Generation of a highly inducible Gal4-->Fluc universal reporter mouse for in vivo bioluminescence imaging

Abstract

Full understanding of the functional complexity of the protein interactome requires mapping of biomolecular complexes within the cellular environment over biologically relevant time scales. New approaches to imaging interacting protein partners in vivo will allow the study of functional proteomics of human biology and disease within the context of living animals. Herein, we describe a universal transgenic reporter mouse strain that expresses firefly luciferase (Fluc) under the regulatory control of a concatenated Gal4 promoter (Tg(G4F(+/-))). Using an adenovirus to deliver a fused binding-domain-activator chimera (Gal4BD-VP16), induction of bioluminescence in Tg(G4F(+/-)) tissues of up to 4 orders of magnitude was observed in fibroblasts, liver, respiratory epithelia, muscle, and brain. The Tg(G4F(+/-)) reporter strain allowed noninvasive detection of viral infectivity, duration of the infection as well as viral clearance in various tissues in vivo. To demonstrate protein-protein interactions in live mice, the well characterized interaction between tumor suppressor p53 (fused to Gal4BD) and large T antigen (TAg) (fused to VP16) was visualized in vivo by using a two-hybrid strategy. Hepatocytes of Tg(G4F(+/-)) mice transfected with p53/TAg demonstrated 48-fold greater induction of Fluc expression in vivo than noninteracting pairs. Furthermore, to demonstrate the feasibility of monitoring experimental therapy with siRNA in vivo, targeted knockdown of p53 resulted in markedly reduced light output, whereas use of a control siRNA had no effect on protein interaction-dependent induction of Fluc. Thus, this highly inducible Gal4-->Fluc conditional reporter strain should facilitate imaging studies of protein interactions, signaling cascades, viral dissemination, and therapy within the physiological context of the whole animal.

Fig. 1.

Generation and analysis of the reporter transgene. (A) Schematic representation of the transgene vector that was injected into pronuclei of FVB oocytes. (B) HeLa Fluc #43, a cell line stably expressing the Gal4→Fluc reporter construct, was transfected with a positive control vector pM3-VP16 that expresses a fusion of Gal4BD to VP16 (right well). A 43-fold induction was observed when compared with mock-transfected cells (left well). (C) PCR genotyping from a negative mouse (WT), a positive littermate (Tg^G4F(+/−)), and pGL3 as positive control are shown. (D) Mouse embryonic fibroblasts (MEF) of a negative embryo (WT) and a positive embryo (Tg^G4F(+/−)) were treated with either vehicle or 4 × 10⁷ pfu of Ad5Gal4BD-VP16 (Ad-Gal4). Cells were imaged 24 h after infection. MEFs from the positive embryo were induced 3-log fold when compared with the uninducible negative MEFs.

Fig. 2.

Transactivation of the reporter gene in the abdomen. (A) Tg^G4F(+/−) and WT mice were injected i.v. with adenovirus Ad-Gal4 expressing the fusion construct (Gal4BD fused to VP16). Mice were imaged before treatment with virus (day 0) and then followed over time as indicated, starting 24 h postinjection (day 1). Images from a representative Tg^G4F(+/−) mouse are shown. The signal on day 1 was induced over 3 orders of magnitude. As a control, Tg^G4F(+/−) mice were injected i.v. with an off-target adenovirus, Ad-Cre, and imaged over time. No signal was detected (day 2 is shown). (B) Signal progression of Fluc activity in these mice was followed for 14 days until the signal declined back to almost pretreatment levels. Shown is total photon flux plotted over time. Data are presented as mean ± SEM (n = 3 for Tg^G4F(+/−) mice) and mean ± range (n = 2 for WT); error bars for WT are smaller than symbol size. (C) Organs from Tg^G4F(+/−) mice administered Ad-Gal4 or Ad-Cre (as indicated) were also imaged ex vivo (day 2). Signal from the liver of an Ad-Gal4-treated mouse was 3-log greater than the liver of an Ad-Cre-treated mouse. Modest signal was also detectable from the spleen and kidney of the Ad-Gal4-treated mouse, whereas low basal activity was observed in the hearts of both.

Fig. 3.

