Repressor structure and the mechanism of positive control

Both fis-dependent and factor-independent upstream activation of the rrnB P1 promoter are face of the helix dependent

Transcription from the Escherichia coli rrnB P1 promoter is increased by a cis-acting sequence which extends upstream of the -35 hexamer to about -150 with respect to the transcription initiation site, the Upstream Activation Region (UAR). Activation by the UAR involves two components: (1) a trans-acting protein, Fis, which binds to three sites in the UAR between -60 and -150, and (2) the UAR sequences themselves which affect RNA polymerase (RNAP) activity independent of other proteins. We refer to the latter as Factor-Independent Activation (FIA). In addition to its interactions with the -10 and -35 hexamers typical of E. coli promoters, RNAP makes contacts to the -53 region of rrnB P1, which may be related to the FIA effect. We constructed a series of insertion mutants containing integral and non-integral numbers of helical turns at position -46, between the Fis binding sites and the -35 region, and the resulting promoter activities were measured in vitro and in vivo. The data suggest that both Fis-dependent and factor-independent activation are face of the helix dependent: the Fis binding site and the sequences responsible for factor-independent activation must be correctly oriented relative to RNA polymerase in order to activate transcription. These results, in conjunction with other evidence, support a model for the involvement of direct Fis-RNAP interactions in upstream activation. We also demonstrate that RNAP interacts with the -53 region of the rrnB P1 UAR even when these sequences are displaced upstream of the RNAP binding site, and that these interactions correlate with factor-independent activation.

The basic region of myogenin cooperates with two transcription activation domains to induce muscle-specific transcription

Myogenin is a skeletal muscle-specific transcription factor that can activate myogenesis when introduced into a variety of nonmuscle cell types. Activation of the myogenic program by myogenin is dependent on its binding to a DNA sequence known as an E box, which is associated with numerous muscle-specific genes. Myogenin shares homology with MyoD and other myogenic regulatory factors within a basic region and a helix-loop-helix (HLH) motif that mediate DNA binding and dimerization, respectively. Here we show that the basic region-HLH motif of myogenin alone lacks transcriptional activity and is dependent on domains in the amino and carboxyl termini to activate transcription. Analysis of these N- and C-terminal domains through creation of chimeras with the DNA-binding domain of the Saccharomyces cerevisiae transcription factor GAL4 revealed that they act as strong transcriptional activators. These transcription activation domains are dependent for activity on a specific amino acid sequence within the basic region, referred to as the myogenic recognition motif (MRM), when an E box is the target for DNA binding. However, the activation domains function independent of the MRM when DNA binding is mediated through a heterologous DNA-binding domain. The activation domain of the acidic coactivator VP16 can substitute for the myogenin activation domains and restore strong myogenic activity to the basic region-HLH motif. Within a myogenin-VP16 chimera, however, the VP16 activation domain also relies on the MRM for activation of the myogenic program. These findings reveal that DNA binding and transcriptional activation are separable functions, encoded by different domains of myogenin, but that the activity of the transcriptional activation domains is influenced by the DNA-binding domain. Activation of muscle-specific transcription requires collaboration between the DNA-binding and activation domains of myogenin and is dependent on events in addition to DNA binding.

The basic region of myogenin cooperates with two transcription activation domains to induce muscle-specific transcription

Myogenin is a skeletal muscle-specific transcription factor that can activate myogenesis when introduced into a variety of nonmuscle cell types. Activation of the myogenic program by myogenin is dependent on its binding to a DNA sequence known as an E box, which is associated with numerous muscle-specific genes. Myogenin shares homology with MyoD and other myogenic regulatory factors within a basic region and a helix-loop-helix (HLH) motif that mediate DNA binding and dimerization, respectively. Here we show that the basic region-HLH motif of myogenin alone lacks transcriptional activity and is dependent on domains in the amino and carboxyl termini to activate transcription. Analysis of these N- and C-terminal domains through creation of chimeras with the DNA-binding domain of the Saccharomyces cerevisiae transcription factor GAL4 revealed that they act as strong transcriptional activators. These transcription activation domains are dependent for activity on a specific amino acid sequence within the basic region, referred to as the myogenic recognition motif (MRM), when an E box is the target for DNA binding. However, the activation domains function independent of the MRM when DNA binding is mediated through a heterologous DNA-binding domain. The activation domain of the acidic coactivator VP16 can substitute for the myogenin activation domains and restore strong myogenic activity to the basic region-HLH motif. Within a myogenin-VP16 chimera, however, the VP16 activation domain also relies on the MRM for activation of the myogenic program. These findings reveal that DNA binding and transcriptional activation are separable functions, encoded by different domains of myogenin, but that the activity of the transcriptional activation domains is influenced by the DNA-binding domain. Activation of muscle-specific transcription requires collaboration between the DNA-binding and activation domains of myogenin and is dependent on events in addition to DNA binding.

Dual Bar homeo box genes of Drosophila required in two photoreceptor cells, R1 and R6, and primary pigment cells for normal eye development

In the Bar mutation of Drosophila, ommatidial differentiation is known to be suppressed in the anterior portion of the eye. Our structural analysis shows that the Bar region contains a pair of homeo box genes, BarH1 and BarH2. These genes encode polypeptides similar in size and sequence and share a common homeo domain that is identical in sequence except for putative trans-activator-binding sites. We also show, by mosaic analysis and immunostaining with anti-BarH1/BarH2 antibodies, that BarH1 and BarH2 are not only specifically coexpressed but also functionally required in R1/R6 prephotoreceptors and primary pigment cells in developing ommatidia. In R1/R6, the expression of BarH1 and BarH2 appears to be regulated by rough and glass gene products. BarH1 and BarH2 proteins are essential to normal lens formation, formation of three types of pigment cells, and elimination of excess cells from mature ommatidia. Taken together, our results suggest that Bar homeo domain proteins may play key roles in the fate-determination processes of pigment cells and cone cells.

