Development of a Lentivirus Vector-Based Assay for Non-Destructive Monitoring of Cell Fusion Activity

Cell-to-cell fusion can be quantified by endowing acceptor and donor cells with latent reporter genes/proteins and activators of these genes/proteins, respectively. One way to accomplish this goal is by using a bipartite lentivirus vector (LV)-based cell fusion assay system in which the cellular fusion partners are transduced with a flippase-activatable Photinus pyralis luciferase (PpLuc) expression unit (acceptor cells) or with a recombinant gene encoding FLPeNLS+, a nuclear-targeted and molecularly evolved version of flippase (donor cells). Fusion of both cell populations will lead to the FLPe-dependent generation of a functional PpLuc gene. PpLuc activity is typically measured in cell lysates, precluding consecutive analysis of one cell culture. Therefore, in this study the PpLuc-coding sequence was replaced by that of Gaussia princeps luciferase (GpLuc), a secretory protein allowing repeated analysis of the same cell culture. In myotubes the spread of FLPeNLS+ may be limited due to its nuclear localization signal (NLS) causing low signal outputs. To test this hypothesis, myoblasts were transduced with LVs encoding either FLPeNLS+ or an NLS-less version of FLPe (FLPeNLS−) and subsequently co-cultured in different ratios with myoblasts containing the FLPe-activatable GpLuc expression cassette. At different times after induction of cell-to-cell fusion the GpLuc activity in the culture medium was determined. FLPeNLS+ and FLPeNLS− both activated the latent GpLuc gene but when the percentage of FLPe-expressing myoblasts was limiting, FLPeNLS+ generally yielded slightly higher signals than FLPeNLS− while at low acceptor-to-donor cell ratios FLPeNLS− was usually superior. The ability of FLPeNLS+ to spread through myofibers and to induce reporter gene expression is thus not limited by its NLS. However, at high FLPe concentrations the presence of the NLS negatively affected reporter gene expression. In summary, a rapid and simple chemiluminescence assay for quantifying cell-to-cell fusion progression based on GpLuc has been developed.


Introduction
During cell-to-cell fusion, plasma membranes of individual cells merge to form a multinucleated structure called a syncytium. Plasma membrane fusion is a crucial event during, for example, fertilization, syncytiotrophoblast production, skeletal muscle formation, bone remodeling, eye lens development and certain forms of tissue repair [1]. In general, cell fusion is a tightly regulated and highly selective process involving specific cell types. Inappropriate cell fusion has been implicated in tumor development and progression [2].
Cell fusion can be easily observed using microscopic techniques and in many studies the extent of cell fusion is expressed as fusion index, which either stands for the percentage of cells with two or more nuclei or the percentage of nuclei present in syncytia [3]. However, without continuous monitoring, it is impossible to decide by microscopy alone whether multinucleation is caused by cell fusion or the result of karyokinesis without cytokinesis. In addition, cells growing on top of each other can be mistaken for syncytia. Furthermore, as fusion index determinations are generally carried out manually, they are laborious, error-prone and often inaccurate. This has led to the development of methods for quantifying cell fusion independent of microscopic inspection.
Nearly all these methods are based on systems of two components that interact to create a novel detectable signal only after cell fusion [3]. Mohler and Blau, for example, developed a quantitative cell fusion assay based on functional complementation between two biologically inactive b-galactosidase deletion mutants [4]. Another possibility to produce fusion-dependent signals is by applying site-specific recombination systems such as Cre-loxP and FLP-FRT. In these systems, a latent reporter gene is activated by the action of the site-specific DNA recombinase Cre from bacteriophage P1 or flippase/FLP from Saccharomyces cerevisiae, which catalyze the excision and inversion of DNA flanked by 34base pair (bp) recognition sequences (loxP for Cre and FRT for FLP) in a direct or inverted repeat configuration, respectively [5,6].
Gonçalves et al. previously developed a bipartite lentivirus vector (LV)-based cell fusion assay system in which the cellular fusion partners are endowed with a FLP-activatable Photinus pyralis luciferase (PpLuc) expression unit/''gene switch'' (acceptor cells) or with a recombinant gene encoding a molecularly evolved version of FLP (FLPe) with a nuclear localization signal (NLS) derived from the simian virus 40 large T antigen (donor cells) [7]. Fusion between acceptor and donor cells led to the FLPe-dependent generation of a functional episomal PpLuc expression module. This cell fusion monitoring system was successfully used to study the role of the p38 MAPK signaling pathway in myoblast fusion/ myotube formation. However, since PpLuc is a cytoplasmic protein and its substrate D-luciferin is poorly membrane-permeable, this assay requires lysis of the cells prior to luminometry and does not allow repeated analysis of the same cell culture. This prompted us to develop a nondestructive method to quantify cell fusion using the bipartite LV-based cell fusion assay system described by Gonçalves and colleagues as starting point.
The key difference between the new and ''old'' version of the LV-based cell fusion assay system is the replacement of the PpLuc open reading frame (ORF) in the ''original'' gene switch construct by the humanized coding sequence of Gaussia princeps luciferase (GpLuc), which is a secretory protein converting the substrate coelenterazine into coelenteramide plus light. GpLuc also displays a much higher specific luciferase activity than PpLuc and is exceptionally resistant to exposure to heat and strongly acidic and basic conditions [8]. In addition, we hypothesized that in myotubes the spread of nuclear-targeted FLPe (FLPe NLS+ ) beyond the direct surroundings of donor nuclei may be limited due to the presence of the NLS. This would result in the activation of only a fraction of the reporter genes especially in hybrid myotubes containing a relatively low percentage of FLPe gene-positive donor nuclei compared to GpLuc-encoding acceptor nuclei. To test this hypothesis, we generated an LV encoding an NLS-less version of FLPe (FLPe NLS2 ) and compared, in myogenic fusion assays, its ability to activate latent GpLuc genes with that of FLPe NLS+ .

