In vitro transcription of a cloned mouse ribosomal RNA gene

TATA-Binding protein-interacting protein 120, TIP120, stimulates three classes of eukaryotic transcription via a unique mechanism

We previously identified a novel TATA-binding protein (TBP)-interacting protein (TIP120) from the rat liver. Here, in an RNA polymerase II (RNAP II)-reconstituted transcription system, we demonstrate that recombinant TIP120 activates the basal level of transcription from various kinds of promoters regardless of the template DNA topology and the presence of TFIIE/TFIIH and TBP-associated factors. Deletion analysis demonstrated that a 412-residue N-terminal domain, which includes an acidic region and the TBP-binding domain, is required for TIP120 function. Kinetic studies suggest that TIP120 functions during preinitiation complex (PIC) formation at the step of RNAP II/TFIIF recruitment to the promoter but not after the completion of PIC formation. Electrophoretic mobility shift assays showed that TIP120 enhanced PIC formation, and TIP120 also stimulated the nonspecific transcription and DNA-binding activity of RNAP II. These lines of evidence suggest that TIP120 is able to activate basal transcription by overcoming a kinetic impediment to RNAP II/TFIIF integration into the TBP (TFIID)-TFIIB-DNA-complex. Interestingly, TIP120 also stimulates RNAP I- and III-driven transcription and binds to RPB5, one of the common subunits of the eukaryotic RNA polymerases, in vitro. Furthermore, in mouse cells, ectopically expressed TIP120 enhances transcription from all three classes (I, II, and III) of promoters. We propose that TIP120 globally regulates transcription through interaction with basal transcription mechanisms common to all three transcription systems.

Regulation of mammalian ribosomal gene transcription by RNA polymerase I

All cells, from prokaryotes to vertebrates, synthesize vast amounts of ribosomal RNA to produce the several million new ribosomes per generation that are required to maintain the protein synthetic capacity of the daughter cells. Ribosomal gene (rDNA) transcription is governed by RNA polymerase I (Pol I) assisted by a dedicated set of transcription factors that mediate the specificity of transcription and are the targets of the pleiotrophic pathways the cell uses to adapt rRNA synthesis to cell growth. In the past few years we have begun to understand the specific functions of individual factors involved in rDNA transcription and to elucidate on a molecular level how transcriptional regulation is achieved. This article reviews our present knowledge of the molecular mechanism of rDNA transcriptional regulation.

Cloning and structural analysis of cDNA and the gene for mouse transcription factor UBF

The gene and protein structure of the mouse UBF (mUBF), a transcription factor for mouse ribosomal RNA gene, have been determined by cDNA and genomic clones. The unique mUBF gene consists of 21 exons spanning over 13 kb. Two mRNAs coding for mUBF1 and mUBF2 having 765 a.a. and 728 a.a., respectively, are produced by an alternative splicing of exon 8. It specifies 37 amino acids constituting a part of the regions homologous to high mobility group proteins (HMG box 2). A human UBF (hUBF) cDNA obtained by polymerase chain reaction also indicates the presence of two kinds of mRNAs, the shorter form lacking the same region as mUBF2. Comparison of the cDNAs from hUBF and mUBF revealed an unusual conservation of nucleotide sequence in the 3'-terminal non-coding region. We examined the relative amounts of expression of mUBF1 and mUBF2. The eight tissues studied contained both molecular species, although mUBF2 was the predominant form of UBF. The mRNA of mUBF1 was expressed one half of the mUBF2 in quiescent mouse fibroblasts but reached the same amount in growing state.

Structure of the core promoter of human and mouse ribosomal RNA gene. Asymmetry of species-specific transcription

In vitro transcription of the ribosomal RNA gene (rDNA) shows a remarkable species specificity such that human and mouse rDNA cannot use heterologous extracts of each other. The region that is responsible for this specificity has been studied using human-mouse chimeric genes and characteristic structures of both core promoters are presented. When the mouse sequence is substituted by the corresponding human sequence from upstream, the promoter activity in the mouse extract begins to decline at nucleotide -32 or -30, decreasing gradually and is lost completely at -19. A similar gradual decrease was noted for the 3' side substitution, which started at nucleotide -14 and was completed when up to the nucleotide -22 mouse position was replaced by the corresponding sequence from human. Thus, in the mouse rDNA core promoter, the sequence that is involved in species specificity resides only in a stretch encompassing the non-conserved region between the distal conserved sequence (DCS) and the proximal conserved sequence (PCS), plus two altered nucleotides in the PCS. When human rDNA is transcribed with human cell extract, the mouse sequence cannot substitute for the human sequence within the region from nucleotide -43 to +17 without affecting promoter activity significantly. This asymmetry of species specificity is due to the presence of nucleotides -43, +1 and +17, which are sensitive to change in only the human core promoter. The difference in the 5' border is ascribed to the species specificity of a transcription factor TFID, which recognizes this region. But the large difference of the 3' border is apparently due to another factor, possibly RNA polymerase I itself, because this region is not recognized by TFID in either human or mouse. Mammalian rDNA core promoter appears to consist of a tandem mosaic in which three evolutionarily conserved sequences alternate with non-conserved sequences having certain functionally important nucleotides. Not only non-conserved sequences and non-conserved nucleotides in conserved sequences, but also the spacings between the three conserved regions, play a crucial role in species specificity.

