The Nuclear Export Signal within the Adenovirus E4orf6 Protein Contributes to Several Steps in the Viral Life Cycle

The concept of a Rev-like nuclear export pathway influencing adenovirus late gene expression has recently been challenged by a note in the Journal of Virology entitled “Adenovirus Late Gene Expression Does Not Require a Rev-Like Nuclear RNA Export Pathway” (5). The authors summarize that they “have tested the role of NES-mediated RNA export during adenovirus infection, and find that it is not essential for the expression of adenovirus late genes.” For the reasons outlined below, this conclusion seems misleading and not fully justified by the data.

We found that the E4orf6 protein of adenovirus type 5 contains a nuclear export signal (NES), allowing the protein to shuttle between the nucleus and cytoplasm (2). More recently, we observed that a functional NES of E4orf6 is required for multiple functions during infection, such as virus production, DNA replication, accumulation and transport of viral mRNA, and late expression of viral proteins (6). Nucleo-cytoplasmic shuttling of E4orf6 was now confirmed by Eileen Bridge and coworkers (5). However, the authors did not observe a significant reduction in the late expression of viral proteins when comparing wild-type E4orf6 with a mutant lacking the NES. They conclude that the export function of E4orf6 is not critical for late gene expression. However, the experimental approach leading to this conclusion requires critical evaluation. The authors transiently transfected expression plasmids for E4orf6 and mutants, followed by superinfection with E4-deficient adenovirus. They then stained the cells with antibodies to E4orf6 and simultaneously with antibodies against late proteins. The proportion among the E4orf6-positive cells that also expressed late proteins was scored and compared for various E4orf6 mutants. This score did not significantly differ between wild-type E4orf6 and the NES mutant. The problem with this approach is that it selects for those cells that express high levels of E4orf6—at least high enough to allow detection by immunofluorescence. The experiment does not score those cells that express the protein at levels that are too low for immunodetection but high enough to support virus replication. However, such a low level of expression would be what resembles the physiological situation during infection. E4orf6 is expressed by the virus at levels that barely allow immunodetection at late times after infection and essentially preclude detection at earlier times or at low multiplicities of infection (references 1 and 4 and our unpublished observations). Therefore, Rabino et al. (5) have almost certainly observed the effects of E4orf6 and mutants at supraphysiological intracellular concentrations. Our unpublished observations show that indeed, strong overexpression of E4orf6 using newly developed transfection lipids makes the NES function unnecessary for complementation of virus growth. However, this does by no means reflect the situation of an infected cell. In the experiments described in our previous report (6), we expressed E4orf6 using electroporation, which allows a relatively even, albeit weak expression of transfected plasmids. Under these conditions, the NES strongly contributed to virus growth. Instead of counting cells that stain positive for late proteins, we have documented several hallmarks of the viral life cycle. Our results consistently show that DNA replication, virus production, and late protein synthesis (assayed by immunoprecipitation of hexon to allow overall quantitation) as well as RNA levels and distribution all depend to a large extent on the NES within E4orf6 (6).

Second, Rabino et al. (5) report that leptomycin B (LMB) fails to inhibit the synthesis of late viral proteins when added 12 h postinfection, and they conclude that adenovirus late gene expression does not require a Rev-like nuclear RNA export pathway. Again, this conclusion is not fully justified by the data. First, the authors do not provide an unequivocal test for LMB function in their assays. They report an inhibitory effect on cell growth, but this may be related to the general toxicity of LMB and does not provide a proof that nuclear export was abolished by LMB at this concentration. Indeed, the concentration used by the authors (10 nM) was previously shown to inhibit nuclear export of the Rev protein only partially but not entirely (3). Furthermore, the authors do observe a reduction of late protein synthesis when LMB was added shortly after infection. Thus, the data provided by Rabino et al. (5) merely suggest that a nuclear export function of E4orf6 may be of greater importance during the first 12 h after infection than afterward.

Taken together, these considerations and the currently available data suggest that the NES of E4orf6 significantly contributes to several steps in the viral life cycle. One of our goals is the construction of a recombinant adenovirus with a mutated NES in the E4orf6 protein. This is expected to further clarify the role of this export signal in the physiological context of an adenovirus infection.