Introduction to the Traffic Light Reporter Assay

The conventional Traffic Light Reporter (TLR) assay was a dual fluorescent readout to screen for genomic editing activity. It was initially used to generate a flow cytometric readout of homology-directed repair (HDR) mediated gene targeting and mutagenic non-homologous end joining (mutNHEJ) mediated gene disruption [1].

First Iteration and Learning from Leaky Expressions

In the case of prime editing, the assay is modified to utilize fluorescent proteins with premature stop codons that can be remedied through simple prime editing. The reverse transcriptase (RT) linked to the nCas9 protein reverse transcribes the correct amino acid using the reverse transcriptase template found on the prime editing guide RNA (pegRNA).

Design

The premature stop codon (5'-TGA-3') is thus replaced with glycine (5'-GGA-3'), restoring the wild-type mCherry fluorescent protein and allowing the expression of the Clover fluorescent protein downstream. Theoretically, both fluorescent proteins (mCherry and Clover) would thereby be expressed in successfully edited cells, and they can be analysed using flow cytometry.

Build

Following this idea, an initial plasmid for the TLR assay was built. In the initial design, the plasmid Addgene #65777 harbouring a mCherry-stop-Clover design was adopted. Two sets of pegRNAs and ngRNAs targeting the premature stop codon on the mCherry were also designed - one set targeting the sense strand, and the other set targeting the anti-sense strand. (Figure 1).

Figure 1. Prime editing of premature stop codon on mCherry fluorescent protein. A premature stop codon (5'-TGA-3') is found on the C-terminus of the mCherry fluorescent protein, resulting in the null expression of the Clover fluorescent protein downstream. Two sets of pegRNAs and ngRNAs targeting the premature stop codon were thus designed. The expected mechanism was hence to edit the premature stop codon into glycine (5'-GGA'-3'), hence continuing the transcription and translation of Clover downstream. Successfully transfected and edited cells will thus fluoresce both red and green.

Test

HEK293T cells were then transfected with the plasmids required for prime editing on a 96-well plate. A negative control with only transfection of the ngRNA plasmid (#65777 with mCherry-stop-clover) was also set-up (Table 1). Flow cytometry was subsequently utilized to analyse the fluorophore excitation in each cell.

	Well 1	Well 2	Well 3	Well 4	Well 5	Well 6
nCas9-RT construct	-	-	PE2	PEV19	PE2	PFV19
TLR-pegRNA 1	-	-	+	+	-	-
TLR-ngRNA 1	+	-	+	+	-	-
TLR-pegRNA 2	+	-	-	-	+	+
TLR-ngRNA 2	-	+	-	-	+	+

Table 1: Combinations of plasmids transfected into HEK293T cells for TLR assay. In both wells 1 and 2, only the TLR ngRNAs were transfected. They served as negative control for any prime editing activity. Wells 3 and 4 were transfected with the first set of pegRNA and ngRNA, targeting the sense strand to remedy the premature stop codon on the mCherry protein. Wells 5 and 6 were transfected with the second set of pegRNA and ngRNA, targeting the anti-sense strand to remedy the premature stop codon on the mCherry protein.

Learn

From our initial testing, the TLR assay suggests its sufficiency in discriminating the editing efficiency across 2 selected reverse transcriptase (RT) constructs (Figure 2). This can be seen from the difference in the mean fluorescence intensity (MFI) GFP levels between wells 5 (PE2) and 6 (PFV19). However, there was also a significant flaw in the design which led to leaky expression of the Clover fluorescent protein (Figure 3). The start codon found on the Clover fluorescent protein might have allowed for the restart of translation to occur at the Clover despite the initial stop codon on the mCherry. This led to significant expression of the Clover fluorescent protein downstream despite the premature stop codon on the C-terminus of the mCherry fluorescent protein. As such, cells in the negative control were fluorescing both green and red when analysed under flow cytometry, leading to high background noise in our positive samples. As such, our initial TLR assay, whilst might be sufficient to screen for possible editing activity, was insufficient to accurately represent the editing efficiency. This flaw in the design thus motivated an improvement in the TLR assay.

Figure 2: Traffic Light Reporter (TLR) assay as potential validation of prime editing efficiency. A premature stop codon (5'-TGA-3') is found on the C-terminus of the mCherry fluorescent protein, resulting in the null expression of the Clover fluorescent protein downstream. Two sets of pegRNAs and ngRNAs targeting the premature stop codon were thus designed. The expected mechanism was hence to edit the premature stop codon into glycine (5'-GGA'-3'), hence continuing the transcription and translation of Clover downstream. Successfully transfected and edited cells will thus fluoresce both red and green.

