In this protocol, the translating ribosome affinity purification (TRAP) method and the isolation of nuclei tagged in specific cell types (INTACT) method were optimized for the paired interrogation of the cell-specific ovarian transcriptome and epigenome using the NuTRAP mouse model crossed to a Cyp17a1-Cre mouse line.
Assessing cell-type-specific epigenomic and transcriptomic changes are key to understanding ovarian aging. To this end, the optimization of the translating ribosome affinity purification (TRAP) method and the isolation of nuclei tagged in specific cell types (INTACT) method was performed for the subsequent paired interrogation of the cell-specific ovarian transcriptome and epigenome using a novel transgenic NuTRAP mouse model. The expression of the NuTRAP allele is under the control of a floxed STOP cassette and can be targeted to specific ovarian cell types using promoter-specific Cre lines. Since recent studies have implicated ovarian stromal cells in driving premature aging phenotypes, the NuTRAP expression system was targeted to stromal cells using a Cyp17a1-Cre driver. The induction of the NuTRAP construct was specific to ovarian stromal fibroblasts, and sufficient DNA and RNA for sequencing studies were obtained from a single ovary. The NuTRAP model and methods presented here can be used to study any ovarian cell type with an available Cre line.
The ovaries are major players in somatic aging1, with distinct contributions from specific cell populations. The cellular heterogeneity of the ovary makes it difficult to interpret molecular results from bulk, whole-ovary assays. Understanding the role of specific cell populations in ovarian aging is key to identifying the molecular drivers responsible for fertility and health decline in aged women. Traditionally, the multi-omics assessment of specific ovarian cell types was achieved by techniques such as laser microdissection2, single-cell approaches3, or cell sorting4. However, microdissection can be expensive and difficult to perform, and cell sorting can alter cellular phenotypic profiles5.
A novel approach to assess ovarian cell-type-specific epigenomic and transcriptomic profiles uses the nuclear tagging and translating ribosome affinity purification (NuTRAP) mouse model. The NuTRAP model allows the isolation of cell-type-specific nucleic acids without the need for cell sorting by using the affinity purification methods: translating ribosome affinity purification (TRAP) and isolation of nuclei tagged in specific cell types (INTACT)6. The expression of the NuTRAP allele is under the control of a floxed STOP cassette and can be targeted to specific ovarian cell types using promoter-specific Cre lines. By crossing the NuTRAP mouse with a cell-type-specific Cre line, the removal of the STOP cassette causes eGFP-tagging of the ribosomal complex and biotin/mCherry-tagging of the nucleus in a Cre-dependent manner6. The TRAP and INTACT techniques can then be used to isolate mRNA and nuclear DNA from the cell type of interest and proceed to transcriptomic and epigenomic analyses.
The NuTRAP model has been used in different tissues, such as adipose tissue6, brain tissue7,8,9, and the retina10, to reveal cell-type-specific epigenomic and transcriptomic changes that may not be detected in whole-tissue homogenate. The benefits of the NuTRAP approach over traditional cell sorting techniques include the following: 1) the prevention of ex vivo activational artifacts8, 2) the minimized need for specialized equipment (i.e., cell sorters), and 3) the increased throughput and decreased cost of cell-type-specific analyses. In addition, the ability to isolate cell-type-specific DNA and RNA from a single mouse allows for paired analyses that increase the statistical power. Since recent studies have implicated ovarian stromal cells in driving premature aging phenotypes11,12,13, we targeted the NuTRAP expression system to stromal and theca cells using a Cyp17a1-Cre driver. Here, we demonstrate that the induction of the NuTRAP construct is specific to ovarian stromal and theca cells, and sufficient DNA and RNA for sequencing studies are obtained from a single ovary. The NuTRAP model and methods presented here can be used to study any ovarian cell type with any available Cre line.
For the generation of a cell-type-specific ovarian NuTRAP mouse line, the nuclear tagging and translating ribosome affinity purification (NuTRAP) allele has a floxed STOP codon that controls the expression of BirA, biotin ligase recognition peptide (BLRP)-tagged mCherry/mRANGAP1, and eGFP/L10a. When crossed with a cell-type-specific Cre line, the expression of the NuTRAP cassette labels the nuclear protein mRANGAP1 with biotin/mCherry and ribosomal protein L10a with eGFP in a Cre-dependent manner. This allows for the isolation of nuclei and mRNA from specific cell types without the need for cell sorting. The NuTRAPflox/flox can be paired with a cell-type-specific Cre relevant to ovarian cell types to assess this.
All animal procedures were approved by the Institutional Animal Care and Use Committee at the Oklahoma Medical Research Foundation (OMRF). Parent mice were purchased from the Jackson Laboratory (Bar Harbor, ME) and bred and housed under SPF conditions in a HEPA barrier environment on a 14 h/10 h light/dark cycle (lights on at 6:00am) at the OMRF.
