Search

We have analyzed middle repetitive DNA in the albumin and vitellogenin gene families of Xenopus laevis. Mapping specific repetitive DNA sequences derived from introns of the A1 vitellogenin gene reveals that these sequences are scattered within and around the four vitellogenin genes (A1, A2, B1 and B2) and the two albumin genes (74 kd and 68 kd). Three repetitive DNA elements present in the A1 vitellogenin transcriptional unit are also located in introns of the 74 kd albumin gene. This apparently random distribution of middle repetitive DNA in the two gene families suggests that the analyzed sequences are not involved in gene regulation, but rather that they might represent unstable genetic elements. This hypothesis is further supported by the finding that size polymorphism in the A1 vitellogenin gene and in the 74 kd albumin gene is correlated with the presence or absence of repetitive DNA.

Ribosomal RNA and precursor ribosomal RNA from at least one representative of each vertebrate class have been analyzed by electron microscopic secondary structure mapping. Reproducible patterns of hairpin loops were found in both 28 S ribosomal and precursor ribosomal RNA, whereas almost all the 18 S ribosomal RNA molecules lack secondary structure under the spreading conditions used. The precursor ribosomal RNA of all species analyzed have a common design. The 28 S ribosomal RNA is located at or near the presumed 5′-end and is separated from the 18 S ribosomal RNA region by the internal spacer region. In addition there is an external spacer region at the 3′-end of all precursor ribosomal RNA molecules. Changes in the length of these spacer regions are mainly responsible for the increase in size of the precursor ribosomal RNA during vertebrate evolution. In cold blooded vertebrates the precursor contains two short spacer regions; in birds the precursor bears a long internal and a short external spacer region, and in mammals it has two long spacer regions. The molecular weights, as determined from the electron micrographs, are 2·6 to 2·8 × 10 6 for the precursor ribosomal RNA of cold blooded vertebrates, 3·7 to 3·9 × 10 6 for the precursor of birds, and 4·2 to 4·7 × 10 6 for the mammalian precursor. Ribosomal RNA and precursor ribosomal RNA of mammals have a higher proportion of secondary structure loops when compared to lower vertebrates. This observation was confirmed by digesting ribosomal RNAs and precursor ribosomal RNAs with single-strandspecific S 1 nuclease in aqueous solution. Analysis of the double-stranded, S 1 -resistant fragments indicates that there is a direct relationship between the hairpin loops seen in the electron microscope and secondary structure in aqueous solution.

Under certain extraction conditions, 28-S rRNA of Xenopus laevis can form specific dimers by intermolecular renaturation of a particular hairpin loop near the B end of the molecules. This was demonstrated by electron microscopy of isolated aggregates under conditions where rRNA shows a reproducible pattern of secondary structure. In 50% formamide, 15 mM sodium chloride, 1.5 mM trisodium citrate, pH 7.0, the aggregates dissociate into their monomers at a temperature (61 degrees C) at which the most stable hairpin loops of the 28-S rRNA are melting. Precursor rRNAs containing the 28-S rRNA region also form aggregates with 28-S rRNA by the same linkage mechanism. Since such artificial aggregates yield molecular weights higher than expected for true precursors of Xenopus rRNA, they may interfere in precursor rRNA studies. Analysis of whole-cell RNA pulse-labeled with (Me-3H)methionine, or nucleolar RNA pulse-labeled with (3H)uridine in the presence of tubercidin (inhibitor of rRNA processing), as well as electron microscopy of high-molecular-weight RNA fractions confirm that the 40-S precursor rRNA (2.8 X 10(6)) is the primary transcript of the ribosomal genes in Xenopus laevis.

