Improved antibiotic-free plasmid vector design by incorporation of transient expression enhancers.

Methods to improve plasmid-mediated transgene expression are needed for gene medicine and gene vaccination applications. To maintain a low risk of insertional mutagenesis-mediated gene activation, expression-augmenting sequences would ideally function to improve transgene expression from transiently transfected intact plasmid, but not from spurious genomically integrated vectors. We report herein the development of potent minimal, antibiotic-free, high-manufacturing-yield mammalian expression vectors incorporating rationally designed additive combinations of expression enhancers. The SV40 72 bp enhancer incorporated upstream of the cytomegalovirus (CMV) enhancer selectively improved extrachromosomal transgene expression. The human T-lymphotropic virus type I (HTLV-I) R region, incorporated downstream of the CMV promoter, dramatically increased mRNA translation efficiency, but not overall mRNA levels, after transient transfection. A similar mRNA translation efficiency increase was observed with plasmid vectors incorporating and expressing the protein kinase R-inhibiting adenoviral viral associated (VA)1 RNA. Strikingly, HTLV-I R and VA1 did not increase transgene expression or mRNA translation efficiency from plasmid DNA after genomic integration. The vector platform, when combined with electroporation delivery, further increased transgene expression and improved HIV-1 gp120 DNA vaccine-induced neutralizing antibody titers in rabbits. These antibiotic-free vectors incorporating transient expression enhancers are safer, more potent alternatives to improve transgene expression for DNA therapy or vaccination.

The expression of slow myosin during mammalian somitogenesis and limb bud differentiation.

The developmental pattern of slow myosin expression has been studied in mouse embryos from the somitic stage to the period of secondary fiber formation and in myogenic cells, cultured from the same developmental stages. The results obtained, using a combination of different polyclonal and monoclonal antibodies, indicate that slow myosin is coexpressed in virtually all the cells that express embryonic (fast) myosin in somites and limb buds in vivo as well as in culture. On the contrary fetal or late myoblasts (from 15-d-old embryos) express in culture only embryonic (fast) myosin. At this stage, muscle cells in vivo, as already shown (Crow, M.T., and F.A. Stockdale. 1986. Dev. Biol. 113:238-254; Dhoot, G.K. 1986. Muscle & Nerve. 9:155-164; Draeger, A., A.G. Weeds, and R.B. Fitzsimons. 1987. J. Neurol. Sci. 81:19-43; Miller, J.B., and F.A. Stockdale. 1986. J. Cell Biol. 103:2197-2208), consist of primary myotubes, which express both myosins, and secondary myotubes, which express preferentially embryonic (fast) myosin. Under no circumstance neonatal or adult fast myosins were detected. Western blot analysis confirmed the immunocytochemical data. These results suggest that embryonic myoblasts in mammals are all committed to the mixed embryonic-(fast) slow lineage and, accordingly, all primary fibers express both myosins, whereas fetal myoblasts mostly belong to the embryonic (fast) lineage and likely generate fibers containing only embryonic (fast) myosin. The relationship with current models of avian myogenesis are discussed.

Hereditary pituitary dwarfism in mice affects skeletal and cardiac myosin isozyme transitions differently.

The dwarf mutation in mice interferes with the development of those anterior pituitary cells responsible for production of thyroid stimulating hormone, growth hormone, and prolactin. Myosin isozyme transitions in both cardiac and skeletal muscle were also found to be affected in this mutant. Electrophoresis of native myosins demonstrated that the fetal (V3) to adult (V1) ventricular cardiac isozyme transition was completely blocked in dwarf mice; in contrast, the neonatal to adult fast myosin transition in hind limb skeletal muscle was slowed but not totally inhibited. The persistence of neonatal myosin heavy chain for up to 55-75 d after birth in dwarf mice, as compared with 16 d in normal mice, was directly demonstrated by polypeptide and immunopolypeptide mapping. Morphological examination of 18-36-d-old dwarf skeletal muscles by optical and electron microscopy revealed a relative immaturity, but no signs of gross pathology were evident. Immunocytochemical analysis showed that the abnormal persistence of neonatal myosin occurs in most of the fibers. Multiple injections of thyroxine restored a normal isozyme complement to both cardiac and skeletal muscles within 11-15 d. Therefore, the effects of the dwarf mutation on myosin isozymes can be explained by the lack of thyroid hormone in these animals. Because the synthesis of growth hormone is not stimulated by thyroid hormone in dwarf mice as it would be in normal animals, these results demonstrate that thyroid hormone promotes myosin isozyme transitions independent of growth hormone production.

Neonatal and adult myosin heavy chain isoforms in a nerve-muscle culture system.

