In the past two decades, numerous RNA tagging methods have been developed for live-cell imaging of RNA molecules (reviewed in (Moon et al., 2016; Rath and Rentmeister, 2015)). Among them, the MS2-GFP system, using the highly specific interaction between the MS2 coat protein (MCP) and the MS2 binding site (MBS) (Beach et al., 1999; Bertrand et al., 1998), has been widely used for single mRNA imaging. In the MS2-GFP system, a target mRNA is tagged with multiple repeats of MBS, and MCP fused with a fluorescent protein (MCP-FP) is co-expressed in the same cell. The number of MBS repeats directly determines the amplification factor of the signal; 24 repeats of MBS are commonly used for single RNA tracking with a high signal-to-noise ratio (SNR) (Fusco et al., 2003; Park et al., 2010). This method enables the monitoring of dynamic changes in mRNA transcription, localization, translation, and decay within live cells (reviewed in (Tutucci et al., 2018a; Vera et al., 2016)).

However, concerns have been recently raised about possible artifacts of the MS2-GFP system when used in Escherichia coli (E. coli ) (Golding and Cox, 2004) and Saccharomyces cerevisiae (S. cerevisiae)(Garcia and Parker, 2015; 2016; Haimovich et al., 2016; Heinrich et al., 2017). The tight binding between MBS and MCP may hinder the proper degradation of the tagged mRNA, which could lead to the accumulation of mRNA decay fragments. This raises an important issue because it can complicate the interpretation of live-cell imaging data for monitoring the dynamics of mRNA movement and associated processes. In this context, Garcia and Parker presented evidence that MBS stem-loops bound by MCP block the exonuclease XrnI activity on the mRNA substrate in budding yeast strains (Garcia and Parker, 2015). By using Northern blot analysis, they examined the degradation patterns of the MBS-tagged QCR8 and PGK1 mRNAs expressed from TET-off plasmids and the endogenous MFA2 and ASH1 mRNAs labeled with the MBS cassette knocked in the 3′ untranslated region (UTR). In this study, a substantial accumulation of 3′ mRNA fragments containing the MBS was observed when MCP-3×GFP was co-expressed in the yeast cells. This population of mRNA decay fragments hybridized to the probes targeting the 3′ side of the MBS array and required the cytoplasmic 5′ to 3′ exonuclease XrnI for its accumulation.

Supporting Garcia and Parker, Heinrich et al. also reported that the introduction of the MBS or its orthologous PP7 binding site (PBS) stem-loops can affect the processing and localization of PGK1 mRNA in budding yeast (Heinrich et al., 2017). Using single molecule fluorescence in situ hybridization (smFISH), they found that 3′-end stem-loop-containing fragments accumulate in processing bodies (P-bodies) and that the number of nascent mRNAs increases upon stem-loop tagging in glucose-depleted conditions. PBS-tagged FBA1 and GFA1 mRNAs also showed a similar aberrant enrichment in the P-bodies. However, the increase in the nascent mRNA number was observed only for the abundant mRNA species, such as PGK1 and FBA1 mRNAs, but not for the moderately expressed mRNAs, such as GFA1 mRNA.

Based on RNA-binding protein purification and identification (RaPID) (Slobodin and Gerst, 2010) and real-time quantitative polymerase chain reaction (RT-qPCR) experiments, Haimovich et al. noted that the accumulation of 3′ decay products may not be a general phenomenon for endogenously expressed MS2-tagged mRNAs; rather, it could be an artifact arising from the overexpression of the tagged mRNAs (Haimovich et al., 2016). However, Northern blot analysis still detected 3′ decay fragments from several endogenously expressed MS2-labeled mRNAs in yeast (Garcia and Parker, 2016). Yet, it is still an open question whether this result can be generalized to other mRNAs and other organisms.