Materials and Methods
Plasmids DNA constructions were carried out with enzymes from Fermentas (Fisher Scientific, Landsmeer, the Netherlands) or from New England Biolabs (Bioké, Leiden, the Netherlands) by using established procedures [9] or following the instructions provided with specific reagents.
The ligation mixtures were introduced in chemocompetent cells of Escherichia coli strain GeneHogs (Life Technologies Europe, Bleiswijk, the Netherlands) or GT115 (InvivoGen, San Diego, CA). Large-scale plasmid purifications were performed using JETSTAR 2.0 Plasmid Maxiprep kits (Genomed, Löhne, Germany) according to the manufacturer's instructions.

Cells
The culture and differentiation conditions of the murine Bmi1and human TERT-immortalized human myoblasts (iDMD myoblasts) have been described previously [13].  Fig. 1C), pLV.GS.GpLuc.v1, pLV.GS.PpLuc and pLV.GS.GpLuc.v6, respectively. The 293T cells were transfected with one of the LV shuttle constructs and the packaging plasmids psPAX2 (Addgene; plasmid number: 12260) and pLP/VSVG (Life Technologies Europe) at a molar ratio of 2:1:1. To concentrate and purify the LV particles, producer cell supernatants were layered onto 5-ml cushions of 20% (wt/vol) sucrose in phosphate-buffered saline (PBS) and centrifuged at 15,000 rotations per minute for 2 h at 4uC in an SW32 rotor (Beckman Coulter Nederland, Woerden, the Netherlands). Prior to ultracentifugation, producer cell supernatants were clarified by low speed centrifugation and filtration through 0.45-mm pore-sized cellulose acetate filters (Pall Netherlands, Mijdrecht, the Netherlands). For more details about the SIN-LV production method, see [15]. The titers of the resulting LV stocks were determined using the RETROTEK HIV-1 p24 Antigen ELISA kit (ZeptoMetrix, Franklin, MA) following the instructions provided by the manufacturer. To derive functional titers from these measurements a conversion factor of 2.5 transducing units (TUs) per pg of HIV-1 p24 protein was used.

Cell transductions
Cryopreserved LV.FLPe NLS+ .PurR-transduced iDMD myoblasts ( [7]; hereinafter referred to as myoblasts-FLPe NLS+ ) were thawed and cultured in the presence of puromycin (Life Technologies Europe) at a final concentration of 0.4 mg/ml to prevent transgene silencing. FLPe NLSexpressing iDMD myoblasts were generated by overnight (620 h) exposure of 10 5 cells in a well of a 24-well cell culture plate (Greiner Bio-One, Alphen aan den Rijn, the Netherlands) to 30 TUs of LV.FLPe NLS2 .PurR per cell in 500 ml of growth medium in a humidified atmosphere of 5% CO 2 /95% air at 37uC. The next day, the cell monolayer was rinsed three times with 1 ml of PBS after which fresh culture medium was added. At 3 days post transduction, the culture of LV.FLPe NLS2 .PurR-treated cells (hereinafter referred to as myoblasts-FLPe NLS2 ) as well as a control culture of untransduced iDMD myoblasts were given medium containing 0.8 mg/ml of puromycin. Within a week, all cells in the culture of untransduced iDMD myoblasts had died while the cells in the LV.FLPe NL-S2 .PurR-treated culture were nicely expanding. The myoblasts-FLPe NLS2 were passaged once a week (split ratio 1:3) in growth medium containing 0.4 mg/ml of puromycin. Myoblasts GS.GLuc , myoblasts GS.PLuc and myoblasts GS.GLuc+ were generated likewise by exposure of iDMD myoblasts to LV.GS.GpLuc.v1, LV.GS.PpLuc and LV.GS.GpLuc.v6, respectively. Before being used for co-culture experiments, the cells were passaged at least three times to rule out secondary transduction of the FLPeexpressing myoblasts in the co-cultures with luciferase-encoding SIN-LVs [16].