When adult mouse muscle fibers are co-cultured with embryonic mouse spinal cord, the muscle regenerates to form myotubes that develop cross-striations and contractions. We have investigated the myosin heavy chain (MHC) isoforms present in these cultures using polyclonal antibodies to the neonatal, adult fast, and slow MHC isoforms of rat (all of which were shown to react specifically with the analogous mouse isoforms) in an immunocytochemical assay. The adult fast MHC was absent in newly formed myotubes but was found at later times, although it was absent when the myotubes myotubes were cultured without spinal cord tissue. When nerve-induced muscle contractions were blocked by the continuous presence of alpha-bungarotoxin, there was no decrease in the proportion of fibers that contained adult fast MHC. Neonatal and slow MHC were found at all times in culture, even in the absence of the spinal cord, and so their expression was not thought to be nerve-dependent. Thus, in this culture system, the expression of adult fast MHC required the presence of the spinal cord, but was probably not dependent upon nerve-induced contractile activity in the muscle fibers.

Co-expression of multiple myosin heavy chain genes, in addition to a tissue-specific one, in extraocular musculature.

We have investigated the developmental transitions of myosin heavy chain (MHC) gene expression in the rat extraocular musculature (EOM) at the mRNA level using S1-nuclease mapping techniques and at the protein level by polypeptide mapping and immunochemistry. We have isolated a genomic clone, designated lambda 10B3, corresponding to an MHC gene which is expressed in the EOM fibers (recti and oblique muscles) of the adult rat but not in hind limb muscles. Using cDNA and genomic probes for MHC genes expressed in skeletal (embryonic, neonatal, fast oxidative, fast glycolytic, and slow/cardiac beta-MHC), cardiac (alpha-MHC), and EOM (lambda 10B3) muscles, we demonstrate the concomitant expression at the mRNA level of at least six different MHC genes in adult EOM. Protein and immunochemical analyses confirm the presence of at least four different MHC types in EOM. Immunocytochemistry demonstrates that different myosin isozymes tend to segregate into individual myofibers, although some fibers seem to contain more than one MHC type. The results also show that the EOM fibers exhibit multiple patterns of MHC gene regulation. One set of fibers undergoes a sequence of isoform transitions similar to the one described for limb skeletal muscles, whereas other EOM myofiber populations arrest the MHC transition at the embryonic, neonatal/adult, or adult EOM-specific stage. Thus, the MHC gene family is not under the control of a strict developmental clock, but the individual genes can modify their expression by tissue-specific and/or environmental factors.

The translation of globin messenger RNA with use of muscle initiation factors.

The ability of embryonic chicken muscle initiation factors to translate rabbit globin messenger RNA in an efficient, fractionated cell-free system has been examined. Although muscle factors stimulate leucine incorporation to only 15--35% the levels achieved with rabbit reticulocyte initiation factors, they synthesize more than one globin chain per mRNA molecule and both alpha and beta globin are produced. Increasing the ribosome concentration and adding the polyamine spermidine to the system produce stimulatory effects which are quantitatively and qualitatively similar for both factor preparations. The lower efficiency of synthesis of muscle factors relative to reticulocyte factors is also apparent when mRNA from encephalomyocarditis virus or embryonic chicken muscle polysomes are used in the cell-free system. These results do not support a specific restriction in the capacity of muscle factors to translate globin mRNA. Furthermore, the similarity of the effects of presumed non-specific components on the activity of muscle and reticulocyte factors suggests that globin synthesis in the cell-free system may be controlled in a similar fashion for both preparations.

The majority of calf muscle cell messenger RNAs contain poly(A).

Previous studies from our laboratory have investigated messenger RNA metabolism in calf muscle cells in tissue culture. The analysis of mRNA was based on its poly(A) content. We have now examined directly the proportion of mRNA which contains poly(A) in these cells. After separation of poly(A)+ -and poly(A) - -RNA on oligo(dT) -cellulos, the two fractions were translated in a reconstituted, heterologous cell-free protein-synthesizing system and the products were compared with those from the translation of total RNA. The great majority of mRNA form either prefusion or postfusion cultures was poly(A)- containing; quantitative determinations show that about 70-90% of the actin mRNA is poly(A)-containing. In order to determine if a large fraction of the calf muscle mRNA can be translated by a heterologous cell-free system, [3H]-POLY(A)+ -RNA was added to reticulocyte lysates and the formation of initiation complexes was followed. These experiments suggest that the bulk of calf muscle cell mRNA would be utilized in such a system and justify the use of cell-free systems to examine the poly(A) content of total mRNA. Thus, differential polyadenylation does not seem to be an important aspect of mRNA metabolism in cultured muscle cells. The previous study of mRNA in these cells, based on poly(A) content, is apparently a valid measure of overall mRNA metabolism.