Structural determinant of the species-specific transcription of the mouse rRNA gene promoter

Mammalian ribosomal DNA (rDNA) transcription has a certain species specificity such that, both in vivo and in vitro, human rDNA cannot be transcribed by mouse machinery and vice versa. This is due to a species-dependent transcription factor, TFID (Y. Mishima, I. Financsek, R. Kominami, and M. Muramatsu, Nucleic Acids Res. 10:6659-6670, 1982). On the basis of the information obtained from 5' and 3' substitution mutants, we prepared a chimeric gene in which the mouse sequence from positions -32 to -14 was inserted into the corresponding location of the human rDNA promoter. The chimeric gene could be transcribed by mouse extracts nearly as efficiently as the wild-type mouse promoter. The chimeric gene could also sequester transcription factor TFID at an efficiency similar to that for the mouse promoter. Partially purified mouse TFID that could not protect the human rDNA promoter against DNase I produced a clear footprint on this chimeric gene that was similar to that on mouse rDNA promoter. The basic structure of the mouse rDNA core promoter is discussed in relation to the interaction with TFID.

Structural determinant of the species-specific transcription of the mouse rRNA gene promoter

Mammalian ribosomal DNA (rDNA) transcription has a certain species specificity such that, both in vivo and in vitro, human rDNA cannot be transcribed by mouse machinery and vice versa. This is due to a species-dependent transcription factor, TFID (Y. Mishima, I. Financsek, R. Kominami, and M. Muramatsu, Nucleic Acids Res. 10:6659-6670, 1982). On the basis of the information obtained from 5' and 3' substitution mutants, we prepared a chimeric gene in which the mouse sequence from positions -32 to -14 was inserted into the corresponding location of the human rDNA promoter. The chimeric gene could be transcribed by mouse extracts nearly as efficiently as the wild-type mouse promoter. The chimeric gene could also sequester transcription factor TFID at an efficiency similar to that for the mouse promoter. Partially purified mouse TFID that could not protect the human rDNA promoter against DNase I produced a clear footprint on this chimeric gene that was similar to that on mouse rDNA promoter. The basic structure of the mouse rDNA core promoter is discussed in relation to the interaction with TFID.

A model for regulation of mammalian ribosomal DNA transcription. Co-ordination of initiation and termination

Based on recent experimental data about transcription initiation and termination, a model for regulation of mammalian ribosomal DNA transcription is developed using a simple kinetic scheme. In this model, the existence of the transition pathway from the terminator to the promoter increases the rate of ribosomal RNA precursor synthesis. In addition to this 'non-transcribed spacer' traverse of RNA polymerase I, the co-ordination of initiation and termination allows a rapid on/off switch transition from the minimum to the maximum rate of ribosomal RNA precursor synthesis. Furthermore, taking account of the participation of two factors in the termination event, we propose a plausible molecular mechanism for the co-ordination of initiation and termination. This co-ordination is emphasized by repetition of the terminator unit.

Primary processing of mammalian rRNA involves two adjacent cleavages and is not species specific

The primary transcript of the mouse rRNA gene is rapidly processed at nucleotide approximately +650 both in vivo and in vitro. Using run-off transcription in a mouse cell extract as well as S1 nuclease and primer extension analysis of cellular RNA, we demonstrated that this primary processing actually results in the formation of two species of downstream RNA which differ in length by approximately 6 nucleotides, indicating the existence of two closely positioned alternative processing sites. The 200-base-pair region just 3' to the mouse processing site has a striking 80% sequence homology with a region of the human rRNA external transcribed spacer, and S1 nuclease analysis of human cellular RNA has demonstrated that an analogous rRNA processing occurs at the 5' border of the homologous human region. Unlike rDNA transcriptional initiation, however, the primary rRNA processing is not highly species specific, for the transcript of a chimeric gene containing the human processing region adjacent to a mouse rDNA promoter was synthesized and correctly processed in a mouse cell extract. This result confirms that mouse and human rRNA undergo a common primary processing event which is evidently directed by sequences within the 200-base-pair conserved sequence region.

Localization of the in vivo and in vitro transcription initiation site and comparative analysis of the flanking sequences in the two main size classes of Ascaris lumbricoides rDNA

An accurate in vitro transcription system which utilizes the cloned 8.8 and 8.4 kb size classes of Ascaris rRNA genes (pAlr8 and pAlr13) and two kinds of cellular extracts from Ascaris oogonia has been established. Both rDNA containing plasmids are efficiently transcribed in vitro by RNA polymerase I from a unique site of rDNA which corresponds to the in vivo initiation site. The in vitro transcription product has a triphosphorylated 5'-end and starts on a G localized 414 bp (pAlr8) upstream of the beginning of the mature 18S rRNA. The promoter region has been delimited by testing the in vitro template activity of a series of restriction fragments. The region essential for the accuracy of initiation is contained within nucleotides -72 to +65, but full efficiency of transcription requires the additional presence of the region from nucleotides +66 to +84. The sequences upstream from position -72 do not appear to modulate the efficiency of specific in vitro initiation. Furthermore, the sequences flanking the transcription initiation site from position -1500 to +570 have been determined in the two cloned representatives of the two rDNA main size classes.

Primary processing of mammalian rRNA involves two adjacent cleavages and is not species specific

The primary transcript of the mouse rRNA gene is rapidly processed at nucleotide approximately +650 both in vivo and in vitro. Using run-off transcription in a mouse cell extract as well as S1 nuclease and primer extension analysis of cellular RNA, we demonstrated that this primary processing actually results in the formation of two species of downstream RNA which differ in length by approximately 6 nucleotides, indicating the existence of two closely positioned alternative processing sites. The 200-base-pair region just 3' to the mouse processing site has a striking 80% sequence homology with a region of the human rRNA external transcribed spacer, and S1 nuclease analysis of human cellular RNA has demonstrated that an analogous rRNA processing occurs at the 5' border of the homologous human region. Unlike rDNA transcriptional initiation, however, the primary rRNA processing is not highly species specific, for the transcript of a chimeric gene containing the human processing region adjacent to a mouse rDNA promoter was synthesized and correctly processed in a mouse cell extract. This result confirms that mouse and human rRNA undergo a common primary processing event which is evidently directed by sequences within the 200-base-pair conserved sequence region.

