|[Frontiers in Bioscience 1, d19-29 March 1, 1996]|
MICROINJECTION STRATEGIES FOR THE STUDY OF MITOGENIC SIGNALING
IN MAMMALIAN CELLS
Ned J.C.Lamb, Cecile Gauthier-Rouviere and Anne Fernandez.
Cell Biology Unit, C R B M , CNRS-INSERM, 1919, Route de Mende, F-34033, Montpellier Cedex. France.
Received 15/12/95; Accepted 30/01/96; On-line 03/01/96
One of the principal difficulties in manipulating intracellular metabolism stems from the need for purified active components in a form suitable for microinjection. An alternative strategy to microinjecting purified protein components, involves directly overexpressing cellular components through microinjection of DNA sequences in expression vectors under the control of efficient mammalian promoters. Two different types of plasmid DNA can be used to modulate cellular protein levels in either an additive or substractive manner: coding DNA constructs placed in a sense orientation after the promoter can be used to induce rapid elevation in the intracellular level of a given protein, whereas an antisense orientated DNA sequence will suppress the expression of a given protein.
As with most microinjection tools, the use of expression plasmids has both advantages and disadvantages (table 1). The major advantage of plasmid expression vectors is the production of soluble, homogenous, correctly folded and post translationally modified proteins within cells. This precludes the purification and isolation of proteins, which often leads to partial degradation and modification or inactivation of the protein. Finally, it is possible to couple genes of interest to short peptide tags which subsequently allows easy identification of the overexpressed protein within cells. Short epitope tags such as the hemagglutinin (HA) tag derived from influenza viral proteins can be particularly useful to determine the cellular localization of the overexpressed proteins by immunocytochemistry using anti-HA antibodies.
Whereas, in principle there is no difference between plasmid transfection and microinjection, the latter offers a number of advantages. Primarily, the efficiency of "transfection" by microinjection is close to 100% of the injected cells and therefore does not require any selection process. Secondly, the number of copies of the plasmid injected per cell can be accurately manipulated since the concentration of plasmid DNA in the needle is known, . Thirdly, different plasmids can be microinjected in close proximity on the same dish, allowing direct and simultaneous comparison of their effects under the same experimental conditions.
Three types of eukaryotic expression vector can be used. The first class is derived from eukaryotic viruses and contains deregulated highly efficient promoter sequences that lead to continuous and uncontrolled expression of DNA sequences placed downstream from the promoter region. The second class, termed inducible, contain promoters derived from various existing mammalian promoter sequences that respond to the addition of heavy metal ions or glucocorticoids. The third class comprise plasmids in which a specific promoter sequence from a known mammalian gene such as the c-fos promoter, is placed 5' from the gene of interest. These promoters respond in a manner identical to that of the normal gene which they regulate, leading to the expression of the protein of interest instead of Fos. To date, the first class of vector has been the most commonly used for specific overexpression of proteins. Within this group, two promoters have been predominantly used: the first derived from the polyoma SV40 virus and the other derived from the cytomegalovirus. Both classes of promoter are sensitive to the presence of growth factors, an important point to bear in mind when using them to study early events in cell activation.
There are some other limitations in the use of such plasmids for microinjection. First, they produce unregulated expression of the target protein, which within 2 to 3 hours of injection may lead to levels of protein in the cell that are far beyond physiological levels. As a result, any effect(s) and/or change in localization may in such cases, be artifactual. Second, since there is a lag time of 3-4 hours between microinjection and expression, plasmid injection is inappropriate for studies of highly dynamic intracellular processes.
