Expression of dihydrofolate reductase, and of the adjacent EIb region, in an Ad5-dihydrofolate reductase recombinant virus

A Broad Overview and Review of CRISPR-Cas Technology and Stem Cells

The pinnacle of four decades of research, induced pluripotent stem cells (iPSCs), and genome editing with the advent of clustered, regularly interspaced, short palindromic repeats (CRISPR) now promise to take drug development and regenerative medicine to new levels and to enable the interrogation of disease mechanisms with a hitherto unimaginable level of model fidelity. Autumn 2014 witnessed the first patient receiving iPSCs differentiated into retinal pigmented epithelium to treat macular degeneration. Technologies such as 3D bioprinting may now exploit these advances to manufacture organs in a dish. As enticing as these prospects are, these technologies demand a deeper understanding, which will lead to improvements in their safety and efficacy. For example, precise and more efficient reprogramming for iPSC production is a requisite for wider clinical adoption. Improving awareness of the roles of long non-coding RNAs (lncRNAs) and microRNAs (miRNAs) and genomic epigenetic status will contribute to the achievement of these aims. Similarly, increased efficiency, avoidance of off-target effects, and expansion of available target sequences are critical to the uptake of genome editing technology. In this review, we survey the historical development of genetic manipulation and stem cells. We explore the potential of genetic manipulation of iPSCs for in vitro disease modeling, generation of new animal models, and clinical applicability. We highlight the aspects that define CRISPR-Cas as a breakthrough technology, look at gene correction, and consider some important ethical and societal implications of this approach.

Development of Sendai virus vectors and their potential applications in gene therapy and regenerative medicine

Gene delivery/expression vectors have been used as fundamental technologies in gene therapy since the 1980s. These technologies are also being applied in regenerative medicine as tools to reprogram cell genomes to a pluripotent state and to other cell lineages. Rapid progress in these new research areas and expectations for their translation into clinical applications have facilitated the development of more sophisticated gene delivery/expression technologies. Since its isolation in 1953 in Japan, Sendai virus (SeV) has been widely used as a research tool in cell biology and in industry, but the application of SeV as a recombinant viral vector has been investigated only recently. Recombinant SeV vectors have various unique characteristics, such as low pathogenicity, powerful capacity for gene expression and a wide host range. In addition, the cytoplasmic gene expression mediated by this vector is advantageous for applications, in that chromosomal integration of exogenous genes can be undesirable. In this review, we introduce a brief historical background on the development of recombinant SeV vectors and describe their current applications in gene therapy. We also describe the application of SeV vectors in advanced nuclear reprogramming and introduce a defective and persistent SeV vector (SeVdp) optimized for such reprogramming.

Gene transfer into neural cells in vitro using adenoviral vectors

Adenoviral vectors are excellent vehicles to transfer genes into the nervous system due to their ability to transduce dividing and nondividing cells, their ability to be grown to very high titers, and their relatively large insert capacity. Also, adenoviral vectors can sustain very long-term transgene expression in the CNS of rodents and in neurons and glial cells in culture. Successful gene transfer into the nervous system is dependent on the development, production, and quality control of vector preparations, which need to be of the highest quality. This unit provides protocols to clone, rescue, amplify, and purify first-generation adenoviral vectors. Detailed quality control assays are provided to ensure that vector preparations are devoid of contamination from replication-competent adenovirus and lipopolysaccharides. Also included are methodologies related to adenoviral-mediated gene transfer into neurons and glial cells in culture, and the analysis of transgene expression using immunocytochemistry, enzymatic assays, and fluorescence-activated cell sorting (FACS) analysis.

Wild-type adenoviruses from groups A-F evoke unique innate immune responses, of which HAd3 and SAd23 are partially complement dependent

Alternative human and non-human Ad serotype vectors are currently studied for gene therapy and/or vaccine applications to capitalize upon their likely ability to avoid pre-existing immunity to HAd5. However, relatively little attention has been given to the nature and scope of innate immune responses generated by alternative Ad serotypes. In this study, we characterized several innate immune responses after intravenous administration of wild-type Ad serotypes HAd31, HAd3, HAd5, HAd37, SAd23 and HAd41, representing groups A-F, respectively. Notably, biodistribution studies revealed significant differences between the serotypes, with high levels of HAd3 genomes found in the liver and lung, and HAd37 genomes found in the spleen after systemic administration. Relative to similar treatments with other Ad serotypes, HAd3 and SAd23 induced altered innate immune responses, illustrated by induction of higher levels of cellular gene transcription in several tissues, and higher plasma levels of cytokines and chemokines. We also investigated whether complement interactions have a role in HAd3- and SAd23-induced responses. We confirmed complement dependent gene transcription, plasma cytokine/chemokine responses, and liver toxicities incurred after administration of HAd3 and SAd23. This study highlights the potential benefits and/or limitations to the proposed use of alternative Ad serotypes for gene therapy or vaccine applications.

Enhancement of protein expression in insect cells by a lobster tropomyosin cDNA leader sequence

We describe a cis element that dramatically increases the expression levels of exogenous genes in baculovirus-infected insect cells. This 21 bp sequence element is derived from a 5' untranslated leader sequence of a lobster tropomyosin cDNA (L21). By using a transfer vector carrying L21, the expression levels of tropomyosin and luciferase were 20- and seven-fold higher with L21 than without L21, respectively. L21 has both the Kozak sequence and the A-rich sequence found in the polyhedrin leader sequence. We assume that both sequence elements are essential for the enhancement of protein expression in the baculovirus-based expression system.

