Here, we describe the immunoprecipitation of ribosomes and associated RNA from select populations of adult male mouse germ cells using the RiboTag system. Strategic breeding and careful immunoprecipitation result in clean, reproducible results that inform on the germ cell translatome and provide insight into the mechanisms of mutant phenotypes.
Quantifying differences in mRNA abundance is a classic approach to understand the impact of a given gene mutation on cell physiology. However, characterizing differences in the translatome (the whole of translated mRNAs) provides an additional layer of information particularly useful when trying to understand the function of RNA regulating or binding proteins. A number of methods for accomplishing this have been developed, including ribosome profiling and polysome analysis. However, both methods carry significant technical challenges and cannot be applied to specific cell populations within a tissue unless combined with additional sorting methods. In contrast, the RiboTag method is a genetic-based, efficient, and technically straightforward alternative that allows the identification of ribosome associated RNAs from specific cell populations without added sorting steps, provided an applicable cell-specific Cre driver is available. This method consists of breeding to generate the genetic models, sample collection, immunoprecipitation, and downstream RNA analyses. Here, we outline this process in adult male mouse germ cells mutant for an RNA binding protein required for male fertility. Special attention is paid to considerations for breeding with a focus on efficient colony management and the generation of correct genetic backgrounds and immunoprecipitation in order to reduce background and optimize output. Discussion of troubleshooting options, additional confirmatory experiments, and potential downstream applications is also included. The presented genetic tools and molecular protocols represent a powerful way to describe the ribosome-associated RNAs of specific cell populations in complex tissues or in systems with aberrant mRNA storage and translation with the goal of informing on the molecular drivers of mutant phenotypes.
Analysis of a cell or tissue's RNA abundance, as examined by microarray or RNA-sequencing, has proven a powerful tool to understand the molecular drivers of mutant phenotypes. However, these relatively incomplete analysis tools may not inform on the physiology or proteome of the cell, especially in systems where many mRNAs are stored prior to translation such as neural and germ cells. In these systems, defining the population of mRNAs being actively translated into protein, or the cell's translatome, may be a better indicator of the cell's actual physiological state1,2. For example, germ cells at various stages of developement transcribe RNAs that are stored for later translation, driven either by developmental3 or signaling cues4. This process is exemplified by the mRNAs encoding protamines, wherein the mRNA transcript is detectable days before the protein is made1,2,5,6. Likewise, neural cells transcribe RNA in the nucleus and transport it down the axon, as is the case with β-actin7. In addition to these specialized cell systems, steady-state transcriptomes are unlikely to be informative in models where RNA storage, ribosome biogenesis, or mRNA translation are impacted. Multiple other factors may also impact a cell's steady-state RNA pool include mRNA decay and the activity of RNA binding proteins. In these cases, robust tools to analyze ribosome-associated RNAs or mRNAs under active translation are more likely to yield biologically relevant results.
To that end, several methods have been developed for identifying ribosome-associated or actively translated messages. Polysome profiling, which provides a snapshot of ribosome associated transcripts, has been used for many decades to isolate intact RNAs associated with either ribosomal subunits or mono-, di-, and poly-ribosome complexes3. Briefly, collected cell lysates are applied to a linear sucrose gradient and centrifuged as high speed, resulting in separation of the ribosome subunits, intact ribosomes, and polysomes by size. Traditionally, this technique has been applied to study one or a few mRNAs, but recently this method has been combined with RNA-seq studies to elucidate the function of potential translational regulators8,9 While a powerful way to differentiate between actively translating and non-translating mRNAs10, polysome profiling does require time-consuming methods (gradient fractionation and ultracentrifuge) and can require a good deal of sample making the analysis of rare cell populations challenging.