As a case study in a mammalian system, we have examined the endogenously expressed MBS-tagged β-actin mRNA using Actb-MBS (Lionnet et al., 2011) and MBS×MCP mice (Park et al., 2014). The Actb-MBS and MBS×MCP mice have been extensively used to investigate transcription (Cho et al., 2016; Kalo et al., 2015; Lionnet et al., 2011), nuclear export (Grunwald and Singer, 2010), nuclear and cytoplasmic transport (Lee and Park, 2018; Monnier et al., 2015; Park et al., 2014; Smith et al., 2015; Song et al., 2018), protein-RNA interactions (Eliscovich et al., 2017; Wu et al., 2015), and localization and local translation of β-actin mRNA (Buxbaum et al., 2014; Katz et al., 2012; 2016; Park et al., 2012; Yoon et al., 2016). We used Northern blot analysis following the approach used for budding yeast (Garcia and Parker, 2015) and found no noticeable accumulation of 3′ decay products containing the MBS in either mouse embryonic fibroblasts (MEFs) or various organ tissues from the mouse models. Our results suggest that the Actb-MBS and MBS×MCP mice can be used without concern for possible accumulation of MBS RNA decay fragments.

Animals

Animal care and experiments were performed in accordance with the protocols approved by the Institutional Animal Care and Use Committee (IACUC) at Seoul National University. The Actb-MBS knock-in and MBS×MCP hybrid mouse models were housed under specific-pathogen-free (SPF) conditions in the research animal facility at Seoul National University. For comparison of different organs, 3–8-week-old male mice were used.

Tissue cultures

The wild-type (WT) and Actb-MBS MEFs were cultured from E14 embryos and immortalized by transfecting a plasmid expressing SV40 large T antigen as described previously (Lionnet et al., 2011). The MBS×MCP MEF cell line was generated by infecting the Actb-MBS MEF with a lentiviral vector expressing MCP-GFP under the ubiquitin C promoter. Cells were cultured in Dulbecco’s modification of Eagle’s medium (DMEM) supplemented with 10% FBS, 1% Glutamax, and 1% Pen-Strep at 37°C and 5% CO₂.

RNA extraction and Northern Blot

To extract total RNA from cell cultures, we added TRI reagent (or TRIzol) directly to the cells after removing the media and homogenized them by pipetting up and down (Rio et al., 2010). For RNA extraction from mouse organs, tissues were snap-frozen with liquid nitrogen and pulverized by mortar and pestle, followed by the addition of TRI reagent. RNA quality was assured via spectrophotometry and agarose gel electrophoresis.

For Northern blotting of β-actin mRNA, a total of 10–15 μg RNA was separated in 1.0% agarose gels and transferred to positively charged nylon membrane (Roche) using a semi-dry transfer kit (BioRad). After 30 min of prehybridization in SES1 solution (0.5 M Sodium phosphate, 7% (w/v) SDS, 1 mM EDTA, pH 7.2), the blot was incubated overnight with the ³²P end-labeled probes. For washing the membranes, 20×SSPE solution (3.6 M NaCl, 0.2 M NaH₂PO₄, 20 mM EDTA, pH 7.4) was prepared and diluted to 6×SSPE solution. After washing three times with 6×SSPE solution (30 min at 37°C and twice more at 42°C), the membranes were placed in a phosphor screen cassette to obtain autoradiography images. The oligonucleotide probe sequences used in this study are as follows.

• GSP1: CAAAACAATGTACAAAGTCCTCAGCCACATTTGTAG AACTTTGGGGGATG

• GSP2: CCCAGGGAGACCAAAGCCTTCATACATCAAGTTGG GGGGACAAAAAAAAG

• GSP3: ACATCAAGTTGGGGGGACAAAAAAAAGGGAGGC CTCAGACC

• LK51: TTTCTAGGCAATTAGGTACCTTAGGATCTAATGAACCC GGGAATACTGCAG

Real-time quantitative polymerase chain reaction (RT-qPCR)

The WT and Actb-MBS MEF cells were cultured at a density of 2.2 × 10⁶ cells per 100 mm dish followed by the addition of 100 μM DRB (5,6-Dichlorobenzimidazole 1-β-D-ribofuranoside)(Sigma). Total RNA was extracted from cells at different time points of the DRB treatment using the TRIzol reagent (Invitrogen) and then treated with DNase I (Roche) at 37°C for 20 min to digest any genomic DNA followed by addition of 8 mM EDTA and heat-inactivation at 75°C for 10 min. cDNA was prepared from the total mRNA by reverse transcription using TOPscript Reverse Transcriptase enzyme (enzynomics) with 100 μM random hexamers. The reaction was done with the SimpliAmp Thermal Cycler (Applied Biosystems) using the following cycling conditions: annealing at 25°C for 10 min, reaction at 42°C for 60 min and inactivation at 95°C for 5 min.

