TATA-dependent enhancer stimulation of promoter activity in mice is developmentally acquired

Herpes simplex virus (HSV) thymidine kinase (tk) promoter activity depends on four transcription factor binding sites, one of which is a TATA box sequence, and the presence of either a cis-acting enhancer sequence or a transactivator protein. Studies presented here show that this TATA box was required for promoter activity only after cells began to differentiate and then only when promoter activity was stimulated by either an enhancer or a transactivator. When the HSV tk promoter was utilized by mouse embryos from the one-cell to eight-cell stage of development or by undifferentiated mouse embryonic stem cells, disruption of the HSV tk TATA box by site-specific mutations did not reduce promoter activity. This was true even when HSV tk promoter activity was stimulated strongly by either the embryo-responsive polyomavirus F101 enhancer or its natural transactivator, the HSV ICP4 gene product. However, stimulated expression was dependent on a distal Sp1 DNA binding site. Similarly, disruption of the TATA box did not reduce tk promoter activity in primary mouse embryonic fibroblasts or in immortalized 3T3 mouse fibroblasts; in fact, promoter activity was increased up to 2.6-fold. However, in these differentiated cells, stimulation of the HSV tk promoter by either the F101 enhancer or ICP4 protein required the TATA box. HSV tk promoter activity also was dependent on its TATA box in the mouse oocyte, a terminally differentiated cell with an endogenous transactivating activity. These results reveal that the need for a TATA box is developmentally acquired and depends on at least two parameters: the differentiated state of the cell and stimulation of the promoter by either an enhancer or a transactivator.

Papillomavirus contains cis-acting sequences that can suppress but not regulate origins of DNA replication

Bovine papillomavirus (BPV) DNA has been reported to restrict its own replication and that of the lytic simian virus 40 (SV40) origin to one initiation event per molecule per S phase, which suggests BPV DNA replication as a model for cellular chromosome replication. Suppression of the SV40 origin required two cis-acting BPV sequences (NCOR-1 and -2) and one trans-acting BPV protein. The results presented in this paper confirm the presence of two NCOR sequences in the BPV genome that can suppress polyomavirus (PyV) as well as SV40 origin-dependent DNA replication as much as 40-fold. However, in contrast to results of previous studies on SV40, most of the suppression of the PyV origin was due to NCOR-1, a 512-bp sequence that functioned independently of distance or orientation with respect to the PyV origin and that was not required for BPV DNA replication. Moreover, NCOR-1 alone or together with NCOR-2 did not restrict the ability of the PyV ori to reinitiate replication within a single S phase and did not require any BPV protein to exert suppression. Furthermore, NCOR-1 did not suppress BPV origin-dependent DNA replication except in the presence of PyV large tumor antigen (T-ag). Since NCOR-1 suppression of PyV origin activity also varied with T-ag concentration, suppression of origins by NCOR sequences appeared to require papovavirus T-ag. Therefore, it is unlikely that NCOR sequences are involved in regulating BPV DNA replication. When these results are taken together with those from other laboratories, BPV appears to be a slowly replicating version of papovaviruses rather than a model for origins of DNA replication in eukaryotic cell chromosomes.

Requirements for DNA transcription and replication at the beginning of mouse development

In mice, the first round of DNA replication occurs in fertilized eggs (1-cell embryos), while the onset of zygotic gene transcription begins approximately 20 hours after fertilization, a time that normally coincides with formation of a 2-cell embryo. One approach to investigating the mechanisms that control these developmentally regulated events has been to microinject plasmid DNA into the nuclei of mouse oocytes and embryos in order to determine the requirements for unique DNA sequences that regulate transcription and replication. The results from these and other studies have revealed two important mechanisms that regulate the beginning of animal development. The first is a time dependent "zygotic clock" of unknown detail that delays the onset of transcription, regardless of whether or not a 2-cell embryo is formed. The second is a mechanism that represses the activity of promoters and origins of replication specifically in maternal pronuclei of oocytes and 1-cell embryos, and in all nuclei of 2-cell embryos, regardless of their parental origin or ploidy. This repression is linked to chromatin, but the striking ability to relieve this repression with specific embryo-responsive enhancers first appears with formation of a 2-cell embryo. The need for a TATA-box to mediate enhancer stimulation of promoter activity appears even later when cell differentiation becomes evident. Thus, a biological clock delays transcription until both paternal and maternal genomes are replicated and remodeled from a post-meiotic state to one in which transcription is repressed by chromatin structure in a manner that can be relieved by cell-specific enhancers at appropriate times during development.

