Using Caenorhabditis elegans as a
Toxicology Model for Novel Antibiotic Compounds
by Janelle Sangalang, Alexa Garcia, Jessica Sangalang, Jennifer Kerr
Notre Dame of Maryland University
4701 N Charles Street
Baltimore, MD 21210
Abstract: As more microorganisms become resistant to known antibiotics, the need for new antimicrobial agents becomes more urgent. Efficiently identifying potential new antibiotic candidates remains difficult. This study aimed to assess the toxicity of novel antibiotic compounds extracted from soil bacterial isolates against a mutant strain of Caenorhabditis elegans: DC19 [bus-5(br19)]X. The bus-5 mutants have fragile cuticles that allow for increased permeability of chemical agents. Novel antibiotics were methanol-extracted from bacterial lawns and reconstituted in methanol.
An in vivo toxicity assay of four crude antibiotic extracts was conducted against bus-5 C. elegans. Toxicity was measured by counting the number of dead worms following exposure to the antibiotic extracts (N2#4, N1#5, N2#5, or D1#4), methanol alone, or water at days 0, 3, and 7. Changes in locomotion were monitored on day 7 as the number of tail flicks/minute (TF/min). Worm survival was significantly reduced for the N2#4 antibiotic compound (50%) on day three. Locomotion differences (8+/-4.2 TF/min) were found for compound N2#5 when comparing it to methanol (solvent) alone (18.9 +/-6.4 TF/min). C. elegans cuticle mutants (bus-5) were successfully used to screen for toxicity against four novel antibiotic extracts, presenting a cost-effective strategy to screen new antibiotics for initial non-mammalian animal toxicity.
Keywords: Caenorhabditis elegans; model host; toxicity; predictive toxicology; antibiotic screening
Acknowledgements: I would like to thank the Notre Dame of Maryland University’s Committee for Faculty Research and Development for funding this research, Marrisia Moore and Pat Bell, our lab managers, for providing the necessary equipment and resources to make this research possible, and Alexa Garcia and Jessica Sangalang for assisting with this research. I’d especially like to thank Dr. Jennifer Kerr for guiding me in this endeavor and for being my mentor
Several studies have strongly suggested the promise of Caenorhabditis elegans, a small, transparent nematode, as a versatile model system for toxicity. They have previously shown to be predictive of rat and mouse LD50 (the lethal dose concentration that results in killing 50% of the sample population) ranking and have displayed conservation of mode of toxic action between these nematodes and mammals (Hunt 2016). One study screened compounds for anti-infective properties by testing them against C. elegans infected with a human bacterial pathogen: Enterococcus faecalis. They were able to identify 28 novel compounds and natural product extracts with anti-infective properties leading to worm survival during the bacterial infection (Moy et al. 2009). Yet another study looked at the impact of ethanol exposure in C. elegans by monitoring locomotion defects (Davies et al. 2015). C. elegans are microscopic roundworms that feed on soil bacteria and fungi. This host is approximately 1 mm in length as an adult, has a simple anatomy, and a fully sequenced genome (~100Mb). In addition, maintaining C. elegans is easy and cost-effective (Hunt 2016; Kong et al. 2016). When the worms develop into adults, they can become males or self-fertile hermaphrodites (Figure 1). Each hermaphrodite is typically capable of producing 300 genetically identical offspring. Since the worms can be rapidly generated, screening can be conducted on large scales with a cost effective budget (Hunt 2016).
C. elegans have a transparent cuticle that allows internal structures to be seen without needing dissection or dyes (Hunt 2016). Up until now, one key limitation of nematode use in toxicity testing is that this cuticle is tough and made up of complex layers of collagen, which prevents the efficient uptake of small molecules, such as antibiotics (Johnstone 2000). Key work by Xiong et al. (2017) compared various cuticle mutants (agmo-1, bus-5, bus-8, bus-16 and bus-17) of C. elegans to determine their cuticle sensitivity, permeability, and overall fitness affect. The bus-5 mutant was identified to be the most suitable strain for general chemical toxicity exposure due to its permeable cuticle with minimal fitness consequences, except for slightly skiddy movement (Xiong et al. 2017). The mutant strain DC19 [bus-5(br19)]X has a severe missense mutation and was shown to be sensitive to certain chemicals, including boric acid (Figure 2), due to the presence of a fragile cuticle (Xiong et al. 2017). The Bus (bacterially unswollen) mutants were originally shown to be resistant to abdominal swelling during infection with the nematode pathogen Microbacterium nematophilumwere (Gravato-Nobre et al. 2005). The bus-5 gene encodes an enzyme that is required for polysaccharide, dTDP-rhamnose production, which is thought to be an important cuticle component (Feng et al. 2016). Based on this work, we decided to use this C. elegans strain, DC19 [bus-5(br19)]X for our antibiotic toxicity tests.
