Hepatitis B virus (HBV) codes for three forms of surface protein. The minor, large form is translated from transcripts specified by the preS1 promoter, while the middle and small forms are translated from transcripts specified by the downstream S promoter. When the large surface protein is overexpressed, the secretion of both subviral and virion particles is blocked within the secretory pathway. We show here that overexpression of the large surface protein leads to up to a 10-fold activation of the S promoter but not of an unrelated promoter. Neither the middle nor the small surface protein, nor a secretable form of the large surface protein, activates the S promoter, but agents that induce endoplasmic reticulum (ER) stress have an effect similar to that of the large surface protein. The large surface protein also activates the S promoter in the context of the entire viral genome. Therefore, it appears that HBV has evolved a feedback mechanism, such that ER stress induced by accumulation of the large surface protein increases the synthesis of the middle and small surface proteins, which in combination with the large surface protein can form mixed, secretable particles. In addition, like other agents that induce ER stress, the large surface protein can activate the cellular grp78 and grp94 promoters, raising the possibility that it may alter the physiology of the host cell.

In the course of chronic infection, hepatitis B virus mutants can sometimes be found circulating in the serum as the predominant species. One class of such mutants contains in-frame deletions in the S promoter region. By transfecting hepatoma cells with wild-type or mutant viral genomic DNA, we have shown that one such mutant gives rise to extremely small amounts of S transcripts, as expected, and therefore expresses very little of the middle and small surface (viral envelope) proteins that are translated from these transcripts. In addition, this mutant gives rise to greater-than-wild-type levels of the preS1 transcripts, which are translated into the large surface protein. Because the large surface protein, unlike the other forms of surface protein, is incompetent for secretion, cells transfected with the mutant viral DNA contain large amounts of 20-nm particles within dilated perinuclear vesicles. Therefore, this and similar S promoter mutants may be one contributing factor in the pathogenesis of ground-glass cells, which are hepatocytes containing nonsecretable viral surface proteins within dilated vesicles and are commonly found during chronic hepatitis B. Interestingly, DNA-containing virion particles are secreted into the medium by cells transfected with the mutant DNA, in amounts that are slightly larger than those secreted from wild-type-transfected cells, apparently because the amount of large surface protein is insufficient to block virion secretion. This finding may explain how such mutants can become the predominant circulating species in the serum, especially if there are selection pressures against the wild-type virus.

Hepatitis B virus small surface protein is synthesized as a transmembrane protein of the rough endoplasmic reticulum (RER) and then buds into the lumen in the form of subviral particles that are secreted. The closely related large surface protein is also targeted to the RER but is retained in a pre-Golgi compartment and cannot be secreted. It has been assumed that the large surface protein remains as a transmembrane RER protein and hence cannot form particles, possibly because of binding to a host factor on the cytosolic face of the RER membranes. We have reexamined this question and found the following results. (i) The retained large surface protein is associated not with RER but, rather, with a more distal compartment. (ii) Electron microscopy reveals intravesicular 20-nm particles, similar to those formed by the small surface protein. (iii) The large surface protein colocalizes with and binds to calnexin, an ER chaperone protein. Therefore, our results indicate that the large surface protein is capable of budding and forming particles, and hence its intracellular retention cannot be attributed to a cytosolic factor. We interpret the data as evidence that the large surface protein is retained by virtue of interacting with calnexin, a component of what is considered the quality control mechanism of the ER.

Texas were compared serotypically and electropherotypically. All isolates were determined to be serotype 3 by reaction with hyperimmune antiserum to the serotype 3 H-2 strain of equine rotavirus. All displayed RNA electrophoretic migration patterns related to that of the H-2 strain but distinct from that of serotype 5 strain H-1. A serologic survey of 184 mares in Kentucky, which was done to determine the incidence of H-1 and H-2 infections, showed geometric mean serum neutralizing titers to the H-2 strain of equine rotavirus to be significantly higher than those to the H-1 strain. These data suggest that the serotype 3 H-2 strain is the dominant equine rotavirus in Kentucky and perhaps elsewhere in the United States. We were unable to produce confirmational evidence that the H-1 strain occurs as a natural infection in the United States.

In a previous study (S. Zheng, G. N. Woode, D. R. Melendy, and R. F. Ramig, J. Clin. Microbiol. 27:1939-1945, 1989), it was predicted that the VP7 serotype 6 bovine rotavirus strains NCDV and B641 do not share antigenically similar VP4s. In this study, gene 4 and the VP7 gene of B641 were sequenced, and the amino acid sequences were deduced and compared with those of NCDV and bovine rotavirus strain UK. Amino acid sequence homology in VP7 between the three strains was greater than 94%, confirming their relationship as VP7 serotype 6 viruses. VP4 of B641 showed amino acid homology to UK of 94% but only 73% homology to NCDV. Sequence comparison of a variable region of VP8 demonstrated amino acid homology of 53% between B641 and NCDV, whereas B641 and UK were 89% homologous in this region. These results confirm the earlier prediction that although the same serotype by VP7 reactivity, B641 and NCDV represent different VP4 serotypes. This difference in VP4 may have contributed to the lack of homotypic protection observed in calves, implicating VP4 as an important antigen in the active immune response to rotavirus infection in bovines.