The medial amygdala sends projections to the cMPOA.

(A) (left) Schematic diagram of “virgin” male mice, intermediate males called “FGE” by mating and cohabitation with partner female, and “paternal” mice, that experienced delivery of partner females and cohabitation with pups for three days. (right) Relative frequencies of behavioral patterns of virgin, FGE, and paternal mice toward pups during four consecutive days of pup exposure.

(B) upper, medial amygdala in the parasagittal section referring to the Mouse Brain Atlas of Franklin & Paxinos (2007). lower, Representative images of the Me stained with anti-Cart visualized with Alexa Fluor 488 and BDA visualized with Alexa Fluor 568-conjugated streptavidin (magenta). Scale bar = 200 µm.

(C) MPOA in the parasagittal section.

(D) Representative images of the Me injected with the tracer BDA in the MPOA. Parasagittal sections are shown. Fixed sections were visualized with Alexa Fluor 568-conjugated streptavidin (magenta). In parallel, fixed sectioned were immunostained with anti-NPI and visualized with Alexa Fluor 647 (green). Scale bar = 200 µm.

(E) Image of cFos visualized by DAB staining in the mice showed parenting and infanticide. Scale bar = 500 µm.

(F) Expression distribution of cFos-positive neurons in each region delimited by the 100 µm grid. Red-shifted color indicates a higher density of cFos-positive neurons. paternal (n = 8), virgin (n = 3)

(G) AAV5-EF1α-DIO-eYFP was injected into the Me of Cartpt-Cre mice. Representative images of eYFP emission from Me and MPOA. MeCartpt neurons sent projection fibers into the MPOA. Scale bar = 200 µm.

(H) Representative images of the MePD in Cartpt-Cre: Ai9 reporter mice. Among MePD neurons immunostained with anti-NeuN antibody visualized with a secondary antibody conjugated to Alexa Fluor 647 (blue), 21.2% Cartpt neurons (magenta) were tdTomato positive (1 male, 1 female). Scale bar = 300 µm.

(I) Relative frequencies of behavioral patterns of FGE Cartpt-Cre: Ai9 reporter mice toward pups during four consecutive days of pup exposure.

(J) Representative images of cFos induction in the MePD following pup exposure or neutral object to FGE Cartpt-Cre: Ai9 reporter mice. Fixed sections were stained with anti-cFos and visualized with Alexa Fluor 647 (green). Arrowhead indicates the tdTomato-cFos-double-positive neurons. Scale bar = 30 µm.

(K) The ratio of c-Fos-positive neurons out of all tdTomato-positive MeCartpt neurons following pup or object exposure. object (n = 3), infanticide (n = 4), parenting (n = 4). Statistical analysis by one-way ANOVA, **P < 0.01.

Synapse plastic changes in the cMPOA neuron that occur in the process leading to paternal expression.

(A) AAV5-EF1α-DIO-eYFP or AAV5-EF1α-DIO-ChR2-eYFP was injected into the Me of Cartpt-Cre mice. After the expression of eYFP with/without ChR2, whole-cell patch clamp recordings were performed from cMPOA neurons in parasagittal sections of virgin and FGE mouse brains.

(B) Schematic image of the recording from the cMPOA neuron in a parasagittal section. Input fibers from MeCartpt expressed ChR2 were stimulated by blue light application.

(C) Diagram of dominant responses of IPSC and EPSC in the opposite current direction as that of GABAergic and glutamatergic ionotropic receptors.

(D) Representative traces of blue light-evoked IPSCs and EPSCs in IPSC-and EPSC-dominant cMPOA neurons. When MeCartpt neurons were injected with the control AAV virus, no responses to optical activation were observed in cMPOA neurons.

(E) Relative frequencies of excitatory-and inhibitory-dominant cMPOA neurons in virgin mice (n = 12 cells; 4 animals), FGE mice (n = 14 cells; 3 animals), and FGE mice with GDPβS (1 mM) in the pipette solution (n = 9 cells; 2 animals). A significant difference in of excitatory-to inhibitory-dominant cMPOA neurons ratio between virgin and FGE group mice was observed (Fisher’s exact probability test, **P = 0.0048). A significant difference in the ratio of excitatory- and inhibitory-dominant cMPOA neurons between FGE mice and FGE mice with GDPβS (1 mM) in the pipette solution (FGE + GDPβS group) was also observed (Fisher’s exact probability test, #P = 0.0131).

(F) Representative traces of blue light-evoked IPSCs in IPSC-dominant cMPOA neurons. The application of TTX blocked light-evoked IPSC. Additional application of 4AP recovered IPSC, which was sensitive to picrotoxin.