An alternative approach to examining the translatome is ribosome profiling, which identifies the portions of transcripts directly associated with the ribosome. In brief, ribosome associated RNA fragments are generated via RNase protection assay, individual ribosome complexes separated via sucrose gradient, and associated RNA fragments isolated and converted to RNA-seq tags amenable to deep sequencing11. One of the key benefits to ribosome profiling is the ability to determine the specific locations of the ribosomes at the time of isolation which allows identification of translation start sites, calculation of ribosomal occupancy and distribution, and identification of ribosome stalls12. However, this method has several inherent drawbacks, including equipment needs (gradient fractionator and ultracentrifuge), a relatively complex protocol that require extensive troubleshooting, and computational issues not easily handled by the inexperienced user11. Importantly, ribosome profiling is primarily applied to isolated cells in culture and requires substantial material, making its application to mixed cell-type tissues or sorted cells from in vivo limited.
The mammalian RiboTag method, developed by Sanz et al. in 200913, eliminates a number of issues inherent to both polysome and ribosome profiling. In this method, ribosome-associated RNAs can be isolated from specific cell types allowing for investigation of cell-type specific translatomes in complex tissues without additional isolation techniques and specialized equipment, as is necessary in other methods13,14. The basis of the RiboTag method is a transgenic mouse model (RiboTag) carrying a modified locus for the 60S ribosomal subunit protein gene Rpl22. This locus (Rpl22-HA) includes a floxed terminal exon followed by an additional copy of the terminal exon amended to include a C-terminal hemagglutinin (HA) tag within the coding region. When crossed to a mouse expressing a cell-specific Cre Recombinase, the floxed exon is removed allowing the expression of HA-tagged RPL22 in a cell-specific manner (Figure 1). The HA tag can then be used to immunoprecipitate (IP) ribosomes and their associated RNAs from the selected cell type.
In addition to the initial publication that developed the technique, several other laboratories have utilized the RiboTag model. Diverse tissues such as mouse Sertoli and Leydig cells15, microglia16, neurons17,18, oocytes19, and cultured cells20 have been profiled and studied. Though this protocol is clearly able to successfully isolate ribosomes and the associated RNAs across a diverse tissue types there are still areas needing improvement, especially when applied to mutant systems. In particular, common reliance on fresh tissue results in increased technical variation which can greatly reduce the power of the analysis. Secondly, confident identification of differentially translated targets is made more challenging when high immunoprecipitation background occurs from non-Cre recombined cell types as previously reported13. While Sanz et al. engineered the basic premise of the technique, in this manuscript the Snyder laboratory presents optimization of the protocol to reduce these concerns, rendering the method more practical and efficient.
The intent of this article is to explain the steps for breeding mice expressing cell-specific HA-tagged ribosomes, immunoprecipitating these ribosomes from flash-frozen samples, and purifying their associated RNAs for further downstream analyses. As the mammalian male germ cell and the infertility mutation studied provide unique challenges, efforts have been made to illuminate ways this technique can be adapted to other cell systems.
All animal use and handling practices have been approved by Rutgers University's Internal Animal Care and Use Committee (IACUC).
1. Mouse breeding
2. Sample collection
3. Preparation of solutions
NOTE: Preparation of solutions and all subsequent steps must be done under stringently conditions.
4. Preparation of tissue
5. Homogenization of sample
6. Equilibration of beads
7. Preclearing sample
8. Incubation of antibody
9. Incubation of beads
10. Washing of beads
11. RNA extraction
12. Quantification and sample analysis
Previous reports have suggested non-specific immunoprecipitation from cells lacking Cre14. In order to determine if this was the case in our modified protocol, IP efficiency was determined in samples derived from animals carrying both Cre and Rpl22-HA and animals carrying only one but not the other with the expectation that without both a Cre-driver and Rpl22-HA, immunoprecipitated RNA should be minimal. As shown in Figure 3, very little RNA is isolated from samples lacking either Cre or Rpl22-HA demonstrating the effectiveness of this protocol to reduce IP background and isolate genuine HA-tagged ribosome RNAs. Further, both Cre-only and Rpl22-HA-only samples represent suitable negative controls.
Given the potential for reagent source to significantly impact the efficiency of IP, a series of antibodies and RNA isolation protocols were tested (Figure 4) in Cre and Rpl22-HA positive samples. These results demonstrate reagent selection can have a significant impact on IP efficiency thus any changes to reagent selection should be done so with care.
