CoChat

Stochastic mRNA Synthesis in Mammalian Cells

Open access
Raj, Arjun, Peskin, Charles S, Tranchina, Daniel, Vargas, Diana Y, Tyagi, Sanjay · 2006 · PLoS Biology

The DOI matches its Crossref record; 40 of the 40 references it cites were found in a registry or an index; none of the 59 sources its claims cite is checked for support yet.

Abstract

Individual cells in genetically homogeneous populations have been found to express different numbers of molecules of specific proteins. We investigated the origins of these variations in mammalian cells by counting individual molecules of mRNA produced from a reporter gene that was stably integrated into the cell's genome. We found that there are massive variations in the number of mRNA molecules present in each cell. These variations occur because mRNAs are synthesized in short but intense bursts of transcription beginning when the gene transitions from an inactive to an active state and ending when they transition back to the inactive state. We show that these transitions are intrinsically random and not due to global, extrinsic factors such as the levels of transcriptional activators. Moreover, the gene activation causes burst-like expression of all genes within a wider genomic locus. We further found that bursts are also exhibited in the synthesis of natural genes. The bursts of mRNA expression can be buffered at the protein level by slow protein degradation rates. A stochastic model of gene activation and inactivation was developed to explain the statistical properties of the bursts. The model showed that increasing the level of transcription factors increases the average size of the bursts rather than their frequency. These results demonstrate that gene expression in mammalian cells is subject to large, intrinsically random fluctuations and raise questions about how cells are able to function in the face of such noise.

Overview

CoChat Summary

TL;DR

Single-molecule FISH in mammalian CHO cells showed that mRNA is synthesized in infrequent, intense transcriptional bursts, producing large cell-to-cell variation. These bursts are intrinsically random, can affect an entire genomic locus, occur for natural genes, and can be buffered at the protein level by slow protein degradation.

Key findings

  • mRNA numbers varied massively from cell to cell: mean ≈40 mRNA molecules per cell and variance ≈1,600 molecules squared, inconsistent with a Poisson distribution expected from constant-rate synthesis.
  • Cells with active transcription foci had an average of 244 mRNA per cell versus 33 mRNA per cell in cells without foci (p < 10⁻⁴); cells with active sites also had a higher nuclear mRNA fraction (35%, 17 cells) than cells without active sites (25%, 22 cells) (p = 0.0093).
  • Bursts were intrinsically random rather than due to global extrinsic factors: two reporters at separate loci showed uncorrelated expression (R = 0.056, p = 0.57), whereas two reporters at the same locus showed strong correlation (R = 0.89, p = 1.2 × 10⁻³⁸). RNA polymerase II large-subunit mRNA also burst (variance over 50 times the mean) and was uncorrelated with reporter mRNA (R = 0.083, p = 0.41).
  • Model fitting indicated that increasing transcription factor levels or binding sites increases average burst size rather than burst frequency; mRNA half-life was measured as approximately 4 ± 1 h. Actively degraded GFP correlated strongly with mRNA (R > 0.78, p < 2.6 × 10⁻²²), while nondegraded CFP correlated less strongly (R = 0.35, p = 1.9 × 10⁻⁴).

Methods

  • Reporter genes with 32 tandem probe-binding sites (M1 or M2) were stably integrated into CHO cells under tetracycline-responsive promoters with 1× or 7× tet operator sites; doxycycline controlled transcription factor binding. Single-molecule FISH with fluorescently labeled probes counted individual mRNA molecules, including natural mRNA encoding the large subunit of RNA polymerase II.
  • Cells were imaged by wide-field fluorescence microscopy with 0.3-μm Z-steps; custom MATLAB software identified and counted spots in three dimensions. Protein levels were quantified by fluorescence imaging, and mRNA half-life was measured by quantitative RT-PCR.
  • A stochastic gene activation/inactivation model was fitted to mRNA count distributions by maximum likelihood; protein dynamics were simulated using Gillespie’s stochastic simulation algorithm.

