Here, we present an interactome capture protocol applied to Arabidopsis thaliana leaf mesophyll protoplasts. This method critically relies on in vivo UV crosslinking and allows for the isolation and identification of plant mRNA-binding proteins from a physiological environment.
RNA-binding proteins (RBPs) determine the fates of RNAs. They participate in all RNA biogenesis pathways and especially contribute to post-transcriptional gene regulation (PTGR) of messenger RNAs (mRNAs). In the past few years, a number of mRNA-bound proteomes from yeast and mammalian cell lines have been successfully isolated through the use of a novel method called "mRNA interactome capture," which allows for the identification of mRNA-binding proteins (mRBPs) directly from a physiological environment. The method is composed of in vivo ultraviolet (UV) crosslinking, pull-down and purification of messenger ribonucleoprotein complexes (mRNPs) by oligo(dT) beads, and the subsequent identification of the crosslinked proteins by mass spectrometry (MS). Very recently, by applying the same method, several plant mRNA-bound proteomes have been reported simultaneously from different Arabidopsis tissue sources: etiolated seedlings, leaf tissue, leaf mesophyll protoplasts, and cultured root cells. Here, we present the optimized mRNA interactome capture method for Arabidopsis thaliana leaf mesophyll protoplasts, a cell type that serves as a versatile tool for experiments that include various cellular assays. The conditions for optimal protein yield include the amount of starting tissue and the duration of UV irradiation. In the mRNA-bound proteome obtained from a medium-scale experiment (107 cells), RBPs noted to have RNA-binding capacity were found to be overrepresented, and many novel RBPs were identified. The experiment can be scaled up (109 cells), and the optimized method can be applied to other plant cell types and species to broadly isolate, catalog, and compare mRNA-bound proteomes in plants.
Eukaryotes use multiple RNA biogenesis regulatory pathways to maintain cellular biological processes. Among the known types of RNA, mRNA is very diverse and carries the coding capacity of proteins and their isoforms1.The PTGR pathway directs the fates of pre-mRNAs2,3. RBPs from different gene families control the regulation of RNA, and in PTGR, specific mRBPs guide mRNAs through direct physical interactions, forming functional mRNPs. Therefore, identifying and characterizing mRBPs and their mRNPs is critical to understanding the regulation of cellular mRNA metabolism2.Over the past three decades, various in vitro methods – including RNA electrophoretic mobility shift (REMSA) assays, systematic evolution of ligands by exponential enrichment assays (SELEX) based on library-derived constructs, RNA Bind-n-Seq (RBNS), radiolabeled or quantitative fluorescence RNA binding assays, X-ray crystallography, and NMR spectroscopy4,5,6,7,8,9 – have been widely applied to studies of RBPs, mainly from mammalian cells. The results of these studies of mammalian RBPs can be searched via the RNA-binding Protein DataBase (RBPDB), which collects the published observations10.
Although these in vitro approaches are powerful tools, they determine the bound RNA motifs from a given RNA pool of sequences and therefore are limited in their ability to discover new target RNAs. The same is true for computational strategies to predict genome-wide RBPs, which are based on the conservation of protein sequence and structure15. To overcome this, a new experimental method has been established that allows for the identification of the RNA motifs that an RBP of interest interacts with, as well as for the determination of the precise location of binding. This method, called "crosslinking and immunoprecipitation" (CLIP), is composed of in vivo UV crosslinking followed by immunoprecipitation11. Early studies have shown that photoactivation of DNA and RNA nucleotides can occur at excitation UV wavelengths greater than 245 nm. The reaction through thymidine seems to be favored (rank in order of decreased photoreactivity: dT ≥ dC > rU > rC, dA, dG)12. Using UV light with a wavelength of 254 nm (UV-C), it was observed that covalent bonds between RNA nucleotides and protein residues are created when in the range of only a few Angstroms (Å). The phenomenon is therefore called the "zero-length" crosslinking of RNA and RBP. This can be followed by a stringent purification procedure with little background13,14.