In order to test this protocol in the context of an RNA binding protein mutant, wildtype-Cre+Rpl22-HA+ (wildtype) and gene of interest-/--Cre+Rpl22-HA+ (Gene of interest-/-) testis were examined for the effectiveness of the RiboTag system to isolate ribosome associated RNA (Figure 5). When RNA concentration of wildtype input and IP was compared to Gene of interest-/- input and IP, no significant difference was seen between genotypes. For both genotypes, however, the input concentration was significantly higher than the IP concentration, indicating that there was more RNA in the input sample (Figure 5B). This result is expected, as not all the mRNAs present in the cell are associated with the ribosome at any given time, especially in the case of germ cells where RNAs may be transcribed long before they are translated.
In order to confirm the quality of the RNA samples sent for sequencing, samples were run on a bioanalyzer. RNA integrity numbers (RINs), normally calculated as the ratio of 28S and 18S rRNA peaks, were compared across samples. In total RNA pools, RIN values are expected to be near 10 with a higher RIN correlated to higher inferred sample integrity and quality. While the IP samples had lower RIN than the inputs, the RINs were still within an acceptable range and were not dependent on sample genotype (Figure 5C). The reduced RIN values for IPed samples are likely a result of RNA degradation though very minor given the relatively small decrease in average nucleotide length of analyzed RNAs. Given the length of the protocol and temperatures required for the immunoprecipitation some degradation is expected. It is also possible the reduced RIN and RNA length is a function of enrichment for non-rRNA species, such as mRNAs.
Figure 1: The RiboTag method. The premise of the method is biologically simple. A new Exon 4 is inserted into the sequence for the Rpl22 locus downstream of the original Exon 4. In the presence of a Cre driver, loxP sites on either side of the original Exon 4 are cut, excising the floxed exon. The HA-tagged Exon 4 is now incorporated into Rpl22 mRNA, generating an HA tagged RPL22 in cells expressing CRE. Please click here to view a larger version of this figure.
Figure 2: Sample breeding scheme for RiboTag mice. The breeding scheme used to generate experimental animals is shown. A two-pronged scheme was utilized. In generation 1, two parallel sets of breeder trios were established. One side combined the allele of interest (carried maternally as this specific mutation results in male infertility) with the hemizygous Cre and on the other combining the allele of interest, again carried maternally, with Rpl22-HA. Then, in generation 2, males from the first pairing that carry the allele of interest and the Cre are crossed to female offspring of the second pairing that carry the allele of interest and Rpl22-HA. The genotypes of the resulting offspring are determined, and experimentally relevant animals selected (in this case either wildtype or homozygous mutant carrying both the Cre and Rpl22-HA alleles). It is important to note for germ cell expressing Cre-drivers, the Cre and Rpl22-HA alleles should not be breed together until the experimental generation. Exposure of the Rpl22-HA allele to germ cell-expressed Cre in breeding generations will drive germ-cell Rpl22-HA excision. Any resulting offspring will globally express Rpl22-HA thus preventing cell-specific ribosome isolation. In the system described herein, an n = 4 per genotype provided sufficient statistical power for downstream analyses. Biological replicate number should be determined for each experimental system. Please click here to view a larger version of this figure.
Figure 3: Confirmation of negative controls. Samples that are Cre+ and Rpl22-HA+ show higher RNA pulldown than samples that are Cre+ or Rpl22-HA+ alone. There is no significant difference between Cre+ and Rpl22-HA+ sample RNA pulldown efficacy, indicating that samples lacking either the Cre or the Rpl22-HA are suitable negative controls. A "+" indicates that the corresponding allele or transgene was present in these samples and a "-" denotes its absence. IP/input represents the ratio of RNA immunoprecipitated (IP) over total (input) RNA. Value calculated from concentration in nanograms. ** indicates p < 0025. Please click here to view a larger version of this figure.
