The Spl protein activates transcription from many eukaryotic promoters. Spl can act in vivo from enhancer sites that are distal to the promoter and exhibit synergistic interaction with promoter-proximal binding sites. To investigate possible protein-protein interactions between DNA-bound Spl molecules, we have used electron microscopy to visualize the DNA-protein complexes. At the SV40 promoter, we observed the expected localized interaction at the Spl sites; in addition, we found that DNA-bound Spl served to associate two or more DNA molecules. At a modified thymidine kinase promoter, we observed a localized interaction at each of two binding locations that were separated by 1.8 kbp; in addition, we noted a substantial fraction of DNA molecules in which the distant binding regions were joined by a DNA loop. As judged by studies with mutant Spl proteins, the distant interactions depended on the glutamine-rich regions of Spl required for transcriptional activation. We conclude that DNA-bound Spl can self-associate, bringing together distant DNA segments. From the correlation between DNA looping in vitro and synergistic activation of the modified thymidine kinase promoter shown previously in vivo, we suggest that Spl exerts its transcriptional synergism by a direct protein-protein association that loops the intervening DNA. Our experiments support the DNA-looping model for the function of transcriptional enhancers.[Key Words: Spl protein; DNA looping; DNA binding; transcription]Received December 26, 1990; revised version accepted January 23, 1991.The controlled initiation of transcription in prokaryotes and eukaryotes depends on the action of regulatory proteins from sites that are too distant for a direct interaction with promoter-bound proteins on linear DNA. Three principal models have been proposed for positive regulation from distant (enhancer) sites (Dynan and Tjian 1985;Echols 1986;Ptashne 1986). In the first, the regulatory protein (or RNA polymerase) associates with the DNA at the enhancer site and then traverses the DNA to the promoter site (scanning model). In the second, the enhancer-binding protein initiates a change in DNA structure that is propagated from the enhancer to the promoter, thereby activating transcription (structural transmission model). In the third, the enhancerbound regulatory protein activates transcription by a direct protein-protein interaction with RNA polymerase at the promoter (or other proteins that contact polymerase) (DNA-looping or nucleoprotein model).DNA looping is currently favored as the most likely mechanism for distant regulatory interactions. The interaction between DNA-bound proteins has been firmly established as the central structural feature for regulated 3Present address:

The NtrC protein activates transcription of the ginA operon of enteric bacteria by stimulating the formation of stable "open" complexes by RNA polymerase (oMholoenzyme form). To regulate the ginA promoter, NtrC binds to sites that have the properties of transcriptional enhancers: the sites will function far from the promoter and in an orientation-independent fashion. To investigate the mechanism of enhancer function, we have used electron microscopy to visualize the interactions of purified NtrC and RNA polymerase with their DNA binding sites and with each other. Under conditions that allow the formation of open complexes, about 30% of DNA molecules carry both RNA polymerase and NtrC bound to their specific sites. Of these, about 15% form looped structures in which NtrC and the RNA polymerase-promoter complex are in contact. The length of the looped DNA is that predicted from the spacing that was engineered between the enhancer and the ginA promoter (390 base pairs). As expected for activation intermediates, the looped structures disappear when RNA polymerase is allowed to transcribe the DNA. We conclude that the NtrC enhancer functions by means of a direct association between DNA-bound NtrC and RNA polymerase (DNA-looping model). Association of DNA-bound proteins appears to be the major mechanism by which different types of site-specific DNA transactions are localized and controlled.Enhancer sequences have been defined by their function in activating transcription from relatively long distances in an orientation-independent manner (1, 2). There are many wellcharacterized examples of enhancers controlling eukaryotic cellular and viral genes (3-5). Although initially defined in eukaryotes, regulation of transcription from distant sites has been observed for many prokaryotic promoters (6, 7). The site for positive regulation of the glnA operon of enteric bacteria is a particularly well-defined example of a prokaryotic enhancer (8-10). To investigate the mechanism of enhancer action, we have studied the activation of transcription from the glnA promoter by the NtrC enhancer-binding protein.Three principal models have been proposed for enhancer action (11-13). In the first, the regulatory protein (or RNA polymerase) associates with the DNA at the enhancer site and then traverses the DNA to the start site for RNA synthesis (entry-site or scanning model) (1,14). In the second model, the enhancer-binding protein facilitates transcription by initiating a change in DNA structure that is propagated from the enhancer site to the promoter (e.g., a site-specific DNA topoisomerase) (15). In the third model, the enhancerbound regulatory protein stimulates transcription by a direct protein-protein interaction with RNA polymerase at the promoter site (or with other proteins that contact polymerase) (DNA-looping model) (11-13, 16).The DNA-looping model for enhancer action has been considered attractive for several reasons. The interaction between DNA-bound proteins has been for some time an established principle for contr...

