cell-free supernatant was diluted 1:1000 with an initial inoculum of 13,000 spores/mL, whereas an inoculum of 650 spores/mL needed no dilution in the behavior assays. Activity in the cell-free supernatant degrades over time.

Extraction and modification of fatty acids

A 20 mL aliquot of bacterial culture was pelleted, and the resulting supernatant was filtered through a 0.22 µm filter. The cell-free supernatant was extracted sequentially with a 1:5 mix of methanol/methyl tert-butyl ether using a 40 mL then an 80 mL portion of the organic solvents. The combined organic layers were concentrated under reduced pressure. Fatty acid derivatization was performed according to Narayana (Narayana et al., 2015). Briefly, the extraction concentrate was dissolved in 100 µL acetonitrile (ACN). A 50 µL portion of a 1 mg/mL solution of carbonyldiimidazole in ACN was added, and the tube was inverted for mixing for 2 min. Next, 50 µL of a 50 mg/mL solution of 3-hydroxymethyl-1-methylpyridinium in 95:5 ACN/Et3N was added to the reaction tube and inverted for mixing for an additional 2 min. The tube was heated at 40°C for 20 min. to complete the derivatization. A control of YEME medium spiked with decanoic acid was also extracted, derivatized and analyzed.

UPLC/Electrospray Ionization-MS analysis

The LC-MS analysis was performed on an Agilent 6530 Accurate Mass QToF LC/MS with a dual electrospray ionization source (Agilent) connected to a Agilent 1290 Infinity II UPLC front end equipped with an SB C18 column (1.8 um, 3.0 × 100 mm, 828975–302 [Agilent]). The gas temperature was 350°C. The VCap was set to 4000V. The instrument was operated in positive ion mode under MS conditions. Absorbance threshold was set at 5000 and the relative threshold was set at 1%. Data were collected and stored in centroid form. The mobile phase system was made up of Solvent A (95:5 water:ACN + 0.1% formic acid) and Solvent B (95:5 ACN:water +0.1% formic acid), and a flow rate of 0.2 mL/min. The column was equilibrated for 10 min at 40% B before samples were run. The method and gradient used to elute the derivatized fatty acids were: 0’ to 2’ 40% B to 62% B, 2’ to 3’ held at 62% B, 3’ to 5’ 62% B to 100% B, 5’ to 7’ held at 100% B, 7’ to 8’ 100% B to 40% B, 8’ to 13’ 40% B. All gradients were linear. The first 1.2 min of the run were sent to waste to remove underivatized tags. MS data werecollected from 1.2 min until the end of the run. Analysis was performed using Agilent MassHunter Qualitative Analysis Software. Exact masses were calculated using ChemDraw and analyzing standards. Total ionization curves (TICs) were extracted using these exact masses ± 100 ppm, the preset default by MassHunter to generate electrospray ionization curves (EICs). EICs were manually integrated and the area under the peak was calculated by MassHunter.

Phasmid (tail) avoidance assay

To examine the response of the phasmid neurons to repellents, tail dry-drop tests were performed as previously described (Park et al., 2011). Assays were performed on pre-dried plates (incubated overnight at 30°C), so that the drops quickly dry into the media, allowing sensation while eliminating wicking along the body of animals that might stimulate the head neurons. Briefly, a nose touch is administered to a day one gravid adult hermaphrodite to induce backward movement, then a capillary is used to administer a drop of either a control M13 buffer or an experimental sample behind the worm. The time that the animal backs into the drop before stopping is recorded. For Figures 12, we tested the response of at least 40 adults in M13 buffer and at least 40 adults to 0.6 mM SDS (diluted in M13) on the same day. The relative response index was calculated by dividing the average backing time into the experimental treatment by the average backing time into 0.6 mM SDS, so as to normalize the SDS response index to 100%. For Figure 4A and C, the relative response index was calculated as previously described (Park et al., 2011). Briefly, for the mutant animals, we divide the average backing time into the experimental treatment by the average backing time into buffer, and normalized this ratio with the ratio of the average response time of wild-type animals to SDS and to buffer on the same day.

wildtype:ExperimentaltreatmentwildtypeBufferwildtypeSDSwildtypeBufferwildtype=ExperimentaltreatmentwildtypeSDSwildtypeMutant:ExperimentaltreatmentmutantBuffermutantSDSwildtypeBufferwildtype

For buffer bars in Figure 4, the experimental treatment was buffer.

Most of the replication in our studies, including the phasmid avoidance assays described in this section and the amphid avoidance assays and chemotaxis assays described below, was biological. Different worms are used for every measurement in a sample group, as animals might otherwise adapt to treatments or nose touches. While there may also be sources of technical variation, such as slight imprecision in time measurements, we observed little variation in backup time for individual worms moving into saline buffer, indicating that the amount of technical variation is lprobably very low.

Amphid (head) avoidance assay

To examine the response of the amphid neurons to nonvolatile repellent, we adapted a behavioral assay developed by Hilliard and colleagues (Hilliard et al., 2002). As with the phasmid avoidance assay, tests were performed on pre-dried plates. A capillary was used to place a dry drop of either a control buffer or an experimental sample in front of a forward-moving worm. If a worm terminated forward movement and initiated backward movement after the head had contacted the dry drop area, we considered it a response. If the worm continued to move forward through the dry drop after the head region had moved into the area of the drop, then we considered that a failure to respond. Wild-type animals responded to the previously known repellent SDS the majority of the time. In the negative control buffer M13, the wild-type animals responded only approximately 20% of the time.

Chemotaxis assay

To examine the response of nematodes to volatile cues, chemotaxis assays were conducted as previously described (Bargmann et al., 1993). Diacetyl was diluted in ethanol at 1:1000. For bacterial cultures, 1 µl of cultured bacteria was used as the experimental stimulant, whereas the medium in which the bacteria were grown was used as a control. Chemotaxis indexes were then calculated using the equation: chemotaxisindex=#experimental#control#total.

