C. Weissmann

Structural relationship of human interferon alpha genes and pseudogenes

We have isolated and characterized DNA segments containing IFN-alpha-related sequences from human lambda and cosmid clone banks. We describe six linkage groups comprising 18 distinct IFN-alpha-related loci, and report the nucleotide sequences of nine chromosomal IFN-alpha-genes with intact reading frames, as well as of five pseudogenes. Taking into account as yet unsequenced genes as well as clones described by others, there are now seven linkage groups and 23 loci, of which 15 correspond to potentially functional genes and six to non-functional genes; two loci remain unsequenced. Eighteen additional sequences are likely to be allelic to the above. The finding that at least two IFN-alpha genes appear to be natural hybrids of other IFN-alpha genes, and that two distinct IFN-alpha loci have completely identical coding sequences, although their flanking regions are different, is evidence for information exchange between the individual genes.

Purification of globin messenger RNA from dimethylsulfoxide-induced friend cells and detection of a putative globin messenger RNA precursor☆

Abstract Procedures are described that permit the detection and isolation of a specific messenger RNA as well as its precursor from total cell extracts. DNA complementary to the mRNA was elongated by the addition of dCMP residues and annealed with labeled cell RNA. The elongated DNA with RNA hybridized to it was isolated by chromatography on a poly(I)-Sephadex column. The method was used to isolate 32P-labeled globin mRNA from labeled Friend cells, a mouse erythroleukaemic cell line, induced with dimethylsulfoxide to synthesize hemoglobin. 32P-labeled globin mRNA isolated by this procedure was estimated to be 80% pure by hybridization analysis and sedimented as a single peak at 10 S. Partial sequences were determined for 16 oligonucleotides derived from the purified 32P-labeled globin mRNA by RNAase T1 digestion. The partial sequences for nine oligonucleotides corresponded to those predicted from the amino acid sequences of α and β globin; the other oligonucleotides were presumably derived from non-translated regions. In order to detect a possible precursor to globin mRNA, RNA from induced Friend cells pulse-labeled with [32P]phosphate for 20 minutes was centrifuged through a sucrose gradient and the resulting fractions were analyzed for globinspecific sequences. Two peaks of globin-specific RNA were detected, a larger one at 10 S, the position of mature globin mRNA, and a smaller one at 15 S.

Site-directed mutations in the initiator region of the bacteriophage Qβ coat cistron and their effect on ribosome binding☆

Qβ plus strands with a 70 S ribosome bound to the coat cistron initiation site were used as template for Qβ replicase. Minus strand synthesis proceeded until the replicase reached the ribosome. The ribosome was removed and elongation was continued in a substrate-controlled, stepwise fashion. The nucleotide analog N4-hydroxyCMP was introduced into the positions complementary to the third and fourth nucleotides of the coat cistron. The minus strands were elongated to completion, purified and used as template for Qβ replicase. The final plus strand preparation consisted of four species, with the sequences -A-U-G-G- (wild type), -A-U-A-G- (mutant C3), -A-U-G-A- (mutant C4) and -A-U-A-A- (mutant C3C4) at the coat initiation site. The ribosome binding capacity of the mutant RNAs relative to wild type was <0.1 (C3), 3.2 (C4) and 0.3 (C3C4). The finding that mutant C3 no longer formed an initiation complex suggests that the interaction of the ribosome binding site with fMet-tRNA plays an essential role in the formation of the 70 S initiation complex. The fact that mutant C4 RNA bound more efficiently than wild type, and that mutant C3C4 RNA showed substantial ribosome binding capacity whereas the single mutant C3 did not, can be explained by assuming that an A residue following the A-U-G triplet interacts with a complementary U residue in the anticodon loop sequence. In the case of C3C4 this additional base-pair may offset the reduced codon-anticodon interaction resulting from the modification of the A-U-G codon.