A strategy complementary to CLIP is to combine in vivo UV crosslinking with protein identification to describe the landscape of RBPs. A number of such genome-wide mRNA-bound proteomes have been isolated from yeast cells, embryonic stem cells (ESCs), and human cell lines (i.e., HEK293 and HeLa) using this novel experimental approach, called "mRNA interactome capture"18,19,20,21. The method is composed of in vivo UV crosslinking followed by mRNP purification and MS-based proteomics. By applying this strategy, many novel "moonlighting" RBPs containing non-canonical RBDs have been discovered, and it has become clear that more proteins have RNA-binding capacities than previously supposed15,16,17. The use of this method allows for new applications and for the ability to answer new biological questions when investigating RBPs. For example, a recent study has investigated the conservation of the mRNA-bound proteome (the core RBP proteome) between yeast and human cells22.
Plant RBPs have already been found to be involved in growth and development (e.g., in the post-transcriptional regulation of flowering time, the circadian clock, and gene expression in mitochondria and chloroplasts)24,25,26,27,28,29. Furthermore, they are thought to perform functions in the cellular processes responding to abiotic stresses (e.g., cold, drought, salinity, and abscisic acid (ABA))31,32,33,34. There are more than 200 predicted RBP genes in the Arabidopsis thaliana genome, based on RNA recognition motif (RRM) and K homology (KH) domain sequence motifs; in rice, approximately 250 have been noted35,36. It is notable that many predicted RBPs seem to be unique to plants (e.g., no metazoan orthologs to approximately 50% of predicted Arabidopsis RBPs containing an RRM domain)35, suggesting that many may serve new functions. The functions of most predicted RBPs remain uncharacterized23.
The isolation of mRNA-bound proteomes from Arabidopsis etiolated seedlings, leaf tissue, cultured root cells, and leaf mesophyll protoplasts through the use of mRNA interactome capture has recently been reported38,39. These studies demonstrate the strong potential of systematically cataloging functional RBPs in plants in the near future. Here, we present a protocol for mRNA interactome capture from plant protoplasts (i.e., cells without cell walls). Arabidopsis thaliana leaf mesophyll protoplasts are the major type of leaf cell. The isolated protoplasts allow optimal access of UV light to the cells. This cell type can be used in assays that transiently express proteins for functional characterization40,41. Furthermore, protoplasting has been applied to several other plant cell types and species42,43,44 (e.g., Petersson et al., 2009; Bargmann and Birnbaum, 2010; and Hong et al., 2012).
The method encompasses a total of 11 steps (Figure 1A). Arabidopsis leaf mesophyll protoplasts are first isolated (step 1) and are subsequently UV irradiated to form crosslinked mRNPs (step 2). When protoplasts are lysed under denaturing conditions (step 3), the crosslinked mRNPs are released in lysis/binding buffer and pulled down by oligo-d(T)25 beads (step 4). After several rounds of stringent washes, the mRNPs are purified and further analyzed. The denatured peptides of mRBPs are digested by proteinase K before the crosslinked mRNAs are purified and the RNA quality is verified by qRT-PCR (steps 5 and 6). After RNase treatment and protein concentration (step 7), the protein quality is controlled by SDS polyacrylamide gel electrophoresis (SDS-PAGE) and silver staining (step 8). The difference in protein band patterns can easily be visualized between a crosslinked sample (CL) and a non-crosslinked sample (non-CL; the negative control sample from protoplasts that is not subjected to UV irradiation). The identification of proteins is achieved through MS-based proteomics. The proteins from the CL sample are separated by one-dimensional polyacrylamide gel electrophoresis (1D-PAGE) to remove possible background contamination, are "in-gel digested" into short peptides using trypsin, and are purified (step 9). Nano reverse-phase liquid chromatography coupled to mass spectrometry (nano-LC-MS) allows for the determination of the amount of definitive proteins in the mRNA-bound proteome (step 10). Finally, the identified mRBPs are characterized and cataloged using bioinformatic analysis (step 11).
1. Arabidopsis Leaf Mesophyll Protoplast Isolation
NOTE: Arabidopsis leaf mesophyll protoplasts are essentially isolated as described by Yoo et al., 2007, with several modifications40.
2. In Vivo mRNA-protein Crosslinking by UV Irradiation
NOTE: Keep the non-CL sample tube on ice. The CL sample must be immediately UV irradiated.