Figure 4: Reagents impact protocol success. (A) Flash frozen tissue results in immunoprecipitation efficiency similar to that of fresh tissue. (B) IP efficiency for two commercial antibodies were determined, designated as Antibody 1 and Antibody 2 (Table of Materials). When tested, Antibody 1 was more efficient at pulling down RNA than Antibody 2 which appeared to be unable to differentiate between negative (not expressing either Cre or Rpl22-HA+) and positive controls (expressing both Cre and Rpl22-HA+). Dots indicative of the ratio of IPed versus input RNA for individual biological replicates. (C) When RNA extraction kits (Table of Materials) were compared, Kit 2 significantly outperformed Kit 1. Though both kits IPed a similar amount of RNA from negative controls, Kit 2 resulted in a significantly higher RNA yield from positive samples. * indicates p < 0.05, ** p < 0.025, *** p < 0.01. Please click here to view a larger version of this figure.
Figure 5: Application of the method to a mutant model. (A) Sample Gene of interest genotype confirmation via Western blotting using a custom in-house antibody against the gene of interest protein demonstrates Gene of interest-/- (M/M) males fail to produce the associated protein. GAPDH shown as a loading control. (B) Graphical bioanalyzer output from paired input and IPed samples for wildtype and mutant samples. (C) Comparison RNA integrity numbers (RIN) by sample type and genotype. (D) Average nucleotide length of RNA species analyzed by bioanalyzer by sample type and genotype. * indicates p < 0.05, ** p < 0.025, *** p < 0.01, N.S. not significant. N = 4 per genotype. Please click here to view a larger version of this figure.
Figure 6: Example of Cre driver confirmation. Quantification of a gene translated early in germ cell development (Stra8) and two genes translated late in germ cell development (Tnp1 and Prm1) analyzed by qRT-PCR of RNA immunoprecipitated from RPL22-HA driven by two different germ cell Cres, Stra8-iCre (expressed early in germ cell development) and Hspa2-Cre (expressed later in germ cell development, specifically after Stra8 translation). Here, HA-IP of Stra8 is achieved with the early germ cell Cre driver but not with the later germ cell Cre driver demonstrating cell-specificity of the immunoprecipitation. In contrast, HA-IP of transcripts translated in late germ cells is achieved by both Stra8-iCre and Hspa2-Cre. This is expected as cells that express Stra8-iCre will generate HA-tagged RPL22 throughout the entirety of their development while those expressing Hspa2-Cre will only do so during the later portions of their development. Please click here to view a larger version of this figure.
Understanding the translatome of a particular cell type is invaluable for more accurately understanding a cell's physiology in the normal or mutant state. Special benefit is seen in systems wherein translation is uncoupled from transcription, such as in neural tissue where translation occurs very far from transcription, or in germ cells where transcription occurs long before translation. Relative to other methods of translatome analysis, the RiboTag system's biggest advantage comes from the use of the Cre recombinase system. This allows the freedom to target any cell population that has a relevant Cre driver. Secondly, the RiboTag IP as described herein is effective and much less technically challenging and time-consuming than either polysome or ribosome profiling. Lastly, RiboTag IP can be easily performed at the benchtop.
There are a number of critical steps in this protocol. Chief among them is the establishment of the RiboTag mouse line and generation of experimental animals for study. As for all genetic models, careful tracking of individuals within the mouse colony as well as careful genotyping protocols should be applied. PCR genotyping as per Sanz et al. should include primers targeting the loxP site 5' of the wildtype exon 4 to distinguish between wildtype alleles (260 bp) and mutant (290 bp)13. For the case of RiboTag analysis in germ cell models, very specific breeding requirements should be adhered to. First, in the case of mutants that result in infertility, thoughtful breeding strategies should include methods to optimize the number of offspring with the desired genotype in intermediate and experimental generations. Second, in the case of germ cell specific Cres, care should be taken regarding Cre-zygosity given the sensitivity of germ cells to Cre toxicity. In the germ cell, high levels of Cre protein can have deleterious effects22, prohibiting the use of Cre/Cre animals in the breeding scheme. Lastly, when using germ cell Cres, it is important to isolate the RiboTag allele from the Cre until the final generation as Cre expression in the germ cell of an intermediate generation will result in offspring expressing Rpl22-HA globally.