Regulation of muscle cell growth and differentiation by the MyoD family of helix-loop-helix proteins

The skeletal muscle cell system provides a powerful model for exploring the mechanistic basis for the antagonism between cell growth and differentiation. The recent identification of the MyoD family of muscle-specific transcription factors now offers opportunities to dissect at the molecular level of the mechanisms through which defined cell type-specific transcription factors can activate an entire differentiation program as well as to unravel the mechanisms through which growth factor and oncogenic signals can disrupt cellular differentiation. Because the mechanisms for growth factor signaling and induction of cell proliferation are conserved in diverse cell types, it is tempting to speculate that the molecular mechanisms responsible for the antagonism between cell proliferation and differentiation in muscle cells are also operative in other cell types. Resolution of this question, however, must await identification of the regulatory factors that specify cell fate in other lineages.

RNA polymerase bound to the PR promoter of bacteriophage lambda inhibits open complex formation at the divergently transcribed PRM promoter. Implications for an indirect mechanism of transcriptional activation by lambda repressor

We demonstrate that RNA polymerase bound at the PR promoter of bacteriophage lambda can repress transcription initiation from the divergently transcribed PRM promoter in vitro. Using abortive initiation and run-off transcription experiments we show that inactivating mutations introduced into either the -10 or -35 regions of PR result in a significant increase in the rate of formation of transcriptionally competent complexes at the PRM promoter. This is due primarily to an increase in the rate constant for the isomerization of closed to open complexes. Gel shift and DNase I footprinting experiments were employed to further define the mechanism by which PR sequences mediate PRM repression. From these assays we were able to conclude that the formation of an open complex at the PR promoter did not exclude RNA polymerase from binding at PRM. Rather, initiation at PRM was impaired because closed complexes must isomerize in the presence of an open complex already situated at the PR promoter. Extensive evidence has been obtained previously indicating that lambda repressor activates transcription directly by contacting RNA polymerase situated at the PRM promoter. Results presented here raise the possibility that an additional mechanism could be operative, whereby lambda repressor indirectly activates PRM transcription by excluding RNA polymerase from the PR promoter.

Regulatory crosstalk at composite response elements

Transcriptional regulatory factors from different families interact with each other when bound to DNA at composite response elements. This level of communication has two striking consequences: ubiquitous factors can effect cell specificity, and closely related factors from a given family can produce very different regulatory patterns.

Functional map of the alpha subunit of Escherichia coli RNA polymerase: two modes of transcription activation by positive factors

The role of the alpha subunit of Escherichia coli RNA polymerase in transcription activation by positive factors was investigated using two reconstituted mutant RNA polymerases (containing C-terminally truncated alpha subunits) and three positive factors [the cAMP receptor protein (CRP), OmpR, and PhoB]. The mutant RNA polymerases did not respond to transcription activation by activator proteins that bind upstream of the respective promoters. Transcription by these mutant enzymes was, however, activated in the cases where activators bind to target sites that overlap the promoter -35 region. Two different mechanisms are proposed for the positive control of transcription by activator proteins, one requiring the C-terminal domain of the alpha subunit, and the other not requiring it.

Muscle-specific transcriptional activation by MyoD

We focus on the mechanism by which MyoD activates transcription. Previous experiments showed that when the 13-amino-acid basic region of E12 replaced the corresponding basic region of MyoD, the resulting MyoD-E12Basic chimeric protein could bind specifically to muscle-specific enhancers in vitro and form dimers with E12, but could not activate a cotransfected reporter gene or convert 10T1/2 cells to muscle. Here we show that back mutation of this chimeric protein (with the corresponding residues in MyoD) re-establishes activation, and we identify a specific alanine involved in increasing DNA binding and a specific threonine required for activation. Using a reporter gene containing MyoD-binding sites located downstream from the transcription start site, we show that MyoD-E12Basic can bind in vivo and thereby inhibit expression of the reporter. In vivo binding is also supported by the fact that the addition of the "constitutive" VP16 activation domain to MyoD-E12Basic restores full trans-activation potential. The normal MyoD-activation domain maps within the amino-terminal 53 residues and can be functionally replaced by the activation domain of VP16. The activity of the MyoD-activation domain is dramatically elevated when deletions are made almost anywhere in the rest of the MyoD molecule, suggesting that the activation domain in MyoD is usually masked. Surprisingly, MyoD-E12Basic can activate transcription in CV1 and B78 cells (but not in 10T1/2 or 3T3 cells), suggesting that the activation function of the basic domain requires a specific factor present in CV1 and B78 cells. We propose that to function, the masked MyoD-activation domain requires the participation of a second factor that recognizes the basic region. We refer to such a factor as a recognition factor.

Mutagenesis of the myogenin basic region identifies an ancient protein motif critical for activation of myogenesis

Myogenin is a muscle-specific nuclear factor that acts as a genetic switch to activate myogenesis. Myogenin, MyoD, and a growing number of proteins implicated in transcriptional control share sequence homology within a basic region and an adjacent helix-loop-helix motif. Here we identify by site-directed mutagenesis a 12-amino acid subdomain of the myogenin basic region essential for binding of DNA and activation of myogenesis. The basic region of the widely expressed helix-loop-helix protein E12 is conserved at 8 of these 12 residues and can mediate DNA binding when placed in myogenin, but it cannot activate myogenesis. Replacement of each of the four nonconserved residues of the myogenin basic region with the corresponding residues of E12 reveals two adjacent amino acids (Ala86-Thr) that can impart muscle specificity to the basic region. These residues are specific to, and conserved in, the basic regions of all known myogenic helix-loop-helix proteins from Drosophila to man, suggesting that they constitute part of an ancient protein motif required for activation of the myogenic program.

