This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Tollefson, A. E.
Right arrow Articles by Wold, W. S. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Tollefson, A. E.
Right arrow Articles by Wold, W. S. M.

 Previous Article  |  Next Article 

Journal of Virology, July 2003, p. 7764-7778, Vol. 77, No. 14
0022-538X/03/$08.00+0     DOI: 10.1128/JVI.77.14.7764-7778.2003
Copyright © 2003, American Society for Microbiology. All Rights Reserved.

Mutations within the ADP (E3-11.6K) Protein Alter Processing and Localization of ADP and the Kinetics of Cell Lysis of Adenovirus-Infected Cells

Ann E. Tollefson, Abraham Scaria,{dagger} Baoling Ying, and William S. M. Wold*

Department of Molecular Microbiology and Immunology, St. Louis University Health Sciences Center, St. Louis, Missouri 63104

Received 10 July 2002/ Accepted 21 April 2003


arrow
ABSTRACT
 
ADP (also known as E3-11.6K protein) is synthesized abundantly in late adenovirus infection and is required for efficient lysis of infected cells and release of viral progeny at the end of the viral replication cycle. ADP is a type III bitopic NendoCexo nuclear membrane and Golgi glycoprotein that is produced at high levels in late adenovirus infection (>24 h postinfection). We show pulse-chase and other studies indicating that ADP undergoes a complex process of N- and O-linked glycosylation and proteolytic cleavage. In order to further characterize ADP, a series of 23 deletion and point mutations has been constructed in the adenovirus serotype 2 adp gene and then built into a wild-type adenovirus background. These mutants were analyzed for processing and intracellular localization of ADP. Mutation of the single predicted N glycosylation site eliminated N glycosylation. Deletion of a region in ADP rich in serine and threonine residues reduced O glycosylation. In general, mutations within the lumenal domain of ADP resulted in lower protein stability; immunofluorescence assays indicated that these ADPs were primarily present in the Golgi apparatus. Viruses with mutations within the cytoplasmic-nucleoplasmic domain of ADP showed normal glycosylation patterns and protein abundance for ADP, but the protein was often found throughout cellular membranes rather than being localized specifically to the nuclear membrane and Golgi apparatus. The ADP virus mutants were analyzed by cell viability assays to determine the kinetics of cell lysis following infection of human A549 cells. In general, viruses with mutations within the lumenal domain of ADP display greatly reduced efficiencies of cell lysis. Viruses with large deletions in the cytoplasmic-nucleoplasmic domain of ADP retain much of their ability to lyse infected cells.


arrow
INTRODUCTION
 
Human adenoviruses (Ads) encode several proteins that regulate cell viability (reviewed in references 35, 39, and 40). In order to ensure the survival of the cell during infection, the Ad early region 3 (E3) transcription unit encodes several "stealth" proteins that counteract host immunosurveillance (reviewed in references 14, 18, 35, 38, and 39). These proteins prevent the premature death of the infected cell; maintaining cell viability allows virus replication to occur and ensures the production of large numbers of Ad virions. However, the progression of infection also requires the release of virions from the infected cell and additional rounds of productive infection.

A specific and abundant late protein, named adenovirus death protein (ADP) and also known as E3-11.6K, has been associated with efficient release of Ad serotype 2 (30, 31). ADP is produced in small amounts from scarce E3 mRNAs (d and e) at early times postinfection (p.i.). However, ADP synthesis is ~400 times greater in late stages of infection (29 to 32 h p.i.) (32). During late stages of infection, ADP is synthesized from mRNAs that contain the Ad major late tripartite leader; thus, ADP should be considered a true late protein (32). However, it is not a structural protein of the virion (32). ADP is the only E3 protein that is predominantly a late protein.

Evidence for the function of ADP in virus release was initially suggested by the formation of small plaques by viruses in which this gene was deleted. When viruses in which the genes for the E3 proteins are individually deleted were compared, only deletion of adp resulted in small plaques that were slow to develop (30). While normal amounts of virus progeny were found intracellularly, much less virus was found extracellularly (31). Data indicated that ADP is required for efficient cell lysis of Ad-infected cells, and it was proposed that this process mediates the release of virus progeny (30, 31).

While cells infected with wild-type (WT) Ads begin to lose viability by 2 to 3 days p.i., cells infected by viruses in which adp is deleted do not begin to lyse until 6 to 7 days p.i. (30, 31). Both WT and ADP- viruses cause host protein synthesis shutoff, but synthesis of viral proteins continues for an extended time in cells infected with adp mutants. WT Ad infection results in degradation of cellular DNA and RNA; this is not seen for infections with adp mutant viruses (31). Cells infected with ADP- viruses exhibit typical Ad cytopathic effect in that the cells round up and detach from the surface of the tissue culture dish. However, the ADP- virus-infected cells remain intact and metabolically active for a greatly extended time. Electron microscopy indicates that the nuclei of the infected cells become very swollen with large numbers of progeny virus particles. However, the plasma membrane and nuclear membrane remain intact; the cells exhibit much less vacuolarization than cells infected with WT Ad (30). Cells infected with WT Ad do not show morphological features of apoptosis, and the DNA degradation does not result in formation of "DNA ladders" (31). ADP mediates loss of viability in a variety of human tumor cell lines infected with Ads (30).

ADP was initially identified as a 13,000-molecular-weight (13K)-14K doublet in immunoprecipitations of radiolabeled early proteins from Ad type 2 (Ad2)-infected cells (36). Subsequently, immunoprecipitation and analysis by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) of ADP from late infection revealed multiple diffuse bands of ~13 to 14K, ~19K, and ~27 to 31K (32). Immunoblots indicated that these various bands all represent forms of ADP (32). The Ad5 ADP (93 residues) showed band patterns similar to those of the Ad2 protein (101 residues), although the bands had higher mobility, as would be expected for a protein with eight fewer residues (32).

The three major sets of ADP bands (~13 to 14K, ~19K, and ~27 to 31K) were shown to be the products of posttranslational modification. The Ad2 ADP is Asn glycosylated with high-mannose sugars cotranslationally and becomes resistant to endoglycosidase H treatment, indicating further processing to complex oligosaccharides (25), a process occurring in the medial and trans compartments of the Golgi apparatus. The bands with the lowest mobilities on SDS-PAGE (~27 to 31K) were labeled with mannose and glucosamine (25). The Ad2 ADP is a type III bitopic integral membrane glycoprotein. Indirect immunofluorescence assay indicates that over the course of infection the abundance of ADP increases in the Golgi apparatus and endoplasmic reticulum (ER), but ultimately, ADP accumulates in the nuclear membrane late in infection (>30 h p.i.) (25). The Ad2 ADP has a hydrophobic sequence between residues 41 and 62 that is the putative signal-anchor sequence; Triton X-114 extraction confirmed that ADP is an integral membrane protein (25). The signal-anchor sequence and the glycosylation features are conserved between Ad2 and Ad5 (25). The Ad5 and Ad2 ADPs have been reported to be palmitoylated (12).

For Ads, only the E3 region encodes integral membrane proteins. Work with virus mutants has revealed that localization of these proteins to specific cellular membranes is required for function. We would expect that ADP also has sorting signals for localization to its functional site(s). Characterization of mutants should provide insight into the sorting and function of ADP. In this study, ADP processing and subcellular localization were evaluated in WT Ad infections and in infections with viruses containing ADP mutations. These extensive structure-function studies have identified specific domains for protein processing, exit from the Golgi apparatus, and subcellular localization. In addition, analysis and quantitation of loss of cell viability (cell death) after infection with the ADP mutants have been used to determine the degrees of ADP function for these ADP mutants.


arrow
MATERIALS AND METHODS
 
Cell lines. Suspension cultures of human KB cells were grown in Joklik modified minimal essential medium (J-MEM; Gibco BRL) with 5% equine serum (HyClone). Human A549 cells were grown in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum (FBS; HyClone) at 37°C with 6% CO2. Chinese hamster ovary (CHO) cells and ldlD-14 cells, a mutant CHO cell line defective in O glycosylation (15) (obtained from M. Krieger, Massachusetts Institute of Technology, Cambridge, Mass.), were grown in Ham's F-12 medium (Gibco BRL) containing 7.5% FBS.

Viruses. Ad2 was received from Maurice Green (St. Louis University). rec700 is the WT parental virus for the viruses (with 700 numbers) used in this study. rec700 is an Ad5-Ad2-Ad5 recombinant virus consisting of the Ad5 EcoRI A (map positions 0 to 76), Ad2 EcoRI D (map positions 76 to 83), and Ad5 EcoRI B (map positions 83 to 100) restriction fragments (37). rec700 has the Ad2 version of ADP, a protein predicted from its DNA sequence to consist of 101 amino acids (aa). Virus mutants were generated by oligonucleotide mutagenesis as described previously (26). In the following list of mutants, the Greek symbol {Delta} refers to deletion, and the numbers indicate the amino acids that are deleted, e.g., {Delta}4-11 indicates that aa 4 to 11 are missing from the protein. The mutants included in this study have been represented schematically (see Fig. 3, which also summarizes their phenotypes with regard to ADP processing and induction of cell lysis).



View larger version (50K):
[in this window]
[in a new window]
  		FIG. 3. Schematic of viruses with ADP mutations. A panel of viable virus mutants with ADP deletions and point mutations was generated to evaluate the processing, localization, and function of ADP. For WT ADP (rec700), the proposed major sites for O glycosylation (five serines plus threonines) and the single N glycosylation site are shown, as are the signal-anchor and basic proline (BP) domains. The mutations present in the various viruses are indicated by the schematic. The numbers refer to the amino acid order in the ADP protein. Letters followed by a number indicate that the amino acid is mutated, e.g., L96L97L98 indicates that the Leu residues at positions 96, 97, and 98 in ADP are mutated (to Ala residues). The three columns at the right summarize the overall phenotypes of the mutants with respect to the processing of ADP, as well as the effects that the mutations have on the ability of ADP to promote cell lysis. Estimates of relative cell lysis as judged from plaque development (Fig. 7), trypan blue exclusion (Fig. 8), and lactate dehydrogenase release (not shown) data are indicated as +++++ for rec700 (WT) to - for adp-null mutants. G, Golgi apparatus; nm or NM, nuclear membrane; er or ER, endoplasmic reticulum; memb., membranes. Boldface uppercase italics indicate the primary site of localization. Lowercase letters indicate minor sites of localization.

