HSV has studied epithelial and neuronal cell junctions for many millions of years in order to spread through mucosal, ocular and neuronal tissues. These viruses move swiftly and efficiently between infected and uninfected cells by cannibalizing the cell’s intracellular sorting machinery to direct viral progeny to specific cell surfaces i.e. epithelial and synaptic junctions (reviewed in Johnson and Huber 2002). For example, several viral membrane proteins, e.g. gE/gI and gB, bind clathrin adaptor proteins, acidic cluster binding proteins PACS-1 and PACS-2, as well as other components of the trans-Golgi network (TGN)/endosomal sorting machinery. These interactions determine movement of progeny viruses (produced in the nucleus) to cell junctions. We have shown that nascent HSV particles are specifically directed to epithelial cell junctions and are not found on apical surfaces but instead accumulate in the spaces between cultured epithelial cells (see figure one and Johnson et al. 2001). We are attempting to define intracellular events and cellular components that participate in this process. A model explaining how sorting in the TGN can promote cell-to-cell spread of HSV is shown in figure two. HSV glycoproteins, gE and gI form a heterodimer known as gE/gI and this complex localizes to the TGN by virtue of numerous sorting domains in the cytoplasmic domains of these proteins.

Sorting of HSV proteins and directed transport of virus particles also occurs in neurons. Virus infecting neurons at axon tips move to the nucleus along microtubules. Later (often after latent virus reactivates) viral progeny are specifically delivered down axons via other microtubule motors, and appear to be sorted away from dendrites. We are studying HSV transport within sensory neurons following infection in the cornea. Moreover, we can directly inject HSV into the trigeminal (TG) ganglia and characterize virus movement back to the cornea (figure three A). In many cases, we track viruses by using recombinants expressing GFP or fusion proteins in which GFP or RFP is fused to a viral membrane protein (figure three B: See green axons of neurons infected in the TG ganglia and projecting into the cornea).

US9 apparently contains sequences that couple HSV structural proteins onto microtubule motors causing these viral building blocks to move down axons. Exiting cells, HSV assembles, not in the cell body, but at axon tips. We are also using the US9- mutant and other mutants to study HSV keratitis. One particularly interesting question relates to the origin of HSV antigens that cause this inflammatory disease in the cornea. There is clearly persistent HSV in the cornea and this could trigger immune responses that ultimately damage the cornea. Alternatively, infectious HSV that reactivates in latently-infected neurons could reinfect the cornea producing immunity that causes corneal scarring. Since the US9- mutant cannot return to the cornea from infected neurons, we can test which of these two possibilities is correct using mouse models.

The second major focus of our research involves recognition of viruses by the host’s immune system and evasion of this immunity. In order to persist or remain latent for decades, herpesviruses have evolved many carefully crafted mechanisms to modulate immune responses or avoid recognition. Some years ago, we identified the first herpesvirus protein that blocks detection of infected cells by CD8+ T lymphocytes. HSV ICP47 inhibits TAP, the Transporter associated with Antigen Presentation, a cellular protein that pumps antigenic peptides into the ER where they are loaded onto MHC class I proteins (figure six). Class I proteins move from the ER through the Golgi apparatus to cell surface and present antigenic peptides to the T cell receptors of T lymphocytes. By blocking TAP, ICP47 inhibits recognition of HSV-infected cells by T cells. Since those studies, a large number of viral inhibitors of TAP and other steps in the MHC class I antigen presentation have been described. More recently, we have probed HCMV for other, novel immune evasion mechanisms. HCMV has about 2.5 times the coding capacity of HSV, so that there is the potential to express a huge number of immunomodulatory molecules. Moreover, HCMV replicates very slowly, increasing vulnerability to host immune responses. To deal with the immune system HCMV expresses an impressive array of immune evasion and immunomodulatory proteins. Moreover, HCMV infects a relatively large number of cell types in the body, including several that express MHC class II proteins. We became interested in whether HCMV might block, not only the MHC class I pathway, but also the MHC class II antigen presentation pathway. 

By reducing class I and II proteins on the cell surface, viruses expose themselves to natural killer (NK) cells that recognize cells with reduced MHC proteins. To combat NK cells, HCMV expresses an MHC class I homologue, UL18, that apparently replaces the lost surface class I. Recently, we characterized a second HCMV protein, UL16, that reduces exposure to another NK recognition receptor, NK2G-D. NK2G-D recognizes MICA and MICB which are MHC class I proteins (see figure eight). Normally, HCMV reduces surface MHC class I by inhibiting the class I pathway, but also triggers high level transcription of the MICA and MICB genes that are regulated by heat shock promoters. The appearance of MICA and MICB on the surfaces alerts NK and T cells with NK2GD receptors that HCMV is present. These lymphocytes become activated and would kill the HCMV-infected cells, unless HCMV counters this. We found that UL16 binds MICB and mislocalizes the cellular protein, reducing recognition by NK cells (Dunn et al 2003). Our continuing studies involve mapping and identification of other novel viral inhibitors of the immune system and detailed characterization of how these viral proteins function. The work involves the cell biology of antigen presentation pathways, use of recombinant virus expression vectors, and characterization of T and NK cell recognition mechanisms.