Transactivation of the reporter gene in multiple tissues. (A) Tg^G4F(+/−) and WT mice inhaled 2 × 10⁶ pfu per mouse of Ad-Gal4. Mice were imaged before treatment with virus (day 0) and then followed over time, starting with 24 h post inhalation treatment. Images from a representative Tg^G4F(+/−) mouse are shown. (B) Signal progression of the mice was followed for 9 days until the signal declined back to almost pretreatment levels and is shown as total photon flux plotted over time. Data are presented as mean ± SEM (n = 3 each); error bars for WT are smaller than symbol size. (C) Tg^G4F(+/−) mice administered Ad-Gal4 or off-target Ad-Cre (as indicated) were imaged through an open chest cavity on day 3. Signal was detectable from the nasal passages and lung of only the Ad-Gal4-treated mouse. (D) Tg^G4F(+/−) and WT mice were injected with 2.3 × 10⁵ pfu per mouse of Ad-Gal4 (right calf muscle) and with vehicle (left calf muscle). Mice were imaged on day 1 (24 h after virus injection) and followed over time. Images from a representative Tg^G4F(+/−) mouse are shown. (E) Signal progression of Fluc activity was followed for 4 days and is shown as total photon flux plotted over time. Data are presented as mean ± range (n = 2 each); error bars for WT are smaller than symbol size. (F) Tg^G4F(+/−) mice were injected identically with Ad-Gal4 (right calf muscle) or with Ad-Cre (left calf muscle). Muscles were excised on day 3 and imaged ex vivo. (G) Tg^G4F(+/−) and WT mice were injected intracranially with 2 × 10⁶ pfu per mouse of Ad-Gal4 2 mm lateral and posterior of the bregma and 3 mm below the dura. Mice were imaged before treatment with virus (day 0) and then followed over time, starting with 24 h postinjection. Images from a representative Tg^G4F(+/−) mouse are shown. (H) Signal progression was followed for 9 days and is shown as total photon flux plotted over time. Data are presented as mean ± SEM (n = 3 for Tg^G4F(+/−) mice) and without error bars for WT (n = 1).

Fig. 4.

Imaging protein–protein interactions in vivo using a Gal4→Fluc reporter mouse. (A) Hepatocytes of Tg^G4F(+/−) mice were transfected in vivo by hydrodynamic somatic gene transfer with different plasmid combinations of pGal4BD-p53 together with pVP16-TAg or off-target pVP16-CP, or each plasmid separately as indicated. A plasmid for Renilla luciferase was used as transfection control. Mice were imaged before (pretreatment) and 24 h after transfection for expression of firefly luciferase (Fluc with D-luciferin) and Renilla luciferase (Rluc with coelenterazine). Images of representative mice are shown. CP; polyoma virus coat protein. (B) Data are plotted for the different plasmid combinations as controlled for transfection efficiencies (Fluc/Rluc; left axis) and as fold-induction (right axis) 24 h after transfection. Data are presented as mean ± range for plasmid combinations (n = 2), and without error bars for single plasmids (n = 1). Four independent experiments showed similar results.

Fig. 5.

Imaging shRNAi abrogation of protein–protein interactions in vivo. (A) Hepatocytes of Tg^G4F(+/−) mice were transfected in vivo by hydrodynamic somatic gene transfer with plasmid combinations of pGal4BD-p53 together with pVP16-TAg plus shRNAi-control or shRNAi-p53. A plasmid for Renilla luciferase was used as transfection control. Mice were imaged 24 h after transfection for expression of firefly luciferase (D-luciferin) or Renilla luciferase (coelenterazine). Images from representative mice are shown. (B) Data are plotted for the shRNAi-control and shRNAi against p53. Total photon flux was normalized for transfection efficiencies (Fluc/Rluc). Data are presented as mean ± SEM for shRNAi-53 (n = 3) and mean ± range for shRNAi-control (n = 2). Three independent experiments showed similar results. (C) PCRs of DNA isolated from livers of Tg^G4F(+/−) mice that had been injected with plasmid combinations of pGal4BD-p53 together with pVP16-TAg (mouse 1) or pGal4BD-p53 together with pVP16-TAg plus shRNAi-p53 (mouse 2) demonstrate equal delivery of the plasmids encoding the interacting proteins.