Co-culture establishment and maintenance
Co-cultures containing a total number of 2610 5 cells were established in wells of 24-well culture plates by mixing myoblasts-FLPe NLS+ or myoblasts-FLPe NLS2 with myoblasts GS.GLuc at the indicated ratios. Following an incubation period of about 72 h when the cell monolayers had reached 90-100% confluency, the growth medium was substituted by 400 ml of either differentiation medium or fresh growth medium. At specified time points thereafter, the culture medium (400 ml) was collected and stored at 280uC for luciferase assay. The co-cultures were then either terminated or further incubated at 37uC in a water-saturated atmosphere of 5% CO 2 /95% air. To compare the performance of the newly developed LV.GS.GpLuc.v1-based cell fusion assay system with that of the previously described LV.GS.PpLuc-based cell fusion quantification method [7], myoblasts GS.GLuc or myoblasts GS.PLuc were cocultured with myoblasts-FLPe NLS+ in different ratios in 24-well culture plates containing 2610 5 cells per well. Samples (culture fluid for cultures containing myoblasts GS.GLuc and cell lysates for cultures containing myoblasts GS.PLuc ) were harvested 96 h and 120 h after induction of myogenic differentiation. Exactly the same approach was used to compare the LV.GS.GpLuc.v1-and LV.GS.GpLuc.v6-based cell fusion assays.

Immunocytology
At different time points after the initiation of differentiation, 1:1 co-cultures of myoblasts-FLPe NLS2 and myoblasts GS.GLuc were fixed by incubation for 30 minutes at room temperature (RT) in PBS containing 4% formaldehyde. To permeabilize the cells, they were exposed for 10 minutes at RT to 0.1% Triton X-100 in PBS. Next, cells were incubated overnight at 4uC with mouse antiskeletal muscle troponin I (skTnI) primary antibody (HyTest, Turku, Finland; clone 12F10) diluted 1:100 in PBS +0.1% donkey serum (Santa Cruz Biotechnology, Santa Cruz, CA) followed by a 2-h incubation at RT with Alexa Fluor 568-conjugated donkey anti-mouse IgG (H+L) secondary antibody (Life Technologies Europe) diluted 1:400 in PBS +0.1% donkey serum. Counter-staining of nuclei was performed with 10 mg/ml Hoechst 33342 (Life Technologies Europe) in PBS. Cells were washed three times with PBS after fixation, permeabilization and incubation with primary antibody, secondary antibody and DNA-binding fluorochrome. To minimize photobleaching, coverslips were mounted in Vectashield mounting medium (Vector Laboratories, Burlingame, CA). Pictures were taken with a fluorescence microscope equipped with a digital color camera (Nikon Eclipse 80i; Nikon Instruments Europe, Amstelveen, the Netherlands) using NIS Elements software (Nikon Instruments Europe).