Development and characterization of bovine x human hybrid cell lines that efficiently support the replication of both wild-type bovine and human adenoviruses and those with E1 deleted

The 293 cell line that was generated by transforming human embryonic kidney cells with human adenovirus type 5 (HAV5) early region 1 (E1) sequences is an excellent host for generating and growing HAV5 recombinants with E1 deleted, but it does not support the replication of bovine adenovirus type 3 (BAV3). Madin-Darby bovine kidney (MDBK), an established bovine cell line, is an excellent host for growing and plaquing BAV3. For the purpose of combining the unique characteristics of these two cell lines (293 and MDBK), we generated a number of bovine x human hybrid (BHH) cell lines. Comparison of three BHH hybrid clones-BHH3, BHH8, and BHH2C-with 293-Puro (puromycin-resistant 293 cells) and MDBK-Neo (G418-resistant MDBK cells) cell lines for total cellular DNA content, species-specific surface markers, isoenzyme analysis, and karyotyping indicate that they are hybrid in nature. BHH clones constitutively expressed the E1 proteins (E1A, E1B-21kDa, and E1B-55kDa) of HAV5 and efficiently supported the replication of both wild-type and replication-incompetent bovine or human adenoviruses. Transient gene expression experiments with a plasmid encoding the bacterial beta-galactosidase gene demonstrated that BHH cell hybrids seem to have better transfection efficiencies than either of the parental cell lines. These cell lines will be useful for isolating and growing replication-competent human or bovine adenovirus recombinants with E1 deleted and for the study of cellular or viral factors important for viral replication. The development of somatic cell hybrids appears to be a simple way of combining some of the desirable characteristics present separately in two parental cell lines.

Generation of infectious genome of bovine adenovirus type 3 by homologous recombination in bacteria

The widely used technique of generating adenovirus vectors by homologous recombination in mammalian cells is usually not very efficient. This communication describes a simple method of generating a plasmid containing the full-length genome of an adenovirus by homologous recombination in bacteria. Following transfection of a suitable mammalian cell line with the full-length adenovirus genome, infectious virus progeny could easily be generated. Using this technique the generation of adenovirus recombinants would be efficient and straightforward.

Surfactant protein A-producing cells in human fetal lung are good targets for recombinant adenovirus-mediated gene transfer

Local delivery of Escherichia coli beta-galactosidase gene (beta-gal) to surfactant protein-A (SP-A)-producing cells by a replication-defective recombinant adenovirus (AdCMV.beta-gal) was tested in human 8-12-wk-old fetal lung explants cultured in Waymouth's medium. In contrast to uninfected explants, direct addition of 0.8-1.6 x 10(6) plaque-forming units of AdCMV.beta-gal resulted in beta-galactosidase (beta-Gal)-specific staining of the pulmonary epithelium. SP-A localization by indirect immunofluorescence showed positive specific staining of the beta-Gal+ lung epithelial cells, demonstrating that recombinant-defective adenoviruses efficiently transfer reporter genes to fetal lung SP-A+ cells. The reporter gene expression in SPA+ cells persisted for more than 1 mo. No apparent alteration of morphology, phenotype, and growth was observed. The in vitro human lung model described may be useful for testing DNA constructs for vector-mediated gene therapy, as an approach to the treatment of congenital defects and neonatal disorders, such as respiratory distress syndrome and bronchopulmonary dysplasia.

Quantitative determination of adenovirus-mediated gene delivery to rat cardiac myocytes in vitro and in vivo

To optimize the use of modified adenoviruses as vectors for gene delivery to the myocardium, we have characterized infection of cultured fetal and adult rat cardiac myocytes in vitro and of adult cardiac myocytes in vivo by using a replication-defective adenovirus carrying the chloramphenicol acetyltransferase (CAT) reporter gene driven by the cytomegalovirus promoter (AdCMVCATgD). In vitro, virtually all fetal or adult cardiocytes express the CAT gene when infected with 1 plaque-forming unit of virus per cell. CAT enzymatic activity can be detected in these cells as early as 4 hr after infection, reaching near-maximal levels at 48 hr. In fetal cells, CAT expression was maintained without a loss in activity for at least 1 week. Using in vitro studies as a guide, we introduced the AdCMVCATgD virus directly into adult rat myocardium and compared the expression results obtained from virus injection with those obtained by direct injection of pAdCMVCATgD plasmid DNA. The amount of CAT activity resulting from adenovirus infection of the myocardium was orders of magnitude higher than that seen from DNA injection and was proportional to the amount of input virus. Immunostaining for CAT protein in cardiac tissue sections following adenovirus injection demonstrated large numbers of positive cells, reaching nearly 100% of the myocytes in many regions of the heart. Expression of genes introduced by adenovirus peaked at 5 days but was still detectable 55 days following infection. Adenoviruses are therefore a very useful tool for high-efficiency gene transfer into the cardiovascular system.

Monitoring foreign gene expression by a human adenovirus-based vector using the firefly luciferase gene as a reporter