Plasmid microinjection has been used to study the consequences of overexpressing dominant negative forms of the ubiquitous mitogen activated kinase (MAP-kinase). This kinase is activated by a serial cascade initiated at the membrane by activation of receptor tyrosine kinases. This results in a signal to MAP kinase which gets phosphorylated on tyrosine and threonine residues and subsequently becomes activated (30). Two expression vectors were used, encoding either wild type MAP kinase or a mutant form in which essential tyrosine and threonine residues had been eliminated by mutation to phenylalanine and alanine. These gene constructs were placed downstream of a CMV promoter and subsequently microinjected into rat fibroblasts. Cells were microinjected with the expression plasmids whilst still proliferating and then placed in serum free medium for 24 hours to shut down the mitogen activated pathways. The reinitiation of proliferation upon serum addition was then assessed by measuring passage through S-phase. The level of expression of MAP kinase was probed by immunofluorescence using either antibodies against MAP kinase or antibodies directed against the human influenza HA tag (monoclonal 12CA5). The simultaneous immuno-detection of BrdU incorporation (5bromo-deoxyuridine, an analogue of thymidine which incorporates into newly synthesized DNA when added to the culture medium), allowed determination as to which cells exhibited DNA synthesis. As anticipated, complete abolition of the proliferative response was observed in cells microinjected with the MAP kinase mutant plasmid. However, a similar inhibition was seen with both the wild type plasmid and the vector alone. This point illustrates a possible pitfall in microinjecting plasmid DNA, particularly using the CMV vector system: an excess of vector alone frequently leads to an inhibition of cell proliferation. The reason is unclear, but is dependent upon the plasmid size, expression efficiency and the nature of the promoter region. SV40 driven expression vector plasmids produced a similar response although with less inhibition on cell proliferation when more than 20-30 copies of the plasmid were injected per cell. However, careful titration analysis using lower concentration of the CMV-driven plasmid showed that, when down to 2 copies of plasmid were injected into the nucleus of each cell, both the plasmid vector and the wild type kinase plasmid inhibited DNA synthesis in at least 80% of the cells. Whereas cells injected in their nuclei expressed the plasmid much more efficiently than when injected into the cytoplasm (in comparison to nuclear injection which yielded a consistent 85 - 95% expression, cytoplasmic microinjection yielded no more than 30% expression of the protein), none of the cells injected with control plasmids in the cytoplasm were inhibited in their passage through DNA synthesis. Therefore, in these experimental conditions, only cytoplasmic injection allows to identify a specific inhibition of cell proliferation by the dominant negative mutant of MAP kinase.
These data illustrate one of the most important points when using plasmid microinjection (and microinjection in general): the control for an experiment involving microinjection must always be very careful and well thought through. Such control is a necessary key to the interpretation of any microinjection result. In addition, for the study of protein kinases and phosphatases, plasmid injection strategy should be confronted, when possible, with injection of purified active proteins because cells may counteract the plasmid mediated overexpression as it occurs by overexpressing a corresponding regulatory or inhibitory subunit. Transfection approaches are even more prone to such problems because they require longer time scale experiments. In contrast, injection of highly purified, highly active kinases and phosphatases temporarily circumvent the entire cellular regulatory mechanism by completely skewing the intracellular metabolic balance for 1-2 hours.
The use of microinjection in antisense experiments illustrates the perfect combination between microinjection, biochemistry and immunology. This technique was first used to demonstrate in a formal manner the absolute requirement of the protein cyclin A in DNA synthesis at a time when it was believed that this protein is uniquely involved in the mitotic activation (31). Microinjection of an SV40 expression vector encoding the full length anti sense cyclin A RNA into synchronized mammalian cells completely suppressed cyclin A expression at S-phase. This was shown immunofluorescence of cyclin A protein and simultaneous inhibition of DNA synthesis. This effect was exclusive and specific to cyclin A since antisense to the mitotic cyclin, cyclin B, did not produce a similar block. However, to provide formal proof that the effect observed with the injection of antisense RNA was exclusively related to depletion of cyclin A, cells were subsequently microinjected with purified cyclin A proteins. This specifically reversed the block to DNA synthesis. This technique of antisense rescue represents probably the most formal demonstration for a specific effect that can be provided by microinjection at the present time. In our hands, it is also possible to rescue cells injected with antibodies, but this involves the use of extremely high and unphysiological levels of antigen.