Zfp260 choreographs the early stage osteo-lineage commitment of skeletal stem cells

Animal models

Igs2^{em1(CAG-LSL-ZsGreen-wpre-pA-CAG-RSR-tdTomato-wpre-pA)Smoc} (H11^{LSL-ZsGreen-RSR-tdTomato}) (NM-KI-200319), Ctsk^{em1(2A-CreERT2-WPRE-polyA)Smoc} (Ctsk^creER) (NM-KI-200067) with C57BL/6 J background were purchased from the Shanghai Model Organisms Center, China. C57BL/6J-Zfp260^em1Cflox/Cya (Zfp260^flox/flox) (S-CKO-09613) was obtained from Cyagen, USA, with its construction strategy displayed in sFig. 8. All mice were used for analysis regardless of sex. All mice were housed at the Laboratory Animal Center of Tongji University with constant temperature (22 ± 2°C) and humidity (55 ± 10%) in a 12-h light cycle. All animal studies were approved by the Institutional Animal Care and Use Committee of Tongji University.

In vivo modeling and tamoxifen induction

All mice were used for analysis regardless of sex. For descriptive and functional studies, 7wk-old mice were randomly assigned into different experimental groups. For the conditional genetical deletion of Zfp260 or labeling of Ctsk⁺ periosteal stromal cells, Zfp260^fl/fl, Ctsk^creERZfp260^fl/fl, Ctsk^creERH11^{LSL-ZsGreen-RSR-tdTomato}, Ctsk^creERH11^{LSL-ZsGreen-RSR-tdTomato}Zfp260^fl/fl mice were inraperitoneally administered with Tamoxifen (50 mg/kg) (T5648, Merck) for 3 consecutive days at 7wk-old in accordance with the experimental design. Mice were modeling at 8wk-old under the anesthesia of 2% isoflurane, followed by sterilization. For MSFL, the surgery was performed based on the following processes⁴. After making the median section, the 1/2 tungsten steel round bur was applied to drill holes in the center of the nasal bone without breaking the nasal membrane. A mini implant (WEGO) with diameter of 0.6 mm and length of 0.8 mm was implanted, covered by a 5 mm × 5 mm*0.1 mm Teflon film. For fracture, a full fracture was created at the mid-shaft and fixed by the insertion of a sterilized 22-gauge needle into the medullary cavity³³.

Bulk RNA sequencing and analysis

For the RNA sequencing of FACS-sorted cells (Lin^–ZsGreen⁺ cells of Ctrl or Zfp260-cKO groups derived from the fracture callus or the newly formed tissues in MSFL), total RNA was extracted with their integrity numbers >7.0, measured by Agilent 2100 Bioanalyzer. Similar to our previous study, the library was constructed in accordance with the instructions of the low input Trio RNA-Seq^TM library preparation Kit (0357-32, Nugen), with the insert length varied around 350 bp. Novaseq 6000 was used for high throughput sequencing with the strategy of PE150. To acquire the transcriptomic features of the full fracture healing process, the dataset GSE152677 (ctrl, 1 d, 3 d, 5 d, 7 d, 14 d) was applied and further analyzed in this study³⁴. As for the MSFL healing process, the laser microdissection (LCM) based bulk-seq dataset CRA007231 applied in our previous research was re-analyzed in this study⁴. For data processing, the raw fastq files were trimmed and aligned using Hisat (https://www.psc.edu/resources/software/hisat-2) with the mm10 genome and nomarlized to TPM³⁵. Mfuzz (http://mfuzz.sysbiolab.eu) was used for time course analysis³⁶. Cytoscape 3.7.1 (https://cytoscape.org/index.html) was used to depict the protein–protein interaction network. Principle component analysis (PCA) was applied to displayed the correlations among different cell types.

Sample selection, library preparation and single cell RNA sequencing (scRNA-seq)