FIGURE 1. C. elegans anatomy. Major anatomical features of a hermaphrodite (A) and male (B) viewed laterally. (A) The dorsal nerve cord (DNC) and ventral nerve cord (VNC) run along the entire length of the animal from the nerve ring. Body wall muscles are shown for two of the four quadrants. (B) The nervous system and muscles are omitted in this view so that the pharynx and intestine are clearly visible. From: (Corsi et al. 2005).
FIGURE 2. Comparison of chemical sensitivity of wild type (N2) and cuticle mutant bus-5(br19) to boric acid. Mutant C. elegans bus-5(dc19) displays rapid death and disintegration in comparison to the parent N2 strain. Scale bar is 1mm. Image modified from (Xiong et al. 2017) to remove boric acid concentrations 8.4, 9.6, and 12.
As more bacteria become resistant to known antibiotics, the need for new antimicrobial agents becomes more urgent (Kupferschmidt 2016). An answer to this issue could lie among the dirt. Various soil microorganisms have been found to produce antibiotics; in fact, most of the antibiotics that are used and prescribed today originated from bacteria that live in soil (Kupferschmidt 2016). Genetic plasticity allows bacteria to respond to environmental threats, such as competition from neighbouring bacteria in the soil (Habboush 2019). Bacteria can evolve quickly through spontaneous mutations in their DNA or by acquiring new sets of genes through horizontal transfer from another species (Kupferschmidt 2016), potentially leading to new or alternative antibiotic pathways. An ongoing issue in trying to bring new antibiotics to market is the high cost and prolonged timeline of development from initial identification of a promising antibacterial compound to the ‘ready to prescribe’ stage (Towse 2017). Determining the initial toxicity of candidate antibiotics is an important step in this process that is not yet well defined in a streamlined system. This study aims to demonstrate a low cost, time efficient non-mammalian, small animal model strategy to test the toxicity (by monitoring mortality and locomotion defects) of novel antibiotic compounds using the cuticle mutant strain of C. elegans: bus-5 (br19).
Material and Methods
Strains and Media The Caenorhabditis elegans strain DC19 [bus-5(br19)]X was obtained from the C. elegans Genetics Center (CGC, University of Minnesota, USA). C. elegans strains were cultivated as described (Stiernagle 2006) and maintained at 23 °C on nematode growth media (NGM) agar. The bacterial strain Escherichia coli (OP50) was used to feed the C. elegans populations except during starvation conditions when no food was provided. Bacterial strains Enterobacter aerogenes 51697 or Escherichia coli 11775 were used to screen soil isolates for antibacterial activity. Bacterial soil isolates were maintained on 50% nutrient level trypticase soy agar (50% TSA; BD, Difco) or trypticase soy broth (TSB; BD Difco). Soil dilution plates were supplemented with the anti-fungal cycloheximide (50% TSA-CHX; 100 μg/mL). Isolation and Screening of Antibiotic Producing Bacteria Soil samples were collected from Baltimore, Maryland, USA (39.349995, -76.619772) and (39.350135, -76.620433). One gram of soil was serially diluted in 1x phosphate buffered saline (PBS) and 100 microliters of each dilution was spread plated on 50%TSA-CHX plates from 10-1 to 10-6 dilutions. Plates were incubated at room temperature for 1 week. Unique colony morphologies were patch plated using sterile toothpicks onto 50% TSA plates, yielding a bacterial soil grid ‘library.’ Grid libraries were incubated overnight at room temperature. Bacterial lawns were created on 50%TSA plates using sterile swabs with an overnight 5 mL TSB culture of E. coli 11775 or E. aerogenes 51697. Each grid plate was replica plated onto the lawns and incubated at room temperature for 3 days. Zones of inhibition were recorded (Figure 5). Soil isolates that were initially recorded with a positive zone of inhibition were streaked for isolation and were retested using the same method to confirm a clear zone of inhibition. Four antibiotic producing soil isolates were then selected for antibiotic extraction (N2#4, N1#5, N2#5, or D1#4).