NOTE: In this demonstration, we use a Cyp17iCre+/−(Strain # 028547, The Jackson Laboratory) male paired with a NuTRAP female (Strain # 029899, The Jackson Laboratory). The desired Cyp17-NuTRAP progeny (Cyp17iCre+/−; NuTRAPflox/WT) will express the NuTRAP allele under the control of the Cyp17a1 promoter in ovarian stromal cells. For this manuscript preparation, we used 4-month-old female Cyp17-NuTRAP mice (n = 4). DNA was extracted from mouse ear punch samples for genotyping to confirm the inheritance of the Cyp17iCre and NuTRAP transgenes and for the PCR detection of 1) generic Cre, 2) Cyp17iCre, and 3) NuTRAP floxed using the primers listed in the Table of Materials, as previously described7,8.
1. Mouse ovarian dissection
2. Isolation of nuclei from specific ovarian cell types
3. Isolation of mRNA from specific ovarian cell types
A schematic of the TRAP and INTACT protocols is shown in Figure 1. Here the specificity of the Cyp17-NuTRAP mouse model to ovarian stromal/theca cells is demonstrated by immunofluorescent imaging and RNA-Seq from TRAP-isolated RNA. First, immunofluorescence imaging of the eGFP signal in the ovary and localization of the eGFP signal to the theca and stromal cells were performed. Briefly, 5 µm sections were deparaffinized with a xylene and ethanol gradient. For better eGFP signaling, the eGFP protein was stained using goat anti-GFP primary antibody and Alexa 488 donkey anti-goat secondary antibody, and the nuclei were stained with DAPI15. Images were taken on a fluorescence microscope at 20x magnification (Figure 2A). Next, TRAP was performed on ovaries from 4-month-old Cyp17-NuTRAP mice to isolate ribosome-bound RNA specifically from the theca/stromal cells. The TRAP-isolated RNA (positive fraction) and whole-tissue RNA (input) were used to construct stranded RNA sequencing libraries to sequence on a next-generation sequencer. RNA-Seq trimming, alignment, and normalization (DESeq2) were performed, as previously described7,8. The principal component analysis (PCA) showed strong separation of the input and positive fraction in the first component, suggesting distinct transcriptomic profiles (Figure 2B). There were 2,009 differentially expressed genes between the input and positive fractions, as seen in the volcano plot (paired t-test, Benjamini-Hochberg multiple testing correction FDR < 0.05, fold change >2; Figure 2C). Of the differentially expressed genes, the enrichment of stromal and theca cells markers (Col3a1, Star, Fbn1, C1s) and the depletion of oocyte (Gdf9, Oosp1, Ooep), granulosa (Amhr2, Serpine2, Eml5), immune cell (C1qa, C1qb, Fcerg1), and smooth muscle cell (Mfap5) genes16 could be observed (Figure 2D).
Figure 1: Isolation of cell-type-specific nuclei and mRNA from the Cyp17-NuTRAP model using the INTACT and TRAP methods. (A) The generation of the Cyp17-NuTRAP line is achieved by crossing a NuTRAPflox/flox female with a Cyp17iCre+/+ male. (B) Schematic of the TRAP and INTACT methodologies to isolate nuclei and polysomes from the stromal/theca cells of the Cyp17-NuTRAP mouse. Please click here to view a larger version of this figure.
Figure 2: Validation of Cyp17-NuTRAP mouse model. (A) Immunofluorescence was performed on the paraffinized ovaries of Cyp17-Cre+/−NuTRAPflox/WT and Cyp17-Cre−/−NuTRAPflox/WT mice. eGFP protein was stained using goat anti-GFP primary antibody and Alexa 488 donkey anti-goat secondary antibody, and the nuclei were stained with DAPI. Images were taken on a fluorescence microscope upon 20x magnification. (B–D) The input and TRAP positive fraction RNA was isolated from Cyp17-Cre+/−NuTRAPflox/WT mouse ovaries (n = 4), used to prepare stranded RNA-Seq libraries, as previously done7, and sequenced in a paired-end manner. RNA-Seq trimming, alignment, and normalization (DESeq2) were performed, as previously done7,8. (B) The principal component analysis (PCA) of all the expressed genes showed a separation of the TRAP input and positive fractions in the first component (51.4% explained variance). (C) Volcano plot of differentially expressed genes between the input and positive fractions (paired t-test, Benjamini-Hochberg multiple testing correction FDR < 0.05, fold change >2). (D) The comparison of the TRAP positive fraction to the input (log2[fold change ([ositive fraction/Input)]) showed an overall enrichment of the stromal/theca cell marker genes and the depletion of the oocyte, granulosa, immune, and smooth muscle cell (SMC) marker genes in the TRAP positive fraction (ratio paired t-test, Benjamini-Hochberg multiple testing correction FDR < 0.10, n = 4). The bars represent the mean log2[fold change (positive fraction/input)] ± SEM. Please click here to view a larger version of this figure.
