Recombinant DNA clones containing a mammalian cDNA with appropriate expression signals can be expressed in bacterial cells to give high-level expression of mammalian proteins, but the biological properties of the expressed proteins may not be fully representative of the native molecules. Expressing mammalian proteins in mammalian cells has the obvious advantages that correct protein folding and post-translational modification are generally not an issue, and it is possible to analyze downstream signals and cellular effects. Once transfected into mammalian cells, a transgene with appropriate expression signals can be expressed for short or long periods, according to the expression systems used, as listed below.
• Transient expression. Here, the transgene remains as an independent, extrachromosomal genetic element (an episome) within the transfected cells. In cultured cells, expression of the introduced gene often reaches a maximal level about two or three days after transfection of the expression vector into the mammalian cell line, but that may be sufficient to carry out certain types of gene analysis. Thereafter, expression can diminish rapidly as a result of cell death or loss of the expression construct. (But in long-lived, nondividing cells, expression of non integrated transgenes may be maintained for long periods.)
• Stable expression. Here, the aim is to have the transgene integrate into the host cell’s genome. That often requires a virus vector (retrovirus vectors are highly efficient at integrating into chromosomal DNA) but plasmid expression vectors can also randomly integrate at low frequencies, and stable cell lines can be developed with an integrated gene of interest. The advantage here is that if the transgene has integrated so that the introduced gene is expressed, all descendant cells will contain this gene and expression can be maintained over many cell generations. Neighboring inhibitory regulatory elements and closed chromatin structure may result in silencing of the transgene (“position effects”). The structure of the transgenic locus may also have an effect when more than one copy of the transgene inserts at the same chromosomal location.
Transient expression: the example of COS cells
Stable expression systems in mammalian cells, typically based on plasmid sequences that integrate within a chromosome, have delivered kilograms of complex proteins in industrial-scale bioreactors, but they have required large investments in time, resources, and equipment. As an alternative, large-scale transient expression systems have been developed for producing “recombinant” proteins in mammalian cells.
In addition to delivering protein expression, some mammalian cell lines have found major applications in screening the effects of in vitro manipulations on transcriptional and post-transcriptional control sequences. A good example is provided by COS cells, stable cell lines originally derived from an African green monkey kidney fibroblast cell line, CV-1. When CV-1 cells are infected with the monkey virus SV40, the normal SV40 lytic cycle ensues. However, when CV-1 cells were transformed with a strain of SV40 with a defective replication origin, a segment of the SV40 genome became integrated into the chromosomes of CV-1. The resulting COS cells (CV-1 with defective origin of SV40) stably express the SV40-encoded large T antigen, the only viral protein required for activation of the SV40 origin of replication.
By expressing the SV40 large T antigen in a stable way, COS cells permit any introduced circular DNA with a functional SV40 origin of replication to replicate independently of the host cell’s chromosomes, with no clear size limitation. When transient expression vectors are transfected into COS cells, permanent cell lines do not result because massive vector replication makes the cells nonviable. Even though only a small proportion of cells are successfully transformed, the amplification of the introduced DNA to high copy numbers in those cells compensates for the low take-up rate.
Stable expression and marker-selection systems
Genes introduced into mammalian cells can stably integrate into host chromosomal DNA, but the process is very inefficient. The rare stably transformed cells must be isolated from the background of nontransformed cells by selection for some marker. Two broad approaches have been used: functional complementation of mutant host cells and dominant selectable markers.
In functional complementation, host cells have a genetic deficiency that prevents a particular function; the original function can be restored if a functional copy of the defective gene is supplied by transformation by a vector. The transgene and the marker can be transferred as separate molecules by a process known as co-transformation. An example is the use of a thymidine kinase gene marker with cells that are genetically deficient in thymidine kinase (TK−). Thymidine kinase is required to convert thymidine into thymidine monophosphate (TMP or thymidylate), but TMP can also be synthesized by enzymatic conversion from dUMP. The drug aminopterin blocks the dUMP → TMP reaction, so cells cannot grow in the presence of this drug unless they have a source of thymidine and a functional thymidine kinase gene. Selection for TK+ cells is usually achieved in HAT (hypoxanthine, aminopterin, thymidine) medium (Figure 1).
Fig1. Purine and pyrimidine salvage and the basis of HAT selection in thymidine kinase marker assays. Cultured animal cells can normally synthesize purine nucleotides and thymidylate (TMP) by de novo pathways that involve certain amino acid substrates. De novo synthesis of purine nucleotides begins from glutamate. It involves the eventual synthesis of IMP (inosine monophosphate), which can be converted separately to AMP and GMP. De novo synthesis of pyrimidine nucleotides begins from aspartate. It is dependent on synthesizing UMP and dUMP, which is converted to TMP by thymidylate synthetase; UTP is deaminated to give CTP (not shown). An activated form of tetrahydrofolate is crucially required to provide a methyl or formyl group at three key steps that can be blocked by antifolate agents such as aminopterin, shutting down the de novo pathways. However, purine nucleotides and thymidylate can also be synthesized from pre-formed bases by salvage pathways that are not blocked by antifolate drugs such as aminopterin. The three salvage-pathway enzymes are shown in pale pink boxes: thymidine kinase (TK), hypoxanthine–guanine phosphoribosyltransferase (HGPRT), and adenine phosphoribosyltransferase (APRT). The use of HAT medium (containing hypoxanthine, aminopterin, and thymidine) therefore selects for cells with active salvage-pathway enzymes. Mutant tk− cells are genetically deficient in thymidine kinase and will not survive if grown in HAT medium unless they contain vector molecules containing a functional thymidine kinase gene. PRPP, 5-phosphoribosyl-1-pyrophosphate.
The major disadvantage of complementation markers is that they are specific for a particular mutant host-cell line in which the corresponding gene is nonfunctional. As a result, they have largely been superseded by dominant selectable markers, which confer a phenotype that is entirely novel to the cell and hence can be used in any cell type. Markers of this type are usually drug-resistance genes of bacterial origin that can confer resistance to drugs known to affect both eukaryotic and bacterial cells. For example, the aminoglycoside antibiotics (which include neomycin and G418) are inhibitors of protein synthesis in both bacterial and eukaryotic cells. The neomycin phosphotransferase (neo) gene confers resistance to aminoglycoside antibiotics, so cells that have been transformed by the neo gene can be selected by growth in media containing G418.