To identify key transcription factors specifically expressed in the osteogenic lineage cells at the single-cell level, we selected scRNA-seq data (Ctrl and Fracture-PFD14) published by Kishor et al. (GSE154247), which excluded CD45⁺ and CD117⁺ cells from the homeostatic and post-fracture samples were chosen³⁷. As data excluded hematopoietic and lymphoid cells from the callus, we further sorted CD31⁺ and CD45⁺ cells (lymphoid, and endothelial cells) and Lin^–CD140α⁺ cells (stromal cells) from the mid-shaft of the femur in 8-week-old mice using flow cytometry, followed by scRNA-sequencing (detailed procedures in the ‘Flow Cytometry Sorting’ section). For the MSFL samples, based on previous studies⁴, we performed single-cell sequencing on the digested tissues of the nasal respiratory mucosa and the newly formed region at 3, 6, 9, and 14 days after MSFL surgery (CRA007231). For the aforementioned CD31⁺ and CD45⁺ cells and Lin^– CD140α⁺ cells of the femur, we utilized the Chromium Next GEM Single Cell 3’ Kit version 3.1 (1000075, 10x Genomics) for scRNA-seq, following the manufacturer’s instructions. In brief, single live cells were resuspended in PBS containing 0.04% BSA (A1933, Sigma-Aldrich) to a final concentration of 500–1200 cells/ml, as determined by a TC20 cell counter (BioRad). Ten thousand cells were encapsulated in droplets to generate nanoliter-scale gel beads in EMulsion (GEMs). GEMs were then reverse transcribed in a C1000 Touch Thermal Cycler (Bio-Rad) programmed at 53 °C for 45 min, 85 °C for 5 min, and held at 4 °C. After reverse transcription and cell barcoding, emulsions were broken and cDNA was isolated and purified with Cleanup Mix containing DynaBeads (37002D, Thermo Fisher Scientific) and SPRIselect reagent (B23318, Beckman), followed by PCR amplification. The amplified cDNA was then used for 3’ gene expression library construction, and single-cell RNA libraries were sequenced using an Illumina NovaSeq 6000 sequencer with 150 bp paired-end reads.

scRNA-seq data analysis

Consistent with our previous study, the Cell Ranger toolkit v3.1 (10x Genomics) was applied to aggregate raw data, filter low-quality reads, align reads to the mouse reference genome (mm10), assign cell barcodes, and generate a unique molecular identifier (UMI) matrix. The raw UMI matrix was processed to exclude genes detected in fewer than 10 cells and cells with fewer than 200 genes. We further quantified the number of genes and UMI counts for each cell, maintaining high-quality cells with thresholds of 500-120,000 UMIs and 400-8000 genes. To exclude cells with a high mitochondrial proportion, we used the default parameter of 20%, ensuring that most of the heterogeneous cell types were included for downstream analysis. To minimize the effect of mitochondrial genes on subsequent analysis, a regression was performed prior to clustering. Scrublet³⁸ (https://github.com/AllonKleinLab/scrublet) was applied to remove potential doublets, with an expected doublet rate of 6%, and cells with doublet scores above 90% were excluded. Batch normalization was performed using CCA2, which models the UMI counts with a regularized negative binomial model to remove sequencing depth variation, while adjusting variance by pooling information across genes with similar abundances. Dimension reduction and unsupervised clustering were conducted following the workflow of the Python toolkit Scanpy³⁹. Briefly, dispersion-based methods were employed to detect the top 20 highly variable genes (HVGs). The normalized dispersion was obtained by scaling with the mean and standard deviation of the dispersions for genes. Percentages of mitochondrial gene counts and cell cycle-related gene counts were regressed out from the normalized expression matrices. For clustering 56,781 cells (Fracture) and 63,527 cells (MSFL) in total, PCA was performed on the variable gene matrix to reduce noise. Fifty components were calculated, and the top 40 PCs were used for downstream analyses⁴⁰. Then, the Leiden algorithm⁴¹ was applied to identify cell clusters, with a resolution parameter of 1, referring to the construction of the TISCH database⁴². Notably, the same principal components were also used for non-linear dimension reduction to generate the Uniform Manifold Approximation and Projection (UMAP) for visualization.

Following the first round of unsupervised clustering, we annotated each cluster based on their marker genes. In the fracture dataset, the major cell types included stem cells and progenitors (Stem/Pro), osteo-lineage cells (OLC), osteochondro-lineage cells (OCLC), smooth muscle cells (SMC), monocytes or macrophages (Mono/Mac), endothelial cells (EC), B cells, T cells, erythroblasts (Ery), neutrophils (Neu), eosinophils or basophils (Eos/Baso) and megakaryocyte (Mega). For MSFL, the major cell types remained consistent with previous research, including chondrocytes, osteo-lineage cells (OLCs), pericytes, neural progenitor cells (NPCs), neutrophils, macrophages, epithelial cells (ECs), T cells, sustantacular cells (Sus), horizontal basal cells (HBC), immature olfactory sensory neurons (iOSN) and mature olfactory sensory neurons (mOSN). To depict the curve of gene variations with cell hierarchy from a higher resolution perspective, the OLC cluster in MSFL, along with the Stem&Pro, OLC and OCLC clusters in the fracture dataset, underwent a second round of clustering, following the same process as in the first round. The hierarchical subclusters were annotated based on key markers indicating osteogenesis (Alpl, Col1a2, Runx2, Sp7, Dmp1), chondrogenesis (Sox9, Col2a1, Acan) and stemness (Klf4, Prrx1, Ctsk).