The mutants analyzed are dl735 ({Delta}4-11), dl736 (ADP frameshift mutation; the frame is shifted after the initial 32 bp of the adp sequence), dl737 ({Delta}29-45), dl738 ({Delta}46-60), dl732 (an upstream mutation which increases somewhat the amount of ADP synthesized) (26), dl714 ({Delta}71-94), dl716 ({Delta}79-101), dl717 ({Delta}81-88) (1), dl712 (deletion of the entire adp gene) (3), dl715 ({Delta}76-89), dl735.1 ({Delta}18-22), dl736.1 ({Delta}11-26), dl734.7 (Asn14 replaced with Ser, destroying the putative N glycosylation site), pm734 ({Delta}1-40; initiating at Met41), pm734.1 ({Delta}1-48; ADP which has Met1 and Met41 in ADP mutated to Leu so that ADP would initiate at Met49), pm734.2 (Met41 to Leu), pm734.3 (Met49 to Leu), pm734.4 (Met56 to Leu and Cys52 to Arg), pm734.5 (Met1 and Met49 to Leu; ADP will initiate at Met41), pm734.6 (Met1 and Met56 to Leu and Cys52 to Arg; ADP will initiate at Met41), and dl715.1 ({Delta}63-70). The oligonucleotide used to generate the deletion in the dl735.1 adp gene was 5'-CGC AAC ACC ACT GCT GCC CTA AAT TTA CCC-3'. The oligonucleotide used to generate the deletion in the dl736.1 adp gene was 5'-CGC CCA CAA CGG ATC CCC AAG TTC ATG C-3'. The Asn14 mutation to Ser14 was constructed by using an oligonucleotide in which the AAC codon was replaced by TCC. For the Met-to-Leu mutations, oligonucleotides were used in which the relevant ATG was replaced by TTG. The initial round of mutagenesis replaced each individual Met residue in a full-length background to generate pm734 (Met1 to Leu), pm734.2 (Met41 to Leu), pm734.3 (Met49 to Leu), and pm734.4 (Met56 to Leu; subsequent sequencing indicated a second mutation in pm734.4 such that Cys52 was changed to Arg in addition to the Met-to-Leu mutation). A second round of mutagenesis was used to form the truncated products (pm734.1, pm734.5, and pm734.6); the oligonucleotide used to generate pm734 (Met1 to Leu) was used to create a missense mutation in place of the initial Met codon.

The constructions of pm734.8 and pm734.9 were performed as follows. The following primers were used for PCR to introduce mutations (underlined) into the coding sequences of ADP: ADP1 (5'-TGA TCA AAC CCA GCT TCA GCT TGG-3'), ADP2 (5'-CTC GAG GAA TCA TGT CTC ATT TAA TCA TAC TGA GCA GCT GCA GAA CAT GGT TTC AGA CCG-3') (for mutation of Leu96Leu97Leu98 to Ala96Ala97Ala98), ADP3 (5'-GGC AAC AAA TAA GCC AGA GAA TAA TAA GGC AAA C-3') (Met56 to Leu), and ADP4 (5'-ACT CGA GGA ATC ATG TCT CA-3'). Primers ADP1 and ADP2 were used to make the pm734.9 mutation (Leu96Leu97Leu98 to Ala96Ala97Ala98). The pm734.8 mutation (Met56 to Leu) was made in two steps. In the first round of PCR, primers ADP1 and ADP3 were used. In the second round of PCR, the first PCR product and primer ADP4 were used to generate a mutated full-length ADP. The final PCR products were cloned into the pED plasmid, which contains the Ad2 EcoRI D fragment. The mutations in the plasmid were sequenced before they were built into the viral genome. To build the mutation into the viral genome, the Ad2 EcoRI D fragment was ligated between the Ad5 EcoRI A and EcoRI B fragments. This DNA was transfected into A549 cells, and the resulting plaques were screened by restriction enzyme analysis of the viral DNA in the Hirt supernatant. The mutations were confirmed by sequencing. Virus stocks were prepared in suspension cultures of human KB cells and purified by banding in CsCl, and titers were determined on A549 cells as described previously (29).

Pulse-chase of Ad2 ADP. Suspension cultures of human KB cells were infected with Ad2 at 50 PFU per cell in serum-free J-MEM (107 cells per time point). After the 1-h adsorption period, the cells were diluted to 4 x 105 per ml and incubated with rotation for ~25 h. The cells were pelleted; rinsed two times with DMEM (Met- Cys-) containing 10 mM HEPES, pH 7.2; resuspended at 2 x 107 per ml in this medium; and incubated at 37°C for 1 h prior to being labeled. The cells were labeled with 100 µCi of Met (Tran35S; ICN) per 107 cells in Met- Cys- DMEM for 20 min in the absence of serum. The cells were then diluted in 40 ml of ice-cold J-MEM supplemented with 5x concentrations of Met and Cys and 5% equine serum. The cells were pelleted, rinsed two additional times, and resuspended in the above-mentioned medium; 107 cells were aliquoted per microcentrifuge tube. The tubes were rotated at 37°C until they were collected at the appropriate time points. At each time point, the cells were pelleted, rinsed twice with phosphate-buffered saline (PBS)-phenylmethylsulfonyl fluoride (PMSF; 1 mM), and extracted with Iso-hi-pH buffer (0.14 M NaCl, 1 mM MgCl2, 10 mM Tris-HCl, pH 8.5) containing 0.5% Nonidet P-40 (NP-40) and 1 mM PMSF; nuclei were removed by microcentrifugation.

For immunoprecipitation, half of the extract was diluted in immunoprecipitation buffer (150 mM NaCl, 5 mM EDTA, 50 mM Tris-HCl [pH 7.4], 0.005% NP-40, 0.02% [wt/vol] sodium azide, 1% bovine serum albumin [BSA]), and 5 µl of rabbit antipeptide antibody against residues 87 to 101 of the Ad2 sequence were added (32). After several hours of rotation at 4°C, protein A-Sepharose beads were added. The immunoprecipitates were washed five times with high-salt buffer (0.5 M NaCl, 1 mM EDTA, 10 mM Tris-HCl [pH 7.4], 0.5% NP-40, 1% sodium deoxycholate) and twice with 50 mM Tris, pH 6.8, and then eluted with 2x Laemmli buffer (125 mM Tris-HCl [pH 6.8], 20% glycerol, 4% SDS, 10% 2-mercaptoethanol, 0.0025% [wt/vol] bromophenol blue). The immunoprecipitates were run on 10 to 18% gradient gels by SDS-PAGE with 14C-labeled molecular weight markers (Bethesda Research Laboratories, Gaithersburg, Md.), and the gels were prepared for fluorography.

Infection, labeling, and extraction of rec700 ADP for immunoprecipitation with antibodies against residues 2 to 16 or 63 to 77 of the Ad2 sequence. KB cells were infected with rec700 (5 x 107 cells were infected with 170 PFU per cell). The cells were labeled (in Cys- DMEM) with 250 µCi of [35S]Cys (NEN-Dupont) from 18 to 23 h p.i. The cells were rinsed with PBS-PMSF and extracted with Iso-hi-pH buffer containing 0.5% NP-40. ADP was immunoprecipitated as previously described (32) with rabbit antipeptide antiserum raised against residues 2 to 16 or 63 to 77 of the Ad2 ADP sequence. The immunoprecipitates were run on a 10 to 21% gradient gel by SDS-PAGE, and the gels were fluorographed.

Infection and labeling of CHO and ldlD-14 cell lines. CHO and ldlD-14 cell lines were plated in 60-mm-diameter dishes in Ham's F-12 medium containing 3% newborn calf lipoprotein-deficient serum prepared as described previously (17). After 24 h, the medium was changed to F-12-ITS+ (insulin, transferrin, selenium, linoleic acid, and BSA; Collaborative Research, Inc., Bedford, Mass.). After 2 days, the cells were infected with Ad2 at 400 PFU per cell in F-12-ITS+ medium. The cells were labeled from 24 to 30 h p.i. with 60 µCi of Tran35S per 60-mm-diameter dish in 1 ml of Cys- Met- DMEM (10 mM HEPES, pH 7.2). ITS+ was added to a final concentration of 1% after 1 h of labeling. For the tunicamycin-treated samples, 2 µl of tunicamycin stock (10 mg/ml in dimethyl sulfoxide; final concentration, 20 µg/ml) was added 25 min before labeling was begun. The cells were harvested, extracted with Iso-hi-pH buffer-0.5% NP-40 containing protease inhibitors (1 mM PMSF and 0.24 mg of leupeptin/ml), and the nuclei were removed by microcentrifugation. ADP was immunoprecipitated with rabbit antipeptide antiserum against residues 87 to 101 of the Ad2 sequence as previously described (32). Samples were run on SDS-PAGE (10 to 18% gradient), and the gels were fluorographed.

Infection with ADP mutant viruses, protein labeling, immunoprecipitation, and SDS-PAGE. Human KB cells in suspension culture (107 cells/infection) were infected with the relevant viruses at 50 (see Fig. 5A and C) or 100 (see Fig. 5B) PFU/cell in 4 ml of serum-free J-MEM for 1 h. The cells were diluted to 5 x 105/ml in J-MEM (final concentration, 5% equine serum) and then incubated at 37°C with rotation for ~24 h. The cells were pelleted, rinsed, and resuspended in 1 ml of DMEM (Cys- Met-) (Gibco BRL) and transferred to microcentrifuge tubes, and the proteins in each sample were radiolabeled. The cells were radiolabeled with 50 µCi of Expre35S35S (NEN-Dupont) for 4 h (see Fig. 5A), the cells were radiolabeled from 26 to 30 h p.i. with 50 µCi of Tran35S (ICN) and 50 µCi of 35S-Cys (NEN-Dupont) to compensate for decreased numbers of Met residues in the point (single-amino-acid) mutants (see Fig. 5C), or the cells were labeled from 23.5 to 28 h p.i. with 75 µCi of Expre35S35S (NEN-Dupont) to compensate for the apparent lower stability of the ADPs in the lumenal-domain mutants (see Fig. 5B). In each case, 50 µl of equine serum was added 45 to 60 min after labeling was begun.