Calcium imaging

Calcium-imaging experiments were conducted as previously described (Krzyzanowski et al., 2013). Briefly, microfluidic devices graciously provided by H. Gemrich and S. Kato at the University of California San Francisco were used to trap and immobilize the worms, exposing their nose to a stream of the control buffer (M13) or stimulant diluted in buffer (M13). Experimental animals were exposed first to buffer, then switched to stimulant after 15 s, then switched back to buffer after 60 s of exposure. Buffer control animals were instead exposed to buffer from the second channel (see Statistical Analysis). Fluorescent images were taken at 400X magnification every half second. ImageJ64 (Abramoff et al., 2004; Rasband, 2009) and the StackReg plugin were used to measure fluorescence intensity and to subtract background. The percent change in fluorescence intensity based on the intensity at the 15 s timepoint was calculated. High-inoculum S. avermitilis extract (13,000 spores/mL) diluted 1:50 in M13 buffer and 1 mM dodecanoic acid in M13 were used.

Statistical analysis

To calculate an appropriate sample size, we drew on our previous work using phasmid assays (Park et al., 2011), which showed that most effect sizes for mean back-up time comparisons are large. Using Cohen’s d = 0.8 (which corresponds to a large effect), we calculated sample sizes of approximately 26 animals per group (for significance level alpha = 0.05, and power = 0.8). There was no reason to restrict the sample sizes for most genotypes, therefore we used samples sizes of at least 40 for most genotypes. From a pilot study using amphid assays, a representative effect size corresponded to sample proportions of approximately 0.3 and 0.8, leading to a sample size estimate of 14 animals per group (for significance level alpha = 0.05 and power = 0.8). Sample size estimates were obtained using the pwr package in R. There was no reason to restrict the sample sizes to 14 for most genotypes, therefore we used samples sizes of at least 40 for most genotypes.

Results are reported in the form of P-values in the figures (*, p<0.05; **, p<0.01; ***, p<0.001; NS, p>0.05), and exact P-values are reported in the source data files for the figures. P-values provide precise information about whether two populations differ significantly in either means or proportions.

For the amphid (head) avoidance assay, statistical significance was determined by two-independent-sample z-tests, followed by the Hochberg procedure. Independent sample z-tests can be used to compare two population proportions if the sample sizes are sufficiently large. The Hochberg procedure is a standard procedure applied to adjust for the tendency to reject a null hypothesis incorrectly when multiple comparison are made, and can only conservatively increase P-values. A multiple comparison procedure is required whenever several conclusions are drawn from the same group of data in order to assure that the overall error rate for a type I error is still bound by the chosen significance level (here 0.05, 0.01, 0.001). For phasmid (tail) response assays in Figures 1 and 2, a one-way ANOVA model to compare the means of three or more populations was fitted in R. Tukey’s post-hoc test was performed in R to compare the means of the individual treatments while simultaneously adjusting for multiple comparisons. For the phasmid (tail) response assays reported in Figure 4, data representing mean length or means of percentages were analyzed with two-sample t-tests followed by the Hochberg procedure for multiple comparisons. For the pair-wise comparison of population means in the chemotaxis assays, Tukey’s post-hoc test was used to adjust for multiple comparisons.

Effect sizes (Cohen’s d) for major results include the following: wild-type phasmid response to S. avermitilis cell-free supernatant compared with buffer – 2.8; wild-type amphid response to S. avermitilis compared with buffer – 1.4; wild-type phasmid response to dodecanoic acid compared with buffer – 2.1; wild-type amphid response to dodecanoic acid compared with buffer –1.4; srb-6 phasmid response to dodecanoic acid compared to wild-type phasmid response to dodecanoic acid – 1.8; srb-6 amphid response to dodecanoic acid compared to wild-type amphid response to dodecanoic acid – 0.54.

For calcium-imaging experiments, we normalized the two different compound treatments to animals exposed to buffer in the first channel, then buffer in the second channel, then buffer in the first channel again (buffer control) by subtracting the averaged time series in the buffer control worms from each individual treatment time series. Then we computed the area under the curve during exposure to stimulus (15–75 s) for each individual normalized worm, and compared wild-type to srb-6 using a two-sample t-test. We did this separately for the two different treatments.

RNAi feeding experiments

The RNAi feeding screen was performed using the Ahringer library (Fraser et al., 2000) according to the Source BioScience protocol using the RNAi-sensitive strain eri-1(mg366) (Kennedy et al., 2004). L4s were picked onto plates seeded with each RNAi strain, and their adult progeny were tested for defects in responding to dodecanoic acid exposure at the tail. Assays were performed on at least two days for each RNAi clone.

Cloning and transgenics

To generate the pocr-2::srb-6 construct, the srb-6 cDNA was amplified from an N2 cDNA library, adding 5' NcoI and 3' SacI sites, subcloned into the NcoI-SacI fragment of pSMdelta (a generous gift from S. McCarroll, Harvard University, Cambridge, MA), and sequenced to generate pSMdelta::srb-6. The srb-6 sequence was identical to that predicted on Wormbase.org (WormBase web site, 2017). The ocr-2 promoter was amplified from N2 genomic DNA, adding 5' SphI and 3' SmaI sites, and subcloned into the SphI-SmaI fragment of pSMdelta::srb-6 to generate pocr-2::srb-6. To generate transgenic strains, pocr-2::srb-6 and the podr-1::GFP (37) coinjection marker were injected into srb-6(gk925263) ox6 unc-4(e120) trans-heterozygotes, non-Unc GFP-positive F1 hermaphrodites were singled out, F2 plates with no Unc F2 progeny were selected, and their progeny assayed for their response to S. avermitilis and dodecanoic acid and then sequenced. Defects in sensing S. avermitilis and dodecanoic acid were functionally assessed in non-array-carrying animals. To generate the calcium marker, the srb-6 promoter (Troemel et al., 1995) was amplified from genomic DNA, adding 5’ AfeI and 3’ XmaI sites, and inserted into the AfeI-XmaI fragment of flp-6::GCaMP6 to generate psrb-6::GCaMP6. To generate transgenic strains, psrb-6::GCaMP6 and the punc-122::DsRed coinjection marker were injected into unc-4(e120) heterozygotes. Wild-type and unc-4 mutants were isolated, srb-6(gk925263) ox6 was crossed into unc-4 mutants, srb-6 was homozygosed by pushing out the linked unc-4 gene, and the resulting lines were sequenced.