To explain the connectivity between cell populations and reveal the pathways of cell differentiation, trajectory from osteoprogenitors to osteocytes in MSFL and from skeletal stem cells to mature osteoblasts in the fracture dataset were inferred using PAGA with RNA velocity-directed edges, utilizing the scVelo toolkit⁴³. To calculate velocities, the ‘Dynamic modeling’ was applied with default settings⁴⁴. Cells were organized based on the predicted trajectory using the ‘velocity pseudotime’ function, with the starting point located in the farthest cells within the osteoprogenitor cluster in MSFL or skeletal stem cell cluster in the fracture dataset compared to the other clusters.

Histology

The tissue treatment process was detailed in our previous study⁴⁵. Briefly, tissues were fixed in 4% PFA for 24 h and decalcified in 10% EDTA solution for 2 weeks at 4 °C. For paraffin embedded nasal specimens, the blocks were sectioned into 4 μm slices, followed by deparaffinized in xylene and gradient rehydration before staining. For cryosectioning of fracture samples, the specimens were dehydrated overnight in 30% sucrose and sectioned into 10 μm slices. Masson’s trichrome (G1340, Solarbio) and Toluidine blue staining (E670105, Sangon) were performed using commercial staining kits following the manufacturer’s instructions. Images of regular staining were captured using a Nikon ECLIPSE Ci-E microscope with a 10x objective.

Multiplex immunohistochemistry and confocal imaging

Multiplex immunohistochemistry staining was performed using the TSA-based system (Thermo Scientific) as reported previously⁴. Briefly, after antigen retrieval, peroxidase, and non-specific signal blockages, multicolor IHC commenced with sequential incubation of different primary antibodies, HRP modified secondary antibodies, and TSA reagents with different fluoresceins, followed by the stripping of the previous round of antibody and incubation of the next one. The primary antibodies used in mIHC (dilution 1:400 for all antibodies) included goat anti-mouse/human/rat Itgav (AF1219, Novus Biologicals), mouse anti-mouse/rat CD90 (NB100-65543, Novus Biologicals), mouse anti-mouse/human CD105 (NBP2-22122, Novus Biologicals), rabbit anti-human/mouse/rat CD200 (AF2724, Novus Biologicals), rabbit anti-mouse/human/rat Runx2 (ab236639, Abcam), rabbit anti-mouse/human/rat Sox9 (ab185966, Abcam), rabbit anti-mouse/human Alpl (MA5-24845, Invitrogen), rabbit anti-mouse/human/rat Zfp260 (ABE295, Merck), mouse anti-human/mouse/rat p300 (NB100-616, Novus Biologicals), rabbit anti-human/mouse MED1 (NB100-2574, Novus Biologicals), rabbit anti-human/mouse BRD4 (NBP2-76393, Novus Biologicals), mouse anti-human/mouse/rat Prkca (NB600-201, Novus Biologicals), rabbit anti-V5 tag (13202, CST), mouse anti-Collagen type I (67288-1-Ig, proteintech), rabbit anti-Collagen type II (28459-1-AP, proteintech). The tyramide reagents used in mIHC included AF350 (B40952, Invitrogen), AF488 (B40953, Invitrogen), AF546 (B40954, Invitrogen), AF594 (B40957, Invitrogen) and AF647 (B40958, Invitrogen). Images were acquired using a Leica SP8 confocal microscopy with the adjustable wavelength receiving module to separate the spectrum of the channels and avoid overlap.

Image analysis

Imagej 1.54 h was applied for the analysis of multicolor images. To visually demonstrate the co-staining regions of transcription factors Zfp260, Runx2, and Sox9 with SSC lineage cells in morphology, the mIHC was applied. Due to the difficulties in distinguishing cell clusters with multiple merging channels, the clusters were segmented in the bone calluses of fracture and newly formed tissues of MSFL based on the marker panels reported in the previous researches, including SSC (Itgav⁺THY1^–CD200⁺CD105^–), BCSP (including pre-BCSP (Itgav⁺THY1^–CD200^–CD105^–) and BCSP (Itgav⁺THY1^–CD105⁺)), THY (Itgav⁺THY1⁺CD200^–CD105⁺), BLSP (Itgav⁺THY1⁺CD105^–), PCP (Itgav⁺THY1⁺CD200⁺CD105⁺) and their co-staining regions with Zfp260, Runx2, and Sox9 signals. At the final presentation, Itgav⁺ cells were colored in gray and clusters of SSC/BCSP/THY&BLSP/PCP were labeled in purple, with Zfp260/Runx2/Sox9 marked in yellow. The operating pipeline was as follows. Each channel image was adjusted to 8-bit. The threshold of each channel was adjusted to obtain the positive staining area, stored as the positive ROI. The ‘Invert’ command was used to select the negative staining area, followed by the recording of negative ROI. By stacking ROI regions of certain channels, the ROI of a specific cluster was obtained and displayed by using the command ‘creating mask’. Signal areas smaller than 5×5 dpi were excluded as non-cellular noise, with the remaining signal region colored by gray, purple, or yellow as previously mentioned. To calculate the co-concurrence percentage of each transcription factor in each cluster, the areas of co-staining TF and certain clusters were measured separately using the ‘measure’ command, followed by the calculation with the formula Area_TF&Cluster/Area_Cluster. The percentage of three different regions was evaluated in each mIHC staining image. Similarly, the co-staining percentage of ZsGreen⁺ cells labeled by rAAV9 with Alpl and Runx2 or Krt14 and Runx2 was calculated.

