RNA-protein interactions lie at the heart of many cellular processes. Here, we describe an in vivo method to isolate specific RNA and identify novel proteins that are associated with it. This could shed new light on how RNAs are regulated in the cell.
RNA-binding proteins (RBPs) play important roles in every aspect of RNA metabolism and regulation. Their identification is a major challenge in modern biology. Only a few in vitro and in vivo methods enable the identification of RBPs associated with a particular target mRNA. However, their main limitations are the identification of RBPs in a non-cellular environment (in vitro) or the low efficiency isolation of RNA of interest (in vivo). An RNA-binding protein purification and identification (RaPID) methodology was designed to overcome these limitations in yeast and enable efficient isolation of proteins that are associated in vivo. To achieve this, the RNA of interest is tagged with MS2 loops, and co-expressed with a fusion protein of an MS2-binding protein and a streptavidin-binding protein (SBP). Cells are then subjected to crosslinking and lysed, and complexes are isolated through streptavidin beads. The proteins that co-purify with the tagged RNA can then be determined by mass spectrometry. We recently used this protocol to identify novel proteins associated with the ER-associated PMP1 mRNA. Here, we provide a detailed protocol of RaPID, and discuss some of its limitations and advantages.
RNA-binding proteins (RBPs) represent about 10% of S. cerevisiae proteins1,2 and about 15% of mammalian proteins3-5. They are implicated in many cellular processes such as mRNA post-transcriptional processing and regulation, translation, ribosome biogenesis, tRNA aminoacylation and modification, chromatin remodeling, and more. An important subgroup of RBPs is the mRNA-binding proteins (mRNPs)6,7. In the course of mRNA maturation, different RBPs bind the transcript and mediate its nuclear processing, export out of the nucleus, cellular localization, translation and degradation6-8. Thus, the distinct set of RBPs bound to a particular transcript at any time point determines its processing and ultimately its fate.
The identification of RBPs associated with an mRNA could significantly improve our understanding of processes underlying their post-transcriptional regulation. Diverse genetic, microscopic, biochemical and bioinformatics methods have been used to identify proteins involved in mRNA regulation (reviewed in9-11). However, only a few of these methods enable the identification of proteins associated with a particular target mRNA. Of note is the Yeast Three Hybrid system (Y3H), which utilizes the mRNA of interest as bait to screen an expression library in yeast cells. Positive clones are usually observed through a growth selection or reporter expression12-14. The key advantage of this method is the large number of proteins that can be scanned in a cellular environment and the ability to measure the strength of the RNA-protein interaction. Drawbacks include the relatively large number of false positive results due to non-specific binding, and the high potential for false negative results due, in part, to misfolding of the fusion protein prey or the bait RNA.
An alternative to the genetic approach is affinity purification of RNA with its associated proteins. Poly A-containing mRNAs can be isolated through the use of oligo dT columns, and their associated proteins are detected by mass spectrometry. The RNA-protein interaction is conserved in its cellular context by crosslinking, which makes short-range covalent bonds. The use of the oligo dT column yields a global view of the entire proteome that is associated with any poly A-containing mRNA3,5,15. However, this does not provide a list of proteins that are associated with a particular mRNA. Very few methods are available to accomplish such an identification. The PAIR method entails the transfection of nucleic acid with complementarity to the target mRNA16,17. The nucleic acid is also attached to a peptide, which allows crosslinking to RBPs in close vicinity to the interaction site. After crosslinking, the RBP-peptide-nucleic acid can be isolated and subjected to proteomics analysis. Recently, an aptamer-based methodology was successfully applied to extracts from mammalian cell lines18. An RNA aptamer with improved affinity to streptavidin was developed and fused to a sequence of interest (AU-rich element (ARE) in this case). The aptamer-ARE RNA was attached to streptavidin beads and mixed with cell lysate. Proteins that associated with the ARE sequence were purified and identified by mass spectrometry (MS). Although this method detected associations that occur outside the cellular settings (i.e., in vitro), it is likely to be modified in the future so as to introduce the aptamers into the genome and thus enable the isolation of proteins associated with the mRNA while in the cellular milieu (i.e., in vivo). In yeast, where genetic manipulations are well established, the RaPID method (developed in Prof. Jeff Gerst's lab) provides a view of in vivo associations19. RaPID combines the specific and strong binding of the MS2 coat protein (MS2-CP) to the MS2 RNA sequence, and of the streptavidin-binding domain (SBP) to streptavidin conjugated beads. This enables efficient purification of MS2-tagged mRNAs through streptavidin beads. Moreover, expression of 12 copies of MS2 loops allows up to six MS2-CPs to bind simultaneously to the RNA and increase the efficiency of its isolation. This protocol was therefore suggested to enable the identification of novel mRNA-associated proteins once the eluted samples are subjected to proteomics analysis by mass spectrometry.