Phylogenetic analysis

The phylogenetic tree in Figure 1A was constructed using the Maximum Likelihood method based on the Hasegawa-Kishino-Yano model (Hasegawa et al., 1985). The tree with the highest log likelihood is shown. The percentage of trees in which the associated taxa clustered together is shown next to the branches. Initial tree(s) for the heuristic search were obtained automatically by applying Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances estimated using the Maximum Composite Likelihood (MCL) approach, and then selecting the topology with superior log likelihood value. The analysis involved five nucleotide sequences. There were a total of 1753 positions in the final dataset. Evolutionary analyses were conducted in MEGA7 (Kumar et al., 2016).

The C. elegans SRB-6 protein sequence record P54141.1 (aka NP_496199.1) was retrieved from NCBI. A pre-computed BLAST (RRID:SCR_004870) search was retrieved (https://www.ncbi.nlm.nih.gov/sutils/blink.cgi?pid=1711518). A BLAST search conducted against non-redundant database (nr) was re-run using the same settings. A dataset containing exactly 100 protein sequences was retrieved by BLAST with E-value <0.0001. A phylogenetic tree was built using the online web-server http://www.phylogeny.fr/

References

  1. 1
  2. 2
    Image processing with imagej
    1. MD Abramoff
    2. PJ Magelhaes
    3. SJ Ram
    (2004)
    Biophotonics International 11:36–42.
  3. 3
  4. 4
  5. 5
    The genetics of caenorhabditis elegans
    1. S Brenner
    (1974)
    Genetics 77:71–94.
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
  20. 20
  21. 21
  22. 22
  23. 23
  24. 24
  25. 25
  26. 26
  27. 27
  28. 28
  29. 29
  30. 30
  31. 31
  32. 32
  33. 33
  34. 34
    ImageJ
    1. WS Rasband
    (2009)
    U S National Institutes of Health, Bethesda, Maryland, USA.
  35. 35
  36. 36
  37. 37
    Nematocidal Value of Some Fatty Acids
    1. AC Tarjan
    2. PC Cheo
    (1956)
    Kingston, RI: Agricultural Experiment Station, University of Rhode Island.
  38. 38
  39. 39
  40. 40
  41. 41
  42. 42
  43. 43
  44. 44
    WormBase web site
    1. WormBase web site
    (2017)
    Release WS258.
  45. 45

Decision letter

  1. Oliver Hobert
    Reviewing Editor; Howard Hughes Medical Institute, Columbia University, United States

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

Thank you for submitting your article "C. elegans avoids toxin-producing Streptomyces using a seven transmembrane domain chemosensory receptor" for consideration by eLife. Your article has been reviewed by two peer reviewers, and the evaluation has been overseen by Oliver Hobert as Reviewing Editor and a Senior Editor. The following individual involved in review of your submission has agreed to reveal her identity: Leslie B Vosshall (Reviewer #1).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

The reviewers appreciate the description of a potential mechanism by which C. elegans nematodes detect and avoid toxic bacteria. The paper catalogues the species of bacteria that induce avoidance responses, use analytical chemistry to identify the bioactive repellents, and discover the sensory neurons and receptors responsible. The work is an engaging read on many levels: behavioral neuroscience, chemical ecology, and chemoreceptor biology.

The main experimental issue that needs to be addressed is based on the fact that the srb-6 phenotype is based on a single allele and no rescue data is provided. Since the focus of srb-6 action is also not explored, the reviewers suggest addressing both issues simultaneously by rescuing the srb-6 phenotype with available neuron-type specific promoters. This would confirm that srb-6 is indeed responsible for the phenotype and assess where the gene acts.

Other points that need to be fixed, but may not require further experimentation:

1) Statistical analysis throughout seems flawed. Rather than comparing everything to the saline buffer control and reporting significance relative to this, can the authors do ANOVA analysis and show us which of the data points differ from each other? Conventional presentation is to label plots that differ at some significance level with different letters.

2) It's not clear from the experiments what the actual link is between Streptomyces avoidance and dodecanoate (dodecanoic acid? see comment below) avoidance. First, most compounds that were tested elicited an avoidance response in this assay, and presumably many other bacterial metabolites that were not tested would also elicit a similar response. Second, what concentrations of metabolites were tested? This information does not appear to be provided. Was dose-response analysis performed?

3) There is confusion about the actual compounds that were tested. The authors refer to "dodecanoate" and "decanoate" as the tested compounds. But this isn't actually a compound. Does this mean sodium dodecanoate and sodium decanoate, or dodecanoic acid and decanoic acid? Based on the chemical structures in Figure 2, I'm guessing it was the acids that were tested. But then in the Discussion, "dodecanoate" is mentioned in the same paragraph as "dodecanoic acid" as if these are two different compounds. The authors need to specify exactly which chemicals they are testing and refer to them consistently throughout the paper. A table with CAS numbers would be helpful. Also, these compounds are referred to as "water-soluble cues" when I don't think they are very water-soluble.

4) Results paragraph three: This paragraph is confusing. Is there a reason to think that PHA and PHB only sense SDS and similar compounds? Also, how does the conclusion that Streptomyces is unlikely to act by physically disrupting the nematode follow from the fact that the surfactant activity of SDS is not required for avoidance? If PHA and PHB express multiple chemoreceptors, presumably they can sense a number of unrelated compounds.

5) Figure 2:

a) It would help to show molecular structures for C and E in addition to D.

b) It would be helpful to show statistics relative to the SDS positive control in addition to the buffer negative control. Based only on comparisons to the buffer control, it's hard to tell which compounds are as effective as SDS in eliciting an avoidance response.

c) In E, why are only a subset of the tested compounds listed in the legend? Also, why was a two-sample z-test used? I thought a z-test was for cases where you know there are two different populations embedded within the larger population (e.g. males vs. females).

6) Figure 3: The legend needs to be more detailed to be comprehensible to a non-chemist.

7) Figure 4: The authors mention that they did an expression screen, but no details are provided. Was this by transcriptional profiling? Using existing data available on WormBase? Were other receptors found that were also expressed in the PHA and PHB neurons in addition to a subset of amphid neurons? Since there is no information about the expression screen, it's not clear how unique srb-6 really is. A rationale should be provided for why srb-6 was picked for analysis.

