Enhanced single RNA imaging reveals dynamic gene expression in live animals

  1. Yucen Hu
  2. Jingxiu Xu
  3. Erqing Gao
  4. Xueyuan Fan
  5. Jieli Wei
  6. Bingcheng Ye
  7. Suhong Xu
  8. Weirui Ma

  • Curated by eLife

    eLife logo

    eLife assessment

    In this manuscript, the authors devised a new, useful mRNA-imaging approach by combining MS2 and SunTag labeling systems. The authors showed that this new method can be used to image the activation of gene expression and endogenous mRNA dynamics in live C. elegans. While the application in C. elegans has great future potential, this study is incomplete because it lacks essential characterization of the new imaging method to demonstrate that it does not interfere with RNA expression.

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Imaging endogenous mRNAs in live animals is technically challenging. Here, we describe an MS2-based signal amplification with the Suntag system that enables live-cell RNA imaging of high temporal resolution and with 8xMS2 stem-loops, which overcomes the obstacle of inserting a 1300 nt 24xMS2 into the genome for the imaging of endogenous mRNAs. Using this tool, we were able to image the activation of gene expression and the dynamics of endogenous mRNAs in the epidermis of live C. elegans .

Article activity feed

  1. Version published to 10.7554/elife.82178 on eLife
  2. eLife

    Author Response

    Reviewer #1 (Public Review):

    The authors devised a new mRNA imaging approach, MASS, and showed that it can be applied to investigate the activation of gene expression and the dynamics of endogenous mRNAs in the epidermis of live C. elegans. The approach is potentially useful, but this manuscript will benefit by addressing the following questions:

    We thank the reviewer for spending time reviewing our manuscript and for the insightful comments.

    Major comments:

    1. In Figure 1-figure supplement 1, the authors claimed that MASS could verify the lamellipodia-localization of beta-actin mRNAs. However, the image showed the opposite of the authors' claim as the concentration of beta-actin mRNA was lower in lamellipodia than the rest of the cytosol. This result disagreed with ref. 17 (Katz, Z.B. et al., Genes and Development, 2012). Hence, the authors cannot make the statement that "MASS can be readily used to image RNA molecules in live cells without affecting RNA subcellular localization". To thoroughly test this notion, the authors should image beta-actin mRNA using MASS and the conventional MS2 system side by side and calculate the polarization index in the same way as shown in Katz, Z.B. et al., Genes and Development, 2012.

    We noticed that b-ACTIN mRNAs were less polarized in our image compared to that shown in Katz, Z.B. et al. (Genes and Development, 2012). It is likely due to different cell lines being used. In the previous study, mouse embryonic fibroblasts (MEFs) were used. In our initial experiment, HeLa cells were used. Our data showed b-that ACTIN mRNAs labeled with MASS could be localized to the lamellipodia.

    As suggested by the reviewer, we performed new experiments to image b-ACTIN mRNAs using MASS and the conventional MS2 system side by side in NIH3T3 cells, a mouse fibroblast cell line (MEF cells are not available in our lab). We did not find cells with extensively polarized b-ACTIN mRNAs localization, potentially due to different cell lines. We, therefore, did not calculate the polarization index. However, we found that b-ACTIN mRNAs detected by both methods showed a similar localization pattern. These new data suggest that MASS does not affect RNA subcellular localization. We added the new results and updated Figure 1-figure supplement 3.

    1. The experiments that validate this new RNA imaging method are not sufficient. The authors need to systematically compare MASS and the MS2 system, including their RNA signal intensity, signal-to-background ratio.

    We have systematically compared MASS and the conventional MS2 system, including signal intensity and signal-to-noise ratio, and measured the velocities of mRNA movement. We found that MASS showed a similar signal-to-noise ratio and higher signal intensity to the conventional MS2 system. We have now revised the information in the text on pages 4 and 5, and in Figure 1-figure supplement 4, 5, and 6.

    1. In line with this, does beta-actin mRNA display the same behavior as in (Figure 1C-F) when the mRNA was imaged with the MS2 system? The movies do not indicate the type of motility expected of mRNA. For instance, it seems that almost all of the GFP dots, which are presumably single beta-actin mRNAs, stayed stationary over a time course of tens of seconds (Movie 1). This seems to be very different from what has been observed before. It's not clear that the dots are real mRNAs molecules. This further stresses the importance for them to compare their new imaging system with the conventional MS2 application.

    We noticed that the mobility of b-ACTIN mRNAs vary in different cells. It is possible that the mobility of mRNAs was regulated in a cell context-dependent manner.

    To confirm that the GFP foci detected are real mRNA molecules, we performed MASS combined with single-molecule RNA FISH. We found that MASS detected a similar number of GFP foci compared to the spots detected by smFISH. In addition, the majority (72%) of GFP foci colocalized with the smFISH spots of b-ACTIN-8xMS2 mRNAs. It is reported that not all MS2 stem-loop will be bound by the MCP (Wu et al., Biophysical journal 2012). As only 8xMS2 was used in MASS, it is likely that some mRNAs were not entirely bound by MCP and were not detected. On the other hand, only sixteen probes were used in the smFISH experiment, and it is possible that some mRNAs were miss labeled by smFISH. Therefore, 100% colocalization of MASS foci with the smFISH spots was hard to achieve. Thus these results suggest that GFP dots are real mRNA molecules. We have added the new data in Figure 1, Figure 1-figure supplement 1, and the text on page 3.