A 2-step RT-qPCR reaction was done with 100 μM of each primer (mouse β-actin: FW 5′-CCACTGCCGCATCCTCTTCC-3′, REV 5′-CTCGTTGCCAATAGTGATGACCTG-3′; mouse GAPDH: FW 5′-CATGGCCTTCCGTGTTCCTA-3′, REV 5′-GCGGCACGTCAGATCCA-3′) and TOPreal qPCR 2× PreMIX (SYBR Green with high ROX) (enzynomics) using the StepOnePlus Real-Time PCR System (Thermo Fisher). Samples were run in triplicates and analyzed using the StepOnePlus Real-Time PCR software. β-actin mRNA levels were analyzed by the 2^−ΔΔCt method (Livak method) and normalized to the GAPDH mRNA levels.

To compare the size distribution of the labeled and unlabeled β-actin mRNA, we extracted RNA from the cells and tissues of WT, Actb-MBS (Lionnet et al., 2011), and MBS×MCP mice (Park et al., 2014). In the Actb-MBS mouse, the cassette containing 24 repeats of MBS (24×MBS) is knocked-in to the 3′UTR of the Actb gene 441 bp downstream of the stop codon and after the zipcode (Kislauskis et al., 1994) responsible for mRNA localization. The MCP mouse is a transgenic mouse that expresses MCP-GFP in every cell type. The Actb-MBS and MCP mice were crossed and bred to double homozygosity to generate the MBS×MCP mouse line (Park et al., 2014). In the MBS×MCP mouse, all endogenous β-actin mRNAs are labeled with up to 48 MCP-GFPs because each MBS binds a dimer of MCP-GFP. Previously, Lionnet et al. confirmed the successful knock-in of the 24×MBS and the proper size of the full-length modified mRNA transcript (Lionnet et al., 2011). The 24×MBS cassette is made of 12 repeats of a tandem MBS doublet unit. Each unit has two 19-ribonucleotide (nt)-long stem-loops that are connected by a 20-nt spacer, and a 51-nt linker connects the doublet units. For the Northern blot analysis, we used three gene-specific probes that hybridize to the 5′ side (GSP1) and the 3′ side (GSP2 and GSP3) of the 24×MBS cassette and the LK51 probe that hybridizes to the 51-nt linker (Fig. 1). Because of the lower sequence specificity of GSP2 and GSP3, we designed two probes for the 3′ side of the 24×MBS cassette. While each GSP probe has only one hybridization site in the β-actin transcript, the LK51 probe has 11 hybridization sites in the modified full-length mRNA, providing a higher signal intensity and sensitivity.

First, we compared the Northern blot images of the immortalized MEF cell lines obtained from the WT, homozygous Actb-MBS, and double homozygous MBS×MCP mice (Fig. 2). Unlike the yeast model studies (Garcia and Parker, 2015, 2016), we could not observe the accumulation of the 3′ decay products in any of the three cell lines. Some smearing due to general RNA degradation was observed with the LK51 and 18S rRNA probes because they had a much stronger signal than that of the GSP probes (Fig. 2A). However, there were no noticeable 3′ decay fragments that hybridized with the GSP2 probe in either the Actb-MBS or MBS×MCP MEF cell lines compared with the WT MEF. By increasing the photomultiplier tube (PMT) voltage, we confirmed that little difference was found in the degradation patterns of the labeled and unlabeled β-actin transcripts (Fig. 2B). The overall stability of the WT and MBS labeled β-actin mRNA was also similar to each other at different time points after inhibition of transcription by DRB (p ≤ 0.05 by ANOVA, Fig. 2C).

Next, we performed Northern blot analysis for whole brain lysates from the WT and MBS×MCP mice. We randomly chose three mice from each group and sacrificed them to obtain brain samples. The total RNA extract from the brain samples were probed with the LK51, GSP1, and GSP2 probes (Fig. 3). There was no noticeable difference between the WT and the MS2-GFP-labeled β-actin mRNA transcripts in the brain tissue samples.

Finally, we examined the β-actin mRNA degradation patterns in several different organs from the WT, Actb-MBS, and MBS×MCP mice. Again, no clear organ-specific feature was noted in regard to the integrity of the mRNA (Fig. 4). Six mouse organs, including brain, testis, kidney, liver, heart, and lung, showed no noticeable accumulation of the mRNA decay product in the Actb-MBS and MBS×MCP mouse lines compared with the WT mice. This finding confirmed that MCP-GFP binding negligibly influences the RNA decay in this mouse model system.