We have constructed a helper-independent adenovirus type 5-luciferase recombinant (Ad5-Luc 3) containing the firefly luciferase gene flanked by simian virus 40 (SV40) regulatory sequences inserted in the early region 3 (E3) of the Ad5 genome. Expression of luciferase in cells infected with Ad5-Luc3 was relatively efficient. In HeLa cells approximately 20 micrograms luciferase per 10(6) cells was made by 36 h post-infection and a 62 kilo-Dalton (kDa) luciferase band was clearly visible in a [35S]methionine-labeled Ad5-Luc 3-infected cell extract analyzed directly by sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE). The results of experiments in which cultured cells were infected with Ad5-Luc 3 in the presence or absence of 1-beta-D-arabinofuranosyl cytosine (AraC) showed that the majority of luciferase expression was dependent on viral DNA replication. This suggested that the enzyme was probably translated primarily from mRNA derived from transcripts expressed from the major late promoter of Ad5. An anti-luciferase antibody was raised in a rabbit and used to further characterize the luciferase expressed in HeLa cells infected with Ad5-Luc 3 by immunoprecipitations and Western blot analyses. The half-life of luciferase expressed in HeLa cells infected with Ad5-Luc 3 was calculated to be approximately 6-8 h by pulse chase analysis. Luciferase is likely to be a useful marker for monitoring virus dissemination and gene expression in experimental animals because assays for enzymatic activity are extremely sensitive and backgrounds are low in all tissues. In mice inoculated intraperitoneally (i.p.) with Ad5-Luc 3, luciferase activity was detected in the liver, spleen, kidney, and lung. A single i.p. inoculation of mice with Ad5-Luc 3 was sufficient to raise anti-luciferase antibody and Ad5 neutralizing antibody which persisted for at least 8 weeks. Even in the presence of circulating anti-luciferase and Ad5 neutralizing antibodies, luciferase activity could be detected in the livers, spleens, and kidneys of mice inoculated i.p. a second time with Ad5-Luc 3.

Nucleocytoplasmic transport: the influenza virus NS1 protein regulates the transport of spliced NS2 mRNA and its precursor NS1 mRNA

Influenza virus unspliced NS1 mRNA, like retroviral pre-mRNAs, is efficiently exported from the nucleus and translated in the cytoplasm of infected cells. With human immunodeficiency virus (HIV), the transport of viral pre-mRNAs is facilitated by the viral Rev protein. We tested the possibility that the influenza virus NS1 protein, a nuclear protein that is encoded by unspliced NS1 mRNA, has the same function as the HIV Rev protein. Surprisingly, using transient transfection assays, we found that rather than facilitating the nucleocytoplasmic transport of unspliced NS1 mRNA, the NS1 protein inhibited the transport of NS2 mRNA, the spliced mRNA generated from NS1 mRNA. The efficient transport of NS2 mRNA from the nucleus to the cytoplasm occurred only when the synthesis of the NS1 protein was abrogated by amber mutations. The NS1 protein down-regulated the export of NS2 mRNA whether or not it was generated by splicing, indicating that the NS1 protein acted directly on transport. Actinomycin D chase experiments verified that the NS1 protein acted on the transport and not on the differential stability of NS2 mRNA in the nucleus as compared to the cytoplasm. In addition, the NS1 protein inhibited the transport of NS1 mRNA itself, which contains all of the sequences in NS2 mRNA, particularly when NS1 mRNA was released from the splicing machinery by mutating its 3'-splice site. Our results indicate that the NS1 protein-mediated inhibition of transport requires sequences in NS2 mRNA. The transport of the viral PB1 protein, nucleocapsid protein, hemagglutinin, membrane protein, and M2 mRNAs was not affected by the NS1 protein. When the NS2 mRNA sequence was covalently attached to the PB1 mRNA, the transport of the chimeric mRNA was inhibited by the NS1 protein. Our results identify a novel function of the influenza virus NS1 protein and demonstrate that post-transcriptional control of gene expression can also occur at the level of the nucleocytoplasmic transport of a mature, spliced mRNA.

Regulation of the extent of splicing of influenza virus NS1 mRNA: role of the rates of splicing and of the nucleocytoplasmic transport of NS1 mRNA

Influenza virus NS1 mRNA is spliced by host nuclear enzymes to form NS2 mRNA, and this splicing is regulated in infected cells such that the steady-state amount of spliced NS2 mRNA is only about 10% of that of unspliced NS1 mRNA. This regulation would be expected to result from a suppression in the rate of splicing coupled with the efficient transport of unspliced NS1 mRNA from the nucleus. To determine whether the rate of splicing of NS1 mRNA was controlled by trans factors in influenza virus-infected cells, the NS1 gene was inserted into an adenovirus vector. The rates of splicing of NS1 mRNA in cells infected with this vector and in influenza virus-infected cells were measured by pulse-labeling with [3H]uridine. The rates of splicing of NS1 mRNA in the two systems were not significantly different, strongly suggesting that the rate of splicing of NS1 mRNA in influenza virus-infected cells is controlled solely by cis-acting sequences in NS1 mRNA itself. In contrast to the rate of splicing, the extent of splicing of NS1 mRNA in the cells infected by the adenovirus recombinant was dramatically increased relative to that occurring in influenza virus-infected cells. This could be attributed largely, if not totally, to a block in the nucleocytoplasmic transport of unspliced NS1 mRNA in the recombinant-infected cells. Most of the unspliced NS1 mRNA was in the nuclear fraction, and no detectable NS1 protein was synthesized. When the 3' splice site of NS1 mRNA was inactivated by mutation, NS1 mRNA was transported and translated, indicating that the transport block occurred because NS1 rRNA was committed to the splicing pathway. This transport block is apparently obviated in influenza virus-infected cells. These experiments demonstrate the important role of the nucleocytoplasmic transport of unspliced NS1 mRNA in regulating the extent of splicing of NS1 mRNA.