Heterogenic mRNAs with an identical protein-coding region of the human embryonic myosin alkali light chain in skeletal muscle cells.

The formation of human myotubes in culture is accompanied by the induction of developmentally regulated, muscle-specific genes. We have studied the expression of human myosin light chain proteins and mRNAs during myogenesis in culture, in particular the skeletal embryonic myosin light chain 1 (MC1emb), which is indistinguishable from MLC1 of adult atrial cardiac muscle (MLC1A) as has been shown for rodent and bovine MLC1emb. We have identified distinct MLC1emb/MLC1A mRNAs in cultured human skeletal muscle cells that differ in their 5' and 3' untranslated regions but contain identical protein-coding regions. The alternative 3' untranslated region is detectable also in RNA of human atria. The different MLC1emb RNAs are likely to be encoded by one gene. It appears that the two MLC1emb 5' untranslated regions of the human gene are specific for man. In the mouse, only one 5' untranslated region of the MLC1emb gene has been detected.

The transcriptional activity of a muscle-specific promoter depends critically on the structure of the TATA element and its binding protein.

We have previously characterized the proximal promoter of the mouse IIB myosin heavy chain (MyHC) gene, which is expressed only in fast-contracting glycolytic skeletal muscle fibers. We show here that the substitution into this promoter of a non-canonical TATA sequence from the IgH gene results in inactivity in muscle cells, even though TATA-binding protein (TBP) can bind strongly to this mutated promoter. Chemical foot-printing data show, however, that TBP makes different DNA contacts on this heterologous TATA sequence. The inactivity of such a non-canonical TATA motif in the IIB promoter context appears to be caused by a non-functional conformation of the bound TBP-DNA complex that is incapable of sustaining transcription. The conclusions imply that the precise sequence of the promoter TATA motif needs to be matched with the specific functional class of upstream activator proteins present in a given cell type in order for the gene to be transcriptionally active.

DNA vaccination with HuD inhibits growth of a neuroblastoma in mice.

Some patients with small cell lung cancer (SCLC) or neuroblastoma develop an immune response against HuD, a human homologue of the Drosophila protein, elav, which is expressed in the nucleus and to a lesser degree the cytoplasm of neurons and tumor cells. This immune response is characterized by antibodies (anti-Hu) that at high titers are associated with a disease called paraneoplastic encephalomyelitis/sensory neuronopathy, in which infiltrates of T cells are found in the tumor and nervous system. Although all SCLCs express HuD, anti-Hu antibodies are identified in only 17% of patients with SCLC, usually at low titers, and are associated with indolent tumor growth. To determine whether the anti-Hu immune response causes indolent tumor growth, we developed an animal model using HuD DNA immunization. We found that a plasmid coding for a secreted form of HuD induced a strong and specific anti-Hu response. Immunized animals were challenged by s.c. implantation of a neuroblastoma cell line that constitutively expresses HuD. When compared with controls, mice immunized with the secreted HuD showed significant tumor growth inhibition (51% reduction volume; P = 0.0012), and 14% of them had complete tumor rejection. Tumors from these animals showed three times more CD3+ lymphocytic infiltrates than those from control mice and had a higher CD8+:CD4+ ratio. None of the animals developed neurological deficits or neuropathological evidence of nervous system pathology. In this mouse model of neuroblastoma, DNA immunization with HuD resulted in tumor growth inhibition but did not induce neurological disease. This model closely mimics the clinical course of more indolent tumor growth seen in patients with the anti-Hu immune response.

DNA-mediated immunization in mice induces a potent MHC class I-restricted cytotoxic T lymphocyte response to the hepatitis B envelope protein.

The particulate form of the major envelope or surface (S) protein of hepatitis B virus (HBV) can be taken up by antigen-presenting cells and processed for class I presentation as an exogenous protein. We have used several DNA plasmid vectors expressing the HBV envelope proteins to determine whether these sequences are able to induce cytotoxic T lymphocyte (CTL) responses in BALB/c mice after intramuscular DNA injection. A potent and specific induction was obtained, which can be detected ex vivo using either specific or nonspecific (interleukin-2) stimulation in cell culture, and the DNA-primed CTL responses are stronger than those obtained with protein injection with either stimulation protocol. The CTL response induced by DNA-based immunization is both canonical and highly specific as indicated by the nature of the epitope presented (amino acids 28-39), the class I allele used (Ld), and the T lymphocytes involved (CD8+). The CTL response is initiated between 3 and 6 days after DNA injection. By 6-12 days after a single DNA injection, ex vivo cytolytic activity is nearly maximal, and similar high levels of activity can still be detected 4 months after injection. The possibility is discussed that the unusual mode of delivery of the antigen to the immune system provided by in situ expression might allow HBV envelope antigen to be taken up and processed for class I presentation by in situ expression might allow HBV envelope antigen to be taken up and processed for class I presentation as an exogenous protein in addition to activating potentially the classical endogenous pathway.