Subcellular fractionation and western blotting
Myoblasts-FLPe NLS+ and myoblasts-FLPe NLS2 were cultured separately in 24-well cell culture plates at a density of 2610 5 cells per well. Following an incubation period of 72 h when the cell monolayers had reached 90-100% confluency, the growth medium was substituted by 400 ml of either differentiation medium or fresh growth medium. Ninety-six h later, cell fractionation was carried out as described by Suzuki et al. [17] with the following modifications. Cell pellets were suspended in 97.5 ml of ice-cold 0.1% NP40 in PBS. One-third of the lysate was removed as ''whole cell lysate'' and mixed with 5 ml of 106 NuPAGE Sample Reducing Agent and 12.5 ml of 46 NuPAGE LDS Sample Buffer (both from Life Technologies Europe). The rest of the lysate was briefly centrifuged at 4uC after which 32.5 ml of the supernatant was removed as ''cytosolic fraction'' and supplemented with 5 ml of 106NuPAGE Sample Reducing Agent and 12.5 ml of 46 NuPAGE LDS Sample Buffer. The remaining supernatant was removed and the pellet was washed with and suspended in 30 ml PBS, after which 5 ml of 106 NuPAGE Sample Reducing Agent and 12.5 ml of 46NuPAGE LDS Sample Buffer were added to produce the ''nuclear fraction''. Nuclear fractions and whole cell lysates were sonicated for 2 times 10 seconds at 200 Hz using a Soniprep 150 ultrasonic disintegrator (Measuring and Scientific Equipment, London, United Kingdom). After incubating the samples for 1 minute at 100uC, 10 ml of whole cell lysate, 10 ml of cytosolic fraction and 5 ml of nuclear fraction were applied to a NuPAGE Novex 12% Bis-Tris gel (Life Technologies Europe). Following electrophoretic separation, the proteins were transferred to a polyvinylidene difluoride membrane (Amersham Hybond P; GE Healthcare Europe, Diegem, Belgium) by wet electroblotting. Next, the membrane was incubated with 2% ECL AdvanceTM blocking agent (GE Healthcare Europe) in PBS-0.1% Tween 20 (PBST) for 1 h at RT and probed with rabbit anti-FLP (1:200; Diagenode, Seraing, Belgium; CS-169-100), mouse anti-glyceraldehyde 3-phosphate dehydrogenase (GAPDH; 1:10,000; Merck Millipore, Billerica, MA; clone 6C5) or rabbit anti-lamin A/C (1:10,000; Santa Cruz Biotechnology; sc-20681) primary antibodies overnight at 4uC, followed by a 1-h incubation with appropriate horseradish peroxidase-conjugated secondary antibodies (Santa Cruz Biotechnology). GAPDH served as cytoplasmic marker protein and lamin A/C antibody was used as nuclear marker protein. Target protein signals were visualized using the SuperSignal West Femto Maximum Sensitivity Substrate Kit (Thermo Scientific, Rockford, IL) and chemiluminescence was measured with the ChemiDoc XRS imaging system (Bio-Rad Laboratories, Veenendaal, the Netherlands).

FLPe functionality test
To test the functionality of the FLPe molecules encoded by LV.FLPe NLS+ .PurR and LV.FLPe NLS2 .PurR, myoblasts GS.GLuc were transduced with LV.FLPe NLS+ .PurR, LV.FLPe NLS2 .PurR or LV.PurR. Myoblasts GS.GLuc were seeded in a 24-well cell culture plate at a density of 10 5 cells per well and exposed for 20 h to 75 ml per well of concentrated vector stock diluted in growth medium to a final volume of 500 ml. Next, the cell monolayers were rinsed three times with 1 ml of PBS after which 400 ml fresh growth medium was added. At 24 h after the removal of the inoculum, the culture medium was collected and transiently stored at 280uC for subsequent analysis of luciferase activity. The cells were overlaid with 400 ml of fresh growth medium, which was harvested 24 h later for storage at 280uC until luciferase activity measurement.

Luciferase assay
After thawing the GpLuc-containing samples on ice, 50 ml of each sample was transferred to a well of a white opaque 96-well flat-bottom microtiter plate (OptiPlate-96; PerkinElmer, Groningen, the Netherlands) for chemiluminescence measurements. The native coelenterazine (Promega Benelux, Leiden, the Netherlands) stock solution (5 mg/ml in acidified methanol) was diluted 1,000 times in phenol red-free Dulbecco's modified Eagle's medium (Life Technologies Europe) and equilibrated for 1 h in the dark at RT before starting the measurements. The luciferase activity was measured at RT with the aid of a Wallace 1420 VICTOR 3 multilabel plate reader with automatic injection system (PerkinElmer). Immediately after automated addition of 20 ml of substrate to a well, substrate and sample were mixed by shaking for 1 second (double orbital, 0.1 mm, normal speed). PpLuc activity was measured in cell lysates as previously described [7]. For each condition, three independent samples were measured in three series of measurements.

Statistical analysis
Different experimental groups were compared using the independent samples t-test. Differences among means were considered significant at P#0.05. Graphs were prepared in GraphPad Prism version 5 (GraphPad Software, La Jolla, CA).