Activation of the trpBA promoter of Pseudomonas aeruginosa by TrpI protein in vitro

We have developed an in vitro transcription system in which purified TrpI protein and indoleglycerol phosphate (InGP) activate transcription initiation at the trpBA promoter (trpPB) and repress initiation at the trpI promoter (trpPI) of Pseudomonas aeruginosa. The phenotypes resulting from mutations in the -10 region of both promoters indicate that the -10 region consensus sequence in P. aeruginosa is probably the same as that in Escherichia coli. Furthermore, in the absence of TrpI and InGP, the activities of the two promoters are inversely correlated: down mutations in trpPI lead to increased activity of trpPB, and up mutations in trpPB cause a decrease in trpPI activity. These results are a consequence of the fact that the two promoters overlap, so that RNA polymerase cannot form open complexes with both promoters simultaneously. Thus, in theory, by preventing RNA polymerase from binding at trpPI, TrpI protein could indirectly activate trpPB. However, oligonucleotide-induced mutations that completely inactivate trpPI do not relieve the requirement for TrpI and InGP to activate trpPB. Therefore, activation of trpPB is mediated by a direct effect of TrpI on transcription initiation at trpPB. In addition, the oligonucleotide-induced mutations in trpPI alter site II, the weaker of two TrpI binding sites identified in DNase I and hydroxyl radical footprinting studies (M. Chang and I. P. Crawford, Nucleic Acids Res. 18:979-988, 1990). Since these mutations prevent full activation of trpPB, we conclude that specific base pairs in site II are required for activation.

Cassette mutagenesis implicates a helix-turn-helix motif in promoter recognition by the novel RNA polymerase sigma factor sigma 54

Cassette mutagenesis has been used to study the role of a helix-turn-helix (HTH) motif in the novel RNA polymerase sigma factor sigma 54 of Klebsiella pneumoniae. Of the four residues which are predicted to be solvent-exposed in the second helix, the first (Glu-378) tolerated all substitutions, and some mutations of this residue increased expression from sigma 54-dependent promoters. Certain substitutions in the third exposed residue (Ser-382) produced a promoter-specific phenotype and all substitutions in the fourth residue (Arg-383) inactivated the protein, identifying this residue as being likely to be involved in base-specific interactions with the promoter. In vivo footprinting indicated that the inactive HTH mutants of sigma 54 were defective in interaction with both the -24 and -12 regions of the glnAp2 promoter.

Analysis of the DNA-binding and activation properties of the human transcription factor AP-2

The mammalian transcription factor AP-2 is a sequence-specific DNA-binding protein expressed in neural crest lineages and regulated by retinoic acid. Here we report a structure/function analysis of the DNA-binding and transcription activation properties of the AP-2 protein. DNA contact studies indicate that AP-2 binds as a dimer to a palindromic recognition sequence. Furthermore, cross-linking and immunoprecipitation data illustrate that AP-2 exists as a dimer even in the absence of DNA. Examination of cDNA mutants reveals that the sequences responsible for DNA binding are located in the carboxy-terminal half of the protein. In addition, a domain mediating dimerization forms an integral component of this DNA-binding structure. Expression of AP-2 in mammalian cells demonstrates that transcriptional activation requires an additional amino-terminal domain that contains an unusually high concentration of proline residues. This proline-rich activation domain also functions when attached to the heterologous DNA-binding region of the GAL4 protein. This study reveals that although AP-2 shares an underlying modular organization with other transcription factors, the regions of AP-2 involved in transcriptional activation and DNA binding/dimerization have novel sequence characteristics.

Mutational analysis of a bacteriophage P4 late promoter

Transcription from the late Psid promoter of satellite bacteriophage P4 is dependent on the bacterial RNA polymerase carrying the sigma 70 subunit and is positively regulated by the product of the P4 delta gene or the ogr gene of helper bacteriophage P2. Through deletion and mutational analyses of the Psid promoter, we identified mutations in the -10 region and in a region of hyphenated dyad symmetry centered around position -55 that inactivate Psid. Most of these mutations alter base pairs that are highly conserved in the five other delta-activated P4 and P2 late promoters. We propose that the P4 delta and P2 ogr gene products bind the -55 region of the P4 and P2 late promoters.

RNA polymerases from Pseudomonas aeruginosa and Pseudomonas syringae respond to Escherichia coli activator proteins

The activities of RNA polymerases (RNAPs) from Pseudomonas aeruginosa and Pseudomonas syringae were compared with that of Escherichia coli RNAP. All three enzymes are able to initiate transcription at the trpBA promoter of P. aeruginosa and at the coliphage lambda promoters, pRM and pRE, in response to heterospecific activators (TrpI protein, repressor, and cII protein, respectively). However, both Pseudomonas polymerases have less stringent requirements for promoter recognition in the absence of activators than does E. coli RNAP.

Interconversion of the DNA-binding specificities of two related transcription regulators, CRP and FNR

In Escherichia coli, FNR and CRP are homologous transcriptional regulators which recognize similar nucleotide sequences via DNA-binding domains containing analogous helix-turn-helix motifs. The molecular basis for recognition and discrimination of their target sites has been investigated by directed amino acid substitutions in the corresponding DNA-recognition helices. In FNR, Glu-209 and Ser-212 are essential residues for the recognition of FNR sites. A V208R substitution confers CRP-site specificity without loss of FNR specificity, but this has adverse effects on anaerobic growth. In contrast, changes at two (V208R and E209D) or three (V208R, S212G and G216K) positions in FNR endow a single CRP-site binding specificity. In reciprocal experiments, two substitutions (R180V and G184S) were required to convert the binding specificity of CRP to that of FNR. Altering Asp-199 in FNR failed to produce a positive control phenotype, unlike substitutions at the comparable site in CRP. Implications for the mechanism of sequence discrimination by FNR and CRP are discussed.

