A miR-29a/b₁ knockout mouse line was established using CRISPR/Cas9 gene editing technology and was supplied by Shanghai Center for Model Organisms (SMOC) (17). miR-29a/b₁ ^−/− homozygous animals and their wild-type littermates were obtained by mating corresponding heterozygotes with each other. Genomic DNA was extracted from tail biopsies, using magnetic bead DNA isolation Kit (DE0596D, EmerTher, Shanghai). PCR was adopted for genotyping using 2 × Taq Plus Master Mix (P212-01, Vazyme) under the following conditions: denaturation at 98°C for 2 minutes, then 35 cycles of 98°C for 10 seconds, annealing at 63°C for 15 seconds, and extension at 68°C for 60 seconds. Primers used for genotyping are listed in Table S1 .

All animals were housed in a specific pathogen-free environment (12 h light/12 h dark with lights on at 7.00 h at 21 ± 2°C) with food and water ad libitum. This study was performed in strict accordance with institutional guidelines and approved by the Institutional Animal Care and Use Committee of Shanghai Model Organisms, and the IACUC permit number is 20090002.

8-week-old miR-29a/b₁ KO and wild-type virgin female or male mice were bred with wild-type male or female mice with known fertility at a proportion of ♀2: ♂1, and vaginal plug formation was examined every morning for 20 consecutive days. Pregnant female mice were separated and pups were recorded, while non-pregnant mice continued to mate. Female or male mice that did not conceive within 1 month of mating were defined as infertile.

Female miR-29a/b₁ KO and wild-type mice from the age of 3-week-old were examined twice daily with respect to on vaginal opening as a marker of rodent sexual maturity. The date of vaginal opening in each mouse was recorded. Female miR-29a/b₁ KO and wild-type (8-10 weeks old for each genotype) mice were caged individually for 3 weeks and at least two full estrous cycles were obtained in each mouse.

Vaginal smears were collected daily and the determination of estrous cycle was evaluated microscopically with the vaginal epithelium. The vaginal epithelium obtained from the vaginal opening by gently eluting 10 μl of physiological saline solution 2-4 times, then the vaginal epithelium transferred onto a microscopic slide and dried at room temperature and fixed with 100% methanol. The slides were stained with Wright’s Giemsa (BASO) stain and examined with light microscopy. Proestrus cells are well-formed nucleated epithelial cells. Animals with 85% superficial epithelial cells were considered to be estrus. During metestrus, cornified squamous epithelial cells often in fragments, as well as leukocytes, may be observed. Otherwise, the predominant presence of leukocytes in the cytological smear was identified as diestrus.

Adult (8‐10 weeks) miR-29a/b₁ KO and control females in diestrus morning were injected subcutaneously with pentobarbital (effective dose 320 mg/kg). Mice were deeply anesthetized and placed on a heating pad. The back skin was shaved and cleaned. About 1.0 cm long incision was made through the muscle layer above the ovaries on each side of the midline. Through the incision, the ovaries were gently pulled outside the body and removed by cauterization below the oviduct. The skin incision was closed with sutures. The mice were left to recover on a heating pad. Adult sham-operated mice were in diestrus on the day of recording as determined by vaginal cytology. Sham‐treated animals were processed in the same way, except for the intact ovaries retained. Mice were killed 7 days post-surgery, and their serum were measured for LH and FSH levels.

For hormone measurement, orbital blood was collected in the morning (10.00 h-11.00 h) and evening (18.00 h-19.00h) (18) from freely‐moving conscious animals during randomly estrous cycle stages, and were kept at room temperature for 30 minutes. Serum was obtained by centrifuging for 15 minutes at 3000 g at 4°C and was stored at -80°C until analysis. Serum levels of hormone and pituitary proteins LH level were analyzed by Shanghai WESTANG BIO-TECH cooperation using enzyme-linked immunosorbent assay (ELISA). The minimum detectable level of the LH assay was 0.1mIU/ml and the intra-and inter-assay coefficients of variation were 9.9% and 8.3%. The minimum detectable level of the FSH assay was 1mIU/ml and the intra-and inter-assay coefficients of variation were 9.8% and 8.6%, The minimum detectable level of the estrogen assay was 30 pg/ml and the intra-and inter-assay coefficients of variation were 9.3% and 8.5%, The minimum detectable level of the progesterone assay was 0.2ng/ml and the intra-and inter-assay coefficients of variation were 9.5% and 8.3%, The minimum detectable level of the testosterone assay was 0.1ng/ml and the intra-and inter-assay coefficients of variation were 9.4% and 8.2%, respectively.