8) Along the same lines as previous point: Have other GRCR mutants with similar expression pattern been tested for Streptomyces avoidance; if so, this data could be provided as control? Similarly, have other srb genes been tested? It would also be nice to see data for a screen of other srb mutants to see if the defect seen with srb-6 is really specific to srb-6.

9) The authors show that srb-6 mutants are defective in their response to S. avermitilis. Have Streptomyces species been tested? If srb-6 is the primary receptor gene required for avoidance of Streptomyces, you'd expect srb-6 mutants to be defective in their avoidance of all three species.

10) Figure 4—figure supplement 1: It would help to highlight srb-6 so it is easier to find it in the phylogenetic tree. Also, aren't all of the srb genes nematode-specific?

11) Were the behavioral experiments performed across multiple days to control for day-to-day variation in behavior?

12) It would improve presentation clarity if the authors de-cluttered their plots by eliminated extraneous vertical lines to mark off the axis units

13) Figure 1—figure supplement 2 contains bar plots, which should be converted to box plots or dot plots so that readers can see something closer to the raw data

14) Figure 2A and 3A: hard to see the black text on top of the cyan semi-circle. Maybe play with color to enhance clarity

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Thank you for submitting your revised article "C. elegans avoids toxin-producing Streptomyces using a seven transmembrane domain chemosensory receptor" for consideration by eLife. Your article has been reviewed by two peer reviewers and the evaluation has been overseen by a Reviewing Editor and a Senior Editor. The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

It was generally agreed upon that the revisions did not properly address the concerns raised in the original review. The specific problems are as follows:

1) Figures 4 D and E need to be modified to include data from the mutant. Mutant rescue cannot be interpreted without side-by-side data from the mutant vs. wild type vs. mutant rescue. From the graph, it looks like the authors did not test wt, srb-6, and ocr-2::srb-6 in parallel. However, in the methods section, the authors say that they injected the rescue construct into srb-6 mutants, and in the behavioral assays they tested "non-array carrying animals (data not shown)". Based on this statement, it's possible that the authors did test wt, srb-6 (non-array carrying animals), and ocr-2::srb-6 in parallel, but for some reason didn't show this. If this really is what they did, they should show the data and do the appropriate statistics to show that srb-6 mutants are different from wt and the rescue line when tested in parallel. If this is not what they did, the authors need to repeat the experiment and test the three genotypes in parallel.

2) Figure 4 D: where is the missing "wild type buffer" control?

3) Since the authors have not individually rescued srb-6 function in each of the 5 neurons in which it is expressed, they should soften language and interpretations about circuitry accordingly.

4) The authors' description of the rescue is ambiguous. In the main text, they say that they used a promoter that "drives expression selectively in PHA and PHB." However, they chose a promoter that is expressed in PHA, PHB, ASH, ADL, ADF, and AWA. The language implies to me that srb-6 is only expressed in the phasmid neurons, but this is not the case. In the Materials and methods, they also talk about "cell-specific rescue," which again is ambiguous.

https://doi.org/10.7554/eLife.23770.019

Author response

The main experimental issue that needs to be addressed is based on the fact that the srb-6 phenotype is based on a single allele and no rescue data is provided. Since the focus of srb-6 action is also not explored, the reviewers suggest to address both issues simultaneously by rescuing the srb-6 phenotype with available neuron-type specific promoters. This would confirm that srb-6 is indeed responsible for the phenotype and assess where the gene acts.

We agree that this is an interesting experiment. We cloned and sequenced the srb-6 cDNA, which has a sequence identical to that predicted on Wormbase.org. Troemel et al., 1995 reported that srb-6 is strongly expressed in PHA and PHB neurons in the tail, and ASH and ADL neurons in the head, and also weakly expressed in ADF neurons in the head, which are presynaptic to ASH. We therefore searched for a promoter that was expressed in ASH, ADL and ADF neurons and as few other pairs of head neurons as possible. The well characterized gene ocr-2 is expressed in these three pairs of cells and only one other, the AWA olfactory neurons. ocr-2 is additionally expressed in PHA and PHB neurons in the tail. We therefore generated an pocr-2::srb-6 transgene, allowing us to perform cell-specific rescue assays in the head and tail with the same transgenic animals. The ability of srb-6 mutants to respond to S. avermitilis and dodecanoic acid exposure at the tail and head, respectively, was rescued by addition of the pocr-2::srb-6 transgene. These results are shown in Figure 4D and Figure 4E and are described in the Results section with the following text:

“In the tail, the limited expression of the srb-6 GPCR in PHA and PHB neurons indicates that srb-6 likely functions in these neurons to drive the tail response. […] Taken together, these data and previous results indicate that the circuit responsible for mediating the response to Streptomyces and dodecanoic acid exposure at the head may differ from the circuit responsible for the response to SDS.”

In the Discussion, we also added the following text:

“Rescue of gene expression in these cells and one additional amphid olfactory neuron largely restores the response of srb-6 mutants to Streptomyces and dodecanoic acid exposure at the tail and head, respectively. […] While ASH, ADL and ASK mediate the response to SDS exposure at the head, ASH, ADL and ADF likely mediate the response to Streptomyces and dodecanoic acid exposure at the head, indicating distinct response circuits for these structurally similar compounds.”

In addition, srb-6 RNAi produces a similar phenotype to that observed for the srb-6 allele used in these studies (see response to reviewer comment 7), which should also help to address this concern.

Other points that need to be fixed, but may not require further experimentation:

1) Statistical analysis throughout seems flawed. Rather than comparing everything to the saline buffer control and reporting significance relative to this, can the authors do ANOVA analysis and show us which of the datapoints differ from each other? Conventional presentation is to label plots that differ at some significance level with different letters.

We performed 1-factor ANOVA analysis for relative response index data in Figure 1C, Figure 2B, C, and D, Figure 4B, and Figure 4D. In order to better illustrate the similarities and differences in backing times to SDS and buffer, respectively (as requested in the similar reviewer comment #5 part 2), we modified the appropriate figure panels to include both of these comparisons (Figure 1C, Figure 2B, Figure 2C, and Figure 2D). In addition, we now include tables of Tukey’s post-hoc test values for all pairwise comparisons for each of these figures in the source data files associated with these figures.

