Mice with the targeted mutation of p140Cap have been previously described (Repetto et al., 2014). The mutation was maintained in homozygosity (as the animals are healthy, despite reduced fertility) in the transgenic unit of the Molecular Biotechnology Center (University of Turin, Turin, Italy) and animals were given water and food ad libitum. Genetic screening was performed by PCR as previously described (Repetto et al., 2014). Procedures were conducted in conformity with national and international laws and policies, protocols were approved by the internal Ethical Committee and authorized by the Italian Ministry of Health. To assess the fertility of both sexes, three possible mating combinations between p140Cap KO and WT P60 female and male mice were carried out. Measurements included the number of days required to produce the first litter, the numbers of litters born, and the number of pups per litter, recorded over 90 days.

Assessment of vaginal opening and examination of estrus cyclicity were carried out as previously described (Caligioni, 2009). Vaginal lavage of p140Cap KO females and their control littermates was performed every day (9:00 A.M. to 11:00 A.M.) for 20 consecutive days using 0.9% saline. Smears were observed under the microscope and the phase identified as diestrus/metestrus if they predominantly contained leukocytes, as proestrus if they predominantly contained nucleated cells, and as estrus if they predominantly contained cornified epithelial cells. An estrus cycle was considered normal when the vaginal lavage had leukocytes for 2 days followed by 1 day of nucleated cells and 1–2 days of cornified cells.

Ovaries were collected from P120 p140Cap KO and WT mice, fixed in 4% (w/v) PFA in PBS, and stored at 4°C. Paraffin-embedded ovaries were sectioned at a thickness of 5 μm and stained with hematoxylin-eosin. Corpora lutea and preovulatory follicles were counted on photomicrographs from every tenth section throughout the ovary.

Pregnant mare’s serum gonadotropin (5 IU/mouse) was inoculated intraperitoneally in P60 p140Cap KO and WT animals. 48 h later, human chorionic gonadotropin (5 IU/mouse) was administered intraperitoneally to stimulate ovulation. After 22 h, mice were sacrificed, the ovaries were removed, cells of the cumulus oophorus were eliminated by treatment with 0.3% hyaluronidase, and the oocytes were collected and counted. KP-54 (Tocris) was injected intraperitoneally (1 nmol in 100 μl of PBS/mouse) in P45 (for the determination of c-Fos-ir GnRH neurons) and P60 (for the quantification of serum LH levels) p140Cap KO and WT females in the diestrus phase of the estrous cycle. For the quantification of serum LH levels, animals were deeply anesthetized and blood was collected from the retro-orbital cavity 1 h after the injection. For the determination of c-Fos-ir GnRH neurons, animals were deeply anesthetized and transcardially perfused 90 min after the injection, and brains processed as described below.

For the preparation of adult and P10 p140Cap KO and WT brains, mice were deeply anesthetized with Avertin (30 μl pure Avertin in 400 μl of PBS/mouse), transcardially perfused with 10 ml of PBS and then with 10 ml of 4% (w/v) PFA in PBS, and brains were dissected. For kisspeptin immunolabeling, P100 p140Cap KO and WT females in the proestrus phase of the estrous cycle were analyzed. For VGLUT-VGAT/GnRH double immunolabeling, P60 p140Cap KO and WT females in the diestrus phase of the estrous cycle were analyzed. For VGLUT-VGAT/KP double immunolabeling, P120 p140Cap KO and WT females in the proestrus phase of the estrous cycle were analyzed. For the preparation of embryonic brains, E14.5 embryos were collected by cesarean cut, and brains were dissected. After dissection, embryonic, P0, P10, and adult brains were post-fixed overnight at 4°C in 4% (w/v) PFA in PBS, placed overnight at 4°C in 30% (w/v) sucrose in PBS for cryoprotection, embedded in OCT blocks, and stored at –80°C until analysis. P0, P10, and adult brains were sliced into free-floating coronal sections of 30 μm using a cryotome (Leica CM1950). Free-floating sections were collected in PBS in multiwell plates and stored at –20°C in a cryoprotectant solution [30% (v/v) glycerol and 30% (v/v) ethylene glycol in 0.2 M phosphate buffer, pH 7.4] until processed for immunolabeling. Embryonic brains were sliced into sagittal sections of 16 μm using a cryotome, mounted on glass slides, and stored at –20°C until processed for immunolabeling.