Microscopic analysis of cell fusion kinetics
Cultured myoblasts can be prompted to fuse with each other by withdrawing mitogens from the culture medium. This causes a time-dependent accumulation of nuclei in syncytial structures called myotubes or myosacs depending on whether these structures are elongated or rounded. To get a first impression of the cell-to-cell fusion kinetics of the genetically modified iDMD myoblasts, 1:1 co-cultures of myoblasts-FLPe NLS2 and myoblasts GS.GLuc were exposed to myogenic differentiation conditions. As shown in the upper panel of Fig. 4, the myoblasts started to fuse 48 h after serum withdrawal resulting in the formation of myotubes/sacs. Both the percentage of nuclei present in myotubes/sacs as well as the size of the syncytia increased with time until 120 h following serum removal, after which the cells started to detach from the surface of the culture plates. The fusion process was accompanied by the accumulation of sarcomeric proteins as evinced by the results of the skTnI-specific immunostaining depicted in the lower panel of Fig. 4.

Immunodetection of FLPe in LV.FLPe NLS+/2 .PurRtransduced iDMD myoblasts
To compare FLPe protein level and intracellular distribution between myoblasts-FLPe NLS+ and myoblasts-FLPe NLS2 , western blot analysis was performed on whole cell lysates as well as on nuclear and cytosolic cell fractions (Fig. 5A). As expected from the presence at its amino terminus of the SV40 NLS, FLPe NLS+ (predicted molecular weight: 49.7 kilodaltons) had a slightly lower gel mobility than FLPe NLS2 (predicted molecular weight: 48.6 kilodaltons). Both under growth and differentiation conditions, the steady-state level of FLPe NLS+ was considerably higher than that of FLPe NLS2 even though the nucleotide sequences upstream of the FLPe start codon are very similar and both proteins contain a ''destabilizing'' amino acid residue (serine in FLPe NLS2 versus alanine in FLPe NLS+ ; [18]) immediately downstream of the initiator methionine. Fig. 5A also reveals that a larger fraction of FLPe NLS+ molecules than of FLPe NLS2 molecules resides in the nucleus (nuclear-to-cytosolic ratios under differentiation conditions of 8.4 and 3.1, respectively) consistent with the presence in FLPe NLS+ of an SV40 NLS.  To compare the ability of FLPe NLS+ and FLPe NLS2 to activate the GpLuc gene switch upon cell fusion, myoblasts GS.GLuc were cocultured with myoblasts-FLPe NLS+ or myoblasts-FLPe NLS2 at different ratios (i.e. 95:5, 90:10, 75:25, 50:50, 25:75, 10:90 and 5:95). Monocultures of myoblasts-FLPe NLS+ , myoblasts-FLPe NLS2 or myoblasts GS.GLuc exposed to growth or differentiation medium and co-cultures of FLPe-expressing myoblasts and myoblasts GS.-GLuc maintained in growth medium served as negative controls. Based on the results of the microscopic analysis of cell fusion activity (Fig. 4), the culture medium was harvested 96 h after induction of myogenic differentiation. It should be noted, however, that the kinetics of cell fusion progression slightly differed between individual experiments probably reflecting small differences in the myoblast populations used for different experiments. Luciferase activity in the medium of the fusogenic cell cultures depended on the ratio of myoblasts GS.GLuc and myoblasts-FLPe, showed a similar trend for myoblasts-FLPe NLS+and myoblasts-FLPe NLScontaining co-cultures and was highest when co-cultures contained 50-95% myoblasts GS.GLuc (Fig. 7A). The peak of GpLuc activity was reached at myoblast GS.GLuc :myoblast-FLPe ratios of 90:10 and 75:25 for myoblasts-FLPe NLS+ and myoblasts-FLPe NLS2 , respectively (Fig. 7A). Interestingly, at low myoblast GS.GLuc :myoblast-FLPe ratios (i.e. 10:90 and 5:95) the luciferase activity was significantly higher for myoblasts-FLPe NLS2 than for myoblasts-FLPe NLS+ (Fig. 7A). Myoblast cultures kept under growth conditions and myoblast-FLPe monocultures maintained in differentiation medium yielded luminescence signals close to or at background levels. The monocultures of myoblasts GS.GLuc did, however, secrete detectable amounts of GpLuc under differentiation conditions although the signal intensity was much lower than that produced by serum-deprived co-cultures containing 50-90% myoblasts GS.GLuc . For the co-cultures containing 50-90% myoblasts GS.GLuc shifting from growth to differentiation medium resulted in a .100-fold increase in luciferase activity (Fig. 7B).