We recently utilized RaPID to identify novel proteins associated with the yeast PMP1 mRNA20. PMP1 mRNA was previously shown to be associated with the ER membrane and its 3' untranslated region (UTR) was found to be a major determinant in this association21. Thus, RBPs that bind PMP1 3' UTR are likely to play an important role in its localization. RaPID followed by liquid chromatography-mass spectrometry/mass spectrometry (LC-MS/MS) resulted in the identification of many new proteins that interact with PMP120. Herein, we provide a detailed protocol of the RaPID methodology, important controls that need to be done, and technical tips that may improve yield and specificity.
Note: Insert a sequence consisting of 12 MS2-binding sites (MS2 loops; MS2L) into the desired genomic locus, usually between the open reading frame (ORF) and the 3' UTR. A detailed protocol for this integration is provided elsewhere22. Verify proper insertion and expression by PCR, northern analysis or RT-PCR20,23. It is important to verify that the integration did not intervene with the synthesis of the 3'UTR. In addition, a plasmid-expressing MS2-CP fused to SBP under the expression of an inducible promoter (methionine depletion) should also be introduced into the cells24. An identical strain, excluding the introduced MS12 loops, should be used as a control for the detection of non-specific signals.
1. RaPID Purification
2. RNA Extraction
3. Protein Preparation for Western Blot or Mass Spectrometry Analysis
RaPID enables the isolation of a specific target RNA with its associated proteins. Critical for its success is keeping the RNA intact as much as possible, thereby obtaining a sufficient amount of proteins. To determine the isolation efficiency and quality of RNA, northern analysis is performed (Figure 1A). Northern analysis has the advantage of directly reporting the efficiency and quality of RaPID. Thus, the relative amounts of full length and degradation products can be determined in a single run. Ribosomal RNAs (rRNAs) are detected easily in the ethidium bromide staining, and the lack of rRNAs in the elution sample indicates the stringency of purification. The stringency and specificity of the purification are further demonstrated by the lack of a signal of an untagged mRNA in the elution samples (ACT1 bottom panel). The apparent signal in the FPR1 lane, which is not of the size of ACT1, is a leftover from previous hybridization with the MS2L probe. The northern analysis also reveals that the input RNA is somewhat more degraded compared to RNA that was purified by the hot phenol protocol (c/o ACT1 probe panel). This is attributed to the lengthier and more complicated RaPID protocol. Nevertheless, a significant amount of full-length, tagged RNA is isolated specifically, as revealed by the strong signal in the elution fraction (MS2L probe).
Protein samples can be run on SDS-PAGE and stained by silver stain prior to proteomics analysis (Figure 1B). A sample isolated from control, untagged cells (-MS2L) is helpful in distinguishing non-specific associations from RNA-dependent ones. Therefore, only bands that are stronger in the tagged sample (+MS2L) are cut out of the gel and taken for LC-MS/MS. Western analysis with general RBPs is also recommended to indicate the efficiency of protein co-purification (Figure 1C). Herein, we used Yef3, which is known to interact with PMP120, or GFP, which indicates the overall efficiency and specificity of the isolation. Interestingly, the efficiency of Yef3 isolation appears to differ between PMP1 and FPR1, and is much lower compared to GFP.
Figure 1. Specific Isolation of MS2L-tagged RNAs and Identification of Associated Proteins. Results of two different RNAs that were subjected to RaPID (PMP1 and FPR1) are presented. (A) Northern blot analysis of RNA samples purified from a RaPID experiment. RNA was run on an agarose gel and stained with ethidium bromide (upper panel). The gel was blotted to nylon membrane and subjected to hybridization with the indicated probes (lower panels). The samples analyzed are: fast-cell lysis (hot phenol [step 1.4]), RaPID cell lysis (input [step 1.13]) and after elution with biotin (elution [step 1.24]). (B) Silver stain of proteins from RaPID with MS2-tagged and untagged cells. Protein samples from elution fractions of MS2-tagged PMP1 (+MS2L) or untagged (-MS2L) control were run in SDS-PAGE and silver stained. Bands with differential intensity (indicated by asterisks) were cut out of the gel and determined by mass spectrometry analysis. The arrow indicates the MS2-CP-GFP-SBP fusion protein, which demonstrates equal protein loading. Irrelevant lanes were cropped out for clarity. (C) Western analysis of positive controls. Western blot with antibodies that recognize the Yef3 or GFP moiety of the MS2-CP fusion protein was conducted. Irrelevant lanes were cropped out for clarity. Please click here to view a larger version of this figure.