TLR-pegRNA 1 and TLR-ngRNA 1 only	TLR-pegRNA 2 and TLR-ngRNA 2 only

Figure 3: High frequencies of leaky expression of Clover detected in negative control in TLR assay. In both negative controls, only the TLR-pegRNAs and TLR-ngRNAs were transfected in the absence of the nCas9-RT construct. Despite the lack of the nCas9-RTs, there is significant expression of the Clover protein, which suggests the lack of stringency in our initial design. a: Transfection of HEK293T cells with TLR-pegRNA 1 and TLR-ngRNA 1 revealed the detection of Clover fluorescent protein despite the premature stop codon installed in the mCherry fluorescent protein. b: Transfection of HEK293T cells with TLR-pegRNA 2 and TLR-ngRNA 2 also revealed the detection of Clover fluorescent protein despite the premature stop codon installed in the mCherry fluorescent protein.

Second Iteration and Learning from Optimization

In our second iteration, the principle behind the TLR assay remains the same – prime editing fluorophores with a premature stop codon to exhibit fluorescence in successfully edited cells. However, there is motivation to improve the design such that it is more stringent and can better discriminate the editing efficacies of nCas9-RT constructs in simple and complex edits.

Design

Two different approaches were thus design – a simple edit to restore the wild-type fluorescent protein, and a complex edit to restore the wild-type fluorescent protein. The rationale behind the two types of edits was to also resolve the editing efficacies of prime editing better in simple and complex edits.

Build

In our first approach (approach 1), an eGFP protein with a premature stop codon was cloned into a separate pCVL TLR plasmid (Addgene #31483) (Figure 4) [1]. This was complemented with the designing of PETLR-pegRNA 1 and PETLR-ngRNA 1 to target the premature stop codon. The premature stop codon (5’-TGA’-3’) is thus edited into threonine (5’-ACA-3’), restoring the wild-type eGFP and expression of eGFP.

Figure 4: Cloning of eGFP protein with a premature stop codon on position 62 with simple edit into the pCVL plasmid (Addgene #31483)
Left: The eGFP with a premature stop codon requiring a simple edit is cloned into the plasmid vector pCVL Traffic Light Reporter 2.1 (VF2468 ZFN target) Ef1a (Addgene #31483).
Right: The premature stop codon on position 62 found in the eGFP prevents the translation of the full length eGFP protein. PETLR-pegRNA 1 and PETLR-ngRNA 1 targeting the premature stop codon were designed. HEK293T cells with successful edits would thus swap the stop codon into threonine, restoring the wild-type eGFP. HEK293T cells with successful edits would thus express the full length eGFP, and fluorescence green under flow cytometry analysis.

In our second approach (approach 2), an eGFP requiring a complex edit to restore the wild-type protein (hereafter denoted as “TLR gene”) was integrated into the genome of HEK293T cells (Figure 5). This integration of the TLR gene into the HEK293T also allowed for the mimicking of prime editing on endogenous target sites. Two separate cell lines were also created – a heterogenous cell line (TLR WT) and a homogeneous cell line derived from the former (TLR 1D10).

Figure 5: Cloning of eGFP protein with a premature stop codon with complex edit required. The eGFP with a premature stop codon requiring a complex edit is integrated into the HEK293T cells. The premature stop codon found in the eGFP prevents the translation of the full length eGFP protein. TLR-76 pegRNA and 777BFP-ngRNA targeting the premature stop codon were designed. HEK293T cells with successful edits would thus express the full length eGFP, and fluorescence green under flow cytometry analysis.

Test

In approach 1, HEK293T cells were transfected with the custom pCVL plasmid (Figure 4a), nCas9-RT constructs, as well as PETLR-pegRNA 1 and PETLR-ngRNA 1. A negative control with only pCVL plasmid transfected was also set up (Table 2).

	Well 1	Well 2	Well 3	Well 4	Well 5	Well 6
nCas9-RT construct	-	-	PE2	PE1	PE497	PFV19
PETLR-pegRNA 1	-	-	+	+	+	+
PETLR-ngRNA 1	+ (83ng)	+ (1ng)	+	+	+	+

Table 2: Combinations of plasmids transfected into HEK293T cells for improved TLR assay. In both wells 1 and 2, only the PETLR ngRNA 1 was transfected. They served as negative control for any prime editing activity. Wells 3-6 were transfected with different nCas9-RT constructs.

In approach 2, the TLR WT and TLR 1D10 cells were transfected with various nCas9-RT constructs, TLR-76 pegRNA and 777BFP ngRNA. A negative control with zero plasmids transfected was also set up to allow for comparisons.