Transcription enhancer factor-1 (TEF-1) DNA binding sites can specifically enhance gene expression at the beginning of mouse development

In an effort to identify transcriptional elements that are recognized at different stages of early mouse development, polyomavirus (PyV) enhancer mutations were selected for their ability to support PyV transcription and replication in various mouse undifferentiated embryonal carcinoma (EC) and embryonic stem (ES) cell lines. Several of these enhancer mutations were then isolated, sequenced and tested for their ability to stimulate the PyV early gene promoter in plasmid DNA that was either transfected into EC, ES and fibroblast cell lines, or injected into the nuclei of mouse 1-cell and 2-cell embryos. EC, ES and fibroblast cell lines showed clear preferences for different enhancer configurations, and cleavage-stage embryos (2- to 8-cell stage) strongly preferred the same enhancer configuration favored by ES cells. This 'embryo responsive' (ER) enhancer configuration was characterized by a tandem duplication of the region containing a single point mutation that created a DNA binding site for Transcription Enhancer Factor-1 (TEF-1). ER enhancers stimulated the PyV promoter up to 350-fold in embryos, and were up to 74-fold more active than the wild-type PyV enhancer. Most of the activity from PyER enhancers could be duplicated in 2-cell embryos by synthesizing only the tandemly repeated sequence. Comparison of these synthetic enhancers with ER enhancers confirmed that TEF-1 DNA binding sites were highly preferred in ES cells and cleavage-stage embryos, and suggested that ER enhancer activity resulted primarily from cooperative interaction between either two closely spaced TEF-1 DNA binding sites or two TEF-1 DNA binding sites separated by a third, as yet unidentified, transcription factor binding site. These results provide a prototype of a mammalian embryo responsive enhancer, and suggest that TEF-1 plays an important role in activation of gene expression at the beginning of mammalian development.

Requirements for promoter activity in mouse oocytes and embryos distinguish paternal pronuclei from maternal and zygotic nuclei

Fertilization of mouse eggs produces a 1-cell embryo containing both a paternal and maternal pronucleus. These two nuclei combine during the first mitosis to form the zygotic nuclei of 2-cell embryos. This transition is accompanied by the onset of transcription and the decline of maternal mRNA-dependent gene expression. To determine how changes in nuclear composition affect gene expression, plasmid DNA containing a promoter and an enhancer that function throughout a broad host range was injected into nuclei of oocytes and embryos. The requirements for promoter activity in paternal pronuclei of 1-cell embryos were distinct from those in maternal or zygotic nuclei: (1) Paternal pronuclei permitted high levels of promoter activity relative to maternal or zygotic nuclei. (2) Butyrate, an agent that alters chromatin structure, stimulated promoter activity in maternal or zygotic nuclei, but not in paternal pronuclei. (3) The embryo-responsive polyomavirus F101 enhancer also stimulated promoter activity, but only after formation of a 2-cell embryo. Either butyrate or the F101 enhancer stimulated promoter activity in zygotic nuclei to the level observed in paternal pronuclei. Stimulation also was observed with 2-cell embryos containing nuclei of only maternal or paternal origin, but their transcriptional capacity was more limited. These and other results support the hypothesis that the need for enhancers in 2-cell embryos results from repression by chromatin structure, and the role of enhancers is to relieve this repression.