The 68K and 74K albumin genes of Xenopus laevis arose by duplication approximately 30 million years ago. Electron microscopic analysis showed that both genes contain 15 coding sequences. The lengths of corresponding coding sequences are almost identical and are extremely similar to those of mammalian albumin genes. A block of four coding sequences, which in mammals codes for one protein domain, is repeated three times. The corresponding introns are usually different in length and have therefore diverged as a result of insertion/deletion events. The extensive homology between these gene sequences is neither confined to nor most extensive in the coding sequences and similar amounts of homologous sequences are found in the flanking DNAs as in the gene regions. Various structures were formed in the 5'-flanking DNA by mutually exclusive pairing of different homology regions. Analysis of the two 74K albumin gene sequences isolated suggests that the X. laevis genome may contain one 68K albumin gene and two very closely related 74K albumin genes.

cDNA clones containing sequences complementary to Xenopus laevis albumin mRNA have been identified in a collection of cDNA clones made from poly(A)+ RNA prepared from male Xenopus laevis liver. Although all the albumin cDNA clones crosshybridise, restriction enzyme and heteroduplex analysis show that there are 2 closely related albumin mRNA sequences. The 2 albumin mRNAs are only mismatched by 8% but could be isolated by positive selection using stringent hybridization conditions. Oocytes injected with the 2 purified mRNAs, secreted either the 68,000 or 74,000 dalton albumin into the culture medium showing that the 2 albumins of X. laevis serum are encoded in the 2 closely related mRNAs. Measurements of the abundance of albumin mRNA show that the 2 albumin mRNAs together account for about 9% of total poly(A)+ RNA in male Xenopus laevis liver but the mRNA coding for the 74,000 dalton mRNA is about twice as abundant as that coding for the 68,000 dalton mRNA.

Estrogen treatment of Xenopus males leads to the appearance of a new species of poly (A)-containing RNA in the liver, at a stage when large amounts of the estrogen-induced yolk precursor protein, vitellogenin, is produced. This estrogen-induced RNA sediments at 28 S and migrates on gels in aqueous solution with an apparent molecular weight of 2.0 X 10(6). Contour length measurements under denaturing conditions in the electron microscope reveal a molecular weight of 2.34 X 10(6) compared to the mouse 28-S rRNA. Labeling experiments show that the estrogen-induced RNA has a stability than the average liver poly(A)-containing RNA and represents 10-20% of the poly(a)-containing RNA in the cytoplasm after 24 h of labeling. Hybridization of complementary DNA, synthesized on the isolated estrogen-induced RNA, with its template suggests a complexity corresponding to a single species of poly(A)-containing RNA of such a high molecular weight. Hybridization of the complementary DNA with cytoplasmic poly (A)-containing RNA from estrogen-treated Xenopus males and control toads show that the estrogen-induced RNA constitutes 12-15% of all cytoplasmic poly(A)-containing RNA, and is at least 2000-fold less abundant in untreated males. Size, complexity and abundance of the estrogen-induced RNA are characteristics expected for a mRNA coding for vitellogenin.

The structural organization of the two closely related vitellogenin genes A1 and A2 has been determined and compared by electron microscopy. In both genes the mRNA-coding sequence of 6 kb is interrupted 33 times, leading to a total gene length of 21 kb for gene A1 and 16 kb for gene A2. Thus both genes have a mean exon length of 0.175 kb, while the mean intron length is 0.45 kb in gene A1 and 0.31 kb in gene A2. Because the introns interrupt the structural sequence at homologous positions in genes A1 and A2, we suggest that these two genes are the products of a duplication of an ancestral gene which has an intron-exon arrangement similar to that of the extant genes. Since the duplication event, the sequence and length of the analogous introns have changed rapidly, whereas homologous exons have diverged to an extent of only 5% of their sequences. The results suggest different mechanisms of evolution for exons and introns. While the exons evolved primarily by point mutations, such mutations, as well as deletion, insertion and duplication events, were important in the evolution of the introns.

