Avian retroviral vector pFuTraP expressing ASCT2
For ectopic expression of wild-type and mutated ASCT2, we chose to use the chicken cell system, which provided several advantages. i. The absence of the ASCT2 gene in the chicken genome (16) eliminates any background of endogenous ASCT2 gene product and its interference with the ectopically expressed protein. ii. Efficient ectopic expression of ASCT2 can be driven by a versatile avian replication-defective retroviral vector, which stably integrates in the chicken cell genome. iii. The interaction of Syncytin-1 with ASCT2 can be simulated by entry of a chimeric avian leukosis-based virus (ALV) carrying the Syncytin-1 envelope. iv. ASCT2-expressing and Syncytin-1-enveloped retroviral vectors can be easily produced in previously prepared chicken packaging cells and in the DF-1 chicken cell line, respectively.
We constructed a dual-fluorescence ASCT2 expression vector, designated here as pFuTraP-hA2, that contains wild-type human ASCT2 fused to the Green Fluorescent Protein from Aequorea coerulescens (AcGFP), followed by an IRES-iRFP713 cassette (Fig. 1a). We then transduced chicken DF-1 cells with VSV-G-pseudotyped viral particles carrying the FuTraP-hA2 genome and obtained cells expressing two proteins of the FuTraP system, specifically the ASCT2-AcGFP fusion protein, and iRFP713 protein. ASCT2-AcGFP permits exact quantification of the ASCT2 protein, which can be directly localised in the cell. iRFP713 translation occurs from the same mRNA as the ASCT2-AcGFP fusion protein but is initiated at the Internal Ribosome Entry Site (IRES) sequence. The fluorescence intensity of iRFP713 thus reflects the mRNA level of the FuTraP transcript and allows us to enrich the successfully modified cells regardless of the ASCT2 expression levels. After the transduction of DF-1 cells with the wild-type FuTraP-hA2 we sorted the cell population with efficient expression of the iRFP713 fluorescent protein (see the FuTraP-hA2 dot-plot in Fig. 1d in comparison to non-modified DF-1 cells used as negative control).
Next, to explore the role of region C that has been suggested as a retroviral docking site, we designed pFuTraP-d5 to pFuTraP-d22 mutants containing progressively extended deletions of five to 22 amino acids from the ASCT2 receptor region C (Fig. 1b). The largest deletion removed the entire region C. We also constructed a pFuTraP-N212Q mutant with abolished glycosylation within region C (Fig. 1b). Similarly to the wild-type FuTraP-hA2, we transduced DF-1 cells with the individual FuTraP mutants. After cell sorting, we obtained a similar fraction of iRFP713-positive cells for all transduced FuTraP variants, showing a comparable efficiency of our technique (Fig. 1d). The mean fluorescence intensity (MFI) of iRFP713 was about two times higher in the deletion mutants in comparison to wild-type ASCT2 (Fig. 1d,e), reflecting differences in the mRNA production or stability.
The ASCT2 protein level affected by deletion of region C
The fluorescence intensity of AcGFP, which was C-terminally fused to the wild-type or mutant receptor, monitored expression of the ASCT2 protein. Since it was difficult to separate the AcGFP-positive and AcGFP-negative populations in some of the mutants (Fig. 1d), we evaluated the MFI of AcGFP for the entire cellular population (Fig. 1f). MFI of the non-modified DF-1 cell line served as the negative control. We observed that deletion of seven or more amino acids of region C led to a substantial decline of the ASCT2 protein amount when compared to the wild-type ASCT2 (Fig. 1f). The protein level of the FuTraP-d22 mutant showed no difference compared to the DF-1 negative control. The ratio of AcGFP intensity that reflected the level of the protein (Fig. 1f) to the iRFP713 intensity that reflected the mRNA level (Fig. 1e) indicated that progressive deletions of five or more amino acids of region C decreased the protein production/stability (Fig. 1g). On the other hand, according to AcGFP to iRFP713 ratio, the protein amount of FuTraP-N212Q glycosylation mutant was comparable to the wild-type, at least in the avian system (Fig. 1g). Our results revealed that region C was essential for the protein expression/stability of ASCT2.
