To investigate the olfactory ligands that result in puberty acceleration, we utilized an in vivo bioassay with inbred, BALB/cJ, female mice. After weaning, chemosensory stimuli were directly applied to the nares of female juveniles twice daily for six days (see Materials and Methods for details). On day seven, both uterine and body masses were measured. Consistent with the Vandenbergh effect [5], [9], male urine promotes a significant development of the uterus of BALB/cJ females as indicated by increased wet uterine mass (Fig. 1A, C) without effect on total body mass (Fig. 1B). C57BL/6J females failed to demonstrate uterine weight increase under similar conditions (data not shown), a result which underscores the known influence of genetic background on pheromone-mediated behaviors [33]. The puberty acceleration bioassay in the BALB/cJ strain provides a screen by which putative pheromones and chemosensory neurons may be investigated.

The molecular profile of the puberty accelerating chemosensory neurons has not been identified, but surgical ablations of either the VNO or the main olfactory system have indicated that puberty accelerating pheromones are detected by unknown neurons within the VNO [34]–[36]. The function of TrpC2, the primary sensory channel of VNO neurons, is required to detect and respond to some pheromones [32], [37]–[39]. In order to analyze the role of TrpC2 function in pheromone-mediated uterine weight gain, we fully introduced the TrpC2 mutation [32] into the BALB/cJ genetic background (see Materials and Methods). Littermate progeny of TrpC2+/- (BALB/cJ) breeders were exposed to urine and their uterine weight was assayed blind of genotype (Fig. 1D, E). Subsequent genotyping revealed that TrpC2 null congenic females failed to increase uterine mass in response to the pheromone(s) in male urine (Fig. 1D, E), indicating that TrpC2 function is necessary to detect and respond to pheromones that increase uterine weight.

We next evaluated the bioactivity of pheromones that were previously reported to accelerate puberty: IAA, IBA, HMH, SBT, β-farnesene, EEARSM, and Mups. We first directly investigated the activity of the Mups. Male urine was ultrafiltrated to size fractionate the larger Mup proteins from small molecules. SDS-PAGE analysis revealed that the high molecular weight (HMW) retentate contained detectable proteins (Fig. 2A), which were the expected size of Mups [40]. When analyzed in the bioassay, we found the low molecular weight (LMW) filtrate and whole urine to each increase uterine mass, while the Mup-containing HMW retentate showed no significant bioactivity (Fig. 2B). This finding is consistent with a previous study evaluating the puberty accelerating effect of a recombinant Mup [18]. The absence of protein from LMW fractions and the lack of bioactivity in the HMW fraction argue against the role of Mups as pheromones that accelerate reproductive organ development. All of the remaining previously identified pheromones are smaller than 3000Da and are expected to be present in LMW urinary fractions.

We used three separate approaches to directly evaluate the extent to which other previously identified puberty inducing pheromones increase uterine mass. In all of our experiments, we analyzed the ligands using the concentrations and vehicles previously reported for bioactivity [10], [17], [18], [41]. First, we assessed the sufficiency of each pheromone diluted in water as previously described. Exposure of BALB/cJ females to commercially available synthetic pheromones, β-farnesene, and blended IAA and IBA, failed to significantly increase uterine mass (Fig. 2C). Custom synthesized pheromones, HMH, SBT, and hexapeptide EEARSM also lacked the ability to significantly increase uterine weight when individually tested (Fig. 2D).

Second, while these pheromones were not reported to function as an obligate blend, some invertebrates respond to blends of structurally related sex pheromones with potentiated bioactivity [42], [43]. We therefore evaluated whether the bioactivity of these previously identified pheromones was more apparent as a blend. Mice were exposed to a mixture of IAA, IBA, β-farnesene, SBT, HMH and EEARSM. This complex blend of pheromones failed to increase uterine mass (Fig. 2E). Lastly, SBT has been reported to additionally act as a signaling pheromone, but only when added to urine from castrated males (which alone does not generate bioactivity) [22]. Therefore, we aimed to determine if the previously identified pheromones accelerate puberty in conjunction with other unidentified male odors that may be present in castrated urine. In the first two approaches we observed aversion to IAA and IBA and considering the toxicity of inhaled aliphatic amines in mice [44], these ligands were excluded from this final direct assessment. As previously described [22], we added β-farnesene, SBT, HMH, and EEARSM into urine from castrated males, providing urinary context. However, we found both the urine from castrated males alone and castrated urine supplemented with the blend of pheromones similarly unable to significantly increase uterine weight (Fig. 2F). In total, all three methods of direct analyses of previously identified pheromones did not accelerate BALB/cJ reproductive organ development in our bioassay.