Immunohistochemistry to detect GnRH was performed as previously reported (Giacobini et al., 2008). For GnRH/p140Cap, GnRH/VGAT, GnRH/VGLUT, and c-Fos/GnRH double immunostainings, sections were blocked in 12% Normal Goat Serum and 2% Bovine Serum Albumin in 0.5% TritonX-100 in PBS for 1 h at room temperature (RT) and subsequently incubated with primary antibodies in 0.1% TritonX-100 in PBS for 48 h at 4°C. Sections were then washed in PBS, incubated with the secondary antibodies and DAPI (1:1000) diluted in 0.2% TritonX-100 in PBS for 2 h at RT, and washed three times in PBS. Staining for KP/GnRH was carried out as described elsewhere (Naulé et al., 2014). For KP/VGLUT and KP/VGAT double immunostainings, sections were blocked in 2% Normal Goat Serum and 0.2% Bovine Serum Albumin in 0.2% TritonX-100 in PBS for 1 h at RT. Sections were then incubated with anti-KP primary antibody in 0.2% TritonX-100 in PBS for 48 h at 4°C, incubated with anti-sheep secondary antibody diluted in 0.2% TritonX-100 in PBS for 2 h at RT, incubated with anti-VGLUT/VGAT primary antibody in 0.2% TritonX-100 in PBS for 48 h at 4°C, incubated with anti-guinea pig secondary antibody diluted in 0.2% TritonX-100 in PBS for 2 h at RT, and finally washed three times in PBS. GnRH/activated caspase 3 immunostainings were carried out as described elsewhere (Macchi et al., 2017). Sections were mounted on glass slides using 1,4-diazabicyclo [2.2.2] octane (Sigma-Aldrich) and examined. In all steps, p140Cap KO and WT sections were processed in parallel at the same time. The total number of GnRH-ir neurons in the adult brain was determined as described elsewhere (Herbison et al., 2008). Determination of the total number of GnRH-ir neurons in the embryonic brain was carried out by sectioning whole brains in sagittal orientation at 16 μm. GnRH-ir neurons were counted in every third section through the brain, and the number obtained was multiplied by 3 to get the total number of GnRH neurons. Quantification of GnRH immunoreactivity (voxel counts) was carried out on anatomically matched sections of the median eminence [plates 47-50 of the Mouse Brain Atlas of Paxinos et Franklin (Paxinos and Franklin, 2001)]. Determination of the number of kisspeptin neurons was carried out in anatomically matched sections of the anteroventral periventricular region (plates 28–29) and the rostral (plate 30) and caudal (plates 31–32) regions of the RP3V. Quantification of kisspeptin immunoreactivity (voxel counts) was carried out on anatomically matched sections of the arcuate nucleus (plate 41-45). Determination of the number of VGLUT-ir punctae in the OVLT and POA was carried out on anatomically matched sections (plates 25-27).

In situ hybridization was carried out as previously described (Oleari et al., 2021). Briefly, PFA-fixed 20 μm thick cryosections were incubated with digoxigenin (DIG)-labeled anti-sense riboprobes for mouse Gnrh1 at 65°C. mRNA expression was revealed with AP-conjugated anti-DIG antibody (1:1500; Roche), 4-Nitro blue tetrazolium chloride solution, and 5-Bromo-4-chloro-3-indolyl phosphate disodium salt (1:1000, Roche).

Primary antibodies used are: rabbit anti-GnRH (Beauvillain and Tramu, 1980), sheep anti-KP AC053 (Franceschini et al., 2013), sheep anti-GnRH (Skrapits et al., 2015), mouse anti-p140Cap (Grasso et al., 2018, see Supplementary Figure 1 for specificity test), guinea pig anti-Vesicular GABA Transporter (VGAT, Synaptic System, used 1:500), guinea pig anti-Vesicular glutamate Transporter (VGLUT, Synaptic System, used 1:500), rabbit anti-c-Fos (sc-52 Santacruz, used 1:500) and rabbit anti-cleaved caspase 3 (Cell Signaling Technology, used 1:400, Whittaker et al., 2021). Secondary antibodies used are: Alexa-Fluor 488 goat anti-rabbit (Invitrogen, used 1:400), Cy3 goat anti-rabbit (Invitrogen, used 1:800), Alexa-Fluor 568 goat anti-mouse (Invitrogen, used 1:400), FITC anti-guinea pig (Sigma F6261, used 1:100), Cy3 donkey anti-sheep (Jackson, used 1:1000). Peroxidase-conjugated secondary antibodies were obtained from GE Healthcare.