Site-directed mutagenesis: Generation of an extracistronic mutation in bacteriophage Qβ RNA

Abstract The preparation in vitro and chemical characterization of bacteriophage Qβ RNA with an extracistronic mutation, a G → A transition in the 16th position from the 3′-terminus, is described. The 5′-terminal region of the Qβ minus strand was synthesized in vitro up to position 14 (inclusive) by using ATP and GTP as the only substrates. The mutagenic nucleotide analog N4-hydroxyCMP was then incorporated into position 15 instead of CMP. The minus strand was completed with the four standard ribonucleoside triphosphates, purified and used as a template for the synthesis of plus strands. Of the plus strand product, 33% had a G → A transition in the 16th position from the 3′-end (which corresponds to position 15 of the minus strand), as shown by nucleotide sequence analysis of the terminal T1 oligonucleotide. The modified RNA was efficiently replicated by Qβ replicase and a preparation containing 55% of the mutant RNA was obtained. The general approach to directed mutagenesis outlined above should allow the introduction of mutations into the 5′ and 3′-terminal regions of Qβ RNA as well as into the intercistronic sequences.

Preparation and characterization of form V DNA, the duplex DNA resulting from association of complementary, circular single-stranded DNA

Abstract Complementary circular single strands of plasmid PβG or bacteriophage PM2 DNA but not of single-stranded φX174 DNA associate under hybridisation conditions, giving rise to a two-stranded complex. This DNA, which we call form V, has well-defined physico-chemical properties. It sediments as a sharp peak in neutral sucrose gradients; its electrophoretic mobility in agarose gels is between that of covalently closed (form I) and denatured DNA. In the electron microscope form V appears as highly folded duplex molecules indistinguishable from form I. However, increasing concentrations of ethidium bromide which lead to relaxation and recoiling of form I DNA have no comparable effect upon form V. At 260 nm form V PβG DNA has a hypochromicity of 18.6%, as compared to 23.4% in the case of PβG form II DNA and 10.5% in the case of single-stranded φX174 DNA. The thermal melting of form V is non-cooperative with gradual increase in absorbance similar to that of single-stranded DNA. The circular dichroism spectrum of form V DNA differs from that of form I, circular nicked (form II) and single-stranded φX174 DNA in that it shows a negative band at 295 nm and a shift for the main positive band from 273 to 266 nm. We propose that form V consists of right-handed Watson-Crick type double-helices which are compensated by an equal number of left-handed duplex turns and negative supercoils. Wo cannot decide whether left-handed duplex turns are stabilised by base-stacking and hydrogen bonding, as for example in the models described by Rodley et al. (1976) or Sasisekharan & Pattabiraman (1976), or whether they are merely compensatory turns without inherent stability.

The 3′-termini of bacteriophage Qβ plus and minus strands

The 3′-end groups of bacteriophage Qβ plus and minus strands of different origin have been determined by terminal labelling with KB8H4 (or NaB3H4) following oxidation with sodium periodate. Qβ minus strands, as well as Qβ plus strands (extracted from the viral particle or from infected bacteria), terminate predominantly in adenosine, and to a lesser degree in cytidine. Qβ RNA synthesized in vitro by the purified Qβ replicase system also has mostly adenosine at the 3′-terminus, whether (a) Qβ RNA, (b) Qβ RNA from which the 3′-terminal nucleotide had been removed chemically, or (c) denatured double-stranded Qβ RNA (in the absence of factors), was used as template. The same enzyme system, however, was unable to transfer an adenylate residue from [α-32P]ATP to Qβ RNA lacking the 3′-terminal adenosine. It is concluded that free, 3′-adenosineless Qβ RNA is not an intermediate in Qβ synthesis and that only nascent Qβ RNA can serve as an acceptor for the 3′-terminal pA.

Physical mapping of Qβ replicase binding sites on Qβ RNA

Abstract Using the recently developed benzyldimethylalkylammoniumchloride-spreading method for nucleic acid molecules (Vollenweider et al. , 1975), crosslinked complexes of Qβ replicase and Qβ RNA were visualized by electron microscopy. Two types of complexes were observed: one showed a single loop in the RNA, with the replicase bound at its base. This structure appears to come about by the crosslinking of one replicase molecule to two sites located at 29.5 ± 2.5 and at 65 ± 3.5 map units from one end of the RNA. The other type of complex consists of an extended RNA strand with a single replicase molecule attached to one of the two sites revealed by the looped complexes. In no case was more than one replicase molecule associated with an RNA strand. In contrast to these findings with Qβ replicase, DNA-dependent RNA polymerase of Escherichia coli bound randomly to Qβ RNA. At the enzyme-to-RNA ratio used, up to five enzyme molecules attached to one RNA chain, and the RNA did not form specific loops at the polymerase binding sites.

