The past decade has seen rapid progress in deciphering the essential role of the mammalian accessory olfactory system in chemical communication and the regulation of social behaviors [12, 57], but the functional significance of the anatomical and molecular segregation into VSNs that express either of two different G protein subunits, Gαo or Gαi2, has not been resolved. Elucidating the specific behavioral roles of each of these VSN populations is required to understand whether and how these subsystems work together to represent the sensory environment and how exactly they control behavioral responses .
To address these questions, we have developed a mouse strain that harbors a conditional deletion of the Gnao1 gene under the control of the promoter of the Omp gene . We have shown that behavioral responses that depend on a functional main olfactory system are normal in these mice and that the mutation has no impact on Gαo expression in OMP-positive neurons that are present in some reproductive CNS brain nuclei. We observed no obvious changes in the amount and distribution of OMP-positive cells in hypothalamic and amygdaloid nuclei in Gαo mutants and there was no obvious overlap between Gαo and OMP immunoreactivity in both B6 mice and a mouse line expressing an OMPCre-eRosa26τGFP reporter. Furthermore, blood estradiol and progesterone levels, ovary morphology, and general fertility parameters were all normal in the Gαo mutants. Thus, our conditional Gαo-mutant mice constitute an appropriate model to examine the role of Gαo-expressing VSNs in pheromone-stimulated behavioral responses of female mice. Deletion of critical signaling molecules such as Gαo, Trpc2, and Gγ8 causes a significant reduction in the number of basal VNO neurons [13, 41, 58]. Given that we do not know whether this cell loss reflects loss of specific V2R-expressing VSNs, we cannot fully rule out the possibility that some of the phenotypes identified here reflect dominant or neomorphic phenotypes. Nonetheless, our results clearly define important functions of Gαo-expressing VSNs in different behavioral repertoires.
A key result of this report is that Gαo signaling impacts on a much wider range of pheromone-dependent behaviors than previously anticipated. For example, our results reveal an unexpected delay in the initiation of puberty (Figure 1C,D) and an altered estrous cycle (Figure 2) in the mutant mice, even without active stimulus presentation. This suggests that sensory input via Gαo-expressing VSNs is required for the normal display of puberty onset and the regulation of ovulatory signals . Furthermore, selective ablation of Gαo conferred insensitivity to male urinary pheromones that facilitate the display of female reproductive behaviors: urine-stimulated puberty acceleration was defective in Gαo-mutant females (Figure 1A,B). The chemical nature of the puberty-accelerating pheromone(s) is still unclear but at least two reports have linked puberty acceleration to MUPs or MUP-derived peptides [33, 34]. MUPs activate basal V2R-positive VSNs and their VNO detection is lost following Gαo deletion [22, 41]. Thus, these studies are consistent with our findings showing defective puberty acceleration in Gαo mutants. On the other hand, several reports have indicated a role for small organic molecules in puberty acceleration [15, 37–39] and some of these molecules are known to activate apical V1R/Gαi2-expressing VSNs which function normally in the Gαo mutants [41, 43]. One possible explanation for these seemingly divergent results is that Gαo- and Gαi2-expressing subsets of VSNs could both be involved in these effects. Such a scenario is not without precedent: the display of male territorial aggression and maternal aggression also seems to depend on the activation of both Gαi2- and Gαo-expressing VSNs [36, 41, 59]. Besides a lack of effect of male urine to induce uterine growth, we observed that Gαo-mutant females showed larger uteri in the absence of stimulation (Figure 1B). One potential explanation for this result is that the basal VSNs are required for the Lee-Boot effect  in which female urine may suppress uterine maturation. If so, in the absence of suppression, uterine weight will increase regardless of the stimulation. However, Gαo mutants displayed a delayed first estrus (Figure 1C) and unstimulated adults did not show more frequent estrous cycles (Figure 2), as would be expected in mice deficient for the Lee-Boot effect. Therefore, we cannot currently confirm a direct dependency of this effect on Gαo signaling.