    We measured the velocity of (b-ACTIN mRNA movement tracked by MASS and the conventional MS2 system. We added this information in Figure 1-figure supplement 5 and to the text on pages 4 and 5. With the conventional MS2 system, we observed similar behavior to those observed by MASS.

    1. The authors claimed that a major advantage of MASS is that it has only 8xMS2 stemloops (350 nt) and overcomes "the previous obstacle of the requirement of inserting a long 1,300 nt 24xMS2". This statement lacks experimental support in this manuscript. The authors need to quantitatively compare the genomic tagging efficiency of 8xMS2 and 24xMS2.

    It has been reported by several decent studies that the knock-in efficiency decreases dramatically with increasing insert size. For example:

    ~10-fold decrease of knockin frequency with a 1085 bp compared to a 767 bp insertion of DNA fragment (Extended Data Fig.8. Wang, J. et al. Nature methods, 2022).

    ~30-fold decrease of knockin frequency with an 1122 bp compared to a 714 bp insertion of DNA fragment (Figure 3 and Table S1. Paix, A. et al. PNAS, 2017).

    In this study, we did not directly examine the knock-in efficiency of 8xMS2 and 24xMS2. Based on published data from other laboratories, we assumed that the efficiency of the knock-in of 8xMS2 (350 nt) would be higher than that of 24xMS2 (~1300 nt).

    1. MASS has the same strategy as SunRISER (Guo, Y. & Lee, R.E.C., Cell Reports Methods, 2022). Both methods use Suntag to amplify signals of MS2- or PP7-tagged RNA. The authors need to elaborate the discussions and describe the similarities and differences of the two studies. In particular, the Guo paper needs to be properly referenced.

    We have cited the paper and discussed the similarities and differences between our method and the SunRISER (page 7). Taking both studies together, Guo and we demonstrated that it is an efficient strategy to combine the MS2 system and the Suntag system as a signal amplifier for long-term and endogenous mRNA imaging in live cells.

    1. In Guo, Y. & Lee, R.E.C., Cell Reports Methods, 2022, they showed that 8XPP7 with 24XSunTag configuration led to fewer mRNA per cell (Figure 5B of the Cell Reports Methods paper). Does MASS, which has 8xMS2 with 24xSunTag, similarly lead to few mRNAs? The authors should compare the number of mRNAs detected by MASS and the conventional MS2, or by FISH.

    We compared the number of mRNAs detected by MASS and by smFISH. We performed MASS combined with single-molecule RNA FISH and found that MASS detected a similar number of GFP foci compared to the spots detected by smFISH.

    In addition, the majority (72%) of GFP foci colocalized with the smFISH spots of b-ACTIN8xMS2 mRNAs. It is reported that not all MS2 stem-loop will be bound by the MCP. As only 8xMS2 was used in MASS, it is likely that some mRNAs were not entirely bound by MCP and were not detected. On the other hand, only sixteen probes were used in the smFISH experiment, and it is possible that some mRNAs were miss labeled by smFISH. Therefore, 100% colocalization of MASS foci with the smFISH spots was hard to achieve. These data indicated that MASS could label the majority of mRNAs from a specific gene in live cells.

    We have added the new data in Figure 1, Figure 1-figure supplement 1, and the text on page 3.

    Reviewer #2 (Public Review):

    Hu et al. developed a new reagent to enhance single mRNA imaging in live cells and animal tissues. They combined an MS2-based RNA imaging technique and a Suntag system to further amplify the signal of single mRNA molecules. They used 8xMS2 stem-loops instead of the widely-used 24xMS2 stem-loops and then amplified the signal by fusing a 24xSuntag array to an MS2 coat protein (MCP). While a typical 24xMS2 approach can label a single mRNA with 48 GFPs, this technique can label a single mRNA with 384 GFPs, providing an 8-fold higher signal. Such high amplification allowed the authors to image endogenous mRNA in the epidermis of live C. elegans. While a similar approach combining PP7 and Suntag or Moontag has been published, this paper demonstrated imaging endogenous mRNA in live animals. Data mostly support the main conclusions of this paper, but some aspects of data analysis and interpretation need to be clarified and extended.

    Strengths:

    Because the authors further amplified the signal of single mRNA, this technique can be beneficial for mRNA imaging in live animal tissues where light scattering and absorption significantly reduce the signal. In addition, the size of an MS2 repeat cassette can be reduced to 8, which will make it easier to insert into an endogenous gene. Also, the MCP24xSuntag and scFv-sfGFP constructs can be expressed in previously developed 24xMS2 knock-in animal models to image single mRNAs in live tissues more easily.

