The first mating experience in males decreased the latency to mate
To investigate how the first mating experience in males could influence mating behavior in medaka fish, we performed mating tests using naïve males and experienced males who had mated with females more than 7 times. To prepare naïve males, we separated juvenile males into groups and bred them without any females until performing the mating tests. We also prepared sexually mature females (> 3 months after hatching) for the mating test. The mating tests were carried out using 11 fixed pairs for 7 days (7 times) (Fig. 1A, 1B). The latency to mate was defined as the interval from “releasing the male” to “crossing with spawning”. To evaluate the effect of the first mating experience on the latency to mate, we used a GLMM, which revealed that the latency of naïve males significantly decreased after the first mating experience and biased the distribution for 5 days, but not for 7 days (Fig. 1C, Table S1, GLMM; day1 vs day3 estimate ± s.e. = 0.5094 ± 0.192, z.ratio = 2.646, P = 0.0406, day1 vs day5 estimate ± s.e. = 0.8494 ± 0.189, z.ratio = 4.488, P = < 0.0001, day1 vs day7 estimate ± s.e. = 0.4231 ± 0.219, z.ratio = 1.934, P = 0.2137, day3 vs day5 estimate ± s.e. = 0.3401 ± 0.191, z.ratio = 1.781, P = 0.2825, day3 vs day7 estimate ± s.e. = − 0.0862 ± 0.217, z.ratio = 0.398, P = 0.978, day5 vs day7 estimate ± s.e. = − 0.4263 ± 0.218, z.ratio = − 1.951, P = 0.2067). In experienced males, on the other hand, the mating experience had no significant effect on the latency to mate (Fig. 1C, Table S1). Together, these findings indicated that the male mating experience influenced medaka mating behavior only in naïve males. We confirmed that these tendencies were reproduced in experiments performed at a different laboratory (Fig. S1, Table S2, S3). Therefore, we concluded that the first mating experience in naïve males decreased the latency to mate.
The first mating experience shortened the latency to a first courtship display
We further examined which behavioral component could influence the latency to mate in naïve males in repeated mating tests. First, we counted the number of courtship displays and measured the latency to the first courtship display, revealing that the latency to the first courtship display was significantly decreased only in naïve males, and not in experienced males (Fig. 2A, Tables S1, GLMM; day1 vs day3 estimate ± s.e. = 0.0973 ± 0.189, z.ratio = 0.515, P = 0.9555, day1 vs day5 estimate ± s.e. = 0.6015 ± 0.186, z.ratio = 3.229, P = 0.0068, day1 vs day7 estimate ± s.e. = 0.7165 ± 0.201, z.ratio = 3.562, P = 0.0021, day3 vs day5 estimate ± s.e. = 0.5043 ± 0.171, z.ratio = 2.951, P = 0.0167, day3 vs day7 estimate ± s.e. = 0.6192 ± 0.198, z.ratio = 3.135, P = 0.0093, day5 vs day7 estimate ± s.e. = 0.115 ± 0.196, z.ratio = 0.586, P = 0.963). On the other hand, the number of courtship displays did not significantly change in either naïve or experienced males (Fig. 2B, Table S1). Next, we compared the latency to mate from the first courtship display, which negatively correlated with the degree of female receptivity12–14. The latency was significantly decreased only in naïve males (Fig. 2C, Table 1, GLMM; day1 vs day3 estimate ± s.e. = 1.4486 ± 0.331, z.ratio = 4.379, P = 0.0001, day1 vs day5 estimate ± s.e. = 1.5358 ± 0.346, z.ratio = 4.44, P = 0.0001, day1 vs day7 estimate ± s.e. = 0.3013 ± 0.4, z.ratio = 0.754, P = 0.875, day3 vs day5 estimate ± s.e. = 0.872 ± 0.326, z.ratio = 0.268, P = 0.9933, day3 vs day7 estimate ± s.e. = − 1.1472 ± 0.377, z.ratio = 3.041, P = 0.0126, day5 vs day7 estimate ± s.e. = − 1.2344 ± 0.385, z.ratio = − 3.207, P = 0.0073), and strongly suggested that the female mating experience with the naïve males enhanced female receptivity. Furthermore, we analyzed other behavioral elements (wrappings and wrapping rejections), and found no significant change between the number of the 2 behavioral elements and the number of mating in either naïve or experienced males (Fig. S3, Table S1). Next, we analyzed behavioral transitions in the mating test, and found that the number of courtship displays significantly decreased between day1 vs day3 and day1 vs day5 in experienced males (Fig. S5, Tables S1, GLMM; day1 vs day3 estimate ± s.e. = 0.636 ± 0.234, z.ratio = 2.717, P = 0.0333, day1 vs day5 estimate ± s.e. = 0.820 ± 0.251, z.ratio = 3.26, P = 0.0061, day1 vs day7 estimate ± s.e. = 0.42 ± 0.223, z.ratio = 1.881, P = 0.2361, day3 vs day5 estimate ± s.e. = 0.184 ± 0.283, z.ratio = 0.649, P = 0.9158, day3 vs day7 estimate ± s.e. = − 0.216 ± 0.258, z.ratio = − 0.838, P = 0.8365, day5 vs day7 estimate ± s.e. = − 0.400 ± 0.273, z.ratio = − 1.464, P = 0.4592). In addition, we confirmed that these tendencies were reproduced in experiments performed at a different laboratory (Fig. S4, S6, Tables S2, S3). Accordingly, we revealed that the first mating experience shortened the latency to the first courtship display only in naïve males.