Origins of DNA replication that function in eukaryotic cells

This past year has seen a significant increase in our understanding of eukaryotic origins of replication, of the proteins that identify these origins, of DNA sequences that promote their unwinding, and of transcription factors that stimulate origin activity. DNA replication begins at specific sites in both simple and complex genomes, but origins in complex genomes may include nuclear structure as well as DNA sequence.

Analysis of gene expression in mouse preimplantation embryos demonstrates that the primary role of enhancers is to relieve repression of promoters

Enhancers are generally viewed simply as extensions of promoters, lacking a function of their own. However, previous studies of mouse preimplantation embryos revealed that 1-cell embryos can utilize enhancer-responsive promoters efficiently without an enhancer, whereas 2-cell embryos require an enhancer to achieve the same levels of expression. This suggested that enhancers relieved a repression in 2-cell embryos that is absent in 1-cell embryos. Results presented here demonstrate first that the ability of 1-cell embryos to dispense with enhancers does not result from the absence of specific activation proteins. Under conditions where GAL4-VP16 activated a GAL4-dependent promoter in both embryos, GAL4-VP16 activated a GAL4-dependent enhancer only in 2-cell embryos. Moreover, the role of an enhancer is not to compensate for either changes in promoter requirements, or for reduced levels of promoter-specific transcription factors. Linker-scanning mutations in a natural promoter revealed that both embryos utilized the same promoter elements, and comparison of different promoters revealed that these embryos have equivalent transcriptional capacities. In addition, titration experiments revealed less Sp1 activity in 1-cell embryos where enhancers are dispensable than in 2-cell embryos where enhancers are required. Therefore, we propose that the primary function of enhancers, first evident with formation of a mouse 2-cell embryo, is to prevent repression of weak promoters, probably by altering chromatin structure. Consistent with this hypothesis is the fact that butyrate, an agent that alters chromatin structure, stimulated promoters in 2-cell embryos, but not in 1-cell embryos.

Striking homology between mouse and human transcription enhancer factor-1 (TEF-1)

Images

Origins of DNA replication in metazoan chromosomes

In metazoan chromosomes, initiation of DNA replication occurs primarily at specific sites (0.5-3 kb; the OBR) using the same replication fork mechanism favored by simple genomes. Nucleosome segregation is distributive. However, initiation events also occur at other sites randomly distributed throughout a larger initiation zone (8-55 kb). These nonspecific initiation events presumably occur at a lower frequency than those at the OBR. Although origins in native chromosomes appear to be genetically determined, ori sequences have been convincingly demonstrated only during programmed gene amplification. It seems likely that metazoan origins include both specific DNA sequences as well as chromatin structure and nuclear organization, something that may be difficult to reproduce with plasmid DNA. Definitive answers will come only with a functional assay for initiation that exhibits the sequence-specific characteristics observed in vivo.

Site-specific initiation of DNA replication in metazoan chromosomes and the role of nuclear organization

We have asked whether or not Xenopus eggs or egg extracts, which have previously been shown to replicate essentially any DNA molecule, will preferentially utilize a known mammalian OBR. Our results reveal that Xenopus egg extracts can preferentially initiate DNA replication at sites chosen in vivo by the hamster cell, provided that the DNA substrate is presented to the extract in the form of a nucleus rather than bare DNA. Thus, site-specific initiation of DNA replication in metazoan cell chromosomes appears to be determined by nuclear organization as well as DNA sequence. We have also considered whether or not BPV, which was previously reported to regulate its copy number through negative as well as positive cis-acting sequences, provides a suitable paradigm for cellular origins. The BPV genome was found to contain cis-acting sequences that can suppress DNA replication driven by a lytic virus such as PyV. However, this suppression did not require any BPV protein, did not limit PyV origin activity to one initiation event per S phase, and did not affect BPV origin activity. These results, together with data from other laboratories, strongly suggest that BPV is simply a slow-replicating version of SV40 and PyV and therefore is not an appropriate model to explain how initiation of cellular DNA replication is limited to once per cell cycle.