Use of the LV.FLPe NLS+/2 .PurR/LV.GS.GpLuc-based cell fusion assay system to analyse cell fusion progression
To investigate the utility of the LV.FLPe NLS+/2 .PurR/ LV.GS.GpLuc-based cell fusion assay system to follow cell fusion progression, myoblasts GS.GLuc were mixed with myoblasts-FLPe NLS+ or with myoblasts-FLPe NLS2 at a ratio of 50:50. After the cell cultures had become nearly confluent, they were either given fresh growth medium or exposed to differentiation medium. This was followed by the periodic collection of culture fluid for luciferase measurements using two different approaches. In one experiment, the culture medium was left on the cells for different time periods (i.e. from 0-24, 0-36, 0-48, 0-60, 0-72, 0-84, 0-96, 0-108 and 0-120 h) before being harvested for luminometry (''cumulative assay''; Fig. 8). In the other experiment, the culture  Fig. 9). As shown in Fig. 8A, following an initial slow increase, the luciferase activity in the culture medium of the serum-deprived co-cultures rose sharply at late times (.72 h) after initiation of differentiation. Co-cultures of myoblasts GS.GLuc and myoblasts-FLPe NLS2 produced better results than the combination of myoblasts GS.GLuc and myoblasts-FLPe NLS+ (Fig. 8A,B) in spite of the much higher FLPe concentration in myoblasts-FLPe NLS+ than in myoblasts-FLPe NLS2 (Fig. 5A). These findings were corroborated by the data derived from the ''kinetics assay'' (Fig. 9).
. On the basis of the previous results, another experiment was carried out in which we directly compared the performance of FLPe NLS+ and FLPe NLS2 at different myoblast GS.GLuc :myoblast-FLPe ratios (i.e. 95:5, 75:25, 25:75 and 5:95) and different time points (i.e. 72, 96 and 120 h after serum withdrawal). The culture medium was refreshed just before the start of the first sampling interval (i.e. at 48 h after serum removal) and after each round of sample collection. This experiment confirmed that at high myoblast GS.GLuc :myoblast-FLPe ratios FLPe NLS2 was nearly as efficient as FLPe NLS+ at inducing reporter gene expression while at low myoblast GS.GLuc :myoblast-FLPe ratios FLPe NLS2 gave rise to more RLUs (Fig. 10A) and to higher signal-to-noise ratios  (Fig. 10B). In accordance with the experiment presented in Fig. 7, the co-cultures consisting of 75% myoblasts GS.GLuc and 25% myoblasts-FLPe yielded the highest signals both in absolute (Fig. 10A) and relative (Fig. 10B) terms. Also in line with the previous experiments was the finding that most GpLuc accumulation takes place between 96 and 120 h after serum removal.  In the next experiment, a direct comparison was made between the previously described LV.GS.PpLuc-based quantitative cell fusion assay system [7] and the new LV.GS.GpLuc-based method to quantify cell-to-cell fusion. Consistent with the much higher light output of GpLuc than of PpLuc [8], LV.GS.GpLuc yielded up to 23-fold higher signals than LV.GS.PpLuc (Fig. 11A). However, the LV.GS.PpLuc-based cell fusion assay system appeared to be approximately twice as sensitive as its LV.GS.GpLuc-based counterpart at detecting myoblast-to-myoblast fusion at 120 h after initiation of differentiation (Fig. 11B). The difference in sensitivity between the GS.GpLuc.v1-and LV.GS.PpLuc-based cell fusion assay systems was even bigger for the samples collected at 96 h after serum removal especially at the lowest two myoblast GS.Luc :myoblast-FLPe NLS+ ratios (i.e. when FLPe levels are highest).

