miRNA-1175 downregulates a long non-coding natural antisense RNA and promotes long term memory

Abstract

Single-trial induced long-term memories (LTMs) require fast removal of inhibitory memory constraints following learning. We propose that the interplay between short non-coding RNAs (sncRNAs) and long non-coding RNAs (lncRNAs) is engaged in this process, suggesting the existence of a new and unexpected pathway for LTM formation. In the mollusc Lymnaea, this pathway involves the functional interaction between a typical sncRNA, Lym-miR-1175, and a lncRNA, Lym-NOS1AS. In the current study, we validate this hypothesis and provide evidence that: (1) LTM formation in Lymnaea is associated with the timed and targeted changes in the expression of Lym-miR-1175; (2) Lym-miR-1175 is a negative regulator of the Lym-NOS1AS; (3) Lym-miR-1175 is required for single-trial induced LTM; (4) Lym-miR-1175 and Lym-NOS1AS are co-expressed in a key neuron of the Lymnaea memory network. Thus, our data indicate an important role of the interaction between a miRNA and a lncRNA in one-trial learning.

Keywords: Brain; Long non-coding RNAs; Long-term memory; MiRNA; Natural antisense transcripts; Nitric oxide synthase.

Fig. 1

A LTM-related Lym-miRNA-1175 exhibits specific training-induced changes in its expression. (a) Schematic representation of the experiment to identify miRNAs involved in single-trial reward conditioning. (b) A schematic diagram showing that Lym-miR-1175 (red) has a putative target site (grey) located within the Lym-NOS1AS (diagram not drawn to scale). (c) Schematic representation of the experiment conducted to investigate whether single-trial reward conditioning is associated with specific changes in Lym-miR-1175 expression. A single pairing of amyl acetate, a conditioned stimulus (CS) and sucrose, an unconditioned stimulus (US) was employed to produce LTM. (d, e) Results of quantitative RT-PCR performed on the cerebral and buccal ganglia, respectively. The relative levels of Lym-miRNA-1175 expression in ganglia from unpaired control (UP, white bars, the CS and the US were separated by an interval of 1 h) and conditioned (paired, P, dark grey bars) animals dissected at 1 h, 2 h, 4 h and 6 h after training were calculated as 2^−ΔΔCt. All data in this figure are shown as means ± SEM (n = 20 in each group, collected in 5 tubes each containing 4 either BG or CG). Asterisk indicates statistically significant difference (p = 0.036, t = 2.85, df = 4.87, unpaired two-tailed t-test with Welch’s correction) between the ‘conditioned’ and ‘unpaired’ cerebral ganglia at 1 h after training. See also Supplementary Table S1.

Fig. 2

Lym-miR-1175 downregulates Lym-NOS1AS NAT. (a) Schematic diagram of the experiments showing that overexpression of miR-1175 in the stably transfected HEK293 GFP/NOS1AS cell line results in the decreased level of the GFP/Lym-NOS1AS transgene expression. Hatched box shows the putative binding site within the Lym-NOS1AS. It is important to note that the seed region of the Lym-miR-1175 exhibits 100% complementarity to the target site located within the Lym-NOS1AS (the non-Watson–Crick G–U base pairs are shown by dots). Asterisk indicates significant difference between the control (white bar, NCmim) and the experimental (grey bar, 1175mim) samples (Mann-Whitney U Test, z-score 2.50672, p = 0.012). (b) Schematic diagram of the experiments showing that overexpression of miR-1175 in stably transfected HEK293 GFP/NOS1ASmut cell line does not change the level of the GFP/Lym-NOS1ASmut transgene expression. Nucleotide substitutions in the Lym-NOS1ASmut, which lower the level of complementarity to the Lym-miR-1175, are shown by lowercase letters. The Mann-Whitney U Test reveals no significant difference between the control (white bar, NCmim) and the experimental (grey bar, 1175mim) samples. Of note, the whole procedure was run 5 times to enable statistics and all data in this figure are shown as means±SEM. See also Supplementary Table S2. The approximate positions of primers used in RT-PCR experiments are shown by half arrows.

Fig. 3

The role of Lym-miR-1175 in LTM. (a) Injection of snails with Invivofectamine does not impair LTM formation. Snails injected with Invivofectamine were trained and then tested for feeding response to the amyl acetate (CS) at 24 h after training. One-way ANOVA with post-hoc Tukey HSD test revealed no significant changes between the mean feeding responses of the Control Conditioned snails (light grey bar, CC, n = 16) and the conditioned animals injected with InvivoFectamine (dark grey bar, IF, n = 22, p = 0.11). Note that the feeding response of Naive animals (white bar, N, n = 17) is significantly lower than the response of the CC group (n = 16, p < 0.00001) and the IF group (n = 22, p < 0.001). (b) Blocking of Lym-miR-1175 impairs LTM after single-trial conditioning in vivo. All snails were tested for feeding response to the CS at 24 h after training. To test if the Lym-miR-1175 inhibitor down regulates LTM formation, we used the unpaired two-sample one-tailed t-test. It showed that the mean feeding response of the animals injected with the Lym-miR-1175 Inhibitor (red bar, INH) is significantly lower than the response of the snails injected with the Negative Control Inhibitor (black bar, NCI, n = 20, p = 0.035, t = 1.86). All data in these figures are shown as means ± SEM. See also Supplementary Table S3.

Fig. 4

Co-expression of Lym-miR-1175, Lym-NOS1AS NAT and Lym-NOS1 mRNA in the cerebral giant cell is in favour of the existence of a new pathway involved in LTM formation. (a) Lym-miR-1175, Lym-NOS1AS NAT and Lym-NOS1 mRNA are co-expressed in the cerebral giant cell (CGC). (ai) The results of RT-PCRs conducted on RNA extracted from isolated CGCs indicate the presence of Lym-NOS1AS NAT and Lym-miR-1175. Of note, PCR products of the expected sizes were detected using standard DNA electrophoresis in a mini gel, and their identity was confirmed by cloning and sequencing. The ‘RT−’ lanes represent the outcome of the control experiments in which reverse transcriptase was omitted. See also Supplementary Fig. 1. (aii) In situ hybridization reveals the presence of Lym-NOS1 mRNA in the cytoplasm of the CGC. (b) Schematic diagram shows the proposed role of the Lym-miR-1175/Lym-NOS1AS/Lym-NOS1 pathway in single-trial induced LTM.