To calculate the nuclear translocation efficiency of Zfp260 or transfected Zfp260-V5, the nuclear or cytosolic region were selected, with respective ROI established. After measuring the MOI values of the above regions, the values were substituted into the formula MOI_nucleus/MOI_cytosol to calculate the translocation efficacy. Randomly selected 30 cells of each group were calculated.

For the co-localization analysis of Zfp260, Brd4, p300, and Med1 within the cell nucleus, the ‘straight line’ tool was applied to select ROI. Within each signal channel, the command ‘plot profile’ was applied to obtain a matrix table of fluorescence intensity changes along the selected ROI for each channel. The matrix table of each channel was plotted as a curve with distance (μm) on the x-axis and fluorescence intensity on the y-axis.

Flow cytometry cell sorting or analysis

For the analysis of SSC lineage cells, the callus of fracture mice (5 mice per biological repeat, n = 2 per time point), along with the newly formed tissues of MSFL models (15 mice per biological repeat, n = 1 per time point) were dissected under stereomicroscopy. For scRNA-seq, the mid-shaft of mice femurs (5 8-week-old mice) were dissected. The tissues were digested in DPBS containing 0.3% collagenase, 0.3% dispase II and 0.01% DNase I at 37 °C for 30 min in a shaking bath, followed by filtering with 70 μm cell strainers (352350, BD Falcon). After centrifugation at 4 °C at 300 × g for 10 min, the supernatant was discarded, and cells were suspended in red blood cell lysis buffer (00433357, eBioscience) at 4 °C for 5 min. After neutralization and centrifugation, the cell pellets were resuspended by staining buffer containing the CD16/32 blocking antibody (156603, Biolegend) and incubated for 30 min, followed by staining with antibodies. All antibodies were diluted at 1:200 except for the anti-CD45 antibody, which was diluted at 1:800. Antibodies used included: CD140a (Pdgfra)-BV605 (APA5, Biolegend), CD31-PE/Dazzle^TM 594 (MEC13.3, Biolegend), CD45-PE/Dazzle^TM 594 (30-F11, Biolegend), TER119-PE/Dazzle^TM 594 (TER-119, Biolegend), CD105-APC/Cy7 (MJ7/18, Biolegend), CD200-APC (OX-90, Biolegend), CD51 (Itgav)-biotin (RMV-7, Biolegend), BV421-Streptavidin (405226, Biolegend), Ly-51-PE/Cy7 (6C3, Biolegend), CD90.1 (Thy1.1)-PerCP/Cy5.5 (OX-7, Biolegend), CD90.2 (Thy1.2)-PerCP/Cy5.5 (30-H12, Biolegend). Zombie Aqua^TM was applied to distinguish live and dead cells. BD FACSAria^TM III was applied for cell sorting, while the BD-LSRFortessa was used for FACS analysis. The gating strategy for mouse skeletal stem cell lineages referred to previous research, with signal compensation performed on FMO controls. For sorting Lin^–ZsGreen⁺ cells from Ctsk^creER;H11^{RSE-tdTomato-LSL-zsGreen} and Ctsk^creER;H11^{RSE-tdTomato-LSL-zsGreen};Zfp260^fl/fl modeling mice, the cells were gated from single live cells, Lin^– cells and ZsGreen⁺ cells. For scRNA-sequencing, single live Lin^–CD140α⁺ cells and single live CD31⁺ and CD45⁺ cells were sorted and applied for the subsequent sequencing separately.

Micro-CT analysis

Mice modeled by fracture and MSFL from different treatment groups were sacrificed and perfused with 4% PFA. Following an additional 24 h of fixation, their femurs and nasal bones were dissected for the subsequent radiological analysis without decalcification. Two-dimensional (2D) images of fracture scanning sections, reconstructed 3D images of whole femurs or newly formed tissues of MSFL, and calculated structural indices were obtained using Micro-CT 50 (Scanco Medical) with a resolution of 10 μm. Quantitative analysis of the callus by fracture and newly formed tissues by MSFL were performed using the associated analyzing software. Abbreviations used for morphometric parameters were as follows: BMD, bone mineral density; BV, bone volume.