Following cycle stage determination by vaginal smear microscopic analysis, mice were anesthetized in the afternoon of proestrus and at the diestrus stage, and blood was collected from the retro-orbital cavity. Serum LH was measured using Rodent LH ELISA kit (ERKR7017, Endocrine Technologies). Serum E2 was measured using Rodent estradiol ELISA kit (ERKR7011, Endocrine Technologies).

Images were captured using a Nikon microscope (Eclipse 80i) and 2 × /0.06 NA, 10 × /0.30 NA, and 20 × /0.50 NA objectives (Nikon) equipped with a digital camera (CX 9000; MBF Bioscience). For observation coupled with confocal analysis, a laser-scanning Fluoview confocal system (IX70; Olympus) and 10 × /0.30 NA, 20 × /0.70 NA, and 60 × /1.25 NA objectives (Olympus) were used. Subsequent analysis of digitized images was performed with ImageJ (NIH, Bethesda, Maryland¹) software. For the determination of c-Fos-ir GnRH neurons, c-Fos-ir/GnRH-ir double-stained neurons in the hypothalamus were counted and expressed as a percentage of the total number of GnRH-ir neurons. Quantification of kisspeptin and GnRH-ir fiber density was carried out by voxel counts on a set of 10 serial image planes (z step size = 1 μm). KP-ir, VGLUT-ir, and VGAT-ir punctae on GnRH neurons were counted manually in a set of 20 serial image planes (z step size = 0.5 μm). Only GnRH neurons whose cell bodies were entirely present in the slice depth were considered for the analysis. The density of VGLUT-ir punctae in the OVLT and POA was calculated using the function “Find Maxima” of ImageJ. Photoshop (Adobe) software was used to process, adjust and merge the photomontages.

Total RNA was extracted from dessected hypothalami and pituitary glands from p140Cap KO and WT mice using Trizol Reagent (Ambion, Life Technologies Italia) and its concentration was determined with a NanoDrop™ 1100 (NanoDrop Technologies, Wilmington, DE, United States). Total RNA was reverse transcribed with high-capacity cDNA reverse transcriptase (#4368813, Applied BioSystem) according to the manufacturer’s instructions and amplified with specific primers. Taqman PCR reactions were performed using the Universal Probe Library system (Roche Italia, Monza, Italy) and quantified with the Molecular Analyst software (Bio-Rad Laboratories). The 18S rRNA pre-developed TaqMan assay (#4319413, Applied Biosystems) was used as an internal control. The expression of the target genes was calibrated against the values obtained in WT animals. Primers and probes used:

GnRH f 5′- CCCTTTGACTTTCACATCCAA-3′

GnRH r 5′- CGCAACCCATAGGACCAGT-3′ [probe #19]

LH f 5′- GTCCCAGGACTCAACCAATG-3′

LH r 5′- AACACCTGCTGGTGGTGAA-3′ [probe #10].

For the statistical comparison, GraphPad Prism software (GraphPad Software Inc.) was used. For each experiment, the statistical test used is reported in the figure legends. Shapiro-Wilk and Kolmogorov-Smirnov tests were used to test for normality, F test was used to test for equality of variance, and results were evaluated to choose the appropriate statistical test.

We noticed that p140Cap KO mice have a lower mating success as compared to WT, even if they do not show any apparent behavioral deficit that could be accounted for hypofertility. To quantify the hypofertility and to assess whether both male and female fertility was affected, we set up a continuous mating protocol of 90 days, comparing different breeding combinations of P60 mice. In matings involving WT females and KO males, the number of days required to produce the first litter (Figure 1A), the fertility index (Figure 1B), determined as the number of litters obtained in 90 days, and the number of pups/litter (Figure 1C), were similar to matings between WT animals. In contrast, in matings involving KO females and WT males, both the number of days required to produce the first litter (Figure 1A) and the fertility index (Figure 1B) were significantly lower, while the number of pups/litter was normal (Figure 1C). These results suggest that the loss of p140Cap significantly alters female, but not male, fertility.