Regulation of the extent of splicing of influenza virus NS1 mRNA: role of the rates of splicing and of the nucleocytoplasmic transport of NS1 mRNA

Influenza virus NS1 mRNA is spliced by host nuclear enzymes to form NS2 mRNA, and this splicing is regulated in infected cells such that the steady-state amount of spliced NS2 mRNA is only about 10% of that of unspliced NS1 mRNA. This regulation would be expected to result from a suppression in the rate of splicing coupled with the efficient transport of unspliced NS1 mRNA from the nucleus. To determine whether the rate of splicing of NS1 mRNA was controlled by trans factors in influenza virus-infected cells, the NS1 gene was inserted into an adenovirus vector. The rates of splicing of NS1 mRNA in cells infected with this vector and in influenza virus-infected cells were measured by pulse-labeling with [3H]uridine. The rates of splicing of NS1 mRNA in the two systems were not significantly different, strongly suggesting that the rate of splicing of NS1 mRNA in influenza virus-infected cells is controlled solely by cis-acting sequences in NS1 mRNA itself. In contrast to the rate of splicing, the extent of splicing of NS1 mRNA in the cells infected by the adenovirus recombinant was dramatically increased relative to that occurring in influenza virus-infected cells. This could be attributed largely, if not totally, to a block in the nucleocytoplasmic transport of unspliced NS1 mRNA in the recombinant-infected cells. Most of the unspliced NS1 mRNA was in the nuclear fraction, and no detectable NS1 protein was synthesized. When the 3' splice site of NS1 mRNA was inactivated by mutation, NS1 mRNA was transported and translated, indicating that the transport block occurred because NS1 rRNA was committed to the splicing pathway. This transport block is apparently obviated in influenza virus-infected cells. These experiments demonstrate the important role of the nucleocytoplasmic transport of unspliced NS1 mRNA in regulating the extent of splicing of NS1 mRNA.

The adenovirus L4 100-kilodalton protein is necessary for efficient translation of viral late mRNA species

When screening a number of adenovirus type 5 (Ad5) temperature-sensitive mutants for defects in viral gene expression, we observed that H5ts1-infected 293 cells accumulated reduced levels of newly synthesized viral late proteins. Pulse-labeling and pulse-chase experiments were used to establish that the late proteins synthesized in H5ts1-infected cells under nonpermissive conditions were as stable as those made in Ad5-infected cells. H5ts1-infected cells contained normal levels of viral late mRNAs. Because these observations implied that translation of viral mRNA species was defective in mutant virus-infected cells, the association of viral late mRNAs with polyribosomes was examined during the late phase of infection at a nonpermissive temperature. In Ad5-infected cells, the majority of the viral L2, L3, L4, pIX, and IVa2 late mRNA species were polyribosome bound. By contrast, these same mRNA species were recovered from H5ts1-infected cells in fractions nearer the top of polyribosome gradients, suggesting that initiation of translation was impaired. During the late phase of infection, neither the polyribosome association nor the translation of most viral early mRNA species was affected by the H5ts1 mutation. This lesion, mapped by marker rescue to the L4 100-kilodalton (kDa) nonstructural protein, has been identified as a single base pair substitution that replaces Ser-466 of the Ad5 100-kDa protein with Pro. A set of temperature-independent revertants of H5ts1 was isolated and characterized. Either true reversion of the H5ts1 mutation or second-site mutation of Pro-466 of the H5ts1 100-kDa protein to Thre, Leu, or His restored both temperature-independent growth and the efficient synthesis of viral late proteins. We therefore conclude that the Ad5 L4 100-kDa protein is necessary for efficient initiation of translation of viral late mRNA species during the late phase of infection.

Quenching of the amidolytic activity of one-chain tissue-type plasminogen activator by mutation of lysine-416

In contrast to most other serine proteases, tissue-type plasminogen activator (t-PA) possesses enzymatic activity as the one-chain zymogen form. The hypothesis that lysine residues 277 or 416 may be involved in stabilization of an active conformation of one-chain t-PA via salt-bridge formation with aspartic acid residue 477 was tested by site-directed mutagenesis. Four recombinant t-PA mutants were constructed. The amidolytic activities of these analogues were compared to that of authentic t-PA. Substitution of arginine-275 provided an analogue [( R275G]t-PA) resistant to plasmin cleavage. The amidolytic activity of [R275G]t-PA was comparable to that of authentic one-chain t-PA, and so was the activity of [R275L,K277L]t-PA, in which additional substitution of lysine residue 277 was carried out. This suggested that its presence was nonessential for obtaining one-chain t-PA activity. In contrast, substitution of lysine residue 416 to obtain [K416S]t-PA and [K416S,H417T]t-PA resulted in substantial quenching of amidolytic one-chain activity. As expected, the amidolytic activities of the two-chain forms were less affected by the substitution. Involvement of lysine residue 416 in one-chain t-PA activity was also indicated by decreased activities of [K416S]t-PA and [K416S,H417T]t-PA with plasminogen as the substrate. The one-chain activity of the lysine residue 416 substitution analogues was partially restored in the presence of fibrin. This could indicate that strong ligands such as fibrin might provide an alternative stabilization of the active conformation of one-chain t-PA.

Protection of mice against lethal challenge with herpes simplex virus by vaccination with an adenovirus vector expressing HSV glycoprotein B

Increasing attention has been focused on the use of recombinant mammalian viruses as potential vaccines. Recombinant human adenoviruses are one of the more promising vaccine vectors because they can be easily constructed and because live adenovirus vaccines have been administered orally to large numbers of military recruits without adverse reactions. In order to examine the efficacy of human adenoviruses as vaccines we have studied the immunity induced by a recombinant adenovirus vector, AdgB2, which induces high level expression of herpes simplex virus (HSV) glycoprotein B (gB) in human and murine cells. Mice inoculated with AdgB2 produced antibodies specific for gB which neutralized HSV in the presence of complement. Although mice inoculated with AdgB2 showed no ill-effects after AdgB2 inoculation and we were unable to detect replication of human adenoviruses in mice, the mice were protected from a lethal challenge with HSV after a single inoculation with AdgB2.