View larger version (244K):
[in this window]
[in a new window]
  		FIG. 5. ADP synthesized by Ads with mutations in the adp gene. KB cells in suspension culture were infected with the indicated viruses, and proteins were labeled with [35S]Met and [35S]Cys. Proteins were extracted, ADP was immunoprecipitated, and the immunoprecipitates were analyzed by SDS-PAGE. The nature of the mutation for each virus is shown above each lane. (A to C) Cells were infected with 50 (A and C) or 100 (B) PFU of the mutant viruses/cell. At ~24 h p.i., the cells were labeled with 50 (A and C) or 75 (B) µCi of [35S]Met for 4 h. In addition to the [35S]Met, 50 µCi of [35S]Cys was added for the Met point mutants (C) to compensate for the decreased number of Met residues. ADP was immunoprecipitated from cell extracts using a rabbit polyclonal antibody to residues 87 to 101 of the Ad2 ADP sequence. Mock, mock infected. (D) Immunoblot of rec700, pm734.8 (Met56 to Leu), and pm734.9 (Leu96Leu97Leu98 to AlaAlaAla). KB cells in suspension culture were infected with 50 PFU of the indicated viruses/cell and were harvested at 28.5 h p.i. For the immunoblot, a rabbit antipeptide antiserum against residues 63 to 77 of the Ad2 ADP sequence was used.

After being labeled, the cells were pelleted, rinsed with ice-cold PBS-PMSF (1 mM), and extracted with Iso-hi-pH buffer containing 1 mM PMSF and 0.5% NP-40 (see Fig. 5A and C) or extracted with NP-40 lysis buffer (250 mM NaCl, 0.1% NP-40, 50 mM HEPES, pH 7.0) (see Fig. 5B). ADP was immunoprecipitated as previously described (32) using a rabbit antipeptide antiserum specific for residues 87 to 101 of the Ad2 ADP sequence. The immunoprecipitates were run on SDS-PAGE and prepared for fluorography as described previously (33). Ten to 18% gradient gels (see Fig. 5A and C) and a 15% gel (see Fig. 5B) are shown.

Immunoblotting. KB cells were infected with 50 PFU of viruses/cell. At 28.5 h p.i., the cells were collected by centrifugation, rinsed twice with ice-cold PBS, and lysed on ice for 30 min with lysis buffer (10 mM Tris-HCl, pH 7.4, 0.4% deoxycholate, 66 mM EDTA, 1.0% NP-40, and 1 mM PMSF). The lysates were cleared by microcentrifugation (Eppendorf) at 12,000 rpm for 3 min. The supernatants were collected, and the protein concentrations were quantitated with the Bio-Rad DC protein assay kit; 0.25 µg of each sample was electrophoresed on SDS-15% PAGE. The proteins were electrophoretically transferred to an Immobilon-P membrane, and the membrane was blocked with 5% nonfat milk. The blot was probed with an antipeptide antiserum directed against aa 63 to 77 of the Ad2 ADP sequence (1:400 dilution) and subsequently incubated with horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin G (IgG) (1:4,000; Cappel-ICN). The bands were visualized using the ECL protocol (Amersham Pharmacia).

Indirect immunofluorescence microscopy. A549 cells were plated on no. 1 glass coverslips (Corning) in 35-mm-diameter tissue culture dishes 2 days preinfection. Cells at 5 x 105/dish were infected at 50 PFU/cell. At 30 h p.i., the cells were fixed in methanol (-20°C) for 10 min and subsequently rehydrated in PBS. The fixed virus-infected cells were incubated with the primary antibody (rabbit antipeptide antiserum to residues 87 to 101 of the Ad2 ADP sequence) at a 1:400 dilution. The secondary antibody was goat anti-rabbit IgG (whole-molecule)-fluorescein isothiocyanate conjugate (Cappel-ICN) at a 1:50 dilution. Primary and secondary antibodies were diluted in PBS (pH 7.4)-1% BSA-0.1% sodium azide. Coverslips were mounted on slides in Elvanol (13) containing p-phenylenediamine to prevent fading. Immunofluorescence was viewed by epifluorescence on a Nikon Optiphot microscope. A Nikon DXM1200 digital camera was used to record images of the infected cells using Nikon ACT-1 software.

Plaque development assays. Plaque assays were done as described previously (29). Briefly, A549 cell monolayers were infected in triplicate with 10-fold serial dilutions of CsCl-banded Ad stocks; after adsorption, the cells were overlaid with DMEM-Noble agar-2% FBS. The WT parental virus (rec700) and an ADP mutant (either dl712 or pm734.1) were included in all assays as controls. Macroscopic plaques were counted at 2- to 5-day intervals for ~30 days p.i. The data were then plotted as the number of plaques on a given day as a percentage of the final plaque number to give the kinetics of plaque development. Typically, the results are based on 200 to 500 total plaques per virus.

Trypan blue exclusion. For the trypan blue exclusion experiment for lumenal ADP mutants (see Fig. 8A), A549 cells at 1.4 x 106 per 60-mm-diameter dish were infected at 20 PFU/cell for 1 h in 1 ml of serum-free DMEM. Four milliliters of DMEM (10% FBS) were added per dish after 1 h of virus adsorption (final concentration, 8% FBS). For the trypan blue exclusion experiments for point mutants and cytoplasmic deletion mutants (see Fig. 8B and C), A549 cells (9.1 x 105 per 35-mm-diameter dish) were infected at 25 PFU per cell for 1.5 h in 1 ml of serum-free DMEM. At the end of the adsorption period, 1.5 ml of DMEM (10% FBS) was added (final FBS concentration, 6%). To compare additional ADP mutants, A549 cells at 2.0 x 106 per 60-mm-diameter dish (see Fig. 8D) were infected at 20 PFU per cell in 1 ml of serum-free DMEM. At the end of a 1.5-h adsorption period, 4 ml of DMEM (10% FBS) was added (final concentration of serum, 8% FBS). At daily intervals, the medium was removed and attached cells were trypsinized and pooled with cells in the supernatant so that all cells would be included. A 5% volume of trypan blue (0.4% in saline; Gibco BRL) was added (final concentration, 0.02% trypan blue), and viable and nonviable cells were counted (typically 500 to 700 cells/sample).



View larger version (30K):
[in this window]
[in a new window]
  		FIG. 8. Loss of cell viability is altered for human A549 cells infected by viruses with ADP mutations. (A) A549 cells were infected at 20 PFU per cell with the indicated viruses with ADP lumenal mutations. At daily intervals, cell viability was assayed by trypan blue exclusion. rec700 is the WT ADP control. pm734.1 is the ADP- control. The virus mutations were dl735 ({Delta}4-11), pm734.7 (Asn14 to Ser), pm734 ({Delta}1-40), pm734.1 (missense mutations of Met1 and Met41 such that ADP would initiate at Met49), and dl737 ({Delta}29-45). (B) A549 cells were infected at 25 PFU per cell with viruses with point mutations of Met residues. The mutations were as follows: rec700 (WT control), pm734.1 ({Delta}1-48), pm734.2 (Met41 to Leu), pm734.3 (Met49 to Leu), pm734.4 (Met56 to Leu and Cys52 to Arg), and pm734 ({Delta}1-40). (C) A549 cells were infected at 25 PFU per cell with viruses containing deletions of the transmembrane or cytoplasmic sequence of ADP. The viruses were as follows: pm734.1 ({Delta}1-48), dl738 ({Delta}46-60), dl715.1 ({Delta}63-70), dl714 ({Delta}71-94), dl715 ({Delta}76-89), dl716 ({Delta}79-101), and dl717 ({Delta}81-88). (D) A549 cells were infected at 25 PFU per cell with viruses containing point mutations in the transmembrane or cytoplasmic sequence of ADP. Controls were rec700 (WT), pm734.1 (ADP-), and dl712 (ADP-). Point mutations were as follows: pm734.4 (Met56 to Leu and Cys52 to Arg), pm734.8 (Met56 to Leu), and pm734.9 (Leu96Leu97Leu98 to Ala96Ala97Ala98). Panels B and C are from the same experiment.


arrow
RESULTS
 
Pulse-chase assay indicates that ADP is processed to bands of lower mobility and, subsequently, bands of higher mobility. The Ad2 WT ADP appears as multiple bands on SDS-PAGE with apparent molecular weights of ~27 to 31K, ~19K, and ~13 to 14K (25). It is not known whether one or all of these products are the functional forms or whether there are precursor-product relationships. In order to clarify the putative sequential processing that is occurring, a pulse-chase experiment was conducted. Suspension cultures of human KB cells infected with Ad2 were labeled with Tran35S for 20 min at 26 h p.i. and then chased in the presence of excess unlabeled Met and Cys. Cells were collected at intervals, proteins were extracted, and ADP was immunoprecipitated and then separated on SDS-PAGE. Pulse-chase data for ADP indicated that the ~19K band is the initially synthesized form; it was the only product apparent at the end of the 20-min labeling period (designated zero hour of pulse-chase) (Fig. 1). This should be the full-length protein, N glycosylated at Asn14 (the NH2 terminus of ADP is presumed to be oriented into the lumen). After 30 to 60 min, a series of polypeptide bands with lower mobility appeared at 20K and stepwise up to ~27 to 31K (Fig. 1). It is likely that the smallest of these bands represents the modification of high-mannose N-linked oligosaccharides to complex oligosaccharides and that the largest of the bands represents modification by O glycosylation. ADP has serines and threonines within the lumenal region (residues 1 to 40) that are potential sites for O-linked glycosylation. Some of the bands could also arise from conversion of the N-linked high-mannose sugars to negatively charged complex oligosaccharides (25).



View larger version (90K):
[in this window]
[in a new window]
  		FIG. 1. ADP undergoes posttranslational modification, as indicated by pulse-chase analysis of Ad2 ADP. Human KB cells in suspension culture were infected with Ad2 at 50 PFU/cell. At 26 h p.i., the cells were labeled for 20 min with Tran35S, and then at the indicated hours of chase (top), proteins were extracted. ADP was immunoprecipitated with rabbit antipeptide antiserum against residues 87 to 101 of the Ad2 sequence and separated by SDS-PAGE (10 to 18% gradient gel). The upper, middle, and lower arrows indicate the ~27 to 31K, ~19K, and ~13 to 14K groups of bands of processed ADP that are discussed in the text.

The bands with the highest mobilities (~13 to 14K) began to appear 1 to 2 h postlabeling (Fig. 1). It is postulated that these polypeptides arise by proteolytic cleavage. These bands are no longer diffuse, suggesting that the glycosylated sequences have been removed. Previously, it had been shown that these polypeptide bands are not labeled with radioactive sugars (25). The presence of these bands only after 1 to 2 h would indicate that they do not arise by internal initiation at Met41. The bands first appeared subsequent to putative O glycosylation and modification of N-linked sugars; this may indicate that cleavage normally occurs only subsequent to N and O glycosylation. At 10 h postlabeling, all polypeptides remained quite stable, suggesting that there is not a total precursor-to-product conversion. Even 24 h postlabeling, all forms of ADP continued to be present with relative abundances similar to those seen at 10 h (data not shown).

