A novel fluorescent reporter sensitive to serine mis-incorporation

Figures & data

Figure 1. Design of a mistranslation sensitive fluorescent protein. To detect mistranslation events, a mutant tRNA^Ser with a proline anticodon is co-expressed with a fluorescently inactive mCherry Ser151Pro mutant (a). If no mistranslation event occurs at position 151, then the mCherry is accurately made and does not fluoresce. A mistranslation event at position 151 that incorporates serine instead of proline will restore fluorescent activity. Ser151 interacts with the mCherry chromophore (Met71-Tyr72-Gly73, PDB code 2H5Q [31]) via a hydrogen bond (b). We anticipated that mutation at Ser151 would interfere with the chromophore and disrupt mCherry fluorescence.

Figure 2. Growth on solid media of wild-type or mistranslating E. coli expressing mCherry variants. E. coli BL21 cells were transformed with plasmids expressing either wild-type mCherry with a Ser151 AGC codon or mutant mCherry with a Pro151 CCA codon. Each strain was co-transformed with a second plasmid bearing either wild-type tRNA^Ser, derived from the GCT 1–1 tRNA^Ser gene (a) or the GGA 1–1 tRNA^Ser gene (b), or a mutant version of each tRNA^Ser gene that has a UGG proline-decoding anticodon. mCherry with a Pro151 codon co-expressed with either of the tRNA^Ser_UGG variants demonstrated fluorescence rescue resulting from mistranslation. For the negative controls, a wild-type tRNA^Ser was co-expressed with a mutant mCherry Pro151 that did not fluoresce above the background. Cultures from three biological replicates were aliquoted in serial dilution as indicated in Methods.

Figure 3. Growth of E. coli strains expressing wild-type or mistranslating tRNAs. Using a 96-well microplate, cell growth (A₆₀₀) was recorded over a 16-hour time course for (a) E. coli cells co-expressing mCherry with tRNA^Ser_GCU (blue circles), mCherry with mutant tRNA^Ser_UGG (red triangles), mCherry Ser151Pro with tRNA^Ser_UGG (Orange triangles), or mCherry Ser151Pro with wild-type tRNA^Ser_GCU (grey circles). Cell growth (c) is also shown for E. coli cells expressing mCherry with tRNA^Ser_GGA (blue circles), mCherry with a corresponding mutant tRNA^Ser_UGG (red triangles), Ser151Pro with mutant tRNA^Ser_UGG (Orange triangles), Ser151Pro with tRNA^Ser_GGA (grey circles). At each time point, fluorescence (ex 570 nm, em 610 nm) intensity was also recorded (b, d). The inset panels (b, d) show a zoomed in view of the fluorescence recovery due to serine mis-incorporation at Pro151.

Table 1. Ser mis-incorporation levels in cells and purified mCherry proteins

mCherry variant	tRNA^Ser anticodon	Maximal fluorescence per cell	Relative rate of fluorescence per cell	Relative specific fluorescence of purified mCherry	MS/MS Peptide Count # S, P151
WT	GCT	91 ± 11%	78 ± 23%	80.0 ± 0.5%	-
WT	UGG	100 ± 30%	100 ± 32%	100 ± 1%	-
Ser151Pro	GCT	0.00 ± 0.03%	0.01 ± 0.01%	0.010 ± 0.003%	-
Ser151Pro	UGG	12 ± 2%	21.0 ± 1.5%	4.1 ± 0.2%	S 5 P 17
		Maximal level of error per cell/codon:	Relative rate of error per cell/codon:	Level of error/codon in soluble fraction:
Error levels at CCA151		12 ± 2%	21.0 ± 1.5%	4.1 ± 0.2%	29%
WT	GGA	72 ± 20%	67 ± 13%	98 ± 3.5%	-
WT	UGG	100 ± 9%	100 ± 9%	100 ± 3.5%	-
Ser151Pro	GGA	0.00 ± 0.01%	0.00 ± 0.05%	0.005 ± 0.003%	-
Ser151Pro	UGG	15.0 ± 0.5%	23 ± 5%	9.7 ± 0.1%	S 9 P 19
		Maximal level of error per cell/codon:	Relative rate of error per cell/codon:	Level of error/codon in soluble fraction:
Error levels at CCA151		15.0 ± 0.5%	23 ± 5%	9.7 ± 0.1%	47%