Regulation of poly(A) site selection in adenovirus

We have investigated the mechanisms involved in the early-to-late RNA-processing switch which regulates the mRNA species generated from the adenovirus major late transcription unit (MLTU). In particular, polyadenylation choice mechanisms were characterized by using a reconstructed adenovirus E1A gene as a site for insertion of MLTU poly(A) regulation signals (L1 and L3). Adenovirus constructs containing the variant poly(A) recognition elements were used to compare E1A poly(A) signal utilization with wild-type MLTU (L1 to L5) utilization. In both early and late stages of infection, either polyadenylation site (L1 or L3) is capable of being utilized when presented as the only operational poly(A) site. In an early infection, a virus which contains multiple elements presented in tandem (L13) uses the first poly(A) site, L1, preferentially (ratio of L1 to L3, 8:1) in both E1A and MLTU loci. Transcription termination is not involved in restricting the utilization of the downstream L3 site. In a late infection, when each of the five MLTU poly(A) sites is used, a switch also occurs for the E1AL13 construct, with utilization of both the L1 and L3 poly(A) sites. The switch from early to late was not the result of altered processing factors in the late infection, as demonstrated by superinfecting the E1AL13 construct into cells which had already entered a late stage of infection. The superinfecting virus gave an L1-only phenotype; therefore, a cis mechanism is involved in adenovirus poly(A) regulation.

Expression of bicistronic measles virus P/C mRNA by using hybrid adenoviruses: levels of C protein synthesized in vivo are unaffected by the presence or absence of the upstream P initiator codon

The measles virus (MV) P/C mRNA is functionally bicistronic. Translation is presumed to initiate at both the first and second 5'-proximal AUG codons, leading, respectively, to synthesis of the P and C polypeptides from different overlapping reading frames. To study the function and differential expression of these polypeptides, we have constructed hybrid human adenoviruses capable of expressing high levels of P and C together or of C alone. Cloned cDNA corresponding to the MV P/C gene was coupled to the adenovirus type 2 (Ad2) major late promoter, most of the Ad2 tripartite leader sequence, and the simian virus 40 3'-end processing signal and then used to replace most of the E1a-E1b region of the Ad5 genome in two hybrid adenoviruses: one (Ad5MV/PC13) which contained both 5'-proximal AUG codons of the P/C mRNA and another (Ad5MV/C3) which retained only the second. The sequence context for the P protein initiator AUG codon in Ad5MV/PC13 was made more favorable (GAGAUGG) than the relatively unfavorable context (CCGAUGG) seen in the native MV P/C mRNA. After infection of 293 cells (which provide complementary E1a-E1b functions), both viruses directed equal amounts of P/C-specific mRNA transcription. Ad5MV/PC13 directed the synthesis of both P and C proteins, while Ad5MV/C3 directed the synthesis of C protein alone. Ad5-expressed P protein was phosphorylated, while C was not. C protein had a similar diffuse cytoplasmic localization in both MV and Ad5-infected cells. Ad5MV/C3 and Ad5MV/PC13 directed equal amounts of C protein expression in 293 cells at a level approximately 15 times greater than that seen in MV-infected cells. Thus the level of C protein expression was unaffected by the presence or absence of an out-of-frame upstream AUG codon in a favorable sequence context. This observation cannot be explained by the scanning model for ribosomal initiation and suggests that ribosomes may be binding directly at an internal mRNA site at or near the initiator AUG codon for the C protein.

Abundant expression of herpes simplex virus glycoprotein gB using an adenovirus vector

Herpes simplex virus type 1 (HSV-1) glycoprotein B (gB) is a major component of infected cell membranes and virion envelopes. Glycoprotein B is known to be essential for entry of viruses into cells and may play important roles in virus-induced cell fusion and other alterations in cell morphology. In order to study the biochemical and immunological properties of gB in isolation from other HSV-1 polypeptides we have constructed human adenovirus vectors capable of expressing high levels of gB. The gB gene was coupled to the SV40 early promoter and inserted into the E3 region of two adenovirus vectors, one in which the E1 region was deleted (AdgB-1) and another which contained E1 sequences (AdgB-2). In AdgB-1 the orientation of the chimeric gB-SV40 gene was right to left, i.e., opposite to the direction of late and E3 mRNA transcription, whereas in AdgB-2 the orientation was left to right. Human 293 cells which express E1 functions supported replication of AdgB-1 and gB was expressed in these cells but not in mouse cells and only at very low levels in human cells other than 293. Replication of AdgB-2 was not limited to 293 cells and the virus was able to induce synthesis of gB at levels equal to or higher than those expressed in HSV-1-infected human or mouse cells. Microscopic examination of AdgB-2-infected cells revealed extensive vacuolization in a manner completely uncharacteristic of adenovirus-infected cells, and fluorescent antibody staining indicated that gB was not only present at the cell surface but also concentrated in the cytoplasmic vacuoles.

Optimizing gene expression in BPV-transformed cells: effects of cell type on enhancer/promoter interaction

We have compared several combinations of enhancers and promoters in expressing the chloramphenicol acetyl transferase gene in transient assays, in mouse C127, the most widely used host cell for the bovine papilloma virus (BPV) expression vector. Of the various combinations tested, the unit comprised of the SV40 enhancer and adenovirus type 2 major late promoter (MLP) was the most active in BPV transformed C127 cells. We further demonstrate that untransformed and BPV transformed C127 cells respond differently to the various enhancer/promoter combinations tested. Moving the SV40 enhancer closer to the cap site of a complete MLP (from -414 to -107) reduced potentiation to less than half in BPV transformed cells. The level of potentiation with enhancer at either site was similar in human HeLa cells. In BPV transformed C127 cells, the SV40 enhancer and the MLP (at the -414 site) supports 4-5 fold greater levels of expression than the murine sarcoma virus (MSV) enhancer/mouse metallothionein (MT) promoter which has previously been extremely effective in BPV vectors. Our findings provide a basis for the improvement of the BPV vector system in supporting increased levels of expression of proteins of important therapeutic application.