    The authors performed control experiments by omitting each one of the four elements of the system: MS2, MCP, 24xSuntag, and scFV. These control data confirm that the observed GFP foci are the labeled mRNAs rather than any artifacts or GFP aggregates. And the constructs were tested in two model systems: HeLa cells and the epidermis of C. elegans. These data demonstrate that the technique may be used across different species.

    We thank the reviewer for spending time reviewing our manuscript and for the insightful comments.

    Weaknesses:

    Although the paper has strength in providing potentially useful reagents, there are some weaknesses in their approach.

    Each MCP-24xSunTag is labeled with 24 GFPs, providing enough signal to be visualized as a single spot. Although the authors showed an image of a control experiment without MS2 in Figure 1B, the authors should at least mention this potential problem and discuss how to distinguish mRNA from MCP tagged with many GFPs. MCP-24xSunTag labeled with 24 GFPs may diffuse more rapidly than the labeled mRNA. Depending on the exposure time, they may appear as single particles or smeared background, but it will certainly increase the background noise. Such trade-offs should be discussed along with the advantage of this method.

    With MCP-24xSuntag, in theory, there will be up to 24 GFP molecules tethered to one MCP molecule, which may lead to the formation of GFP puncta. However, under our imaging conditions (100 ms to 500 ms) with a spinning disk confocal microscopy, puncta of MCP24xSuntag were not detected. As the reviewer suggested, it might be because MCP24xSuntag is diffusing too fast to be detected as a spot.

    For the signal-to-noise ratio, we did more experiments and analyses. We imaged overexpressed b-ACTIN mRNAs using the conventional 24xMS2 system or MASS with different repeats of Suntag arrays (MCP-24xSuntag, MCP-12xSuntag, MCP-6xSuntag). For the conventional 24xMS2 system, we followed the previous protocol that added a nuclear localization signal (NLS) to MCP, and b-ACTIN mRNAs were nicely detected with a signal-to-noise ratio of 1.21.

    We found that MASS showed a comparable or better signal-to-noise ratio than the conventional 24xMS2 system. (MASS with MCP-24xSuntag: 1.79, MASS with MCP12xSuntag: 1.48, MASS with MCP-6xSuntag: 1.42). These data indicate that using Suntag as a signal amplifier did not increase background noise.

    Also, more quantitative image analysis would be helpful to improve the manuscript. For instance, the authors can measure the intensity of each GFP foci, show an intensity histogram, and provide some criteria to determine whether it is an MCP-24xSuntag, a single mRNA, or a transcription site. For example, it is unclear if the GFP spots in Figure 2D are transcription sites or mRNA granules.

    Under our imaging conditions, MCP-24xSuntag was not detected as GFP foci.

    We performed MASS combined with single-molecule RNA FISH and found that MASS detected a similar number of GFP foci compared to the spots detected by smFISH.

    In addition, the majority (72%) of GFP foci colocalized with the smFISH spots of b-ACTIN8xMS2 mRNAs. It is reported that not all MS2 stem-loop will be bound by the MCP. As only 8xMS2 was used in MASS, it is likely that some mRNAs were not entirely bound by MCP and were not detected. On the other hand, only sixteen probes were used in the smFISH experiment, and it is possible that some mRNAs were miss labeled by smFISH. Therefore, 100% colocalization of MASS foci with the smFISH spots was hard to achieve. These data indicated that MASS could label the majority of mRNAs from a specific gene in live cells.

    We have added the new data in Figure 1, Figure 1-figure supplement 1, and the text on page 3.

    The GFP spots in Figure 2D are not transcription sites, as they were localized in the cytoplasm, not in the nucleus. We imaged exogenous BFP-8xMS2 mRNAs in the epidermis of C. elegans and found that the size of the GFP foci of endogenous C42D4.38xMS2 mRNAs is larger than that of BFP-8xMS2 mRNAs. Those data suggest that the GFP spots in Figure 2D (C42D4.3-8xMS2 mRNA) are mRNA granules. We added those new data in Figure 2-figure supplement 5 and the text on page 7.

    Another concern is that the heavier labeling with 24xSuntag may alter the dynamics of single mRNA. Therefore, it would be desirable to perform a control experiment to compare the diffusion coefficient of mRNAs when they are labeled with MCP-GFP vs MCP- 24xSuntag+scFv-sfGFP.

    We thank the reviewer for raising this critical issue. We have performed live imaging of bACTIN mRNA using the conventional 24xMS2 system or MASS with different lengths of Suntag arrays (MCP-24xSuntag, MCP-12xSuntag, MCP-6xSuntag). We then measured the velocity of mRNA movement in each imaging condition. We found that compared to the conventional 24xMS2 system, mRNA labeled with MCP-24xSuntag or by MCP-12xSuntag showed a smaller velocity, indicating that heavier labeling affected mRNA movement speed.
    In contrast, we found that mRNAs labeled with MCP-6xSuntag showed a similar velocity to that tagged with the conventional 24xMS2 system. Those data pointed out that when MASS is used to measure the speed of mRNA movement, a short Suntag array (MCP6xSuntag) should be used. We added those new data in Figure 1-figure supplement 5 and to the text on pages 4, 5.