Sequestration analysis for RNA polymerase I transcription factors with various deletion and point mutations reveals different functional regions of the mouse rRNA gene promoter

We compared the ability of various deletion and substitution mutants of the mouse rRNA gene promoter to bind essential factors required for accurate transcription initiation by RNA polymerase I. Different amounts of a competitor template were first incubated with a mouse cell extract containing the whole complement of factors and RNA polymerase I, and then a tester template was added for the second incubation. Transcription was started by adding nucleoside triphosphates (one labeled), and the accurate transcripts were determined on a gel. The results indicated that the ability of 5' deletion mutants to sequester essential factors decreased almost concurrently with the impairment of in vitro transcription activity, whereas when the promoter sequence was removed from the 3' side, the transcription activity decreased earlier and more drastically than the sequestration ability. Similar, though not identical, results were obtained by preincubation with fraction D separated on a phosphocellulose column, indicating that the major factor which was sequestered was TFID, the species-dependent transcription initiation factor that binds first to the promoter in the initiation reaction (H. Kato, M. Nagamine, R. Kominami, and M. Muramatsu, Mol. Cell. Biol. 6:3418-3427, 1986). Compilation of the data suggests that a region inside the 5' half of the core promoter (-40 to -1) is essential for the binding of TFID. The 3' half of the promoter (-1 to downstream) is not essential for the binding of TFID but is highly important for an efficient transcription initiation. A strong down-mutant with a one-base substitution at -16 (G to A) had a reduced ability to bind to TFID, whereas a null mutant with a single base substitution at -7 (G to A) showed a binding ability similar to that of the wild-type promoter when tested with whole-cell extract. This null mutant, however, could not sequester the TFID well when incubated with fraction D alone, suggesting that the binding of TFID with this mutant is unstable in the absence of another factor(s) present in cell extract. The factor is not TFIA, which binds after TFID, because the addition of fraction A containing TFIA did not cause TFID to bind to the mutant. The availability of different mutants having lesions at different steps of transcription initiation will provide a powerful tool for the dissection of the initiation reaction of the RNA gene.

Sequestration analysis for RNA polymerase I transcription factors with various deletion and point mutations reveals different functional regions of the mouse rRNA gene promoter

We compared the ability of various deletion and substitution mutants of the mouse rRNA gene promoter to bind essential factors required for accurate transcription initiation by RNA polymerase I. Different amounts of a competitor template were first incubated with a mouse cell extract containing the whole complement of factors and RNA polymerase I, and then a tester template was added for the second incubation. Transcription was started by adding nucleoside triphosphates (one labeled), and the accurate transcripts were determined on a gel. The results indicated that the ability of 5' deletion mutants to sequester essential factors decreased almost concurrently with the impairment of in vitro transcription activity, whereas when the promoter sequence was removed from the 3' side, the transcription activity decreased earlier and more drastically than the sequestration ability. Similar, though not identical, results were obtained by preincubation with fraction D separated on a phosphocellulose column, indicating that the major factor which was sequestered was TFID, the species-dependent transcription initiation factor that binds first to the promoter in the initiation reaction (H. Kato, M. Nagamine, R. Kominami, and M. Muramatsu, Mol. Cell. Biol. 6:3418-3427, 1986). Compilation of the data suggests that a region inside the 5' half of the core promoter (-40 to -1) is essential for the binding of TFID. The 3' half of the promoter (-1 to downstream) is not essential for the binding of TFID but is highly important for an efficient transcription initiation. A strong down-mutant with a one-base substitution at -16 (G to A) had a reduced ability to bind to TFID, whereas a null mutant with a single base substitution at -7 (G to A) showed a binding ability similar to that of the wild-type promoter when tested with whole-cell extract. This null mutant, however, could not sequester the TFID well when incubated with fraction D alone, suggesting that the binding of TFID with this mutant is unstable in the absence of another factor(s) present in cell extract. The factor is not TFIA, which binds after TFID, because the addition of fraction A containing TFIA did not cause TFID to bind to the mutant. The availability of different mutants having lesions at different steps of transcription initiation will provide a powerful tool for the dissection of the initiation reaction of the RNA gene.

Formation of the transcription initiation complex on mammalian rDNA

Steps for the formation of transcription initiation complex on the human rRNA gene (rDNA) in vitro were analyzed with partially purified transcription factors and RNA polymerase I. The reaction requires at least two factors besides RNA polymerase I for maximal efficiency. Preincubation and short-pulse analyses of the accurate transcripts revealed the following steps. First, the species-dependent factor, designated TFID, bound to the rDNA template, forming a preinitiation complex (PIC-1) which was resistant to a moderate concentration (0.015 to 0.02%) of Sarkosyl. Other factors, designated TFIA and RNA polymerase I, were then added to convert it to the final preinitiation complex PIC-3. This complex incorporated the first two nucleoside triphosphates of the starting site to complete the initiation complex (IC), which was resistant to a high concentration (0.2%) of Sarkosyl. Binding of TFID was rate limiting in the overall initiation reaction in vitro. Together with the kinetics of incorporation, the results are interpreted to mean that TFID, one bound, remains complexed with rDNA together with TFIA as the PIC-2 for many rounds of transcription by RNA polymerase I. Thus, the formation of PIC-2 may be a prerequisite for the stable opening of rDNA for transcription in vivo.

