Fluorescence fluctuation spectroscopy enables quantitative imaging of single mRNAs in living cells

Abstract

Imaging mRNA with single-molecule sensitivity in live cells has become an indispensable tool for quantitatively studying RNA biology. The MS2 system has been extensively used due to its unique simplicity and sensitivity. However, the levels of the coat protein needed for consistent labeling of mRNAs limits the sensitivity and quantitation of this technology. Here, we applied fluorescence fluctuation spectroscopy to quantitatively characterize and enhance the MS2 system. Surprisingly, we found that a high fluorescence background resulted from inefficient dimerization of fluorescent protein (FP)-labeled MS2 coat protein (MCP). To mitigate this problem, we used a single-chain tandem dimer of MCP (tdMCP) that significantly increased the uniformity and sensitivity of mRNA labeling. Furthermore, we characterized the PP7 coat protein and the binding to its respective RNA stem loop. We conclude that the PP7 system performs better for RNA labeling. Finally, we used these improvements to study endogenous β-actin mRNA, which has 24xMS2 binding sites inserted into the 3' untranslated region. The tdMCP-FP allowed uniform RNA labeling and provided quantitative measurements of endogenous mRNA concentration and diffusion. This work provides a foundation for quantitative spectroscopy and imaging of single mRNAs directly in live cells.

Figure 1

Normalized brightness of coat proteins. (A) Schematic of the coat protein constructs. (B) The brightness of CP-EGFP measured in U2OS cells is plotted as a function of CP concentration. From the data, it is clear that PCP-EGFP (triangles) dimerizes at a much lower concentration than MCP-EGFP (diamonds) does . The data were fit to Eq. 1 to obtain the dissociation constant of the coat protein (410 nM for MCP-EGFP and <20 nM for PCP-EGFP). (C) The normalized brightness of tdCP-EGFP stays at unity at different concentrations, indicating that the tandem dimers are behaving as monomers.

Figure 2

Normalized brightness of mRNA. (A) mRNA constructs used in the experiment. The mRNAs have a CFP open reading frame. After the stop codon, 24xPBS or 24xMBS is inserted into the 3′ UTR. (B) CFP-24xPBS is cotransfected with tdPCP-EGFP and mCherry in U2OS cells and measured for 3 min at a wavelength of 1010 nm. The two-species fit of the data reports the brightness of the mRNA. The normalized mRNA brightness, which measures the number of EGFPs on the mRNA, is plotted as a function of the total concentration of EGFP, determined by dividing the total fluorescence intensity by the EGFP brightness. The data indicate that the average number of EGFP on mRNA is 23 ± 5, implying that 24 PBS are fully occupied. (C) The same experiments were performed as in B except that tdPCP-EGFP was substituted by PCP-EGFP. The normalized brightness of mRNA saturates at 48 at high PCP concentration, but at low concentration the PP7 stem loops are not fully occupied. (D and E) The same experiments were performed on CFP-24xMBS cotransfected with tdMCP-EGFP (D) or MCP-EGFP (E). The normalized brightness of mRNA does not change with concentration for tdMCP-EGFP (the average is 13 ± 2), but it is approximately half of the expected full occupation number, 24. For MCP-EGFP, the mRNA brightness increases with the concentration of MCP and saturates at 26 ± 3.

Figure 3

MBS-MEF cells stably expressing (A) MCP-EGFP or (B) tdMCP-EGFP are imaged on an epifluorescence microscope with an excitation wavelength of 488 nm. To assist the comparison, both images are scaled with the same black/white levels. (A) The signal of MCP-EGFP-labeled mRNA depends on the concentration of MCP. The upper-left cell has higher fluorescence intensity in the nucleus and more detectable mRNA than the lower-right cell. These two cells are in the same imaging field. (B) mRNA is uniformly labeled with tdMCP-EGFP. The cell has similar fluorescence intensity in the nucleus as the dimmer cell in panel A, but mRNA molecules are brightly and uniformly labeled. The scale bar is 5 μm. (C) The detected mRNA number in the cytoplasm (normalized by the size of the cytoplasm) is plotted as a function of fluorescence intensity in the nucleus. Each symbol is a measurement of a single cell. To facilitate comparison, the same criterion for spot detection was used for all images. With tdMCP labeling, the detected mRNA number does not depend on the expression level of tdMCP (triangles). However, for MCP-labeled mRNA (diamonds), the detected mRNAs increase with the concentration of MCP and only reach the tdMCP-detected mRNA level at high concentration.

Figure 4

FFS measurements of endogenous β-actin mRNA. (A) Diffusion constant of β-actin mRNA in the cytoplasm and nucleus. MBS-MEF was infected with lentivirus to stably express tdMCP-EGFP. First, the MBS-MEF cell was measured in cytoplasm for 3 min. The autocorrelation function was fit with a two-species diffusion model (Eq. 3; see Fig. S2 for fit) and the mRNA diffusion constant was measured. Second, to measure the diffusion property of mRNA in the nucleus, MBS-MEF cells were stimulated with 20% serum after serum starvation overnight. FFS measurements were conducted in the nucleus immediately after serum stimulation. The photon counting traces were split into 5-min segments. The autocorrelation curves were calculated from the segments and fit to Eq. 3 to obtain the diffusion constant of mRNA. The scatter plot of the diffusion constants of β-actin mRNA is shown. In the cytoplasm, the diffusion constant of β-actin mRNA ranges from 0.15 to 0.74 μm²/s, with an average of 0.35 μm²/s. The diffusion constant of mRNA in the nucleus is larger than in cytoplasm, with an average of 0.72 μm²/s. (B) Concentration of β-actin in cytoplasm. The MBS-MEF cell was measured for 3 min in cytoplasm. The data are fit by a two-species TIFCA model, which provides the concentration of β-actin mRNA. The histogram of mRNA concentration is plotted. In the inset, the scatterplot of the concentration is also shown. The concentration ranges from 1 to 30 nM, with an average of 11 nM. (C) The MBS-MEF was serum-stimulated as described for panel A. The data were subjected to a two-species TIFCA fit, and the concentration of β-actin mRNA is plotted as a function of time. Each dotted curve represents a measurement of a single cell. The average response of these cells is plotted as solid lines.