Specific transcription factors stimulate simian virus 40 and polyomavirus origins of DNA replication

The origins of DNA replication (ori) in simian virus 40 (SV40) and polyomavirus (Py) contain an auxiliary component (aux-2) composed of multiple transcription factor binding sites. To determine whether this component stimulated replication by binding specific transcription factors, aux-2 was replaced by synthetic oligonucleotides that bound a single transcription factor. Sp1 and T-antigen (T-ag) sites, which exist in the natural SV40 aux-2 sequence, provided approximately 75 and approximately 20%, respectively, of aux-2 activity when transfected into monkey cells. In cell extracts, only T-ag sites were active. AP1 binding sites could replace completely either SV40 or Py aux-2. Mutations that eliminated AP1 binding also eliminated AP1 stimulation of replication. Yeast GAL4 binding sites that strongly stimulated transcription in the presence of GAL4 proteins failed to stimulate SV40 DNA replication, although they did partially replace Py aux-2. Stimulation required the presence of proteins consisting of the GAL4 DNA binding domain fused to specific activation domains such as VP16 or c-Jun. These data demonstrate a clear role for transcription factors with specific activation domains in activating both SV40 and Py ori. However, no correlation was observed between the ability of specific proteins to stimulate promoter activity and their ability to stimulate origin activity. We propose that only transcription factors whose specific activation domains can interact with the T-ag initiation complex can stimulate SV40 and Py ori-core activity.

Guide to identification of origins of DNA replication in eukaryotic cell chromosomes

Several experimental approaches for identification of origins of DNA replication have been developed recently that allow, for the first time, unique initiation sites in mammalian chromosomes to be mapped at single-copy loci. A brief description of the rationale, advantages, and limitations has been provided for each approach, as well as information that can help the reader choose the method(s) most suitable for a particular system. The various methods are divided into three groups: (1) analysis of nascent DNA strands, (2) analysis of DNA structures, and (3) analysis of origin activity (i.e., ability to support autonomous replication). It is hoped that this information will serve as a practical guide for identifying new origins of replication.

T-antigen binding to site I facilitates initiation of SV40 DNA replication but does not affect bidirectionality

SV40 origin auxiliary sequence 1 (aux-1) encompasses T-antigen (T-ag) binding site I and facilitates origin core (ori-core) activity in whole cells or cell extracts. Aux-1 activity depended completely upon its sequence, orientation and spacing relative to ori-core. Aux-1 activity was lost either by inserting 10 base pairs between aux-1 and ori-core or by placing either orientation of aux-1 on the opposite side of ori-core. Reversing the orientation of aux-1 in its normal position actually inhibited replication. Easily unwound DNA sequences that stimulate yeast or E. coli origins of replication could not replace aux-1. Aux-1 did not affect bidirectional replication. Replication remained bidirectional even when aux-1 was inactivated, and deletion of aux-1 did not affect selection of RNA-primed DNA synthesis initiation sites in the origin region: the transition from discontinuous to continuous DNA synthesis that marks the origin of bidirectional replication occurred at the same nucleotide locations in both wild-type and aux-1 deleted origins. These results support a model for initiation of SV40 DNA replication in which T-ag binding to aux-1 (T-ag binding site I) facilitates the efficiency with which T-ag initiates replication at ori-core (T-ag binding site II) without affecting the mechanism by which initiation of DNA replication occurs.