    The authors could briefly explain about the genes c42d4.3 and mai-1. Why were these specific genes chosen to study gene expression upon wound healing? Did the authors find any difference in the dynamics of gene expression between these two genes?

    The function of C42D4.3 and mai-1 is currently not known. Through mRNA deep sequencing, It has been shown that the expression level of C42D4.3 and mai-1 was quickly increased after wounding of the epidermis of C. elegans. We, therefore, choose those two mRNAs for imaging. We added more information about C42D4.3 and mai-1 to the text on page 6.

    We observed similar dynamics of gene expression between C42D4.3 and mai-1 (Video 7 ,8, 9).

    Reviewer #3 (Public Review):

    It is a brilliant idea to combine the MS2-MCP system with Suntag. As the authors stated, it reduces the copies of the MS2 stem loops, which can create challenges during cloning process. The Suntag system can easily amplify the signal by several to tens of folds to boost the signal for live RNA tagging. One of the best ways to claim that MASS works better than the MS2 system by itself is to compare their signal-to-noise ratios (SNRs) within the same model system, such as HeLa cells or the C. elegans epidermis. Because the authors' main argument is that they made an improvement in live RNA tagging method, it is necessary to compare it with other methods side-by-side. The authors claim that MASS can significantly improves the efficiency of CRISPR by reducing the size of the insert, it still requires knocking in several transgenes, which can be even more challenging in some model systems where there are not many selection markers are available. Another possible issue is that the bulky, heavy tagging (384 scFv-sfGFP along with 24xSuntag) can affect the mobility or stability of the target mRNAs. If it also tags preprocessed RNA in the nucleus, it may affect the RNA processing and nuclear export. A few experiments to address these possibilities will strengthen the authors' arguments. I am proposing some experiments below in detailed comments.

    We thank the reviewer for spending time reviewing our manuscript and for the insightful comments.

    1. For the experiments with HeLa cells, it is not clear whether the authors used one focal plane or the whole z-stack for their assessment of mRNA kinetics, such as fusion, fission, and anchoring. If it was from one z-plane, it was possible that many mRNAs move along the z-axis of the images to assume kinetics. If the kinetics is true, is it expected by the authors? Are beta-actin mRNAs bound to some RNA-binding proteins or clustered in RNP complexes?

    One focal plane was used in the experiments showing mRNAs' fusion, fission, and anchoring behavior. We have now added this information in the figure legend of figure 1. Yes, b-ACTIN mRNA are bound to specific RNA-binding proteins, for example, ZBP1, and it has been reported that ZBP1 forms granules with b-ACTIN mRNAs (Farina, K.L., et al., Journal of cell biology, 2003).

    1. Some quantifications on beta-actin mRNA kinetics, such as a plot of their movement speed or fusion rate, etc., would help readers better understand the behaviors of the mRNAs and assess whether the MASS tagging did not affect them.

    We thank the reviewer for raising this critical issue. We have performed live imaging of bACTIN mRNA using the conventional 24xMS2 system or MASS with different lengths of Suntag arrays (MCP-24xSuntag, MCP-12xSuntag, MCP-6xSuntag). We then measured the velocity of mRNA movement in each imaging condition. We found that compared to the conventional 24xMS2 system, mRNA labeled with MCP-24xSuntag or by MCP-12xSuntag showed a smaller velocity, indicating that heavier labeling affected mRNA movement speed.
    In contrast, we found that mRNAs labeled with MCP-6xSuntag showed a similar velocity to that tagged with the conventional 24xMS2 system. Those data pointed out that when MASS is used to measure the speed of mRNA movement, a short Suntag array (MCP6xSuntag) should be used. We added those new data in Figure 1-figure supplement 5 and the text on pages 4 and 5.

    1. Using another target gene for MASS tagging would further confirm the efficacy of the system. Assuming the authors generated a parental strain of HeLa cell, where MCP24xSuntag and scFv-sfGFP are already stably expressed (shown in Fig. 1B), CRISPR-ing in another gene should be relatively easy and fast.

    For exogenous genes, in addition to b-ACTIN, we imaged mRNAs from three more genes, C-MYC, HSPA1A, and KIF18B, with MASS in HeLa cells. For endogenous genes, we imaged C42D4.3 and mai-1 in the epidermis of C. elegans. These data indicated that MASS is able to image both exogenous and endogenous mRNAs in live cells. We have now added those new data in Figure 1-figure supplement 2, Figure 2-figure supplement 2, and to the text on pages 3, 4, and 6.

    1. Adding a complementary approach to the data presented in Fig. 1, such as qRT-PCR for beta-actin, with or without the MASS system would ensure the intense tagging did not affect the mRNA expression or stability.