Factors and nucleotide sequences that direct ribosomal DNA transcription and their relationship to the stable transcription complex

We have studied the protein components and nucleic acid sequences involved in stably activating the ribosomal DNA (rDNA) template and in directing accurate transcription of mammalian rRNA genes. Two protein components are necessary to catalyze rDNA transcription, and these have been extensively purified. The first, factor D, can stably associate by itself with the rDNA promoter region and is responsible for template commitment. The second component, factor C, which appears to be an activated subset of polymerase I, can stably bind to the factor D-rDNA complex but not to the rDNA in the absence of factor D. A third component which had been previously identified as a rDNA transcription factor is shown to be a RNase inhibitor. Extending our earlier observation that the approximately 150-base-pair mouse rDNA promoter consists of a minimal essential region (residues approximately -35 to approximately +9) and additional upstream stimulatory domains, we now report that each of these promoter domains acts to augment the binding of the polymerase I transcription factors. A minimum core region (residues approximately -35 to approximately -15) is capable of stable complex formation and of binding transcription factor D. Factor C can also bind to this D-core region complex.

Factors and nucleotide sequences that direct ribosomal DNA transcription and their relationship to the stable transcription complex

We have studied the protein components and nucleic acid sequences involved in stably activating the ribosomal DNA (rDNA) template and in directing accurate transcription of mammalian rRNA genes. Two protein components are necessary to catalyze rDNA transcription, and these have been extensively purified. The first, factor D, can stably associate by itself with the rDNA promoter region and is responsible for template commitment. The second component, factor C, which appears to be an activated subset of polymerase I, can stably bind to the factor D-rDNA complex but not to the rDNA in the absence of factor D. A third component which had been previously identified as a rDNA transcription factor is shown to be a RNase inhibitor. Extending our earlier observation that the approximately 150-base-pair mouse rDNA promoter consists of a minimal essential region (residues approximately -35 to approximately +9) and additional upstream stimulatory domains, we now report that each of these promoter domains acts to augment the binding of the polymerase I transcription factors. A minimum core region (residues approximately -35 to approximately -15) is capable of stable complex formation and of binding transcription factor D. Factor C can also bind to this D-core region complex.

Formation of the transcription initiation complex on mammalian rDNA

Steps for the formation of transcription initiation complex on the human rRNA gene (rDNA) in vitro were analyzed with partially purified transcription factors and RNA polymerase I. The reaction requires at least two factors besides RNA polymerase I for maximal efficiency. Preincubation and short-pulse analyses of the accurate transcripts revealed the following steps. First, the species-dependent factor, designated TFID, bound to the rDNA template, forming a preinitiation complex (PIC-1) which was resistant to a moderate concentration (0.015 to 0.02%) of Sarkosyl. Other factors, designated TFIA and RNA polymerase I, were then added to convert it to the final preinitiation complex PIC-3. This complex incorporated the first two nucleoside triphosphates of the starting site to complete the initiation complex (IC), which was resistant to a high concentration (0.2%) of Sarkosyl. Binding of TFID was rate limiting in the overall initiation reaction in vitro. Together with the kinetics of incorporation, the results are interpreted to mean that TFID, one bound, remains complexed with rDNA together with TFIA as the PIC-2 for many rounds of transcription by RNA polymerase I. Thus, the formation of PIC-2 may be a prerequisite for the stable opening of rDNA for transcription in vivo.

Transcription of eukaryotic ribosomal RNA gene

Remarkable advances have been made in the identification of the promoter regions for the ribosomal RNA genes from lower and higher eukaryotes. There has been some progress in the elucidation of the factors that control transcription of the ribosomal RNA gene. The characterization of the transcription factors are crucial for the understanding of the molecular mechanisms of ribosomal gene expression.

Transcription of mouse rDNA terminates downstream of the 3' end of 28S RNA and involves interaction of factors with repeated sequences in the 3' spacer

RNA polymerase I terminates transcription of mouse rDNA 565 bp downstream of the 3' end of mature 28S rRNA. This specific termination event can be duplicated in a nuclear extract system. RNA molecules with authentic 3' ends are transcribed from ribosomal minigene constructs provided the templates retain a minimal length of downstream spacer sequences. The nucleotide sequence of the region of transcription termination contains a set of repetitive structural elements consisting of 18 bp conserved nucleotides surrounded by stretches of pyrimidines. Termination in vivo occurs within the first element. This site is preferentially used in vitro at low template concentrations. At increasing DNA concentrations a termination site within the second repetitive element is used. Competition experiments with defined 3'-terminal fragments suggest that transcription termination by RNA polymerase I requires interaction of some factor (or factors) with the repetitive structural elements in the 3' nontranscribed spacer.

Transfection of mouse ribosomal DNA into rat cells: faithful transcription and processing

Truncated mouse ribosomal DNA (rDNA) genes were stably incorporated into rat HTC-5 cells by DNA-mediated cell transfection techniques. The mouse rDNA genes were accurately transcribed in these rat cells indicating that there is no absolute species specificity of rDNA transcription between mouse and rat. No more than 170 nucleotides of the 5' nontranscribed spacer was required for the accurate initiation of mouse rDNA transcription in rat cells. Further, the mouse transcripts were accurately cleaved at the 5' end of the 18S rRNA sequence, even though these transcripts contained neither the 3' end of mouse 18S rRNA nor any other downstream mouse sequences. Thus, cleavage at the 5' end of 18S rRNA is not dependent on long range interactions involving these downstream sequences.

Characterization of mouse 45S ribosomal RNA subspecies suggests that the first processing cleavage occurs 600 +/- 100 nucleotides from the 5' end and the second 500 +/- 100 nucleotides from the 3' end of a 13.9 kb precursor

Mouse fibroblasts labeled 1-9 h with 3H-uridine contained radioactive 45S rRNA subspecies of 13.9, 13.3, and 12.8 kb, as determined by hybrid-selection with rDNA plasmids and by electrophoresis in agarose-formaldehyde. The 13.9 kb subspecies contained 5' and 3' terminal rDNA sequences known from the work of Grummt and colleagues to be at or near the ends of the primary transcript. The 13.3 kb subspecies contained the 3' terminal sequence but lacked the 5' terminal sequence. The 12.8 kb subspecies lacked both terminal sequences. Washed nuclei produced one discrete species of 13.9 kb. The results suggested that synthesis of the primary transcript terminated 500 +/- 100 nucleotides beyond the 3' end of 28S rRNA, that the first processing cleavage occurred 600 +/- 100 nucleotides from the origin of synthesis, and the second cleavage occurred near the 3' end of 28S rRNA. Changes in relative radioactivities among the subspecies after serum stimulation or after treatment with low concentrations of cycloheximide suggesting that processing was not perfectly coupled with synthesis and that cycloheximide inhibited one cleavage more than others.