Antibody against residues 2 to 16 of the ADP sequence recognizes only the ~19K ADP band. To further define which polypeptide bands represent which portions of ADP, Ad2 ADP was immunoprecipitated with rabbit antipeptide antiserum against residues 2 to 16 or residues 63 to 77 of the ADP sequence. The ADP band pattern seen by SDS-PAGE following immunoprecipitation with antibody against residues 63 to 77 (Fig. 2A) was essentially identical to that seen with the antibody generated against residues 87 to 101 (Fig. 1). The antibody against aa 63 to 77 showed similar multiple bands on immunoblots (32) (see Fig. 5D). Rabbit antipeptide antibody against residues 2 to 16 of the Ad2 sequence recognized only a subset of the ~19K protein products (Fig. 2A, left lane). This may indicate that N glycosylation at Asn14 and/or potential O glycosylation sites within the 2-to-16 region of the protein cause steric interference for recognition of the site by the antibody. The antibody did not recognize the 13 to 14K bands, again suggesting that the lumenal portion of the protein is removed by proteolysis (as suggested by the pulse-chase data in Fig. 1).



View larger version (54K):
[in this window]
[in a new window]
  		FIG. 2. Characterization of the different forms of ADP visible on SDS-PAGE. (A) Only the ~19K form of ADP is recognized by antibody to residues 2 to 16 of the Ad2 sequence. The rec700 ADP was immunoprecipitated using rabbit antipeptide antiserum against residues 2 to 16 (lane P2-16) or 63-77 (lane P63-77) of the Ad2 ADP sequence. Lane P63-77 has forms similar to those seen in Fig. 1 (in which ADP was immunoprecipitated with antipeptide antiserum raised against residues 87 to 101 of the Ad2 sequence), while the 2-16 antibody recognized only the ~19K form of ADP. Samples were run on 10 to 21% gradient SDS-PAGE. (B) ADP is O and N glycosylated. ADP was immunoprecipitated from Ad2-infected CHO (lanes a and c) or ldlD-14 (lanes b and d) cells. ldlD-14 cells (a CHO derivative) are deficient in O glycosylation. For lanes c and d, labeling was done in the presence of tunicamycin (+ Tun) to inhibit N glycosylation. Proteins were labeled from 24 to 30 h p.i. ADP was immunoprecipitated from cellular extracts with rabbit antiserum against residues 87 to 101 of the Ad2 sequence; samples were run on 10 to 18% gradient SDS-PAGE. The solid arrows indicate the positions of the WT ADP bands at 13 to 14K, ~19K, and ~27 to 31K. The open arrows indicate the bands seen in the presence of tunicamycin.

We previously reported that the ~19K band was labeled with [3H]mannose, while the 27 to 31K bands had higher incorporation of 3H-glucosamine (25). These bands became resistant to endoglycosidase H digestion 4 to 10 h postlabeling. Tunicamycin data showed that ADP was N glycosylated (25). To determine further whether the diffuse bands with the lowest mobilities are O glycosylated as previously postulated (25), virus infections were conducted in CHO and ldlD-14 cells. The ldlD-14 cell line, derived from the parental CHO cell line, has a mutation that prohibits O glycosylation of proteins (15, 16). Processing of ADP in CHO cells was very similar to that seen in human cells, giving rise to bands of ~27 to 31K, ~19K, and 13 to 14K (Fig. 2B, lane a). Immunoprecipitates from CHO cells (lane a) contained the ~27 to 31K ADP bands; these bands were not present in extracts from ldlD-14 cells (lane b). This would support the premise that the ~27 to 31K bands are normally O glycosylated. Reduced amounts of the 13 to 14K bands were obtained from the ldlD-14 cell line (lane b). This could suggest that O glycosylation facilitates the processing and cleavage to these forms.

Some samples were treated with tunicamycin prior to and during the labeling period to inhibit N-linked glycosylation. No band was apparent at ~19K, and the bands with the lowest mobilities (normally found at ~27 to 31K) had higher mobilities (~25 to 26K) and appeared to be less diffuse when tunicamycin was added to the CHO infection (Fig. 2B, lane c). This indicates that the upper diffuse bands are normally O and N glycosylated in untreated cells; in tunicamycin-treated CHO cells, the form with higher mobility represents O glycosylation in the absence of N glycosylation. It is apparent that this O-linked form can be processed to the 13 to 14K products, indicating that N glycosylation is not a prerequisite for subsequent cleavage (also see the data for the ADP mutant virus pm734.7 below). There was no change in size for the ~13 to 14K bands with tunicamycin treatment or when O glycosylation was blocked; this indicates that these bands are not glycosylated. For ADP from ldlD-14 cells, the absence of N and O glycosylation resulted in bands of 15 to 16K, as well as the 13 to 14K bands (Fig. 2B, lane d). The bands at 16K could represent the primary translation product that is neither N nor O glycosylated and that has not undergone proteolytic processing. It is probable that the 13 to 14K bands are due to proteolytic processing. The heterogeneity of these small bands suggests additional modification, such as addition of fatty acids; Hausmann et al. (12) have shown that the Ad5 ADP is palmitoylated. The cysteines identified as targets for palmitoylation in the Ad5 ADP are conserved in the Ad2 ADP. Previous attempts to detect phosphorylation were negative (25).

Construction of virus mutants in which adp has been altered by mutagenesis. From earlier research, it was known that deletions within the coding sequence of ADP could greatly alter splicing patterns in E3, resulting in altered abundances of the E3 proteins (26). Therefore, it was necessary to generate point mutations and small deletions to distinguish ADP function from differences generated by alternative pre-mRNA splicing. Deletion and point mutations were engineered into the adp gene by mutagenesis of the ADP sequence. Virus mutants were analyzed in viral infections for the synthesis, stability, posttranslational modification, and intracellular localization of ADP. The degree of ADP function was determined by plaque and cell viability assays. Figure 3 is a schematic of ADP (top), as well as the mutations that were constructed in ADP. The phenotypes of the mutants as determined by this study are indicated. Figure 4 shows the general phenotypes of some of the mutants with respect to plaque size. The synthesis and intracellular localizations of the mutant proteins are shown in Fig. 5 and 6, respectively. The abilities of the mutant viruses to induce cell lysis were measured in two different assays, a plaque development assay (Fig. 7) and a trypan blue exclusion assay (Fig. 8).



View larger version (50K):
[in this window]
[in a new window]
  		FIG. 4. Plaque morphology of viruses containing ADP mutations. Photographs were taken of representative dishes 14 days p.i. When plaque size for the mutants is compared to that for rec700 (the WT parental virus) and dl712 (in which adp is deleted), the plaques show a range of sizes. (A) Dishes are from the same experiment shown in Fig. 7A; (B) dishes are from the same experiment shown in Fig. 7B.



View larger version (102K):
[in this window]
[in a new window]
  		FIG. 6. Indirect immunofluorescence assay of ADP for adp mutants. Human A549 cells growing in monolayers were infected with 50 PFU of the indicated viruses/cell. The cells were fixed in methanol at 30 h p.i. The primary antibody was rabbit antipeptide antiserum specific for residues 87 to 101 of the Ad2 sequence. The secondary antibody was goat anti-rabbit IgG (fluorescein isothiocyanate conjugate). dl717 is shown at higher magnification. The images were taken with a 60x lens, except that for dl717 for which a 100x lens was used.



View larger version (36K):
[in this window]
[in a new window]
  		FIG. 7. The kinetics of plaque development on human A549 monolayers are altered for many Ad ADP mutants. The y axis shows the number of plaques observed on any given day of the plaque assay (x axis) as a percentage of the total number of plaques that were observed on the final day (days 25 to 30) of the plaque assay. The WT parental (rec700) and dl712 (total deletion of ADP) or pm734.1 (mutation of Met1 and Met41 in the ADP sequence; this ADP is nonfunctional) viruses are included as controls in each plaque assay. Each point represents the percentage of the final plaques that were macroscopically visible on the indicated day p.i. (A) dl732 (which overexpresses WT ADP due to altered mRNA splicing), rec700 (WT), dl712 ({Delta}ADP), pm734 ({Delta}1-40), dl735 ({Delta}4-11), dl736 (frameshift after the first 10 codons of ADP), dl738 ({Delta}46-60), pm734.2 (Met41 to Leu), and dl716 ({Delta}79-101). (B) rec700 (WT), dl712 ({Delta}ADP), dl737 ({Delta}29-45), pm734.3 (Met49 to Leu), pm734.4 (Met56 to Leu and Cys52 to Arg), and pm734.1 (Met1 and Met41 to Leu). (C) rec700 (WT), dl712 ({Delta}ADP), dl714 ({Delta}71-94), dl715 ({Delta}76-89), dl716 ({Delta}79-101), dl717 ({Delta}81-88), pm734.7 (Asn14 to Ser), and dl715.1 ({Delta}63-70). (D) rec700 (WT), pm734.1 (Met1 and Met41 to Leu), pm734.4 (Met56 to Leu and Cys52 to Arg), pm734.8 (Met56 to Leu), and pm734.9 (Leu96Leu97Leu98 to Ala96Ala97Ala98).

Mutations within the adp gene result in viruses with altered plaque sizes. It has been reported that deletion of the adp gene results in a small-plaque phenotype on A549 cells (30, 31). The release of infectious progeny from an infected cell results in infection and morphological changes in surrounding cells (rounding and/or death), producing a macroscopically visible plaque. Viruses (i.e., adp mutants) in which cell killing and virus release are less efficient will have smaller plaques that become visible with delayed kinetics. Direct observation of representative plaque assay dishes 14 days p.i. indicated substantial variation in plaque size for a panel of viruses in which adp had been mutated (Fig. 4). Some mutants (dl732, dl716, and pm734.3) exhibited larger plaque sizes than the WT parental virus, rec700 (Fig. 4). dl732 encodes a WT ADP that is overproduced due to an upstream mutation that increases the abundance of the mRNA for ADP; the overexpression of ADP by dl732 is associated with increased plaque size (Fig. 4). Other viruses (e.g., dl738, dl735, and pm734.1) had small plaques that were similar in size to those of dl712, a mutant in which the adp gene is completely deleted. In addition, some mutants had intermediate plaque sizes or had a wider range of plaque sizes.