ASCT2 cell surface localisation and receptor function
The correct display of the ASCT2 protein on the cell surface is crucial for its receptor function. To track ASCT2 localisation, we used confocal microscopy on a selected panel of ASCT2 mutants. Wild-type human ASCT2 localised preferentially to the cell surface, which was visualised after fluorescent staining of cellular membrane (Fig. 2, blue channel, middle right). Similarly, both the FuTraP-d5 deletion mutant and the FuTraP-N212Q glycosylation mutant preferentially localised to the cell surface (Fig. 2). Corresponding to the FACS results, the FuTraP-d11, FuTraP-d15, and FuTraP-d22 mutants displayed a sharp reduction in the ASCT2 protein amount, while their cellular localisation could not be determined (Fig. 2). In contrast to AcGFP, iRFP713 was distributed throughout the entire cell, including the nucleus, in all FuTraP variants (Fig. 2, red channel, left).
To further examine the ASCT2 display and its binding to Syncytin-1, we adapted immunoadhesin, a soluble fusion protein that combines the target-binding region of a ligand with the Fc region of an IgG (27–29). We prepared the soluble form of Syncytin-1 (sS1) consisting of Syncytin-1 RBD fused with the heavy chain of rabbit IgG (Fig. 3a). Selected panel of cells expressing ASCT2 variants were incubated with sS1 and the binding to the ASCT2 receptor on the cell surface was visualised by means of an Alexa Fluor 594-conjugated anti-IgG antibody. Non-modified DF-1 cells labelled with sS1 served as a negative control. Flow cytometry analysis showed a specific shift in Alexa Fluor 594 staining, demonstrating the binding of sS1 to the wild-type FuTraP-hA2 (Fig. 3b). The FuTraP-d5 deletion mutant and the FuTrap-N212Q glycosylation mutant bound sS1 even better than the wild-type. The FuTraP-d7 binding of sS1 was similar to the wild-type. The results further showed that FuTraP-d11, FuTraP-d15, and FuTraP-d22 deletions of region C led to a decreased interaction of ASCT2 mutants with the sS1, which was similar to the negative control (Fig. 3b). The results confirmed the surface localisation of the ASCT2 receptor expressed from the pFuTraP vector, hence demonstrated the functional interaction of the receptor with the soluble form of Syncytin-1. These observations further supported the decreased surface display of ASCT2 mutants missing 11 or more amino acids of region C.
ASCT2 receptor mediates entry of Syncytin-1-enveloped virus
To explore the ASCT2 capacity to mediate retroviral cellular entry, we adapted the human Syncytin-1 glycoprotein to act as a functional envelope protein of infectious avian retrovirus. We constructed the chimeric replication-competent pMCAS(Sync1-MSC16)dsRed vector, in which Syncytin-1 glycoprotein replaced the original envelope of ALV. After elimination of a cryptic splicing acceptor site and shortening the cytoplasmic domain of Syncytin-1 to 16 amino acids (10) (Fig. 4a), we obtained the infectious MCAS(Sync1-MSC16)dsRed virus produced in the supernatant of transfected DF-1 cells. The virus reached a titre of 104 IU/mL and was used to infect the cells with the multiplicity of infection 0.1 to 0.4. The Syncytin-1-enveloped virus specifically entered cells expressing the wild-type FuTraP-hA2 as we identified by dsRed fluorescence detected microscopically or by flow cytometry (14% dsRed-positive cells, Additional File 1: Fig. S1). Based on the virus titre and percentage of dsRed-positive cells, we calculated the cellular sensitivity to viral infection, which corrects for the possible simultaneous co-infections of the same cell (Fig. 4b). Importantly, no viral infection of non-modified DF-1 chicken cells was detected (Fig. 4b, Additional File 1: Fig. S1). It is of note that the infectivity of the replication-competent retrovirus with Syncytin-1 envelope validates the supposed original function of Syncytin-1 as an envelope of ancestral exogenous retrovirus.
To determine whether deletion mutants can serve as a receptor for the Syncytin-1-enveloped virus, we infected DF-1 cells carrying ASCT2 variants with the MCAS(Sync1-MSC16)dsRed virus. Both deletion mutants FuTraP-d5 and FuTraP-d7 and glycosylation mutant FuTraP-N212Q conferred higher sensitivity to infection in comparison to the ASCT2 wild-type FuTraP-hA2 (Fig. 4b). On the other hand, cells carrying FuTraP-d11 and larger deletions displayed reduced sensitivity to infection compared to cells carrying the wild-type FuTraP-hA2 (Fig. 4b). However, all deletion mutants, including those with a deletion of the entire region C, displayed some level of susceptibility to infection when compared to non-modified DF-1 chicken cells (Fig. 4b).