Limitations

  • The paper states that it is impossible to determine whether increased average burst size arises from decreased gene inactivation or increased transcription of the activated gene, and that it is unclear why gene activation rate increases at higher doxycycline concentrations; this may be due to factors not included in the model or cell behavior at high doxycycline.
  • The paper states methodological limits: Cy5.5 and Alexa 594 probes had poor photostability, so some dual-color experiments counted mRNA in only one optical slice; the high-mRNA cell line required integrated fluorescence rather than particle counting, which is less accurate at low molecule numbers; and single-copy YFP produced too few proteins to detect after fixation, preventing simultaneous mRNA/protein measurement in that case.
  • Evident from the text: the study is largely in one mammalian cell type (CHO) with engineered reporters plus one natural gene, and the model assumes transcription/translation reaction randomness is negligible relative to gene activation/inactivation; direct evidence for chromatin remodeling as the burst mechanism is not provided.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1371/journal.pbio.0040309

Cites

40 of 40 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Figure 1 Detection and Quantification of Single Molecules of mRNA in Individual Cells (A) Schematic diagram depicting the mRNA detection method. Multiple fluorescent probes bind to each mRNA molecule, yielding a bright, localized signal. (B) Merged image of a three-dimensional stack of images from a CHO cell expressing the 7x-tetO gene depicted in (A), where each mRNA is hybridized to FISH probes that bind to the multimeric probe-binding sequence in its 3′-UTR (probe P1-TMR binding to the M1 multimer). Each spot corresponds to a single mRNA molecule. (C) Identification of the spots in the three-dimensional image stack in (C). Each particle found by the image-analysis algorithm is colored differently, showing that the algorithm is accurate and that individual molecules are uniquely identified. The scale bars are 5 μm long.
Figure 1 Detection and Quantification of Single Molecules of mRNA in Individual Cells (A) Schematic diagram depicting the mRNA detection method. Multiple fluorescent probes bind to each mRNA molecule, yielding a bright, localized signal. (B) Merged image of a three-dimensional stack of images from a CHO cell expressing the 7x-tetO gene depicted in (A), where each mRNA is hybridized to FISH probes that bind to the multimeric probe-binding sequence in its 3′-UTR (probe P1-TMR binding to the M1 multimer). Each spot corresponds to a single mRNA molecule. (C) Identification of the spots in the three-dimensional image stack in (C). Each particle found by the image-analysis algorithm is colored differently, showing that the algorithm is accurate and that individual molecules are uniquely identified. The scale bars are 5 μm long.CC BY · PubMed Central
Figure 2 Cell-to-Cell Variation of mRNA Numbers in Clonal Cell Lines (A) Schematic diagram of the doxycycline-controllable promoters and the reporter genes that they control. Doxycycline binds to the tTA protein, thereby preventing it from binding to the tet operator. (B, C) Representative fields of cells from cell lines E-YFP-M1-1x and E-YFP-M1-7x, containing the 1x-tetO and 7x-tetO promoters, respectively, where each mRNA is hybridized to FISH probe P1-TMR and the image was obtained by merging a three-dimensional stack of images. (D) Two sister cells from cell line E-YFP-M1-7x displaying mRNA hybridized to FISH probe P1-TMR (red) and costained with DAPI (blue). The image represents one focal plane. The scale bars are 5 μm long.
Figure 2 Cell-to-Cell Variation of mRNA Numbers in Clonal Cell Lines (A) Schematic diagram of the doxycycline-controllable promoters and the reporter genes that they control. Doxycycline binds to the tTA protein, thereby preventing it from binding to the tet operator. (B, C) Representative fields of cells from cell lines E-YFP-M1-1x and E-YFP-M1-7x, containing the 1x-tetO and 7x-tetO promoters, respectively, where each mRNA is hybridized to FISH probe P1-TMR and the image was obtained by merging a three-dimensional stack of images. (D) Two sister cells from cell line E-YFP-M1-7x displaying mRNA hybridized to FISH probe P1-TMR (red) and costained with DAPI (blue). The image represents one focal plane. The scale bars are 5 μm long.CC BY · PubMed Central
Figure 3 Statistical Analysis of Per-Cell mRNA Population Distributions (A) Histograms showing the distribution of mRNA molecules per cell for three doxycycline concentrations for both cell lines E-YFP-M1-1x (left) and E-YFP-M1-7x (right). (B) Graphs showing the population mean (top) and noise (defined as the standard deviation divided by the mean [bottom]) as a function of doxycycline concentration. Statistics were taken from the mRNA counts used in (A), and error bars were obtained by bootstrapping. (C) Activation rate (λ/δ) and average number of mRNA molecules produced per burst (μ/γ) for 1x-tetO (blue) and 7x-tetO (red) obtained from fitting the mRNA count data to the model of gene activation and inactivation by the maximum-likelihood method. Error bars reflect 95% confidence intervals.