Various methods use the isolation of specific mRNAs to identify their associated proteins11,34 35. These methods apply in vitro and in vivo strategies to probe RNA-protein interactions. In vitro methods incubate exogenously transcribed RNA with cell lysate to capture RBPs and isolate RNP complexes36,37. An effective approach of this type was presented recently, which enabled the identification of novel proteins that bind a regulatory RNA motif18. A drawback of these methods is the binding of non-specific RBPs because the association occurs outside of the cellular environment. In vivo methods, in which proteins are crosslinked while in their natural settings, may provide a better view of RNA-protein association. Furthermore, crosslinking allows higher stringency during isolation and therefore higher specificity. The PAIR approach17 utilizes the transfection of the antisense sequence to the target RNA in order to direct a peptide moiety to the vicinity of RBPs. The peptide-RBPs can then be crosslinked and isolated through beads that are linked to a sequence that is complementary to the antisense sequence. The PAIR methodology is very advantageous due to its simple application to many cell types without the need for complex genomic manipulations. It is, however, dependent on the efficiency of transfection, and cases where a low percentage of cells accept the antisense will have low efficacy. Furthermore, the isolation of the RBP-peptide-antisense sequence is obtained through magnetic streptavidin beads that are coupled with biotinylated oligonucleotide complementary to the antisense nucleic acid. The multiplicity of interactions (magnetic beads, streptavidin-biotin and base pairing) may limit the stringency and efficiency of the isolation. The RaPID method24 described here provides an alternative to these limitations in several aspects. First, the integration of the MS2 loops into the genomic loci ensures that all cells express it, thus increasing its yield. Second, it uses the high affinity interaction of the MS2-CP-GFP-SBP to twelve MS2 loop sequences and allows them to bind in vivo. Third, the fixation of the RNA-protein interaction enables more stringent washes and reduces the identification of non-specifically bound proteins. Finally, the SBP domain binds streptavidin beads with high affinity and is easily eluted with free biotin.
Insertion of the MS2 loop into the RNA is advisable between the ORF and the 3' UTR to minimize translation perturbation and to ensure expression of all loops. Six to 24 MS2 loop repeats were used previously for the visualization of mRNA localization38-40. However, the insertion of a long sequence of MS2 loops may cause changes in the processing and structure of the RNA; it might destabilize the transcript or change the repertoire of RBPs bound to it. In our experience, 12 loops are sufficient to get an efficient elution of MS2-tagged RNA20. In this respect, we note that we usually observed additional shorter transcripts in our northern hybridizations. Northern analyses with different probes revealed that these shorter forms included the MS2L and part of the 3' UTR20, indicative of premature transcription termination. Such cases are likely to reduce the efficiency of the isolation of proteins associated with the 3' UTR. Therefore, the number of inserted loops and their location within the RNA should be balanced between high efficiency pull-down and transcript stability.
The RNA has to remain intact throughout the protocol to enable efficient isolation and detection of the maximal array of bound proteins. Chances for external RNase contamination can be reduced by wearing gloves, ensuring a clean work area, preparing solution with RNase-free water, and working concisely and efficiently to reduce protocol time. Nevertheless, we found that the greatest contributor to RNA degradation is the release of cellular RNases to the lysate upon cell lysis. It was proposed that cryogenic grinding of yeast cells with a mechanical mill results in improved RNA purification41. This can be tried in cases where significant degradation is observed with this protocol.
A common drawback of pull-down assays is the isolation of numerous proteins that bind nonspecifically to the column. Therefore, the use of a yeast strain with untagged RNA is an important control. This strain expresses the MS2-CP-GFP-SBP fusion protein to further exclude proteins that bind to the fusion protein and not the target RNA. We note that this does not exclude proteins that might bind the MS2 loop itself, and these need further validation with the protein pull-down assay. Figure 1B shows that we were able to identify several proteins that bind specifically to PMP1 mRNA, and some were subsequently validated by a co-IP assay20.
The RaPID methodology is currently limited to yeast systems, utilizing its well-established site-specific integration methodologies. Site-specific integration protocols are currently being developed for genes in other systems using the CRISPR-Cas system42,43. These will be of great importance in expanding the utilization of RaPID to additional systems. Another future application of RaPID is for the isolation of RNAs that are associated with an mRNA of interest. This can be easily achieved, by subjecting the isolated material to RNA-seq rather than LC-MS/MS. Considering the important roles of small RNAs in regulating mRNA expression, this application is likely to be of great importance. Finally, improvements in mass spectrometry (MS) analysis and the use of quantitative MS methods such as SILAC will result in the quantitative measure of mRNA-bound proteins under different conditions. RaPID can be used with yeast grown in media enriched with heavy isotopes and compared to cells grown under different conditions (e.g., stress) with unlabeled media44,45. Applying the RaPID method together with quantitative MS analysis will yield accurate measures of changes in the RBP repertoire that are associated with a particular mRNA.