Direct gene transfer into skeletal muscle in vivo: factors affecting efficiency of transfer and stability of expression.

Striated muscle is the only tissue found to be capable of taking up and expressing reporter genes that are transferred in the form of plasmid DNA. Thus, direct gene transfer is a potential method of gene therapy for the primary inherited myopathies. However, results to date have had insufficient and too variable expression to consider using direct gene transfer in human trials. We have determined that much of the variability of expression is due to nonuniform distribution of substances injected into skeletal muscle in vivo, and have developed a model to ameliorate this. Preinjection of muscles with a relatively large volume of hypertonic sucrose improves the distribution of injected substances and results in significantly less variable expression of reporter genes for luciferase or beta-galactosidase; the coefficient of variation for mean luciferase activity was reduced from about 120% to 25%. Expression is not directly proportional to dose, but is more so if the muscles are preinjected with sucrose than not. Expression is higher and less variable if DNA is injected in a larger than a smaller volume. The choice of promoter appears to be particularly important. Luciferase reporter gene expression from the SV40 promoter was transient and low, whereas expression driven by the Rous sarcoma virus (RSV) promoter was high and sustained, such that a 1,000-fold difference in expression could be observed. The mechanism of gene uptake is still unknown, but our findings indicate that fibers damaged by the injection procedure do not take up and express plasmid DNA.

Plasmid DNA is superior to viral vectors for direct gene transfer into adult mouse skeletal muscle.

Direct gene transfer into skeletal muscle offers several therapeutic possibilities. We assessed direct intramuscular injection of recombinant plasmids, adenovirus, or retrovirus in normal or regenerating muscles of mice. The incorporation and expression of reporter genes introduced by any of these three vectors is greater in regenerating than in mature muscle. In regenerating muscle, pure DNA and adenovirus result in equivalent numbers of fibers expressing reporter gene (> 10%), but adenovirus also induces considerable cellular infiltration. In mature muscle, recombinant DNA is better than adenovirus. Retrovirus failed to infect mature muscle fibers and was less effective than plasmid DNA or adenovirus in regenerating muscle. The surprisingly high relative efficiency of pure plasmid DNA suggests that this method will provide a simple, safe and viable alternative for gene therapy involving muscle tissue.

A comparative analysis of intermediate filament proteins in bovine heart Purkinje fibres and gastric smooth muscle.

The intermediate filament (IF) composition of muscle cells of various sources is still a controversial issue. In the present study, the IF composition of bovine Purkinje fibres (PFs), atrial and ventricular myocardium, and gastric smooth muscle (SM) has been compared using biochemical and immunocytochemical methods. The Mr of the major IF subunit protein in all four tissues was 55,000. In two-dimensional (2-D) electrophoresis gels of Triton-treated ordinary atrial and ventricular myocardium and the gastric muscular wall, two or three isoelectric isoforms were seen, whereas in PFs up to seven isoforms caused by phosphorylation were observed. In immunofluorescence studies antibodies against the Mr 55,000 subunit of PFs and gastric SM, respectively, both showed identical reactivity with PFs, atrial and ventricular myocytes, gastric SM cells and some SM cells in intramyocardial and gastric muscular wall blood vessels. A small amount of vimentin (Mr 57,000) was also detected in 2-D gel electrophoresis in all four tissues as well as in immunoblotting of PFs with antibodies to vimentin. Immunofluorescence studies using both polyclonal and monoclonal antibodies to vimentin showed that vimentin was present in the endothelium and SM cells of both intramyocardial and gastric muscular wall vessels, sometimes together with desmin in the vascular SM cells, but was never seen in PF, atrial, ventricular or gastric SM cells proper. As expected, vimentin was present in interstitial tissue, i.e., fibroblasts and capillaries. However, interestingly, the monoclonal antibodies, which recognized different antigenic determinants of vimentin, did not give identical staining patterns. Especially the staining of the vascular SM cells differed. Since this staining pattern did not change upon denaturation and unmasking experiments, it seems that the organization of vimentin in different mesenchymal cell types varies. Vimentin was also detected in isolated PFs but here it was located solely in the contaminating interstitial tissue. Thus, desmin is the sole IF protein expressed in PFs, in atrial and ventricular myocytes and in gastric SM cells proper; vimentin alone being present in the interstitial tissue cells, whilst in vascular SM cells desmin and vimentin are coexpressed in various proportions. The variation in number of isoforms of desmin and the heterogeneity in staining of mesenchymal tissues with monoclonal vimentin antibodies probably indicates that the IF cytoskeletons are differently organized in various cell types, even though they contain IFs of the same class.