Characteristics of wild-type and ADP-null virus infections. rec700 is an Ad5-Ad2-Ad5 chimera with Ad2 sequences for much of E3, including the adp gene. rec700 is the WT parental virus for the ADP mutants in this study. In order to have a completely null ADP mutant, mutant pm734.1 was constructed by mutating both the Met1 and Met41 codons of the ADP sequence. A putative protein initiating at Met49 would lack a portion of the signal-anchor sequence and would be postulated to initiate more poorly or to be unstable. The pm734.1 ADP was not detectable by immunoprecipitation (Fig. 5C, lane f). pm734.1 has a plaque development phenotype that approximates the adp deletion mutant dl712 (Fig. 4 and 7); therefore, pm734.1 was used as a null mutation in these experiments. Since pm734.1 has only point mutations, there is no alteration in the splicing patterns of the E3 mRNAs.

For WT rec700, three groups of bands were observed on SDS-PAGE (Fig. 5A, lane b, and B, lane a): the middle protein band, containing exclusively N-linked high-mannose oligosaccharides; the upper band, consisting of heterogeneous ADP molecules with O-linked and complex N-linked oligosaccharides (25); and the lower bands that arise by posttranslational proteolytic processing. This band pattern for rec700 is very similar to that seen with Ad2 ADP (Fig. 1) (31).

For rec700 infection, ADP localized sharply to the nuclear membrane, and in addition, there was strong staining of ADP in the Golgi apparatus (WT) (Fig. 6A). There was very little staining of the ER and no detectable staining of the plasma membrane. This immunofluorescence pattern is typical of ADP at very late stages (30 to 40 h) of Ad2 infection (25). It was postulated previously that the site of action of ADP in promoting cell death is the nuclear membrane and/or the Golgi apparatus-trans-Golgi network (TGN) (30, 31). No immunostaining was observed for pm734.1 (Fig. 6B), indicating that the P87-101 antibody has no background under these staining conditions (mock-infected cells also were not stained [not shown]).

As described previously (31), the efficiency of virus release can be estimated from the kinetics of plaque development in a plaque assay (see also Fig. 7). Viruses (i.e., adp mutants) in which cell killing and virus release are less efficient will have smaller plaques that become visible with delayed kinetics. The kinetics of plaque formation can be determined as the number of plaques seen on a given day as a percentage of the final plaque number. With rec700, >80% of all plaques were visible by 9 to 10 days p.i. (31). However, when adp is deleted, a much smaller percentage (10 to 30%) of the final plaques were visible 9 to 10 days p.i., and typically the final plaque number was not reached until ~4 weeks p.i. (30, 31).

In the series of experiments described in this report, rec700-induced plaques were readily visible ~5 days p.i., and they exceeded 90% of the final plaque number by 11 days p.i. (Fig. 7A). Overproduction of ADP increases the rate of plaque formation (dl732 [Fig. 7A]). Previous data suggest that this is due to earlier cell lysis and more efficient release of virions from the cells (4-6, 30, 31). When adp is deleted completely (dl712) or has a frameshift early in its reading frame (dl736; only the first 10 residues of ADP are in frame), the rate of plaque formation is extremely compromised (Fig. 7A). At 6 days p.i., there were no plaques for dl712 and only 2% of the final plaque number for dl736; it required 21 to 25 days for dl712 and dl736 to reach 90% of the final plaque number (Fig. 7A).

Cell viability assays were used to further characterize ADP function. A549 cell viability and cell lysis were determined at daily intervals p.i. by trypan blue exclusion. Lactate dehydrogenase release was assayed in parallel with similar results (data not shown). Each experiment included rec700 as the WT parental virus control and pm734.1 as the ADP- control. pm734.1 was totally defective in promoting cell lysis in the plaque development (Fig. 7B and D) and trypan blue (Fig. 8) assays. These results are very similar to those seen for viruses in which the entire adp gene is deleted (e.g., dl712 [Fig. 8D]) (31). Trypan blue exclusion assays indicated that rec700-infected A549 cells began to lose viability by 2 to 3 days p.i. (Fig. 8) while pm734.1-infected cells do not begin to lose viability until 5 to 7 days p.i.

Viruses encoding lumenal ADP mutations exhibit altered ADP glycosylation, lowered stability of ADP, and delayed loss of cell viability following infection. Mutation of the single N-linked glycosylation site at Asn14 to Ser (pm734.7) resulted in the synthesis of an ADP containing O-linked oligosaccharides (Fig. 5A, lane c, and B, lane b). The middle band corresponding to the high-mannose N-glycosylated species was not produced, as would be expected. The mobility and sharpness of the upper band were increased, indicating that the band normally contains both N-linked and O-linked sugars. The proteolytic processing products of ~13 to 14K were not affected. This is consistent with the CHO and ldlD-14 cell data: in CHO cells treated with tunicamycin, bands were present at ~26K and 13 to 14K (Fig. 2B). The pm734.7 ADP localized more to the Golgi apparatus-TGN than to the nuclear membrane (Fig. 6C), suggesting a delay in processing or decreased stability of ADP. pm734.7 showed ~20% fewer plaques at 6 to 12 days p.i. than were shown by rec700 (Fig. 7C). pm734.7 gave an intermediate phenotype in the trypan blue exclusion assay (Fig. 8A).

For dl735 ({Delta}4-11), the N-glycosylated species, the proteolytic products, and a small amount of heterogeneous upper species were obtained (Fig. 5A, lane d, and B, lane c). These data suggest that the major sites of O glycosylation are near the NH2 terminus; the deletion in dl735 appears to remove the major O-linked glycosylation sites (the slowly migrating species have complex N-linked oligosaccharides); this ADP appears to be less stable than the WT. The mobility of the upper bands is increased due to the deletion, but the smallest bands (~13 to 14K) migrate at the same size as for wild-type ADP. Thus, the deletion is not found in these smallest bands, consistent with the putative removal of the lumenal sequences by proteolysis to generate the ~13 to 14K bands. The dl735 ADP localized almost exclusively in the Golgi apparatus-TGN with very little staining of the ER or nuclear membrane (not shown). dl735 formed plaques only slightly more rapidly than dl712 (in which adp is deleted) (Fig. 7A). By trypan blue analysis, dl735 showed greatly extended survival of infected cells, with >70% of the cells remaining viable 6 to 7 days p.i. (Fig. 8A).

Two other mutations in the lumenal domain result in proteins with very low abundances on SDS-PAGE. The ADP of pm735.1 ({Delta}18-22) is probably very unstable (Fig. 5B, lane d). dl736.1 ({Delta}11-26) also showed much reduced levels of ADP (Fig. 5B, lane e), suggesting instability of ADP. The deleted sequence in dl736.1 includes the N glycosylation site at Asn14. Bands of lower mobility were present at ~20 to 21K, which would suggest retention of O glycosylation sites. There was less formation of the proteolytic products at 13 to 14K (Fig. 5B, lane e). Because these mutants appear to have very unstable ADP proteins, the data on subcellular localization and function of the ADPs is not presented.

The dl737 ADP ({Delta}29-45) is stable despite the relatively large deletion in the protein sequence (Fig. 5A, lane e). This ADP is both N and O glycosylated, as indicated by the middle and upper bands; these bands have higher mobilities because residues 29 to 45 are deleted. The amounts of the proteolytically processed bands (~13 to 14K) were reduced or, in other gels, almost nonexistent. This supports the suggestion that the preferred proteolytic cleavage site is within the 29 to 45 region of ADP (near Met41). Diffuse bands of >30K suggest that additional glycosylation may be occurring for this ADP. It is possible that additional sites undergo O glycosylation when proteolytic cleavage is reduced or absent. The dl737 ADP localized to the nuclear membrane and Golgi apparatus in a manner similar to that of rec700 (Fig. 6D). Thus, residues 29 to 45 are not required for ADP to localize to the nuclear membrane, Golgi apparatus, or TGN. dl737 formed plaques at a normal to somewhat accelerated rate (Fig. 7B). In the trypan blue exclusion assay, death induced by dl737 was slightly delayed initially but later appeared relatively similar to death induced by rec700 (Fig. 8A). The majority of cells lost viability between 2 and 5 days p.i. for rec700 and dl737 infections.

Deletion of the entire lumenal sequence has been studied in the point mutant pm734. In this mutant, the codon for Met1 was replaced by a codon for Leu, thereby destroying the initiation codon for ADP. pm734 expresses a truncated form of ADP, presumably initiating internally at Met41 in the adp sequence. The species corresponding to residues 41 to 101 was observed in significant abundance (Fig. 5C, lane e). Note that the pm734 ADP comigrates with the lower group of bands that are generated by proteolytic processing; these bands may arise by cleavage near Met41 and/or the transmembrane region. The putative transmembrane region would be residues 41 to 62. By indirect immunofluorescence, the pm734 ADP localized more to the Golgi apparatus-TGN than to the nuclear membrane (Fig. 6E). Note that several cells in the field show only Golgi apparatus staining. Localization of the pm734 ADP to the Golgi apparatus suggests that signals in the aa 41 to 101 sequence are sufficient for transport of the newly synthesized ADP to the Golgi apparatus from the ER. This smaller ADP still retained some function. For pm734, 29% of plaques were visible 9 days p.i. compared to 81% for the WT (rec700) and 10% for the adp-null mutant dl712 (Fig. 7A). The phenotype of pm734 was intermediate in plaque development (Fig. 7) but was very defective in the trypan blue exclusion assay (Fig. 8A and B) (20 or 25 PFU/cell, respectively). Deletion of important sequences near the NH2 terminus, slow transport, or instability may explain why the pm734 ADP was defective in promoting cell lysis.

In summary, mutations within the lumenal region of ADP (residues 1 to 40) frequently change the pattern of N and O glycosylation. Apparently, O-linked (dl735) and N-linked (pm734.7) glycosylation (or residues 4 to 26 and Asn14) are required for efficient sorting of ADP from the Golgi apparatus-TGN to the nuclear membrane and/or for stabilization of ADP. Except for dl737, these ADP lumenal mutant viruses show reduced rates of plaque formation and delayed cell lysis.