Considering that the lack of p140Cap impacted only the reproductive axis in females and not in males, female mice were further characterized. To determine whether hypofertility was associated with ovulation deficiency, we evaluated ovarian morphology and follicle classification in P120 p140Cap KO and WT mice. Histological examination of p140Cap KO ovaries revealed a significant reduction in the number of pre-ovulatory follicles and corpora lutea as compared to WT (Figures 2A–C).

The defect observed in the ovarian compartment is apparently in contrast with the unaltered litter size observed in p140Cap KO mice. However, considering that several parameters can affect the number of pups at each pregnancy, such as efficiency of implantation, placenta functionality, and frequency of miscarriage, the reduced ovulation efficiency may be compensated, resulting in normal litter size. The impaired ovulation may depend on the loss of p140Cap expression in the ovarian tissue. Therefore, p140Cap expression was assessed by Western blot analysis on tissue extracts from the ovary and total brain of p140Cap KO and WT mice. As shown in Figure 2D, while p140Cap is highly expressed in the brain, it is undetectable in ovary extracts, indicating that the ovarian phenotype is non-cell-autonomous and may depend on the upstream hormonal axis.

To test if the reduced ovulation in p140Cap KO females was due to an impaired hormonal stimulation, a classical protocol of exogenous stimulation with Pregnan’ Mare’s Serum Gonadotropin (PMSG) and Human Chorionic Gonadotropin (HCG) was performed. As shown in Figure 2E, a relevant number of oocytes was found in both p140Cap KO and WT mice, indicating that in p140Cap KO females the ovarian tissue is still able to respond to an exogenous hormonal stimulation. However, while super-ovulation led to a mean of 20 oocytes/ampulla in WT mice, p140Cap KO females produced < 10 oocytes/ampulla. These results indicate that, although the ovarian tissue of KO mice may be less responsive, it is still able to undergo effective ovulation upon proper stimulation.

Ovulation of mature follicles in the ovary is induced by a large burst of GnRH-induced LH secretion—the preovulatory LH surge—which is triggered by increased estradiol (E2) levels. To assess the hormonal status in p140Cap KO mice, serum E2 and LH were quantified by ELISA in diestrus and proestrus phases. The results show that E2 and LH serum levels in p140Cap KO females do not increase during proestrus (Figures 2F,G), meaning that p140Cap KO females do not show the expected E2 and LH surge during proestrus, implying that these mice have a defective LH and E2 production in the proestrus stage.

We evaluated vaginal opening as an external index of puberty onset. Vaginal opening in female rodents is dependent on E2 levels and it reflects the maturation of the mouse female genital tract at the time of puberty, dependent on the activation of the HPG axis (Kennedy and Armstrong, 1973; Rodriguez et al., 1997). Weaned p140Cap KO and WT P21 females were daily examined for vaginal opening and the appearance of the first estrus, by flushing of vaginal secretion. Both vaginal opening and first estrus were significantly delayed in p140Cap KO females (Figures 3A,B). We also analyzed estrous cyclicity of p140Cap KO female mice by daily inspection of vaginal cytology. We found that p140Cap KO animals spend only 5% of time in the estrus phase, while WT females spend more than 20% of time (Figure 3C and Supplementary Figure 2), indicating that p140Cap KO mice have altered estrous cyclicity.

The endocrine and ovarian phenotypes described above clearly point to a central defect of GnRH neurons. Thus, we set forth to examine the expression and localization of p140Cap in the nasal-olfactory region of the developing embryo, and in the adult hypothalamus, by double immunostaining with anti-p140Cap and anti-GnRH antibodies, followed by confocal microscopy. At earlier embryonic ages, co-expression was clearly detected in a subpopulation (about 25–30%) of GnRH-ir neurons in the olfactory region and ventral forebrain (Figure 4A). When we analyzed sections of the adult hypothalamus, p140Cap was not detected in GnRH-ir neurons (Figure 4B), although it was detected in the vast majority of cells in this region.

Finally, we examined the expression of p140Cap on tissue extracts from the pituitary gland and hypothalamus by western blot analysis. While p140Cap is undetectable in pituitary gland extracts, it is highly expressed in the hypothalamus (Figure 4C), consistently with the immunostaining results.