Regardless of cell system, a number of recommended controls are possible to verify robustness of both Cre expression and specificity. Proper expression of your Cre and expected excision of Rpl22-HA can be determined using multiple methods. In the first, tissue isolated from experimental animals can be stained for HA using either immunohistochemistry or immunofluorescence15. This method is optimal in that it requires no a priori knowledge of translated mRNAs in the target cell types. The second method, an example of which is demonstrated in Figure 6, requires some knowledge of translationally regulated messages in the target cell types. In this method, enrichment for a known translationally regulated mRNA can be confirmed from the selected Cre-driver using quantitative PCR of IPed versus input RNA. Likewise, robust Rpl22-HA expression can be confirmed by comparison of Cre+/Rpl22+ samples (positive controls) with either Cre-/Rpl22+ or Cre+/Rpl22- samples which act as effective negative controls (see Figure 3). These comparisons can either by done on total IPed RNA or enrichment for a known translationally regulated mRNA assessed by qRT-PCR or some other quantitative method.
Common problems in the execution of the protocol tend to only become apparent when RNA yield is unexpectedly low or high. The most common cause for these failures arise from incorrect genotyping of individuals. We recommend retaining additional tissue from collected sample to confirm genotypes of all samples in the case of unexpected RNA yields. Once confirmed, other possibilities should be considered including the possibility of RNase contamination or sample degradation. Careful sample storage and handling and maintenance of RNase free zones within the laboratory can greatly reduce or eliminate these issues. Although RNA isolation issues are another potential problem in this protocol, the use of commercial kits greatly reduces this issue though care should be taken to ensure they are maintained as RNase free and contain no expired solutions. Lastly, as with all antibody-based procedures, variations in lot, storage conditions, concentration, or even shipping have the potential to negatively impact the quality of the antibody and the subsequent success of the pulldown. As a result, following careful and repeated testing, we chose the most consistent and efficient antibody available.
This protocol contains two major modifications relative to other published RiboTag protocols that significantly enhance the likelihood of success. One is the ability to use flash frozen tissue, thereby mitigating any issues involving lot, technician, or technical run variations. Samples can be collected and stored allowing isolation of HA-ribosomes from all samples at once, lowering what may be major sources of variation. Second, the addition of the preclear step substantially reduces the sort of background reported by other users of the RiboTag system. Recently, a protocol report by Sanz et al. indicates the presence of high background due to abundance of ribosome-associated transcripts from non-Cre-driven cells14. Our protocol remedies this issue by including a preclearing step, effectively eliminating the presence of RNA in Cre negative IP samples.
Like all systems, the inherent limitations of the RiboTag system should be kept in mind. When using uncharacterized Cre drivers, analysis of expression should be performed prior to experimental sample production. From the perspective of translation, several nuances of this method should be noted. First, RiboTag does not allow differentiation between mRNAs poised to translate and those actively translating. As such, current RiboTag-based methods do not allow the quantification of translation efficiency as a function of individual mRNAs. If translation efficiency is of interest, it may be measured on a cell-specific basis if the RiboTag method is combined with other translatome analysis tools such as polysome profiling. Secondly, it is fundamental to take into account total RNA abundance changes stemming from individual or genotype dependent variance. It is for this reason that careful isolation of input RNA accompany immunoprecipitation and samples derived from each remain paired throughout any downstream analyses. Lastly, and in regards to RiboTag-based analysis, it should be remembered that association with a ribosome does not necessarily prove translation is occurring. Secondary methods of analysis should be performed to confirm translational regulation in targets of interest.