Analysis of cDNA clones synthesized from vitellogenin mRNA of X. tropicalis revealed three different types of cDNA clones, i.e. A, A* and B. A and A* clones have a sequence divergence of about 6% and are both related to X. laevis vitellogenin cDNAs of subgroup A1 as well as A2 with a sequence divergence of 6-9%. B clones however, are related to X. laevis cDNA clones of subgroup B1 and B2 with a sequence divergence of about 7%. While the A and B clones correspond to vitellogenin mRNAs of similar abundance, A* clone is complementary to a vitellogenin mRNA about 100 fold less abundant than A and B mRNAs although all three vitellogenin mRNAs are encoded by single copy genes. Furthermore, two forms of A* mRNA were found. One of the two is lacking an internal fragment of about 900 bp. Since this DNA fragment is highly repeated in the genome, we suggest that this A* clone was synthesized from a processing intermediate of the A* precursor vitellogenin mRNA.

The switch from larval to adult globin synthesis during amphibian metamorphosis was analyzed by polyacrylamide gel electrophoresis in acid urea-Triton X-100. 4–6 slowly migrating components were detected in larvae and at least 4 faster components appear at metamorphosis and persist in the adult stage. To relate the globin subunits to the mRNA sequences, isolated as cDNA clones, mRNAs were hybrid-selected from total RNA of larval and adult erythrocytes and translated in a wheat germ system. The translation products were identified by comparison with the globin pattern of hemolysates

The genome of eukaryotes consists of single-copy DNA interspersed with middle repetitive DNA. Both types of DNA are known to be transcribed (reviewed in ref. 1). Several reports have shown that some nuclear RNA is composed of single-copy DNA transcripts covalently linked to sequences transcribed from repetitive DNA2–4, but the significance of these nuclear transcripts is unknown; there is no convincing evidence that these nuclear transcripts containing repetitive DNA sequences are precursors to mRNAs. To elucidate this question we have analysed the structure of the two related vitellogenin genes A1 and A2 of Xenopus laevis5,6. The expression of these genes is initiated in liver cells by oestrogen (reviewed in refs 7–9). Preliminary experiments demonstrated that the flanking region as well as several introns of these genes contain middle repetitive DNA9. We have now identified and mapped six different repetitive DNA sequences within the transcribed introns of the A1 vitellogenin gene. Further characterization of these sequences implies that they are evolutionary unstable. Analysis of nuclear RNA revealed that transcripts of the six sequences are also found in non-vitellogenic liver cells, suggesting that these repetitive DNAs are also present in transcriptional units active in the absence of oestrogen.

Cloning of vitellogenin cDNA of Xenopus laevis revealed that vitellogenin is encoded in a small family of genes representing two distantly related main groups A and B, each comprising two more closely related subgroups A1, A2, and B1, B2 respectively. To characterize the proteins derived from these genes we have isolated the corresponding mRNAs by hybridizing, under stringent conditions, cytoplasmic poly(A)-containing RNA from the liver of estrogen-stimulated Xenopus to filter-bound cDNA clones containing sequences specific for all four vitellogenin genes. Hybridization of the isolated mRNAs with nick-translated cDNA clones revealed that contamination of the mRNAs by those of the other main group was less than 0.1 %. Melting curves of the hybrids prepared with the isolated mRNAs and cDNA clones specific for the four vitellogenin genes showed that the isolated vitellgenin mRNAs are also specific for the four subgroups. Analysis of R loops formed between isolated mRNAs and cDNA clones representing the corresponding subgroup further indicated about 10 % cross-contamination between the more closely related mRNAs. In a reticulocyte lysate each of the four mRNAs coded for a 200000-Mr protein immunoprecipitable by monospecific vitellogenin antibody. From these results we conclude that the four different mRNAs A1, A2, B1 and B2, which all can be isolated efficiently, code for vitellogenin and are expressed simultaneously in response to estrogen stimulation.