The authors have nothing to disclose.
We thank Prof. Jeff Gerst and Boris Slobodin for their helpful advice in setting up the RaPID protocol and providing the necessary plasmids. We also thank Dr. Avigail Atir-Lande for her help in establishing this protocol and Dr. Tamar Ziv from the Smoler Proteomics Center for her help with the LC-MS/MS analysis. We thank Prof. T.G. Kinzy (Rutgers) for the YEF3 antibody. This work was supported by grant 2011013 from the Binational Science Foundation.
Tris | sigma | T1503 | |
SDS | bio-lab | 1981232300 | |
DTT | sigma | D9779 | |
Acidic Phenol (pH 4.3) | sigma | P4682 | |
Acidic Phenol: Chloroform (5:1, pH 4.3) | sigma | P1944 | |
Chloroform | bio-lab | 3080521 | |
Formaldehyde | Frutarom | 5551820 | |
Glycine | sigma | G7126 | |
NP-40 | Calbiochem | 492016 | |
Heparin | Sigma | H3393 | |
Phenylmethylsulfonyl Flouride (PMSF) | Sigma | P7626 | |
Leupeptin | Sigma | L2884 | |
Aprotinin | Sigma | A1153 | |
Soybean Trypsin Inhibitor | Sigma | T9003 | |
Pepstatin | Sigma | P5318 | |
DNase I | Promega | M610A | |
Ribonuclease Inhibitor | Takara | 2313A | |
Glass Beads | Sartorius | BBI-8541701 | 0.4-0.6mm diameter |
Mini BeadBeater | BioSpec | Mini BeadBeater 16 | |
Guanidinium | Sigma | G4505 | |
Avidin | Sigma | A9275 | |
Streptavidin Beads | GE Healthcare | 17-5113-01 | |
Bovine serum albumin (BSA) | Sigma | A7906 | |
Yeast tRNA | Sigma | R8508 | |
Biotin | Sigma | B4501 | |
Yeast extract | Bacto | 288620 | |
peptone | Bacto | 211677 | |
Glucose | Sigma | G8270 | |
1 x Phosphate-Buffered saline (PBS) | |||
0.2 M NaOH | |||
4 x Laemmli Sample Buffer (LSB) | 0.2 M Tris-Hcl pH 6.8, 8% SDS, 0.4 M DTT, 40% glycerol, 0.04% Bromophenol-Blue. | ||
Hot phenol lysis buffer | 10 mM Tris pH 7.5, 10 mM EDTA, 0.5% SDS | ||
3 M Sodium Acetate pH 5.2 | |||
100% and 70% Ethanol (EtOH) | |||
RNase-free water | |||
RaPID lysis buffer | 20 mM Tris pH 7.5, 150 mM NaCl, 1.8 mM MgCl2, 0.5% NP-40, 5 mg/ml Heparin, 1 mM Dithiothreitol (DTT), 1 mM Phenylmethylsulfonyl Flouride (PMSF), 10 µg/ml Leupeptin, 10 µg/ml Aprotinin, 10 µg/ml Soybean Trypsin Inhibitor, 10 µg/ml Pepstatin, 20 U/ml DNase I, 100 U/ml Ribonuclease Inhibitor. | ||
2x Cross-linking reversal buffer | 100 mM Tris pH 7.4, 10 mM EDTA, 20 mM DTT, 2 % SDS. | ||
RaPID wash buffer | 20 mM Tris-HCl pH 7.5, 300 mM NaCl, 0.5% NP-40 | ||
0.5 M EDTA pH 8 | |||
Silver Stain Plus Kit | Bio-Rad | 161-0449 | For detecting proteins in polyacrylamide gels |
SD selective medium | 1.7 g/l Yeast nitrogen base with out amino acids and ammonium sulfate, 5 g/l Ammonium sulfate, 2% glucose, 350 mg/l Threonine, 40 mg/l Methionine, 40 mg/l Adenine, 50 mg/l Lysine, 50 mg/l Tryptophan, 20 mg/l Histidine, 80 mg/l Leucine, 30 mg/l Tyrosine, 40 mg/l Arginine | ||
Anti-eEF3 (EF3A,YEF3) | Gift from Kinzy TG. (UMDNJ Robert Wood Johnson Medical School) | 1:5,000 | |
Anti GFP antibody | Santa Cruz | sc-8334 | 1:3,000 |
Anti rabbit IgG-HRP conjugated | SIGMA | A9169 | 1:10,000 |