At least six nucleotides preceding the AUG initiator codon enhance translation in mammalian cells

Sequences flanking the AUG initiator codon influence its recognition by eukaryotic ribosomes. From a comparison of several hundred mRNA sequences, CCA/GCCAUGG emerged as the consensus sequence for initiation in higher eukaryotes. Systematic mutagenesis of a cloned preproinsulin gene confirmed the facilitating effect of A or G in position -3 (i.e. 3 nucleotides upstream from the AUG codon), C in positions -1 and -2, and G immediately following the AUG codon. The analysis of a new set of mutants now reveals that sequences slightly farther upstream are also influential, the optimal context for initiation being (GCC)GCCA/GCCAUGG. Possible mechanistic implications of the repeating GCC motif are discussed.

Expression of human pancreatic polypeptide precursors from a dicistronic mRNA in mammalian cells

The ability of eukaryotic ribosomes to reinitiate translation at downstream AUG codons on polycistronic mRNAs was used to select transfected CHO clones that secreted a precursor to the human pancreatic polypeptide (PP). In the in vitro constructed transcription unit, a viral promoter directed the synthesis of a dicistronic mRNA. The PP cDNA was placed in the 5'-part of this transcript, while a DHFR cDNA was placed 3' to the PP. This dicistronic expression unit was transfected into CHO cells, and methotrexate-resistant colonies were selected. RNA-blots verified that the PP precursor and DHFR were expressed from the same dicistronic mRNA. The CHO cells synthesized the hormone precursor and secreted it through the constitutive secretory pathway.

Overproduction of polyomavirus middle T antigen in mammalian cells through the use of an adenovirus vector

To overproduce biologically active polyomavirus middle T antigen, we used an adenovirus vector and human 293 cells as hosts. Two helper-independent recombinant adenoviruses were isolated that contain a hybrid transcription unit, in differing orientations, at a site in the adenovirus genome from which the E1a and most of the E1b transcription units have been deleted. The hybrid transcription unit consists of the adenovirus type 2 major late promoter and tripartite leader and a cDNA segment capable of encoding polyomavirus middle T antigen and accompanying 3' RNA-processing signals. Both recombinant viruses were stable and replicated to high titers in human 293 cells. The polyomavirus sequences were expressed, predominantly at late times after infection of 293 cells, to yield mRNAs that encoded middle T antigen. One of the recombinant viruses also expressed a middle T antigen-related protein in 293 cells. The latter was translated from one of several novel mRNA species that resulted from aberrant splicing and incomplete RNA processing of precursor RNA transcripts. Comparison of the amount of middle T antigen produced in 3T6 cells infected with polyomavirus with that in 293 cells infected with either of the recombinant adenoviruses, under optimal conditions for each system, revealed at least a 10-fold greater yield of the protein on a per-cell basis in the latter system than in the former. The recombinant-virus-encoded middle T antigen was biologically active, as evidenced by its ability to associate with and serve as a substrate for human pp60c-src. The functionality of the middle T antigen was further confirmed by demonstrating that both recombinant viruses efficiently transformed Rat-1 cells. These recombinant viruses will be useful to overproduce middle T antigen and to introduce the polyomavirus oncogene into a wide variety of mammalian cells.

Abundant expression of polyomavirus middle T antigen and dihydrofolate reductase in an adenovirus recombinant

A modular gene with a cDNA encoding the polyomavirus middle T antigen positioned behind the adenovirus type 2 major late promoter and tripartite leader was substituted for the E1a region in an adenovirus vector. Permissive human cells infected with this recombinant produce middle T protein at levels as high as those of the most abundant late adenoviral proteins, e.g., hexon or fiber. This level represents at least a 40-fold increase over that observed in a polyomavirus lytic infection of murine cells. Partial proteolytic mapping showed that this protein has the same primary structure as middle T protein produced in polyomavirus-infected murine cells. The adenovirus recombinant-generated middle T protein exhibited in vitro kinase activity, although at an approximately 10-fold-lower specific activity than that of middle T protein from polyomavirus-infected murine cells. Comparison of the expression levels of this middle T antigen-containing adenovirus vector with a similar construction encoding dihydrofolate reductase suggested that the translation efficiency of the inserted gene was dependent upon the proximity of its initiation codon to the tripartite leader. We tested this possibility by comparing three dihydrofolate reductase recombinants among which the spacing between the initiation codon and tripartite leader varied from 188 to 36 nucleotides. The efficiency of expression of dihydrofolate reductase protein dramatically increased as this spacing was reduced.

Cell-type-specific synthesis of murine immunoglobulin mu RNA from an adenovirus vector

The mouse immunoglobulin heavy-chain mu constant region gene was cloned into the early region 1B of an adenovirus type 5 vector to allow reproducible kinetics of expression of the mu gene in the presence of continuous host protein synthesis after infection by the recombinant. The immunoglobulin-adenovirus recombinant is helper independent in infecting human fibroblastic and B- and T-cell lines and expresses mu in a cell-type-specific manner. By Northern blot analysis, correctly polyadenylated and spliced E1B-mu S and E1B-mu m mRNAs are found to be equally abundant at steady state in fibroblasts. In contrast, and appropriately, only E1B-mu S mRNAs accumulate in a lambda light-chain-secreting myeloma cell line. Analysis of nascent transcripts pulse labeled in isolated nuclei demonstrates equimolar polymerase loading throughout the mu region in all cell types infected by mu-Ad. Thus, correct polyadenylation and splicing of E1B-mu S and E1B-mu m in fibroblasts does not require transcription termination in the region separating the mu S and mu m polyadenylation sites. Furthermore, differential expression of mu transcripts in the background of myeloma cells is regulated at the level of RNA processing and does not require the presence of the immunoglobulin heavy-chain enhancer or promoter element.

