Quantitative analysis of recombination between YFP and CFP genes of FRET biosensors introduced by lentiviral or retroviral gene transfer

Abstract

Biosensors based on the principle of Förster (or fluorescence) resonance energy transfer (FRET) have been developed to visualize spatio-temporal dynamics of signalling molecules in living cells. Many of them adopt a backbone of intramolecular FRET biosensor with a cyan fluorescent protein (CFP) and yellow fluorescent protein (YFP) as donor and acceptor, respectively. However, there remains the difficulty of establishing cells stably expressing FRET biosensors with a YFP and CFP pair by lentiviral or retroviral gene transfer, due to the high incidence of recombination between YFP and CFP genes. To address this, we examined the effects of codon-diversification of YFP on the recombination of FRET biosensors introduced by lentivirus or retrovirus. The YFP gene that was fully codon-optimized to E.coli evaded the recombination in lentiviral or retroviral gene transfer, but the partially codon-diversified YFP did not. Further, the length of spacer between YFP and CFP genes clearly affected recombination efficiency, suggesting that the intramolecular template switching occurred in the reverse-transcription process. The simple mathematical model reproduced the experimental data sufficiently, yielding a recombination rate of 0.002-0.005 per base. Together, these results show that the codon-diversified YFP is a useful tool for expressing FRET biosensors by lentiviral or retroviral gene transfer.

Figure 1. Recombination of FRET biosensors during lentiviral or retroviral infection.

(A) Schematic representation of the recombination between YFP and CFP genes in FRET biosensors in the process of lentiviral or retroviral gene transfer. Two copackaged genomic RNAs encoding FRET biosensors are included in a virus particle. After infection, cells express only YFP or CFP. (B) FRET biosensors with different YFP variants. A PKA FRET biosensor, AKAR3EV, is composed of YPet (YFP), a FHA1 domain, linker, PKA substrate, nTurquoise-GL (CFP), and a nuclear export sequence (NES). In this study, YPet is replaced with h100YPet, H75-e25YPet, h50-e50YPet, h25-e75YPet, e100YPet, e75-h25YPet, e50-h50YPet, and e25-h75YPet.

Figure 2. Working model for the recombination in FRET biosensors.

(A) The recombination of FRET biosensors with h25-e75YPet generates CFP, which includes the critical amino acid substitution of Y66W from GFP. (B) The recombination of FRET biosensors with e75-h25YPet generates GFP or YFP, which includes the critical amino acid substitution of T203Y from GFP.

Figure 3. Recombination between the YFP and CFP genes by lentiviral gene transfer.

(A–H) A549 cells were infected with lentivirus encoding 8 different FRET biosensors as shown in Fig. 1B. At least 4 days after infection, the cells were imaged with an epi-fluorescence microscope. The average fluorescence intensities of CFP and YFP are represented as a log-log plot. Each dot corresponds to an A549 cell. Three hundred cells were analyzed from two independent experiments. Red lines are the fitted line with the e100YPet data. Orange and cyan arrowheads indicate the T203Y and Y66W positions, respectively.

Figure 4. Recombination between the YFP and CFP genes by retroviral gene transfer.

(A–H) A549 cells were infected with retrovirus encoding 8 different FRET biosensors as shown in Fig. 1B. At least 4 days after infection, the cells were imaged with an epi-fluorescence microscope. The average fluorescence intensities of CFP and YFP are represented as a log-log plot. Each dot corresponds to an A549 cell. Three hundred cells were analyzed from two independent experiments. Red lines are the fitted line with the e100YPet data. Orange and cyan arrowheads indicate the T203Y and Y66W positions, respectively.

Figure 5. Effect of a short spacer between YFP and CFP on recombination.

(A) Schematic representation of a FRET biosensor with YPet, GGSGG linker, nTurquoise-GL, and NES. (B,C) HeLa cells were infected with lentivirus encoding FRET biosensor with a short spacer, 15 bases (B) and full spacer, 812 bases (C). At least 4 days after infection, 300 cells were imaged with an epi-fluorescence microscope, and represented as in Fig. 3. Note that panel C is the same graph as in Supplementary Fig. 2A. Orange and cyan arrowheads indicate the T203Y and Y66W positions, respectively.

Figure 6. Computer simulation of recombination between the YFP and CFP genes.

The recombination of YFP and CFP genes in A549 cells infected with the indicated lentivirus was simulated by computer with a recombination rate of 0.0043 (/base), which showed maximal likelihood estimation. To reproduce the experimental data, 9 parameters were extracted from the experimental data set in Fig. 3 (see Supplementary Fig. S6 and the Methods for details). Red lines were fitted with the e100YPet data. Orange and cyan arrowheads indicate the T203Y and Y66W positions, respectively.