Identification of two polypeptide segments of CCAAT/enhancer-binding protein required for transcriptional activation of the serum albumin gene

We used molecular genetic methods to generate systematically altered forms of CCAAT/enhancer-binding protein (C/EBP). The aim of our experiments was to identify regions of C/EBP that contribute to its capacity to activate transcription from the promoter of the serum albumin gene in cultured hepatoma cells. Earlier experiments had shown that the DNA-binding domain must remain intact for C/EBP to activate albumin transcription. We now provide evidence of two additional elements of C/EBP that are required for its gene-activating role. One such element occurs within a 28-residue region located close to the amino terminus of the protein. The other maps to a broader, more internal region of the protein and appears to exhibit functional redundancy. These newly defined elements of C/EBP exhibit two characteristics of "activation" domains delineated in studies of other gene regulatory proteins. First, they play no obvious role in the capacity of C/EBP to bind to its DNA substrate. Second, they retain function after being appended onto the DNA-binding domain of a different protein. Neither of these putative activating elements is characterized by overt distinction in either charge or preponderance of any particular amino acid. The more amino-terminal element does, however, exhibit several features suggesting that it may assume an alpha-helical structure. These studies offer observations and reagents that will be valuable for future studies concerning the physiologic function of C/EBP.

Escherichia coli metR mutants that produce a MetR activator protein with an altered homocysteine response

Using an Escherichia coli lac deletion strain lysogenized with a lambda phage carrying a metH-lacZ gene fusion, we isolated trans-acting mutations that result in simultaneous 4- to 6-fold-elevated metH-lacZ expression, 5- to 22-fold-lowered metE-lacZ expression, and 9- to 20-fold-elevated metR-lacZ expression. The altered regulation of these genes occurs in the presence of high intracellular levels of homocysteine, a methionine pathway intermediate which normally inhibits metH and metR expression and stimulates metE expression. P1 transductions and complementation tests indicate that the mutations are in the metR gene. Our data suggest that the mutations result in an altered MetR activator protein that has lost the ability to use homocysteine as a modulator of gene expression.

A mechanism for synergistic activation of a mammalian gene by GAL4 derivatives

In prokaryotes and eukaryotes many gene activators work synergistically. For example, two dimers of lambda repressor interact to promote binding of these proteins to DNA, a reaction that is crucial at the repressor concentrations found in lysogens. In this case one of the bound dimers activates transcription, evidently by touching RNA polymerase. In another example, the yeast transcriptional activator GAL4, which can stimulate transcription in many eukaryotes, binds to multiple sites on DNA to activate transcription synergistically; the presence of two such sites can elicit a level of transcription more than twice that found with a single site. In this paper we show that synergistic activation by each of several GAL4 derivatives involves a mechanism different from that illustrated by the lambda repressor: multiple activator molecules can work synergistically under conditions in which their binding sites on DNA are saturated. The accompanying paper shows that under similar conditions of activator excess, GAL4 derivatives work synergistically with a heterologous mammalian gene activator. These results support the idea that eukaryotic activators can cooperate not by directly interacting but by simultaneously touching some component(s) of the transcriptional machinery.

Activation of the bacteriophage Mu lys promoter by Mu C protein requires the sigma 70 subunit of Escherichia coli RNA polymerase

Bacteriophage Mu C protein, a product of the middle operon, is required for activation of the four Mu late promoters. To address its mechanism of action, we overproduced the approximately 16.5-kilodalton C protein from a plasmid containing the C gene under the control of a phage T7 promoter and ribosome-binding site. A protein fraction highly enriched for Escherichia coli RNA polymerase (E sigma 70) and made from the overproducing strain was able to activate transcription in vitro from both the tac promoter (Ptac) and a Mu late promoter, Plys. The behavior of Plys was similar in vivo and in vitro; under both conditions, transcription was C dependent and the RNA 5' ends were identical. When anti-sigma 70 antibody was added to C-dependent transcription reactions containing both Ptac and Plys templates, transcription from both promoters was inhibited; transcription was restored by the addition of excess E sigma 70. This result suggests that C-dependent activation of Plys requires sigma 70. Further supporting evidence was provided by a reconstitution experiment in which an E sigma 70-depleted fraction containing C was unable to activate transcription from Plys unless both purified sigma 70 and core polymerase were added. These results strongly suggest that C is not a new sigma factor but acts as an activator for E sigma 70-dependent transcription.

Mutations that increase the activity of a transcriptional activator in yeast and mammalian cells

Activating region I of GAL4 protein is a stretch of amino acids, positioned adjacent to the DNA-binding region, that activates transcription in yeast and, as we show here, in mammalian cells. Here we describe mutations located throughout a 65-amino acid region that increase the activation function of region I. Most of these mutations replace positively charged amino acids in the region with neutral ones, although we also describe substitutions at one position that do not alter the charge of the region. Mutations of region I that alter the activation function in yeast have similar effects on activation when assayed in mammalian cells. When individual mutations that raise the acidity of the activating region are recombined, the activities of the mutant proteins increase with increasing negative charge in both yeast and mammalian cells. These results extend and modify the correlation between acidity and activation and suggest that the requirements for a strong activating region are conserved in yeast and mammals.

Integration host factor is necessary for lysogenization of Escherichia coli by bacteriophage P2

Whether infection by bacteriophage P2 results in lysogenization of the host or vegetative growth of the phage depends upon a race between transcription from the repressor promoter Pc and the early promoter Pe; transcription from these promoters is mutually exclusive, since the Pc repressor Cox is formed from the Pe transcript and the Pe repressor C from the Pc transcript. The involvement of integration host factor (IHF) in the lysogenization of Escherichia coli K12 by P2 was tested by comparing wild-type and IHF-deficient (himA and himD) mutants. No lysogenic clones were formed following infection of the mutant bacteria. A switch plasmid that contains Pc-C-cat and Pe-cox-kan was used to test the choice for expression of Pc versus Pe. In the wild-type K12 bacteria, 20% of the clones expressed Pe transcription and 80% Pc transcription, whereas all transformed IHF-defective clones expressed transcription from Pe only. The effects of IHF on the in vivo expression of the Pe and Pc promoters were only marginal. The IHF protein was found to bind upstream of the Pe promoter, where a potential ihf sequence is located.