Structure of a Neurospora RNA polymerase I promoter defined by transcription in vitro with homologous extracts

A Neurospora in vitro transcription system has been developed which specifically and efficiently initiates transcription of a cloned Neurospora crassa ribosomal RNA gene by RNA polymerase I. The initiation site of transcription (both in vitro and in vivo) appears to be located about 850 bp from the 5' end of mature 17S rRNA. However, the primary rRNA transcripts are normally cleaved very rapidly at a site 120-125 nt from the 5' end in vitro and in vivo. The nucleotide sequence surrounding the initiation site has been determined. The region from -16 to +9 exhibits partial homology to the corresponding sequences from a wide variety of organisms including yeast, but the most striking similarity is to the initiation region from Dictyostelium discoideum which displays 73% homology to the Neurospora sequence from -23 to +47. The Neurospora sequences from -96 to +97 have been shown to be sufficient for transcription. This region contains two sequences displaying 8/9 bp matches to elements of the 5S rDNA promoter.

Presence of a limited number of essential nucleotides in the promoter region of mouse ribosomal RNA gene

Point mutations are introduced into a mouse rDNA fragment containing the promoter region by a sodium bisulfite method and the mutants are tested for the ability of accurate transcription initiation in vitro. The results indicate that the change, G to A, at -7 completely eliminates the promoter activity, and those at -16 and at -25 decrease it to about 10% and 50%, respectively. On the other hand, the substitutions at +9, +4, -2, -9 and -39 do not alter the template activity significantly. It is concluded that there are limited but distinct nucleotides that are essential for the transcription initiation of this gene. This sort of absolute requirement for single specific bases is not reported in protein coding genes transcribed by RNA polymerase II. We propose that these rigid recognition signals which we have found are the molecular basis for the strong species-dependency of the transcription machinery of RNA polymerase I system. A model is presented in which a transcription factor interacts with the rDNA promoter from one side of the DNA double-helix with essential contacts at these bases.

Transcription of herpes simplex virus tk sequences under the control of wild-type and mutant human RNA polymerase I promoters

We studied RNA polymerase I transcription in cells transfected with a plasmid, prHuTK, containing the herpes simplex virus tk gene fused to a human rRNA promoter. Primer extension analysis of tk RNA isolated from COS cells transfected with prHuTK reveals that transcription from the RNA polymerase I promoter is highly efficient and initiates at the same position used for the synthesis of endogenous rRNA in HeLa cells. The RNA products derived from prHuTK are distinguishable from normal RNA polymerase II transcripts of tk in that they are not polyadenylated, are extremely unstable, and are found predominantly in the nucleus. Moreover, the transcription observed is resistant to 300 micrograms of alpha-amanitin per ml. These results strongly suggest that prHuTK transcription is under the control of the human rRNA promoter and RNA polymerase I. To further characterize the activity of the human rDNA promoter in vivo, a series of 5' and 3' deletion mutants was tested in this transfection assay. The deletion analysis indicates that a core region of ca. 40 base pairs overlapping the initiation site is critical for transcription. In addition, a region between nucleotides -234 and -131 upstream from the core sequence serves to modulate the efficiency of transcription. Insertion into prHuTK of additional ribosomal nontranscribed spacer DNA or the simian virus 40 enhancer element has no apparent effect on the promoter activity. Surprisingly, RNA polymerase II transcripts synthesized at low levels from two start sites within the core control element of the wild-type RNA polymerase I promoter are activated upon deletion of upstream RNA polymerase I promoter sequences. However, these RNA polymerase II transcripts are not expressed from the endogenous rRNA promoter.

Transcription of herpes simplex virus tk sequences under the control of wild-type and mutant human RNA polymerase I promoters

We studied RNA polymerase I transcription in cells transfected with a plasmid, prHuTK, containing the herpes simplex virus tk gene fused to a human rRNA promoter. Primer extension analysis of tk RNA isolated from COS cells transfected with prHuTK reveals that transcription from the RNA polymerase I promoter is highly efficient and initiates at the same position used for the synthesis of endogenous rRNA in HeLa cells. The RNA products derived from prHuTK are distinguishable from normal RNA polymerase II transcripts of tk in that they are not polyadenylated, are extremely unstable, and are found predominantly in the nucleus. Moreover, the transcription observed is resistant to 300 micrograms of alpha-amanitin per ml. These results strongly suggest that prHuTK transcription is under the control of the human rRNA promoter and RNA polymerase I. To further characterize the activity of the human rDNA promoter in vivo, a series of 5' and 3' deletion mutants was tested in this transfection assay. The deletion analysis indicates that a core region of ca. 40 base pairs overlapping the initiation site is critical for transcription. In addition, a region between nucleotides -234 and -131 upstream from the core sequence serves to modulate the efficiency of transcription. Insertion into prHuTK of additional ribosomal nontranscribed spacer DNA or the simian virus 40 enhancer element has no apparent effect on the promoter activity. Surprisingly, RNA polymerase II transcripts synthesized at low levels from two start sites within the core control element of the wild-type RNA polymerase I promoter are activated upon deletion of upstream RNA polymerase I promoter sequences. However, these RNA polymerase II transcripts are not expressed from the endogenous rRNA promoter.