Animals received an intraperitoneally injection with 125ng/g (19) exogenous GnRH (L7134, Sigma-Aldrich, St Louis, MO, USA) or saline vehicle. Twenty minutes after GnRH or saline injection, orbital blood was collected, and the resultant serum samples were stored at −80°C for subsequent human-LH radioimmunoassay (RIA, performed by Beijing North Institute Biological Technology, Beijing, China) (20), with sensitivity and intra- and inter-assay coefficient of variation for LH of 0.5 mIU/ml, 15% and 20%, respectively.

Superovulation: To induce superovulation, 8-week-old mice were intraperitoneally injected with 5 IU pregnant mare serum gonadotropin (PMSG, Sigma) at afternoon (15:00h-16:00 h), followed by 5 IU human chorionic gonadotropin (hCG, Sigma) 48 hour later to trigger oocyte maturation and ovulation. Female mice were mated with 10-week-old fertile wild-type males 16 h after injection and checked for vaginal plug formation the next morning.

Oocyte collection: Super ovulated or natural mated mice with a visible plug were sacrificed by cervical dislocation, the ovaries were removed and the ampulla was collected. Oocytes were harvested in M2 media and quantified by microscopy (Nikon SMZ800) following brief digestion in hyaluronidase (800IU/ml, Sigma) to strip cumulus and pipetting for 30-60 s. Oocytes were washed 5 times with PBS. The washed oocytes were transferred to M16 media and cultured overnight, and two-cell stage embryos were counted in the next morning.

Total pituitary (P) protein from wild-type and miR-29a/b₁ KO females (8 weeks, n=3) were isolated and labelled with iTRAQ reagents 114, 115, 116, 117, 118, 119, 120 or 121, respectively, followed by Liquid Chromatography with tandem mass spectrometry (LC-MS/MS) (Shanghai Wayen Biotechnologies Inc.). The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE (21) partner repository with the dataset identifier PXD017106.

Wild-type and miR-29a/b₁ KO were euthanized and transcardially perfused with cold saline, followed by 4% paraformaldehyde 0.1 M phosphate buffer (PFA). Brain, pituitary and ovary were collected and fixed overnight at 4°C. Paraffin-embedded ovary samples were serially sectioned at 4 μm-thick sections. Brain coronal (20 μm) slices were cut with a Leica CM1950 following by dehydration in 30% sucrose saline solution.

Pituitary and ovary stained with hematoxylin and eosin using standard histological techniques (Servicebio). Stained sections were scanned using LEICA CTR6000 with a 10X, 20X and 40X objective. Ovarian follicles at different developmental stages were classified and quantified in serial sections according to the Pedersen and Peters method (22). To avoid double counting of follicles across sections, only follicles containing oocyte with a clearly visible nucleus were scored (23), and follicles were counted in every fifth serial section. Any follicle also appearing in the adjacent lookup section was not counted. The entire section was analyzed without subsampling. Each ovary was coded with no information about genotype group for blind counters and prevent bias. The mean count per section was calculated. All follicle types were summed together to determine the total number of follicles.

For immunohistochemistry, sections were subjected to antigen retrieval by incubation in 10 mM sodium citrate, pH 7.0, for 10 minutes at 95°C. The endogenous peroxidase activity of the sections was quenched with 3% H₂O₂ treatment (Sangon Biotech, Shanghai). Immunohistochemical staining was performed using mouse anti-Lutropin beta antibody (1:500, SANTA CRUZ, sc-373941) or rabbit anti-GnRHR antibody (24, 25) (1:100, Proteintech, 19950-1-AP) and HRP-conjugated donkey anti-mouse IgG (1:1000, ThermoFisher, A16017) or donkey anti-rabbit IgG (1:1000, ThermoFisher, A16035) for lutropin and GnRHR antibody.