Since the worms utilized for the amphid assays are not the same in different treatment groups (e.g. buffer, SDS, dodecanoic acid, etc.), and we are interested in comparing proportions of worms that react to the treatment across the different groups, we are using independent sample Z-tests for proportions to establish significance. In order to better illustrate the similarities and differences of proportions to SDS and buffer, respectively (and address reviewer comment #5 part 2), we modified the appropriate figure panels to include both of these comparisons (Figure 1E and Figure 2E). Additionally, we included tables of all pairwise comparisons for each of these figures in the source data files associated with these figures. All p-values are adjusted by the Hochberg method for multiple comparisons.

2) It's not clear from the experiments what the actual link is between Streptomyces avoidance and dodecanoate (dodecanoic acid? see comment below) avoidance. First, most compounds that were tested elicited an avoidance response in this assay, and presumably many other bacterial metabolites that were not tested would also elicit a similar response.

We propose in this manuscript that dodecanoic acid secreted by Streptomyces in a contributor to C. elegans avoidance of these toxin-producing species. It is likely that other secreted factors also contribute to this response. In support of this, we demonstrate that C. elegans avoids the species that secrete dodecanoic acid, species of Streptomyces, but do not avoid the control species that does not secrete dodecanoic acid, B. subtilis. We also identify a G protein-coupled receptor, srb-6, that is required for the response to both Streptomyces and dodecanoic acid. To clarify this for the readers, we have revised text in the Results: “While additional cues from Streptomyces likely contribute to the C. elegans response, our observation of dodecanoic acid supports the hypothesis that it is a cue produced by Streptomyces.” Also please note in the Discussion: “C. elegans rapidly avoids secretions from the bacteria, which include dodecanoic acid.”

To address the point that most compounds that were tested elicited an avoidance response in this assay, our goal was to understand the structure-activity relationship of compounds related to SDS, the most potent chemical activator of the phasmid neurons known, and a potent activator of ASH and ADL in the head in hermaphrodites. We predicted that most related compounds would have some activity. Still, we found that certain components were critical for detection. For example, when we removed the carboxylate or sulfate group (e.g. 1-dodecanol (Figure 2C, 2E) or 1-decanol (Figure 2D)), we lost all activity.

To address the point that presumably many other bacterial metabolites that were not tested would also elicit a similar response, we propose that the toxic Streptomyces secretion dodecanoic acid is avoided, but non-toxic secretions from bacteria that serve as a food source for C. elegans would be unlikely to elicit an escape response. In support of this, we show that C. elegans avoided secretions from Streptomyces, but they do not rapidly avoid secretions from E. coli or B. subtilis. However, as stated above, we agree that other secretions from Streptomyces likely contribute to the avoidance response. We have softened language in the paper to reflect the possibility that other bacterial secretions could be involved in this avoidance.

Second, what concentrations of metabolites were tested? This information does not appear to be provided. Was dose-response analysis performed?

Thank you for catching this accidental deletion. We have included concentration information to the captions of Figure 2 and Figure 4. Dose-response analysis for SDS, decanoic acid, and dodecanoic acid was also added as Figure 2—figure supplement 2.

3) There is confusion about the actual compounds that were tested. The authors refer to "dodecanoate" and "decanoate" as the tested compounds. But this isn't actually a compound. Does this mean sodium dodecanoate and sodium decanoate, or dodecanoic acid and decanoic acid? Based on the chemical structures in Figure 2, I'm guessing it was the acids that were tested. But then in the Discussion, "dodecanoate" is mentioned in the same paragraph as "dodecanoic acid" as if these are two different compounds. The authors need to specify exactly which chemicals they are testing and refer to them consistently throughout the paper. A table with CAS numbers would be helpful. Also, these compounds are referred to as "water-soluble cues" when I don't think they are very water-soluble.

We apologize for the confusion and thank the editors for their suggestion of adding CAS numbers. We have added a table with the CAS numbers (Figure 2—figure supplement 1). In addition, we used the terms for the carboxylic acids in the text, but have added the following statement to make it clear that at pH 7, the carboxylate form would be the predominant form in solution: “Both SDS and these carboxylic acids have long hydrophobic alkyl tails and negatively charged head groups at pH 7.” In the figures, we have maintained the terms and structures for the carboxylate forms, because these are the relevant forms when the compounds are dissolved in the pH 7 assay buffer. We feel that this is important to demonstrate the structural similarities among these molecules. For further clarity, in each appropriate figure panel (Figure 2D, 2E, 4B–E) we added in the statement that compounds are represented as the predominant form in a pH 7 solution, the carboxylates. To prevent confusion, we also removed the three uses of the term “water-soluble” from the text.

4) Results paragraph three: This paragraph is confusing. Is there a reason to think that PHA and PHB only sense SDS and similar compounds? Also, how does the conclusion that Streptomyces is unlikely to act by physically disrupting the nematode follow from the fact that the surfactant activity of SDS is not required for avoidance? If PHA and PHB express multiple chemoreceptors, presumably they can sense a number of unrelated compounds.

We acknowledge that this point was not well explained in the original manuscript. We reworked this paragraph to clarify. It now reads:

“Previously, the PHA and PHB-class neurons in the tail were shown to act in a defined circuit respond to the man-made detergent SDS, which does not appear in nature. […] Therefore, we looked to define the structural elements of SDS that are responsible for the avoidance response.”

We do not think that PHA and PHB only sense SDS and similar compounds. However, SDS is the compound that elicits the most robust phasmid response known, and few other compounds have been fully characterized with respect to these neurons. This is why we chose it as a starting point for our studies. To make this more clear in the text, we added the following sentence: “Few compounds have been fully characterized with respect to these neurons, however dilute concentrations of sodium dodecyl sulfate (SDS) have been shown to elicit a rapid and robust avoidance response in hermaphrodites, the primary sex found in nature.”

