Search

We grew two-dimensional crystals of HeLa cell prosomes, also called multicatalytic proteinases (MCP) and proteasomes, for a structure determination by electron microscopy. The molecules were arranged in side views in these crystals. The crystals have p21 plane group symmetry with one particle per unit cell. This symmetry confirms previously published evidence indicating that eukaryotic prosome-MCPs are dimers of two identical halves. Structure factors from six crystals each comprising more than 1000 unit cells were combined to generate a 1.5-nm projection map. We discovered that while the general cylindrical shape of HeLa prosome-MCPs resembles the shape of the archaebacterial Thermoplasma acidophilum proteasomes, the internal structure differs significantly. We propose that because of different subunit composition, the architecture of HeLa prosome-MCPs differs from the basic architecture of related particles previously reported.

The concentration of glutamate dehydrogenase (GDH) varies strongly between different organs and between different regions within organs. To permit further studies on the regulation of GDH expression, we isolated and characterized the rat gene encoding the GDH protein. This gene contains 13 exons and spans approximately 34 kbp. The GDH gene is present as a single, autosomally located copy in the Wistar rat genome, but shows an extensive restriction-fragment-length polymorphism for several enzymes. Promoter activity of the 5'-flanking sequence is shown by transient transfection experiments. The 5'-flanking sequence contains a TTAAAA sequence at position -29, instead of a consensus TATA box and, like many other TATA-less promoters, is characterized by a very high G + C content. In addition, consensus sequences for the binding sites of the transcription factors Sp1 and Zif268 are present in the G + C-rich upstream region.

The Adenovirus DNA-binding protein (DBP) binds to single-stranded (ss) DNA as well as to double-stranded (ds) DNA and forms multimeric protein-DNA complexes with both. Gel retardation assays indicate rapid complex formation for both DNAs. DBP rapidly dissociates from dsDNA, indicating a dynamic equilibrium, whereas the ssDNA-DBP complex is much more stable. We investigated the complex between DBP and dsDNA in more detail. Electron microscopical analysis shows thick filament-like and beaded structures in which the length of the DNA is not significantly altered. Cryo-electron micrographs suggest the presence of interwound protein fibres around the DNA. Ligase-mediated cyclization, but not linear multimerization, of DBP-saturated DNA fragments exceeding the persistence length was severely inhibited. This suggests that DNA may be organized by DBP into a rigid structure. Under those conditions, DBP induces distinct changes in the circular dichroism spectrum of the DNA, indicative of structural DNA changes. No bending or twisting of the complex was observed. Hydroxyl radical footprinting showed that the breakdown pattern of DNA at saturating DBP concentrations is much more regular than the protein-free DNA. This suggests the removal of tertiary structures, which may be related to the effects of DBP on enhanced NFI binding and chain elongation during Adenovirus DNA replication. Using purified proteins in an in vitro replication system, we correlate the structural changes with the effects of DBP on enhancement of NFI-binding as well as on DNA replication.

Duck erythroblasts prosomes were analysed by small angle neutron scattering (SANS), dynamic light scattering and (cryo-)electron microscopy. A molecular weight of approximately 720,000 +/- 50,000, a radius of gyration of 64 +/- 2 A and a hydrodynamic radius of approximately 86 A were obtained. Electron micrographs show a hollow cylinder-like particle with a diameter of 120 A, a height of 170 A and a diameter of 40 A for the cavity, built of four discs, the two outer ones being more pronounced than those in the center. Results from SANS indicate less then 5% of RNA in the purified prosomes, but nuclease protection assays confirm its presence.