Deletion of the transmembrane region abolishes the ability of ADP to induce cell death; viruses with deletions within the cytoplasmic-nucleoplasmic region of ADP show ADP stability and altered subcellular distribution of ADP and vary in their abilities to cause cell death ADP contains an internal hydrophobic domain (~41 to 62) that is believed to mediate the insertion of ADP into membranes and to serve as a signal-anchor domain (25, 31). The deletion in dl738 ({Delta}46-60) abolishes the signal-anchor sequence; this ADP should no longer be incorporated into membranes. Attempts to immunoprecipitate this ADP from cell extracts failed, perhaps because the dl738 ADP was either unstable or no longer soluble for extraction. Indirect immunofluorescence suggested that this ADP appears to form aggregates in the cytoplasm (Fig. 6F). The apparent nuclear staining could be due to the basic-proline sequence of ADP (residues 63 to 70), which is similar to nuclear localization signals, or may simply indicate passive diffusion because of the small size of the protein. dl738 displayed significantly delayed plaque formation (Fig. 7A) and complete loss of ADP function when assayed by trypan blue exclusion (Fig. 8C).

A number of deletions had previously been constructed in the cytoplasmic region of ADP. Data on ADP immunoprecipitated from infections with some of these mutants (dl714, dl716, and dl717) were published previously (32). These mutants all produce ADPs that have normal N and O glycosylation and that appear to be very stable.

The ADPs of dl715.1 ({Delta}63-70), dl714 ({Delta}71-94), and dl715 ({Delta}76-89) all showed more rapid migration of all bands, including the ~13 to 14K proteolytically cleaved products (Fig. 5A, lanes g, h, and i for dl715.1, dl714, and dl715, respectively); this would indicate that all bands present contain the C-terminal region (the antibody is raised against residues 87 to 101 of the Ad2 ADP) and all bands are shifted in size (higher mobility is due to the deletions). Similar results were obtained for dl716 ({Delta}79-101) and dl717 ({Delta}81-88) (using an antipeptide antibody to aa 63 to 77 of the Ad2 ADP sequence) (32).

The dl714 and dl715 ADPs localized not only to the nuclear membrane and Golgi apparatus-TGN but also to the ER and plasma membrane (Fig. 6G and H, respectively). The dl715.1 ADP localized not only to the nuclear membrane and Golgi apparatus-TGN but also to the ER and, to a lesser extent, to the plasma membrane (Fig. 6I). Overexpression of ADP by dl732 resulted in expression of ADP in the ER in addition to its localization in the Golgi apparatus and the nuclear membrane but not in the plasma membrane (not shown). Therefore, plasma membrane staining is not attributable to a high abundance of ADP.

Deletions in the mutants dl714 and dl716 had little or no effect on the ability of ADP to promote cell lysis. Surprisingly, dl714 and dl716 produced larger plaques than rec700 (WT) (Fig. 4) and more rapid plaque formation (Fig. 7C). dl714 and dl716 showed slightly delayed loss of viability compared to rec700 in trypan blue exclusion assays (Fig. 8C). Earlier immunoprecipitations and immunoblots had suggested that the dl714 and dl716 ADPs were more abundant than the rec700 ADP (32); higher expression could potentially compensate for the reduced specificity of localization.

The basic-proline domain appears to be important but not essential to the cell lysis-promoting activity of ADP. dl715.1 had slower plaque formation than rec700 (Fig. 7C) and an intermediate phenotype by the cell viability assay (Fig. 8C). In a separate experiment (data not shown), dl715.1 still showed much less killing activity when used for infection at 100 PFU per cell.

By immunofluorescence, the dl715 ADP was present in the plasma membrane, as well as in other cellular membranes (Fig. 6H). dl715 formed plaques at a rate similar to that of rec700 (Fig. 4 and 7C) and had an intermediate phenotype by the trypan blue exclusion assay (Fig. 8C).

The dl717 ADP appears to be abundant and is glycosylated normally, but by immunofluorescence assay, it exhibits a more vesicular pattern than WT ADP (Fig. 6J). The mutation in dl717 ({Delta}81-88) may interfere with a transport signal that is normally present. dl717 had slower plaque formation than rec700 (Fig. 7C). In the trypan blue exclusion assay, dl717 had a greatly delayed loss of viability compared to rec700 (Fig. 8C).

One additional mutation was made in the extreme C terminus of ADP. This was within the region that is no longer conserved between Ad2 and Ad5 due to a frameshift (36). The Ad2 ADP contains three consecutive leucine residues near the C terminus. Dileucine motifs frequently play a role in intracellular sorting (reviewed in reference 34). To determine whether these leucines were required for ADP trafficking, the mutant pm734.9 was made. Leu96Leu97Leu98 was replaced by Ala96Ala97Ala98 at the C terminus of ADP. The pm734.9 ADP was abundant and showed normal processing patterns (see the immunoblot in Fig. 5D). An alternate antibody was used to detect the pm734.9 ADP because the mutation disrupted recognition by the P87-101 antipeptide antiserum used in Fig. 5A, B, and C. Indirect immunofluorescence assay, using an alternate C-terminal rabbit polyclonal antibody, indicated good nuclear membrane localization (not shown). Plaque formation for pm734.9 was initially delayed but by day 10 was comparable to that of the WT (Fig. 7D); there was some delay in loss of infected-cell viability relative to rec700 (Fig. 8D).

In summary, ADPs that contain deletions in the cytoplasmic region (residues ~63 to 101) are stable. The presence of complex N-linked and O-linked glycosylation also indicates that these proteins have been inserted into membranes and have traversed the Golgi apparatus. The mobilities of the ~13 to 14K proteolytic processing products are increased in a manner consistent with the size of the deletions; this indicates that the cleavage event that produces these proteins occurs upstream of aa 63. In general, deletions in the C-terminal region of the ADP sequence did not inhibit plaque formation and cell killing as significantly as deletions within the lumenal domain.

Some point mutations in adp resulted in small-plaque phenotypes and delayed loss of viability after infection. Our data indicate that multiple bands of ADP arise by posttranslational processing of a product that has initiated at Met1. When Met1 is removed, translation of ADP can initiate at a downstream AUG, presumably Met41. In order to investigate whether translation initiates at other AUGs in the ADP mRNA, each of the other Met codons was mutated. The main purpose of this study was to allow us to distinguish between translation initiation products and ADP processing products. The results suggest that few, if any, of the many ADP species are due to initiation at Met41 or Met49 when the Met1 codon is present. With pm734.2 (Met41 to Leu) (Fig. 5C, lane b) and pm734.3 (Met49 to Leu) (Fig. 5C, lane c), the pattern and quantity of ADP bands were similar to those of rec700 (WT). By indirect immunofluorescence assay, the intracellular localizations of the pm734.2 and pm734.3 ADPs were similar to those of the WT (not shown). There was some increase of ER staining for the pm734.2 ADP. Both pm734.2 (Fig. 7A) and pm734.3 (Fig. 7B) formed plaques at rates similar to rec700. Mutation of Met56 to Leu (pm734.8) also showed normal kinetics of plaque formation (Fig. 7D). Viruses with mutations of Met41 (pm734.2), Met49 (pm734.3), or Met56 (pm734.8) in the full-length background were near WT in function in the trypan blue assay (Fig. 8B and C). These results indicate that Met41, Met49, and Met56 play little if any role in the cell lysis-promoting properties of ADP.

The initial attempt to make the Met56-to-Leu mutation inadvertently resulted in a double point mutation in which the Met56 codon was converted to a Leu codon but the Cys52 codon was mutated to an Arg codon. This mutant has been studied because it has an interesting phenotype. With pm734.4, only the middle band corresponding to the species with N-linked high-mannose oligosaccharides was observed (Fig. 5C, lane d, and A, lane f) (25). The pm734.4 ADP is not O glycosylated and does not show the highest mobility cleavage products (~13 to 14K). The pm734.4 ADP was stable and localized primarily to the nuclear membrane (Fig. 6K); the Golgi apparatus did not stain. The lack of Golgi apparatus staining correlates with the observed lack of O glycosylation in this mutant (Fig. 5A, lane f, and C, lane d). ADP could reach the nuclear membrane by lateral diffusion from the ER. This result, as well as indirect immunofluorescence (Fig. 6K), indicates that the pm734.4 ADP is no longer transported to the Golgi apparatus and TGN, where it would obtain O-linked and complex N-linked oligosaccharides. The lack of ~13 to 14K bands suggests that ADP must be sorted to the Golgi apparatus and/or be O glycosylated in order for these bands to form; proteolytic cleavage may normally occur subsequent to O glycosylation of ADP. pm734.4 was severely defective in plaque development (Fig. 7B and D) and nearly totally defective in trypan blue exclusion (Fig. 8B and C). The pm734.4 phenotype was markedly different from WT rec700 or pm734.8 (the Met56 mutation alone). The pm734.8 ADP protein exhibited a strong Golgi apparatus and good nuclear membrane localization (Fig. 6L). This further supports the hypothesis that the defective ADP localization in pm734.4 was due to the inadvertent second mutation of Cys52 to Arg.

In addition to the Met mutations generated in the full-length ADP background, several point mutations of Met residues were generated in the pm734 ({Delta}1-40) background. Mutation of the Met at position 41 in the pm734 background (pm734.1) has been described above. The pm734.5 ADP (Met1 and Met49 to Leu), which should initiate at Met41, showed mobility similar to that of the ADP bands of pm734 but had only one more prominent band at ~14K (Fig. 5C, lane g). pm734.6 (Met1 and Met56 to Leu and Cys52 to Arg), which should also initiate at Met41, had bands with higher mobilities (Fig. 5C, lane h). Perhaps this mutation prevents modification near Met56 (potentially altered palmitoylation). ADPs from all the mutants in a pm734 background seem to have lower abundance; this may indicate that these ADPs are less stable or that initiation at Met41 is far less efficient then at Met1. Indeed, the sequence of the Met1 initiation codon, GAGAUGC, is similar to the "Kozak sequence" consensus (A,G)XXAUG(A,G). Met41, Met49, and Met56 all deviate from the consensus sequence. In Fig. 5C, the infected cells had been labeled with [35S]Cys, as well as [35S]Met, to compensate for the decreased number of Met residues for the ADPs with these point mutations. Because of the low abundance and stability of these ADPs, their localizations and functions are not described here.


arrow
DISCUSSION
 
Pulse-chase analysis of ADP indicates that the protein appears to be N glycosylated cotranslationally. The protein is then O glycosylated, and subsequently, a portion of the protein is proteolytically processed such that the lumenal portion of ADP is removed.