Isolation, culture and treatment of periosteal stem cells (PSCs)

PSCs were isolated from intact femur, tibia and maxillofacial bones from 8wk-old Ctsk^creER;H11^{RSE-tdTomato-LSL-zsGreen} and Ctsk^creER;H11^{RSE-tdTomato-LSL-zsGreen};Zfp260^fl/fl mice. The mice were labeled with Tamoxifen (50 mg/kg) (T5648, Merck) for 3 consecutive days at 7 weeks of age. Adherent soft tissues were removed and the periosteum was scratched and digested in DPBS containing 0.3% collagenase, 0.3% dispase II, and 0.01% DNase I at 37 °C for 30 min in a shaking bath. Following digestion, FACS cell sorting was performed to isolate single live Lin^–ZsGreen⁺Itgav⁺THY1^–6C3^–CD105^–CD200⁺ cells. After a two-week culture period, all clones were selected and passaged to P1. The P1 PSCs were applied for all experiments, except for co-immunoprecipitation or ChIP in this study.

Due to the necessary quantity of cells, PSCs for co-immunoprecipitation or ChIP would undergo an immortalization-based amplification approach. The P1 PSCs were infected by lentivirus expressing human telomerase reverse transcriptase (hTERT). The cells were then further expanded to satisfy the quantity need for the subsequent pull-down and ChIP assays.

GO 6976, a potent PKCalpha inhibitor, was used for the inhibition experiments. PSCs were pretreated with 1 μM GO 6976 (HY-10183, MCE) 6 h before osteogenic induction in the experimental group, with an equal amount of DMSO supplemented in the control group.

For osteogenic induction, PSCs were cultured in osteogenic induction medium (α-MEM containing 10% FBS, 10 nM dexamethasone, 50 μg/mL ascorbic acid, 1% penicillin/streptomycin and 10 mM β-glycerophosphate). Alizarin Red S staining was performed after 14 or 21 days of induction. For the observation of nuclear translocation and immunoprecipitation, induction durations of 6 or 24 h were administered according to the experimental design.

Total RNA extraction and quantitative real-time PCR (qRT-PCR)

Total RNA was extracted from FACS-sorted cell clusters (SSC, BCSP, THY + BCSP and PCP) using centrifuge tubes, following the manufacturer’s instructions (12183016, Invitrogen). For in vitro cultured PSCs, RNA extraction was performed using TRIzol (15596018, Invitrogen) as previously described. First strand cDNA synthesis was applied using the RevertAid First Strand cDNA Synthesis Kit (K1622, Thermo Scientific). Quantitative real-time PCR (qRT-PCR) was performed using PowerTrack^TM SYBR Green Master Mix (A46012, Applied Biosystems). Detection was conducted on a QuantStudio 7 Flex (Thermo Scientific), with β-Actin as the internal control. The primer sequences were listed in the Supplementary Data 1.

Chromatin immunoprecipitation (ChIP)

PSCs with different treatments were subjected to ChIP according to the manufacturer’s protocol of Pierce Magnetic ChIP kit (26157, Thermo Scientific). Briefly, cells were crosslinked with 1% formaldehyde for 10 min, followed by quenching with 125 mM glycine for 5 min. Cells were then washed three times with ice-cold PBS, and the pellets were collected. Cell lysis was performed for 10 min on ice. Nuclei were collected by centrifugation and digested with MNase to ensure DNA fragments ranged from 150 bp to ~1000 bp. Soluble chromatin was immunoprecipitated overnight with magnetic beads with antibodies for V5 (ab15828, Abcam), H3K27ac(8173, CST), or Brd4 (13440, CST), along with appropriate isotype controls. For each reaction, 4 μg of antibody was used. Immunoprecipitated beads were collected, followed by protein digestion and DNA cleanup.

For wild-type PSCs transfected with Zfp260-V5 and immunoprecipitated with anti-V5, DNA-library construction was performed using the VAHTS Universal DNA Library Prep Kit for Illumina V4 (ND610, Vazyme) according to the manufacturer’s protocol. Novaseq 6000 was applied for high-throughput sequencing with a strategy of PE150. For wild-type and Zfp260^-/- PSCs, immunoprecipitation was carried out using anti-H3K27ac or anti-Brd4 antibodies. For Zfp260^-/- PSCs transfected with Zfp260-V5 and its mutants (Y173F, S182G, S197G, Y173F-S182G, Y173F-S197G, S182G-S197G and triple mutants), immunoprecipitation was performed with anti-H3K27ac, anti-Brd4, or anti-V5 antibodies. Quantitative PCR (qPCR) was then performed on purified ChIP and input DNAs at target loci, with enrichment compared to respective isotype control IgG. The primer sequences for target loci were listed in the Supplementary Data 1.