The NuTRAP mouse model6 is a powerful transgenic labeling approach for the paired interrogation of the transcriptome and epigenome from specific cell types that can be adapted to any cell type with an available Cre driver. Here, we demonstrate the specificity of the Cyp17-NuTRAP mouse model in targeting ovarian theca and stromal cells. The Cyp17-NuTRAP model can be used to further elucidate the theca and stromal cell-specific epigenetic mechanisms involved in ovarian aging, cancer, and disease.
When selecting a Cre driver, it is recommended to validate the cell specificity of the Cre-mediated NuTRAP expression using several orthogonal approaches. The Cyp17a1-Cre is constitutively expressed. The NuTRAP model can also be crossed with inducible Cre lines to avoid the transient expression of genes in non-target cells during development. Since most ovarian cells are highly sensitive to hormonal fluctuations, estrous cycle staging should also be considered17.
The NuTRAP approach has many benefits when compared to traditional cell sorting techniques. First, cell sorting techniques rely on the creation of single-cell suspensions, which can cause the induction of ex vivo activation artifacts8. In addition, cell sorting relies on cell surface markers, which may change with age or disease states, thus bringing into question whether the populations of cells being compared between two experimental groups are actually equivalent. Cell surface marker labeling also relies on the identification of reliable antibodies for immunostaining18.
Over the past decade, there has been an expansion in single-cell/nuclei transcriptomic19,20,21,22,23 and epigenomic24,25,26 techniques to explore the heterogeneity of cell populations. While these approaches have led to a greater appreciation of cellular heterogeneity, they often suffer from a lack of sensitivity and genomic coverage, which can limit the depth of analyses. For example, single-cell transcriptomic (scRNA) techniques only detect ~1,000-5,000 genes per cell27, whereas the NuTRAP technique used in this study detected 14,980 genes. In addition, several scRNA techniques rely on 3' or 5' tagging20,21,22,28,29,30, which does not allow for the analysis of specific RNA isoforms available using the TRAP-seq methodology.
While single-cell transcriptomics has become commonplace in molecular biology (reviewed in Aldridge and Teichmann31), the epigenetic field has lagged behind. Most single-cell epigenomic techniques rely on the FACS deposition of cells or microfluidic techniques followed by library preparation from single cells. As such, the throughput for single-cell epigenomic techniques is generally lower than for the droplet encapsulation methods commonly used for single-cell RNA-Seq32. The recent addition of a single-cell (sc) or single-nuclei (sn) assay for transposable-accessible chromatin (ATAC-Seq) to the 10x genomics family of single-cell solutions represents the first droplet-based method allowing for the interrogation of an epigenomic endpoint (chromatin accessibility) at single-cell resolution. Multiplexing techniques, such as CoBATCH-Seq24 for histone modifications, have increased the throughput dramatically but not to the level of droplet-based approaches. Single nuclei methyl-sequencing (snmC-Seq) has been performed using fluorescence-activated nuclei sorting into individual wells and BS-Seq library preparation25,26. However, the harsh bisulfate conversion leads to sparse genome-wide coverage, which necessitates large genomic binning during data analysis (~100 kb) and causes the genomic context necessary to make robust scientific conclusions to be lost. The NuTRAP mouse model allows for the sensitive analysis of the epigenome and transcriptome of specific cell types.
Regarding the TRAP procedure, there are some considerations to ensure successful isolation. First, the ovary must be cut prior to homogenization in order to rupture the external membrane and allow for complete dissociation. The ovary is a fibrous tissue, especially in older mice, making it difficult to fully homogenize. However, during the optimization of this protocol, aggressive homogenization during the TRAP protocol resulted in contamination of the RNA isolates, with visible DNA bands displayed by capillary electrophoresis. To resolve this issue, 2 mM spermidine was included in the TRAP homogenization buffer in order to stabilize the nuclear membrane33. Additionally, we only homogenized the TRAP samples with the loose pestle A to prevent the disruption of the nuclear membrane. It is important to check the RNA samples on a Bioanalyzer or TapeStation following isolation to ensure an RNA integrity number >7 before proceeding to sequencing applications.
