The genetic basis of parental care evolution in monogamous mice

Abstract

Parental care is essential for the survival of mammals, yet the mechanisms underlying its evolution remain largely unknown. Here we show that two sister species of mice, Peromyscus polionotus and Peromyscus maniculatus, have large and heritable differences in parental behaviour. Using quantitative genetics, we identify 12 genomic regions that affect parental care, 8 of which have sex-specific effects, suggesting that parental care can evolve independently in males and females. Furthermore, some regions affect parental care broadly, whereas others affect specific behaviours, such as nest building. Of the genes linked to differences in nest-building behaviour, vasopressin is differentially expressed in the hypothalamus of the two species, with increased levels associated with less nest building. Using pharmacology in Peromyscus and chemogenetics in Mus, we show that vasopressin inhibits nest building but not other parental behaviours. Together, our results indicate that variation in an ancient neuropeptide contributes to interspecific differences in parental care.

Extended Data Figure 1. Parental behaviours in undisturbed home cages for three days following the birth of a litter

The fraction of time an animal was engaged in each behaviour averaged across 5-min samples for each hour, for 16 h of light and 8 h of dark parts of the day separately. Horizontal lines denote the mean. *, P<0.05; NS, not significant by Mann-Whitney U test.

Extended Data Figure 2. Parental behaviours towards own pups and heterospecific pups

The behaviour of parents was measured across four consecutive days, alternating the pup species each day (randomizing the pup that was given on day 1). Grey lines connect an individual’s behaviour. Blue and red lines denote the median for fathers and mothers, respectively. *, P<0.05; **, P<0.01; NS, not significant by paired t-test or Wilcoxon signed-rank test (nest quality).

Extended Data Figure 3. QTL mapping of parental behaviours

Non-parametric interval mapping of a) six parental behaviours in males (n=419) and females (n=350), b) the subset of F2 animals that performed a behaviour (i.e. excluding the animals that did not huddle or lick their pups for the duration of all three trials). Sample sizes for huddling: males = 259, females = 300; licking: males = 319, females = 313. c) Haley-Knott regression on nqrank normalized values of all F2 individuals, using sex as a covariate. Plots show the difference in lod scores for the scan with sex as an interactive and as an additive covariate minus the scan with sex as an additive covariate alone. The artificial narrow peaks at the ends of chromosomes result from lack of genotype imputation by Multiplex Shotgun Genotyping at chromosome ends (Nest quality, chromosome 22; Latency to approach, chromosome 9; Latency to handle, chromosome 14). a–c) Dashed lines denote the P=0.05 genome-wide significance level determined by 1,000 permutations.

Extended Data Figure 4. QTL effect plots

Phenotype means (±SEM) against genotypes at peak QTL markers reported in Figure 4 and Extended Data Figure 3. Above each graph is the chromosomal position of the QTL peak. *, significant QTL-by-sex interaction. The percent variance explained is given for each QTL and is underlined if the QTL was significant in a QTL analysis of that sex (Extended Data Figure 3a,b).

Extended Data Figure 5. Behaviour of sexually naïve and parental animals

a) Sexually naïve animals were tested with 4–6 day old conspecific pups, and parents with their own 4–6 day old pups. Box plots indicate median, interquartile range, and 10^th to 90^th percentiles. P. maniculatus (man, magenta) and P. polionotus (pol, green). *, P<0.05; **, P<0.01; ***, P<0.001; ****, P<0.001; NS, not significant by Mann-Whitney U or Fisher’s exact tests (retrieving and infanticide). b) Nest quality (mean ±SEM, and each pair in magenta circles for P. maniculatus and green squares for P. polionotus) in the two weeks following the birth of a litter. Existing nests were removed from the parents cage at the time of weaning and 5 g of new cotton nesting material (Nestlet) was provided. Litters were born 1–4 days after weaning the previous litter, and nest quality was evaluated once a day in the home cage, where both mother and father were present. ****, P<0.0001 effect of species in a two-way ANOVA including species and time as factors. There was no significant effect of time or species-by-time interaction.

Extended Data Figure 6. RNAseq analysis of P. maniculatus and P. polionotus hypothalamus

a) Clustering dendogram of RNAseq samples by Euclidian distances of transcript expression levels. Samples cluster by species but not by sex. b) Strength of differential expression between species. Each circle represents a gene and is colour-coded in magenta or green if its differential expression was significant at a 5% FDR. c) Relationship between average gene expression level and differential expression between species. In both b and c, genes that have been associated with parental care in previous studies (by physiological/pharmacological studies or induced mutations)^– are labelled; the gene is also boxed if located inside parental behaviour QTLs identified in this study.

Extended Data Figure 7. Expression analysis of genes in the chromosome 4 nest-building QTL

a) Twenty-three genes that are differentially expressed in the hypothalamus at 5% FDR between P. maniculatus and P. polionotus, sorted by FDR-adjusted P-value. No significant differences between sexes for any gene. For each gene, its average expression level in each species and sex is shown on the left, the fold-difference in expression between the species in the middle, and the FDR-adjusted P-value on the right. b) Allele-specific expression of the 15 genes from a for which interspecific genetic variation allows this analysis. Mean fraction of reads (±SEM) matching the P. maniculatus allele from RNAseq of the hypothalamus of 12 male and 12 female P. maniculatus × P. polionotus F1 hybrids. There are no significant differences between males and females so the sexes were combined. *, P<0.05; **, P<0.01; ***, P<0.001; ****, P<0.0001 by a linear model that measures allelic bias: y = log(maniculatus allele counts gene_x)-log(polionotus allele counts gene_x) ~ sex).

