Cytomegalovirus is transmitted with blood and organs from seropositive individuals, although the particular leukocyte population harboring latent or persistent virus remains poorly characterized. Murine cytomegalovirus, tagged with the Escherichia coli lacZ gene, was used to identify cells in which virus replicates during acute infection of immunocompetent mice. Recombinant murine cytomegaloviruses, RM461, RM460, and RM427, were constructed to express ,1-galactosidase under control of the human cytomegalovirus iellie2 promoter/ enhancer. The lacZ gene was inserted between the ie2 and sggl genes in RM461 and RM460, disrupting a 0.85-kb late transcript that was found to be dispensable for replication in cultured cells as well as for infection of mice. In BALB/c mice, lacZ-tagged and wild-type viruses exhibited a similar 50%o lethal dose and all had the capacity to latently infect the spleen. Peripheral blood mononuclear phagocytes were the major infected leukocyte cell type, as demonstrated by the ability of infected cells to adhere to glass and to phagocytize latex beads; however, these cells did not exhibit typical monocyte markers. Plaque assay for virus and 5-bromo-4chloro-3-indolyl-,1-D-galactopyranoside (X-Gal) staining of frozen sections of organs from infected mice revealed that the major target organs included the spleen, adrenal glands, liver, and salivary glands, although tissues as diverse as brown fat and lungs were also involved. Individual blue-staining cells were readily identified in all infected tissues. These studies identified a mononuclear phagocyte, possibly a macrophage or dendritic cell precursor, as the vehicle of virus dissemination during acute infection, and demonstrate the

Natural killer (NK) cells are an important early mediator of host immunity to murine cytomegalovirus (MCMV) infection. However, MCMV has evolved mechanisms to elude recognition and clearance by NK cells. We have identified an MCMV immune evasion protein that impairs NKG2D-mediated NK cell antiviral activity. Infection of BALB/c 3T3 cells with the Smith strain of MCMV resulted in strong down-regulation of H60, a high affinity ligand for NKG2D, from the surface of virus-infected cells. The MCMV m155 protein specifically down-regulated H60 without affecting expression of the other known NKG2D ligands, RAE-1 and MULT-1. Treatment with the proteasome inhibitors lactacystin or epoxomicin reversed m155 down-regulation of H60. An MCMV mutant virus lacking m155 was severely attenuated in BALB/c mice; however, treatment with neutralizing anti-NKG2D monoclonal antibody or with NK-depleting anti-asialo GM1 antisera restored virulence of the mutant virus. Thus, down-regulation of H60 by m155 is a powerful mechanism of inhibiting NKG2D-mediated antiviral function.

A transposon derived from Escherichia coli Tn3 was introduced into the genome of murine cytomegalovirus (MCMV) to generate a pool of viral mutants. We analyzed three of the constructed recombinant viruses that contained the transposon within the M25, M27, and m155 open reading frames. Our studies provide the first direct evidence to suggest that M25 and M27 are not essential for viral replication in mouse NIH 3T3 cells. Studies in cultured cells and Balb/c mice indicated that the transposon insertion is stable during viral propagation both in vitro and in vivo. Moreover the virus that contained the insertion mutation in M25 exhibited a titer similar to that of the wild-type virus in the salivary glands, lungs, livers, spleens, and kidneys of the Balb/c mice that were intraperitoneally infected with these viruses. These results suggest that M25 is dispensable for viral growth in these organs and the presence of the transposon sequence in the viral genome does not significantly affect viral replication in vivo. The Tn3-based system can be used as a mutagenesis approach for studying the function of MCMV genes in both tissue culture and in animals.

The murine cytomegalovirus alpha (immediate-early) gene product, IE2(391aa), a protein that is related to the human cytomegalovirus US22 protein family, had previously been shown to be dispensable for viral growth in cell culture. In transient assays, however, this protein was found to transactivate the murine CMV ie1/ie3 and ie2 promoters, as well as a number of other promoters. Transactivation was mediated via promoter-proximal elements rather than through elements located upstream in the enhancer region. This activation predicted that ie2 would play a role in regulating gene expression; however, ie2 mutants did not exhibit altered growth or latency in the mouse. ie2-deficient viruses reached peak titers in spleen, salivary glands, lungs, liver, kidneys, pancreas, peripheral blood leukocytes, and adrenal glands that were comparable to wild-type virus. When assayed by spleen explant culture, ie2-deficient viruses yielded reactivation levels similar to wild type. Thus, the murine CMV ie2 gene encodes a regulatory protein that is dispensable for viral infection of cells in culture as well as for interaction with tissues in the infected BALB/c mouse.

Murine cytomegalovirus carrying a deletion mutation disrupting the expression of a gene dispensable for growth in cultured cells was found to disseminate poorly in the mouse. The mutation resulted in a dramatic decrease in the expression of a 1.5-kb major and a 1.8-kb minor 1a transcript from a region adjacent to the ie2 gene in the viral genome. Nucleotide sequence determination indicated that 323 bp, including a predicted polyadenylation signal, was deleted from this 1a gene. In cultured cells, the plaque morphology and growth characteristics of the mutant were similar to those of parental or rescued wild-type viruses. Following intraperitoneal inoculation of BALB/c mice, growth of the mutant in the salivary gland was dramatically reduced 10,000-fold, while growth in the liver and spleen was not dramatically affected. The 13 gene was thus denoted sggl (salivary gland growth gene 1). Neither intranasal infection nor direct inoculation into the salivary glands completely overcame the restriction of growth in this organ, suggesting that the sggl gene encoded a determinant of tissue tropism. To investigate the impact of the sggl mutation on virus dissemination via the blood, the virus titer in peripheral blood leukocytes was determined. No difference was found between the sggl mutant and rescued wild-type virus. Thus, murine cytomegalovirus sggl gene products appear to be involved in entry or replication of virus in salivary gland cells.