We imagine that transcriptional activators work by one or both of the following mechanisms: recruiting an essential component to DNA and/or changing the conformation of some pre-bound target. In vitro experiments have led to the suggestion that gene activation in eukaryotes involves an ordered, multistep assembly of the transcription complex on DNA, various steps of which might be catalyzed by activators, either by direct recruitment or by induced conformational change. My recent work, taken with that of others, suggests that a transcriptional activator works by recruiting a pre-formed RNA Polymerase II holoenzyme complex to DNA, and that such recruitment can be achieved by interaction between the DNA-bound activator and any target present on the surface of the holoenzyme. With my future research, using a combination of genetic and biochemical assays, I will pursue analysis of the mechanisms of transcriptional activation in eukaryotes.

The selective use of genes, a phenomenon known as gene regulation, underlies the processes of development and of cellular differentiation in eukaryotes. Proper regulation of genes is vital to the organism. Indeed, incorrect gene expression is the major cause of cancer, and many of the oncogenes responsible for causing cancer encode factors which control messenger RNA synthesis. While many steps are required to express a gene, much of gene expression is controlled at the transcriptional level. The mechanism of transcriptional activation in eukaryotes has been the focus of great attention for many years; it is also the heart of my research. The problem, reduced to its simplest form, is how does a transcriptional activator when bound to DNA upstream, or even downstream, of a gene stimulate transcription of that gene by RNA polymerase II. The answer to this question has remained elusive in part due to the many proteins (transcription factors) that form the general transcriptional machinery, as well as the many different types of activator proteins that have been identified (for reviews see refs. 1 to 3).
   Different activators have been shown to interact with one or another of various general transcription factors, including TATA-binding protein (TBP), TBP-associated factors (TAFs), TFIIB, TFIIA, and TFIIH (4-8), and efforts now center on determining the biological relevance of these interactions. In vitro, the general transcription factors can assemble in an ordered fashion on a promoter along with RNA polymerase II to form the so-called preinitiation complex (9-11). These findings have led to the suggestion that transcriptional activators might function by catalyzing various steps of this ordered assembly process on DNA. It has been proposed, for example, that activators recruit one or more of the general transcription factors to DNA, or induce conformational changes in target proteins that either initiate the process of assembling the transcriptional apparatus, or that trigger transcription initiation and/or elongation (12-15). The more recent report that, in yeast as well as in mammals, RNA polymerase II and several transcription factors can form a complex, called the Pol II holoenzyme, independently of DNA, raises the possibility that simple recruitment of the holoenzyme would suffice for gene activation (16-20).
   The results of experiments that my colleagues and I have performed suggest that simple recruitment of the Pol II holoenzyme is indeed a plausible mechanism of transcriptional activation (21). Our results can be summarized as follows: the yeast transcription factor GAL11 is a component of the Pol II holoenzyme. It is important to emphasize that, despite its name, the role of GAL11 in transcription is not restricted to genes or activators involved in galactose metabolism. In yeast cells bearing a mutant form of this protein, called GAL11P (potentiator), certain derivatives of the transcriptional activator GAL4 lacking any classical activating region work as strong activators. We have shown, using a combination of genetic and biochemical analysis, that the mutation in GAL11P promotes a novel interaction between GAL11 and a portion of the dimerization region (not the activating region) of GAL4 (GAL4dim). As far as we know, GAL11 is not the natural target of classical activators, and the region of GAL11 affected by the P mutation is evidently functionally inert in ordinary cells, suggesting that this mutation is of no functional significance beyond creating a fortuitous target-site on the holoenzyme for an otherwise inactive DNA-binding protein. The effect on gene activation of the GAL4dim-GAL11P molecular interaction can be mimicked by directly fusing GAL11 to a DNA-binding domain. Furthermore, by using different GAL11P alleles, we have also shown that the strength of the interaction between the two components, as quantitated in vitro, correlates with the degree of gene activation elicited in vivo (22). The simplest explanation for our results is that tethering GAL11 to DNA recruits the Pol II holoenzyme and thereby activates gene transcription. More complicated explanations, such as induction of relevant conformational changes, or enzymological modification, are difficult to imagine for this instance of gene activation. Our results suggest that interaction between a DNA-bound protein and a single component (perhaps any component) of the holoenzyme is sufficient to trigger gene activation. This interpretation implies that activation of transcription does not require direct interaction of the activating region with, for example, machinery that helps remove nucleosomes (23).

It is widely recognized that, while in vitro experiments are required for studying molecular interactions in a controlled setting, a genetic approach is indispensable to address the biological relevance of such interactions. I have decided to continue to work principally with yeast because this organism provides the best model system for combining genetic and biochemical analyses of fundamental molecular processes such as gene activation.
   The results obtained by studying the GAL4dim-GAL11P interaction and its effect on gene transcription raise the possibility that natural (classical) activators might also work by recruitment of the Pol II holoenzyme, potentially through interactions with any exposed component. With the next generation of experiments I will address this issue, namely whether the mechanism of gene activation by GAL4dim/GAL11P can be considered a paradigm for natural activation in general. Any attempt to pursue this goal will also require the identification of Pol II holoenzyme components contacting GAL11, and the characterization of their reciprocal interactions and functions within the complex.
   Our system provides an excellent tool for investigating mechanisms of transcriptional activation because a defined, single interaction between a DNA-binding protein and one component of the holoenzyme triggers gene expression. This simplicity, in contrast to the complexity of the interactions thought to occur between natural activators and their targets, is advantageous for gaining insight into in vivo phenomena such as synergism in transcriptional activation. This system could also provide insight into the mechanistic similarities (or differences) between the transcription initiation and reinitiation processes. Furthermore, we have recently developed an assay that should allow us to test whether a remote enhancer communicates with a gene promoter via DNA looping. This assay has been developed in yeast. However, we want to modify it such as to apply it also to mammalian cells, in which enhancers can work at a greater distance than in yeast.
   Through our research, I want to address the following specific questions:
a) How similar are the mechanisms of activation by GAL4dim/ GAL11P and classical activators?
b) What are the Pol II holoenzyme components interacting with GAL11? What are the functions of these proteins?
c) Are there GAL11 homologous proteins in higher eukaryotes?
d) Is the observed phenomenon of synergism in transcriptional activation the result of simultaneous interactions between activators and their different targets on the surface of the holoenzyme?
e) Does each initiation event involve recruitment of the entire transcription complex to DNA?
f) Do remote transcriptional enhancers work via DNA looping?