Figure 3 Statistical Analysis of Per-Cell mRNA Population Distributions (A) Histograms showing the distribution of mRNA molecules per cell for three doxycycline concentrations for both cell lines E-YFP-M1-1x (left) and E-YFP-M1-7x (right). (B) Graphs showing the population mean (top) and noise (defined as the standard deviation divided by the mean [bottom]) as a function of doxycycline concentration. Statistics were taken from the mRNA counts used in (A), and error bars were obtained by bootstrapping. (C) Activation rate (λ/δ) and average number of mRNA molecules produced per burst (μ/γ) for 1x-tetO (blue) and 7x-tetO (red) obtained from fitting the mRNA count data to the model of gene activation and inactivation by the maximum-likelihood method. Error bars reflect 95% confidence intervals.CC BY · PubMed Central
Figure 4 Cell-to-Cell Variations in mRNA Numbers in a Cell Line with Multiple Reporter Gene Integrations at the Same Gene Locus (A) Representative field from cell line L-GFP-M1-7x, generated by lipofection, where the mRNA was hybridized to FISH probe P1-TMR; the image was obtained by merging a three-dimensional stack of images. (B) Histogram showing the distribution of mRNA molecules per cell over for cell line L-GFP-M1-7x when grown in media containing no doxycycline. (C) Graphs showing the population mean (top) and noise (defined as the standard deviation divided by the mean [bottom]) as a function of doxycycline concentration. Error bars were obtained by bootstrapping.
Figure 4 Cell-to-Cell Variations in mRNA Numbers in a Cell Line with Multiple Reporter Gene Integrations at the Same Gene Locus (A) Representative field from cell line L-GFP-M1-7x, generated by lipofection, where the mRNA was hybridized to FISH probe P1-TMR; the image was obtained by merging a three-dimensional stack of images. (B) Histogram showing the distribution of mRNA molecules per cell over for cell line L-GFP-M1-7x when grown in media containing no doxycycline. (C) Graphs showing the population mean (top) and noise (defined as the standard deviation divided by the mean [bottom]) as a function of doxycycline concentration. Error bars were obtained by bootstrapping.CC BY · PubMed Central
Figure 5 Dual-Reporter Experiments Showing Variations Are Intrinsically Random and Can Affect an Entire Gene Locus (A) Schematic depicting dual-reporter integration experiments, where the two reporters are either integrated in separate locations in the genome (left) or at the same locus (right). (B) Two-color overlay showing a merged three-dimensional stack of images of the cell line in which two different reporter genes (one expressing mRNA containing the M1 sequence array and the other expressing mRNA containing the M2 sequence array) are integrated into separate loci (cell line L-GFP-M1-7x+L-CFP-M2-7x); green corresponds to the signal from GFP-M1 mRNA and red corresponds to the signal from CFP-M2 mRNA (using probes P1-TMR and P2-Alexa-594, respectively). The relative mRNA levels of both genes were quantified by counting the mRNA of each color in a single optical slice in each cell (inset). (C) Two-color overlay showing a merged three-dimensional stack of images of the cell line in which the two distinct reporter genes are integrated into the same locus (cell line L-YFP-M1-CFP-M2), where the same FISH probes were used as in (B). The relative mRNA levels of both genes were quantified by counting the mRNA of each color in a single optical slice in each cell (inset). The scale bars are 5 μm long.
Figure 5 Dual-Reporter Experiments Showing Variations Are Intrinsically Random and Can Affect an Entire Gene Locus (A) Schematic depicting dual-reporter integration experiments, where the two reporters are either integrated in separate locations in the genome (left) or at the same locus (right). (B) Two-color overlay showing a merged three-dimensional stack of images of the cell line in which two different reporter genes (one expressing mRNA containing the M1 sequence array and the other expressing mRNA containing the M2 sequence array) are integrated into separate loci (cell line L-GFP-M1-7x+L-CFP-M2-7x); green corresponds to the signal from GFP-M1 mRNA and red corresponds to the signal from CFP-M2 mRNA (using probes P1-TMR and P2-Alexa-594, respectively). The relative mRNA levels of both genes were quantified by counting the mRNA of each color in a single optical slice in each cell (inset). (C) Two-color overlay showing a merged three-dimensional stack of images of the cell line in which the two distinct reporter genes are integrated into the same locus (cell line L-YFP-M1-CFP-M2), where the same FISH probes were used as in (B). The relative mRNA levels of both genes were quantified by counting the mRNA of each color in a single optical slice in each cell (inset). The scale bars are 5 μm long.CC BY · PubMed Central
