The polypeptides of adenovirus. VII. Further studies of early polypeptides in vivo and localization of E2 and E2A to the cell plasma membrane

Requirements for the association of adenovirus type 2 E3/19K wild-type and mutant proteins with HLA antigens

The E3/19K protein of human adenovirus type 2 is a resident of the endoplasmic reticulum (ER). Immediately after synthesis, it associates with major histocompatibility complex class I antigens and prevents their intracellular transport and cell surface expression. We have generated several C-terminal deletion mutants of the E3/19K protein that are preterminated at various positions on both sides of the membrane-spanning segment of the protein. One of these mutants is terminated at the luminal side of the membrane (M310), and two are terminated in the hydrophobic segment (M374 and M392), whereas mutant M621 is terminated on the cytoplasmic side of the ER membrane. The M310, M374, and M392 mutants are soluble proteins. They do not associate with HLA antigens in transfected 293 cells, and they are, to some extent, secreted into the medium. The M621 mutant protein is integrated in the ER membrane, associates immediately after its synthesis with HLA antigens, and exits from the ER. By using either an in vitro translation system supplemented with microsomes or overexpression in insect cells, we showed that M374 and E3/19K are able to associate with HLA antigens. These results indicate that the conformation of the luminal part of the E3/19K protein is not grossly altered by the mutations. Rapid transport of the M374 mutant out of the ER and partial degradation of this protein may prevent the interaction with HLA class I antigens in transfected 293 cells.

The E3/19K protein of adenovirus type 2 binds to the domains of histocompatibility antigens required for CTL recognition

The E3/19K protein of human adenovirus type 2 binds to HLA class I antigens and blocks their terminal glycosylation and cell surface expression. The nature of this interaction is non-covalent and involves neither disulfide bridges between the two molecules nor their carbohydrates. The murine H-2 Kd antigen associates with the E3/19K protein in a similar fashion to human HLA antigens whereas the allelic product H-2 Kk does not. Hybrid genes between the Kd and Kk alleles were constructed and their products were expressed in embryonic kidney cells together with the E3/19K protein. This allowed us to identify the alpha 1 and alpha 2 domains as the essential structures of the histocompatibility antigens for binding the viral protein. Interestingly, these domains are also crucial for T cell recognition. The implications for the evolution of adenoviruses and their ability to cause persistent infections are discussed.

An adenovirus type 2 glycoprotein blocks cell surface expression of human histocompatibility class I antigens

The adenovirus type 2 encoded protein E3/19K binds to human histocompatibility class I antigens (HLA). This association occurs both in adenovirus-infected cells and in cells that have been transfected with the gene encoding the E3/19K protein. The formation of the HLA-E3/19K complex prevents the HLA antigens from being correctly processed by inhibiting their terminal glycosylation. This effect is specific for HLA antigens and does not generally involve the glycosyltransferases. Furthermore, the HLA-E3/19K association dramatically reduces the cell surface expression of the HLA antigens. This reduced level of antigens might influence the cytotoxic T cell response. Therefore, our results show a possible molecular mechanism whereby adenoviruses, and perhaps other viruses, delay or escape the cellular immune system of the host.

Homology of adenoviral E3 glycoprotein with HLA-DR heavy chain

The Mr 19,000 adenovirus glycoproteins coded by the E3 region that is expressed on the cell membrane and presumably binds to HLA class I antigen shows sequence homology to the alpha-chain of human HLA-DR and to the kappa-chain of human IgM. The homology extends to a conserved domain present in all of the major histocompatibility complex antigens, beta 2-microglobulins, and immunoglobulins. Predicted beta-sheet secondary structures are similar between the Mr 19,000 adenovirus protein and these immune system proteins. Evolutionary and functional implications of the homology are discussed.