Bending of DNA by gene-regulatory proteins: construction and use of a DNA bending vector

The binding of a protein to its specific sequence, borne on a DNA fragment, retards the mobility of the fragment in a characteristic way during gel electrophoresis. If the protein induces bending in the DNA, the contortion can also be monitored by gel electrophoresis, because the amount of retardation of the mobility of the DNA-protein complex is dependent upon the position and the degree of the bend induced in the DNA fragment [Wu and Crothers, Nature 308 (1984) 509-513]. We have constructed a plasmid, pBend2, which can generate a large number of DNA fragments of identical length in which the protein-binding nucleotide sequence is located in circular permutations. The vector contains two identical DNA segments containing 17 restriction sites in a direct repeat spanning a central region containing cloning sites. The protein-binding sequence is inserted at one of these cloning sites. To investigate the functional significance of bending, we have compared, using pBend2, the cAMP.cAMP-receptor protein (CPR)-induced bending of CRP-binding sites found in five different genes of Escherichia coli. We have also shown that the bacteriophage lambda 0R1 operator DNA is bent when complexed with the CI or Cro repressor of the phage.

Mutations that alter transcriptional activation but not DNA binding in the zinc finger of yeast activator HAPI

Transcription of eukaryotic genes requires an interaction between transcription factors that bind to the TATA box region, and transcriptional activators that bind to upstream activating sequences (UASs) or enhancers. Several yeast upstream transcriptional activators, such as GCN4, GAL4 and HAP1, seem to contain separate domains for binding to DNA and activating transcription. The expression of the cytochrome genes CYC1 and CYC7 is controlled by HAP1, which binds to dissimilar DNA sequences in UAS1 of CYC1 and the UAS of CYC7. HAP1 has a zinc-finger DNA-binding domain between amino-acid residues 1 and 148, and a highly acidic C-terminal activation domain between residues 1,308 and 1,483 (ref. 10). A mutant allele of the HAP1 gene, HAP1-18, leads to a change in Ser 63 to Arg 63, immediately adjacent to the zinc finger in the DNA-binding domain. The HAP1-18 mutation specifically abolishes the ability of the protein to bind to UAS1, but greatly increases the ability of the protein to activate transcription of CYC7. We now report that this increase in activation is mediated solely by the CYC7 UAS and the HAP1-18 protein, and also, that it is not caused by an altered binding affinity of the protein for the CYC7 UAS. Furthermore, even by substituting other amino acids at position 63 and over-expressing the resulting derivatives in vivo we were unable to increase activity at the UAS of CYC7 to the level obtained with HAP1-18. This rules out the possibility that the HAP1-18 mutation increases transcriptional activation by abolishing competition by UAS1 and UAS1-like sites for the protein. We thus conclude that HAP1-18 is a better activator of transcription than the wild-type protein when bound to the UAS of CYC7. Moreover, our findings indicate that in addition to the acidic activation domain, the zinc-finger DNA-binding domain participates directly in the activation of transcription.

The structure of the Antennapedia homeodomain determined by NMR spectroscopy in solution: comparison with prokaryotic repressors

The structure of the Antennapedia homeodomain from Drosophila melanogaster was determined by nuclear magnetic resonance spectroscopy in solution. It includes three well-defined helices (residues 10-21, 28-38, and 42-52) and a more flexible fourth helix (residues 53-59). Residues 30-50 form a helix-turn-helix motif virtually identical to those observed in various prokaryotic repressors. Further comparisons of the homeodomain with prokaryotic repressors showed that there are also significant differences in the molecular architectures. Overall, these studies support the view that the third helix of the homeodomain may function as the DNA recognition site. The elongation of the third helix by the fourth helix is a structured element that so far appears to be unique to the Antennapedia homeodomain.

The Oct-1 homoeodomain directs formation of a multiprotein-DNA complex with the HSV transactivator VP16

The herpes simplex virus transactivator VP16 participates in the formation of a multiprotein-DNA complex with the ubiquitous octamer-motif-binding factor Oct-1. Complex formation is dependent on specific amino acids in the Oct-1 homoeodomain which are in positions analogous to positive control mutations in helix 2 of the lambda phage repressor helix-turn-helix motif, indicating that this structure is an ancient target for protein-protein interactions mediating transcriptional control.

The evolutionary conservation of eukaryotic gene transcription

The basic components required for eukaryotic gene transcription have been highly conserved in evolution. Structural and functional homology has now been documented among promoters, promoter factors, regulatory proteins, and RNA polymerases from eukaryotes as diverse as yeast and mammals. The ability of these proteins and DNA sequences to function across phylogenetic boundaries demonstrates that common molecular mechanisms underlie gene control in all eukaryotic cells, and provides the basis for powerful new approaches to the study of eukaryotic gene transcription.

Mutations in the glucocorticoid receptor zinc finger region that distinguish interdigitated DNA binding and transcriptional enhancement activities

Mammalian glucocorticoid receptors bind specifically to glucocorticoid response element (GRE) DNA sequences and enhance transcription from GRE-linked promoters in mammalian cells and in yeast. We randomly mutagenized a segment of the receptor encompassing sequences responsible for DNA-binding and transcriptional regulation and screened in yeast for receptor defects. The mutations all mapped to a 66-amino-acid subregion that includes two zinc fingers; in general parallel phenotypes were observed in yeast and animal cells. Mutants defective for DNA binding also failed either to enhance or to repress transcription. However, several mutations in the second finger selectively impaired enhancement; we suggest that such 'positive control' mutants may alter protein-protein contacts required for transcriptional activation.