(G) left, Schematic image of the recording from cMPOA neuron in a parasagittal section. right, Representative traces of sIPSC in cMPOA from virgin, FGE and paternal mice.

(H) Frequency of sIPSC in the cMPOA neuron of virgin (n =38 from 11 animals), FGE (n = 12 from 4 animals), and paternal mice (n = 14 from 6 animals). Statistical analysis by Kruskal-Wallis test, *P = 0.0207,

(I) Cumulative probability plots of the inter-event interval of the sIPSCs. Statistical analysis by Kolmogorov-Smirnov test, P < 0.0001 (virgin vs FGE), P = 0.0227 (virgin vs paternal),

(J) Amplitude of sIPSC in the cMPOA neuron of virgin (n =38 from 11 animals), FGE (n = 12 from 4 animals). Statistical analysis by One way ANOVA, P = 0.0666.

(K) Cumulative probability plots of the amplitude of the sIPSCs. Statistical analysis by Kolmogorov-Smirnov test, P > 0.05

(L) left, Schematic image of the recording from the cMPOA neuron in a parasagittal section. Input fibers from MePD were electrically stimulated. right, Input−output curves of stimulus-evoked IPSPs from virgin (n = 18 cells; 6 animals), FGE (n = 14 cells; 4 animals), paternal (n =8 cells; 3 animals), virgin with GDPβS (1 mM) in the pipette solution (n = 13 cells; 4 animals), paternal with GDPβS (n = 16 cells; 4 animals). Statistical analysis by two-way RM ANOVA, P = 0.0016.

(M) left, summary of experimental procedure. right, RNA-seq identification of candidate gene expressed in the cMPOA neurons. Heatmap showing a z-scored TPM of selected genes for neuronal cluster-mediated social behavior in the MPOA21. virgin n = 4 (n = 4), FGE (n = 4), and paternal (n = 4).

(N) Representative images of the MPOA. Galanin mRNA Alexa Fluor 647 -conjugated hairpin DNA (#S23) (magenta), streptavidin conjugated with FITC (green) and DAPI (blue). (N1-N3) magnified images corresponding area in (N), Scale bar = 200 µm for (N, left) and 20 µm for (N1-N3)

(O) Representative images of cMPOA neurons after whole-cell patch-clamp recording. Biocytin was infused from recording pipette. Fixed sections visualized with FITC -conjugated streptavidin (green) and in situ HCR for Galanin mRNA Alexa Fluor 647 -conjugated hairpin DNA (#S23) (magenta). Scale bar = 20 µm.

(P) Ratio of Galanin positive cMPOA neurons targeted for whole-cell patch-clamp recording.

(Q) Input−output curves of stimulus-evoked IPSPs from the Galanin positive cMPOA of virgin mice (n = 12 cells; 7 animals) and FGE mice (n = 7 cells; 5 animals). Values are presented as the mean ± standard error. (bottom) Representative traces used to construct input−output curves (40, 80, 120, 160, and 200 μA stimuli). Statistical analysis by two-way RM ANOVA, P = 0.0397, Holm-Sidak post hoc test, *P < 0.05, **P < 0.01 vs. virgin group.

The modulation of Me inputs into the cMPOA.

(A) AAV10-EF1α-DIO-vLWO-eGFP or AAV10-EF1α-DIO-eYFP were injected into the Me of Cartpt-Cre mice and whole-cell recordings were obtained from fluorescent-labeled Me neurons.

(B) Representative resting membrane potential traces recorded from a MeCartpt neuron expressing vLWO. Green light was applied for 20 s (green bar).

(C) Effects of green light application on the resting membrane potential in the MeCartpt neurons expressing vLWO. The data are shown as the averaged potentials between −6 s and −1 s before the onset of green light application (pre-green light) and between 5 s and 10 s after the onset of green light application (green light). Lines represent the data obtained from individual neurons (n = 5), and bars represent the averaged data obtained from five neurons. Statistical analysis by two-tailed paired t-test, **P = 0.0017, compared with pre-green light application.

(D) Representative resting membrane potential traces recorded from a MeCartpt neuron expressing control eYFP. Green light was applied for 20 s (green bar). Five animals were excluded due to exclusion criteria.

(E) Effects of green light application on the resting membrane potential in the MeCartpt neurons expressing control eYFP. Lines represent the data obtained from individual neurons (n = 10), and bars represent the averaged data obtained from nine neurons. Statistical analysis by two-tailed paired t-test, P = 0.6308, compared with pre-green light application.

(F) AAV5-EF1α-DIO-ChR2(H134R)-mCherry and AAV10-EF1α-DIO-vLWO-eGFP-5HT1A was injected into the Me of Cartpt-Cre mice and whole-cell recordings were obtained from cMPOA neurons.