This protocol describes the isolation of ribosome-associated RNAs from the germ cells of male mice using the RiboTag model. Not accounting for mouse breeding and sample collection, the protocol takes two days, with three hours the first day, best done in the afternoon, an overnight incubation, followed by five hours of work the subsequent morning. It is strongly recommended that preparation of stock solutions (HB and HSWB) as well as tissue grinding be done in advance. The overall success of the protocol is reliant on correct genotyping and stringently RNase-free conditions. The ability to examine the translatome of specific cell types will allow future studies to better understand the relationship between transcription, translation, and the proteome in myriad cell types and mutant backgrounds.
The authors have nothing to disclose.
This work was funded by NIH grant NICHD R00HD083521 to EMS and internal support from Rutgers University to EMS.
1 mL mechanical pipette | Preference of researcher | ||
1,4-Dithio-DL-threitol (8% | Alfa Aesar | A15797 | |
1.7 mL or 2mL tubes | Preference of researcher | ||
10 mL conical tubes | Preference of researcher | ||
10 mL serological pippettes | Preference of researcher | ||
10 uL mechanical pipette | Preference of researcher | ||
20 uL mechanical pipette | Preference of researcher | ||
200 uL mechanical pipette | Preference of researcher | ||
5 mL serological pipettes | Preference of researcher | ||
50 mL conical tubes | Preference of researcher | ||
Anti-HA tag antibody | Abcam | ab18181 | Antibody 1 |
Anti-HA tag antibody | Antibodies.com | A85278 | Antibody 2 |
B6.FVB-Tg(Stra8-icre)1Reb/LguJ Mice | The Jackson Laboratory | 17490 | Or mice carrying Cre driver of choice |
B6N.129-Rpl22tm1.1Psam/J Mice | The Jackson Laboratory | 11029 | |
Benchtop centrifuge | Preference of researcher | ||
C1000 Touch thermal cycler | BioRad | 184-1100 | Or equivalent thermal cycler |
Centrifuge 5424 R | Eppendorf | 5404000537 | Or equivalent refrigerated centrifuge |
Cyclohexamide | Sigma Aldrich | C7698-1g | |
Dissection scissors | Preference of researcher | ||
Dynabeads Protein G for Immunoprecipitation | Invitrogen by ThermoFisher Scientific | 10009D | |
DynaMag-2 Magnet rack | Invitrogen by ThermoFisher Scientific | 12321D | |
E.Z.N.A. Total RNA Kit 1 | OMEGA | 6834-01 | Kit 1 |
Heat block | Preference of researcher | ||
Heparin | Sigma Aldrich | 84020 | |
Magnesium Chloride (MgCl2) | Sigma Aldrich | M9272-500g | |
Microdissection forceps | Preference of researcher | ||
Microdissection scissors | Preference of researcher | ||
MiRNeasy kit | Qiagen | 217004 | Kit 2 |
NanoDrop One Microvolume UV-Vis Spectrophotometer with Wi-Fi | ThermoFisher Scientific | ND-ONE-W | Or equivalent spectrophotometer |
NP-40 Alternative – CAS 9016-45-9 – Calbiochem | Millipore Sigma | 492016 | |
Pierce Protease Inhibitor Tablets, EDTA-free | ThermoFisher Scientific | A32965 | |
Potassium (KCl) | Sigma Aldrich | P3911-2.5kg | |
RNase Inhibitor, Murine | New England BioLabs Inc. | M0314 | |
SI vortex-genie 2 | Scientific Industries | SI-0236 | Or equivalent benchtop vortex |
Tips for 1 mL mechanical pipette | Preference of researcher | ||
Tips for 10 uL mechanical pipette | Preference of researcher | ||
Tips for 20 uL mechanical pipette | Preference of researcher | ||
Tips for 200 uL mechanical pipette | Preference of researcher | ||
Tris Base (White Crystals or Crystalline Powder/Molecular Biology) | ThermoFisher Scientific | BP152-500 | |
Tube Revolver / Rotator | ThermoFisher Scientific | 88881001 | Or equivalent rotator |
VWR Powerpette Plus pipet controller | VWR | 75856-450 | Or equivalent pipette controller |