For immunofluorescence, brain and pituitary sections permeabilized by incubation with 0.1% Triton X-100 in PBS for 10 minutes at room temperature. After permeabilization, the sections were washed three times in PBST, and blocked with 5% normal donkey serum in PBS for 1h at room temperature, then were incubated with mouse anti-Lutropin beta antibody (1:1000, SANTA CRUZ, sc-373941), rabbit anti-GnRH1 (1:500, Immunostar, PA1-121) or rabbit anti-GnRHR antibody (26) (1:100, Proteintech, 19950-1-AP) overnight at 4°C, followed by staining with Alexa Fluor 647-conjugated donkey anti-mouse antibody (Invitrogen Molecular Probes) or Alexa Fluor 594-conjugated donkey anti-rabbit (Invitrogen Molecular Probes) antibody and DAPI dye to stain nuclei. The mouse liver and lung tissues were selected to negative control for GnRHR and Lutropin beta antibody respectively.

Stained sections were scanned using the 40X objective of a Zeiss Confocal microscope (LSM880). The area fractions of positive cells relative to entire area were determined using ImageJ (Fiji, NIH) software. Cell location was mapped to the atlas (27).

Total RNA was isolated using TRIzol (Tiangen Biotech, Beijing) according to the manufacturer’s instructions and kept at -80°C subsequent for use. For microRNA measurement, 2 μg total RNA was transcribed into cDNA using the miRcute Plus miRNA First-Strand cDNA Synthesis Kit (Tiangen Biotech, Beijing). Expression level of mature miR-29a, miR-29b and miR-29c were measured using miRcute Plus miRNA qPCR Detection (Tiangen Biotech, Beijing). U6 snRNA was used for normalization.

For mRNA measurement, total RNA (2 μg) from each sample was transcribed by using EasyScript First-Strand cDNA Synthesis SuperMix (TransGen Biotech, Beijing), and mRNA levels of target genes were detected using TransStart Tip Green qPCR SuperMix (TransGen Biotech, Beijing) according to the manufacturer’s instructions. Murine β-actin was used as a reference to normalize target gene expression levels. Real-time PCR amplification was performed using the Realplex system (Applied Biosystems QuantStudio3, ThermoFisher Scientific). The sequences of the specific primers used are listed in Supplementary Material, Table S1 . RNA levels were calculated using the 2^−ΔCT method, where CT is the cycle threshold (28). Melting curve analysis for each primer set revealed only one peak for each product, and the sizes of PCR products were confirmed by comparing sizes with a commercial ladder after agarose gel electrophoresis. PCR products were further confirmed by sequencing.

Mice were euthanized and tissues were collected. Total tissue protein was extracted using RIPA buffer (ThermoFisher scientific) containing protease and phosphatase inhibitor cocktails (Selleck Chemicals). Protein concentration was quantified using the Enhanced BCA Protein Assay Kit (Beyotime). Protein (20 μg) from each sample was separated on 4%-20% SDS-PAGE (GenScript) and transferred onto nitrocellulose membranes (GE Healthcare). Membranes were blocked with Western BLoT Blocking Buffer (Protein Free) (Takara) for 1 h at room temperature and then incubated with primary antibodies, Lutropin beta (1:1000, SANTA CRUZ, sc-373941) or anti-β-actin (1:1000, Santa Cruz, sc-47778) diluted in Western BLoT Immuno Booster PF (Takara) at 4°C overnight. After washing with TBST three times, membranes were incubated with fluorescent-conjugated secondary antibody for 1 h (1:10000, LI-COR Biosciences). Quantitative detection of protein expression was then performed using the Odyssey Infrared Imaging system (LI-COR Biosciences) and analyzed with Image J software (National Institutes of Health, Bethesda, MD, USA).

Data analysis was performed using GraphPad Prism 7 (GraphPad software Inc.). Data are expressed as the mean ± SEM. Difference in mean values between two groups were analyzed using the Student’s t-test (continuous variables) or Mann–Whitney test (discrete variables). For comparisons involving more than two groups, ANOVA (continuous variables) or Kruskal–Wallis (discrete variables) with post hoc testing was used, and survival profiles were constructed by Kaplan-Meyer survival analysis. Statistically significant differences are shown with asterisks (*p < 0.05, **p < 0.01, ***p < 0.001, and **** p < 0.0001).