In vitro evidence that eukaryotic ribosomal RNA transcription is regulated by modification of RNA polymerase I

We have utilized a cell-free transcription system from Acanthamoeba castellanii to test the functional activity of RNA polymerase I and transcription initiation factor I (TIF-I) during developmental down regulation of rRNA transcription. The results strongly suggest that rRNA transcription is regulated by modification, probably covalent, of RNA polymerase I: (1) The level of activity of TIF-I in extracts from transcriptionally active and inactive cells is constant. (2) The number of RNA polymerase I molecules in transcriptionally active and inactive cells is also constant. (3) In contrast, though the specific activity of polymerase I on damaged templates remains constant, both crude and purified polymerase I from inactive cells have lost the ability to participate in faithful initiation of rRNA transcription. (4) Polymerase I purified from transcriptionally active cells has the same subunit architecture as enzyme from inactive cells. However, the latter is heat denatured 5 times faster than the active polymerase.

Processing of the external transcribed spacer of murine rRNA and site of action of actinomycin D

The primary rRNA transcript contains a large external transcribed spacer (ETS) approximately 4,000 nucleotides in length. We have used subcloned DNA probes derived from the 5' end of the ETS in conjunction with Northern blot analysis of murine nuclear RNA to examine processing of this region. In agreement with the results of previous investigators, we find that the large rRNA precursor lacks part of the ETS region. These ETS sequences are also missing from subsequent rRNA processing intermediates. Experiments using actinomycin D confirm that the excision of portion of the ETS is an early event in rRNA processing. In addition, in the presence of actinomycin D small RNA species accumulate which hybridize to a probe specific for the 5' end of the ETS. The length of these abbreviated transcripts defines a region of rDNA which is probably a target for this drug.

Localization and structure of endonuclease cleavage sites involved in the processing of the rat 32S precursor to ribosomal RNA

The initial endonuclease cleavage site in 32 S pre-rRNA (precursor to rRNA) is located within the rate rDNA sequence by S1-nuclease protection mapping of purified nucleolar 28 S rRNA and 12 S pre-rRNA. The heterogeneous 5'- and 3'-termini of these rRNA abut and map within two CTC motifs in tSi2 (internal transcribed spacer 2) located at 50-65 and 4-20 base-pairs upstream from the homogeneous 5'-end of the 28 S rRNA gene. These results show that multiple endonuclease cleavages occur at CUC sites in tSi2 to generate 28 S rRNA and 12 S pre-rRNA with heterogeneous 5'- and 3'-termini, respectively. These molecules have to be processed further to yield mature 28 S and 5.8 S rRNA. Thermal-denaturation studies revealed that the base-pairing association in the 12 S pre-rRNA:28 S rRNA complex is markedly stronger than that in the 5.8 S:28 S rRNA complex. The sequence of about one-quarter (1322 base-pairs) of the 5'-part of the rat 28 S rDNA was determined. A computer search reveals the possibility that the cleavage sites in the CUC motifs are single-stranded, flanked by strongly base-paired GC tracts, involving tSi2 and 28 S rRNA sequences. The subsequent nuclease cleavages, generating the termini of mature rRNA, seem to be directed by secondary-structure interactions between 5.8 S and 28 S rRNA segments in pre-rRNA. An analysis for base-pairing among evolutionarily conserved sequences in 32 S pre-rRNA suggests that the cleavages yielding mature 5.8 S and 28 S rRNA are directed by base-pairing between (i) the 3'-terminus of 5.8 S rRNA and the 5'-terminus of 28 S rRNA and (ii) the 5'-terminus of 5.8 S rRNA and internal sequences in domain I of 28 S rRNA. A general model for primary- and secondary-structure interactions in pre-rRNA processing is proposed, and its implications for ribosome biogenesis in eukaryotes are briefly discussed.

A conserved sequence element is present around the transcription initiation site for RNA polymerase A in Saccharomycetoideae

To identify DNA elements involved in the initiation of rRNA transcription in yeast we located the start site of the rRNA operon of Kluyveromyces lactis and Hansenula wingei, both members of the Saccharomycetoideae, by S1 nuclease analysis and determined the surrounding nucleotide sequences. Comparison of these sequences with those of Saccharomyces carlsbergensis, S. cerevisiae and S. rosei (all belonging to the same yeast subfamily) reveals an identical sequence at the site of transcription initiation from position +1 to +7 which is part of a larger conserved region extending from position -9 to +23; the conserved heptanucleotide sequence is supposed to constitute an important part of the promoter for yeast RNA polymerase A. The non-transcribed spacers (NTS) upstream of position -9 have diverged strongly with the exception of two short elements around positions -75 and -135. The external transcribed spacer (ETS) downstream of position +23 is largely conserved between K. lactis, S. rosei and S. carlsbergensis except for a divergent region around position +75. On the other hand, the ETS of H. wingei has diverged significantly.

Determination of the promoter region of mouse ribosomal RNA gene by an in vitro transcription system

Sequences required for a faithful and efficient transcription of a cloned mouse ribosomal RNA gene (rDNA) are determined by testing a series of deletion mutants in an in vitro transcription system utilizing two kinds of mouse cellular extract. Deletion of sequences upstream of -40 or downstream of +52 causes only slight reduction in promoter activity as compared with the "wild-type" template. For upstream deletion mutants, the removal of a sequence between -40 and -35 causes a significant decrease in the capacity to direct efficient initiation. This decrease becomes more pronounced when the deletion reaches -32 and the sequence A-T-C-T-T-T, conserved among mouse, rat, and human rDNAs, is lost. Residual template activity is further reduced as more upstream sequence is deleted and finally becomes undetectable when the deletion is extended from -22 down to -17, corresponding to the loss of the conserved sequence T-A-T-T-G. As for downstream deletion mutants, the removal of the sequence downstream of +23 causes some (and further deletions up to +11 cause a more) serious decrease in template activity in vitro. These deletions involve other conserved sequences downstream of the transcription start site. However, the removal of the original transcription start site does not abolish the transcription initiation completely, provided that the whole upstream sequence is intact.