We quantitated translation error in 3 ways: 1) We calculated the maximal error level per codon per cell (column 3) by dividing the maximal fluorescence intensity at the end of the time course (Figure 3) by the number of cells recorded at that time using the standard conversion (A₆₀₀ unit = 8 × 10⁸ cells/mL); 2) We quantified the rate over time at which the mis-made protein accumulated per CCA151 codon in cells (column 4) by determining the slope of the linear phase of mCherry fluorescence production (between the 7 and 12 hour time points, Figure 3) to determine the rate of fluorescence increase over time for each strain; and 3) we determined the level of error in the purified mCherry reporters from large preparative cultures, which represents the total error level per CCA151 codon in the soluble protein fraction. The values (column 5) are based on the specific fluorescence measurement of pure mCherry variants (Figure 5). The data were normalized relative to the wild-type mCherry produced in mistranslating cells, which was set to 100%. We found no significant difference between the levels of mCherry produced in normal versus mistranslating cells (Figures 4(c,d) and 5(c, d)). Finally, we included a summary of spectral counts of peptides identified by MS/MS (Figure 6 and Table S3) as containing Ser151 or Pro151 in response to the CCA151 codon in mistranslating cells.

Figure 4. Quantification of mistranslation level and rate in live E. coli cells. Data from cell growth and mCherry fluorescence in microplates (Figure 3) were analysed to compare the maximal fluorescence per cell observed in each of the indicated mCherry producing strain (a, b). The rate of fluorescence increase per cell (c, d) was calculated by measuring the slope of the linear phase of fluorescent mCherry production per cell (Figure 3(b,d)). Data are shown for the tRNA^Ser_GCU derived variants (a, c) and for the tRNA^Ser_GGA derived variants (b, d). Error bars represent ± 1 standard deviation of three biological replicates and three technical replicates. Statistical significance is based on ANOVA single factor analysis (n. s. – not significant, ** p < 0.01).

Figure 5. Fluorescent measurements of purified mCherry variants. mCherry proteins variants were purified and fluorescence (ex 570 nm, em 610 nm) of each serial dilution (a, b) was measured. Based on linear regression of the data (A,B dotted lines), the slope is the specific fluorescence or fluorescence intensity per amount of mCherry protein (c, d). The wild-type or Ser151Pro mCherry variants were produced with co-expression of either the (a) tRNA^Ser_GCU or a UGG anticodon mutant or with the (b) tRNA^Ser_GGA or a UGG anticodon mutant. According to specific fluorescence, the fluorescence recovery or serine mis-incorporation level is (c) 4.1% for the tRNA^Ser_GCU derived proline decoding mutant and (d) 9.7% for tRNA^Ser_GGA derived mutant. Statistical significance is based on ANOVA single factor analysis (n. s. – not significant, ** p < 0.01).

Figure 6. Tandem mass spectrometry analysis reveals Ser mis-incorporation in mistranslating cells at Pro151. MS/MS coverage of purified and trypsin digested mCherry variants is represented as a schematic with each blue bar representing an identified peptide. MS/MS spectra for representative peptide hits are shown (right). Cells co-expressing wild-type mCherry with a mutant tRNA^Ser_UGG (a) showed only Ser151 as expected. Cells co-expressing the mutant mCherry with a wild-type tRNA^Ser (b) showed only Pro151. Mistranslation was readily detected in cells expressing both mutant mCherry and mutant tRNA (c) where multiple high-quality spectra were identified for peptides containing either Pro151 or Ser151.

Table 2. Two-sample KS test results from comparing RMSD distributions of mCherry variants

Variant 1	Variant 2	D-statistic	P-value	Interpretation*
WT	A150G	0.0895	8.96 × 10^−4	Equivalent
WT	S151P	0.3573	8.18 × 10^−55	Different
WT	S151F	0.4729	1.91 × 10^−97	Different
A150G	S151P	0.4365	1.54 × 10^−82	Different
A151G	S151F	0.5896	3.27 × 10^−133	Different
S151P	S151F	0.1552	1.67 × 10^−10	Different

*Interpretation of D-statistic and P-value to state whether the RMSD distributions of the two mCherry variants are equivalent or significantly different at an α-level of 10⁻⁶ (D statistic critical value = 0.0933).

Figure 7. Conformational sampling explains the mechanism of fluorescence inhibition and rescue. A set of 1,000 plausible conformations of the entire mCherry precursor protein were generated in Rosetta for WT, Ala150Gly, Ser151Pro and Ser151Phe mutants. (a) The distribution in chromophore precursor position was compared to that of the mature chromophore in the 2H5Q mCherry crystal structure [31]. (b) The position distribution for each mutant is represented as RMSD values (x-axis) of the chromophore precursor atoms compared to the initial starting model. The Rosetta Score (y-axis), a metric used to estimate relative conformational stability, is also shown for each generated model.

Shu X, Shaner NC, Yarbrough CA, et al. Novel chromophores and buried charges control color in mFruits. Biochemistry. 2006;45:9639–9647.

 PubMed Web of Science ®Google Scholar