    To address this question, we performed more experiments to test whether MASS affected the mRNA expression and stability. Because b-ACTIN mRNA is very stable; thus it is not suitable for measuring mRNA stability. We, therefore, tested three genes, including C-MYC, HSPA1A, and KIF18B, which were reported as medium-stable mRNAs. We found that MASS did not affect the stability of those three mRNAs in HeLa cells. We also tested the expression level and the stability of endogenous C42D4.3 mRNA in the epidermis of C. elegans and found that both the expression and the stability were not affected by MASS. We have now added those new data in Figure 1-figure supplement 2, Figure 2-figure supplement 2, and to the text on pages 3, 4, and 6.

    1. For experiments with the C. elegans epidermis, including at least one more MASS movie clip for C42D4.3 and a movie for mai-1 would be helpful for readers to appreciate the RNA labeling and its dynamics.

    We showed two movies (video 7 and video 8) and the snapshots for C42D4.3 mRNA (Figure 2D and Figure 2-figure supplement 3). We also added a movie (Video 9) for mai-1.

    1. The difference between Fig. 2D and Fig. 2-fig supp. 3 is unclear. The authors should address the different patterns of RNA signal propagation. Is it due to the laser power used too much, resulting in photobleach in Fig. 2D?

    We have noticed the difference between Figure 2D and Figure 2-figure supplement 3. In Figure 2D, GFP foci did not appear within the injury area after wounding. In Figure 2-figure supplement 3, GFP foci quickly appeared within the injury area. Although we kept the laser power setting constant when performing the laser wounding experiment, there are indeed variations in the actual laser power used. As the reviewer suggested, the difference may be due to photobleaching in Figure 2D. Alternatively, it is possible that the location of the injury site or the degree of injury could affect the dynamics of gene expression.

    However, we would like to point out that the dynamics of gene expression pattern in Figure 2D (Video 7) and Figure 2-figure supplement 3 (Video 8) is similar. GFP foci of C42D4.3 mRNAs were first detected around the injury sites. Then GFP foci gradually appeared from the area around the injury site to distal regions.

    1. Movie 7 is the key data the authors are presenting, but there are a few discrepancies between their arguments and what is seen from the movie. The authors say the RNAs are "gradually spread" (the line 120 in the manuscript). However, it seems that the green foci just appear here and there in the epidermis and the majority of them stay where they were throughout the timelapse. This pattern seems to be different from the montage in Fig. 2-fig supp. 3, which indeed looks like the mRNA spots are formed around the lesion and spread overtime. Additional explanation on this will strengthen the arguments. Given the dramatic increase of c42d4.3 mRNA abundance 1 min. after the laser wounding, there must be a tremendous boost of transcription at the active transcription sites, which should be captured as much bigger and fewer green foci that are located inside the nucleus. Is this simply because those nuclear sites are out of focus or in a similar size as mRNA foci? Regardless, this should be addressed in the discussion.

    We apologize for the confusing description of our original data. We wrote "gradually spread", but we did not mean that mRNAs were transcribed at the wounding site and moved to the distal regions. We actually mean that GFP foci first appeared close to the wounding site and more GFP foci gradually appeared at the distal regions. We have changed our writing to "the appearance of GFP foci gradually spreads from the area around the injury site to distal regions".

    For the difference between Figure 2D and Figure 2-figure supplement 3, please see our discussion for comment 6.

    For transcription, we also expected a boost of transcription after wounding. However, we failed to detect the appearance of bigger GFP foci in the nucleus. We agree with the reviewer that this is because the active nuclear sites are out of focus. The epidermis of C. elegans is a syncytium with 139 nuclei located in different regions and focal planes. With our microscopy, we were able to image only one focal plane, in which there are usually only four to ten nuclei. Therefore, it is likely that the nuclei with active transcription were out of focus. We have now discussed this point in the revised manuscript (page 6).

    1. One clear way to confirm that MASS labels mRNAs and does not affect their stability/localization is to compare the imaging data with single-molecule RNA fluorescence in situ hybridization (smFISH) that the Singer lab developed decades ago. The authors can target the endogenous c42d4.3 or mai-1 RNAs using smFISH and compare their abundance and subcellular localization patterns with their data.

    To confirm that the GFP foci detected are real mRNA molecules, we performed MASS combined with single-molecule RNA FISH and found that MASS detected a similar number of GFP foci compared to the spots detected by smFISH. In addition, the majority (72%) of GFP foci colocalized with the smFISH spots of b-ACTIN-8xMS2 mRNAs. It is reported that not all MS2 stem-loop will be bound by the MCP. As only 8xMS2 was used in MASS, it is likely that some mRNAs were not fully bound by MCP and were not detected. On the other hand, only sixteen probes were used in the smFISH experiment, and it is possible that some mRNAs were miss labeled by smFISH. Therefore, 100% colocalization of MASS foci with the smFISH spots was hard to achieve. These data indicated that MASS could detect single mRNA molecules and label the majority of mRNAs from a specific gene in live cells. We have now added the new data in Figure 1, Figure 1-figure supplement 1, and to the text on page 3.