Molecular cloning of the rDNA of Saccharomyces rosei and comparison of its transcription initiation region with that of Saccharomyces carlsbergensis

We have cloned one complete repeating unit of rDNA from Saccharomyces rosei and determined its physical and genetic organization. Heteroduplex analysis of the rDNA units from S. rosei and S. carlsbergensis shows that the nontranscribed spacers are largely nonhomologous in sequence, whereas the transcribed regions are essentially homologous. We also determined the transcription initiation site for the 37S precursor RNA on S. rosei rDNA. Sequence comparison of the region surrounding the site of transcription initiation for the 37S RNA with the corresponding region of S. carlsbergensis revealed extensive homology from position -9 downstream into the external transcribed spacer. Very little homology was observed between position -9 and -55, but some homologous tracts are present upstream from position -55.

Varigated chromatin structures of mouse ribosomal RNA genes

We have employed a chromatin fractionation procedure on micrococcal nuclease-digested nuclei to examine the chromatin structure of mouse ribosomal RNA genes in two systems that differ by at least 14-fold in the level of ribosomal RNA transcription. In a cultured cell line enriched in transcriptionally active ribosomal chromatin, most ribosomal sequences are preferentially sensitive to digestion by micrococcal nuclease, reside in an insoluble chromatin fraction, and lack typical nucleosomal packaging; only minor amounts of ribosomal sequences are packaged into soluble, nucleosomal chromatin. By contrast, in adult liver, which is enriched in transcriptionally inactive ribosomal chromatin, the majority of ribosomal genes are packaged into soluble, nucleosomal chromatin. However, a significant fraction of liver ribosomal chromatin is insoluble and possesses a non-nucleosomal structure. Therefore, within a single cell population or tissue, mouse ribosomal RNA genes are organized into both nucleosomal and non-nucleosomal chromatin structures. We suggest that these structures have functional significance.

Regulation of human ribosomal RNA transcription

We have used a cell-free polymerase I transcription system derived from HeLa cells to study the regulation of human rRNA synthesis. Analysis of deletion mutants spanning the start site of transcription at nucleotide +1 indicates that the control region affecting initiation of human rRNA synthesis is contained within sequences from nucleotides -158 to +18. This promoter region can be subdivided into (i) a central segment of approximately 40 base pair that is required for transcription and (ii) flanking sequences that influence the efficiency of transcription in vitro. We have examined the in vitro transcriptional activity of the human extract under various conditions that are thought to modulate rRNA synthesis in vivo. Cell-free extracts prepared from HeLa cells infected with adenovirus 2 synthesize human rRNA at levels greatly decreased relative to uninfected cell extracts. By contrast, in vitro transcription of human rRNA is stimulated 2- to 3-fold by the addition of purified simian virus 40 large tumor antigen to the transcription reaction. Moreover, a mutant tumor antigen known to be defective for rRNA activation in vivo is incapable of stimulating rRNA synthesis in vitro. The ability to detect these different regulatory phenomena in vitro provides us with an experimental basis for investigating the molecular mechanisms that control rRNA synthesis.

Identification and sequence of the initiation site for rat 45S ribosomal RNA synthesis

The transcription initiation site for rat 45S precursor ribosomal RNA synthesis was determined by nuclease protection mapping with two single-strand endonucleases. S1 and mung bean, and one single-strand exonuclease, ExoVII. These experiments were performed with end-labeled ribosomal DNA from double-stranded pBR322 recombinants and from single-stranded M13 recombinants. Results from experiments using both kinds of DNA and all three enzymes showed that the 5' end of 45S RNA mapped to a unique site 125 bases upstream from the Hind III site in the ribosomal DNA gene. The DNA surrounding this site (designated +1) was sequenced from -281 to +641. The entire sequence of this region shows extensive homology to the comparable region of mouse. This includes three stretches of T residues in the non-coding strand between +300 and +630. Two sets of direct repeats adjacent to these T-rich regions are observed. Comparison of the mouse and human ribosomal DNA transcription initiation sites with the rat sequence reported in this paper demonstrates a conserved sequence at +2 to +16, CTGACACGCTGTCCT. This suggests that this region may be important for the initiation of transcription on mammalian ribosomal DNAs.

Nucleotide sequence of the transcriptional initiation region of Dictyostelium discoideum rRNA gene and comparison of the initiation regions of three lower eukaryotes' genes

The 5' end of the rRNA precursor of D. discoideum was mapped on a cloned rDNA by S1 nuclease protection mapping, and the sequence of about 1240 nucleotides surrounding the transcriptional initiation site of the rRNA gene has been determined. Repeated sequences consisting of 16 nucleotides appeared in the region upstream from the initiation point. Comparison of the nucleotide sequences around the initiation site of rRNA genes in three lower eukaryotes, D. discoideum, Saccharomyces cerevisiae and Tetrahymena pyriformis, indicated that there was little similarity in the nontranscribed spacer regions, but in the transcribed spacer regions near the initiation point, very similar sequences consisting of 9 nucleotides were found.

Nucleotide sequence of the genetically labile repeated elements 5' to the origin of mouse rRNA transcription

We have determined the complete nucleotide sequence of a cloned Balb/c mouse rDNA NTS fragment containing 13 tandem copies of a 135 bp subrepeating segment. This repetitious region (VrDNA) lies close to the origin of ribosomal RNA transcription. Analyses of these VrDNA subrepeats from Balb/c and a related species, Mus pahari, reveal regions of inverted repeat DNA as well as large poly T tracts, either of which may be significant to the generation of the high levels of VrDNA copy number variation found in wild and inbred mice and/or the modulation of rRNA synthesis. Unlike the highly homogeneous subrepeats in the Xenopus laevis NTS repetitious regions, the VrDNA subrepeats differ from one another on the average by about 13%. Sequence analysis and Southern hybridization studies have also shown that, unlike the Xenopus and Drosophila NTS, extensive duplications of sequences found surrounding the mouse rRNA initiation site are found neither in the VrDNA region nor 6 kb further upstream in the NTS.