The results of the characterization of adp mutants can be summarized as follows. Regarding the signals for ADP sorting within the cell, the lumenal domain, residues 1 to 40, appears to be vital for efficient transit of ADP through the Golgi apparatus-TGN and into the nuclear membrane. Residues 41 to 101 appear to be sufficient for transport from the ER to the Golgi apparatus. The most important signal with respect to transit through the Golgi apparatus-TGN appears to be glycosylation of ADP, because mutation of the N glycosylation site (pm734.7), or deletion of the O glycosylation sites at residues 4 to 11 (dl735), caused ADP to be retained to a significant extent in the Golgi apparatus-TGN. On the other hand, deletion of the entire lumenal domain does not preclude transport of ADP to the nuclear membrane, as shown with pm734, although the transport appears to be significantly delayed. It is clear that residues 29 to 45, which are deleted in dl737, play little if any role in the transit of ADP through the Golgi apparatus-TGN, inasmuch as the dl737 ADP was WT for localization to the nuclear membrane. It did appear that glycosylation may stabilize ADP. Recently, it has been reported that O glycosylation may reduce susceptibility to degradation in proteasomes (11, 28).

Mutations of Met residues at Met41, Met49, or Met56 in the transmembrane region did not result in altered processing or localization. Deletion of the transmembrane signal-anchor sequence results in apparent aggregation of ADP in the cytoplasm. ADPs containing deletions of cytoplasmic-nucleoplasmic sequences (residues ~60 to 101) undergo normal glycosylation and proteolytic cleavage and appear to be quite stable, but these ADPs have lost specificity for localization. The ADPs from dl714, dl715, and dl715.1 localize to all or most membranes; the sequences deleted (residues in the 63-to-94 region) may contain a signal or signals that target ADP specifically to the nuclear membrane or that would interact with nuclear or nuclear membrane proteins for retention at these locations. dl717 ({Delta}81-88) showed an increased number of vesicles in the cytoplasm that stain for ADP. Perhaps vesicles mediate the transport of ADP from the Golgi apparatus to the nuclear membrane. The dl717 deletion could reduce specificity for sorting, resulting in an accumulation of sorting intermediates.

ADP is synthesized in very large amounts during late Ad infection, typical of an Ad late protein (32). Thus, ADP probably functions in a stoichiometric manner. When ADP proteins were unstable or were produced in low abundance, the ability of ADP to promote cell lysis was completely or nearly completely abrogated. All mutations in the lumenal domain other than the dl737 deletion ({Delta}29-45) seem to delay export of ADP from the Golgi apparatus and may decrease the stability of the protein. Most of these mutants are severely compromised in the ability to induce cell lysis. Deletion of the signal-anchor domain (residues ~40 to 60) resulted in almost total loss of function in plaque and viability assays; we interpret this result to indicate that ADP must enter membranes in order to function. Moreover, the basic-proline domain (at residues ~63 to 70) undoubtedly plays an important facilitating role. Some differences are seen in the phenotypes of the viruses when plaque assay phenotypes are compared to trypan blue exclusion phenotypes. Our unpublished data suggest that ADP function is influenced by the multiplicity of infection, confluency of the monolayer, overlay effects, and serum concentration.

Relatively few nuclear membrane proteins have been described. Only a few proteins have been identified which localize to the inner nuclear membrane (reviewed in reference 8). These proteins contribute to nuclear organization and the association between the nuclear membrane and nuclear lamina and chromatin. The retention of these proteins in the inner nuclear membrane would be due to binding to ligands or multimerization. ADP does not appear to contain any of these published nuclear membrane localization sequences. ADP signals for localization may be informative in determining sequences for nuclear membrane location. The use of chimeric proteins in a transfection system indicated that the transmembrane and/or the cytoplasmic domains of the Ad5 ADP were sufficient for localization in the nuclear membrane, Golgi apparatus, and ER (9). Our own unpublished data indicate that the Ad2 ADP localizes to the inner nuclear membrane. A few viral proteins, including a number of the envelope glycoproteins of herpesviruses, have been found to localize in the inner nuclear membrane. Inner nuclear membrane localization of herpes simplex virus type 1 (HSV-1) gB also reportedly involved the membrane anchor (10, 23, 24). HSV-1 infection has been reported to disrupt the localization of LBR and nuclear lamins (27). This is postulated to facilitate the egress of the virions from the nucleus.

Meyer et al. (19) recently reported conserved hexameric motifs that targeted human cytomegalovirus gB, HSV-1 gB, and the cellular lamin B receptor to the inner nuclear membrane. Their experiments indicated that RXR was sufficient for localization of the CD8 reporter gene to the inner nuclear membrane. Such an RXR motif is present in the basic-proline region of ADP (aa 63 to 70). However, deletions within the aa 71 to 94 region of ADP (dl714, dl715, and dl717) do not localize specifically in the nuclear membrane; thus, RXR is not sufficient for the localization of ADP to the nuclear membrane.

ADP also localizes to the Golgi apparatus. The localization of the infectious bronchitis virus E protein in the Golgi apparatus is dependent on a sequence in the cytoplasmic tail (2). This protein, like ADP, has palmitoylation of one or two cysteines that are adjacent to the transmembrane domain; however, palmitoylation was not required for Golgi apparatus localization.

ADP may function by interacting with proteins in the nuclear membrane, or it may interact with proteins involved in maintaining nuclear integrity (for example, proteins interacting with nuclear lamins or chromatin), resulting in loss of nuclear integrity and loss of viability. If it is correct that ADP must localize to the nuclear membrane to promote cell lysis, then it is likely that ADP functions by interacting with and blocking the activities of cellular proteins that contact the nuclear membrane. Of interest, the Ad E1B-19K protein has been found to localize to the inner nuclear membrane and to associate with nuclear lamins (22); this is proposed to be a mechanism by which the E1B-19K protein inhibits apoptosis.

It has been reported that nuclear structure is altered and that virions and viral DNA are found in the cytoplasm during late stages of Ad infection (20). ADP might also affect the phosphorylation of inner nuclear membrane proteins. Differences in the phosphorylation of inner nuclear membrane proteins have been shown to occur during the cell cycle and mitosis (7, 41); the nuclear membrane is disrupted during mitosis. It could be expected that release of Ad progeny would be facilitated by disruption of the nuclear membrane and/or by counteracting cellular survival signals.

In addition to its role in Ad biology and as an example of a nuclear membrane protein, ADP is interesting because it increases the efficiency with which replication-competent cancer gene therapy vectors can lyse tumor cells (4-6). Because of the association of ADP with efficient cell lysis, we have constructed a number of replication-competent Ad vectors overexpressing ADP as potential cancer therapeutics (4-6). These vectors have been shown to spread more rapidly in tissue culture monolayers and to suppress human tumor growth in a nude-mouse tumor model. Another group has also placed the adp gene into Ad vectors designed for cancer therapy and has seen improved cell killing (21). The structure-function analyses of ADP described in this report may be useful in understanding the function of ADP in Ad cancer gene therapy vectors.


arrow
ACKNOWLEDGMENTS
 
We thank M. Krieger for ldlD-14 and CHO cells; Shari O'Brien, Lora Wold, and Chris Wells for technical assistance; and Dawn Schwartz for preparation of the figures and manuscript.

This work was supported by grants CA71704, CA24710, and CA58538 from the National Institutes of Health.


arrow
FOOTNOTES
 
* Corresponding author. Mailing address: Department of Molecular Microbiology and Immunology, St. Louis University Health Sciences Center, 1402 S. Grand Blvd., St. Louis, MO 63104. Phone: (314) 577-8435. Fax: (314) 773-3403. E-mail: woldws{at}slu.edu. Back