3. Protoplast Lysis under Denaturing Conditions
4. mRNP Pull-down and Purification by Oligo-d(T)25 Beads
NOTE: All the following described materials and reagents, from step 4.1 to step 4.4, are only for one sample (i.e., one non-CL sample or one CL sample). The protoplast lysate must be added immediately to the tubes containing beads after the bead supernatant lysis/binding buffer is discarded.
5. Proteinase K Treatment and mRNA Purification
6. qRT-PCR Assay
7. RNase Treatment and mRBP Concentration
8. SDS-PAGE and Silver Staining
9. Trypsin Digest of Protein Bands and Peptide Purification
10. Nano-LC-MS
11. Catalog of the Identified mRNA-bound Proteome
We observed a characteristic halo, which surrounds the bead pellet in the CL sample, in wash step 4.3 with wash buffer 2 (Figure 1B). Although it has not been investigated, this phenomenon can probably be explained by the interference of crosslinked mRNP complexes with bead aggregation during the magnetic capture, causing a more diffuse aggregate to form. It indicates that the oligo-d(T)25 bead capture was effective14.
The significantly higher UBQ10 reference mRNA levels than 18S rRNA in both non-CL control and CL samples are shown in Figure 1C. This indicates that mRNAs are enriched in the eluent due to the capture by oligo-d(T)25 beads, which can only bind to poly(A)-tailed mRNAs. The efficiency of the oligo-d(T)25 bead capture was further supported by SDS-PAGE and silver staining (step 8), as shown in Figure 1D after RNase treatment and mRBP concentration (step 7). A difference in protein band pattern between non-CL control and CL sample lanes can clearly be observed, and the protein bands present in the non-CL control sample lane can be explained by the presence of RNase. This illustrates the strong enrichment of mRBPs in the CL sample.
The efficiency of crosslinking can be controlled by varying the time duration of UV irradiation. Sufficient crosslinking but the avoidance of protoplast damage and RNA degradation is optimal. In Figure 1D, specific protein band patterns in all CL samples were obtained when comparing different UV irradiation times (1, 3, and 5 min). Under the same UV intensity (0.00875 W/cm2), the most optimal conditions were found to be 1 or 3 min, due to the indistinguishable protein band intensities. Weaker band intensities were observed with the longer crosslinking time (5 min). We considered 1 min as the optimal condition because a shorter duration of crosslinking has the additional advantage of easier transfer and collection of protoplasts from the Petri dish. Protoplasts can be rapidly precipitated to the bottom of the Petri dish during UV irradiation because they stick to the dish bottom. This makes it more difficult to collect and transfer them to the tubes after longer UV irradiation duration.
Based on the optimized conditions, the identification of proteins from three biological replicates was subsequently achieved by qualitative and quantitative proteomics, which was described in steps 9 and 10. In the analysis by qualitative proteomics (Figure 1E, right), a total of 341 proteins were identified in CL samples, of which 36 were found both in non-CL control and CL samples, and only 8 proteins were detected in the non-CL control sample. This enormous difference in identified protein numbers between both samples is consistent with the different protein band patterns (Figure 1D), supporting the idea that the SDS-PAGE and silver-staining assay are good tools to validate the efficiency of oligo-d(T)25 bead capture. In the analysis by quantitative proteomics (Figure 1E, left), a total of 325 proteins (blue- and green-colored) showed a log2-fold change greater than 2. Among these proteins, 100 proteins (blue-colored) had small p-values above the significance level. Therefore, they were also defined as positive hits. It is notable that the p-values of another 225 proteins (green-colored) were greater than 0.05, below the significance level due to data sparsity. In other words, there were lowly abundant peptides present for each protein, with high variability of peptide intensities. However, because all of them were qualitatively detected only in CL samples, they were considered positive.
Based on the GO and InterPro databases50 (step 11), these 325 proteins were further classified into category I (ribosomal proteins), category II (main RBPs), and category III (candidate RBPs) (Figure 1F). There were 123 ribosomal proteins in category I, while in category II, 70 annotated classical RBPs were observed. These two categories—which contain most annotated proteins linking to molecular RNA binding and RNA biology (Figure 1F, right) and which occupy approximately 38% and 22% of the whole mRNA-bound proteome, respectively—indicate the high efficiency of optimized mRNA interactome capture. The last 40% (132 candidate RBPs) were classified in category III due to the lack of conventional RBDs. Furthermore, most of their roles in RNA binding and RNA biological processes have not been validated (Figure 1F, left). We suppose that proteins from this category could reveal novel functions in RNA regulations.