Epstein-Barr nuclear antigen 1 (EBNA-1) is the only viral protein required to support replication of Epstein-Barr virus during the latent phase of its life cycle. The DNA segment required for latent replication, oriP, contains two essential binding regions for EBNA-1, termed FR and DS, that are separated by 1 kilobase pair. The FR site appears to function as a replicational enhancer providing for the start of replication at the DS site. We have used electron microscopy to visualize the interaction of EBNA-1 with its binding sites and to study the mechanism for communication between the FR and DS sites. We have found that DNA-bound EBNA-1 forms a DNA loop between the FR and DS sites. From these results, we suggest that EBNA-1 bound to the replicational enhancer acts by a DNA-looping mechanism to facilitate the initiation of DNA replication. Occupancy of the DS site alone is highly sensitive to competition with nonspecific DNA. In contrast, occupancy of the DS site by looping from FR is largely resistant to the competitor DNA. These experiments support the concept that enhancers act in cis from nearby sites to provide a high local concentration of regulatory proteins at their target sites and to stabilize regulatory interactions. transcriptional regulation, recent work has demonstrated association ofDNA-bound proteins between sites involved in enhancer function by the bacterial NtrC protein (15), the mammalian Spl protein (16,17), the viral bovine papilloma E2 protein (18), and the Spl and E2 proteins (19). Our work on EBNA-1 has been directed toward two questions: Does replicational enhancement involve the protein-protein association ofDNA-bound EBNA-1? Why do enhancers typically act only from relatively close sites on the same DNA molecule?In the work reported here, we have used electron microscopy to study the interaction of EBNA-1 with its specific binding sites and to examine the ability of DNA-bound EBNA-1 to carry out protein-protein interactions. Our results demonstrate that EBNA-1 bound at the FR and DS sites associates to loop the intervening DNA. As reported in the accompanying paper, Frappier and O'Donnell (20) have also used electron microscopy to show this looping interaction. Based on the sensitivity of the EBNA-1 interaction at the DS site to competition with nonspecific DNA, we suggest that EBNA-1 acts in cis from the FR enhancer to stabilize the interaction of the protein at the DS site. Epstein-Barr virus (EBV) can establish a latent state in which the EBV genome is maintained as a circular plasmid (1, 2). Studies with plasmids derived from EBV DNA have established the functional requirements for replication during the latent state. The Epstein-Barr nuclear antigen 1 (EBNA-1) is the only viral protein required for replication from the EBV latent origin of replication, oriP (3, 4). The oriP region is composed of two essential segments separated by about 1 kilobase pair (kbp): (i) a family of repeats (FR) with 20 tandem copies of a 30-bp sequence and (ii) a dyad symmetry region (DS) with 4...

Communication between distant DNA sites is a central feature of many DNA transactions. Negative regulation of the galactose (ga/) operon of Escherichia coli requires repressor binding to two operator sites located on opposite sides of the promoter. The proposed mechanism for regulation involves binding of the repressor to both operator sites, followed by a protein-protein association that loops the intervening promoter DNA (double occupancy plus association). To assess these requirements in vivo, we have previously converted ga/operator sites to lac and shown that both operator sites must be occupied by the homologous repressor protein (Lac or Gal) for negative regulation of the ga/operon. We have now addressed more directly the need for proteinprotein association by the use of the converted operator sites and a mutant Lac repressor defective in association of the DNA-binding dimers. We have compared the biological and biochemical activity of two Lac repressors: the wild-type (tetramer) I + form, in which the DNA-binding dimer units are tightly associated; and the mutant I "~ repressor, in which the dimer units do not associate effectively. The I ÷ repressor is an efficient negative regulator of the ga/operon in vivo, but the I "~ mutant is an ineffective repressor. Purified I + repressor efficiently forms DNA loops between operator sites that we have visualized by electron microscopy; the I °~ repressor fails to form DNA loops, although the protein binds effectively to both operator sites. From the clear correlation between looping in vitro and repression in vivo, we conclude that regulation of the ga/operon depends on the association of repressor proteins bound to the two operator sites. Repression is likely to involve a DNA-wound nucleoprotein complex in which RNA polymerase is present but unable to carry out a productive interaction with the promoter sequence.

Anthocyanin biosynthesis requires the MYB-bHLH-WD40 protein complex to activate the late biosynthetic genes. LcMYB1 was thought to act as key regulator in anthocyanin biosynthesis of litchi. However, basic helix-loop-helix proteins (bHLHs) as partners have not been identified yet. The present study describes the functional characterization of three litchi bHLH candidate anthocyanin regulators, LcbHLH1, LcbHLH2, and LcbHLH3. Although these three litchi bHLHs phylogenetically clustered with bHLH proteins involved in anthcoyanin biosynthesis in other plant, only LcbHLH1 and LcbHLH3 were found to localize in the nucleus and physically interact with LcMYB1. The transcription levels of all these bHLHs were not coordinated with anthocyanin accumulation in different tissues and during development. However, when co-infiltrated with LcMYB1, both LcbHLH1 and LcbHLH3 enhanced anthocyanin accumulation in tobacco leaves with LcbHLH3 being the best inducer. Significant accumulation of anthocyanins in leaves transformed with the combination of LcMYB1 and LcbHLH3 were noticed, and this was associated with the up-regulation of two tobacco endogenous bHLH regulators, NtAn1a and NtAn1b, and late structural genes, like NtDFR and NtANS. Significant activity of the ANS promoter was observed in transient expression assays either with LcMYB1-LcbHLH1 or LcMYB1-LcbHLH3, while only minute activity was detected after transformation with only LcMYB1. In contrast, no activity was measured after induction with the combination of LcbHLH2 and LcMYB1. Higher DFR expression was also oberseved in paralleling with higher anthocyanins in co-transformed lines. LcbHLH1 and LcbHLH3 are essential partner of LcMYB1 in regulating the anthocyanin production in tobacco and probably also in litchi. The LcMYB1-LcbHLH complex enhanced anthocyanin accumulation may associate with activating the transcription of DFR and ANS.