{dagger} Present address: Genzyme Corporation, Framingham, MA 01701. Back


arrow
REFERENCES
 
    1
  1. Bhat, B. M., H. A. Brady, and W. S. M. Wold. 1985. Virus deletion mutants that affect a 3' splice site in the E3 transcription unit of adenovirus 2. Mol. Cell. Biol. 5:2405-2413.[Abstract/Free Full Text]
    2. 2
  2. Corse, E., and C. E. Machamer. 2002. The cytoplasmic tail of infectious bronchitis virus E protein directs Golgi targeting. J. Virol. 76:1273-1284.[Abstract/Free Full Text]
    3. 3
  3. Deutscher, S. L., B. M. Bhat, M. H. Pursley, C. Cladaras, and W. S. M. Wold. 1985. Novel deletion mutants that enhance a distant upstream 5' splice in the E3 transcription unit of adenovirus 2. Nucleic Acids Res. 13:5771-5788.[Abstract/Free Full Text]
    4. 4
  4. Doronin, K., M. Kuppuswamy, K. Toth, A. E. Tollefson, P. Krajcsi, V. Krougliak, and W. S. M. Wold. 2001. Tissue-specific, tumor-selective, replication-competent adenovirus vector for cancer gene therapy. J. Virol. 75:3314-3324.[Abstract/Free Full Text]
    5. 5
  5. Doronin, K., K. Toth, M. Kuppuswamy, P. Krajcsi, A. E. Tollefson, and W. S. M. Wold. 2003. Overexpression of the ADP (E3-11.6K) protein increases cell lysis and spread of adenovirus. Virology 305:378-387.[CrossRef][Medline]
    6. 6
  6. Doronin, K., K. Toth, M. Kuppuswamy, P. Ward, A. E. Tollefson, and W. S. M. Wold. 2000. Tumor-specific, replication-competent adenovirus vectors overexpressing the adenovirus death protein. J. Virol. 74:6147-6155.[Abstract/Free Full Text]
    7. 7
  7. Ellenberg, J., E. D. Siggia, J. E. Moreira, C. L. Smith, J. F. Presley, H. J. Worman, and J. Lippincott-Schwartz. 1997. Nuclear membrane dynamics and reassembly in living cells: targeting of an inner nuclear membrane protein in interphase and mitosis. J. Cell Biol. 138:1193-1206.[Abstract/Free Full Text]
    8. 8
  8. Georgatos, S. D. 2001. The inner nuclear membrane: simple, or very complex? EMBO J. 20:2989-2994.[CrossRef][Medline]
    9. 9
  9. Ghosh, K., and H. P. Ghosh. 1999. Role of the membrane anchoring and cytoplasmic domains in intracellular transport and localization of viral glycoproteins. J. Biochem. Cell Biol. 77:165-178.
    10. 10
  10. Gilbert, R., K. Ghosh, and H. P. Ghosh. 1994. Membrane anchoring domain of herpes simplex virus glycoprotein gB is sufficient for nuclear envelope localization. J. Virol. 68:2272-2285.[Abstract/Free Full Text]
    11. 11
  11. Han, I., and J. E. Kudlow. 1997. Reduced O glycosylation of Sp1 is associated with increased proteasome susceptibility. Mol. Cell. Biol. 17:2550-2558.[Abstract]
    12. 12
  12. Hausmann, J., D. Ortmann, M. Witt, E. Veit, and W. Seidel. 1998. Adenovirus death protein, a transmembrane protein encoded in the E3 region, is palmitoylated at the cytoplasmic tail. Virology 244:343-351.[CrossRef][Medline]
    13. 13
  13. Heimer, G. V., and C. E. Taylor. 1974. Improved mountant for immunofluorescence preparations. J. Clin. Pathol. 27:254-256.[Free Full Text]
    14. 14
  14. Horwitz, M. 2001. Adenovirus immunoregulatory genes and their cellular targets. Virology 279:1-8.[CrossRef][Medline]
    15. 15
  15. Kingsley, D., K. Kozarsky, L. Hobbie, and M. Krieger. 1986. Reversible defects in O-linked glycosylation and LDL receptor expression in a UDP-Gal/UDP-GalNAc 4-epimerase deficient mutant. Cell 44:749-759.[CrossRef][Medline]
    16. 16
  16. Kozarsky, K., D. Kingsley, and M. Krieger. 1988. Use of a mutant cell line to study the kinetics and function of O-linked glycosylation of low density lipoprotein receptors. Proc. Natl. Acad. Sci. USA 85:4335-4339.[Abstract/Free Full Text]
    17. 17
  17. Krieger, M., D. Reddy, K. Kozarsky, D. Kingsley, L. Hobbie, and M. Penman. 1989. Analysis of the synthesis, intracellular sorting, and functions of glycoproteins using a mammalian cell mutant with reversible glycosylation defects, p. 57-84. In A. M. Tartakoff (ed.), Methods in cell biology. Academic Press, New York, N.Y.
    18. 18
  18. Mahr, J. A., and L. R. Gooding. 1999. Immune evasion by adenoviruses. Immunol. Rev. 168:121-130.[CrossRef][Medline]
    19. 19
  19. Meyer, G., D. Gicklhorn, T. Strive, K. Radsak, and M. Eickmann. 2002. A three-residue signal confers localization of a reporter protein in the inner nuclear membrane. Biochem. Biophys. Res. Commun. 291:966-971.[CrossRef][Medline]
    20. 20
  20. Puvion-Dutilleul, F., S. Besse, E. Pichard, and C. Cajean-Feroldi. 1998. Release of viruses and viral DNA from nucleus to cytoplasm of HeLa cells at late stages of productive adenovirus infection as revealed by electron microscopy in situ hybridization. Biol. Cell 90:5-38.[CrossRef][Medline]
    21. 21
  21. Ramachandra, M., A. Rahman, A. Zou, M. Vaillancourt, J. A. Howe, D. Antelman, B. Sugarman, G. W. Demers, H. Engler, D. Johnson, and P. Shabram. 2001. Re-engineering adenovirus regulatory pathways to enhance oncolytic specificity and efficacy. Nat. Biotechnol. 19:1035-1041.[CrossRef][Medline]
    22. 22
  22. Rao, L., D. Modha, and E. White. 1997. The E1B 19K protein associates with lamins in vivo and its proper localization is required for inhibition of apoptosis. Oncogene 15:1587-1597.[CrossRef][Medline]
    23. 23
  23. Rasile, L., K. Ghosh, K. S. Raviprakash, and H. P. Ghosh. 1993. Effects of deletions in the carboxy-terminal hydrophobic region of herpes simplex virus glycoprotein gB on intracellular transport and membrane anchoring. J. Virol. 67:4856-4866.[Abstract/Free Full Text]
    24. 24
  24. Raviprakash, K. S., L. Rasile, K. Ghosh, and H. P. Ghosh. 1990. Shortened cytoplasmic domain affects intracellular transport but not nuclear localization of a viral glycoprotein. J. Biol. Chem. 265:1777-1782.[Abstract/Free Full Text]
    25. 25
  25. Scaria, A., A. E. Tollefson, S. K. Saha, and W. S. M. Wold. 1992. The E3-11.6K protein of adenovirus is an Asn-glycosylated integral membrane protein that localizes to the nuclear membrane. Virology 191:743-753.[CrossRef][Medline]
    26. 26
  26. Scaria, A., and W. S. M. Wold. 1994. Fine-mapping of sequences that suppress splicing in the E3 complex transcription unit of adenovirus. Virology 205:406-416.[CrossRef][Medline]
    27. 27
  27. Scott, E., and P. O'Hare. 2001. Fate of the inner nuclear membrane protein lamin B receptor and nuclear lamins in herpes simplex virus type 1 infection. J. Virol. 75:8818-8830.[Abstract/Free Full Text]
    28. 28
  28. Su, K., M. D. Roos, X. Yang, I. Han, A. J. Paterson, and J. E. Kudlow. 1999. An N-terminal region of Sp1 targets its proteasome-dependent degradation in vitro. J. Biol. Chem. 274:15194-15202.[Abstract/Free Full Text]
    29. 29
  29. Tollefson, A. E., T. W. Hermiston, and W. S. M. Wold. 1998. Preparation and titration of CsCl-banded adenovirus stocks, p. 1-9. In W. S. M. Wold (ed.), Adenovirus methods and protocols. Humana Press, Totowa, N.J.
    30. 30
  30. Tollefson, A. E., J. S. Ryerse, A. Scaria, T. W. Hermiston, and W. S. M. Wold. 1996. The E3-11.6kDa Adenovirus Death Protein (ADP) is required for efficient cell death: characterization of cells infected with adp mutants. Virology 220:152-162.[CrossRef][Medline]
    31. 31
  31. Tollefson, A. E., A. Scaria, T. W. Hermiston, J. S. Ryerse, L. J. Wold, and W. S. M. Wold. 1996. The adenovirus death protein (E3-11.6K) is required at very late stages of infection for efficient cell lysis and release of adenovirus from infected cells. J. Virol. 70:2296-2306.[Abstract]
    32. 32
  32. Tollefson, A. E., A. Scaria, S. K. Saha, and W. S. M. Wold. 1992. The 11,600-MW protein encoded by region E3 of adenovirus is expressed early but is greatly amplified at late stages of infection. J. Virol. 66:3633-3642.[Abstract/Free Full Text]
    33. 33
  33. Tollefson, A. E., and W. S. M. Wold. 1988. Identification and gene mapping of a 14,700-molecular-weight protein encoded by region E3 of group C adenoviruses. J. Virol. 62:33-39.[Abstract/Free Full Text]
    34. 34
  34. Trowbridge, I. S., J. F. Collawn, and C. R. Hopkins. 1993. Signal-dependent membrane protein trafficking in the endocytic pathway. Annu. Rev. Cell Biol. 9:129-161.[CrossRef][Medline]
    35. 35
  35. Wold, W. S. M., and G. Chinnadurai. 2000. Adenovirus proteins that regulate apoptosis, p. 200-232. In A. J. Cann (ed.), DNA virus replication: frontiers in molecular biology. Oxford University Press, Oxford, United Kingdom.
    36. 36
  36. Wold, W. S. M., C. Cladaras, S. C. Magie, and N. Yacoub. 1984. Mapping a new gene that encodes an 11,600-molecular-weight protein in the E3 transcription unit of adenovirus 2. J. Virol. 52:307-313.[Abstract/Free Full Text]
    37. 37
  37. Wold, W. S. M., S. L. Deutscher, N. Takemori, B. M. Bhat, and S. C. Magie. 1986. Evidence that AGUAUAUGA and CCAAGAUGA initiate translation in the same mRNA region E3 of adenovirus. Virology 148:168-180.[CrossRef][Medline]
    38. 38
  38. Wold, W. S. M., K. Doronin, K. Toth, M. Kuppuswamy, D. L. Lichtenstein, and A. E. Tollefson. 1999. Immune responses to adenoviruses: viral evasion mechanisms and their implications for the clinic. Curr. Opin. Immunol. 11:380-386.[CrossRef][Medline]
    39. 39
  39. Wold, W. S. M., and A. E. Tollefson. 1998. Adenovirus E3 proteins: 14.7K, RID, and gp19K inhibit immune-induced cell death; adenovirus death protein promotes cell death. Semin. Virol. 8:515-523.[CrossRef]
    40. 40
  40. Wold, W. S. M., K. Toth, K. Doronin, M. Kuppuswamy, D. L. Lichtenstein, and A. E. Tollefson. 1999. Adenovirus inhibition of immune-mediated apoptosis. Clin. Immunol. Newsl. 19:1-7.
    41. 41
  41. Yang, L., T. Guan, and L. Gerace. 1997. Integral membrane proteins of the nuclear envelope are dispersed throughout the endoplasmic reticulum during mitosis. J. Cell Biol. 137:1199-1210.[Abstract/Free Full Text]

Journal of Virology, July 2003, p. 7764-7778, Vol. 77, No. 14
0022-538X/03/$08.00+0     DOI: 10.1128/JVI.77.14.7764-7778.2003
Copyright © 2003, American Society for Microbiology. All Rights Reserved.



This article has been cited by other articles:

  • Gros, A., Martinez-Quintanilla, J., Puig, C., Guedan, S., Mollevi, D. G., Alemany, R., Cascallo, M. (2008). Bioselection of a Gain of Function Mutation that Enhances Adenovirus 5 Release and Improves Its Antitumoral Potency. Cancer Res. 68: 8928-8937 [Abstract] [Full Text]  
  • Subramanian, T., Vijayalingam, S., Chinnadurai, G. (2006). Genetic Identification of Adenovirus Type 5 Genes That Influence Viral Spread. J. Virol. 80: 2000-2012 [Abstract] [Full Text]  
  • Pierce, M. A., Svetlanov, A., Horwitz, M. S., Serreze, D. V. (2005). Adenovirus Early Region 3 Transgenes Expressed in {beta} Cells Prevent Autoimmune Diabetes in Nonobese Diabetic Mice: Effects of Deleting the Adenovirus Death Protein 11.6K. J. Virol. 79: 619-621 [Abstract] [Full Text]  

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Tollefson, A. E.
Right arrow Articles by Wold, W. S. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Tollefson, A. E.
Right arrow Articles by Wold, W. S. M.

Copyright © 2009 by the American Society for Microbiology.   For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»