Crude Antibiotic Extraction Two bacterial lawns were created on 50% TSA plates using sterile swabs and an isolated colony of N2#4, N1#5, N2#5, or D1#4. Lawn plates were incubated overnight at room temperature. Afterwards, plates were frozen for 1 hour in a -80oC freezer. Plates were thawed at room temperature, the agar was chopped using a sterile scoopula, and placed into a sterile 250 mL glass jars with 30 mL of methanol. The bottles were capped and left in a shaker (90 rpm) overnight at room temperature. Methanol containing the crude extracts was transferred to 25 mL glass bottles that were individually weighed using an analytical balance. The bottles were reweighed following solvent evaporation (~48 hours) in a chemical fume hood at room temperature. Crude antibiotic compounds were reconstituted in methanol based on weight to a concentration of ~1 μg/µL. Positive antibiotic activity of the extracts were confirmed using the zone of inhibition lawn assay with 5 µL of the extract.
Plate based toxicity assay The bus5(br19) C. elegans cuticle mutants were synchronized, as previously described (Xiong et al. 2017). Briefly, strains were washed from plates with M9 buffer, residual bacteria were removed by washing worms two times with M9 buffer prior to alkaline hypochlorite solution exposure for 4 mins. The bleach solution was removed by washing the eggs three times with M9 buffer. Eggs were left to hatch overnight in 24 well plates with NGM agar and resulted in a population of L1 larvae. L1 larvae were starved for two days. One hundred microliters of M9 buffer was added to individual wells with starved L1 larvae and a 20 μL worm suspension, containing approximately 5-10 worms/well were transferred to a new to a 24-well plate containing 500 μL NGM agar (Figure 3). Thirty microliters of crude antibiotic extracts (N2#4, N1#5, N2#5, or D1#4; concentration ~1 mg/mL) solvent alone control (methanol), or negative control (sterile deionized water) was added to an individual well with 4 replicates per condition. Plates were incubated at 23°C and worm deaths were observed and recorded using a dissection microscope 10x total magnification at days 0, 3, and 7. Also, on day 7, a two-minute video of each well was captured using an Apple iPhone 4 camera phone mounted to the objective. Video was reviewed to determine tail flicks per minute (TF/min; locomotion assay) for 5 randomly selected worms per well for all 24 wells. The toxicity experimental timeline is shown in Figure 4.
FIGURE 3. Plate based toxicity assay. Nematode growth media agar in 24 well plate. Each well contained a 20 µL worm suspension M9 buffer (~5-10 worms/well) and 30 μL of a crude antibiotic extract, solvent alone (methanol) alone, or control (sterile water).
FIGURE 4. Toxicity assay experimental timeline. Worms treatment consisted of exposure to one of four crude antibiotic extracts, methanol only, or sterile water (control).
Bacterial soil isolate patch plates were screened for antibiotic activity by identifying patches that created a clear zone of inhibition (Figure 5) around the soil bacterium patch. The antibiotics produced by soil isolated N2#4, N2#5, and D1#4 were found to be effective against both gram negative bacteria E. coli and E. aerogenes. Isolate N1#5 produced a zone of inhibition against E. aerogenes (Table 1).
FIGURE 5. Bacterial soil isolate patch plates showing antibiotic activity. Representative bacterial lawn with bacterial soil isolates being screened for antibiotics activity. Zone of inhibition = antibiotics activity.
TABLE 1. Antibiotic activity of the crude extracts from 4 bacterial soil isolates. +: zone of inhibition present; -: no zone of inhibition created on bacterial lawn species Enterobacter aerogenes 51697 or Escherichia coli 11775.
Following methanol extraction, worm toxicity was tested using 7-day worm survival in response to antibiotic extract or control (methanol alone or sterile water) exposure (Figure 6). All conditions, except one, showed 100% nematode survival by day 3. Isolate N2#4 had a decrease in survival leaving 50% of the population alive at day 3 with a statistically different survival curve compared to the control methanol, solvent alone (log-rank test; z = 5.32, p < 0.001). By day 7, every condition displayed a decrease in worm population compared to when the compounds were introduced at day 0. The survival percentage for N2#4 on day 7 remained the same as day 3 at 50%. N2#5 notably decreased to only have approximately 36% of the population alive but was not found to be statistically different from the controls.
