Varicella-zoster virus (VZV) is enveloped in the trans-Golgi network (TGN).Here we report that glycoprotein I (gI) is required within the TGN for VZV envelopment. Enveloping membranous TGN cisternae were microscopically identified in cells infected with intact VZV. These sacs curved around, and ultimately enclosed, nucleocapsids. Tegument coated the concave face of these sacs, which formed the viral envelope, but the convex surface was tegument-free. TGN cisternae of cells infected with VZV mutants lacking gI (gI ⌬ ) or its C (gI ⌬C )-or N-terminal (gI ⌬N )-terminal domains were uniformly tegument coated and adhered to one another, forming bizarre membranous stacks. Viral envelopment was compromised, and no virions were delivered to post-Golgi structures. The TGN was not gI-immunoreactive in cells infected with the gI ⌬ or gI ⌬N mutants, but it was in cells infected with gI ⌬C (because the ectodomains of gI and gE interact). The presence in the TGN of gI lacking a C-terminal domain, therefore, was not sufficient to maintain enveloping cisternae. In cells infected with intact VZV or with gI ⌬ , gI ⌬N , or gI ⌬C mutants, ORF10p immunoreactivity was concentrated on the cytosolic face of TGN membranes, suggesting that it interacts with the cytosolic domains of glycoproteins. Because of the gE-gI interaction, cotransfected cells that expressed gE or gI were able to target truncated forms of the other to the TGN. Our data suggest that the C-terminal domain of gI is required to segregate viral and cellular proteins in enveloping TGN cisternae.

The trans-Golgi network (TGN) is putatively the site where varicella-zoster virus is enveloped. gE is targeted to the TGN by selective retrieval from the plasmalemma in response to signaling sequences in its endodomain. gI lacks these sequences but forms a complex with gE. We now find that gI is targeted to the TGN and plasma membrane when expressed in Cos-7 cells; nevertheless, surface labeling revealed that gI is not retrieved from the plasma membrane. TGN targeting of gI depended on the T 338 of its endodomain and was lost when T 338 was deleted or mutated to A, S, or D. The endodomain of gI was sufficient, if it contained T 338 , to target a fusion protein containing the ectodomain of the human interleukin-2 receptor to the TGN. A truncated protein consisting only of the gI ectodomain was secreted and taken up by nontransfected cells. This uptake of the secreted gI ectodomain was blocked by mannose 6-phosphate. Following cotransfection, both gI and gE were retrieved to the TGN from the plasma membrane in 26.7% of cells, neither gI nor gE was internalized in 18.3%, and gE was retrieved to the TGN while gI remained at the plasma membrane in 55%. We suggest that the T 338 of its endodomain is necessary to retain gI in the TGN; moreover, because gI and gE interact, the signaling sequences of each glycoprotein reinforce one another in ensuring that both glycoproteins are concentrated in the TGN yet remain on the cell surface.

Retinoic acid inhibits transformation of cells by polyoma virus middle T oncoprotein. Inhibition of transformation results from a retinoic acid-dependent failure of cells to fully express the c-fos proto-oncogene. Retinoic acid prevents transactivation of the c-fos promoter by disrupting signaling between tyrosine kinases at the plasma membrane and trans-acting factors at the c-fos promoter. We used complementary genetic, biochemical and molecular approaches to demonstrate that: (1) phosphatidylinositol 3-kinase signaling is the principle mechanism of polyoma virus middle T oncoprotein activation of c-fos expression; (2) middle T/phosphatidylinositol 3-kinase transactivation of the c-fos promoter and transformation of cells requires activation of both the small GTP-binding protein Rac and Jun N-terminal kinase; (3) retinoic acid inhibits activation of Jun N-terminal kinase, thereby preventing c-fos transactivation and transformation; and (4) middle T activation of c-fos transcription requires both the serum response element and the promoter proximal cyclic AMP response element. These studies identify a novel target through which retinoids prevent oncogenic transformation.