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

2.1 Oligonucleotide microinjection for in vivo competition in the study of DNA binding transcription factors.

The current dogma is that transcriptional regulation of gene expression is mediated through DNA binding proteins, termed transcription factors. These factors, through their interaction with specific DNA sequences in the 5' regions of genes, modulate the capacity of RNA polymerases to transcribe the downstream 3' coding region. In studies developed principally by Treisman and Gilman (7), it was clearly shown that the transcription factors bound to DNA in a sequence specific manner through precise DNA sequences ranging in size from 7 to 30 nucleotides (6,7).

To examine the role of transcription factors in mediating mitogenic or oncogene signaling in vivo, one strategy involves microinjecting synthetic double-stranded oligonucleotides containing a specific DNA-binding motif into the cytoplasm of cells. These small sequences rapidly migrate into the nucleus where they specifically compete with the endogenous sequences for the DNA binding activity of the related transcription factors (Fig. 1).

The Serum Response Element (SRE) present in the promoter region of several genes including c-fos, requires the binding of SRF, in a dimeric form (together with other accessory proteins), to activate expression of the gene downstream. Expression is prevented by competitive squelching of the SRF transcription factors, which bind to microinjected oligonucleotides corresponding to the sequence of the endogenous SRE. In the SRE sequence shown, the core consensus CC(AT)₆GG (also named CArG box) is underlined.

This technique was used in particular, to study the regulation of c-fos expression and subsequent cell proliferation. Of the major regulatory elements present in the 5' region of the c-fos promoter, the Serum Response Element (SRE), is a region of DNA 300 bp 5' of the transcription initiation site closely juxtapositioned to the TPA response element (TRE). SRE is known to bind the ubiquitous 67 kD mammalian transcription Serum Response Factor (SRF or p67SRF) (8-10), whereas the TRE binds Fos-Jun AP1 complex (reviewed in 11). The study of SRE is of particular interest since it may be implicated in transcriptional control of numerous serum responsive genes. In vivo competition studies were performed using synthetic double stranded oligonucleotides corresponding to the sequence of SRE [which contains the typical core sequence CC(AT)₆GG, also named CArG sequence], or as a control, a mutated oligonucleotide, in which essential oligonucleotide pairs required for interaction with SRF were changed. Using this approach, microinjection of SRE oligonucleotides into synchronized fibroblasts has allowed to show the direct implication of the SRE-containing region not only in c-fos induction following serum addition (12), but also in c-fos repression in quiescent cells together with another short DNA sequence, present in the coding region of c-fos (13). Using a similar strategy, the binding of protein factors at cAMP responsive elements (CRE), was shown to be necessary in mediating the induction of c-fos by cAMP (14).

2.2 Antibody microinjection for in vivo study of transcription factors activity.

While the technique of oligonucleotide microinjection has been highly useful in the analysis of SRE/TRE regulated functions, it is restricted in its application. Oligonucleotides have short half lives in vivo and may present broad specificity by interfering with the binding of several proteins acting at the same promoter site. An alternative method of interfering with transcription factor activity involves microinjection of antibodies directed against the particular transcription factor. Antibodies offer the advantages of relatively long half lives in vivo and more restricted specificity since the antigenic epitope can be specifically targeted to a single protein (see table 1).

The neutralization of SRF in vivo through microinjection of specific antibodies, has shown that activation of SRF (through binding and/or post-translational modification) is required for c-fos induction and subsequent entry into S phase in living cells (15,16).

Moreover, events occurring long after the initial stimulation by serum, also require SRF since DNA synthesis was blocked by microinjecting antibodies to SRF up to 8 hrs after serum addition.

A similar approach, exploiting the advantage of the long half life of microinjected antibodies, has been used to demonstrate the implication of SRF in skeletal muscle differentiation, a process which involves the fusion of myoblasts into myotubes over a period of 3 to 4 days. When injected into myoblasts, anti-SRF antibodies completely prevented the expression of early differentiation markers such as myogenin and subsequent fusion of myoblasts while a majority of surrounding uninjected cells expressed differentiation markers and underwent fusion into myotubes (17).

These examples illustrate the value of microinjection of antibodies to probe and validate in vivo, the role of a transcriptional factor such as SRF which positively regulates diverse processes such as c-fos gene induction, cell cycle progression to S-phase and myoblast differentiation.

The presence of known consensus phosphorylation sites along the protein is indicated. The place of the recently identified (28) nuclear localization signal (NLS) is also marked, and the portion of SRF protein used as a dominant negative mutant of SRF is shown as a shaded box underneath.

2.3 Dominant negative mutant of transcription factors based upon their DNA binding domain

Since the use of antibodies for microinjection requires production of antisera reactive against native proteins, often a tedious process, a complementary strategy to interfere in vivo with transcription factors was developed. The approach consists of engineering a portion of the protein corresponding to its DNA binding domain devoid of transcriptional activation domains.

To test this approach for SRF, a portion of SRF spanning amino acids 113 to 265 containing the DNA binding domain, SRF-DB (see Figure 2), was expressed in bacteria, purified through DNA-affinity and microinjected into living cells.

Microinjection of SRF-DB effectively prevented c-fos expression and DNA synthesis stimulated by growth factors, two events described to require SRF, implying that an excess of SRF-DB injected into living cells can act as a dominant negative mutant (18). This strategy demonstrates the efficiency of a method for generating dominant negative mutants by using the DNA binding region of transcription factors and may allow the future study of events at the DNA/protein level. This approach has other advantages. First, it is not restricted to SRF but can be used in studying all transcription factors or similar DNA-binding proteins for which a defined DNA binding domain has been identified. Second, since the DNA binding sequences are unique for each transcription factor, this approach provides a sharp specificity for the probes. Third, the polypeptide encoding the DNA-binding domain is generally small enough to easily pass through the nuclear membrane and therefore does not require nuclear microinjection (see table 1). Fourth, this technique is insensitive to potential variations in antibody-antigen reactivity in vivo (which may arise with multimeric protein complexes). As such, this ensures that the effects observed are directly related to the injected sequence. Finally, it is possible to develop several DNA binding molecules containing different associated modules (for example dimerization or nuclear localisation domains) to study the relative in vivo contribution of each of these individual domains.