Emetine allows identification of origins of mammalian DNA replication by imbalanced DNA synthesis, not through conservative nucleosome segregation

In the presence of emetine, an inhibitor of protein synthesis, nascent DNA on forward arms of replication forks in hamster cell lines containing either single or amplified copies of the DHFR gene region was enriched 5- to 7-fold over nascent DNA on retrograde arms. This forward arm bias was observed on both sides of the specific origin of bidirectional DNA replication located 17 kb downstream of the hamster DHFR gene (OBR-1), consistent with at least 85% of replication forks within this region emanating from OBR-1. However, the replication fork asymmetry induced by emetine does not result from conservative nucleosome segregation, as previously believed, but from preferentially inhibiting Okazaki fragment synthesis on retrograde arms of forks to produce 'imbalanced DNA synthesis'. Three lines of evidence support this conclusion. First, the bias existed in long nascent DNA strands prior to nuclease digestion of non-nucleosomal DNA. Second, the fraction of RNA-primed Okazaki fragments was rapidly diminished. Third, electron microscopic analysis of SV40 DNA replicating in the presence of emetine revealed forks with single-stranded DNA on one arm, and nucleosomes randomly distributed to both arms. Thus, as with cycloheximide, nucleosome segregation in the presence of emetine was distributive.

Regulation of gene expression in preimplantation mouse embryos: effects of the zygotic clock and the first mitosis on promoter and enhancer activities

Previous studies have reported that promoters requiring enhancers for full activity in mammalian somatic cells also require enhancers when injected into mouse two-cell embryos, whereas the same promoters can be expressed just as efficiently in the absence of an enhancer when injected into arrested one-cell embryos. Experiments were designed to determine whether this phenomenon reflected normal developmental changes at the beginning of mammalian development, or simply differences in the physiological states of these cells under the experimental conditions employed. The activity of three different promoters that function in a wide variety of mammalian cells was measured both in embryos whose morphological development was arrested and in embryos that continued development in vitro. Expression of the injected gene was related to the onset of zygotic gene expression ("zygotic clock"), the phase of the cell proliferation cycle, the use of aphidicolin to arrest cell proliferation, and formation of two-cell embryos in vitro and in vivo. The results demonstrated that promoter activity was tightly linked to zygotic gene expression, while the need for enhancers to stimulate promoter activity depended only on formation of a two-cell embryo. These results further support the hypothesis that the first mitosis induces a general repression of promoters prior to initiation of zygotic gene expression that is relieved specifically by enhancers.

Identification of an origin of bidirectional DNA replication in mammalian chromosomes

Mechanistically, an origin of bidirectional DNA replication (OBR) can be defined by the transition from discontinuous to continuous DNA synthesis that must occur on each template strand at the site where replication forks originate. This results from synthesis of Okazaki fragments predominantly on the retrograde arms of forks. We have identified these transitions at a specific site within a 0.45 kb sequence approximately 17 kb downstream from the 3' end of the dihydrofolate reductase gene in Chinese hamster ovary chromosomes. At least 80% of the replication forks in a 27 kb region emanated from this OBR. Thus, initiation of DNA replication in mammalian chromosomes uses the same replication fork mechanism previously described in a variety of prokaryotic and eukaryotic genomes, suggesting that mammalian chromosomes also utilize specific cis-acting sequences as origins of DNA replication.

Nucleosome assembly in mammalian cell extracts before and after DNA replication

Protein-free DNA in a cytosolic extract supplemented with SV40 large T-antigen (T-Ag), is assembled into chromatin structure when nuclear extract is added. This assembly was monitored by topoisomer formation, micrococcal nuclease digestion and psoralen crosslinking of the DNA. Plasmids containing SV40 sequences (ori- and ori+) were assembled into chromatin with similar efficiencies whether T-Ag was present or not. Approximately 50-80% of the number of nucleosomes in vivo could be assembled in vitro; however, the kinetics of assembly differed on replicated and unreplicated molecules. In replicative intermediates, nucleosomes were observed on both the pre-replicated and post-replicated portions. We conclude that the extent of nucleosome assembly in mammalian cell extracts is not dependent upon DNA replication, in contrast to previous suggestions. However, the highly sensitive psoralen assay revealed that DNA replication appears to facilitate precise folding of DNA in the nucleosome.