Figure 6 Cell-to-Cell Variations in the mRNA Encoding the Large Subunit of RNA Polymerase II (A) Representative field of cell line E-YFP-M1-7x upon performing FISH with differently colored probes for both YFP-M1 mRNA and the mRNA encoding the large subunit of RNA polymerase II. The image is a two-color overlay, where green corresponds to the signal from YFP-M1 mRNA (one optical slice) and red corresponds to the signal from RNA polymerase mRNA (merged three-dimensional image stack). The probes used were P1-Cy5.5 and P3-TMR. (B) Histogram showing the distribution of RNA polymerase mRNA molecules per cell (top) and scatterplot (bottom) showing reporter mRNA levels (quantified by counting the mRNA in a single optical slice) and RNA polymerase mRNA levels (quantified by counting all mRNA) in each cell. Gene activation (λ/δ), inactivation (γ/δ), and transcription rates (μ/δ) (normalized by the mRNA decay rate δ) are given with 95% confidence intervals as indicated. The scale bar is 5 μm long.
Figure 6 Cell-to-Cell Variations in the mRNA Encoding the Large Subunit of RNA Polymerase II (A) Representative field of cell line E-YFP-M1-7x upon performing FISH with differently colored probes for both YFP-M1 mRNA and the mRNA encoding the large subunit of RNA polymerase II. The image is a two-color overlay, where green corresponds to the signal from YFP-M1 mRNA (one optical slice) and red corresponds to the signal from RNA polymerase mRNA (merged three-dimensional image stack). The probes used were P1-Cy5.5 and P3-TMR. (B) Histogram showing the distribution of RNA polymerase mRNA molecules per cell (top) and scatterplot (bottom) showing reporter mRNA levels (quantified by counting the mRNA in a single optical slice) and RNA polymerase mRNA levels (quantified by counting all mRNA) in each cell. Gene activation (λ/δ), inactivation (γ/δ), and transcription rates (μ/δ) (normalized by the mRNA decay rate δ) are given with 95% confidence intervals as indicated. The scale bar is 5 μm long.CC BY · PubMed Central
Figure 7 Propagation of mRNA Variations to Variations in Protein Levels (A) Scatterplot of total GFP and mRNA numbers in individual cells from cell line L-GFP-M1-7x grown under conditions of no doxycycline (blue), 0.08 ng/ml doxycycline (red), and 0.16 ng/ml doxycycline (green). Marginal histograms indicate the distribution of reporter mRNA per cell (top) or total GFP (right) for all the growth conditions. (B) Scatterplot of total CFP and mRNA levels (obtained from a single optical slice) in individual cells from cell line L-GFP-M1-7x+L-CFP-M2-7x grown under conditions of no doxycycline. Marginal histograms indicate the distribution of reporter mRNA per cell (top) or total GFP (right) for all the growth conditions. (C) Histogram of total YFP per cell from cell line E-YFP-M1-7x. YFP was quantified in live cells to minimize loss of fluorescence due to fixation and permeabilization. (D) Scatterplots and associated marginal histograms showing the results of stochastic simulations of the model of mRNA and protein dynamics presented in Protocol S1. The mRNA dynamics were the same for all cases, but the protein degradation rate was increased as indicated. Details of the simulation procedures and exact values of the parameters used are given in the Materials and Methods section.
Figure 7 Propagation of mRNA Variations to Variations in Protein Levels (A) Scatterplot of total GFP and mRNA numbers in individual cells from cell line L-GFP-M1-7x grown under conditions of no doxycycline (blue), 0.08 ng/ml doxycycline (red), and 0.16 ng/ml doxycycline (green). Marginal histograms indicate the distribution of reporter mRNA per cell (top) or total GFP (right) for all the growth conditions. (B) Scatterplot of total CFP and mRNA levels (obtained from a single optical slice) in individual cells from cell line L-GFP-M1-7x+L-CFP-M2-7x grown under conditions of no doxycycline. Marginal histograms indicate the distribution of reporter mRNA per cell (top) or total GFP (right) for all the growth conditions. (C) Histogram of total YFP per cell from cell line E-YFP-M1-7x. YFP was quantified in live cells to minimize loss of fluorescence due to fixation and permeabilization. (D) Scatterplots and associated marginal histograms showing the results of stochastic simulations of the model of mRNA and protein dynamics presented in Protocol S1. The mRNA dynamics were the same for all cases, but the protein degradation rate was increased as indicated. Details of the simulation procedures and exact values of the parameters used are given in the Materials and Methods section.CC BY · PubMed Central