Figure 1: Flowchart and Results of the Optimized Method for Discovering the mRNA-bound Proteome from Arabidopsis Leaf Mesophyll Protoplasts. (A) Major steps of the whole method are listed from 1 to 11. Putative cellular and molecular processes are illustrated by the photo and cartoons. Details for each step have been intensively described in the Protocol. (B) A halo was observed surrounding the beads pellet in the CL sample, but not in the non-CL control sample, during wash step 4.3. (C) Comparison of relative UBQ10 mRNA and 18S rRNA levels in non-CL control and CL samples by qRT-PCR (values are mean ± SD (n = 3); * and **: significant differences with p <0.05 and <0.01). (D) Concentrated protein eluent of non-CL control sample compared with CL samples irradiated by a continuous-wave UV source for 1, 3, and 5 min, with the UV intensity at 0.00875 W/cm2. (E) Identification of mRNA-bound proteome by proteomics. In the quantitative proteomics performed by the Progenesis software package (left), the volcano plot displays the average log2-fold changes (CL/non-CL) and the related adjusted p-values (-log10 (adj. p-values)) of all proteins. In the qualitative proteomics performed by the Peaks software package (right), the proteins in the samples are illustrated in Venn diagrams. The quantity of the proteins is listed in numbers. The FDR at the peptide and protein levels is below 5%. The numbers of proteins inside the light-brown frames are considered positive hits. (F) Classification of three categories from the mRNA-bound proteome. The quantity of identified proteins is listed in numbers in each category. Green-colored regions are annotated proteins linked to RNA binding, while blue-colored regions refer to annotated proteins linked to RNA biology. The blank region indicates proteins with unknown functions linked to RNA binding or RNA biology. All figures above were modified from Zhang et al., 2016. Please click here to view a larger version of this figure.
We successfully applied mRNA interactome capture, developed for yeast and human cells, to plant leaf mesophyll protoplasts. Leaf mesophyll cells are the major type of ground tissue in plant leaves. The major advantage of this method is that it uses in vivo crosslinking to discover the proteins from a physiological environment.
In this protocol, we mainly present a number of optimized experimental conditions (e.g., the number of protoplasts to use as starting material and the duration of UV irradiation)50. The captured proteins in the CL samples can only be detected by SDS-PAGE and silver staining when a minimum of 107 protoplasts (medium-scale) is used (step 1.5.5). Lower concentrations do not yield an observable mRBP pattern on SDS-PAGE and should probably not be used for further analysis by proteomics. Indeed, using more than 107 cells (e.g., 109 in a large-scale experiment) is also possible20. Results from silver-staining assays suggest that UV irradiation for 1 min with 0.00875 W/cm2 of intensity generated by a continuous-wave UV source is optimal (Figure 1D). High efficiency at the critical step (i.e., mRNP pull-down and purification by oligo-d(T)25 beads, step 4) is supported by the results of the RT-qPCR, silver-staining, and MS assays (Figure 1C; 1D; and 1E, right). The combination of qualitative and quantitative proteomics can help to identify the positive RBPs in the overlap region between non-CL control and CL samples (Figure 1E). The majority of identified proteins were RBPs not previously annotated in datasets (Figure 1F). We presented these previously unknown RNA-binding proteins in category III as candidate RBPs50 (Figure 1F). Exploring the binding specificities of these candidate RBPs must be done using other methods, such as the previously mentioned CLIP methods23. One example that investigates the binding specificity of an RBP to regulate its target mRNA transcript in Arabidopsis through the use of CLIP51 can be found in Zhang et al., 2015. In addition, an alternative modified method from PAR-CLIP, called photoactivatable-ribonucleoside-enhanced crosslinking (PAR-CL, 365-nm UV-A) has also been previously recommended for the investigation of mRNA-bound proteomes from yeast and human cells14,23. PAR-CL requires 4Su, which is taken up by cells and incorporated into nascent RNAs during RNA metabolism and can be highly reactive, forming covalent bonds with amino acids52 under UV-A (365-nm) irradiation. Currently, there are no studies focusing on the toxicity of 4sU to the plant cells and the efficient uptake of exogenous 4sU into mesophyll protoplasts53. However, we believe that it will become possible to use both conventional CL (254-nm UV-C) and PAR-CL (365 nm UV-A) on plants, which will contribute to the discovery and validation of diverse RBPs from the physiological environment in the future.