Another surprising finding was the critical role of Gαo signaling in pheromone-induced estrus induction in adult mice. Interestingly, Gαo ablation not only abolished male urine-induced estrus induction, but also seemed to cause a reduction of days in estrus and proestrus (Figure 2C). This result suggests that Gαo-mutant females are not entirely unresponsive to estrus-modifying pheromones but the functional outcome of such chemosignals is altered, perhaps as a result of defective processing or integration with other pheromonal cues. Consistent with this possibility, small organic molecules such as 2-sec-butyl-4,5-dihydrothiazole, dehydro-exo-brevicomin, and α- and β-farnesenes have estrus-inducing effects in mice [61, 62]. These cues are present in male urine, are known to activate VSNs of the apical VNO neuroepithelium , and thus should still be detectable in the absence of Gαo.
Conditional deletion of Gαo also has severe consequences on female sexual receptivity, that is, lordosis behavior. Two measures of lordosis, lordosis quotient and number of females showing lordosis, indicated that this pheromone-stimulated behavioral response was absent or strongly diminished in the mutant mice (Figure 4). Thus, intact Gαo signaling is essential for this innate, female-specific sexual display. These results are consistent with studies demonstrating that the Vmn2r116 receptor is involved in lordosis behavior  and that detection of ESP1 is severely reduced in VSNs lacking Gαo . The fact that lordosis induced by exposure to both B6 and BALB/c mice was diminished in the Gαo mutants indicates that, besides ESP1 and Vmn2r116, other pheromones and V2R receptors are probably involved in this behavior because B6 mice do not secrete ESP1 . We cannot yet completely rule out that the cycling phenotype as observed here impacts on lordosis but, as there was no evidence for ovarian or hormonal imbalance in our experiments (Figure 3), major effects of the cycling phenotype on lordosis seem unlikely.
Importantly, Gαo-mutant females were not only defective in a variety of pheromone-stimulated innate behaviors but also in learned social responses to pheromones. Employing an established paradigm to assess mate recognition (Figure 6), our results provide direct evidence in support of a model in which Gαo-positive VSNs are critically involved in the detection of molecular cues related to genomic individuality. Scent ownership recognition experiments demonstrated directly that this test required contact to chemical cues present in the HMW fraction of urine (Figure 6C,D), consistent with a proposed role for MUPs in this function [3, 35]. Preference for individual male scents requires an associative learning step to provide a linkage between information contained in the volatile and the nonvolatile HMW urinary fractions; we demonstrated here that this learning requires intact Gαo signaling (Figure 6B-D). Of note, Gαo-mutant females could still discriminate the urine of two different males in a habituation-dishabituation test (Figure 6E) and showed a preference for male versus female urine in a two-choice test (Figure 6F), indicating that olfactory discrimination abilities were normal in these mice. Furthermore, defective scent ownership recognition was not due to a loss of gender discrimination: Gαo-mutant females did not display male-typical mating behaviors toward other conspecifics (Figure 5). Such indiscriminate mounting has been reported previously in mice deficient in the cation channel Trpc2 [13, 14, 16]. One of these studies  employed a large open arena, but it is unclear whether the behavioral apparatus impacts on the display of male-like behaviors in Trpc2 mutants.
We were unable to observe a second pheromone-dependent learning paradigm, the Bruce effect, in Gαo-mutant females. Near-maximum non-pregnancy rates occurred with exposure to familiar cues or even without any additional stimulus exposure. We cannot yet determine whether this reflects a failure of the mutant mice to discriminate familiar from unfamiliar cues or whether other deficits such as poor mating performance (lordosis), shorter receptive periods (estrus), and loss of mate recognition capabilities influence the outcome of this test. Most likely, the low pregnancy rates reflect a combination of all of these defects. Remarkably, cGαo-/- mice exhibited high variability on first litter latency, eventually expanding to values of up to 60 days (Figure 3E), which could be consistent with a potential fertility defect. However, on average, first litter latencies and other fertility values in the null mice were not significantly different from the controls. As an explanation for the apparent contradiction between low Bruce effect performance and normal fertility parameters, we believe that the sum of the described reproductive deficiencies may remain unnoticed in a laboratory environment: during the Bruce effect test, males and females are mated for just 24 h in contrast to the fertility monitoring in which breeding pairs remain in permanent contact. However, Gαo-mutant females would be unlikely to stay competitive under natural conditions where animals are subject to time-limited sexual encounters and where optimal reproductive performance is essential for reproductive success.