Extended Data Figure 8. Allele-specific expression of Avp in different brain regions

a) Immnunofluorescence staining of vasopressin in a coronal section of a P. maniculatus male brain, showing the main vasopressin producing nuclei. Red circles illustrate the 1 mm diameter (1.5 mm thick) circular punches used to microdissect these nuclei. b,c) Number of Avp transcripts (b) and allele-specific expression of Avp (c) in each of the microdissected regions in P. maniculatus × P. polionotus F1 hybrid animals, measured by droplet-digital PCR; horizontal line at the mean. **, P<0.01; ***, P<0.001; by ANOVA with Bonferroni correction. SCN: suprachiasmatic nucleus; SON: supraoptic nucleus; PVN: paraventricular nucleus of the hypothalamus; BNST: bed nucleus of the stria terminalis.

Extended Data Figure 9. Relationship between anxiety and nest building

a) Fraction of time in the open arms of an elevated plus maze. Line at the mean. ****, P<0.0001 for difference between species by two-way ANOVA with sex and species as factors. No significant effect of sex or sex-by-species interaction. b) Spearman correlation coefficients among F2s between fraction of time in the open arms and parental behaviours. Handling and approach are promptness to perform those behaviours. c) Linkage (lod score) to chromosome 4 of nest building behaviour and fraction of time in open arms. Males and females combined since there are no major differences in the lod scores between sexes. Red line denotes the location of Avp. Dashed lines denote the P=0.05 genome-wide significance level determined by 1,000 permutations of each trait (n=769).

Extended Data Figure 10. Chemogenetic experiments on vasopressin neurons of Mus musculus

a) Generation of Avp-Cre BAC-transgenic Mus musculus. Top: Schematic illustrating the targeting of the IRES-Cre cassette immediately after the Avp stop codon. Bottom: Immunofluorescence histology of vasopressin (AVP) and Cre in the paraventricular nuclei of the hypothalamus of an Avp-Cre BAC-transgenic Mus musculus. Scale bar, 100 μm. b) A recombinant adeno-associated virus (rAAV) containing a Cre-dependent DREADD was injected into the paraventricular nuclei (PVN) of Avp-Cre transgenic Mus musculus. c,d) Nest-building behaviour for one hour following intraperitoneal injection with 0.9% NaCl or with the DREADD agonist clozapine-N-oxide (CNO) at 10 mg/kg. In c animals expressed the inhibitory Gi-DREADD and in d, the excitatory Gq-DREADD. Males (with blue symbols at the mean ±SEM) are on the left and females (red) are on the right in each panel. Statistical significance determined by repeated-measures ANOVA for quadratic trend.

Figure 1. Parental behaviours of monogamous and promiscuous Peromyscus mice

a) Cladogram of selected Peromyscus species, including known monogamous species (adapted from ref. 8). b) Typical P. maniculatus and P. polionotus parents in the laboratory. c–f) Parental behaviours of P. maniculatus and P. polionotus. Each circle represents the median behaviour of an animal tested three times (c–e) or the fraction of pups retrieved over the three tests (f). Box plots indicate median, interquartile range, and 10^th to 90^th percentiles. Kruskal-Wallis followed by Dunn’s (c) and two-way ANOVA results (d–f): significant effects of species (sp), sex (s), or sex-by-species interaction (x) at P<0.01.

Figure 2. Effect of cross fostering on parental behaviour

a) Design of the cross-fostering experiment between P. maniculatus and P. polionotus. b–e) Parental behaviour of animals raised by their own parents (white) or by parents of the other Peromyscus species (grey). Each circle represents the median behaviour of an animal tested three times. Box plots indicate median, interquartile range, and 10^th to 90^th percentiles. *, P<0.05; **, P<0.01; NS, not significant by two-sided Mann-Whitney U with Bonferroni correction.

Figure 3. Distribution of parental behaviours in each species and their interspecific hybrids

a) Genetic cross design. Female P. maniculatus were mated to P. polionotus males to found the cross (see Methods); the behaviour of females and males of each species is shown here for comparison. b–e) Violin plots show the distribution of behaviours. Each dot represents the mean behaviour of an animal over three trials.

Figure 4. The genetic architecture of parental behaviour

a) Correlation matrix of parental behaviours in F2 hybrids. b) The linkage (lod score) of each behaviour to each marker in each sex and the sex-by-genotype interaction (gray). Boxes delineate the Bayes 0.95 credible interval of QTL of genome-wide significance (α=0.05). Dots denote the allele at the peak of the QTL that correlates with a higher value for that behaviour (for handling and approach, a decrease in latency to perform behaviour). Heterozygotes at QTLs on chromosomes 2 and 12 (females only) retrieve more pups (see Extended Data Fig. 4 and Supplementary Discussion).

Figure 5. Role for vasopressin in parental nest building

a) 23 genes in the nest-building QTL on chromosome 4 are differentially expressed in the hypothalamus of the two species, 15 of which had interspecific variants required for allele-specific analysis in F1 hybrids. b) Expression of the 9 candidate genes from (a) in P. maniculatus (magenta) and P. polionotus (green) hypothalamus. The voom-transformed difference in expression is shown above; colour indicates the species with higher expression. Circles represent each individual. No significant differences between sexes for any gene. c) Parental behaviours of P. polionotus following vehicle, 100 ng vasopressin (AVP) or 100 ng oxytocin (OXT) administration intracerebroventricularly. Each individual (grey) is connected by lines. Red and blue circles with whiskers indicate mean ± SEM *, P<0.05; **, P<0.01; NS, not significant by Friedman followed by Dunn’s compared to Vehicle.