Enhancer of RNA polymerase III gene transcription.

A protein responsible for enhanced transcription by RNA polymerase III was identified in extracts from Xenopus oocytes. This protein, called EP3, interacts with a specific DNA sequence adjacent to the 3'-end of a Xenopus somatic 5S RNA gene and forms a distinct band shift complex with a unique DNase I footprint. Enhanced transcription was observed from both 5S RNA and tRNA reporter genes when EP3 binding sites were inserted at different locations and orientations. Removal of the EP3 protein from an oocyte extract abolished this enhanced transcription. In addition, EP3 was shown to stimulate transcription by increasing the rate of transcription complex assembly. EP3 directly discriminates between the somatic and oocyte 5S RNA gene families and may play a significant role in their differential expression during early Xenopus development.

Interaction of Xenopus TFIIIC with a 5S RNA gene.

Using fractionated Xenopus transcription factors we have identified and characterized a unique protein-DNA complex formed between TFIIIA, TFIIIC and a 5S RNA gene. The formation of this complex was blocked by specific competitor DNAs and by the inactivation of TFIIIC using two different methods. In addition, TFIIIC activity was retained when the complexes were affinity purified using a reversibly immobilized DNA template. The TFIII(A+C)-5S RNA gene complex has a distinct electrophoretic mobility on band-shift gels and a unique DNase I footprint. The characteristic feature of the DNase I footprint is a TFIIIC-dependent extension of the TFIIIA footprint an additional 25 bp toward the 5' end of the gene. This indicates a direct interaction between Xenopus TFIIIC and the template DNA.

Purification of a calcium dependent ribonuclease from Xenopus laevis.

We have purified a Ca2+ dependent ribonuclease from the oocytes of Xenopus leavis. Two properties of this ribonuclease set it apart from other known nucleases. First, Ca2+ was required for ribonuclease activity, and Mg2+ would not substitute. Second, the enzyme specifically degraded RNA and digestion of double or single stranded DNA was not observed. Ca2+ dependent ribonuclease activity of the purified 36-kDa protein was directly observed after renaturation of the protein following electrophoresis in an SDS-Laemmli gel. In addition, the enzyme was shown to have endoribonuclease activity at numerous sites. The Ca2+ dependence suggests that the ribonuclease activity may be modulated by changes in the level of intracellular Ca2+ and thereby provide a direct link to signal transduction systems.

Calcium-dependent inactivation of RNA polymerase III transcription.

Extracts from whole oocytes of Xenopus laevis are widely used as an efficient in vitro system for the transcription of cloned genes by RNA polymerase III. We have found that these extracts no longer support RNA polymerase III transcription in response to a brief incubation in the presence of Ca2+. However, when transcription complexes were first formed on the genes, a subsequent incubation in the presence of Ca2+ had little effect. Fractionation of extracts was used to show that transcription factors (TF) IIIC and, to a lesser extent, TFIIIB, but not RNA polymerase III, were targets of the Ca(2+)-dependent inactivation process. An additional component (not present in fractionated TFIIIC or TFIIIB) was required for the Ca(2+)-dependent destruction of transcription factor activity. The Ca(2+)-dependent inactivation process was blocked by protease inhibitors that inhibit known Ca(2+)-dependent proteases called calpains. These results suggest that TFIIIC and TFIIIB are inactivated by an endogenous calpain. The common use of Ca2+ as a second messenger and the widespread distribution of calpains suggest that the proteolytic degradation of transcription factors may be a general mechanism for the regulation of gene expression.

Bead-shift isolation of protein--DNA complexes on a 5S RNA gene.

Specific protein-DNA complexes formed on a Xenopus 5S RNA gene were isolated and characterized using a novel technique. A DNA template reversibly immobilized on paramagnetic beads was used to capture, affinity purify, and concentrate protein--DNA complexes formed in a whole cell extract. The complexes were then released from the beads in a soluble and transcriptionally active form via restriction enzyme digestion of the DNA. A band-shift gel was used to separate and obtain the DNase I footprints of five individual complexes. Three of the complexes resulted from the independent binding of two proteins, TFIIIA and an unidentified protein binding to a large region just downstream of the 3' end of the gene. Two more slowly migrating complexes contained an additional large central protected region covering most of the gene. The most slowly migrating complex displayed protein interactions over the 5' flanking sequences. The formation of two of these complexes was shown to be dependent on TFIIIC activity. The correlation between transcriptional activity and the formation of these complexes suggests that the observed protein--DNA interactions are important for transcription of 5S RNA genes.

Kinetic control of 5 S RNA gene transcription.

We have determined that the differential transcription of somatic and oocyte-type 5 S RNA genes in a Xenopus laevis oocyte extract is a consequence of vastly different rates of stable complex assembly. Somatic-type 5 S RNA genes sequester a limiting transcription factor much more rapidly than oocyte-type 5 S RNA genes. Once formed, however, transcription complexes on both types of genes are stable, and are transcribed at nearly equivalent rates. The relative rates of stable transcription complex assembly are strongly dependent on the concentration of Mg2+. Kinetic differences in transcription complex assembly provides a key distinguishing feature between these two genes which may be used in the selective repression of oocyte-type 5 S RNA genes during the early development of Xenopus, and may also be utilized in other systems of regulated gene expression.

Transcriptionally inactive oocyte-type 5S RNA genes of Xenopus laevis are complexed with TFIIIA in vitro.