The genes encoding the large (cfxL) and small (cfxS) subunits of ribulose-1,5-bisphosphate carboxylase (RuBisC/O) from Xanthobacter flavus H4-14 were identified and characterized. The RuBisC/O genes are separated by 11 bp and cotranscribed in Escherichia coli from the lac promoter in the order cfxLS. Primer extension and R-loop experiments with RNA isolated from autotrophically grown X. flavus H4-14 showed that transcription of cfxL and cfxS initiated 22 bp upstream from cfxL and resulted in a mRNA of at least 2.3 kb. DNA sequence analysis identified the start of an open reading frame transcribed divergently from cfxL, and displaying significant similarities with genes belonging to the lysR family of transcriptional activators. Downstream from cfxS an additional open reading frame was identified with unknown function. Expression studies showed that the genes encoding fructosebisphosphatase (cfxF) and phosphoribulokinase (cfxP) are located downstream from cfxLS. The cfxF and cfxP genes are cotranscribed in the same direction as cfxLS in the order cfxFP.

From a rat genomic library in phage lambda Charon4A, a complete glutamine synthetase-encoding gene was isolated. The gene is 9.5-10 kb long, consists of seven exons, and codes for two mRNA species of 1375 nucleotides (nt) and 2787 nt, respectively. For both mRNAs, full-length cDNAs containing a short poly(A) tract were identified. The sequences of the entire mRNA and of the exon-intron transitions were determined. The smaller mRNA is identical to the 5' 1375 nt of the long mRNA and contains the entire protein-coding region. The position of the transcription start point was mapped. Within the first 118 bp of promoter sequence, a (T)ATAA-box, a CCAAT-box and an SP1-binding site were identified.

The gene for subunit I of cytochrome c oxidase, contained within the OX13 region of yeast mitochondrial DNA, is split and shows a remarkable variation in structure, which is strain-dependent. The most complex form so far characterized is that of the Saccharomyces cerevisiae strain KL14-4A, in which nine or possibly ten exons are separated by eight to nine introns. At least four of these are facultative, two being absent from S. cerevisiae strain D273-10B (sequenced by Bonitz et al., 1980) and a further two lacking from the gene in Saccharomyces carlsbergensis. The complexity of the gene in KL14-4A is also reflected in its transcript pattern. RNA blot hybridization with isolated and cloned DNA fragments of the OX13 region permits visualization of more than 60 RNAs, which show overlapping and discontinuous hybridization behaviour. In the less complex strains D273-10B and S. carlsbergensis, this number is 20 and 11, respectively. These RNAs are most likely intermediates in processing events leading to the appearance of the mature messenger RNA for cytochrome c oxidase subunit I, which we identify as a 2100-nucleotide transcript (18SE). Most of the processing events are dependent on mitochondrial protein synthesis and do not constitute a single obligatory processing pathway. Like other yeast mitochondrial mRNAs, the 18 S RNA contains a long, untranslated 5' flanking sequence (approximately 400 nucleotides). One unusual aspect of splicing events involving OX13 transcripts is the accumulation of three of the excised introns as single-stranded RNA circles. These abundant and stable transcripts appear to be covalently closed. The simplest assumption is that they arise as (by)-products of splicing, but secondary ligation events have not been excluded. Their function is as yet unknown.

We describe two cloned genomic DNA fragments, both bearing the entire apo very low density lipoprotein II gene. Electron microscopy and restriction enzyme mapping showed that this gene is split into at least four coding sequences by three or more intervening sequences. A very short exon at the 5'-end of the gene is separated by a 1.5-kilobase intron from the second exon, which codes for the AUG initiation codon of the mRNA.

We have analysed the structure of the mtDNAs of six amicronucleate Tetrahymena pyriformis strains, belonging to at least four phenosets, as defined by Borden et al. (Borden, D., Whitt, G.S. and Nanney, D.L. (1973) J. Protozool. 20, 693--700). 2. The mtDNAs of all strains are linear, but they differ in size, in their fragmentation by endonuclease EcoRI and in overall sequence; less than 20% sequence homology was found by DNA-DNA hybridization in all combinations tested, except for the mtDNAs from strains T and ST which are indistinguishable. 3. In spite of these marked sequence differences the mtDNAs of all strains share two structural peculiarities: ragged (gnawed) duplex ends and a duplication-inversion, which varies in length between 0.3 and 1.2 micrometer, depending on the strain. In four strains the duplication-inversion is terminal, allowing formation of single-stranded DNA circles with a duplex tail; in two strains it is subterminal. 4. The ragged ends and sub-terminal position of the duplication-inversion in some of the Tetrahymena mtDNAs do not fit any of the current models for the replication of linear mtDNAs.

