Aversive olfactory learning and associative long-term memory in Caenorhabditis elegans

Abstract

The nematode Caenorhabditis elegans (C. elegans) adult hermaphrodite has 302 invariant neurons and is suited for cellular and molecular studies on complex behaviors including learning and memory. Here, we have developed protocols for classical conditioning of worms with 1-propanol, as a conditioned stimulus (CS), and hydrochloride (HCl) (pH 4.0), as an unconditioned stimulus (US). Before the conditioning, worms were attracted to 1-propanol and avoided HCl in chemotaxis assay. In contrast, after massed or spaced training, worms were either not attracted at all to or repelled from 1-propanol on the assay plate. The memory after the spaced training was retained for 24 h, while the memory after the massed training was no longer observable within 3 h. Worms pretreated with transcription and translation inhibitors failed to form the memory by the spaced training, whereas the memory after the massed training was not significantly affected by the inhibitors and was sensitive to cold-shock anesthesia. Therefore, the memories after the spaced and massed trainings can be classified as long-term memory (LTM) and short-term/middle-term memory (STM/MTM), respectively. Consistently, like other organisms including Aplysia, Drosophila, and mice, C. elegans mutants defective in nmr-1 encoding an NMDA receptor subunit failed to form both LTM and STM/MTM, while mutations in crh-1 encoding the CREB transcription factor affected only the LTM.

Figure 1.

Chemotaxis assay and classical conditioning of C. elegans. (A) Schematic representation of chemotaxis assay of worms to 1-propanol, which was carried out on square agar plates as described in the Materials and Methods. Worms were allowed to move freely on the agar for 10 min at room temperature. Chemotaxis index (CI) values were calculated from the equation shown. (B) CI values of worms to 1-propanol after spaced or massed training with chemicals indicated. Flowcharts of the spaced and massed training protocols used are shown at top. LI values were calculated by using the equation shown. The CI value of reference worms (CI_reference) was the mean value of CI values of worms conditioned with HCl alone and 1-propanol alone. Bars are means ± SEM (n = 9 assays). Asterisks indicate statistically significant differences (**P <0.01) determined by one-way ANOVA with the Bonferroni/Dunn test, in comparison to the CI of naive worms. (C) LI values of worms repeatedly conditioned as indicated on the horizontal axis. At each cycle of the trials, worms were simultaneously conditioned with a solution containing 1.0% 1-propanol and 100 µM HCl (pH 4.0) with a 10-min ITI. Bars are means ± SEM (n = 9 assays). Asterisks indicate statistically significant differences (**P <0.01) determined by one-way ANOVA with Bonferroni/Dunn test, in comparison to the LI of worms trained with 10 conditioning cycles.

Figure 2.

Effects of ISI and ITI lengths on memory acquisition and retention. (A) Effects of ISI lengths on memory acquisition and retention. Flowcharts of backward (a), simultaneous (b), and forward (c) conditioning protocols used are shown at top. In the backward conditioning (a), worms were first stimulated with HCl as a US (dotted arrow), and then with 1-propanol as a CS (solid arrow) for 0 min, 0.5 min, or 2 min after the US stimulation. In the simultaneous conditioning (b), worms were soaked in a solution containing both 1-propanol and HCl. In the forward conditioning (c), worms were first stimulated with 1-propanol and then with HCl for 0 min, 0.5 min, and 2 min after the CS stimulation. These procedures were repeated five times with a 10-min ITI, and then worms were examined for their LI immediately (open bars) and 3 h (closed bars) after the completion of the repetitive conditionings. Data are means ± SEM (n = 10 assays). Asterisks indicate statistically significant differences (*P <0.05; **P < 0.01) determined by one-way ANOVA with Bonferroni/Dunn test, in comparison to the LI of worms simultaneously conditioned in b. (B) Effects of ITI lengths on memory acquisition and retention. A flowchart of the conditioning used is shown at top. Worms were simultaneously stimulated by being soaked in a solution containing both 1-propanol and HCl, followed by various ITI lengths ranging from 0 min through 30 min. These conditioning procedures were repeated five times, and then the worms were tested for LI values immediately (open bars) or 3 h (closed bars) after the completion of the repetitive conditionings. Data are means ± SEM (n = 9 assays). Asterisks indicate statistically significant differences (**P <0.01) determined by one-way ANOVA with Bonferroni/Dunn test, in comparison to the LI of worms conditioned without an ITI.

Figure 3.