A conventional miR-29a/b₁ knockout mouse line (miR-29a/b₁ KO) was previously established using CRISPR/Cas9 methods (17). The genotyping and the expression of miR-29 in different genotypes of mice were detected by PCR and real-time PCR, respectively ( Figure S1 ). To understand the role of miR-29a/b₁ in fertility, the reproductive ability of miR-29a/b₁ KO mice was evaluated. For data in Table 1 , of the 25 females tested, 23 were sterile. The two pregnant miR-29a/b₁ KO female mice gave birth to two offspring each and were not subsequently pregnant again. Among males, 66.7% miR-29a/b₁ KO were still fertile. However, female miR-29a/b₁ KO mice exhibited serious reproductive problems. Vaginal plugs were checked to study the mating behavior. miR-29a/b₁ KO females had a significant lower mating frequency compared to wild-type females ( Table 1 ), suggesting abnormal sexual maturity and estrous cycle. Sexual maturity indicated by vaginal opening occurred 5 days later in miR-29a/b₁ KO female mice (postnatal day 28) compared to wild-type littermates (postnatal day 23) ( Figure 1A ). At the time of puberty onset, the mutant mice are significantly lighter than wild-type mice ( Figure 1B ). Meanwhile, abnormal estrous cycle with less time in estrus and metestrus and significantly more time in diestrus was observed in miR-29a/b₁ KO female mice. ( Figures 1B, C and Figure S2 ). RT-PCR analysis revealed that expression of miR-29a periodically changed in pituitary and ovarian tissues ( Figure S3A ), suggesting that miR-29a/b₁ may play a role in the estrous period in mammals. Taken together, these data illustrate that loss of miR-29a/b₁ induces growth retardation in mutant mice and subfertility in females.

Ovary and uteri weight in miR-29a/b₁ KO females were significantly reduced compared to wild-type females ( Figures 2A, B ), whereas, in males, testis and seminal pouch in miR-29a/b₁ KO mice and wild-type counterparts showed no difference ( Figure S4 ). Fertilized eggs were collected from the oviducts of wild-type and miR-29a/b₁ KO females with vaginal plug after mating with wild-type males. In 20 females miR-29a/b₁ KO mice, only 2 oocytes were found and with no two-cell embryos the next day, while among five wild-type mice, 34 oocytes and 19 two-cell embryos were collected ( Figure 3A ). Histomorphometric analysis revealed that mutant ovaries contained normal primordial follicles, a similar number of secondary follicles with normal oocyte and a thick granulosa cell layer, indicating that the early follicles developed normally, but lacked corpora lutea formation ( Figures 3B–D ). These results suggest that subfertility of the mutant female mice may be caused by an ovulation disorder.

In females, hormonal control of the estrous cycle and ovulation is essential for the establishment of maturation and fertility in mammals (29). Thus, we examined hormone levels in the serum of wild-type and miR-29a/b₁ KO female mice. In the female miR-29a/b₁ KO mice, significant decreases in the serum LH and progesterone (P₄) ( Figure 4A ) were observed, while there was no apparent difference in serum content of follicle-stimulating hormone (FSH) or Testosterone (T) or Estradiol (E₂) compared to wild-type mice ( Figures S5A–C ). Cyp19a₁ and Cyp17a₁ , encoding enzymes involved in estradiol and testosterone synthesis, were expressed at identical levels in ovaries from the two groups of mice, while the Cyp11a, which essential to the level of sex hormones, was significantly decreased in ovaries from mutant mice ( Figure S5D ). These results indicated that impaired corpora lutea formation in miR-29a/b₁ KO mice might be caused by a shortage of LH. This speculation was further confirmed by the superovulation experiment. Ovulation in the mutant mice was rescued by exogenous gonadotropin injection, indicating that responses to LH stimulation were not irreversibly lost in these mutant animals ( Figure 4B ). Ovaries from superovulated adult miR-29a/b₁ KO mice showed normal morphology, and the corpora lutea were formed ( Figure 4C ).

To determine whether the central regulated mechanisms mediating ovulation were altered in miR-29a/b₁ KO mice, females were subsequently treated with an intraperitoneally injection of 125ng/g GnRH or saline vehicle at 10.00 AM. Normal GnRH responsiveness was observed in miR-29a/b₁ KO pituitary, but serum LH level in miR-29a/b₁ KO females remained markedly below the levels observed in wild-type littermates ( Figure 4D ). Furthermore, The GnRHR-immunoreactivity in the pituitary of miR-29a/b₁ KO mice was increased compared to wild-type mice ( Figures 5A, B ), Again, to assess the impact of hyperstimulation with endogenous GnRH modulated by estrogen (30–33), female control and miR-29a/b₁ KO animals were castrated or underwent a sham surgery. Animals were euthanized after 7 days, and serum concentrations of LH and FSH were measured. Consistent with control females, castration resulted in an increase in both LH and FSH compared with sham-operated controls, however, the post-castration rise in LH secretion was blocked in miR-29a/b₁ KO females, while the FSH level was no significant differences in mutant mice serum from controls ( Figure 5C ). LH levels overall were markable lower in miR-29a/b₁ KO females relative to controls. There was no apparent difference in Kiss1and Gnrh1, which stimulating secretion of gonadotropin releasing hormone from the hypothalamus (34–37) and luteinizing hormone from the pituitary (35), respectively ( Figures 5D, E ). These results suggest that ovulation disorder in miR-29a/b₁ KO mice might be caused by dysregulation of related pituitary hormones, especially LH.