Construction of a helper-free recombinant adenovirus that expresses polyomavirus large T antigen

Adenovirus-polyomavirus recombinant viruses were constructed in vitro by inserting a hybrid transcription unit composed of the adenovirus type 2 major late promoter and the early coding region of polyomavirus into the adenovirus type 5 vector Ad5 delta E1/dl309. The vector lacks the E1a and E1b transcription units and contains a unique restriction endonuclease cleavage site in their place. The polyomavirus genomic insert contained a small deletion which precluded the synthesis of functional small and middle T antigen but allowed for the synthesis of large T antigen. One recombinant virus, Ad5PyR39, which contained the hybrid transcription unit in the opposite transcriptional orientation from the overall direction of late-gene transcription, was studied in detail. Ad5PyR39 replicated efficiently without a helper virus in human 293 cells and expressed hybrid mRNAs of the expected size and composition that were translated to yield large T antigen. The large T antigen synthesized in 293 cells was the same size as that produced in mouse 3T6 cells lytically infected with polyomavirus, and this protein bound efficiently and specifically to the large-T-antigen-binding sites in polyomavirus DNA. Moreover, the large T antigen encoded by the recombinant virus proved capable of catalyzing the replication in mouse 3T6 cells of a plasmid containing the polyomavirus origin for DNA replication. Comparison of the amount of large T antigen produced in 3T6 cells infected with polyomavirus with that in 293 cells infected with Ad5PyR39, under optimal conditions for each system, revealed at least a fivefold greater yield of the protein on a per cell basis in the latter system compared with the former. Ad5PyR39 should prove to be useful to isolate large quantities of functional polyomavirus large T antigen for structural and biochemical studies.

In vivo identification of sequence elements required for normal function of the adenovirus major late transcriptional control region

A series of adenovirus type 5 variants were constructed to identify the sequence elements which comprise the major late transcriptional control region in the context of the viral chromosome. The variant chromosomes carried a second copy of DNA sequence derived from the region surrounding the major late mRNA cap site. The reiterated segments replaced the normal transcriptional control region of the E1A gene. By monitoring the rate of E1A transcription subsequent to infection with the variants, it was possible to evaluate the capabilities of the substituted major late elements. A segment derived from -55 to +33 (relative to the major late cap site at +1) functioned for early transcription, in the presence of the E1A enhancer domain, but failed to direct enhanced levels of activity late after infection. A segment from -122 to +33 directed both early and enhanced late transcription. The rate of late E1A transcription directed by this element was about 40% of that displayed by the major late control region at its normal position. Inclusion of additional upstream sequences (to -565) did not increase late transcription rates. Thus, function of the adenovirus major late control region.

Construction of a helper-free recombinant adenovirus that expresses polyomavirus large T antigen

Adenovirus-polyomavirus recombinant viruses were constructed in vitro by inserting a hybrid transcription unit composed of the adenovirus type 2 major late promoter and the early coding region of polyomavirus into the adenovirus type 5 vector Ad5 delta E1/dl309. The vector lacks the E1a and E1b transcription units and contains a unique restriction endonuclease cleavage site in their place. The polyomavirus genomic insert contained a small deletion which precluded the synthesis of functional small and middle T antigen but allowed for the synthesis of large T antigen. One recombinant virus, Ad5PyR39, which contained the hybrid transcription unit in the opposite transcriptional orientation from the overall direction of late-gene transcription, was studied in detail. Ad5PyR39 replicated efficiently without a helper virus in human 293 cells and expressed hybrid mRNAs of the expected size and composition that were translated to yield large T antigen. The large T antigen synthesized in 293 cells was the same size as that produced in mouse 3T6 cells lytically infected with polyomavirus, and this protein bound efficiently and specifically to the large-T-antigen-binding sites in polyomavirus DNA. Moreover, the large T antigen encoded by the recombinant virus proved capable of catalyzing the replication in mouse 3T6 cells of a plasmid containing the polyomavirus origin for DNA replication. Comparison of the amount of large T antigen produced in 3T6 cells infected with polyomavirus with that in 293 cells infected with Ad5PyR39, under optimal conditions for each system, revealed at least a fivefold greater yield of the protein on a per cell basis in the latter system compared with the former. Ad5PyR39 should prove to be useful to isolate large quantities of functional polyomavirus large T antigen for structural and biochemical studies.

Cell-type-specific synthesis of murine immunoglobulin mu RNA from an adenovirus vector

The mouse immunoglobulin heavy-chain mu constant region gene was cloned into the early region 1B of an adenovirus type 5 vector to allow reproducible kinetics of expression of the mu gene in the presence of continuous host protein synthesis after infection by the recombinant. The immunoglobulin-adenovirus recombinant is helper independent in infecting human fibroblastic and B- and T-cell lines and expresses mu in a cell-type-specific manner. By Northern blot analysis, correctly polyadenylated and spliced E1B-mu S and E1B-mu m mRNAs are found to be equally abundant at steady state in fibroblasts. In contrast, and appropriately, only E1B-mu S mRNAs accumulate in a lambda light-chain-secreting myeloma cell line. Analysis of nascent transcripts pulse labeled in isolated nuclei demonstrates equimolar polymerase loading throughout the mu region in all cell types infected by mu-Ad. Thus, correct polyadenylation and splicing of E1B-mu S and E1B-mu m in fibroblasts does not require transcription termination in the region separating the mu S and mu m polyadenylation sites. Furthermore, differential expression of mu transcripts in the background of myeloma cells is regulated at the level of RNA processing and does not require the presence of the immunoglobulin heavy-chain enhancer or promoter element.