We next evaluated the biochemical characteristics of the activity that accelerates pubertal uterine mass increase in the urine LMW fraction. First, we determined the extent to which pheromones that increase uterine mass are volatile. We lyophilized LMW urine, which is devoid of volatile-stabilizing Mups [18], [27] and found it to retain significant bioactivity (Fig. 3A), confirming that the pheromone is not highly volatile [13], [14]. We next fractionated LMW urine by solvent extraction and both the organic and aqueous phases were screened for the ability to increase uterine mass (Fig. 3B). Hydrophobic molecules extracted into the organic phase failed to demonstrate any bioactivity while the aqueous phase significantly accelerated uterine growth (Fig. 3B), confirming that the pheromone is preferentially aqueous [10], [11].

To further identify basic biochemical properties of the pheromone, we continued to use the bioassay to track the activity through several purification steps and collected the pheromone by solvent precipitation to generate a partially purified pheromone “pellet” fraction (see Materials and Methods for details). Attempts to further isolate the pheromone with routine adsorption to C18 reverse-phase solid phase extraction (RP-SPE) columns were unsuccessful. We found the hydrophilic flow-through which was collected as a batch using C18-high pressure liquid chromatography (HPLC) demonstrated significant bioactivity in contrast to hydrophobic compounds that eluted under higher solvent strength (data not shown). We therefore considered that the pheromone might be adsorbed to hydrophilic matrices using aqueous normal phase-SPE (NP-SPE) and evaluated a small panel of normal phase matrices for eluted bioactivity. We found further fractionation of the pheromone “pellet” over an amino column under aqueous normal phase conditions to recover the puberty acceleration bioactivity (Fig. 3C). The ability to isolate the pheromone by NP-SPE methods and the inability of the pheromone to bind hydrophobic C18 columns indicate that the pheromone is hydrophilic, consistent with our results using solvent extraction (Fig. 3B). Importantly, we find both the NP-SPE fraction and total urine to be similarly effective in promoting uterine weight gain (Fig. 3D). Collectively, this analysis indicates that the biochemical characteristics of ligands in the bioactive fraction (Fig. 3) are inconsistent with those of previously identified pheromones (see Table 1).

The constituents of adult male mouse urine vary significantly between strains [45]–[48] and result in strain-specific differences of male urine to stimulate chemosensory neurons [49], [50], attract females [51]–[53] and trigger ovulation [54]. To evaluate the extent to which pheromones which increase uterine mass are produced by different strains of laboratory mice, we assayed male urine from BALB/cJ, C57BL/6J, and CD-1 lineages. We found total urine from all three strains of mice to each significantly increase uterine mass (Fig. 4A). We next similarly prepared NP-SPE fractions from these individual strains and found them each to contain significant bioactivity (Fig. 4B). This suggests that pheromones which function to accelerate pubertal uterine growth emitted from BALB/cJ, C57BL/6J, and CD-1 strains share similar biochemical characteristics.

To further evaluate the nature of our partially purified bioactivity we directly analyzed urine for the presence of previously known pheromones. The GCMS fragmentation patterns of SBT, HMH, and β–farnesene standards (Fig. 4C) as well as published fragmentation patterns of DHB [55] and α-farnesene were acquired prior to sample analyses. IAA, IBA and EEARSM standards were detected by liquid chromatography (LCMS) (Fig. 4E, F). Using these compound standards and published fragmentation data we analyzed total urine of BALB/cJ, C57BL/6J and CD-1 strains by LC- and GCMS for these previously identified pheromones. We found SBT, IAA, and DHB were readily detectable in the organic extracts of BALB/cJ, C57BL/6J and CD-1 male urine and HMH was detected in samples of BALB/cJ and C57BL/6J male urine (Fig. 4D, E). While β–farnesene, IBA, and EEARSM standards were detected by our methods we were unable to find them present in male mouse urine from any of these strains (Fig. 4C, D, E, F). Importantly, when we performed the same analysis on the bioactive NP-SPE fraction from each strain we were unable to detect any of these previously known pheromones (Fig. 4C, D, E, F). These results suggest that the pheromone enriched in the NP-SPE fraction of urine from several laboratory strains is likely to be novel.