Improvement of the GS.GpLuc-based cell fusion assay system
The results presented in Figs. 7 and 11 identify the FLPindependent increase in GpLuc production when shifting from growth to differentiation medium as the main contributor to the reduced signal-to-noise ratio of the LV.GS.GpLuc-based cell fusion assay system as compared to its LV.GS.PpLuc-based counterpart. In search for a possible explanation for the high background signal produced by LV.GS.GpLuc.v1 in comparison to LV.GS.PpLuc, we compared their genetic organization upstream of the Luc start codon. As shown in Figs. 2 A,B and 3 the PpLuc ORF in LV.GS.PpLuc is preceded by an out-of-frame ORF (uORF) starting with 2 ATG codons in a favourable context for translational initiation [19] and ending with a highly efficient stop codon [20] separated by only 7 nucleotides from the PpLuc initiation codon. This specific genetic makeup will be effective in supressing any PpLuc expression directed by mRNAs with 59 ends located upstream of the second ATG codon in the uORF. Oppositely, in LV.GS.GpLuc.v1 the previously mentioned tandem of ATG codons are in-frame with the GpLuc initiation codon allowing the synthesis of an N-terminally extended GpLuc fusion protein. Located further upstream of the GpLuc ORF in LV.GS.GpLuc.v1 is an out-of-frame ORF with suboptimal start and stop codons. LV.GS.GpLuc.v1 thus offers much more possibilities for ''leaky'' Luc expression than LV.GS.PpLuc. To solve this problem, we designed LV.GS.GpLuc.v6. In this construct, the distance between the mMT1 pA and GpLuc ORF is kept very short to minimize the chance of creating transcriptional start sites in the intervening region. As an additional measure to limit leaky GpLuc expression, LV.GS.GpLuc.v6 contains a 21-bp uORF starting immediately upstream of the FRT sequence and ending with an efficient stop codon provided by the FRT sequence. Between the stop codon of the uORF and the PpLuc initiation codon only 20 nucleotides are present comprising the remainder of the FRT sequence and an optimal start site for GpLuc translation.
LV.GS.GpLuc.v6 was used to generate myoblasts GS.GLuc+ carrying the optimized GpLuc gene switch cassette. Next, the performance of the LV.GS.GpLuc.v1-and LV.GS.GpLuc.v6based cell fusion assay systems was compared in an experiment with the same setup as used for the comparison of LV.GS.GpLuc.v1 with LV.GS.PpLuc except for the omission of the 1:1 myoblast GS.GLuc(+) :myoblast-FLPe NLS+ ratio. Luciferase activity in 0-96 h and 0-120 h culture medium of serum-deprived myoblast GS.GLuc+ monocultures was 63-fold lower than in culture medium of differentiating myoblast GS.GLuc monocultures (Fig. 12), demonstrating the effectiveness of the new gene switch design to inhibit leaky GpLuc expression. However, since the improved gene switch design also reduced FLPe-dependent signal output the fold increase in GpLuc activity during myogenic differentiation of myoblast GS.GLuc(+) :myoblast-FLPe NLS+ co-cultures was quite similar for LV.GS.GpLuc.v1 and LV.GS.GpLuc.v6 (Fig. 12B). Still, in comparison to LV.GS.GpLuc.v1 for LV.GS.GpLuc.v6 a much larger part of the increase in GpLuc activity observed in differentiating myoblast GS.GLuc(+) :myoblast-FLPe NLS+ co-cultures is attributable to cell fusion.