FIGURE 6. Survival curves for nematodes with and without antibiotic treatment. Results of killing assays of C. elegans strain bus-5 (dc19). Novel antibiotic compounds were introduced on day 0. Thirty microliters of a single novel antibiotic crude extract was added to 4 replicate wells for each antibiotic, 30 μL of sterile water was added to 4 replicate wells, and 30 μL of solvent only (methanol) was added to 4 replicate wells. Survivor percentages are shown per average of worms (n>30) per condition. Standard deviation shown and * indicates a significant difference from solvent alone, log-rank test; z = 5.32, p < 0.001.
In addition to observing worm survival, toxicity was assessed by observing behavioral defects in motility. This allowed for the capture of non-lethal toxicity effects. Figure 7 illustrates the nematode locomotion rates after 7 days post treatment. Locomotion rates for worms treated with novel antibiotics trended lower than those that were exposed to the controls (water or methanol). N2#5 had a significant decrease in locomotion rate at 8 TF/min compared to controls.
FIGURE 7. Locomotion rates for nematodes treated with antibiotic extracts. Results of average (n=20/condition) locomotion rate (tail flicks/min on day 7) of C.elegans strain bus5. Water only (no treatment) and solvent methanol only were used as controls. ** significantly reduced (p<0.005) the locomotion of bus 5 C. elegans exposed to antibiotics versus solvent alone via t-test. Standard deviation shown.
The toxicity levels of novel antibiotic compounds extracted from soil bacterial isolates were successfully monitored using worm mortality and locomotion defects in a mutant strain of Caenorhabditis elegans: bus-5 (br19) X. The methanol (solvent) alone had no effect on worm survival and locomotion compared to the negative control, sterile water. This indicates any decrease in mortality or locomotion in the experimental treatment groups was likely due to the crude antibiotic toxicity and not the solvent. N2#4 was the only antibiotic extract with a significant difference in its worm survival curve relative to the controls and, therefore, it likely had a toxic effect. The novel antibiotics generally showed decreased nematode survival percentages between 36-55% on day 7 but were not significantly different relative to the controls. These data correlated with a previous study that examined the toxicity of various mycotoxins (antifungals) against C. elegans (Keller 2018). Their data, based on single concentrations of each antifungal in the parent C. elegans strain N2, showed that most of the antifungals yielded 50% nematode survival for each compound by day 7 (Keller 2018). To date, there are no studies that have used the bus-5(dc19) cuticle mutants for antibiotic toxicity. Many toxicity studies have focused on testing metal toxicity (Hunt 2017; Jiang et al. 2016) or anti-infective capabilities (Peterson and Pukkila-Worley 2018). So, this work provides proof of concept for use of these cuticle mutants for antibiotic toxicity testing.
To future leverage the toxicity model, locomotion defects were observed and may suggest that unwanted non-lethal side effects of the novel antibiotics may occur. Sterile water (19.8 +/-6.1 TF/min) and methanol alone (18.9 +/-6.4 TF/min) were used as controls when comparing locomotion to the antibiotic extracts. The novel antibiotics generally decreased nematode locomotion rates when comparing it to the control solvents. However, only N2#5 has significant locomotion differences (8+/-4.2 TF/min), suggesting that this may create unwanted side effects if used as a drug.
Future studies would explore changing the concentrations of the antibiotics, monitoring locomotion at more time points, and using mass spectrometers to potentially identify the antibacterial compounds present in the crude extracts. Other mutant strains of C. elegans could also be tested on to see if similar results will prevail in different genetic backgrounds. This research reinforces the hypothesis that C. elegans can be effectively used to screen for antibiotic toxicity and takes a closer look at the effects of novel antibiotics on bus-5 mutants. Together these data, coupled with the novel antibiotics’ general efficacy in vitro against gram negative bacteria, suggest that compounds from soil isolates N1#5 or D1#4 could be viable for further exploration as novel antibiotics.
The initial in vitro data from these experiments suggested that the extracted antibiotic compounds were good candidates for further exploration as novel antibiotics. Cuticle C. elegans’ mutant strain bus5(dc19) was successfully used to screen for toxicity via worm survival against four antibiotic crude extracts, resulting in identification of one potentially toxic extract from soil isolate N2#4. A locomotion assay further identified a second extract, from soil isolate N2#5, leading to motility defects. Collectively, this work supports worm survival and locomotion monitoring in the cuticle mutant bus5(br19) as a cost effective initial toxicity strategy to screen in vivo for new antibiotic compounds.