The repressor of the PEP:fructose phosphotransferase system is required for the transcription of the pps gene of Escherichia coli

We have cloned the pps gene, coding for PEP synthase, of Escherichia coli. PEP synthase catalyses the ATP-dependent conversion of pyruvate into phosphoenol-pyruvate and is required for gluconeogenesis. The pps gene was cloned by an in vivo cloning method using a mini-Mulac bacteriophage containing a plasmid replicon. Upon expression of the cloned pps gene in the maxicell system a protein with an apparent molecular weight of 84 kDa was synthesized. The position of the pps gene of the plasmid was localized by restriction analysis of isolated transposon insertions and the determination of the PEP synthase activities of the different clones. An operon fusion between the pps gene and the galK gene was constructed. Measurements of the galactokinase activity in Salmonella typhimurium galK and galK fruR mutants showed that the transcription of the pps gene requires the presence of FruR, the repressor of the PEP: fructose phosphotransferase system (PTS) in E. coli and S. typhimurium. To test whether the components of the Fructose PTS, in particular FPr, are involved in the expression of the pps gene, we investigated a S. typhimurium galK strain, containing the fusion plasmid, in which the chromosomal fru operon was inactivated by a transposon insertion. Measurements of the galactokinase activity showed that the absence of the Fructose PTS proteins has no significant influence on the regulation of the pps gene.

Genetic analysis of transcriptional activation and repression in the Tn21 mer operon

Transcription of the Tn21 mercury resistance operon (mer) is controlled by the toxic metal cation Hg(II). This control is mediated by the product of the merR gene, a 144-amino-acid protein which represses transcription of the structural genes (merTPCAD) in the absence of Hg(II) and activates transcription in the presence of Hg(II). We have used a mer-lac transcriptional fusion to obtain regulatory mutants in this metal-responsive system. Some mutants were defective in Hg(II)-induced activation while retaining repression function (a- r+), others were defective in repression but not activation (a+ r-), and some had lost both functions (a- r-). Mutations in three of the four cysteine residues of merR resulted in complete loss of Hg(II)-inducible activation but retention of the repressor function, suggesting that these residues serve as ligands for Hg(II) in the activation process. Other lesions adjacent to or very near these cysteines exhibited severely reduced activation and also retained repressor function. There were two putative helix-turn-helix (HTH) domains in merR, and mutants in each had very different phenotypes. A partially dominant mutation in the more amino-terminal region of the two putative HTH regions resulted in loss of both activation and repression (a- r-), consistent with a role for this region in DNA binding. Mutations in the more centrally located HTH region resulted only in loss of Hg(II)-induced activation (a- r+). Lesions in the central and in the carboxy-terminal regions of merR exhibited both Hg(II)-independent and Hg(II)-dependent transcriptional activation, suggesting that elements important in the activation mechanism may be widely distributed in this relatively small protein. The sole cis-acting mutant obtained with this operon fusion strategy, a down-promoter mutation, lies in a highly conserved base in the -35 region of the merTPCAD promoter.

Genetic structure of the bacteriophage P22 PL operon

The sequence of 1416 base-pairs of the P22 PL operon was determined, linking a continuous sequence from PL through abc2. P22 mutants bearing deletions in the sequenced region were constructed and tested for their phenotypes. Plasmids were constructed to express PL operon genes singly and in combination from Plac UV5. Two previously known genes, 17 and c3, are located within this sequence. In addition, three new genes have been identified: ral, kil and arf. Genes ral and c3 are homologous, as well as functionally analogous, to lambda ral and cIII, respectively. P22 kil, like lambda kil, kills the host cell when it is expressed. The two kil genes, although analogous in cell killing and map location, have no apparent sequence homology. The functions of the P22 and lambda kil genes are unknown; however, P22 kil is essential for lytic growth in the absence of abc. Gene arf (accessory recombination function) is located just upstream from erf; it is essential for P22 growth in the absence of kil or other genes upstream in PL. The growth defect of P22 bearing a deletion that removes arf is complemented by expression of either arf or the lambda red genes from plasmids. Sequences that include the stop codon for gene 17 may form a small stem-loop structure and are nearly identical to lambda sequences that contain the stop codon for ssb, which is near lambda tL 2b. Plasmids that include the P22 structure negatively regulate kil gene expression in cis.

Altered promoter recognition by mutant forms of the sigma 70 subunit of Escherichia coli RNA polymerase

We have systematically assayed the in vivo promoter recognition properties of 13 mutations in rpoD, the gene that encodes the sigma 70 subunit of Escherichia coli RNA polymerase holoenzyme, using transcriptional fusions to 37 mutant and wild-type promoters. We found three classes of rpoD mutations: (1) mutations that suggest contacts between amino acid side-chains of sigma 70 and specific bases in the promoter; (2) mutations that appear to affect either sequence independent contacts to promoter DNA or isomerization of the polymerase; and (3) mutations that have little or no effect on promoter recognition. Our results lead us to suggest that a sequence near the C terminus of sigma 70, which is similar to the helix-turn-helix DNA binding motif of phage and bacterial DNA binding proteins, is responsible for recognition of the -35 region, and that a sequence internal to sigma 70, in a region which is highly conserved among sigma factors, recognizes the -10 region of the promoter. rpoD mutations that lie in the recognition helix of the proposed helix-turn-helix motif affect interactions with specific bases in the -35 region, while mutations in the upstream helix, which is thought to contact the phosphate backbone, have sequence-independent effect on promoter recognition.