Mapping an origin of DNA replication at a single-copy locus in exponentially proliferating mammalian cells

A general method for determining the physical location of an origin of bidirectional DNA replication has been developed recently and shown to be capable of correctly identifying the simian virus 40 origin of replication (L. Vassilev and E. M. Johnson, Nucleic Acids Res. 17:7693-7705, 1989). The advantage of this method over others previously reported is that it avoids the use of metabolic inhibitors, the requirement for cell synchronization, and the need for multiple copies of the origin sequence. Application of this method to exponentially growing Chinese hamster ovary cells containing the nonamplified, single-copy dihydrofolate reductase gene locus revealed that DNA replication begins bidirectionally in an initiation zone approximately 2.5 kilobases long centered about 17 kilobases downstream of the DHFR gene, coinciding with previously described early replicating sequences. These results demonstrate the utility of this mapping protocol for identifying cellular origins of replication and suggest that the same cellular origin is used in both the normal and the amplified DHFR locus.

Simian virus 40 origin auxiliary sequences weakly facilitate T-antigen binding but strongly facilitate DNA unwinding

The complete simian virus 40 (SV40) origin of DNA replication (ori) consists of a required core sequence flanked by two auxiliary sequences that together increase the rate of DNA replication in monkey cells about 25-fold. Using an extract of SV40-infected monkey cells that reproduced the effects of ori-auxiliary sequences on DNA replication, we examined the ability of ori-auxiliary sequences to facilitate binding of replication factors and to promote DNA unwinding. Although the replicationally active form of T antigen in these extracts had a strong affinity for ori-core, it had only a weak but specific affinity for ori-auxiliary sequences. Deletion of ori-auxiliary sequences reduced the affinity of ori-core for active T antigen by only 1.6-fold, consistent with the fact that saturating concentrations of T antigen in the cell extract did not reduce the stimulatory role of ori-auxiliary sequences in replication. In contrast, deletion of ori-auxiliary sequences reduced the efficiency of ori-specific, T-antigen-dependent DNA unwinding in cell extracts at least 15-fold. With only purified T antigen in the presence of topoisomerase I to unwind purified DNA, ori-auxiliary sequences strongly facilitated T-antigen-dependent DNA conformational changes consistent with melting the first 50 base pairs. Under these conditions, ori-auxiliary sequences had little effect on the binding of T antigen to DNA. Therefore, a primary role of ori-auxiliary sequences in DNA replication is to facilitate T-antigen-dependent DNA unwinding after the T-antigen preinitiation complex is bound to ori-core.

Transactivation of the adenovirus EIIa promoter in the absence of adenovirus E1A protein is restricted to mouse oocytes and preimplantation embryos

Undifferentiated mouse embryonal carcinoma (EC) cells are capable of transactivating the adenovirus EIIa promoter in the absence of its normal transactivator, E1A protein, suggesting that EC cells contain an E1A-like activity. In an effort to identify where this activity appears during normal mouse development, mouse oocytes and preimplantation embryos were injected with plasmids containing the EIIa promoter coupled to various reporter genes. These expression vectors were fully active in human 293 cells where E1A is present, but were inactive in differentiated fibroblast cell lines unless cotransfected with the E1A gene. In mouse oocytes and preimplantation embryos, EIIa promoter activity in the absence of adenovirus E1A protein was equivalent to or greater than activity of the HSV thymidine kinase promoter coupled to a strong enhancer. Coinjection of the E1A gene failed to stimulate EIIa activity further, perhaps because c-myc protein, which has been reported to transactivate this promoter, was already present at high levels in mouse oocytes. Activation of the EIIa promoter in the absence of E1A was unique to mouse oocytes and preimplantation embryos because gene expression from an EIIa promoter introduced into transgenic mice was observed only in the adult ovary, and particularly in the oocytes. In addition, post-implantation transgenic embryos failed to express the E1A-activatable reporter gene, thereby indicating that expression from the EIIa promoter is restricted to the relatively undifferentiated stages of oogenesis and preimplantation development. These data suggest that cellular promoters of the class that can be transactivated by E1A may serve uniquely to initiate transcription of genes that are needed for preimplantation development.