CUT & Tag

FACS sorted Lin^–zsGreen⁺ cells from fracture or MSFL models were collected for the Cut & Tag workflow using the Commercial CUT&Tag-IT Assay Kit (53160, Active Motif), following the manufacturer’s protocol. Brief, the FACS sorted cells were centrifuged, resuspended, and captured by ConA beads. Cells were permeabilized with 0.05% digitonin and then incubated overnight with primary antibodies against H3K27ac (8173, CST) or H3K4me1 (5326, CST). A guinea pig anti-rabbit secondary antibody (provided in the kit) was diluted in 1:100 and used for incubation. After a 1-h incubation for assembled pA-Tn5 transposomes binding, the tagmentation buffer was administered to fragment the DNA for 1 h. The fragmented DNA was then enriched and purified for PCR amplification using i5 and i7 index primers for Illumina sequencing. Novaseq 6000 was applied for high-throughput sequencing with a PE150 strategy.

ATAC-seq

FACS sorted cells were used for ATAC-seq with library construction following the manufacturer’s constructions (TD711, Vazyme). Briefly, the cells were collected, washed and lysed using 50 μL of lysis buffer (10 mM Tris-HCl, pH 7.4, 10 mM NaCl, 3 mM MgCl2, and 0.02% digitonin, 0.1% NP-40) on ice for 10 min. For the fragmentation process, 50 μL of fragmentation buffer containing 10 μL 5X TTBL and 4 μL TTE mix V50 was added to the lysed cells. The mixture was incubated at 37 °C for 30 min. The reaction was stopped by adding 5 μL stop buffer at room temperature for 5 min. After fragmentation, DNA was extracted and the library was amplified, followed by size selection with AMPure XP beads (20498100, Beckman) ranging from 200 to 800 bp. All libraries were sequenced on the Illumina NovaSeq 6000 platform.

Analysis of ChIP-seq, CUT&Tag and ATAC-seq

The raw data from ChIP-seq, CUT&Tag, ATAC-seq, and the downloaded dataset (GSM6245143) were analyzed using the HiChIP pipeline⁴⁶. Briefly, FastQC (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/) and Cutadapt (https://cutadapt.readthedocs.io/en/stable/) were used for quality control and adaptor trimming, respectively. Paired-end reads were mapped to the mm10 genome reference using Bowtie2 (https://bowtie-bio.sourceforge.net/bowtie2/index.shtml), followed by duplicate removal with Sambamba⁴⁷. Narrow peaks for H3K4me1, H3K27ac, Brd4, and Zfp260-V5 were identified using the model-based analysis of ChIP-seq (MACS2) (https://pypi.org/project/MACS2/) package with a cutoff of 1.00e-05. For data visualization, BEDTools⁴⁸ and custom scripts were used to generate per-million read density profiles with a 200-bp window size and a 20-bp step size. Signal tracks were visualized using Integrative Genomics Viewer (IGV) software.

The number of reads in the TSS ± 3 kb region of all protein-coding genes was estimated and normalized to 10 million (RP10M), with log2 transformation and quantile normalization applied to each of the ChIP-seq, CUT&Tag, and ATAC-seq libraries. The read density (RPM, reads per milion) over the TSS ± 3 kb region was calculated using the ngs.plot tool (v2.02)⁴⁹. The binding motif of Zfp260-V5 was identified using MEME, and peak-annotated genes were used for the enrichment analysis of Gene Ontology-Biological Process (GO-BP) (http://geneontology.org/) and the Kyoto Encyclopedia of Genes and Genomes (KEGG) (https://www.genome.jp/kegg/).

Super enhancer calling was performed using H3K27ac data from both control and Zfp260-cKO Lin^–zsGreen⁺ cells sorted from fracture and MSFL models. Using the ROSE tool (v1.0.0, http://younglab.wi.mit.edu/super_enhancer_code.html)⁵⁰, super enhancers were identified based on default parameters, including the merging of remaining peaks within a distance of 12.5 kb or less. Compared to the Zfp260-cKO group, the differential super enhancers identified in the control groups were used for GO-BP enrichment analysis.