Finally, we agree that PHA and PHB express multiple chemoreceptors, and presumably can sense a number of unrelated compounds. We explained why we focused on SDS and related compounds in this study above, but it would be very interesting to explore this large topic in future studies.

5) Figure 2:

a) It would help to show molecular structures for C and E in addition to D.

We have added in molecular structures for Figure 2C and 2E in addition to Figure 2D.

b) It would be helpful to show statistics relative to the SDS positive control in addition to the buffer negative control. Based only on comparisons to the buffer control, it's hard to tell which compounds are as effective as SDS in eliciting an avoidance response.

Please see the response to reviewer comment #1, which addresses this directly. In summary, we have added in statistics relative to the SDS positive control in addition to the buffer negative control for all appropriate Figure panels.

c) In E, why are only a subset of the tested compounds listed in the legend? Also, why was a two-sample z-test used? I thought a z-test was for cases where you know there are two different populations embedded within the larger population (e.g. males vs. females).

We have now listed all compounds in the legend for Figure 2E. For the question regarding the use of the two-sample z-test, please see the second paragraph in the response to reviewer comment #1, which addresses this directly.

6) Figure 3: The legend needs to be more detailed to be comprehensible to a non-chemist.

More detail has been added to the Figure 3 legend, and for clarity, we have also added labels into the figure. The legend now reads:

Figure 3.Detection of carboxylates secreted by S. avermitilis, S. costaricanus, and S. milbemycinicus. Carboxylates were extracted from species of Streptomyces and labeled with a pyridinium moiety for easier detection by mass spectrometry. For example, to detect dodecanoate, after Streptomyces cell-free supernatants were extracted and concentrated, dodecanoate and the other carboxylates in the extract were converted to pyridinium esters by treatment with carbonyldiimidazole (CDI) in acetonitrile (ACN).”

7) Figure 4: The authors mention that they did an expression screen, but no details are provided. Was this by transcriptional profiling? Using existing data available on WormBase? Were other receptors found that were also expressed in the PHA and PHB neurons in addition to a subset of amphid neurons? Since there is no information about the expression screen, it's not clear how unique srb-6 really is. A rationale should be provided for why srb-6 was picked for analysis.

Using existing data available on WormBase, we screened for receptors expressed in the PHA and PHB neurons in addition to a subset of amphid neurons. The majority of these genes do not have characterized loss-of-function alleles, therefore we used RNAi to screen for behavioral defects. srb-6 RNAi had the most severe defect. We have added Figure 4—figure supplement 1 to present the results of this RNAi screen. We also revised the appropriate section of the Results to describe the screen. The text now reads:

“To identify the cognate receptor, we performed an expression screen using existing data available on Wormbase.org for chemoreceptors that are expressed in the PHA and PHB neurons and at least one set of amphid neurons. […] We ranked them by the magnitude of the defect, and focused our work on the gene whose knock-down resulted in the most severe phenotype, namely the srb-6 G protein-coupled receptor (GPCR) (Figure 4—figure supplement 1).”

8) Along the same lines as previous point: Have other GRCR mutants with similar expression pattern been tested for Streptomyces avoidance; if so, this data could be provided as control? Similarly, have other srb genes been tested? It would also be nice to see data for a screen of other srb mutants to see if the defect seen with srb-6 is really specific to srb-6.

For testing of other GPCRs with similar expression patterns please see our response to reviewer comment #7. To address the question about other srb genes, only one srb gene has previously reported phasmid expression, srb-3, but RNAi of srb-3 did not yield a severe behavioral defect in our RNAi screen (see response to reviewer comment #7). However to further address this comment, in case expression in the phasmid was not reported, we obtained all RNAi strains available for srb genes from the Ahringer collection (excluding pseudogenes) and performed RNAi. srb-6 again had the most severe defect. The data has been included in Figure 4—figure supplement 1. We added the following sentence to the Results section:

srb-6 similarly had the most severe defect within the srb family of genes (Figure 4—figure supplement 1).”

9) The authors show that srb-6 mutants are defective in their response to S. avermitilis. Have Streptomyces species been tested? If srb-6 is the primary receptor gene required for avoidance of Streptomyces, you'd expect srb-6 mutants to be defective in their avoidance of all three species.

We tested srb-6 mutants for their response to exposure to S. costaricanus, and S.milbemycinicus at the tail and head and found srb-6 mutants to be defective in their avoidance of these species. We have added this data into Figure 4B and Figure 4C.

10) Figure 4—figure supplement 1: It would help to highlight srb-6 so it is easier to find it in the phylogenetic tree. Also, aren't all of the srb genes nematode-specific?

We have highlighted srb-6, its closest homologs within the Caenhorabditis genus, and the monophyletic group containing nematodes outside the Caenhorabditis genus in different colors (now Figure 4—figure supplement 2, entitled “C. elegans srb-6 has homologs within the Caenhorabditis genus and the nematode phylum”). We have also revised and moved the section describing the tree to attempt to make our purpose more clear. The section is now at the end of the Results and reads:

“These circuits in the head and tail of C. elegans and the GPCR required for the avoidance response provide C. elegans with the machinery needed to avoid potentially deadly prey. […] Although the sequences for most nematodes are not complete, the next monophyletic group contains several other species of nematodes with life cycles that include free-living soil-dwelling stages (Figure 4—figure supplement 2), suggesting possible conservation of this mechanism.”

11) Were the behavioral experiments performed across multiple days to control for day-to-day variation in behavior?

The sequence of behavioral experiments was performed across several years. In the phasmid assay, while absolute back-up times into both buffer and compounds are variable from day to day, the ratio of back-up times between SDS and buffer, between other compounds and buffer, and between different compounds and SDS, stays fairly consistent. Therefore to control for day-to-day variation, we work with ratios of individual back-up times in a given compound compared to average backup times in SDS. In the amphid assay, not much variation from day to day was observed for proportions of animals that responded to SDS or buffer. However, both RNAi and chemotaxis assays performed in Figure 1—figure supplement 2 and Figure 4—figure supplement 1 are variable from day-to-day due to variable success of RNAi and possible other sources of variation in chemotaxis assays, so these experiments were performed on more than one day in order to be able to quantify day-to-day variation.