An extract from whole oocytes of Xenopus laevis was shown to transcribe somatic-type 5S RNA genes approximately 100-fold more efficiently than oocyte-type 5S RNA genes. This preference was at least 10-fold greater than the preference seen upon microinjection of 5S RNA genes into oocyte nuclei or upon in vitro transcription in an oocyte nuclear extract. The approximately 100-fold transcriptional bias in favor of the somatic-type 5S RNA genes observed in vitro in the whole oocyte extract was similar to the transcriptional bias observed in developing Xenopus embryos. We also showed that in the whole oocyte extract, a promoter-binding protein required for 5S RNA gene transcription, TFIIIA, was bound both to the actively transcribed somatic-type 5S RNA gene and to the largely inactive oocyte-type 5S RNA genes. These findings suggest that the mechanism for the differential expression of 5S RNA genes during Xenopus development does not involve differential binding of TFIIIA to 5S RNA genes.

Rate of B to Z structural transition of supercoiled DNA.

The forward rate of the B to Z transition induced by negative supercoiling of plasmid DNA containing an alternating C-G sequence has been examined using the binding of Z-DNA-specific antibodies to follow the transition. DNA samples of a plasmid containing a d(pCpG)16 X d(pCpG)16 insert were supercoiled to different extents and appropriate amounts of ethidium were bound to the DNAs to relax them and to keep the alternating C-G sequence in the right-hand helical form. Following the rapid removal of ethidium by passage through a column of cation exchange resin, the DNA becomes negatively supercoiled, which induces the flipping of the helical hand of the C-G insert. The rate of the transition is strongly dependent on the degree of supercoiling. The transition is complete in less than 50 seconds for a DNA with a specific linking difference (superhelical density) sigma of -0.09. For the same DNA, the half-time of the transition is about two minutes at sigma = -0.07 and about a factor of 10 slower at sigma = -0.05.

Transcriptional block caused by a negative supercoiling induced structural change in an alternating CG sequence.

Using supercoiled plasmids containing a (CG)16 sequence downstream of a promoter, it is shown that purified E. coli RNA polymerase can transcribe through the sequence when it is in the B helical form. However, the polymerase together with its nascent transcript is blocked at the boundary of the CG sequence proximal to the promoter when the template is negatively supercoiled to flip the CG sequence to the left-handed Z-form. S1 nuclease mapping of in vivo transcripts from an E. coli gyrase temperature-sensitive mutant harboring the plasmids indicates that the bulk of the transcripts at either permissive or nonpermissive temperatures can proceed through the CG sequence, suggesting that the sequence is normally in the B helical form in vivo. The almost total blockage of transcription in vitro by the (CG)16 sequence in a highly negatively supercoiled DNA is not observed for a d(CA)21 X d(TG)21 insert.

Energetics of B-to-Z transition in DNA.

Analysis by two-dimensional gel electrophoresis of topoisomers of plasmids containing d(pCpG)n . d(pCpG)n inserts, in which n ranges between 8 and 21, shows that the B-to-Z transition within the alternating C-G is readily induced by negative supercoiling and is highly cooperative. The free energy parameters for the transition in dilute aqueous buffers have been evaluated from a statistical mechanical analysis of the data, and these parameters allow prediction of the superhelicities of plasmids at which the transition occurs in alternating C-G inserts over a wide range of lengths. In agreement with the crystal structures, the helical handedness of the B structure in solution and that of the Z structure are shown to be opposite to each other. Furthermore, it is found that the B form of the alternating C-G sequence in solution has a helical periodicity of 10.5 +/- 0.1 base pairs per turn, and the Z form has a helical periodicity of 11.6 +/- 0.3 base pairs per turn. There also appears to be a significant unwinding of the right-handed DNA duplex at each of the B/Z junctions.

Negatively supercoiled plasmids contain left-handed Z-DNA segments as detected by specific antibody binding.

Negative superhelical coiling of covalently closed DNA plasmids facilitates the formation of left-handed Z-DNA segments. This was demonstrated by binding of antibodies specific for Z-DNA. Plasmid pBR322 and two derivatives from it, pLP32 and pLP014, carrying inserts of alternating CG sequences of 32 bp and 14 bp, respectively, were used. Longer inserts required less negative superhelical density to induce the B-Z transitions. Antibody binding to supercoiled plasmids was also visualized by electron microscopy. Cross-linking of the antibody to the negatively supercoiled plasmid and restriction of the DNA with restriction endonucleases demonstrated that the antibodies combine with the CG insert in pLP32. For pBR322, however, evidence suggests that the antibody combines with a section of DNA containing 14 bases with alternating purine and pyrimidine residues with one residue out of alternation: CACGGGTGCGCATG. These cross-linking studies provide evidence for the binding specificity of the anti-Z-DNA antibodies. On the basis of experimental findings, we have calculated the changes in free energy associated with B-Z transitions in superhelical plasmids.

Flipping of cloned d(pCpG)n.d(pCpG)n DNA sequences from right- to left-handed helical structure by salt, Co(III), or negative supercoiling.

Negative supercoiling of plasmid DNAs containing 24--42 base pairs of alternating d(pCpG) inserts is shown to cause the flipping of the helical hand of the inserts from right to left under physiological conditions. For a negatively supercoiled DNA with a fixed linking number, this flipping reduces its superhelicity and, therefore, is accompanied by a shift of its electrophoretic mobility in agarose gel. Quantitation of the mobility shifts indicates that essentially the whole stretch of contiguous alternating d(pCpG) flips to the left-handed helical form when the negative superhelical density (specific linking difference) is greatest than or equal to 0.03.