Verification

Claims the paper makes

0 Valid · 0 Not valid · 22 Not checked · 82 Own results and interpretations

56 more claims

Create a free account to see all 104 claims and to verify the 59 references still open.

See all claims — free account
  • It may state two separate claims as one. 23
  • Probably not a substantial claim. 21

Create a free account to see these claims.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

2 of 2

Citations

Cites

Show all 40 references

Research works printed but not linked: 41

  1. Becskei, 2005, Nat Genet, 37, 937
  2. Blake, 2003, Nature, 422, 633
  3. Elowitz, 2002, Science, 297, 1183
  4. Golding, 2005, Cell, 123, 1025
  5. Raser, 2004, Science, 304, 1811
  6. Rosenfeld, 2005, Science, 307, 1962
  7. Ozbudak, 2002, Nat Genet, 31, 69
  8. Acar, 2005, Nature, 435, 228
  9. Lahav, 2004, Nat Genet, 36, 147
  10. Maamar, 2005, Mol Microbiol, 56, 615
  11. Arkin, 1998, Genetics, 149, 1633
  12. Weinberger, 2005, Cell, 122, 169
  13. Elowitz, 2000, Nature, 403, 335
  14. Becskei, 2001, Embo J, 20, 2528
  15. Swain, 2002, Proc Natl Acad Sci U S A, 99, 12795
  16. Volfson, 2006, Nature, 439, 861
  17. Newman, 2006, Nature, 411, 840
  18. Bar-Even, 2006, Nat Genet, 38, 636
  19. Peccoud, 1995, Theor Population Biol, 48, 222
  20. Bengtsson, 2005, Genome Res, 15, 1388
  21. Norris, 2003, Mol Endocrinol, 17, 193
  22. Walters, 1995, Proc Natl Acad Sci U S A, 92, 7125
  23. Chubb, 2006, Curr Biol, 16, 1018
  24. Ko, 1990, EMBO J, 9, 2835
  25. Vargas, 2005, Proc Natl Acad Sci U S A, 102, 17008
  26. Janicki, 2004, Cell, 116, 683
  27. Gossen, 1992, Proc Natl Acad Sci U S A, 89, 5547
  28. Thattai, 2001, Proc Natl Acad Sci U S A, 98, 8614
  29. Peccoud, 1995, Theor Population Biol, 48, 222
  30. Gillespie, 1977, J Phys Chem, 81, 2340
  31. Paulsson, 2004, Nature, 427, 415
  32. Pedraza, 2005, Science, 307, 1965
  33. Kepler, 2001, Biophys J, 81, 3116
  34. Tumbar, 1999, J Cell Biol, 145, 1341
  35. Memedula, 2003, Curr Biol, 13, 241
  36. Li, 2005, Nat Struct Mol Biol, 12, 46
  37. Hume, 2000, Blood, 96, 2323
  38. Balaban, 2004, Science, 305, 1622
  39. Wernet, 2006, Nature, 440, 174
  40. Robinett, 1996, J Cell Biol, 135, 1685
  41. Femino, 1998, Science, 280, 585

Cited by

Items 901–1,000 of 1,013

Metadata

Authors
Raj, A., Peskin, C. S., Tranchina, D., Vargas, D. Y., & Tyagi, S.
Affiliations
Courant Institute of Mathematical Sciences (US)
Corresponding author
  • Raj, Arjun
  • Tyagi, Sanjay
Venue
PLoS Biology
Published
2006-09-12
Volume
4
Issue
10
Pages
e309
ISSN
  • 1544-9173 (ISSN-L)
  • 1545-7885
OpenAlex
W2141688036
Cited by
1,737
References
40
Citations by year
2026: 54 · 2025: 77 · 2024: 93 · 2023: 82 · 2022: 87 · 2021: 90 · 2020: 109 · 2019: 114 · 2018: 107 · 2017: 101 · 2016: 114 · 2015: 108 · 2014: 86 · 2013: 116 · 2012: 84
FWCI
20.55
Type
journal article
Language
en
Open access
Locations
  • PLoS Biology (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • Rutgers University Community Repository (Rutgers University) (submittedVersion, cc-by)
  • RePEc: Research Papers in Economics (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • Figshare (submittedVersion, cc-by)
Funders
National Institute of General Medical Sciences (GM070357, R01GM070357)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Gene Regulatory Network Analysis
  • Physical Sciences › Physics and Astronomy › Statistical and Nonlinear Physics › stochastic dynamics and bifurcation
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › RNA Research and Splicing
Keywords
stochastic gene expression, mRNA synthesis, transcriptional bursting, intrinsic noise, protein degradation, reporter genes, mammalian cells
MeSH terms
Animals, Gene Expression Regulation, Genetic Vectors, Cricetinae, Models, Genetic, Models, Theoretical, RNA Polymerase II, RNA, Messenger, Stochastic Processes, Transcription, Genetic, Transcriptional Activation, CHO Cells, In Situ Hybridization, In Situ Hybridization, Fluorescence, Genes, Reporter

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

A language model drew the claims from the paper on 2026-10-04, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-04 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.