GST pull-down, mass spectrometry and analysis

The coding sequence of mus musculus Zfp260 was cloned and fused into the pGEX6p-1-GST plasmid using the In-Fusion strategy (639650, Clontech). The pGEX-GST and pGEX-GST-Zfp260 plasmids were transformed into BL21 E. coli, and target protein synthesis was induced by 1 mM IPTG at 16 °C for 24 h. The E.coli pellets were collected, washed, and subjected to thorough ultrasonication and lysis. Centrifugation was performed to separate bacterial debris from the bait protein-containing supernatant. The glutathione magnetic agarose beads (78602, Thermo Scientific) were equilibrated and used to capture GST and GST-Zfp260 in the supernatants for 3 h. The beads were then washed and incubated with whole cell lysates of PSCs at 4 °C overnight. The bound proteins from each group were eluted and visualized by silver staining (24600, Thermo Scientific). The peptides were enzymatically hydrolyzed, extracted, and purified for separation by EASY-nLC 1000 (Thermo Scientific) using an analytical column (C18, 1.9 μm, 75 μm × 20 cm) at a flow rate of 200 nL/min. For mass spectrometry analysis, the Orbitrap Fusion Lumos (Thermo Scientific) was used in Data Dependent Acquisition (DDA) mode.

Identification of peptides and phospho-modifications by database searching were performed by using Proteome Discoverer 2.4 software for subsequent analysis.

Plasmids construction and lentivirus packaging

In vitro manipulation of Zfp260 expression was achieved through lentivirus infection. To overexpress Zfp260, the pCDH-Zfp260-V5 lentivirus expression vector was constructed by cloning and fusing the CDS of mus musculus Zfp260 into the MCS site (EcoRI and BamHI) of the pCDH-V5-Puro plasmid via the In-Fusion strategy (639650, Clontech). For lentivirus packaging, a three-plasmid system (pol/gag, VSV-G and pCDH-V5/pCDH-Zfp260-V5) were transfected into the 293 T cell line (H4-1401, Cyagen Biosciences). The lentivirus was concentrated and purified by ultracentrifugation. The aforementioned viruses were used to infect wild-type PSCs with an MOI of 100. The infected PSCs from each group were applied for subsequent tests according to the experimental design.

To establish point mutations based on pCDH-Zfp260-V5, the QuickMutation Plus point mutation kit (D0208, Beyotime) was applied with primers designed following the manufacturer’s protocol. Taking pCDH-Zfp260-V5 as the template, pCDH-Y173F-V5 (518A-T), pCDH-S182G-V5 (544A-G), pCDH-S197G-V5 (589A-G), pCDH-Y173F-S182G-V5 (518A-T; 544A-G), pCDH-Y173F-S197G-V5 (518A-T; 589A-G), pCDH-S182G-S197G-V5 (544A-G; 589A-G), and pCDH-triple-mut-V5 (518A-T; 544A-G; 589A-G) were constructed for subsequent lentivirus packaging as mentioned above. The aforementioned viruses were used to infect Zfp260^-/- PSCs with an MOI of 100, and subsequent analysis was carried out according to the experimental design.

Plasmids construction and packaging of Adeno-associated virus-9 (rAAV9)

In the in vivo rescue experiment, the pAAV-CMV-MCS-IRES-ZsGreen vector was modified to achieve PSC-specific expression⁵¹. The CMV promoter and enhancer sequences of the original vector were removed by cutting at the NheI and PstI restriction sites. Using high-fidelity enzyme (P510, Vazyme), the promoter sequence of the mouse Ctsk gene (TSS + 2 kb) was cloned into the site via In-Fusion strategy (639650, Clontech), resulting in the pAAV-Ctsk-MCS-IRES-ZsGreen vector. The promoter sequence was listed in the Supplementary Data 1. Using the aforementioned 8 pCDH vectors (pCDH-Zfp260-V5, pCDH-Y173F-V5, pCDH-S182G-V5, pCDH-S197G-V5, pCDH-Y173F-S182G-V5, pCDH-Y173F-S197G-V5, pCDH-S182G-S197G-V5, and pCDH-triple-mut-V5) as templates, the wild-type and all mutant types of Zfp260 coding sequences were cloned into the MCS site, with the Kozak sequence inserted ahead of the CDS region to enhance downstream gene transcription. AAV packaging was performed using the AAV-293T cell line (632273, Takara). The pAAV-RC9 and pAAV-helper packaging vectors were mixed with the pAAV expression vectors (a total of 8 groups) to form a triple-plasmid system. PEI transfection reagents (HY-K2014, MCE) were applied for transfection when cell confluency reached 80%-90%. Cells and supernatant were collected 72 h after transfection. Virus extraction and purification were performed using the AAVpro® Purification kit (6666, Takara) according to the manufacturer’s protocol, followed by titer determination. In the in vivo experiment, each mouse was intravenously administered the rAAV9 at a dose of 4 × 10¹¹GC (2 × 10¹³GC/kg). Two weeks after administration, the infection efficiency in the heart, liver, spleen, lung, kidney, skull, jaw and legs was evaluated using IVIS® Spectrum In Vivo Imaging System (PerkinElmer) in the EGFP channel.