LH is synthesized in and secreted by the pituitary. A lack of miR-29a/b₁ was confirmed in mutant pituitary tissues ( Figure S3B ). The anterior pituitary undergoes rapid proliferation in neonatal mice, subsequently expanding the cells that produce factors required for growth and reproduction (38). Defective anterior pituitary development in animals contributes to many organism-level developmental defects (39). However, there was no difference in pituitary structure, size or position of the anterior pituitary between wild-type and miR-29a/b₁ KO mice ( Figure 6A ). No abnormalities were found upon pathological examination of mutant pituitary tissues ( Figure 6B ). Notably, transcript levels of the Lhβ gene in miR-29a/b₁ KO pituitary did not differ from control animals, but LH protein level and immunoreactivity were even higher in KO mice ( Figures 6C–F ).

To further elucidate the effects of miR-29a/b₁ gene knockout on pituitary function, iTRAQ analysis was performed to compare proteomic changes in the pituitary between mutant and wild-type mice. Total pituitary protein from three biological replicates of each genotype were subjected to LC-MS/MS analysis. The hierarchical clustering profile of differential proteins is shown in the heat map ( Figure 7A ). A total of 163 cellular proteins were statistically significant altered (p<0.05), including 75 upregulated proteins and 88 downregulated proteins ( Figure 7B , Table 2 ). LHβ and FSHβ were significantly increased in the pituitary of miR-29a/b₁ KO mice according to b/y ion signal intensity ( Figure 7C ). Besides, TSHβ and Cga were also markedly upregulated as a result of the miR-29a/b₁ deficiency ( Figures 7C, G ). Through GO analysis, altered proteins identified in this study were found to be involved in a wide range of biological process, and most of the differential proteins were classified in the protein transport processes, which are essential for vesicle-mediated transport in the cytoplasm and exocytosis during plasma infusion (40–42) ( Figure 7D ).

The intersected gene between upregulated expression and miR-29a targets through miRDB (http://mirdb.org) were analyzed, 11 potential direct target transcripts of miR-29a were discovered ( Figure 7E ), and predicted target genes were expected to be upregulated in miRNA loss of-function models ( Figure 7G ). Among them, collagen family Col1a1, Col4a2 and Col5a1 are target genes of miR-29a-3p, and promote cancer cells invasion and migration (43–45). In addition, miR-29a can promote the neurite outgrowth by targeting extracellular matrix-related genes like Fibrillin 1 (Fbn1) and hyaluronan and proteoglycan link protein 1 (Hapln1) (46, 47), which dramatically increased in the pituitary of miR-29a/b₁ KO mice. Hdac4 (48), which is key epigenetic modified writer, may play important roles in the change of gene expression pattern in miR-29a/b₁ gene knockout mice, especially for down-regulated genes. For the 88 down-regulated proteins in the pituitary of miR-29a/b₁ KO mice, a considerable portion of them participate in vesicle-mediated transport and secretion (Ergic1, Fkbp2, Ssr3, Stat5a, Crhbp, Figure 7F ). Notably, Ergic1, encodes a cycling membrane protein, and plays an important role in transport between endoplasmic reticulum and Golgi (49). Absence of trApγ (SSr3) impairs protein translocation into the endoplasmic reticulum and affects transport (50). Myosins were reported as core players in the final stages of regulated secretory pathways (51). Treatment of pituitary cells with the myosin light chain (Myl2/3) kinase inhibitor, wortmannin, attenuated GnRH-induced LH release (52). Further validated by quantitative PCR (qPCR) that the mRNA transcripts of these genes, which were consistent with LC-MS/MS ( Figure 7G ). These results indicated the deficiency of miR-29a/b₁ blocked proteins transportation, leading to impaired pituitary hormone secretion, especially LH released.