Discussion
Apart from being involved in the formation and maintenance of skeletal muscles, bones and the placenta, cell-to-cell fusion plays an important role in numerous other biological processes like fertilization. It has also been implicated in the initiation and progression of cancer [2] and as a driving force in evolution [21]. Moreover, cell-to-cell fusion has been of great value to establish the chromosomal location of specific genes [22], can be used to induce cellular reprogramming [23,24] and is indispensable for generating hybridomas [25]. The involvement of cell-to-cell fusion in a large variety of biological processes and its diverse biotechnological applications have prompted investigations into the mechanisms of cell fusion and the contribution of specific factors to this process. Instrumental to this research is the availability of robust assays to determine cell fusion kinetics and extent. However, most of the existing quantitative cell fusion assays do not allow consecutive analysis of the same cells/tissue. Accordingly, in this paper a new quantitative assay is presented to monitor cell-to-cell fusion. This assay is based on the activation of a latent GpLuc gene after fusion of cells containing this latent reporter gene with cells encoding a recombinase that activates the dormant GpLuc gene. The extent of cell-to-cell fusion is subsequently quantified by simply measuring the enzymatic activity of the luciferase molecules secreted by the cellular fusion products. To the best of our knowledge this is the first assay that allows quantification of cell fusion activity by medium sampling.
To validate the new cell fusion assay it was used to monitor the formation of myotubes/sacs in cultures of serum-deprived human myoblasts. In these experiments, several parameters were varied including the acceptor-to-donor cell ratio and the sample regimen(s) of the cell culture medium. In general, transgene expression increased with increasing fractions of myoblasts GS.GLuc up to the point at which the number of active/nuclear FLPe molecules became limiting (i.e. at myoblast GS.GLuc :myoblast-FLPe ratios of 90:10 for FLPe NLS2 and of 95:5 for FLPe NLS+ ; Fig. 7).
At high myoblast GS.GLuc :myoblast-FLPe ratios LV.FLPe NLS+ was slightly more effective than LV.FLPe NLS2 in activating the latent GpLuc gene most likely due to fact that under differentiation conditions myoblasts-FLPe NLS+ contain 65-fold more nuclear FLPe molecules than myoblasts-FLPe NLS2 (Fig. 5A). In contrast, at low myoblast GS.GLuc :myoblast-FLPe ratios (i.e. when FLPe is no longer limiting) LV.FLPe NLS2 consistently outperformed LV.FLPe NLS+ (Figs. 7 and 10). Collectively, these findings suggest that its NLS does not noticeably hamper the spreading of FLPe NLS+ through myofibers/sacs but that high nuclear FLPe levels may somehow limit reporter gene expression. A possible explanation for the higher GpLuc expression in differentiating cocultures containing large percentages of myoblasts-FLPe NLS2 in comparison to those with large fractions of myoblasts-FLPe NLS+ may be the more frequent occurrence of secondary recombination events in the latter co-cultures leading to the deactivation of functional GpLuc expression modules.
While monocultures of myoblasts GS.GLuc maintained in growth medium displayed very little if any leaky GpLuc expression, considerable amounts of GpLuc were produced by myoblast GS.-GLuc monocultures exposed to differentiation medium. There are several possible explanations for this finding. Firstly, growth and differentiation medium may differently affect light output e.g. by (i) causing different levels of coelenterazine ''auto-oxidation'', (ii) containing different concentrations of chemiluminescence inhibi- tors or (iii) absorbing blue light to a different extent. Possibilities (i) and (iii) can be ruled out since mixing of coelenterazine substrate solution with fresh or myoblasts-FLPe NLS+ -conditioned growth or differentiation medium produced very similar signals (data not shown). This leaves us with the possibility that transcription termination by the mMT1 pA incorporated into the gene switch constructs is not very efficient or that differentiation conditions somehow stimulate transcription initiation in the region located in between the mMT1 pA and the Luc ORFs. For LV.GS.PpLuc and LV.GS.GpLuc.v6 the resulting transcripts may not lead to substantial luciferase production due to the presence of ''decoy'' ORFs immediately upstream of the Luc initiation codons (Figs. 2 and 3). A similar favorable situation does not exist for LV.GS.GpLuc.v1, which may explain the high background signals produced by this construct under differentiation conditions. Even though the luciferase activity in culture medium of differentiating myoblast GS.GLuc+ monocultures is 63-fold lower than in culture medium of differentiating myoblast GS.GLuc monocultures LV.GS.GpLuc.v6 still gives rise to a higher background signal under differentiation conditions than LV.GS.PpLuc (compare Fig. 11 with 12). Considering that the sequences in between the mMT1 pA and the Luc start codon in LV.GS.PpLuc and LV.GS.GpLuc.v1 are nearly identical this may suggest that the GpLuc-coding sequence itself is the source of the relatively high luciferase activity detected in medium of differentiating LV.GS.GpLuc monocultures. If so, the problem could be overcome by switching to another secretory luciferase (e.g. Vargula hilgendorfii luciferase [26], Lucia luciferase (InvivoGen Europe, Toulouse, France) or secretory NanoLuc [27]). Also the fact that GpLuc is a secretory protein with a long half-life (66 days in culture medium) [28] while Ppluc has a relatively short half-life (62 hours in cells) [29] may contribute to the higher background signals associated with LV.GS.GpLuc.v1 and LV.GS.GpLuc.v6 than with LV.GS.PpLuc.
Taken together, in this paper a new assay to quantify (the progression of) cell-to-cell fusion activity is described. Due to its nondestructive nature allowing repeated sampling of the same specimen, this assay will be an attractive alternative to existing quantitative cell fusion assays based on (i) light microscopic assessment of multinucleation, (ii) fluorescence dequenching, (iii) fluorescence resonance energy transfer, (iv) biochemical complementation or (v) activation of reporter genes different from GpLuc including LacZ and PpLuc [3]. Other advantages of the LV.FLPe NLS+/2 .PurR/LV.GS.GpLuc-based cell fusion assay include the simplicity and speed of the analytical procedures and the ability to combine it with (immuno)cytology, real-time microscopy, cell function assays and other methods to study cell behavior.
The sensitivity of the current assay could be improved by changing the human glyceraldehyde 3-phosphate dehydrogenase (hGAPDH) gene promoter driving GpLuc expression for a promoter with higher activity in the cell type(s) under investigation. In addition, the sequences interspersed between the 39 long terminal repeat (LTR) and the GpLuc initiation codon of LV.GS.GpLuc.v6 may be further optimized to minimize leaky GpLuc expression.