Mutations in RNA polymerase II enhance or suppress mutations in GAL4

The activation domains of eukaryotic DNA-binding transcription factors, such as GAL4, may regulate transcription by contacting RNA polymerase II. One potential site on RNA polymerase II for such interactions is the C-terminal tandemly repeated heptapeptide domain in the largest subunit (RPO21). We have changed the number of heptapeptide repeats in this yeast RPO21 C-terminal domain and have expressed these mutant RNA polymerase II polypeptides in yeast cells containing either wild-type or defective GAL4 proteins. Although the number of RPO21 heptapeptide repeats had no effect on the activity of wild-type GAL4, changing the length of the C-terminal domain modified the ability of mutant GAL4 proteins to activate transcription. Shorter or longer RPO21 C-terminal domains enhanced or partially suppressed, respectively, the effects of deletions in the transcriptional-activation domains of GAL4. The same RPO21 mutations also affected transcriptional activation by a GAL4-GCN4 chimera. These data suggest that the activation domains of DNA-binding transcription factors could interact, either directly or indirectly, with the heptapeptide repeats of RNA polymerase II.

Yeast GCN4 transcriptional activator protein interacts with RNA polymerase II in vitro

Regulated transcription by eukaryotic RNA polymerase II (Pol II) requires the functional interaction of multiple protein factors, some of which presumably interact directly with the polymerase. One such factor, the yeast GCN4 activator protein, binds to the upstream promoter elements of many amino acid biosynthetic genes and induces their transcription. Through the use of affinity chromatography involving GCN4- or Pol II-Sepharose columns, we show that GCN4 interacts specifically with Pol II in vitro. Purified Pol II is retained on the GCN4-Sepharose column under conditions in which the vast majority of proteins flow through. Moreover, Pol II can be selectively isolated from more complex mixtures of proteins. Conversely, GCN4 protein, synthesized in vitro or in Escherichia coli, specifically binds to the Pol II-Sepharose column under equivalent conditions. Using deletion mutants, we also show that the DNA-binding domain of GCN4 is both necessary and sufficient for this interaction. We suggest the possibility that this GCN4-Pol II interaction may be important for transcription in vivo.

Role of bacteriophage Mu C protein in activation of the mom gene promoter

The phage Mu C gene product is a specific activator of Mu late gene transcription, including activation of the mom operon. Fusion of the C gene to the efficient translation initiation region of the Escherichia coli atpE gene allowed significant overproduction of C protein, which was subsequently purified and assayed for DNA binding by gel retardation and nuclease footprinting techniques. C protein binds to a site immediately upstream of the -35 region both of the mom promoter and the related phage D108 mod promoter. The location of the mom promoter has been determined by primer extension. Upstream deletions extending more than 3 base pairs into the C-binding site abolished activation of the mom promoter in vivo. In vitro binding of C was not significantly affected by DNA methylation. A second, C-dependent promoter was identified just downstream of the C coding region; comparison with the mom promoter revealed common structural elements.

Bacteriophage Mu late promoters: four late transcripts initiate near a conserved sequence

Late transcription of bacteriophage Mu, which results in the expression of phage morphogenetic functions, is dependent on Mu C protein. Earlier experiments indicated that Mu late RNAs originate from four promoters, including the previously characterized mom promoter. S1 nuclease protection experiments were used to map RNA 5' ends in the three new regions. Transcripts were initiated at these points only in the presence of C and were synthesized in a rightward direction on the Mu genome. Amber mutant marker rescue analysis of plasmid clones and limited DNA sequencing demonstrated that these new promoters are located between C and lys, upstream of I, and upstream of P within the N gene. A comparison of the promoter sequences upstream from the four RNA 5' ends yielded two conserved sequences: the first (tA . . cT, where capital and lowercase letters indicate 100 and 75% base conservation, respectively), at approximately -10, shares some similarity with the consensus Escherichia coli sigma 70 -10 region, while the second (ccATAAc CcCPuG/Cac, where Pu indicates a purine), in the -35 region, bears no resemblance to the E. coli -35 consensus. We propose that these conserved Mu late promoter consensus sequences are important for C-dependent promoter activity. Plasmids containing transcription fusions of these late promoters to lacZ exhibited C-dependent beta-galactosidase synthesis in vivo, and C was the only Mu product needed for this transactivation. As expected, the late promoter-lacZ fusions were activated only at late times after induction of a Mu prophage. The C-dependent activation of lacZ fusions containing only a few bases of the 5' end of Mu late RNA and the presence of altered promoter sequences imply that C acts at the level of transcription initiation.

The MerR heavy metal receptor mediates positive activation in a topologically novel transcription complex

Several physical and chemical signals from the extracellular environment are known to be transduced into changes in gene expression through multiple step pathways; however, mechanisms for triggering cellular responses to heavy metal stress have yet to be elucidated. We demonstrate here one such mechanism that employs a single heavy metal receptor protein, MerR, to directly activate transcription of the bacterial mercuric ion resistance operon. The mercuric ion-MerR complex and E. coli RNA polymerase holoenzyme synergistically bind to the metal responsive promoter in an unprecedented spatial relationship to form transcriptionally competent complexes. The activator binds adjacent to and overlaps with the polymerase molecule between the consensus -35 and -10 promoter regions. Our results support a model for transcriptional activation that includes both effector-induced protein-protein interactions and activator-induced alteration in DNA structure.

Structure of the amino-terminal domain of phage 434 repressor at 2.0 A resolution

The crystal structure of the amino-terminal domain of phage 434 repressor has been solved using molecular replacement methods and refined to an R-factor of 19.3% against data to 2.0 A resolution. The protein comprises five short alpha-helices. Two of these form a helix-turn-helix motif, very similar to those found in related proteins. The protein is remarkably similar to the Cro protein from the same phage.