Because the sensitivity of cells modified with ASCT2 variant may depend on the amount of the receptor, we normalised the cellular sensitivity to infection (calculated from the fraction of dsRed-positive cells, Fig. 4b) to the receptor protein level (AcGFP MFI of non-infected cells measured in the same experiment). The normalised FuTraP-d5 receptor sensitivity was slightly, although significantly lower than the wild-type (Fig. 4c). Interestingly, deletion of seven to 17 amino acids of region C increased the normalised receptor sensitivity to Syncytin-1 in comparison to the wild-type (Fig. 4c). Deletion mutants FuTraP-d19 and FuTraP-d22 normalised to the ASCT2 protein level were still significantly more susceptible to the Syncytin-1-enveloped virus than the non-modified DF-1 cells (Fig. 4c). After normalisation to the ASCT2 protein level, the glycosylation mutant revealed a similar sensitivity to Syncytin-1 as observed in the wild-type (Fig. 4c). These results demonstrate that the deletion of the putative retrovirus docking site did not abolish the ASCT2 receptor interaction with the Syncytin-1 envelope.
Cell-cell fusion triggered by Syncytin-1
Finally, we focused on the cell-cell fusion triggered by the ASCT2-Syncytin-1 interaction. We constructed the non-infectious retroviral pMCAS(3Flag-Sync1-MS)dsRed expression vector. In this construct, the original ALV retroviral envelope was replaced by the Syncytin-1 glycoprotein, which was fused N-terminally to the three-Flag epitope, had mutated the cryptic splicing acceptor site and contained the entire open reading frame (Fig. 5a). After transfection of ASCT2-expressing cells with the pMCAS(3Flag-Sync1-MS)dsRed no infectious viral particles were produced, but importantly, we observed cell-cell fusion that further supported the correct ASCT2 surface localisation and receptor function (Additional File 2: Fig. S2, Additonal File 4: Movie).
To quantify cell-cell fusion, we utilised the NanoLuc Binary Technology (NanoBiT), where the High-Affinity NanoBiT (HiBiT) subunit spontaneously complements the Large NanoBiT (LgBiT) subunit to form a functional NanoLuc luciferase enzyme (Fig. 5b).
We engineered DF-1 cells that stably expressed either LgBiT or HiBiT and modified them to express ASCT2 variants or fusogenic Syncytin-1, respectively. FuTraP-hA2 and FuTraP-d5, FuTraP-d11, FuTraP-d15, FuTraP-d22 were transduced into DF-1/LgBiT cells and the iRFP713-positive populations were separated by cell sorting. The pMCAS(3Flag-Sync1-MS)dsRed vector was transfected into DF-1/HiBiT cells, and a transfection efficiency of 45% was determined according to dsRed fluorescence (Additional File 3: Fig. S3). Anti-Flag cell labelling showed that 40% of the transfected cells expressed Syncytin-1 on the cell surface (Additional File 3: Fig. S3).
Finally, DF-1/LgBiT cells expressing FuTraP variants were seeded together with DF-1/HiBiT cells that had been transfected with Syncytin-1, and cell-cell fusion was quantified as NanoLuc luciferase luminescence (Fig. 5c). DF-1/LgBiT cells without ASCT2 mixed with DF-1/HiBiT-Syncytin-1 cells were used as a negative control. Our results revealed a similar intensity of cell-cell fusion triggered by the wild-type FuTraP-hA2 and FuTraP-d5 mutants (Fig. 5c). In contrast, increasing the extent of deletion within region C led to a decrease in the fusion ability of FuTraP-d11, FuTraP-d15, and FuTraP-d22 mutants. Nevertheless, the fusion activity of the mutants was higher than that of the negative control (Fig. 5c). Our results further confirmed that the interaction of Syncytin-1 with the ASCT2 expressed from the dual-reporter FuTraP system triggered cell-cell fusion. The luciferase fusion assay corroborated our conclusions that progressive deletions of region C reduced ASCT2 cell surface expression but did not disrupt its interaction with the Syncytin-1 envelope glycoprotein.