The type I keratin 19 is unusual in its tissue distribution in that under normal circumstances it does not seem to be restricted, as the other keratins are, to expression in either stratified or simple epithelia. In addition to the previously reported distribution of keratin 19 in human tissues, we have observed keratin 19 in epidermal basal cells, in a defined region of the hair follicle, and in nipple epidermis. We noticed that expression of keratin 19 appears to be especially characteristic of regions of labile or variable cellular differentiation as indicated by the presence of multiple keratin phenotypes in close proximity to each other. Using a monoclonal antibody recognizing keratin 19 (LP2K) to screen a human placenta cDNA expression library, we have isolated, cloned, and sequenced cDNA coding for full-length human keratin 19, as confirmed by its reactivity with several other known anti-keratin 19 monoclonal antibodies and by the near identity of its sequence with that of the bovine keratin 19 homologue. This similarity extends to both proteins being truncated at the C-terminal end to only 13 amino acids beyond the rod domain. Although the amino acid homology over the N-terminal and helical rod domains is particularly high, the human and bovine proteins diverge substantially over the short C-terminal domain, which suggests that this region has no conserved function. Comparison with other type I keratins indicates that the closest evolutionary neighbors of keratin 19 are keratinocyte keratins, probably 13 and 14, and not the simple epithelial keratin 18. Assessing the histochemistry and sequence data together, we propose that the cell may use this apparently deficient keratin as a "neutral" keratin. While unimpaired in its ability to polymerize (keeping the cell integrated into the epithelial sheet via filament-desmosome networks), keratin 19 expression does not irrevocably commit a cell to any one of the local differentiation options. Such predicted differentiational flexibility may also imply vulnerability to transformation.

Very little is known about the human cytomegalovirus functions that activate gamma (late) gene expression. We have investigated the regulation of the human cytomegalovirus gamma gene encoding the ICP36 major late DNA-binding protein family (UL44). Transactivation of the ICP36 gene promoter was found to be absolutely dependent on the trs1 gene product when expressed in cells in conjunction with ie1 and ie2 gene products. Transactivation occurred poorly or not at all when any one of these three transactivators was omitted. TRS1 is a member of the US22 family of proteins and is encoded by a region near the L-S junction of the viral genome within the c repeat and adjacent Us sequences. TRS1 is highly homologous to IRS1, which is encoded from the other copy of the c repeat, and plasmid constructs carrying the irs1 gene were also able to mediate transactivation of the ICP36 promoter. RNA blot analysis of steady-rate RNA throughout infection showed that the trs1 transcript was expressed with the kinetics of an alpha gene but its accumulation was delayed relative to that of ie1 and ie2 transcripts. On the basis of these experiments, TRS1 and IRS1 are proposed to be important intermediaries in the cascade of cytomegalovirus gene expression.

No abstract

No abstract

To generate tools for monitoring processing and folding in keratin intermediate filaments, a group of monoclonal antibodies reacting with the intermediate filament protein keratin 19 were studied using different approaches to define the structure and localization of their epitopes. The binding pattern to bacterially expressed human keratin 19 fragments allowed the definition of minimal amino acid sequences required for antibody binding. The screening of overlapping 15-residue peptides confirmed and further specified the epitope locations for a subset of the tested antibodies. In addition, the epitope of an antibody with apparent species-restricted specificity (LE64) was revealed by isolating and characterizing a full-length keratin 19 clone from a PtK2 cDNA library. Taken together with species cross-reactivity of individual antibodies and sequence information obtained by probing a phage display library, specific amino acid residues could be highlighted as likely to be involved in the antibody binding.