A new approach to the isolation of RNA-DNA hybrids and its application to the quantitative determination of labeled tumor virus RNA☆

Abstract A procedure has been developed by which the hybrid formed between a labeled RNA and complementary DNA can be selectively separated from all other single and double-stranded nucleic acids. We describe the application of this procedure to the quantitative determination of labeled avian tumor virus RNA. Purified DNA complementary to avian myeloblastosis virus RNA is extended at its 3′ terminus with 40 to 60 dCMP residues, using terminal deoxynucleotidyl-transferase. The elongated DNA is annealed with the labeled nucleic acid preparation and the mixture is passed through a column of Sephadex to which poly(I) has been covalently bound. The poly(I) retains the specific RNA-DNA hybrids by virtue of their poly(C) extension. The column is washed with RNAase to degrade nonhybridized RNA, the RNA retained on the column is eluted with formamide and its radioactivity is determined. The background hybridization was reduced to 0.003 to 0.008% by addition of oligo(C) 5.20 to the hybridization mixture and by carrying out the adsorption to the poly(I)-Sephadex column in the presence of poly(U). The hybridization efficiency was about 50%. The content of radioactive Rous sarcoma virus-specific RNA was determined in infected and uninfected cells after labeling with [ 3 H]uridine for two hours. The content of labeled virus-specific RNA in infected cells was 0.6 to 0.9% and 0.05% in uninfected cells. The value found for monkey cell RNA was 0.009%. This method can be used for the detection of hybrids between labeled RNA and complementary DNAs too short to allow quantitation by conventional methods. If the RNAase step is omitted the procedure can be used for the isolation of any RNA for which a complementary DNA is available, as well as for its precursor.

Escherichia coli ribosomes bind to non-initiator sites of Qβ RNA in the absence of formylmethionyl-tRNA

Abstract Escherichia coli ribosomes and Qβ [ 32 P]RNA were incubated with or without fMet-tRNA under protein initiation conditions, treated with RNase A, and centrifuged through a sucrose density gradient. The sample incubated with fMet-tRNA gave a main radioactivity peak in the 70 S region, which consisted predominantly of coat cistron initiator fragments. After incubation without fMet-tRNA, equal amounts of radioactivity were found in the 70 S and the 30 S regions, but in both peaks almost all of the radioactivity was duo to three RNase A-resistant oligonucleotides, A-G-A-G-G-A-G-G-Up (P-2a), A-G-G-G-G-G-Up (P-15) and G-G-A-A-G-G-A-G-Cp (P-4). These three oligonucleotides are derived from three different RNA regions, none of which is close to a protein initiation site. All three fragments show striking complementarity to the 3′-terminal region of E. coli 16 S RNA. Ribosomes incubated with an RNase A digest of Qβ [ 32 P]RNA bound almost exclusively oligonucleotide P-2a; treatment with cloacin DF13 cleaved off a complex consisting of a 49-nucleotide long segment of 16 S rRNA and oligonucleotide P-2a. These experiments show that the interaction of 30 S ribosomes with the “Shine-Dalgarno” region preceding the initiator codon of the Qβ coat cistron is insufficient to direct correct placement of the ribosome on the viral RNA, and that an additional contribution from the interaction of fMet-tRNA with the initiator triplet is required for ribosome binding to the initiator region.

Resynchronization of RNA synthesis by coliphage Qβ replicase at an internal site of the RNA template

Abstract In previous work Qβ replicase has been used to synthesize labelled 5′ terminal segments of Qβ plus or minus strands of defined length. A procedure has now been developed which allows resynchronization of Qβ replicase at an internal position and synthesis of a labelled minus-strand segment complementary to the coat cistron ribosome binding site and the intercistronic region between the A 2 (maturation) and the coat cistron. Resynchronization is accomplished by binding a ribosome to Qβ RNA and allowing Qβ replicase to initiate and elongate up to the ribosome, using unlabelled ribonucleoside triphosphates. The ribosome is dissociated by EDTA treatment and the EDTA is removed. The replicating complex remains functional after this treatment, and addition of labelled substrates leads to synchronized elongation. The radioactive part of the product recovered after a short elongation period with labelled substrates was shown to be complementary to the coat protein ribosome binding site.