12) It would improve presentation clarity if the authors de-cluttered their plots by eliminated extraneous vertical lines to mark off the axis units

We have removed vertical lines to mark off the axis units from all plots.

13) Figure 1—figure supplement 2 contains bar plots, which should be converted to box plots or dot plots so that readers can see something closer to the raw data

The figure has been converted into a dot plot.

14) Figure 2A and 3A: hard to see the black text on top of the cyan semi-circle. Maybe play with color to enhance clarity

We have lightened the cyan color, and are happy to lighten it further if the text is still hard to see on your screen.

We have one additional note. When we performed the cell-specific rescue experiments requested by the reviewers, we used new Streptomyces preparations, and the response rate in the head was a little higher in wild-type animals, possibly due to improved growth conditions. We have revised the data in Figure 1E to be consistent with the new data in Figure 4C, but no text changes were required, as this does not change the interpretation.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

1) Figures 4 D and E need to be modified to include data from the mutant. Mutant rescue cannot be interpreted without side-by-side data from the mutant vs. wild type vs. mutant rescue. From the graph, it looks like the authors did not test wt, srb-6, and ocr-2::srb-6 in parallel. However, in the Materials and methods section, the authors say that they injected the rescue construct into srb-6 mutants, and in the behavioral assays they tested "non-array carrying animals (data not shown)". Based on this statement, it's possible that the authors did test wt, srb-6 (non-array carrying animals), and ocr-2::srb-6 in parallel, but for some reason didn't show this. If this really is what they did, they should show the data and do the appropriate statistics to show that srb-6 mutants are different from wt and the rescue line when tested in parallel. If this is not what they did, the authors need to repeat the experiment and test the three genotypes in parallel.

We have tested wild type, srb-6 (non-transgene) and ocr-2::srb-6; srb-6 in parallel, and performed the appropriate statistics to show that srb-6 mutants are different from wild-type and the rescue lines when tested in parallel, see Figure 4C and D (previously Figure 4D and E) and figure legend.

2) Figure 4 D: where is the missing "wild type buffer" control?

We thank the reviewer for pointing out this potential point of confusion. We have added a wild type buffer control to Figure 4C (previously 4D) and we have revised our Materials and methods section to make more clear that the relative response index measurements, which we previously simply cited, include a normalization to buffer controls (Statistical analysis, final paragraph).

3) Since the authors have not individually rescued srb-6 function in each of the 5 neurons in which it is expressed, they should soften language and interpretations about circuitry accordingly.

We agree and have revised the text accordingly. Specifically, we have softened the language and interpretations about circuitry throughout the manuscript. Major examples include the following: in the Results, we are now careful to say that one or a subset of the cells mediates the response in the tail and in the head. Similarly, in the Discussion, instead of discussing a model in which all of the srb-6 expressing neurons function in the same circuit, we instead mention chemosensory roles of each sensory neuron in which srb-6 is expressed. These changes streamlined the focus of the paper and we thank you for suggesting them. In addition, we performed calcium imaging experiments on ASH neurons in wild-type and srb-6 mutant animals, and found that srb-6 mutants have significantly reduced calcium influx in response to both S. avermitilis and dodecanoic acid, thus specifically implicating this neuron in the head and indicating that srb-6 can act as a receptor for both stimuli (Figure 4E-F).

4) The authors' description of the rescue is ambiguous. In the main text, they say that they used a promoter that "drives expression selectively in PHA and PHB." However, they chose a promoter that is expressed in PHA, PHB, ASH, ADL, ADF, and AWA. The language implies to me that srb-6 is only expressed in the phasmid neurons, but this is not the case. In the Materials and methods, they also talk about "cell-specific rescue," which again is ambiguous.

We thank the reviewers for drawing this potential point of confusion to our attention. We were trying to convey that within the tail region, the promoter selectively drives expression in PHA and PHB neurons. Our previous phrasing, “[…] we cloned the srb-6 cDNA and expressed it under the direction of a promoter that drives expression selectively in the PHA and PHB neurons in the tail region (pocr-2)” could be ambiguous. Therefore we have clarified to the following statement, “[…] we cloned the srb-6 cDNA and expressed it under the direction of the ocr-2 promoter. Within the tail region, this promoter selectively drives expression in the PHA and PHB neurons (25).” It is important to note that in the next paragraph we describe the use of the ocr-2 promoter to rescue srb-6 function in chemosensory neurons in the head, and also use the name of the ocr-2 promoter in Figure 4 and in the Materials and methods. Just in case there might be any similar confusion about expression of the ocr-2 promoter in the head region, we also moved “Within the head” to the beginning of the sentence/paragraph to clarify; the text now reads, “In the head, the ocr-2 promoter drives expression in ASH, ADL and ADF gustatory neurons in addition to AWA olfactory neurons.” Also, in the Materials and methods, we removed the word “cell-specific”. The sentence now reads, “Transgenes used for rescue of srb-6 and maintained as extrachromosomal arrays are iyEx348 and iyEx350

https://doi.org/10.7554/eLife.23770.020

Article and author information

Author details

  1. Alan Tran

    Department of Biological Sciences, San Jose State University, California, United States
    Contribution
    Conceptualization, Formal analysis, Investigation, Visualization, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  2. Angelina Tang

    Department of Biological Sciences, San Jose State University, California, United States
    Contribution
    Conceptualization, Formal analysis, Investigation, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  3. Colleen T O'Loughlin

    Department of Bioengineering and Therapeutic Sciences, University of California San Francisco, San Francisco, United States
    Contribution
    Conceptualization, Formal analysis, Investigation, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  4. Anthony Balistreri

    Department of Chemistry, San Jose State University, California, United States
    Contribution
    Conceptualization, Investigation, Writing—original draft, Writing—review and editing
    Contributed equally with
    Eric Chang, Doris Coto Villa, Joy Li and Aruna Varshney
    Competing interests
    No competing interests declared
  5. Eric Chang

    Department of Biological Sciences, San Jose State University, California, United States
    Contribution
    Conceptualization, Formal analysis, Investigation, Writing—original draft, Writing—review and editing
    Contributed equally with
    Anthony Balistreri, Doris Coto Villa, Joy Li and Aruna Varshney
    Competing interests
    No competing interests declared
  6. Doris Coto Villa