The mRNA coding for vitellogenin, the yolk protein precursor, has been isolated from the liver of estrogen-stimulated Xenopus laevis. The mRNA has a size of 6.3 kilobases (kb). Optimal conditions were investigated for the synthesis of long complementary DNA (cDNA, referring to DNA synthesized in vitro) copies of the mRNA. Temperature, salt concentration, and enzyme-to-RNA ratio were important factors. Double-stranded cDNA with an average size of 2 to 3 kb was inserted into the vector pMB9 by the poly(dA:dT) method, and the recombinant plasmids were amplified in E. coli. Twenty-one clones with vitellogenin inserts ranging from 1 to 3.7 kb were studied. The regions in the RNA from which these clones had been derived were mapped by R-loop analysis in the electron microscope and by hybridization of the cloned DNAs with specific fractions of mRNA. Slightly more than half of the clones were derived from the 3′-terminal portions of the mRNA while the remaining clones are located internally.

Vitellogenin, the yolk protein precursor, is produced in X. laevis liver from a 6.3 kilobase (kb) mRNA. Sequences of this mRNA have been transcribed into cDNA and cloned in E. coli. Some properties of 21 of these cloned DNAs, ranging in size from 1 to 3.7 kb, have been reported by Wahli et al. (1978b). This paper reports restriction endonuclease mapping, cross hybridization, heteroduplex mapping in the electron microscope and heteroduplex melting experiments with these DNAs. We conclude that the cloned DNAs fall into two main groups of sequences which differ from each other in approximately 20% of their nucleotides. Each main group contains two subgroups which differ from each other by about 5% sequence divergence. By hybridizing cloned DNAs with restricted genomic DNA, we showed that sequences corresponding to all four sequence groups are present in a single animal. Furthermore, we have obtained tentative evidence for the presence of large intervening sequences in genomic vitellogenin DNA. Analysis of R loop molecules demonstrated that all four sequences are present in the vitellogenin mRNA population purified from individual animals. While some alternate explanations are not entirely excluded, we suggest that vitellogenin is encoded by a small family of related genes in Xenopus.

cDNA synthesized on purified vitellogenin mRNA from Xenopus liver was hybridized to the template in formamide/urea at 22 degrees C to avoid degradation of the RNA. The hybrids formed were visualized by spreading for electron microscopy. Contour length measurements proved that most of the RNA molecules in the hybrids were still intact showing the expected molecular weight of 2.3 x 10(6). The hybridized cDNA corresponded on the average to 12% of the RNA length. In about 80% of the molecules the cDNA was located at one end. Since cDNA synthesis was primed by oligo(dT), the terminal duplex region marks the 3' end of the vitellogenin mRNA molecule. Internal duplex regions were mainly located at a specific position starting about 2800 nucleotides from the 3' end. Since the cDNA hybridizing at the internal position could specifically be synthesized on a vitellogenin RNA fragment isolated on poly(U)-Sepharose as an oligo(A)-containing RNA, we conclude that cDNA synthesis is not only initiated by the poly(A) of the 3' end, but also by a specific internal sequence.

High-resolution electrophoresis of larval and adult hemoglobins of Xenopus laevis reveals stage-specific differences in the number and mobility of the globin chains. To establish the relationship between the globin chains and the previously described globin genes, the corresponding mRNAs were hybrid-selected from total erythroblast RNA by representative cDNA clones, and translated in vitro. Electrophoretic separation of the translation products allowed identification of a major and a minor α-globin chain in the larval and adult stages. This also holds for the adult β-chains, however in the larval stage a difference in abundance is only detectable in the β-mRNAs, but not in the translation products, because they comigrate. The fact that major and minor globin chains can be assigned to genes, which are located in two clusters, suggests that the related genes are expressed coordinately, but at different levels. Analysis of the globin patterns during development reveals that transition from the larval to the adult globin chains coincides with metamorphosis. Moreover, there is evidence of two globin chains that are only expressed in early larval stages and hence might be related to additional larval β-globin genes of as yet unknown genomic location.