Illuminating the Sites of Enterovirus Replication in Living Cells by Using a Split-GFP-Tagged Viral Protein

van der Schaar, H. M.; Melia, C. E.; van Bruggen, J. A. C.; Strating, J. R. P. M.; van Geenen, M. E. D.; Koster, A. J.; Bárcena, M.; van Kuppeveld, F. J. M.

Having found that the 3A protein was not functionally hampered by the introduction of the GFP(S11) tag, we generated a recombinant CVB3 that encodes 3A(S11aa2). Infections of this virus in BGM cells stably expressing GFP(S1-10) yielded discrete fluorescent signals. However, confirming that the fluorescent foci localize to virusinduced structures demands the higher-resolution images of subcellular structures, in the context of their surroundings, that EM can provide. Establishing this link with ultrastructure is particularly interesting in the case of enteroviruses, since their ROs can adopt various morphologies during infection (i.e., tubules, DMVs, and multilamellar structures) which, as observed by EM, often coexist. In this study, CLEM was employed to confirm that the GFP fluorescent foci localize to ROs and to establish which subset of enterovirus RO morphologies this fluorescence corresponded to. CLEM revealed that the GFP fluorescence was present at bona fide ROs, which took the form of both tubular structures and DMVs resembling those observed previously with wt CVB3 in Vero cells (18) . Split-GFP CLEM thus allows the unambiguous identification of RO morphologies underlying the 3A signal and opens up new possibilities for a better understanding of the requirements for their development. This method also circumvents limitations inherent in other, similar approaches. For instance, while immunogold labeling of proteins is a popular technique that couples an electron-dense gold particle to the protein of interest, allowing it to be visualized by EM, its success depends largely upon the antibody used and the resilience of epitopes during EM sample preparation. In our experience, sample preparation procedures for immunolabeling of CVB3 3A that retain RO membranes unfortunately do so at the expense of viral epitope integrity, resulting in insufficient labeling of 3A (unpublished data).