Memory retention and extinction learning. (A) Memory retention induced by massed or spaced training. Flowcharts of the spaced and massed training protocols used are shown at top. In the spaced training, worms were simultaneously stimulated by being immersed in a solution containing 1-propanol and HCl. This procedure was repeated 10 times with a 10-min ITI, and the LI of the worms was assayed 0 h through 24 h (retention intervals) after the completion of the spaced training (solid line). In the massed training, worms were simultaneously stimulated with 1-propanol and HCl. After this conditioning was repeated 10 times without an ITI, the worms were assayed for LI 0 h, 1 h, and 3 h after the completion of the massed training (broken line). Data points are means ± SEM (n = 9–15 assays). Asterisks indicate statistically significant differences (**P <0.01) determined by one-way ANOVA with Turkey-Kramer's test, in comparison to the LI measured immediately after the trainings. (B) Extinction learning. After the spaced training 10 times simultaneously with 1-propanol and HCl described above in A, worms were transferred to NGM plates seeded with E. coli and were allowed to freely move and eat at 20°C for 6 h. The worms were then conditioned only with the CS (solid line) in the absence of the US as described in “Extinction” of Materials and Methods. This extinction training was repeated one to 10 times as indicated on the horizontal axis. Immediately after the extinction learning, worms were tested for LI. As a control (broken line), worms were also immersed in dH₂O, instead of 1-propanol, at room temperature. Data points are means ± SEM (n = 9 assays). Asterisks indicate statistically significant differences (*P < 0.05; **P <0.01) determined by two-sided Student's t-test, in comparison to LI values of worms after conditioning with dH₂O by the same cycle number.

Figure 4.

Propanol-specific associative learning. (A) Wild-type worms were conditioned 10 times simultaneously with 1-propanol and HCl by spaced (with a 10-min ITI) or massed training. Immediately after the training, worms were assayed for their ability of chemotaxis to 5% 1-propanol, 0.01% benzaldehyde, 0.001% isoamyl alcohol, or 0.001% diacetyl spotted on the edge of chemotaxis agar plates. Note that chemotactic behaviors of the trained worms to benzaldehyde, isoamyl alcohol, and diacetyl were not affected by the training. Data are means ± SEM (n = 9 assays). Asterisks indicate statistically significant differences (**P <0.01) determined by one-way ANOVA with Bonferroni/Dunn test, in comparison to the CI of naive worms (before training). (B) Associative learning of 1-propanol with HCl was specific for 1-propanol. LI values were calculated from the data shown in A. Data are means ± SEM (n = 9 assays). Asterisks indicate statistically significant differences (**P <0.01) determined by one-way ANOVA with Bonferroni/Dunn test, in comparison to the LI of worms assayed with 1-propanol as a stimulus.

Figure 5.

Effect of translation and transcription inhibitors on memory acquisition and retention. (A) A flowchart of spaced training used is shown at top. Worms were cultivated on an NGM plate spread with bacteria, which contained one of the indicated drugs for 2 h, and trained 10 times with a 10-min ITI as shown in the flowchart. During the ITI, worms were placed on an NGM plate with a bacterial lawn, which contains the indicated drug. The worms were tested for their LI by chemotaxis assay immediately (open bars) and 6 h (closed bars) after the completion of the spaced training. (B) A flowchart of massed training used is shown at top. Worms were cultivated for 4 h on an NGM plate spread with a bacterial lawn, which contained one of the indicated drugs, and trained 10 times without an ITI. The worms were tested for their LI by chemotaxis assay immediately (open bars) and 3 h (closed bars) after the completion of the massed training. Data are means ± SEM (n = 9 assays). Asterisks indicate statistically significant differences (**P <0.01) determined by one-way ANOVA with the Bonferroni/Dunn test, in comparison to the LI of worms untreated with drug.

Figure 6.

Sensitivity of memory to disruption. (A) A flowchart of spaced training used is shown at top. Worms were simultaneously stimulated with 1-propanol and HCl. Immediately after repeated conditioning 10 times with a 10-min ITI, the worms were soaked in either room-temperature (RT) dH₂O or ice-cold (IC) dH₂O, and then tested for LI values after being cultivated on NGM plates with a bacterial lawn at 20°C for 0 h (open bars) and 6 h (closed bars). (B) A flowchart of massed training used is shown at top. Worms were conditioned 10 times with a solution containing both 1-propanol and HCl, and then soaked in either room-temperature (RT) dH₂O or ice-cold (IC) dH₂O. Immediately (open bars) or 3 h (closed bars) after the treatment, the worms were tested for LI by chemotaxis assay. Data are means ± SEM (n = 9 assays). Asterisks indicate a statistically significant difference (**P <0.01) determined by one-way ANOVA with Bonferroni/Dunn test, in comparison to the LI of worms untreated.

Figure 7.

C. elegans mutants defective in learning and memory. (A) A massed training protocol of wild-type and mutant worms is shown at top. The worms were trained 10 times by being soaked in a solution containing both 1-propanol and HCl, and then tested for LI values by chemotaxis assay immediately (open bars) and 3 h (closed bars) after the training. (B) A spaced training protocol used for the worms indicated is shown at top. The worms were stimulated by being soaked in a solution containing both 1-propanol and HCl. This conditioning was repeated 10 times with a 10-min ITI. The worms were tested for LI values by chemotaxis assay immediately (open bars) and 6 h (closed bars) after the training. Data are means ± SEM (n = 9 assays). Asterisks indicate statistically significant differences (**P <0.01) determined by one-way ANOVA with the Bonferroni/Dunn test, in comparison to the LI of N2 worms.