    Department of Biological Sciences, San Jose State University, California, United States
    Contribution
    Conceptualization, Formal analysis, Investigation, Writing—original draft, Writing—review and editing
    Contributed equally with
    Anthony Balistreri, Eric Chang, Joy Li and Aruna Varshney
    Competing interests
    No competing interests declared
  7. Joy Li

    Department of Biological Sciences, San Jose State University, California, United States
    Contribution
    Conceptualization, Investigation, Writing—original draft, Writing—review and editing
    Contributed equally with
    Anthony Balistreri, Eric Chang, Doris Coto Villa and Aruna Varshney
    Competing interests
    No competing interests declared
  8. Aruna Varshney

    Department of Biological Sciences, San Jose State University, California, United States
    Contribution
    Conceptualization, Investigation, Writing—original draft, Writing—review and editing
    Contributed equally with
    Anthony Balistreri, Eric Chang, Doris Coto Villa and Joy Li
    Competing interests
    No competing interests declared
  9. Vanessa Jimenez

    Department of Biological Sciences, San Jose State University, California, United States
    Contribution
    Conceptualization, Investigation, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  10. Jacqueline Pyle

    Department of Biological Sciences, San Jose State University, California, United States
    Contribution
    Conceptualization, Investigation, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  11. Bryan Tsujimoto

    Department of Biological Sciences, San Jose State University, California, United States
    Contribution
    Conceptualization, Investigation, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  12. Christopher Wellbrook

    Department of Biological Sciences, San Jose State University, California, United States
    Contribution
    Conceptualization, Investigation, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  13. Christopher Vargas

    Department of Biological Sciences, San Jose State University, California, United States
    Contribution
    Conceptualization, Investigation, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  14. Alex Duong

    Department of Biological Sciences, San Jose State University, California, United States
    Contribution
    Conceptualization, Investigation, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  15. Nebat Ali

    Department of Biological Sciences, San Jose State University, California, United States
    Contribution
    Conceptualization, Investigation, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  16. Sarah Y Matthews

    Department of Chemistry, San Jose State University, California, United States
    Contribution
    Conceptualization, Investigation, Writing—review and editing
    Competing interests
    No competing interests declared
  17. Samantha Levinson

    Department of Chemistry, San Jose State University, California, United States
    Contribution
    Conceptualization, Investigation, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  18. Sarah Woldemariam

    Department of Cell & Tissue Biology, University of California San Francisco, San Francisco, United States
    Contribution
    Conceptualization, Methodology, Writing—review and editing
    Competing interests
    No competing interests declared
  19. Sami Khuri

    Department of Computer Science, San Jose State University, California, United States
    Contribution
    Conceptualization, Formal analysis, Investigation, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  20. Martina Bremer

    Department of Mathematics and Statistics, San Jose State University, California, United States
    Contribution
    Conceptualization, Formal analysis, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  21. Daryl K Eggers

    Department of Chemistry, San Jose State University, California, United States
    Contribution
    Conceptualization, Investigation, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  22. Noelle L'Etoile

    Department of Cell & Tissue Biology, University of California San Francisco, San Francisco, United States
    Contribution
    Conceptualization, Funding acquisition, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  23. Laura C Miller Conrad

    Department of Chemistry, San Jose State University, California, United States
    Contribution
    Conceptualization, Supervision, Investigation, Methodology, Writing—original draft, Writing—review and editing
    For correspondence
    laura.miller.conrad@sjsu.edu
    Competing interests
    No competing interests declared
  24. Miri K VanHoven

    Department of Biological Sciences, San Jose State University, California, United States
    Contribution
    Conceptualization, Supervision, Funding acquisition, Investigation, Visualization, Methodology, Writing—original draft, Project administration, Writing—review and editing
    For correspondence
    miri.vanhoven@sjsu.edu
    Competing interests
    No competing interests declared
    ORCID icon 0000-0001-7714-1555

Funding

National Science Foundation (HRD-1302873)

  • Angelina Tang
  • Sarah Y Matthews

NIH MARC Fellowship (5T34GM008253)

  • Christopher Wellbrook

National Institutes of Health (NS087544)

  • Noelle L'Etoile
  • Miri K VanHoven

National Institutes of Health (5SC3GM118199)

  • Laura C Miller Conrad

National Science Foundation (1355202)

  • Miri K VanHoven

National Institutes of Health (GM089595)

  • Miri K VanHoven

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We would especially like to thank M Fischbach for the use of equipment and facilities; H Gemrich and S Kato for microfluidic devices; J Engebrecht, L Rose and A Deshong for help with RNAi clones; J Ahringer for generation of RNAi strains; V Zavala for bacterial strain preparation; A Farooqi, J Griffin, F Farah, A Madrigal, S Pollock and M. Aguilar for assistance with experiments; S Bros, S Chalasani, E Glater, L Parr, S Rech, and E Skovran for comments and advice on the manuscript; the CGC and ARS Culture Collection for strains; Wormbase; and the National Institutes of Health (NIH) (NS087544 to MV and NL, GM089595 to MV and NS087544 to NL, 5SC3GM118199 to LMC, and 5T34GM008253 MARC fellowship to CW), the National Science Foundation (NSF) (1355202 to MV), and CSU-LSAMP (HRD-1302873 CSU-LSAMP fellowships to A. Tran and S. Matthews) for funding this work. CSU-LSAMP is funded through the National Science Foundation (NSF) under grant #HRD-1302873 and the Chancellor's Office of the California State University. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation or the Chancellor's Office of the CSU.

Reviewing Editor

  1. Oliver Hobert, Reviewing Editor, Howard Hughes Medical Institute, Columbia University, United States

Publication history

  1. Received: December 8, 2016
  2. Accepted: August 21, 2017
  3. Version of Record published: September 5, 2017 (version 1)

Copyright

© 2017, Tran et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 517
    Page views
  • 68
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: PubMed Central, Scopus, Crossref.

Comments

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Cell Biology
    2. Genes and Chromosomes
    Wahid A Mulla et al.
    Research Article Updated