Intracellular localization of adenovirus type 5 tumor antigens in productively infected cells

The intracellular localization of tumor antigens of human adenovirus type 5 (Ad5) during lytic infection of KB cells has been studied. The cells were pulse labeled with [35S]methionine early after infection and early proteins of 58,000 D (58K), 44K, 19K, 18.5K, and 14K detectable by immunoprecipitation with hamster antitumor serum were assayed for association with cytoplasm, nucleoplasm, chromatin, cytosol, cytoskeleton, and membranes. The 44,000 D (44K) tumor antigen encoded in early region 1A (E1A: 0-4.4%) was recovered in approximately equal amounts from cytoplasmic and nucleoplasmic fractions of pulse-labeled cells and within the cytoplasmic compartment was found in the cytosol as well as associated with the cytoskeleton. The E1B-58K (E1B: 4.5-11.2%) antigen was also found to be associated with the cytoplasmic and nucleoplasmic fractions in approximately equal amounts but unlike the E1A-44K showed no affinity for cytoskeletons. Pulse-chase and immunofluorescence experiments suggested the 58K antigen accumulated in the nucleus late in infection. The E1B-19K antigen was found almost exclusively associated with the membrane fraction of infected KB cells and was resolved in polyacrylamide gels into two related species of 18.5K and 19K. Immunofluorescence studies on the E1B 18.5-19K doublet suggested that within a population of infected HeLa cells a small minority seemed to be expressing copious amounts of stainable antigen. Cell fractionation and immunofluorescence studies showed that the E4-14K antigen was a nuclear protein and the only antigen in this study which showed a significant association with a nuclear subfraction composed almost entirely of histones. The implications of these findings for the roles of the Ad5 tumor antigens in lytic infection and transformation are discussed.

Structures of the oligosaccharides of the glycoprotein coded by early region E3 of adenovirus 2

Early region E3 of adenovirus 2 encodes a glycoprotein, E3-gp25K, that is a good model with which to study structure-function relationships in transmembrane glycoproteins. We have determined the structures of the oligosaccharides linked to E3-gp25K. The oligosaccharides were labeled with [2-(3)H]mannose in adenovirus 2-early infected KB cells for 5.5h (pulse) or for 5.5 h followed by a 3-h chase (pulse-chase). E3-gp25K was extracted and purified by chromatography on DEAE-Sephacel in 7 M urea, followed by gel filtration on a column of Bio-Gel A-1.5m in 6 M guanidine hydrochloride. An analysis of the purified protein by sodium dodecyl sulfate-polyacrylamide gel electrophoresis indicated that it was >95% pure. The oligosaccharides were isolated by pronase digestion followed by gel filtration on a column of Bio-Gel P-6, then by digestion with endo-beta-N-acetylglucosaminidase H, followed by gel filtration on Bio-Gel P-6, and finally by paper chromatography. The pulse sample contained equal amounts of Man(9)GlcNAc and Man(8)GlcNAc and small amounts of Man(7)GlcNAc and Man(6)GlcNAc. The pulse-chase sample had predominantly Man(8)GlcNAc and much less Man(9)GlcNAc, indicating that processing of the Man(9)GlcNAc to Man(8)GlcNAc had occurred during the chase period. Thus, Man(8)GlcNAc is the major oligosaccharide on mature E3-gp25K. The structures of these oligosaccharides were established by digestion with alpha-mannosidase, methylation analysis, and acetolysis. The oligosaccharides found had typical high-mannose structures that have been observed in other membrane and soluble glycoproteins, and the branching patterns and linkages of the mannose residues of Man(9)GlcNAc were identical to those of the lipid-linked Glc(3)Man(9)GlcNAc(2) donor. Thus, adenovirus 2 infection (early stages) apparently does not affect the usual cellular high-mannose glycosylation pathways, and despite being virus coded, E3-gp25K is glycosylated in the same manner as a typical mammalian cell-coded glycoprotein.