The authors have nothing to disclose.
We acknowledge the lab of Prof. Joris Winderickx, who provided the UV crosslinking apparatus equipped with the conventional UV lamp. K. G. is supported by the KU Leuven research fund and acknowledges support from FWO grant G065713N.
REAGENTS | |||
0.8 M Mannitol | Sigma | M1902-500G | Primary isotonic Enzyme solution & MMg solution |
2M KCl | MERCK | Art. 4935 | Primary isotonic Enzyme solution & W5 buffer |
0.2 M MES (pH 5.7) (4-morpholineethanesulfonic acid) |
Sigma-aldrich | M2933 | Primary isotonic Enzyme solution, W5 buffer & MMg solution, Filtration sterilization |
Cellulase R10 | Yakult Pharmaceutical Industry Co., Ltd. | CELLULASE “ONOZUKA” R-10, 10 g |
Final isotonic enzyme solution |
Macerozyme R10 | Yakult Pharmaceutical Industry Co., Ltd. | MACEROZYME R-10, 10g | Final isotonic enzyme solution |
10% (w/v) BSA (Bovine Serum Albumin) |
Sigma-aldrich | A7906-100G | Final isotonic enzyme solution & Filtration sterilization |
1M CaCl2 | Chem-Lab NV | CL00.0317.1000 | Final isotonic enzyme solution, W5 buffer & Digestion buffer |
1M NaCl | Fisher Chemical | S/3160/60 | W5 buffer |
2M MgCl2 | Sigma | M8266-100G | MMg solution |
1M LiCl (Lithium Chloride) |
Acros | 199885000 | Lysis/binding buffer, Wash buffer 1, Wash buffer 2 & Low salt buffer |
5% (w/v) LiDS (Lithium Dodecyl Sulphate) |
Sigma-aldrich | L4632-25G | Lysis/binding buffer, Wash buffer 1 & Filtration sterilization |
1M DTT (Dithiothreitol) |
Thermo Fisher Scientific Wash buffer 1 & Wash buffer 2 |
307866 | Lysis/binding buffer, |
1M Tris-HCl (pH 7.5) (Tris(hydroxymethyl)aminomethane, Hydrochloric acid S.G. (HCl)) | Acros & Fisher Chemical |
167620010 & H/1200/PB15 |
Lysis/binding buffer, Wash buffer 1, Wash buffer 2, Low salt buffer & Elution buffer |
0.5 M EDTA (pH 8.0) (Ethylenediaminetetraacetic acid) | Sigma-aldrich | ED-500G | Lysis/binding buffer, Wash buffer 1, Wash buffer 2, Low salt buffer & Elution buffer |
Tween 20 | MERCK | 8.22184.0500 | Regeneration of oligo-d(T)25 beads |
0.1 M NaOH | VWR PROLABO CHEMICALS | 28244.295 | Regeneration of oligo-d(T)25 beads |
1X PBS (pH 7.4) (Phosphate Buffered Saline) containing (NaCl; KCl; Na2HPO4; KH2PO4) |
Fisher Chemical, MERCK, Sigma-aldrich & SAFC |
S/3160/60, Art. 4935, 71640-250G & 60230 |
Regeneration of oligo-d(T)25 beads |
Proteinase K solution (2 μg/μL) | Thermo Fisher Scientific | 11789020 | Protein digestion |
Loading dye | Invitrogen | LC5925 | SDS-PAGE |
qPCR master mix | Promega | A6001 | qRT-PCR assay |
RNase Cocktail | Thermo Fisher Scientific | AM2286 | RNA digestion |
Methanol | Sigma-aldrich | 322415 | Gel fixation and gel destaining |
Acetic acid | Sigma-aldrich | 537020 | Gel fixation and gel destaining |
Coomassie Brilliant Blue R-250 | Thermo Fisher Scientific | 20278 | Gel staining |
1M NH4HCO3 (Ammonium bicarbonate) |
Sigma-aldrich | 09830-500G | Gel hydration & Digestion buffer |
CH3CN (Acetonitrile) |
Sigma-aldrich | 34851-100ML | Gel dehydration & Peptide dissolving solution |
IAA (Iodoacetic acid) |
Sigma-aldrich | I4386-10G | Alkylating agent |