	TLR WT				TLR 1D10
	Well 1	Well 2	Well 3	Well 4	Well 5	Well 6	Well 7	Well 8
nCas9-RT construct	PE2	PE1	PE497	-	PE2	PE1	PE497	-
TLR-76 pegRNA	+	+	+	-	+	+	+	-
777BFP ngRNA	+	+	+	-	+	+	+	-

Table 3: Combinations of plasmids transfected into TLR WT and TLR 1D10 cells for improved TLR assay. In both cell lines (TLR WT and TLR 1D10), the same combination of plasmids was transfected. In the TLR WT cell line, three nCas9-RT constructs were identified (PE2, PE1 and PE497) and transfected together with TLR-76 pegRNA and 777BFP ngRNA (well 1-3). A negative control with zero transfection of plasmids was also set up (well 4). The same replicate was transfected for the TLR 1D10 cell line.

Learn

For both approaches, the initial flaw of leaky expression was eliminated successfully (Figure 6). In the negative controls, minimal green fluorescence was detected, validating the improvement in our second iteration. The reduction in background noise results also meant a more precise comparison of editing efficiency between different nCas9-RT constructs.

Transfection of HEK293T cells with 83ng of PETLR-ngRNA 1 (mCherry) [Table 2 - Well 1]	Transfection of HEK293T cells with 1ug of PETLR-ngRNA 1 (mCherry) [Table 2 – Well 2]
a	b
c	d

Figure 6: Elimination of leaky expression results through screening of negative controls in TLR assay. In both separate approaches, different negative controls were set up to validate the improvement in the second design. Expression of the eGFP reduced significantly, indicating a reduction in background noise and a more accurate testing. a: Transfection of HEK293T cells with only 83ng of PETLR-ngRNA 1 plasmid (mCherry) showed no detectable little to no eGFP fluorescence. b: Transfection of HEK293T cells with 1ug of PETLR-ngRNA 1 plasmid (mCherry) showed reproducible results where there were no detectable eGFP fluorescence. c: Un-transfected HEK293T TLR WT cells showed no detectable eGFP fluorescence. d: Un-transfected HEK293T TLR 1D10 cells also showed no detectable eGFP fluorescence.

Moving onto results, multiple replicates were performed via approach 1 with selected nCas9-RT constructs (Figure 7). Based on the results, the editing efficiencies of the nCas9-RT constructs can be compared to the two wild-types constructs PE1 and PE2.

Figure 7: TLR Assay using PETLR 1 parameters (PETLR-pegRNA 1 and PETLR-ngRNA 1) for selected constructs. P2-FITC A measures the number of HEK293T cells that exhibit green fluorescent (eGFP fluorescence) upon excitation. HEK293T cells were transfected with different combinations of plasmids as shown in Table 2. The cells were subsequently analysed using flow cytometry. Three replicates were also performed, validating the accuracy of the TLR assay design.

An initial testing was also performed on the TLR WT and TLR 1D10 cell lines to validate the design of the TLR assay (Figure 8). From our initial testing results, the TLR assay using the integrated TLR gene produces similar results to the exogenous pCVL plasmid approach.

a

b

Figure 8: Initial results from TLR assay using HEK293T TLR WT and 1D10 cell lines for selected constructs. HEK293T cells were transfected with different combinations of plasmids as shown in Table 3. The cells were subsequently analysed using flow cytometry. P3 FITC-H measures the amount of green fluorescence detected (eGFP fluorescence) in cells that are also excited by the DAPI channel. P3% Parent measures the percentage of cells that fluorescence both green (eGFP) and blue (BFP). P3% Parent is a more precise measurement due to the varying number of TLR gene integrated into the genome.

Combining both approaches, the improved TLR assays eliminated the high background noise present in the results reflected in our initial designed assay. As such, our second iteration provides a more precise and stringent reflection of the nCas9-RT editing efficiency. The TLR assay via PETLR 1 (approach 1) proves to be a more representative assay as seen from the more precise results produced through multiple replicates. Whilst the TLR assay via the integrated endogenous sites (approach 2) can provide a close estimation of the editing efficiency of the different nCas9-RT constructs, further optimization is still required to provide a more precise representation.

Further Actions

Our re-designed approach 1 provides a representative assay for discriminating prime editing efficiency in different constructs, hence no further optimizations are required. However, initial testing via approach 2 has prompted the search of a more optimal cell line that provides a more accurate estimation of editing efficiencies. The successful TLR design (PETLR 1) was further utilised for testing of the alternative RTs and prime editing in stem cells and other chassis.

References

1. Certo, M.T., et al., Tracking genome engineering outcome at individual DNA breakpoints. Nat Methods, 2011. 8(8): p. 671-6.