    We performed more experiments to test whether MASS affected the mRNA expression and stability. Because b-ACTIN mRNA is very stable; thus it is not suitable for measuring mRNA stability. We, therefore, tested three genes, including C-MYC, HSPA1A, and KIF18B, which were reported as medium-stable mRNAs. We found that MASS did not affect the stability of those three mRNAs in HeLa cells. We also tested the expression level and the stability of endogenous C42D4.3 mRNA in the epidermis of C. elegans and found that both the expression and the stability were not affected by MASS. We have now added those new data in Figure 1-figure supplement 2, Figure 2-figure supplement 2, and to the text on pages 3, 4, and 6.

    To test whether MASS affected the mRNA localization, we performed new experiments to image b-ACTIN mRNAs using MASS and the conventional MS2 system side by side in NIH3T3 cells, which is a mouse fibroblast cell line. We found that b-ACTIN mRNAs showed similar localization in both methods. These new data suggest that MASS does not affect RNA subcellular localization. We have now added the new results in Figure 1-figure supplement 2.

    1. One of the main purposes to live image RNAs is to assess their dynamics. Adding some more analyses, such as the movement speed of the foci, would be helpful to show how effective this system is to assess those dynamics features.

    We thank the reviewer for raising this critical issue. We have performed live imaging of bACTIN mRNA using the conventional 24xMS2 system or MASS with different lengths of Suntag arrays (MCP-24xSuntag, MCP-12xSuntag, MCP-6xSuntag). We then measured the velocity of mRNA movement in each imaging condition. We found that compared to the conventional 24xMS2 system, mRNA labeled with MCP-24xSuntag or by MCP-12xSuntag showed a smaller velocity, indicating that heavier labeling affected mRNA movement speed.

    In contrast, we found that mRNAs labeled with MCP-6xSuntag showed a similar velocity to that tagged with the conventional 24xMS2 system. Those data pointed out that when MASS is used to measure the speed of mRNA movement, a short Suntag array (MCP6xSuntag) should be used. We added those new data in Figure 1-figure supplement 5 and to the text on pages 4 and 5.

    Reviewer #4 (Public Review):

    Hu et al introduced the MS2-Suntag system into C. elegans to tag and image the dynamics of individual mRNAs in a live animal. The system involves CRISPR-based integration of 8x MS2 motifs into the target gene, and two transgene constructs (MCP-Suntag; scFv-sfGFP) that can potentially recruit up to 384 GFP molecule to an mRNA to amplify the fluorescent signal. The images show very high signal to background ratio, indicating a large range of optimization to control phototoxicity for live imaging and/or artifacts caused by excessive labeling. The use of epidermal wound repair as a case study provides a simplified temporal context to interpret the results, such as the initiation of transcription upon wounding. The preliminary results also reveal potentially novel biology such as localization of mRNAs and dynamic RNP complexes in wound response and repair. On the other hand, the system recruits a large protein complex to an mRNA molecule, an immediate question is to what extent it may interfere with in vivo regulation. Phenotypic assays, e.g., in development and wound repair, would have been a powerful argument but are not explored. In all, C. elegans is powerful system for live imaging, and the genome is rich in RNA binding proteins as well as miRNAs and other small RNAs for rich posttranscriptional regulation. The manuscript provides an important technical progress and valuable resource for the field to study posttranscriptional regulation in vivo.

    We thank the reviewer for spending time reviewing our manuscript and for the insightful comments.

  3. eLife

    eLife assessment

    In this manuscript, the authors devised a new, useful mRNA-imaging approach by combining MS2 and SunTag labeling systems. The authors showed that this new method can be used to image the activation of gene expression and endogenous mRNA dynamics in live C. elegans. While the application in C. elegans has great future potential, this study is incomplete because it lacks essential characterization of the new imaging method to demonstrate that it does not interfere with RNA expression.

  4. eLife

    Reviewer #1 (Public Review):

    The authors devised a new mRNA imaging approach, MASS, and showed that it can be applied to investigate the activation of gene expression and the dynamics of endogenous mRNAs in the epidermis of live C. elegans. The approach is potentially useful, but this manuscript will benefit by addressing the following questions:

    Major comments:

    1. In Figure 1-figure supplement 1, the authors claimed that MASS could verify the lamellipodia-localization of beta-actin mRNAs. However, the image showed the opposite of the authors' claim as the concentration of beta-actin mRNA was lower in lamellipodia than the rest of the cytosol. This result disagreed with ref. 17 (Katz, Z.B. et al., Genes and Development, 2012). Hence, the authors cannot make the statement that "MASS can be readily used to image RNA molecules in live cells without affecting RNA subcellular localization". To thoroughly test this notion, the authors should image beta-actin mRNA using MASS and the conventional MS2 system side by side and calculate the polarization index in the same way as shown in Katz, Z.B. et al., Genes and Development, 2012.

    2. The experiments that validate this new RNA imaging method are not sufficient. The authors need to systematically compare MASS and the MS2 system, including their RNA signal intensity, signal-to-background ratio.