For the INTACT procedure, the ovary should also be cut prior to homogenization. In addition, we have found that a short, low-speed spin (200 x g for 1.5 min) following homogenization is effective for removing undissociated tissue and blood vessels, as was done previously7,14. If nuclear RNA is a desired endpoint, an RNase inhibitor should be included in the INTACT buffers. Nuclear RNA-Seq or RT-qPCR can be used to test the cell specificity of the INTACT isolations. Streptavidin beads have an incredibly high affinity for biotin, making it difficult to separate the beads from biotin+ nuclei following the INTACT protocol, and this should be considered when planning for downstream applications. In addition to isolating cell-type-specific nuclei from NuTRAP models using INTACT, nuclei can also be sorted based on eGFP expression for downstream applications. The INTACT procedure for isolating nuclei can be used to assess several endpoints, including nuclear transcriptomics, epigenomics, and proteomics.
The aforementioned techniques are great tools for ovarian aging assessments. Ovarian aging is a concern not only from a reproductive perspective but also from a health span perspective. Menopause, which is the cessation of ovarian function, is associated with the acceleration of biological clocks34, and early menopause is associated with a shorter lifespan and increased risk of diseases in women35,36. Oocyte aging is known to be directly related to a decline in ovarian fitness. However, other cell types are also believed to play roles in the decline of ovarian health span and an increase in ovarian senescence and inflammation37. A better understanding of which cells are involved in this process and how to slow it is key. Even though there are undeniable differences between human and murine ovarian dynamics, mouse models remain important tools to investigate ovarian biology. In this context, the Cre-NuTRAP models can be useful tools to identify the most effective cell type targets for attempts to slow ovarian aging and menopause.
The authors have nothing to disclose.
This work was supported by grants from the National Institutes of Health (NIH) (R01AG070035, R01AG069742, T32AG052363), BrightFocus Foundation (M2020207), and Presbyterian Health Foundation. This work was also supported in part by the MERIT award I01BX003906 and a Shared Equipment Evaluation Program (ShEEP) award ISIBX004797 from the United States (U.S.) Department of Veterans Affairs, Biomedical Laboratory Research and Development Service. The authors would also like to thank the Clinical Genomics Center (OMRF) and Imaging Core Facility (OMRF) for assistance and instrument usage.
0.1 M Spermidine | Sigma-Aldrich | 05292-1ML-F | |
1 M MgCl2 | Thermo Scientific | AM9530G | |
10% NP-40 | Thermo Scientific | 85124 | |
100 mg/mL Cycloheximide | Sigma-Aldrich | C4859-1ML | |
2-mercaptoethanol | Sigma-Aldrich | M3148 | |
30 µm cell strainer | Miltenyi Biotec | 130-098-458 | |
All Prep DNA/RNA Mini Kit | Qiagen | 80204 | |
anti-GFP antibody | Abcam | Ab290 | For TRAP and IHC (Rabbit polyclonal to GFP) |
Buffer RLT | Qiagen | 79216 | RNA Lysis Buffer in protocol |
cOmplete, mini, EDTA-free protease inhibitor tablet | Roche | 11836170001 | For TRAP Homogenization Buffer |
Cyp17iCre mouse model | The Jackson Laboratory | 28547 | B6;SJL-Tg(Cyp17a1-icre)AJako/J |
DynaMag-2 magnet | Invitrogen | 12321D | |
Genotyping Primers | IDT | Custom | Generic Cre – Jackson Laboratory protocol 22392, Primers: oIMR1084, oIMR1085, oIMR7338, oIMR7339 |
Cyp17iCre – Jackson Laboratory protocol 30847, Primers: 21218, 31704, 31705, 35663 | |||
NuTRAP – Jackson Laboratory protocol 21509, Primers: 21306, 24493, 32625, 32626 | |||
Halt Protease Inhibitor cocktail (100X) | Thermo Scientific | 1861278 | For NPB Buffer |
M-280 Streptavidin Dynabeads | Invitrogen | 11205D | 2.8 µm bead diameter |
MixMate | Eppendorf | 5353000529 | |
Nuclei Isolation Kit: Nuclei EZ Prep | Sigma-Aldrich | Nuc101 | Contains Nuclei Lysis Buffer and Nuclei Storage Buffer |
1 M HEPES | Gibco | 15630-080 | |
5 M NaCl | Thermo Scientific | AM9760G | |
2M KCl | Thermo Scientific | AM9640G | |
0.5 M EDTA | Thermo Scientific | AM9260G | |
0.5 M EGTA | Fisher Scientific | 50-255-956 | |
NuTRAP mouse model | The Jackson Laboratory | 29899 | B6;129S6-Gt(ROSA)26Sortm2(CAG-NuTRAP)Evdr/J |
Pierce DTT No-Weigh Format | Thermo Scientific | A39255 | |
Protein G Dynabeads | ThermoFisher | 10004D | For TRAP |
RNaseOUT | Invitrogen | 10777019 | |
Sodium Heparin | Fisher Scientific | BP2425 | |
Ultrapure 1M Tris-HCl, pH 7.5 | Invitrogen | 15567-027 | |
VWR Tube Rotator | Fisher Scientific | NC9854190 |