11S and 18S fractions of yeast mitochondrial RNAs, isolated by electrophoresis through agarose gels, have been found by electron microscopy to contain approximately 50% circular molecules. Circles in the 11S fraction have a contour length of 0.36 +/- 0.02 micron, which is approximately equal to the length of the majority of linear molecules also present. Circles in the 18S fraction have an average length of 0.78 +/- 0.11 micron. The size distribution is broader than for the 11S fraction, and we cannot exclude the possibility that more than one size class may be present. The 11S circular RNA forms circular R loops and RNA-DNA hybrids with DNA fragments of the oxi 3 region of mtDNA, which contains the structural gene for subunit 1 of cytochrome oxidase. As judged from the electron micrographs, the complete RNA participates in hybrid formation and the sequences coding for it appear to be continuous. Both 11S and 18S circles withstand treatment with DNAase and pronase. They are not eliminated by treatment with 1 M glyoxal in 50% formamide for 1 hr at 50 degrees C. We conclude that they are covalently closed. The function of the circular RNAs is unknown. They may be active as mRNAs, storage forms, or arise in a cut-and-splice process which generates mRNAs from longer transcripts.

Excised group II introns in yeast mitochondria appear as covalently closed circles under the electron microscope. We show that these circular molecules are branched and resemble the lariats arising through splicing of nuclear pre-mRNAs in yeast and higher eukaryotes. One member of this intron class (aI5c in the gene for cytochrome c oxidase subunit I) is capable of self-splicing in vitro, giving correct exon-exon ligation and resulting in the appearance of both linear and lariat forms of the excised intron. Nuclease digestion of the latter molecules reveals the presence of a complex oligonucleotide with the probable structure AGU, which thus resembles the branch point formed in the spliceosome-dependent reactions undergone by nuclear pre-mRNAs. Unlike group I introns, this group II intron is not demonstrably dependent on GTP for self-splicing and circularization of the isolated, linear intron is not observed. A model accounting for these observations is presented.

We have cloned overlapping segments of the human thyroglobulin gene from a genomic cosmid library. Restriction mapping and electron microscopy show that a region of 38 kb at or near the 3'-end of this gene encodes only 850 nucleotides or 10% of the messenger RNA (mRNA) sequence. The region contains five exons of 130-210 nucleotides, split by introns of 1 to 15-17 kb. This represents the lowest ratio of coding to non-coding DNA (2.2%) found thus far in any eukaryotic gene. Blot hybridization under non-stringent conditions shows the presence of only one copy of this gene in the human genome and the absence of other closely related sequences.

The amino acid sequence of the egg yolk storage protein phosvitin has been deduced from the nucleotide sequence of part of the chicken vitellogenin gene. Of the phosvitin sequence, 210 amino acids including the N-terminal residue are contained on one large exon, whereas the remaining six amino acids are encoded on the next exon. Phosvitin contains a core region of 99 amino acids, consisting of 80 serines, grouped in runs of maximally 14 residues interspersed by arginines, lysines, and asparagines. The serines of the core region are encoded by AGC and AGT codons exclusively and the arginines by AGA and AGG, which results in a continuous stretch of 99 codons with adenine in the first position. The N-terminal quarter of the phosvitin sequence contains 16 serines grouped in a cluster with alanines and threonines and coded mainly by TCX triplets. The C-terminal part includes 27 serines, preferentially coded by AGC and AGT, 13 histidine residues, and the sequence ...Asn-Gly-Ser... at which the carbohydrate moiety of phosvitin is attached. Heteroduplex formation between cloned DNAs from chicken and Xenopus vitellogenin genes shows that the phosvitin sequence contains a stretch of highly conserved sequence.