The POU domain is a bipartite DNA-binding structure

The POU domain (pronounced 'pow') is a highly charged 155-162-amino-acid (aa) region of sequence similarity contained within three mammalian transcription factors. Pt-1 (ref. 2), Oct-1 (ref. 3) and Oct-2 (ref. 4), and the product of the nematode gene unc-86 (ref. 5) which is involved in determining neural cell lineage. This domain consists of two subdomains, a C-terminal homoeo domain and an N-terminal POU-specific region separated by a short nonconserved linker; the sequence relationship shows that the POU homoeo domains form a distinct POU-related family. In the ubiquitous and lymphoid-specific octamer-motif binding proteins Oct-1 and Oct-2, the POU domain is sufficient for sequence-specific DNA binding. Homoeobox domains contain a helix-turn-helix DNA-binding motif, first identified in bacterial repressors. The helix-turn-helix region of the POU domain is important for DNA binding and, in other classes of homoeo-containing proteins, the entire homoeo domain is sufficient for DNA binding; thus the new POU-specific region could be involved in other functions such as protein-protein interactions. Nevertheless, we show here that in fact the POU domain is a novel bipartite DNA-binding structure in which the POU homoeo and POU-specific regions form two subdomains that are both required for DNA binding but are held together by a flexible linker.

The jun proto-oncogene is positively autoregulated by its product, Jun/AP-1

Binding of the human transcription factor Jun/AP-1 to a conserved 8 bp nucleotide sequence (TRE) is responsible for increased transcription of different cellular genes in response to tumor promoters, such as TPA, and serum factors. Enhanced Jun/AP-1 activity in TPA-stimulated cells is regulated by two different mechanisms: a posttranslational event acting on pre-existing Jun/AP-1 molecules, and transcriptional activation of jun gene expression leading to an increase in the total amount of Jun/AP-1. Induction of jun transcription in response to TPA is mediated by binding of Jun/AP-1 to a high-affinity AP-1 binding site in the jun promoter region. Site-specific mutagenesis of this binding site prevents TPA induction and trans-activation by Jun/AP-1. These results clearly demonstrate that jun transcription is directly stimulated by its own gene product. This positive regulatory loop is likely to be responsible for prolonging the transient signals generated by activation of protein kinase C.

Interaction at a distance between lambda repressors disrupts gene activation

The lambda repressor is an activator as well as a repressor of transcription. The activation function is blocked by interaction with another lambda repressor molecule bound upstream on the same DNA molecule. This example of negative control at a distance involves formation of a DNA loop.

Structure of the lambda complex at 2.5 A resolution: details of the repressor-operator interactions

The crystal structure of a complex containing the DNA-binding domain of lambda repressor and a lambda operator site was determined at 2.5 A resolution and refined to a crystallographic R factor of 24.2 percent. The complex is stabilized by an extensive network of hydrogen bonds between the protein and the sugar-phosphate backbone. Several side chains form hydrogen bonds with sites in the major groove, and hydrophobic contacts also contribute to the specificity of binding. The overall arrangement of the complex is quite similar to that predicted from earlier modeling studies, which fit the protein dimer against linear B-form DNA. However, the cocrystal structure reveals important side chain-side chain interactions that were not predicted from the modeling or from previous genetic and biochemical studies.

Two trans-acting regulatory genes (vir and mod) control antigenic modulation in Bordetella pertussis

Expression of virulence factors by Bordetella pertussis is altered by environmental signals (antigenic modulation) and is dependent on an activator encoded by a gene called vir. We have used TnphoA (Tn5 IS50L::phoA) gene fusions to define two sets of genes whose expression is either activated (vag loci) or repressed (vrg loci) by modulation signals. Both groups of genes appear to be regulated by the vir gene product in that, in the absence of modulators, null mutations in vir lead to the repression of vag gene fusions and derepression of vrg gene fusions. Mutants of B. pertussis were isolated that constitutively express virulence factors in the presence of the modulator MgSO4, nicotinic acid, or low incubation temperature. We designate the gene that carries such mutations mod (modulation) and have characterized one (mod-1) of these mod constitutive mutations. A method was developed for the insertional inactivation of the vir gene by using the integration of a suicide replicon. Inactivation of the vir gene in the mod-1 mutant, followed by transcomplementation with the cloned wild-type vir gene, gives the Mod-1 constitutive phenotype, showing that the mod-1 mutation defines a gene distinct from vir. The gene carrying the mod-1 mutation is linked to vir and was cloned on a recombinant cosmid (pLAF-C1) which transcomplements the vir-1::Tn5 mutation in B. pertussis 347. Introduction of pLAF-C1 into vir mutant and vir+ B. pertussis strains also gives the Mod-1 constitutive phenotype, indicating that mod-1 is a dominant allele. These data suggest that the mod gene product could have sensory functions for the environmental signals that affect the expression of vir-regulated genes of B. pertussis. The mod constitutive strains and plasmids described here also have applications in pertussis vaccine development.

The Azorhizobium caulinodans nifA gene: identification of upstream-activating sequences including a new element, the 'anaerobox'

From nucleotide sequencing analyses, the A. caulinodans nifA gene seems to be under dual control by the Ntr (in response to available N) and Fnr (in response to available O2) transcriptional activation/repression systems. Because it fixes N2 in two contexts, the Ntr system might regulate A. caulinodans nif gene expression ex planta, while the Fnr system might similarly regulate in planta. As nifA upstream-activating elements, we have identified: (i) a gpNifA binding site allowing autogenous nifA regulation, (ii) an Ntr-dependent transcription start, presumably the target of gpNifA activation, and (iii) an "anaerobox" tetradecameric nucleotide sequence that is precisely conserved among O2 regulated enteric bacterial genes controlled by the gpFnr transcriptional activator. Because it is precisely positioned upstream of enteric bacterial transcriptional starts, the "anaerobox" sequence may constitute the gpFnr DNA binding site. If so, then a second, Ntr-independent nifA transcription start may exist. We have also deduced the A. caulinodans nifA open reading frame and have compared the gene product (gpNifA) with those of other N2-fixing organisms. These proteins exhibit strongly conserved motifs: (i) sites conserved among ATP-binding proteins, (ii) an interdomain linker region, and (iii) a C-terminal alpha-helix-turn-alpha-helix DNA binding site.