In this report, we examined the size distribution of MS2-GFP labeled β-actin mRNAs that were endogenously expressed in murine cells and tissues. Although recent studies in budding yeast models have raised concerns about aberrant mRNA degradation upon MS2-GFP labeling (Garcia and Parker, 2015, 2016; Haimovich et al., 2016; Heinrich et al., 2017), we did not find such accumulation of mRNA decay fragments in the Actb-MBS knock-in and MBS×MCP mice.

While we were preparing this report, Tuttucci et al. introduced a newer version of MBS called MBSV6 (Tutucci et al., 2018b). MBSV6 has a uridine (U) at position −5 of the loop instead of a cytosine (C) in the previous versions of MBS. The C-to-U replacement reduced the binding affinity between the stem-loop sequence and MCP by 10-fold, which prevented the accumulation of mRNA decay fragments in yeasts. MBSV6 was successfully used for tagging both constitutively expressed mRNAs (MDN1 and DOA1) and highly regulated mRNAs (GAL1 and ASH1) without affecting their normal degradation pattern. Therefore, the use of MBSV6 is preferable for single RNA imaging in yeasts. However, in mammalian cells, the authors did not observe MBS aggregates using either the old or the new version of MBS. Additionally, Horvathova et al. showed that the MS2 and PP7 stem-loops bound to their respective coat proteins did not stabilize any degradation intermediate of the reporter mRNA in HeLa cells (Horvathova et al., 2017).

Because mRNA is relatively short-lived in bacteria and yeasts, degradation of MBS could be rate-limiting for the decay of the labeled mRNA (Tutucci et al., 2018b). The average half-life of mRNA in E. coli is approximately 5 min (Moran et al., 2013) and that in S. cerevisiae is approximately 23 min (Wang et al., 2002). However, the half-lives of most mRNAs in mammalian cells are much longer. The median half-life of mRNA in mouse embryonic stem cells is approximately 7.1 hours (Sharova et al., 2009) and that in human cells is approximately 10 hours (Yang et al., 2003). In particular, β-actin mRNA is well-known to be stable with a half-life of 9 hours (Reuner et al., 1995). Because of the longer half-lives of mRNAs in general, MS2-GFP labeling may be less prone to the accumulation of decay fragments in mammalian cells.

In addition to the lifetime of mRNA, the expression level and the insertion position of the MBS cassette are important factors to consider when using the MS2-GFP tagging system. Both the Parker and Gerst groups agreed that overexpression of MS2-labeled mRNAs caused the accumulation of 3′ decay fragments (Garcia and Parker, 2015; Haimovich et al., 2016). The insertion position of the MBS cassette is typically chosen in the 3′UTR to minimize any perturbation in the transcription and translation of the mRNA. However, there are many cis-regulatory elements and trans-acting factors in the 3′UTR, the disruption of which may cause the mislocalization of the mRNA and dysregulation of mRNA stability. Because the same tagging method could work differently under the context of the experiment, e.g., the model organism, target mRNA species, and the labeling site, it is important to verify whether the labeled mRNA faithfully recapitulates the behavior of the original molecule before further extensive and costly analysis.

In summary, we have confirmed that the Actb-MBS knock-in and MBS×MCP hybrid mouse models can be used to image the dynamics of β-actin mRNA without the worrisome accumulation of mRNA decay fragments. Using Northern blot analysis, we examined the integrity of the labeled mRNA in immortalized mouse embryonic cell lines and various organ tissues. In addition to Northern blot, other experimental techniques, such as single molecule FISH and RT-qPCR, have been used to validate the MS2 and other orthologous tagging systems (Das et al., 2018; Haimovich et al., 2016; Heinrich et al., 2017; Lionnet et al., 2011; Tutucci et al., 2018b). Various fluorescent RNA tagging methods are being continuously developed and refined to visualize diverse biological phenomena in living cells. As exemplified by the MS2-GFP system, it is imperative to validate the labeling technique in the context of each study. Single RNA imaging, when conducted with caution, has much potential to reveal a thus-far unrecognized complexity in the life-long path of RNA inside a cell.

This work was supported by the Creative-Pioneering Researchers Program through Seoul National University and the Howard Hughes Medical Institute (HHMI)-Wellcome International Scholar Awards from the Wellcome Trust [208468/Z/17/Z].