The elongation factor EF-1 alpha is one of the most abundant proteins in eukaryotic cells, where it catalyzes the binding of aminoacyl-tRNA to ribosomes. The genes coding for this protein in the brine shrimp Artemia were analyzed by gene cloning, electron microscopy and chromosomal blot hybridization. There are only a few (about four) copies of one type of gene per haploid genome. These genes contain five exons divided over 10(4) base pairs. Local rearrangements give rise to a number of gene variants. Cross-hybridizations of Artemia cDNA probes with yeast and Drosophila DNA revealed two different yeast EF-1 alpha genes and one or two different Drosophila genes, respectively. Nucleotide sequencing revealed signals for synthesis and processing of EF-1 alpha transcripts as well as the exact location of exons. One interruption in the coding sequence corresponds closely to a splice junction in the gene coding for the homologous chloroplast protein EF-Tu from Euglena gracilis, presumably of prokaryotic origin. The first exon in the chloroplast gene codes for the region of EF-Tu that is homologous to regions of the elongation factor EF-G and of the initiation factor IF2, respectively.

RNA containing the aI3 group I intron of the yeast mitochondrial gene encoding cytochrome oxidase subunit I shows self-splicing in vitro. The excised intron, comprising 1514 nucleotides, is partially split into an upstream portion, containing the intronic reading frame, and a downstream portion, containing the typical group I conserved sequence elements. Full-length intron RNA and intron parts occur in linear and circular form. In the transesterification reactions leading to circle formation, only the guanosine nucleotide added during splicing is removed. Reincubation of isolated, complete circular intron RNA under self-splicing conditions leads to formation of free subintronic RNA circles. Under similar conditions, purified linear intron RNA gives rise to a number of circular and linear products, one of which consists of interlocked subintronic RNA circles. These observations suggest that the intron RNA possesses a dynamic structure in which subtle alterations in folding result in the formation of RNA products with different topology.

A model for the quaternary structure of Escherichia coli RNA polymerase (nucleosidetriphosphate:RNA nucleotidyltransferase, EC 2.7.7.6) is presented. It is based on results from classification of profiles of enzyme molecules, and from application of immuno electron microscopy. Classification of molecules, prepared with the single carbon layer technique, was first achieved for images recorded in dark field with the scanning transmission electron microscope and later on for images recorded in bright-field transmission electron microscopy. It results in five approximately equally sized groups, containing about 80% of the core enzyme profiles. Holoenzyme profiles can be grouped into the same classes, and have approximately the same dimensions (9 nm X 16 nm). Based on the shapes and sizes of the classified profiles, a tentative model for core enzyme has been constructed. Correlation of shadow projections of this model, with the distributions of attachment sites of antibodies against alpha, beta, beta' and sigma over the profiles, has led to models for core and holoenzyme in which the subunits are localized. The model is compared with literature data on the quaternary structure of RNA polymerase.

When transcribing a restriction fragment containing the promoters and the first part of the rrnE operon of Escherichia coli, RNA polymerase holoenzyme starts exclusively on the promoters. Besides run-off transcripts, molecules longer than template-size are formed by terminal strand switch.

We have studied splicing of precursors to the large ribosomal RNA and processing of the excised intron in yeast mitochondria using primer extension with reverse transcriptase and electron microscopy. Structural features of the following intermediates are described: first, a linear RNA carrying a 5'-terminal G that is not encoded in mitochondrial DNA; second, a circular RNA in which the 3' and 5' intron borders are covalently linked. Three nucleotides of the 5' intron border are absent from the site of circle closure. The properties of these intermediates fit remarkably well into the mechanism of self-splicing described for the ribosomal precursor RNA from Tetrahymena nuclei. A new feature of the yeast mitochondrial system is that the excised intron can have one of two destinies, circularization or cleavage at an internal position.