Structural requirements of adenovirus VAI RNA for its translation enhancement function

Recently, by genetic and biochemical approaches, it has been shown that adenovirus VAI RNA is required for efficient translation of viral mRNAs at late times after infection. To understand the nucleotide sequences and the domains of the VAI RNA that are responsible for the role of VAI RNA in enhancement of translation, a mutational analysis of the VAI gene was undertaken. Deletion, substitution, and insertion mutations covering most of the nucleotide sequences of VAI RNA were introduced into the VAI gene at the plasmid level. These mutant genes were then reintroduced into the virus, and growth properties of the mutant viruses were studied. The majority of the mutants retained normal or nearly normal levels of biological function. Mutations in the region between +43 and +53 and between +107 and the 3' end of the gene resulted in a considerable loss of activity. These mutants, however, grew significantly better than did an adenovirus type 5 mutant lacking both functional VAI and VAII genes, indicating that they retain a portion of their activity. Because no one mutation was able to completely abolish the function, we suggest that the VAI RNA may have multiple functional sites for its translation modulation function. These multiple sites may be short oligonucleotide sequences that may interact with cellular or viral components or both during translation.

Effect of the tripartite leader on synthesis of a non-viral protein in an adenovirus 5 recombinant

The EIa region of an Adenovirus 5 recombinant has been substituted by a modular gene encoding dihydrofolate reductase (DHFR). In this recombinant, the mouse DHFR cDNA was positioned behind sequences of the major late promoter and the complete tripartite leader. The leader sequences end in the normal 5' splice site (SS) of the third leader, so that RNA splicing joins the tripartite leader to a 3' splice site immediately upstream of the DHFR cDNA. At late stages of infection, high levels of DHFR mRNAs were synthesized. At early times in the late stage, this mRNA was efficiently translated; however, at later times translation of DHFR decreased probably due to poor competition with other late mRNAs. Synthesis of DHFR protein from an analogous Adenovirus 5 recombinant containing only the first late leader was studied in parallel. Equivalent levels of DHFR mRNA were expressed after infection with this recombinant virus; however, the efficiency of DHFR translation was at least 20 fold lower than that of the DHFR mRNA containing the tripartite leader. This suggests that the tripartite leader sequence is important for translation in the late stage of infection. As reported previously, the Ad5 recombinant containing only the first leader vastly overexpresses polypeptide IX from a novel mRNA, formed by the splicing of the first leader in the modular DHFR gene to the 3' splice site in the EIb region. Cells infected with this recombinant synthesize very little normal mRNA from the EIb region. Here, we demonstrated that coinfection of 293 cells with this recombinant and wild type Adenovirus 5 also results in decreased EIb mRNA synthesis. We propose that the overproduction of polypeptide IX suppresses mRNA expression from the EIb and IX promoter sites, probably by an autoregulation loop active during lytic growth.

Identification of the components necessary for adenovirus translational control and their utilization in cDNA expression vectors

A transient expression system was used to study the role of the adenovirus late and simian virus 40 (SV40) early mRNA leader sequences and adenovirus virus-associated (VA) RNAs in mRNA translation. Hybrid transcription units containing the adenovirus late and SV40 early promoters fused to various coding regions were introduced into monkey COS cells on plasmids containing a SV40 origin of replication. The translational efficiencies of the mRNAs produced from these plasmids were determined after alterations in the viral leader sequences or in the presence of VA RNAs provided by adenovirus infection of the transfected cells or by cotransfection with plasmids containing the VA genes. Efficient translation of mRNA with either adenovirus or SV40 leader sequences is dependent upon the presence of VA RNA. Translational stimulation by VA RNA of mRNAs containing the adenovirus tripartite leader sequences is dramatically reduced if leader exons 2 and 3 are removed or if their orientation is altered. Sequence analysis has indicated a homology between the nontranslated 5' end of SV40 early mRNA and sequences at the border of the 2nd and 3rd tripartite leader exons, which may be responsible for the increased translation of these mRNAs in the presence of VA RNA.

Structural requirements of adenovirus VAI RNA for its translation enhancement function

Recently, by genetic and biochemical approaches, it has been shown that adenovirus VAI RNA is required for efficient translation of viral mRNAs at late times after infection. To understand the nucleotide sequences and the domains of the VAI RNA that are responsible for the role of VAI RNA in enhancement of translation, a mutational analysis of the VAI gene was undertaken. Deletion, substitution, and insertion mutations covering most of the nucleotide sequences of VAI RNA were introduced into the VAI gene at the plasmid level. These mutant genes were then reintroduced into the virus, and growth properties of the mutant viruses were studied. The majority of the mutants retained normal or nearly normal levels of biological function. Mutations in the region between +43 and +53 and between +107 and the 3' end of the gene resulted in a considerable loss of activity. These mutants, however, grew significantly better than did an adenovirus type 5 mutant lacking both functional VAI and VAII genes, indicating that they retain a portion of their activity. Because no one mutation was able to completely abolish the function, we suggest that the VAI RNA may have multiple functional sites for its translation modulation function. These multiple sites may be short oligonucleotide sequences that may interact with cellular or viral components or both during translation.