(G) Representative images of the Me expressing eGFP and mCherry. Scale bar = 200 µm.

(H) Representative traces of blue light-evoked IPSCs in the cMPOA neurons. Red bar indicates red light application (2 s).

(I) Effects of red-light application on the blue light-evoked postsynaptic currents in the cMPOA neurons. Lines represent the data obtained from individual neurons (n = 4), and bars represent the averaged data obtained. Statistical analysis by two-tailed paired t-test ***P = 0.0007, compared with pre-red light application.

(J) AAV10-EF1α-DIO-vLWO-eGFP was injected into the Me of Cartpt-Cre mice and whole-cell recordings were obtained from cMPOA neurons.

(K) Representative trace of green light-induced modulation of sIPSCs in the cMPOA neurons. Green light was applied for 20 s.

(L) Effects of green light application on the frequency of sIPSC in the cMPOA neurons. Lines represent the data obtained from individual neurons (n = 25), and bars represent the averaged data. Statistical analysis by repeated measures ANOVA, P = 0.0396, followed by Dunnett’s multiple comparisons post hoc test *P = 0.0332 compared with pre-green light application. One data showed large sIPSC frequency. It was judged as an outlier in Grubbs’ test. Without this data point, the same repeated measure ANOVA analysis resulted in F (1.733, 39.87) = 4.149, P=0.0279, indicating that the conclusion remains when leaving out this outlier.

(M) Representative trace of sIPSCs in the cMPOA neurons in the presence of TTX (1 μM).

(N) Application of TTX blocked the green light induced changes of IPSCs frequency. Lines represent the data obtained from individual neurons (n = 12), and bars represent the averaged data. Statistical analysis by repeated measures ANOVA, P = 0.6561 followed by Dunnett’s multiple comparisons post hoc test *P = 0.2338 compared with pre-green light application.

Optogenetic manipulation of the MeCartpt−cMPOA input suppressed infanticidal behavior of virgin male mice.

(A) AAV10-EF1α-DIO-vLWO-eGFP-5HT1A or AAV10-EF1α-DIO-eYFP was injected into the Me of Cartpt-Cre mice.

(B) Relative frequencies of paternal/infanticidal behaviors toward pups during manipulation of the MeCartpt-to-cMPOA input by vLWO activation. Significant difference in the ratio showing infanticidal behavior between mice expressing vLWO-eGFP (n = 8, n =5 for post) or eYFP only (n = 10). Analysis of Variance of Aligned Rank Transformed Data, F (1.236, 13.60) = 63.27, P < 0.0001.

(C) Latencies to first sniffing were compared between vLWO and control group mice. Statistical analysis by two-way RM ANOVA, F (1, 16) = 1.642, P = 0.2183.

(D) Survival curve to indicate the probability of committing infanticide on each experimental day. vLWO-eGFP (n = 8, n =5 for post) or eYFP only (n = 10).

(E) AAV10-EF1α-DIO-vLWO-eGFP-5HT1A or AAV10-EF1α-DIO-eYFP was injected into the Me of vGAT-IRES-Cre mice.

(F) Representative images of eYFP expressed in the Me of vGAT-IRES-Cre mice (green). Sections were immunostained with anti-Cart and visualized with a secondary antibody conjugated to Alexa Fluor 594 (magenta). All scale bar = 200 µm.

(G) Relative frequencies of paternal/infanticidal behaviors during vLWO activation at the MevGAT input into cMPOA. Significant difference in the ratio showing infanticidal behavior between mice expressing vLWO-eGFP (n = 8, n =5 for post) or eYFP only (n = 10). Analysis of Variance of Aligned Rank Transformed Data, F (1.236, 13.60) = 5.2140, P = 0.026

(H) Latencies to first sniffing were compared between vLWO and control group mice. Statistical analysis by two-way RM ANOVA, F (1, 16) = 1.642, P = 0.2183.

(I) Survival curve to indicate the probability of committing infanticide on each experimental day. vLWO-eGFP (n = 8, n =5 for post) or eYFP only (n = 10).

Inhibitory synaptic inputs into the BSTrh were potentiated in paternal group mice.

(A) Whole-cell patch clamp recordings were performed on BSTrh neurons (Rec). To observe the evoked synaptic responses, electrical stimulation was delivered through a stimulating electrode (Stim). rh, rhomboid nucleus; pr, principal nucleus; LV, lateral ventricle; ic, internal capsule; f, fornix; D3V, dorsal third ventricle; GP, globus pallidus; ac, anterior commissure; ST, stria terminalis

(B) Input−output curves of stimulus-evoked IPSPs from virgin mice (n = 16 cells; 10 animals), paternal mice (n = 17 cells; 11 animals), and FGE mice (n = 10 cells; 5 animals). Values are presented as the mean ± standard error. (inset) Representative traces used to construct input−output curves (50, 100, 150, and 200 μA stimuli). Statistical analysis by two-way RM ANOVA, P = 0.0083, followed by Holm-Sidak post hoc test, *P < 0.05, **P < 0.01 vs. virgin group.