TFA (Trifluoroacetic acid) |
Sigma-aldrich | 302031-10X1ML | Peptide dissolving solution |
FA (Formic acid) |
Sigma-aldrich | 06554-5G | Peptide extraction |
Trypsin solution (6 ng/μL) | Promega | V5280 | Digestion buffer |
Name | Company | Catalog Number | Yorumlar |
EQUIPMENT | |||
Soil | Peltracom | LP2D | Plant growth |
Vermiculite 3 | Sibli AS | 05VERMICULIET | Plant growth |
Petri dish (150 x 20 mm) | Sarstedt | 82.1184.500 | Carrier for protoplast suspension |
0.22 μm filter | Millipore | SE2M229104 | Homogenization of final isotonic enzyme solution |
Razorblade | Agar Scientific | T585 | Rosette leaf strips |
35-75 μm nylon mesh | SEFAR NITEX | 74010 | Protoplast suspension filtration |
50 mL round bottom tubes | Sigma-aldrich | T1918-10EA | Carrier for protoplast suspension |
Hemocytometer (Bürker hemocytometer) |
MARIENFELD | 650030 | Protoplast cell counting |
UV crosslinking apparatus (HL-2000 HybriLinker) |
UVP, LLC | UVP95003101 | in vivo UV crosslinking |
UV lamp (Sankyo-Denki G8T5) |
SANKYO DENKI |
SD G8T5 | in vivo UV crosslinking |
50 mL glass syringe | FORTUNA Optima | Z314560 | Homogenization of protoplast lysate |
Narrow needle (0.9 x 25 mm) | Becton Dickinson microlance 3 | 2021-04 | Homogenization of protoplast lysate |
Rotator Model L26 | Labinco BV | 26110912 | Sample incubation by rotating |
Oligo-d(T)25 magnetic beads (5 mg/mL) |
New England BioLabs | S1419S | mRNPs and mRNAs binding and pull-down |
Magnetic rack | Invitrogen | CS15000 | mRNPs and mRNAs binding and pull-down |
Centrifugal filter units (Amicon Ultra-4 centrifugal filter units) |
EMD Millipore | UFC800308 | mRBP concentration |
Pierce Silver Stain Kit | Thermo Fisher Scientific | 24612 | Silver-staining assay |
RNA purification kit (InviTrap Spin Plant RNA Mini Kit) |
STRATEC Molecular | 1064100300 | RNA purification |
Spectrophotometer device (NanoDrop 1000 Spectrophotometer) | Thermo Fisher Scientific | ND-1000 | RNA quality and quantity |
Real-Time PCR cycler (StepOne Real-Time PCR cycler) |
Thermo Fisher Scientific | 4376600 | cDNA quantification |
µ-C18 columns (Millipore Zip Tip µ-C18 columns) |
Sigma-aldrich | 720046-960EA | Peptide purification |
Mass spectrometer (Q Exactive Hybrid Quadrupole-Orbitrap Mass Spectrometer) |
Thermo Fisher Scientific | IQLAAEGAAPFALGMAZR | Mass spectrometry-based proteomics |
Liquid chromatography instrument (Ultimate 3000 ultra-high performance liquid chromatography (UHPLC) instrument) | Thermo Fisher Scientific | ULTIM3000RSLCNANO | Mass spectrometry-based proteomics |
C18 column (Easy Spray Pepmap RSLC C18 column) |
Thermo Fisher Scientific | ES800 | Mass spectrometry-based proteomics |
C18 precolumn (Acclaim Pepmap 100 C18 precolumn) |
Thermo Fisher Scientific | 160321 | Mass spectrometry-based proteomics |
Name | Company | Catalog Number | Yorumlar |
Primers for qRT-PCR assay | Sequences | ||
UBQ10 mRNA (Li et al., 2014) |
Fw: AACTTTGGTGGTTTGTGTTTTGG Rv: TCGACTTGTCATTAGAAAGAAAGAGATAA |
||
18S rRNA (Durut et al., 2014) |
Fw: CGTAGTTGAACCTTGGGATG Rv: CACGACCCGGCCAATTA |