    3. In line with this, does beta-actin mRNA display the same behavior as in (Figure 1C-F) when the mRNA was imaged with the MS2 system? The movies do not indicate the type of motility expected of mRNA. For instance, it seems that almost all of the GFP dots, which are presumably single beta-actin mRNAs, stayed stationary over a time course of tens of seconds (Movie 1). This seems to be very different from what has been observed before. It's not clear that the dots are real mRNAs molecules. This further stresses the importance for them to compare their new imaging system with the conventional MS2 application.

    4. The authors claimed that a major advantage of MASS is that it has only 8xMS2 stem loops (350 nt) and overcomes "the previous obstacle of the requirement of inserting a long 1,300 nt 24xMS2". This statement lacks experimental support in this manuscript. The authors need to quantitatively compare the genomic tagging efficiency of 8xMS2 and 24xMS2.

    5. MASS has the same strategy as SunRISER (Guo, Y. & Lee, R.E.C., Cell Reports Methods, 2022). Both methods use Suntag to amplify signals of MS2- or PP7-tagged RNA. The authors need to elaborate the discussions and describe the similarities and differences of the two studies. In particular, the Guo paper needs to be properly referenced.

    6. In Guo, Y. & Lee, R.E.C., Cell Reports Methods, 2022, they showed that 8XPP7 with 24XSunTag configuration led to fewer mRNA per cell (Figure 5B of the Cell Reports Methods paper). Does MASS, which has 8xMS2 with 24xSunTag, similarly lead to few mRNAs? The authors should compare the number of mRNAs detected by MASS and the conventional MS2, or by FISH.

  5. eLife

    Reviewer #2 (Public Review):

    Hu et al. developed a new reagent to enhance single mRNA imaging in live cells and animal tissues. They combined an MS2-based RNA imaging technique and a Suntag system to further amplify the signal of single mRNA molecules. They used 8xMS2 stem-loops instead of the widely-used 24xMS2 stem-loops and then amplified the signal by fusing a 24xSuntag array to an MS2 coat protein (MCP). While a typical 24xMS2 approach can label a single mRNA with 48 GFPs, this technique can label a single mRNA with 384 GFPs, providing an 8-fold higher signal. Such high amplification allowed the authors to image endogenous mRNA in the epidermis of live C. elegans. While a similar approach combining PP7 and Suntag or Moontag has been published, this paper demonstrated imaging endogenous mRNA in live animals. Data mostly support the main conclusions of this paper, but some aspects of data analysis and interpretation need to be clarified and extended.

    Strengths:
    Because the authors further amplified the signal of single mRNA, this technique can be beneficial for mRNA imaging in live animal tissues where light scattering and absorption significantly reduce the signal. In addition, the size of an MS2 repeat cassette can be reduced to 8, which will make it easier to insert into an endogenous gene. Also, the MCP-24xSuntag and scFv-sfGFP constructs can be expressed in previously developed 24xMS2 knock-in animal models to image single mRNAs in live tissues more easily.

    The authors performed control experiments by omitting each one of the four elements of the system: MS2, MCP, 24xSuntag, and scFV. These control data confirm that the observed GFP foci are the labeled mRNAs rather than any artifacts or GFP aggregates. And the constructs were tested in two model systems: HeLa cells and the epidermis of C. elegans. These data demonstrate that the technique may be used across different species.

    Weaknesses:
    Although the paper has strength in providing potentially useful reagents, there are some weaknesses in their approach.

    Each MCP-24xSunTag is labeled with 24 GFPs, providing enough signal to be visualized as a single spot. Although the authors showed an image of a control experiment without MS2 in Figure 1B, the authors should at least mention this potential problem and discuss how to distinguish mRNA from MCP tagged with many GFPs. MCP-24xSunTag labeled with 24 GFPs may diffuse more rapidly than the labeled mRNA. Depending on the exposure time, they may appear as single particles or smeared background, but it will certainly increase the background noise. Such trade-offs should be discussed along with the advantage of this method.

    Also, more quantitative image analysis would be helpful to improve the manuscript. For instance, the authors can measure the intensity of each GFP foci, show an intensity histogram, and provide some criteria to determine whether it is an MCP-24xSuntag, a single mRNA, or a transcription site. For example, it is unclear if the GFP spots in Figure 2D are transcription sites or mRNA granules.

    Another concern is that the heavier labeling with 24xSuntag may alter the dynamics of single mRNA. Therefore, it would be desirable to perform a control experiment to compare the diffusion coefficient of mRNAs when they are labeled with MCP-GFP vs MCP-24xSuntag+scFv-sfGFP.

    The authors could briefly explain about the genes c42d4.3 and mai-1. Why were these specific genes chosen to study gene expression upon wound healing? Did the authors find any difference in the dynamics of gene expression between these two genes?