Mechanisms underlying the asynchronous replacement of myosin light chain isoforms during stimulation-induced fibre-type transformation of skeletal muscle.

During the fibre-type transformation induced by chronic electrical stimulation of rabbit fast-twitch muscle, replacement of the fast forms of the two classes of myosin light chain by their slow isoforms occurs asynchronously. Studies of total cellular myosin light chains and of the slow-to-fast transition now justify the conclusion that the asynchrony is due to switching between the expression of fast and slow genes for the two light chain classes at sequential stages of the transformation process.

Myosin isozyme changes in the heart following constriction of the ascending aorta of a 25-day old rat.

When a constricting band is placed around the ascending aorta of young (25-day old) rats, all chambers of the heart eventually produce hypertrophy. Both the left and right ventricles show strong shifts toward an isozyme pattern in which V3 is predominant, similar to that seen in models where hypertrophy is induced in adult rats. The hypertrophied atria however, show no detectable change in the native myosin isozymes or the light chain subunits.

Effects of hypothyroidism on myosin isozyme transitions in developing rat muscle.

Hypothyroidism was induced in young rats by methylthiouracil treatment of pregnant mothers from 18 days of gestation to 4 weeks after birth. Electrophoretic analysis of native myosin isozymes revealed a persistence of neonatal and embryonic myosin in developing fast and slow muscles up to at least 28 days after birth. The appearance of adult fast myosin was inhibited in 28-day old animals, however adult slow myosin was found in the soleus muscle. Immunocytochemical results on the soleus demonstrate a cellular heterogeneity in the response to hypothyroidism. About half fibers have a normal complement of slow myosin and do not contain neonatal myosin. Only the remaining fibers contain the large amounts of neonatal myosin demonstrated by electrophoresis.

A developmentally regulated disappearance of slow myosin in fast-type muscles of the mouse.

Histochemistry and immunocytochemistry using an antibody to adult rat slow-type myosin demonstrated that about 10% of the fibers in the mouse extensor digitorum longus and semimembranosus muscles contain slow myosin during the first month after birth. In adult animals, these muscles have only 0-08% slow myosin-containing fibers. These results demonstrate a developmentally linked disappearance of an adult-type myosin, and show that the adult phenotype of muscle fibers is not necessarily determined before birth as previously suggested.

Comparative structural analysis of myosin after limited tryptic hydrolysis by use of two-dimensional gel electrophoresis.

Polypeptides obtained by limited tryptic digestion of several myosins have been analyzed by a two-dimensional electrophoresis technique. The different maps thus obtained exhibit some common and distinct features specific of the myosins studied. Myosins from rabbit, fast and slow muscle and cardiac tissue, as well as beef cardiac myosin, have been compared. The polypeptides obtained vary in molecular weight from 120 000 60 15 000. The light chains LC1 and LC2, have disappeared. A peptide which comigrates with fast type LC3 is however found. The fast myosin hydrolyzate is very different from that obtained by the hydrolysis of slow and cardiac myosin. Numerous peptides are common to cardiac and slow myosins. However a few peptides are specific of the two myosin types. In the hydrolyzate from fetal calf myosin, some of the typical slow, fast and cardiac peptides can be found. However several apparently unique fetal peptides are also present. The comparison of the fetal calf tissue myosin hydrolyzate and that of [35S] methionine labeled myosin from myotubes in cultures shows qualitatively a very great homology. Thus the protein synthesized by cultured cells seems to be very similar to or the same as that of the embryonic tissue.

Selective action of colchicine on protein synthesis and release in a clonal line of rat glial cells.

The effect of colchicine on protein synthesis and secretion in stationary cultures of clonal rat glial cells C6 was examined. Colchicine inhibited the synthesis of the brain specific S100 protein in intact cells but not in a cell-free protein synthesizing system derived from these cells. There was no demonstrable effect of the drug on the synthesis of any of the several hundred proteins resolved by a two-dimensional electrophoretic analysis. However, colchicine specifically enhanced the secretion of several proteins of molecular weighs of 30,000 and of 200-300,000 into the medium. Two of the high molecular weight proteins were apparently membrane proteins whose release into the medium was stimulated by the drug.