All procedures were approved by the Institutional Animal Care and Use Committee (IACUC) approval number 06-0298. BALB/cJ female mice were obtained from the TSRI breeding colony at 20–22 days of age and weighed. TrpC2 mutant animals [32] were backcrossed with BALB/cJ animals and the genotype was analyzed with a panel of 300 single-nucleotide polymorphisms (SNPs). This was repeated to generateTrpC2 null animals with a 99% BALB/cJ genetic background; a minimum of 3 additional backcrosses were subsequently performed. All TrpC2 mutant animals used in this study were of the BALB/cJ background. Male urine donors (CD-1, BALB/cJ, and C57BL/6J) were housed in groups of 3-5, including at least one stud male. Fresh urine from males 2–6 months of age was collected in metabolic cages and clarified by centrifugation prior to use. All male urine was from the CD-1 strain unless noted otherwise. Castrated male urine was collected from adult CD-1 males that were castrated at 3–4 weeks of age. Urine was used within 1–48 hrs of collection, kept at 4°C when stored overnight and warmed to ambient temperature prior to use. Upon weaning, female mice were individually caged in a room without males. All mice were maintained in 7.5×11×5” polysulfone cages, ventilated at 60 ACH with a 12∶12 h light/dark cycle (on at 0700 h) in a temperature-controlled room and kept on the Harlan Teklad LM-455 diet.

Prior to each experiment, mice were individually housed. Animals were assigned to weight-matched group (one-way ANOVA, p>0.9). Beginning at 22–24 days of age, BALB/cJ females that weighed 12.0–15.0 g body weight, were treated twice daily with variable chemosensory stimuli painted on the external nares with a synthetic paintbrush dipped into a tube containing 0.5 ml of stimulus (except for IAA and IBA, detailed below). Male urine was collected from mice of the CD-1 strain unless otherwise indicated. To apply the stimulus, mice were restrained by hand in a supine position and slowly stroked with the stimulus until it bubbled at the nostrils, stroked again, and replaced in the cage. On average, cages were disturbed for 13 seconds and mice were handled for 6 seconds, during which ∼15 µl were applied, for each of 12 total exposures. Females were observed to be behaviorally aversive to direct application of 50 mM IAA or IBA (KAF unpublished observations) therefore; these stimuli were applied to a 0.75cm cotton ball and left in the cage as previously described [17]. On day seven, at 28–30 days of age, mice were euthanized and body weight was recorded. Animals were approximately 17 g at the end of each assay. Each uterus was rapidly dissected, removed of fat, and weighed. The progeny of TrpC2^+/- breeders were assayed blind of genotype and tail samples were collected for genotyping.

Mice were genotyped by multiplex PCR. To identify the wild-type TrpC2 gene, we used two primers to amplify the pore region of the ion channel: sense, 5′-GGCCATCTTCTTTTACATATGC, and antisense 5′-GAAACAAGGGAATGCAGG. The deleted gene was replaced with a PGK-neo cassette in the mutant mice [32] which was amplified using the following primers: 5′-CCTGATCGACAAGACCGGCTTC and 5′-GTGTTCCGGCTGTCAGCGCA.

To separate HMW from LMW urinary constituents, fresh urine was passed through Amicon Ultra-4 centrifugal filters with a cutoff of 3 kDa. Following the transfer of the LMW filtrate, the retentate was washed 2–3 times with 3–4 ml PBS and raised to original volume. Putative pheromones were dissolved as previously described. IAA and IBA (Sigma) were suspended in water at 0.05M [17]. β-farnesene (Bedoukian) was suspended in water at 250ppm (∼1.2 mM) [18], [41]. Racemic SBT [2] was suspended in water at 2.6 ppm (∼18 µM) [18]. HMH (a generous gift from Milos Novotny) was suspended in water at 2000 ppm (∼13.7 mM) [18]. The NH₂-EEARSM hexapeptide (TSRI Peptide Synthesis Core Facility, 95% purity) was dissolved in phosphate buffered saline (PBS) at 0.5 mg/ml (∼690 µM) [10]. Pheromones were blended at these concentrations, except for isoamyl- and isobutylamine, which were provided on a cottonball inside the cage as previously described [17].

10 µl crude urine, HMW or LMW urinary fractions were desalted with G-50 spin columns (GE-Amersham) and loaded alongside standard ladder on a 10% acrylamide gel. Following electrophoresis, the gel was stained with Coomassie blue.