(C) Input−output curves of stimulus-evoked EPSPs from virgin mice (n = 9 cells; 6 animals), paternal mice (n = 15 cells; 6 animals), and FGE mice (n = 6 cells; 3 animals). Values are presented as the mean ± standard error. (inset) Representative traces used to construct input−output curves (25 and 100 μA stimuli). Statistical analysis by two-way RM ANOVA, P > 0.05.

(D) Photographs of the MPOA from vehicle-injected and NMDA-lesioned mice. Sections were immunostained with anti-NeuN and visualized with DAB. Arrowhead indicates the ventral edge of lesioned area. Scale bar = 500 μm.

(E) Input−output curves of stimulus-evoked IPSPs from paternal vehicle-injected mice (vehicle; n = 26 from 7 mice, NMDA; n = 15 from 6 mice). Values are presented as the mean ± standard error. (inset) Representative traces used to construct input−output curves (25, 75, 125, and 175 μA stimuli). Statistical analysis by two-way RM ANOVA, P = 0.0032 followed by Holm-Sidak post hoc test, *P < 0.05, **P < 0.01, ***P < 0.001 vs. virgin group.

(F) Input−output curves of stimulus-evoked IPSPs from vehicle-injected virgin mice (vehicle; n = 7 cells; 4 animals) and cMPOA-lesioned virgin mice (NMDA; n = 9 cells; 4 animals). Values are presented as the mean ± standard error. (inset) Representative traces used to construct input−output curves (25, 75, 125, and 175 μA stimuli). Statistical analysis by two-way RM ANOVA, P > 0.05.

(G) Input−output curves of stimulus-evoked IPSPs from virgin mice (n = 10 cells; 4 animals) and paternal mice (n = 9 cells; 4 animals) recorded with GDPβS (1 mM) in the pipette solution. Values are presented as mean ± standard error. (inset) Representative traces used to construct input−output curves (50, 100, 150, and 200 μA stimuli). Statistical analysis by two-way RM ANOVA, P > 0.05.|

(H) Input−output curves of stimulus-evoked EPSPs from virgin mice (n = 9 cells; 5 animals) and paternal mice (n = 10 cells; 4 animals) recorded with GDPβS (1 mM) in the pipette solution. Values are presented as mean ± standard error. (inset) Representative traces used to construct input−output curves (25, 75 and 125 μA stimuli). Statistical analysis by two-way RM ANOVA, P > 0.05.

(I) (top) Paired-pulse ratio of eEPSCs (inter-stimulation interval = 50 ms) from virgin mice (n = 6 cells; 5 animals) and paternal mice (n = 7 cells; 6 animals). Statistical analysis by two-tailed unpaired t-test: P > 0.05. (bottom) Paired-pulse ratio of eIPSCs (inter-stimulation interval = 100 ms) from virgin mice (n = 10 cells; 9 animals) and paternal mice (n = 15 cells; 10 animals). Statistical analysis by two-tailed unpaired t-test, P > 0.05.

Inhibitory synaptic inputs into the BSTrh were potentiated in pup-sensitized mice.

(A) Schematic diagram of continuous pup exposure. Four male mice at postnatal day 28 for the first time were exposed to three pups for 30-60 min, four times per week. After one, three, six, and eight weeks of this protocol, mice were used for the experiment.

(B) Relative frequencies of paternal/infanticidal behaviors of one, three, six, and eight-week pup exposed and control mice.

(C) left, a schematic image of the recording from the cMPOA neuron in a parasagittal section. Input fibers from MePD were electrically stimulated. middle, Input−output curves of stimulus-evoked IPSPs in the cMPOA from pup-sensitized mice (n = 15 cells; 4 animals) and control mice (n = 9 cells; 3 animals). Statistical analysis by two-way RM ANOVA, P = 0.7579, right, representative trace (40, 120, 200 μA stimuli)

(D) left, a Schematic image of the recording from the BSTrh neuron in a parasagittal section. Stria terminalis was electrically stimulated. middle, Input−output curves of stimulus-evoked IPSPs in the BSTth from pup-sensitized mice (n = 16 cells; 4 animals) and control mice (n = 9 cells; 5 animals). Statistical analysis by two-way RM ANOVA, P = 0.0016, right, representative trace (20, 60, 100, 140 μA stimuli)

Values are presented as the mean ± standard error.