To define which component of the HPG axis is affected by the absence of p140Cap, we analyzed the amount of GnRH and LH transcripts in the hypothalamus and pituitary gland of p140Cap KO females. Real-time quantitative PCR analysis showed a marked decrease in both LH expression in the pituitary (Figure 5A) and GnRH expression in the hypothalamus (Figure 5B) of p140Cap KO females, as compared to WT. These data indicate that, in the absence of p140Cap, the hypothalamus is defective in the expression of GnRH mRNA, with a consequent negative impact on LH expression. These results also imply that the defect in p140Cap KO mice may depend on impaired GnRH neuronal function.

Next, we evaluated the number of GnRH neurons in the MS, POA, AHA, and OVLT of p140Cap KO and WT P60 females, and we detected a significant 40% reduction in the total number of GnRH neurons in the mutant mice (Figures 5C,D), which is consistent with the reduced levels of GnRH mRNA. No ectopic localization of GnRH neurons was observed in mutant mice.

To verify that the loss of GnRH-ir neurons resulted from the lack of GnRH mRNA, and not from defects in GnRH prohormone processing, we conducted in situ hybridization. We found a significant reduction in the number of GnRH mRNA-containing neurons in p140Cap KO adult females as compared to WT (Supplementary Figures 3A,B).

GnRH neurons project their axons to the ME, where GnRH is released into the pituitary portal blood for delivery to the anterior pituitary, eliciting the secretion of LH and FSH. To assess if the reduction in the number of GnRH neurons resulted in a reduced number of GnRH innervation in the ME, we analyzed the ME of P60 females, by immunostaining with anti-GnRH. The results show a reduction of GnRH-immunoreactivity in p140Cap KO mice as compared with WT (Figures 5E,F).

A reduced number of hypothalamic GnRH neurons may result from an altered migration during embryonic development, impaired neurogenesis, or altered survival of GnRH neurons. To determine whether the absence of the p140Cap may impact the migration of GnRH embryonic neurons, we analyzed the nasal-olfactory region, the olfactory bulbs, and the ventral forebrain of p140Cap KO and WT E14.5 embryos. As shown in Figure 5G, there was no significant variation in the number of GnRH neurons between p140Cap KO and WT embryos in the various regions, indicating that the absence of p140Cap does not cause migratory defects of immature GnRH neurons during embryonic development. We also found no difference in the total number of GnRH neurons present in the hypothalamus of p140Cap KO and WT mice at E14.5, P0, and P10 (Figure 5H), indicating that the lower number of GnRH neurons in adult p140Cap KO mice does not result from impaired neurogenesis. Finally, to assess if the decreased number of GnRH neurons was due to increased cell death, we labeled adjacent coronal sections of the medial POA (MPOA, where most GnRH neurons reside) in P30 females, with antibodies against GnRH or activated caspase-3, a marker for cells committed to undergo apoptosis. Only a few apoptotic cells were detected in the MPOA of p140Cap KO and WT mice (1-2 cells per animal, Supplementary Figure 3C), suggesting that the loss of GnRH neurons in p140Cap KO mice is not due to increased apoptosis. Therefore, we were not able to determine the mechanism underlying the loss of GnRH neurons in p140Cap KO mice.

Having excluded embryonic migratory defects of GnRH neurons, a reduced number of hypothalamic GnRH neurons may result from post-migratory early-postnatal defects. We examined the morphological maturation of GnRH neurons in the hypothalamus of p140Cap KO and WT mice. We determined the fraction of GnRH neurons showing a unipolar/bipolar or a multipolar morphology in the brains of p140Cap KO and WT P10 female mice. In p140Cap KO brains, 34% of GnRH neurons were multipolar and 66% were bipolar, while in WT brains the fractions were 20% and 80%, respectively (Supplementary Figure 4A). The same analysis was conducted on brains from P60 (young adults) females, but we observed no change in the morphology of GnRH neurons (Supplementary Figure 4B). These results indicate that, in the absence of p140Cap, juvenile hypothalamic GnRH neurons are delayed in their morphological maturation.