To remove urinary volatiles, LMW urine was flash-frozen with liquid nitrogen, lyophilized under vacuum pressure for 24–48 hrs and stored at 4°C until use. Organic extraction of LMW urine was conducted with rigorous shaking in a glass separatory funnel using two volumes chloroform and one volume methanol. Each phase was collected and evaporated under vacuum pressure until chloroform was removed. In order to isolate the pheromone for column chromatography, LMW urine was precipitated with four volumes methanol on ice for 30 min. Following centrifugation, both fractions were dried, resuspended to original volumes, and assayed for the ability to increase uterine weight. While the precipitate had no bioactivity, the methanol solutes significantly accelerated uterine hypertrophy (p<0.001, Tukey's HSD). Thereafter, the precipitant was discarded and the supernatant was further fractionated with five volumes ice cold acetone on ice for 1 hr. Following centrifugation, each fraction was dried, resuspended to original volume, and tested in the bioassay. We found the supernatant to be inactive and acetone precipitate to significantly accelerate uterine hypertrophy (p<0.01, Tukey's HSD). Thereafter, this bioactive “pellet” was dried and stored at −80°C for up to two months. To further isolate the pheromone with NP-SPE, approximately 20 mg of material (obtained from 1 ml crude urine) were suspended in 250 µl 50% acetonitrile and applied to a 500 mg NH₂ column (Supelco), equilibrated with acetonitrile. Several experiments were conducted to determine wash conditions that enabled us to retrieve bioactivity from the column (unpublished observations) and we found 10–15 ml 80% acetonitrile, 10% methanol to be optimal. The wash was discarded. The NP-SPE fraction was eluted with 1.0 ml water and stored at 4° up to two days. Samples were dried in a speed vacuum. All concentrated samples were reconstituted to original volume with water or PBS prior to use.

For GCMS, 0.5 ml chloroform was added to 0.5 ml fresh urine or the NP-SPE elution in a glass vial, vortexed, and stored overnight at −20°C. The lower organic phase was removed from thawed samples and stored at −20°C until analysis that day. 1 µl samples were analyzed by splitless injection onto a 30 m×250 um ID HP5ms column with He₂ carrier gas at a 1.2 ml·min⁻¹ flow-rate. The temperature program began at 50°C, held for 5 minutes, ramped to 300°C at 10°C·min⁻¹ and held for 5 minutes. MS scans were from m/z 50–700 with a 5-minute solvent delay. For LCMS-MS, NP-SPE samples and pheromone standards were analyzed by quadrupole time-of-flight (QTOF) and a triple quadrupole mass spectrometer by single multiple reaction monitoring (SMRM). For QTOF analysis, 5 µl (∼5%) of samples were resolved online with a Sequant ZIC-HILIC column, 2.1 mm×150 mm, using a 50 mM NH₄OAc (buffer A)/acetonitrile (ACN, buffer B) gradient (Time 0 = 10∶90, 3 = 50∶50, 5 = 50∶50, 7 = 85∶15, 10 =  off), followed by 7 min re-equilibration at 200 µl·min⁻¹. The QTOF conditions included a drying gas flow rate of 9 L·min⁻¹ with nebulizer pressure of 25 psi and a 100V fragmentor. The first quad was set to isolate on m/z values of 74.1 and 88.1. Collision energy was set to 15V and product ions of IAA and IBA were targeted with positive mode scans of m/z 50–200. The EEARSM peptide was analyzed by triple quad LCMS using SMRM. 5 µl samples were resolved with the Sequant ZIC-HILIC column, using a 10% ACN, 15 mM NH₄OAc (buffer A)/90% ACN, 15 mM NH₄OAc (buffer B) gradient (Time  = 0∶100, 2 = 0∶100, 10 = 50∶50, 15 = 75∶25, 18 =  off), followed by 5 min re-equilibration at 250 µl·min⁻¹. The EEARSM 722.3 ->464.4 transition state was produced with a 250 V fragmentor and a 32V collision energy.

All data is graphed as box plots, providing the mean, median, and ranges of data, prior to data analysis. Whiskers span the approximated 5–95% range (usually the minima and maxima) and the box encompasses the 25–75% inner-quartile range, bisected by the median line. The mean is represented by a dot. Body weights and log₁₀-transformed uterine weights [9], [18] were analyzed by a Student's t-test or by one-way ANOVA and compared pairwise using Dunnett's (using PBS as the statistical control) or Tukey's HSD (all pairwise comparisons) test as indicated.