  6. eLife

    Reviewer #3 (Public Review):

    It is a brilliant idea to combine the MS2-MCP system with Suntag. As the authors stated, it reduces the copies of the MS2 stem loops, which can create challenges during cloning process. The Suntag system can easily amplify the signal by several to tens of folds to boost the signal for live RNA tagging. One of the best ways to claim that MASS works better than the MS2 system by itself is to compare their signal-to-noise ratios (SNRs) within the same model system, such as HeLa cells or the C. elegans epidermis. Because the authors' main argument is that they made an improvement in live RNA tagging method, it is necessary to compare it with other methods side-by-side. The authors claim that MASS can significantly improves the efficiency of CRISPR by reducing the size of the insert, it still requires knocking in several transgenes, which can be even more challenging in some model systems where there are not many selection markers are available. Another possible issue is that the bulky, heavy tagging (384 scFv-sfGFP along with 24xSuntag) can affect the mobility or stability of the target mRNAs. If it also tags preprocessed RNA in the nucleus, it may affect the RNA processing and nuclear export. A few experiments to address these possibilities will strengthen the authors' arguments. I am proposing some experiments below in detailed comments.

    1. For the experiments with HeLa cells, it is not clear whether the authors used one focal plane or the whole z-stack for their assessment of mRNA kinetics, such as fusion, fission, and anchoring. If it was from one z-plane, it was possible that many mRNAs move along the z-axis of the images to assume kinetics. If the kinetics is true, is it expected by the authors? Are beta-actin mRNAs bound to some RNA-binding proteins or clustered in RNP complexes?
    2. Some quantifications on beta-actin mRNA kinetics, such as a plot of their movement speed or fusion rate, etc., would help readers better understand the behaviors of the mRNAs and assess whether the MASS tagging did not affect them.
    3. Using another target gene for MASS tagging would further confirm the efficacy of the system. Assuming the authors generated a parental strain of HeLa cell, where MCP-24xSuntag and scFv-sfGFP are already stably expressed (shown in Fig. 1B), CRISPR-ing in another gene should be relatively easy and fast.
    4. Adding a complementary approach to the data presented in Fig. 1, such as qRT-PCR for beta-actin, with or without the MASS system would ensure the intense tagging did not affect the mRNA expression or stability.
    5. For experiments with the C. elegans epidermis, including at least one more MASS movie clip for c42d4.3 and a movie for mai-1 would be helpful for readers to appreciate the RNA labeling and its dynamics.
    6. The difference between Fig. 2D and Fig. 2-fig supp. 3 is unclear. The authors should address the different patterns of RNA signal propagation. Is it due to the laser power used too much, resulting in photobleach in Fig. 2D?
    7. Movie 7 is the key data the authors are presenting, but there are a few discrepancies between their arguments and what is seen from the movie. The authors say the RNAs are "gradually spread" (the line 120 in the manuscript). However, it seems that the green foci just appear here and there in the epidermis and the majority of them stay where they were throughout the timelapse. This pattern seems to be different from the montage in Fig. 2-fig supp. 3, which indeed looks like the mRNA spots are formed around the lesion and spread overtime. Additional explanation on this will strengthen the arguments. Given the dramatic increase of c42d4.3 mRNA abundance 1 min. after the laser wounding, there must be a tremendous boost of transcription at the active transcription sites, which should be captured as much bigger and fewer green foci that are located inside the nucleus. Is this simply because those nuclear sites are out of focus or in a similar size as mRNA foci? Regardless, this should be addressed in the discussion.
    8. One clear way to confirm that MASS labels mRNAs and does not affect their stability/localization is to compare the imaging data with single-molecule RNA fluorescence in situ hybridization (smFISH) that the Singer lab developed decades ago. The authors can target the endogenous c42d4.3 or mai-1 RNAs using smFISH and compare their abundance and subcellular localization patterns with their data.
    9. One of the main purposes to live image RNAs is to assess their dynamics. Adding some more analyses, such as the movement speed of the foci, would be helpful to show how effective this system is to assess those dynamics features.

  7. eLife

    Reviewer #4 (Public Review):

    Hu et al introduced the MS2-Suntag system into C. elegans to tag and image the dynamics of individual mRNAs in a live animal. The system involves CRISPR-based integration of 8x MS2 motifs into the target gene, and two transgene constructs (MCP-Suntag; scFv-sfGFP) that can potentially recruit up to 384 GFP molecule to an mRNA to amplify the fluorescent signal. The images show very high signal to background ratio, indicating a large range of optimization to control phototoxicity for live imaging and/or artifacts caused by excessive labeling. The use of epidermal wound repair as a case study provides a simplified temporal context to interpret the results, such as the initiation of transcription upon wounding. The preliminary results also reveal potentially novel biology such as localization of mRNAs and dynamic RNP complexes in wound response and repair. On the other hand, the system recruits a large protein complex to an mRNA molecule, an immediate question is to what extent it may interfere with in vivo regulation. Phenotypic assays, e.g., in development and wound repair, would have been a powerful argument but are not explored. In all, C. elegans is powerful system for live imaging, and the genome is rich in RNA binding proteins as well as miRNAs and other small RNAs for rich posttranscriptional regulation. The manuscript provides an important technical progress and valuable resource for the field to study posttranscriptional regulation in vivo.

  8. Version published to 10.1101/2022.07.26.501631v1 on bioRxiv