Table 1
RNA-seq results. List of differentially expressed genes (DEGs) that were more highly expressed in post-naïve samples (2 matings) than in naïve samples. Genes are ordered from high to low expression using FPKM values as an index.
| gene | log2|FC| | FDR | FPKM | FPKM |
| (naïve) | (post-naïve) |
1 | fkbp5 | 1.89 | 1.010E-18 | 23.68 | 87.57 |
2 | tshba | 2.72 | 3.320E-39 | 6.99 | 45.93 |
3 | hapln2 | 1.12 | 4.760E-05 | 19.22 | 41.74 |
4 | LOC101155558 | 1.40 | 3.284E-03 | 11.44 | 30.26 |
5 | macrosialin | 1.04 | 1.990E-05 | 14.57 | 29.91 |
6 | stat1a | 1.13 | 5.290E-07 | 10.18 | 22.34 |
7 | lgals17 | 1.76 | 1.730E-06 | 4.97 | 16.87 |
8 | dio2 | 1.05 | 9.331E-04 | 7.30 | 15.10 |
9 | lgals3bp | 1.39 | 5.390E-07 | 5.77 | 15.10 |
10 | klf9 | 1.10 | 4.730E-08 | 6.24 | 13.36 |
The behavioral change in naïve males occurred only in fixed pairs
The first male mating experience mainly influenced the latency to mate and the latency to the first courtship display in fixed pairs in 7 continuous mating tests. Here we examined whether this effect was specific for fixed pairs. To compare mating behavior between fixed pairs and swapped pairs, we performed mating tests for 3 days (Fig. 3A, each n = 8). The latency to mate and the latency to the first courtship display were significantly decreased in the fixed group (GLMM; day1 vs day2 estimate ± s.e. = 0.205 ± 0.00818, z.ratio = 2.5, P = 0.0332, day1 vs day3 estimate ± s.e. = 0.345 ± 0.00779, z.ratio = 4.432, P < 0.0001, day2 vs day3 estimate ± s.e. = 0.141 ± 0.0785, z.ratio = 1.791, P = 0.1725), but did not change in swapped pairs (Fig. 3B, 3C, Table S4). The latency to mate from the first courtship display (Fig. 3D, Table S4), numbers of other events (Fig. S7, Table S4), and behavioral transitions (Fig. S8, Table S4) did not significantly change in either fixed or swapped pairs. The latency to mate and the latency to the first courtship display significantly decreased in repeated behavioral experiments (3 times). These findings strongly suggest that the naïve males recognized the first mating partner and thus the latency to mate and the latency to the first courtship display were decreased according to the mating experience.
The first mating experience changed gene expression in the male brain
To evaluate the effect of the first mating experience on the brain gene expression patterns, we compared gene expression profiles using whole brains between the naïve and post-naïve male medaka that had 2 mating experiences (Fig. S9). In the present study, we performed RNA-seq analysis using the whole brain with the pituitary, because in some fish species such as zebrafish and cichlid, social status has significant effects on gene expression at the whole brain level17,18. We identified 10 differentially expressed genes (DEGs) that were upregulated by the mating experience and had a greater than 2-fold change, suggesting that the first mating experience could influence brain gene expression in male medaka. We listed the DEGs in descending order of the expression level (FPKM) (Table 1). Interestingly, we found that 3 genes (tshba, dio2, klf9) of the top 10 DEGs were associated with functional expression of the thyroid hormone system (Table 1). tshba encodes a thyroid stimulating hormone that promotes the synthesis and secretion of inactivated thyroid hormone (T4). dio2 encodes type II iodothyronine deiodinase, which converts inactivated thyroid hormone to activated thyroid hormone (T3). klf9 encodes a transcription factor, Krüppel-like factor 9, which is induced by thyroid hormone (T3)19. These findings indicate that the first mating experience changed the gene expression patterns and upregulated thyroid hormone-related genes. In addition, early-life stress exposure increases fkbp5 expression in the brain and FKBP5 regulates glucocorticoid receptor activity20. High fkbp5 expression and early-life stress interact to increase anxiety-like behavior mediated by AKT signaling in association with hippocampal synaptic plasticity21. Hapln2 (also called Bral1) is essential for formation of the functional extracellular matrix and neuronal conductivity in mice22.