Since the sole reduction in the number of GnRH neurons is unlikely to account for the hypofertility observed in p140Cap KO mice (Herbison et al., 2008), we further analyzed the KP system. The KP system, consisting of two groups of KP neurons located in the arcuate nucleus (ARC) and in the rostral periventricular region of the third ventricle (RP3V), has a crucial role in the control of GnRH neurons activity and pulsatility (Clarkson et al., 2009; Harter et al., 2018). Considering the wide distribution of p140Cap expression in the mouse brain, we hypothesized that a defect in the endogenous KP neuroendocrine system could affect the function/activity of GnRH neurons. Thus, we set forth to determine whether alteration of the KP system could account for the reduced function of GnRH neurons in p140Cap KO females. We double-immunostained anatomically matched hypothalamic sections of p140Cap KO and WT female brains for KP and GnRH and determined the number of KP neurons in the RP3V, the density of KP-ir fibers in the ARC, and the number of KP-ir punctae in close apposition to GnRH neurons. In the absence of p140Cap, we observed no changes in the number of KP neurons in the RP3V (Figures 6A,B), in the density of KP-ir fibers in the ARC (Figures 6C,D), and in the number of KP-ir punctae in close apposition to GnRH neurons (Figures 6E,F).

Next, we decided to further probe the immediate responsiveness of GnRH neurons to exogenous KP, as a way to monitor the presence of a normal KP receptor/transduction machinery. To do this, we used an experimental paradigm previously adopted by other authors (Chen et al., 2014; León et al., 2016): we treated P45 p140Cap KO and WT females in the diestrus phase of the estrus cycle with 1 nmol of KP-54 and, after 1 h, we collected the hypothalamus, on which we determined the fraction of GnRH neurons that were immunoreactive for c-Fos in their nuclei, by immunostaining. The results show that, upon treatment with KP-54, p140Cap KO and WT mice had a similar fraction of GnRH-ir/c-Fos-ir double-positive cells (Figures 6G,H), suggesting unaltered responsiveness.

To further sustain this result, we probed for the surge of LH in the blood of p140Cap KO and WT animals, collected 1 h after treatment with KP-54. Also in this case, we did not observe changes in the LH surge in p140Cap KO versus the WT samples (Figure 6I). Together, these results indicate that, in the absence of p140Cap, hypothalamic GnRH neurons show a normal response to exogenous KP, excluding a cell autonomous defect of GnRH neurons at the level of KP receptor or its signal transduction.

Two other key elements controlling the activity of GnRH neurons are the glutamatergic and the GABAergic synaptic stimulation on these neurons. Studies on immortalized cell lines (such as GT1-7 cells) and on the isolated hypothalamus showed that glutamatergic stimulation positively regulates GnRH release (Spergel et al., 1994; El-Etr et al., 2006). Since p140Cap is expressed in both the pre- and the post-synaptic compartments of glutamatergic synapses and has been shown to be essential for effective glutamatergic synaptogenesis and maturation of dendritic spines (Jaworski et al., 2009; Tomasoni et al., 2013; Repetto et al., 2014), we hypothesized that defects in glutamatergic synapses may contribute to the hypofertility of p140Cap KO female mice.

First, we examined the overall density of glutamatergic (VGLUT-ir) punctae in anatomically matched sections of the OVLT and POA by immunostaining on P60 p140Cap KO and WT animals. We found a significant reduction in their density in the absence of p140Cap (Figures 7A,B), indicating that p140Cap is required for efficient glutamatergic synaptogenesis in these areas. Then, we specifically determined the number of VGLUT-ir punctae in close apposition to the soma of GnRH neurons at P10 and P60 in p140Cap KO and WT mice. We observed a significant reduction in the average number of VGLUT-ir punctae per GnRH neuron at both P10 (Figures 7C,D) and P60 (Figures 7E,F) in p140Cap KO animals, as compared to WT. This result suggests that GnRH neurons of p140Cap KO animals are inefficiently innervated by glutamatergic synapses.

We also examined the density of GABAergic (VGAT-ir) punctae in close apposition to the perisomatic surface of GnRH neurons, comparing p140Cap KO and WT hypothalami, but we observed no significant difference at both P10 (Supplementary Figures 5A,B) and P60 (Supplementary Figures 5C,D). Considering that KP neurons represent the primary modulators of GnRH neurons, we finally assessed the number of VGLUT-ir and VGAT-ir punctae in close apposition to KP neurons in the RP3V of p140Cap KO and WT P120 female mice in the proestrus phase of the estrus cycle. We observed no difference in the number of both VGLUT-ir and VGAT-ir punctae comparing the two genotypes (Supplementary Figures 5E,F).