Corsi, A. K., Wightman, B., & Chalfie, M. (2015). A Transparent Window into Biology: A Primer on Caenorhabditis elegans. Genetics, 200(2), 387–407. https://doi.org/10.1534/genetics.115.176099 Kong, C., Eng, S. A., Lim, M. P., & Nathan, S. 2016. Beyond Traditional Antimicrobials: A Caenorhabditis elegans Model for Discovery of Novel Anti-infectives. Frontiers in microbiology, 7, 1956. https://doi.org/10.3389/fmicb.2016.01956 Davies, A. G., Blackwell, G. G., Raabe, R. C., & Bettinger, J. C. 2015. An Assay for Measuring the Effects of Ethanol on the Locomotion Speed of Caenorhabditis elegans. Journal of visualized experiments: JoVE, (98), 52681. https://doi.org/10.3791/52681 Feng, L., Shou, Q., & Butcher, R. A. (2016). Identification of a dTDP-rhamnose biosynthetic pathway that oscillates with the molting cycle in Caenorhabditis elegans. The Biochemical journal, 473(11), 1507–1521. https://doi.org/10.1042/BCJ201601 Gravato-Nobre, M. J., Nicholas, H. R., Nijland, R., O'Rourke, D., Whittington, D. E., Yook, K. J., & Hodgkin, J. (2005). Multiple genes affect sensitivity of Caenorhabditis elegans to the bacterial pathogen Microbacterium nematophilum. Genetics, 171(3), 1033–1045. https://doi.org/10.1534/genetics.105.045716
Habboush Y, Guzman N. Antibiotic Resistance. [Updated 2020 Feb 10]. In: StatPearls [Internet][cited 2020 July 25]. Treasure Island (FL): StatPearls Publishing; 2020 Jan. Available from: https://www.ncbi.nlm.nih.gov/books/NBK513277/ Hunt P. R. (2017). The C. elegans model in toxicity testing. Journal of applied toxicology: JAT, 37(1), 50–59. https://doi.org/10.1002/jat.3357 Jiang, Y., Chen, J., Wu, Y., Wang, Q., & Li, H. (2016). Sublethal Toxicity Endpoints of Heavy Metals to the Nematode Caenorhabditis elegans. PloS one, 11(1), e0148014. https://doi.org/10.1371/journal.pone.0148014 Johnstone I. L. (2000). Cuticle collagen genes. Expression in Caenorhabditis elegans. Trends in genetics : TIG, 16(1), 21–27. https://doi.org/10.1016/s0168-9525(99)01857-0 Keller, J., Borzekowski, A., Haase, H., Menzel, R., Rueß, L., & Koch, M. (2018). Toxicity Assay for Citrinin, Zearalenone and Zearalenone-14-Sulfate Using the Nematode Caenorhabditis elegans as Model Organism. Toxins, 10(7), 284. https://doi.org/10.3390/toxins10070284 Kupferschmidt K. (2016). Resistance fighters. Science (New York, N.Y.), 352(6287), 758–761. https://doi.org/10.1126/science.352.6287.758.
Moy, T. I., Conery, A. L., Larkins-Ford, J., Wu, G., Mazitschek, R., Casadei, G., Lewis, K., Carpenter, A. E., & Ausubel, F. M. (2009). High-throughput screen for novel antimicrobials using a whole animal infection model. ACS chemical biology, 4(7), 527–533. https://doi.org/10.1021/cb900084v
Peterson, N. D., & Pukkila-Worley, R. (2018). Caenorhabditis elegans in high-throughput screens for anti-infective compounds. Current opinion in immunology, 54, 59–65. https://doi.org/10.1016/j.coi.2018.06.003
Stiernagle T. (2006). Maintenance of C. elegans. WormBook: the online review of C. elegans biology, 1–11. https://doi.org/10.1895/wormbook.1.101.1
Towse, A., Hoyle, C. K., Goodall, J., Hirsch, M., Mestre-Ferrandiz, J., & Rex, J. H. (2017). Time for a change in how new antibiotics are reimbursed: Development of an insurance framework for funding new antibiotics based on a policy of risk mitigation. Health policy (Amsterdam, Netherlands), 121(10), 1025–1030. https://doi.org/10.1016/j.healthpol.2017.07.011
Wright G. D. (2016). Antibiotic Adjuvants: Rescuing Antibiotics from Resistance. Trends in microbiology, 24(11), 862–871. https://doi.org/10.1016/j.tim.2016.06.009
Xiong, H., Pears, C., & Woollard, A. (2017). An enhanced C. elegans based platform for toxicity assessment. Scientific reports, 7(1), 9839. https://doi.org/10.1038/s41598-017-10454-3