Structural organization of rat ribosomal genes restriction endonuclease analysis of genomic and cloned ribosomal DNAs

Structural organization of the rat ribosomal repeating unit was studied using hybridization of blotted restriction fragments of total rat DNA with alpha-32P-labeled cDNA probe synthesized on the 18S and 28S rRNAs. A detailed restriction endonuclease map was constructed, the 18S and 28S rRNA genes mapped and the sizes of the rat ribosomal repeating units determined. Considerable site heterogeneity of rat rDNA was revealed in both nontranscribed and external transcribed spacers. Recombinant phages containing the whole set of transcribed regions and a considerable part of a nontranscribed spacer of the rDNA were selected from the rat gene library. The restriction maps of the cloned rDNA fragments are in good agreement with the map constructed by Southern's technique, add to this map and support the existence of site heterogeneity in ribosomal repeating units. Proximal to the 3'-end of the 28S rRNA gene an internally repetitive region was found, each repeating unit being equal to approx. 150 bp. The site for transcription initiation was mapped 4.0-4.5 kb upstream from the 5'-end of the 18S rRNA gene. Frequently reiterated interspersed sequences were found in the nontranscribed spacer at approx. 2-3 kb distance from both ends of the transcribed region.

In vitro accurate initiation of transcription on the adenovirus type 2 IVa2 gene which does not contain a TATA box

In vitro initiation of transcription has been studied on the IVa2 gene of the adenovirus type 2 ( Ad2 ) which is expressed at intermediate time after infection. This gene does not contain a TATA box around 30 bp upstream from the cap site. By analyses of the in vitro runoff products by size and by single strand nuclease protection mapping, it was concluded that the in vitro initiation occurred accurately at two in vivo cap sites with low but detectable efficiency. By using cloned IVa2 DNA, which is free from the major late promoter, it was suggested that the low efficiency of the in vitro initiation on the IVa2 gene may not be due to close location of the two genes but may rather reflect the efficiency of the promoter itself. The cell-free extracts from both uninfected and infected HeLa cells supported the accurate initiation of transcription in vitro with almost the same efficiency. In addition, the cell free extract directed an artificial initiation of transcription in vitro by recognizing a TATA box located 140 bp upstream from the cap sites.

Fractionation and reconstitution of factors required for accurate transcription of mammalian ribosomal RNA genes: identification of a species-dependent initiation factor

Mouse and human cell extracts (S100) can support an accurate and efficient transcription initiation on homologous ribosomal RNA gene (rDNA) templates. The cell extracts were fractionated with the aid of a phosphocellulose column into four fractions (termed A, B, C and D), including one containing a major part of the RNA polymerase I activity. Various reconstitution experiments indicate that fraction D is an absolute requirement for the correct and efficient transcription initiation by RNA polymerase I on both mouse and human genes. Fraction B effectively suppresses random initiation on these templates. Fraction A appears to further enhance the transcription which takes place with fractions C and D. Although fractions A, B and C are interchangeable between mouse and human extracts, fraction D is not; i.e. initiation of transcription required the presence of a homologous fraction D for both templates. The factor(s) in fraction D, however, is not literally species-specific, since mouse D fraction is capable of supporting accurate transcription initiation on a rat rDNA template in the presence of all the other fractions from human cell extract under the conditions where human D fraction is unable to support it. We conclude from these experiments that a species-dependent factor in fraction D plays an important role in the initiation of rDNA transcription in each animal species.

Nucleotide sequence requirements for specific initiation of transcription by RNA polymerase I

The nucleotide sequence(s) specifying RNA polymerase I initiation has been investigated by studying the transcription of deleted and nondeleted mouse ribosomal RNA gene (rDNA) templates in vitro. The deletion of 5'-flanking sequences upstream from position -- 39 did not affect transcriptional activity, but removal of sequences between positions -- 39 and -- 34 resulted in a 90% decrease of rDNA transcription. The template activity was completely eliminated by the further deletion of nucleotides -- 33 to -- 13. It is concluded that sequences between -- 34 and -- 12, upstream from the transcribed region, represent an essential control region for the initiation of transcription in vitro. Therefore, this region may be functionally analogous to the T-A-T-A box of RNA polymerase II promoters. In addition to this control region, sequences located further upstream (between positions -- 45 and -- 169) may also exert some function in efficient transcription initiation as revealed by competition experiments between wildtype and mutant rDNA templates.

Determination of the transcription initiation site of Tetrahymena pyriformis rDNA using in vitro capping of 35S pre-rRNA

Approximately 700 nucleotide sequences surrounding the transcription initiation site were determined with a cloned rDNA fragment of Tetrahymena pyriformis and the transcription initiation site was localized on these sequences using purified 35S pre-rRNA. A considerable portion of the 35S pre-rRNA was found to be capped in vitro. The 32P-labeled, capped 35S pre-rRNA, on nucleus P1 protection mapping, gave the protection band which is identical in size with that obtained with bulk 35S pre-rRNA. Both reverse transcription extension and nuclease P1 mapping localized the 5'-end of the 35S pre-rRNA at the same adenine nucleotide, 496 base pairs upstream from the HindIII site of the cloned rDNA fragment. Furthermore, sequencing of the 5'-terminal region of the in vitro capped 35S pre-rRNA unambiguously confirmed the above result. The strategy adopted in the present experiment could serve as a general procedure for determining the transcription initiation point even in cases where the concentration of the primary transcript is low.