We have cloned the AMO gene, encoding the microbody matrix enzyme amine oxidase (EC 1.4.3.6) from the yeast Hansenula polymorpha. The gene was isolated by differential screening of a cDNA library, immunoselection, and subsequent screening of a H. polymorpha genomic library. The nucleotide sequence of a 3.6 kilobase stretch of DNA containing the amine oxidase (AMO) gene was determined. The AMO gene contains an open reading frame of 692 amino acids, with a relative molecular mass of 77,435. The 5' and 3' ends of the gene were mapped and show that the transcribed region measures 2134 nucleotides. The derived amino-acid sequence was confirmed by sequencing an internal proteolytic fragment of the purified protein. Amine oxidase contains the tripeptide sequence Ser-Arg-Leu, located 9 residues from the carboxy terminus, which may represent the topogenic signal for protein import into microbodies.

1. We have studied denatured Tetrahymena mtDNA by electron microscopy using the formamide technique. 2. After denaturation all DNA is single stranded, but within a few minutes single-stranded circles with a duplex tail are formed. 3. The duplex tail is 1.3 mum long, i.e. 8 percent of the length of native mtDNA, and it often contains a small single-stranded eye. 4. Digestion of the duplex DNA with exonuclease III of Escherichia coli abolishes its ability to form circles and duplex tails after denaturation. 5. Renaturation of denatured mtDNA leads to the formation of duplex circles with single-stranded section and/or duplex tails. In addition, a minority of duplex circles without apparent tails is formed, but these circles contain a small ambiguous section. 6. We conclude that this mtDNA contains a long terminal duplication-inversion, that could be involved in the replication of this linear mtDNA.

Self-splicing of the precursor to large ribosomal RNA of yeast mitochondria leads not only to circles but also to lariats, structures that have not been observed before as products of self-splicing. Lariats were studied by electron microscopy after hybridization with an RNA complementary to the 3' half of the precursor. This leads to differentiation in at least two classes of lariats that vary in the position of the branch point. In all lariats the tail carries the 3' end, which suggests that a 5' end is used for branch formation with an internal nucleotide. The circles are formed from excised introns. They lack only three nucleotides encoded by mitochondrial DNA along with the 5'-terminal G added in the course of self-splicing. The diverse number of self-splicing products arising in vitro testifies to the considerable reactivity of this intron. The formation of lariats in an RNA catalyzed reaction may have implications for views on the mechanism of splicing of nuclear pre-mRNAs.

1. Isolated mtDNA from Tetrahymena pyriformis strain GL is a linear duplex molecule with an average molecular weight of 32.6 - 10(6) and without internal gaps or breaks. Denaturation of this DNA results in single strands with a duplex hairpin at one end. The length of this hairpin varies between 0 and 5 micrometer within one preparation. 2. Uder renaturation conditions the single strands with hairpins are able to circularize in two ways, depending on the length of the hairpin. Circularization is also observed after partial digestion with exonuclease III of native strain GL mtDNA. 3. All these data fit a model (see Fig.2) in which the DNA is heterogeneous in length at both ends. At the left end a 10-micrometer duplication-inversion is present; part of this duplication-inversion is complementary to a region at the right end of the molecule. 4. The analogy between the structural peculiarities of strain GL mtDNA and of some linear viral DNAs is stressed.

An allele giving rise to a polymorphism within the 3' part of the chicken vitellogenin gene was cloned, sequenced, and compared to the previously cloned allele. The polymorphism is formed by a perfect copy of 343 bp from intron 32 in tandem array with a perfect copy of 244 bp from intron 33; this 587-bp element is inserted in a head-to-tail arrangement in intron 33. We propose a mechanism in which an unequal crossing-over resulted in a vitellogenin gene with two exons 33, one of which was subsequently deleted. Thus, intron 33 was enlarged by the tandem repeats without affecting the protein-encoding sequence of the gene. At the boundaries of the repeated elements, two short direct repeats are found that resemble the recombination signals of immunoglobulin genes. They may have had a key role in the formation of the new allele.