Phos-assay SDS-PAGE

Phos-Assay SDS-PAGE⁵² was performed to detect phosphorylated modifications of Zfp260. An 8% SDS-PAGE gel was prepared by adding 5 mM Phos-assay and 10 mM ZnCl₂ to final concentrations of 25 μM and 50 μM, respectively. Electrophoresis was conducted under a constant current of 30 mA. For MS analysis, Coomassie Brilliant Blue (CBB) staining was performed. For Western immunoblotting, gels were pre-treated with 10 mM EDTA and washed before membrane transfer.

Immunoprecipitation (IP) and western immunoblotting (WB)

The whole cell lysates (WCL) of different cell types subjected to various treatments were prepared according to the study design. To extract WCL, cells were washed twice with ice-cold PBS and lysed with ice-cold IP lysis buffer supplemented with 1 mM PMSF and a protease inhibitor cocktail at the working concentration. The cell lysates were incubated on ice for 30 min before centrifugation at 13,000 × g for 10 min at 4 °C.

For immunoprecipitation, a pre-clear step was performed using Pierce Protein A/G Magnetic Beads (88802, Thermo Scientific) for 2 h at 4 °C before the formal IP process. The magnetic beads were then discarded, and the pre-cleared supernatants from each group, containing equal amounts of protein were subjected to the formal IP process with specific antibodies through overnight incubation, followed by magenetic bead capture.

The detailed procedure for Western immunoblotting has been described in previous research⁵³. Antibodies used in this study at a 1:1000 dilution included: mouse anti-p300 (NB100-616, Novus Biologicals), rabbit anti-MED1 (NB100-2574, Novus Biologicals), rabbit anti-BRD4 (NBP2-76393, Novus Biologicals), rabbit anti-V5 tag (13202, CST), rabbit anti-Flag tag (14793, CST), mouse anti-Prkca (NB600-201, Novus Biologicals), rabbit anti-β Actin (4970, CST), and rabbit anti-Histone H3 (4499, CST). To avoid interference from light and heavy chains, the VeriBlot for IP Detection reagent (conjugated with HRP, ab131366, Abcam) was used at a 1:200 dilution for the detection of immunoprecipitated samples. For input samples, conventional HRP-conjugated secondary antibodies specific to the corresponding species were diluted 1:2000 for WB signal detection.

The protein structure modeling, docking and molecular dynamic simulations

The protein structure files involved in this study were obtained from the PDB database or predicted using AlphaFold2⁵⁴. The proteins were dehydrated and hydrogenated using Pymol (http://www.pymol.org/pymol), including wild-type Zfp260 (AF2, AF-Q62513-F1, average pLDDT Score=80.05). Additionally, the structural models of seven Zfp260 mutants were predicted online using AlphaFold2, with their highest-ranked prediction model selected for subsequent analysis. The average pLDDT values for these mutants were all greater than 80.

For the docking of Zfp260 (wild-type or mutant) with Prkca, the GRAMM-X server was applied⁵⁵. The conformation with the lowest binding energy between Zfp260 and Prkca was analyzed and displayed, and the shortest distance between the amino acids 173, 182 or 197 of Zfp260 and the Prkca catalytic domain (328-668 amino acids) was calculated.

Renal capsule transplantation

As previously described³, eight-week-old male C57BL6/J mice were anesthetized and shaved on the left flank and abdomen prior to the sterilization of the surgical site. A 1-cm incision was made to externalize the kidney, and a 2-mm pocket was created in the renal capsule. A 5-µl Matrigel (Corning, 356231) containing 8,000 cells was implanted beneath the capsule, and the opening was sealed using a cauterizer before repositioning the kidney into the body cavity. The animals were euthanized after four weeks. For FACS, the kidneys were dissected and subjected to digestion in accordance with the part “Flow cytometry cell sorting or analysis”. To assess bone formation, the kidneys were fixed in 4% PFA for six hours and subsequently analyzed using micro-CT.

Statistical analysis

No methods were used to predetermine sample size, and no blinding or randomization was employed for data analysis. Data were presented as means ± SD, with the number of independent experiments indicated in the legends. Independent two-tailed Student’s t test, one-way ANOVA, or two-way ANOVA was applied to analyze the data after determining whether the data were normally distributed. SPSS 26.0 (IBM) was used for all statistical calculations. Significant differences were considered at p values below 0.05 and were indicated in the figures.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

• SEO Powered Content & PR Distribution. Get Amplified Today.
• PlatoData.Network Vertical Generative Ai. Empower Yourself. Access Here.
• PlatoAiStream. Web3 Intelligence. Knowledge Amplified. Access Here.
• PlatoESG. Carbon, CleanTech, Energy, Environment, Solar, Waste Management. Access Here.
• PlatoHealth. Biotech and Clinical Trials Intelligence. Access Here.
• Source: https://www.nature.com/articles/s41467-024-54640-0