Purification and characterization of an early protein (E14K) from adenovirus type 2-infected cells

One adenovirus type 2 (Ad2) early protein, with an apparent molecular weight of 14,000 in sodium dodecyl sulfate-polyacrylamide gels (E14K), was purified to homogeneity. Purification involved fractionation of cytoplasmic extracts, precipitation at low pH, and DEAE-cellulose, phosphocellulose, and hydroxylapatite chromatography. The yield was around 12 microgram of purified protein per 10(9) HeLa cells. The two Ad2 DNA binding proteins with molecular weights of 75,000 and 45,000 (E75K and E45K) were purified by the same procedure. Tryptic peptide analyses indicated that the E14K protein is unrelated to the DNA binding proteins. The purified E14K protein has a high content of basic amino acids and a sedimentation coefficient of 5.5S in the native state, corresponding to a molecular weight of around 95,000. Pulse-chase experiments suggest that the E14K polypeptide is a primary translation product. Immunoprecipitation with a monospecific antiserum against the E14K protein revealed that it is exclusively localized in the cytoplasm of infected cells. E14K started to be synthesized at 2 hpostinfection, with a maximal rate of synthesis at 4 to 6 h postinfection. Immunoprecipitation of cell extracts from four different Ad2-transformed hamster embryo cell lines revealed that only one (Ad2HE4) of them expresses this protein. The adenovirus-simian virus 40 hybrid virus (Ad2ND1) does not express this protein, suggesting that the gene for the E14K protein is located in the part of the Ad2 genome which is deleted in this hybrid virus.

Viral DNA sequences and gene products in hamster cells transformed by adenovirus type 2

Complementary strand-specific adenovirus DNA of full length or from endonuclease BamHI fragments was used as a probe to estimate the fractional representation and abundance of viral sequences in five hamster cell lines (Ad2HE1-5) transformed with UV-inactivated adenovirus type 2. The fraction of the viral genome present in the five transformed cell lines varied from 44% in the Ad2HE5 cell line to 84% in the Ad2HE3 cell line. The number of viral DNA copies per diploid cell equivalent ranged from 1.8 in the Ad2HE1 line to 7.1 in the Ad2HE4 line. In vivo labeling with [35S]methionine followed by immunoprecipitation with an antiserum against adenovirus type 2 early proteins revealed virus-specific polypeptides with molecular weights of 42,000 to 58,000 in extracts from all five hamster cell lines. Several other early viral polypeptides were detected in some of the adenovirus type 2-transformed hamster cell lines.

DNA polymerases in adenovirus type 5-infected and uninfected KB cells. Induction of an alpha-type DNA polymerase in adenovirus type 5-infected and in fast growing cells

DNA polymerase activities in uninfected KB cells or KB cells infected with adenovirus type 5 (Ad5) were compared by chromatography on DNA-cellulose and DEAE-cellulose and by isoelectric focusing. On DNA cellulose three components were found both in infected and in uninfected cells. The major component eluted at 0.15 M NaC1 and contained DNA polymerase alpha. Two minor components were found, one which did not bind to DNA-cellulose and one which bound strongly. This latter component contained DNA polymerase beta as characterized by DEAE-cellulose chromatography and sedimentation studies. No difference in properties between uninfected or Ad5-infected KB cells was found for the beta-polymerase. DEAE-cellulose chromatography of DNA polymerase alpha revealed the presence of two activities eluting at 0.11 and 0.13 M NaC1 designates as alphaI and alphaII, respectively. In Ad5-infected cells alphaII was the major component. In uninfected, stationary cells alphaI was the major component and alphaII was only detectable as a shoulder in the elution profile. However, fast growing, uninfected cells gave a similar pattern as Ad5-infected cells. These results indicate that the observed change of the DNA polymerase pattern after infection with Ad5 is related to the level of DNA synthesis and not to the induction of a viral enzyme.

Early gene expression of adenovirus type 2: R-loop mapping of mRNA and time course of viral DNA, mRNA, and protein synthesis

Adenovirus type 2 DNA was hybridized to early mRNA isolated from the cytoplasm of infected cells prior to the initiation of viral DNA synthesis. Resulting R loops were visualized in the electron microscope, and their positions were oriented with the help of DNA fragments generated by digestion with the restriction endonuclease BamHI. Early RNA was found to map (in order of relative R-loop frequency) with midpoints near positions 0.95, 0.80, 0.03, 0.65, and 0.09 on the conventional adenovirus map. The time of appearance of individual viral mRNA's was compared to the time course of viral protein and DNA synthesis. We present a refined map of adenovirus gene functions which is based on results documented in this and the accompanying study by Meyer et al. (1977), as well as on data published by other laboratories.