Hepatitis delta (delta) virus (HDV), a satellite virus of the hepatitis B virus (HBV), causes a severe form of viral hepatitis in humans. Here we present evidence based on electron microscopy and electrophoretic behaviour that HDV contains a single stranded circular RNA molecule. This is the first animal virus identified with a circular RNA genome. Circular RNAs have only been found in plant viruses. We have obtained a partial complementary DNA clone representing approximately 25% of the total genome of HDV. Analysis of this cDNA revealed similarity to two plant viruses that may explain the origin of the virus.

The physical fate of plasmid DNA after entry into human fibroblasts was studied using Southern hybridisation and electron microscopy. Exposure of the cells (5 X 10(5) per well) to pC194 DNA-CaPi, containing 50 micrograms plasmid DNA, resulted in the occasional formation of interlocked molecules. Exposure to a co-precipitate containing 100 micrograms pC194 plasmid DNA per well resulted in an increase of interlocked molecules by a factor of 10-20 relative to the number of monomers. In addition, new classes of molecules were observed. After prolonged incubation of the cells exposed to the higher DNA concentration, the plasmid DNA was partly contained in structures with a very low electrophoretic mobility. Upon restriction endonuclease digestion of the re-extracted DNA, a pattern of bands was observed, suggesting the involvement of illegitimate recombination between non-random plasmid DNA sequences in the formation of the new classes of molecules.

A series of overlapping recombinant clones, which cover the vitellogenin gene, has been isolated from a phage-lambda linked chicken gene library. The DNA of the overlapping clones spans 28 kb of contiguous DNA sequences in the chicken genome. Electron microscopic analysis of hybrids between vitellogenin mRNA and the genomic clones indicates that the chicken vitellogenin gene has a length of approximately 22 kb, about 3.8 times the size of the mRNA. The mRNA sequence is interrupted by at least 33 intervening sequences (introns). Comparison with the vitellogenin gene A2 from Xenopus laevis (Wahli et al., 1980, Cell 20: 107-117) indicates conservation of the number and length of the exons during evolution. Heteroduplex analysis reveals a short stretch of sequence homology between the genes from chicken and frog.

DNA prepared from bacteriophage GA-1 contains circular DNA molecules, which are converted to linear molecules by treatment with trypsin.

Employing an exonuclease III protection assay we detected a protein in crude HeLa nuclear extracts binding, with apparent sequence specificity, to molecular ends of adenovirus type 2 (Ad2) DNA. This protein, designated nuclear factor IV (NFIV), was purified to homogeneity and was shown to be a hetero-dimer of 72,000 and 84,000 Mr. Binding to terminal Ad2 sequences was strongly enhanced by the presence of either of the sequence-specific DNA-binding proteins nuclear factor I and nuclear factor III. These proteins appeared to function as blockades for translocation of NFIV on DNA, thus producing apparent sequence specificity. In the absence of such a blockade, NFIV moved freely, without energy input, on any double-stranded DNA forming a regular DNA-multimeric protein complex as shown by methidiumpropyl EDTA footprinting and electron microscopy. Binding is completely dependent upon the presence of molecular ends. Evidence was obtained for a two-step mechanism in which termini are recognized by NFIV and used as a starting point for subsequent translocation. The possible functions of the protein in adenovirus DNA replication and in cellular processes requiring DNA termini are discussed.

A goat strain with congenital goiter was studied as a model for human thyroid disoders. These goats were deficient in thyroglobulin (Tg), the precursor protein of the thyroid hormones T3 and T4. RNA coding for Tg (Tg-RNA) was detected in reduced amounts in the goiters and was almost absent from the membranes, where Tg is normally synthesized. This paper describes the preparation and characterization of a goat Tg cDNA plasmid and its use in the study of thyroglobulin gene expression in the goiter. We found that the goiter Tg-RNA is polyadenylated and has the same size as normal 33S goat TG mRNA. Thus, there are no major defects in the mRNA processing. However, in contrast to normal Tg mRNA, the goiter Tg-RNA was not translated into immunoprecipitable Tg subunits when injected into Xenopus oocytes. We conclude therefore that the goiter Tg-RNA has one or more alterations causing a lack of proper translation and/or a decreased cytoplasmic stability.