Multiple degradation pathways of the rpsO mRNA of Escherichia coli. RNase E interacts with the 5' and 3' extremities of the primary transcript

The degradation process of the rpsO mRNA is one of the best characterised in E coli. Two independent degradation pathways have been identified. The first one is initiated by an RNase E endonucleolytic cleavage which allows access to the transcript by polynucleotide phosphorylase and RNase II. Cleavage by RNase E gives rise to an rpsO message lacking the stabilising hairpin of the primary transcript; this truncated mRNA is then degraded exonucleolytically from its 3' terminus. This pathway might be coupled to the translation of the message. The second pathway allows degradation of polyadenylated rpsO mRNA independently of RNase II, PNPase and RNase E. The ribonucleases responsible for degradation of poly(A) mRNAs under these conditions are not known. Poly(A) tails have been proposed to facilitate the degradation of structured RNA by polynucleotide phosphorylase. In contrast, we believe that removal of poly(A) by RNase II stabilises the rpsO mRNA harbouring a 3' hairpin. In addition to these two pathways, we have identified endonucleolytic cleavages which occur only in strains deficient for both RNase E and RNase III suggesting that these two endonucleases protect the 5' leader of the mRNA from the attack of unidentified ribonuclease(s). Looping of the rpsO mRNA might explain how RNase E bound at the 5' end can cleave at a site located just upstream the hairpin of the transcription terminator.

Adjacent sequences influence DNA repair accompanying transposon excision in maize

Mobile elements transposing via DNA intermediates often leave small rearrangements, or "transposon footprints," at sites where they excise. Each excision event leaves its own footprint and, at any given site, these vary in size and sequence. Footprint formation involves DNA repair of sequences flanking the element. We have analyzed the footprints formed by a 2-kb Ds element excising from six different sites in exons of the maize waxy (Wx) gene. We find that groups of footprints left at individual sites are surprisingly nonrandom; different excision products predominate consistently at each site. Less frequent footprints left by each insertion appear related to the predominant type. The data suggest that flanking sequences affect the DNA repair processes associated with element excision. Two models have been proposed to explain footprint formation, one featuring a 5' exonuclease and the other featuring hairpin loop formation and an endonuclease. Our data have interesting implications for both these models. Evidence is also presented to support the presence of a separate excision mechanism that can remove Ac/Ds elements without leaving any footprint and that operates in parallel with the footprint-forming mechanism.

The interaction between lipoamide dehydrogenase and the peripheral-component-binding domain from the Azotobacter vinelandii pyruvate dehydrogenase complex

The sensitivity of lipoamide dehydrogenase (dihydrolipoamide:NAD+ oxidoreductase E3) from Azotobacter vinelandii to inhibition by NADH requires measurement of the activity in the initial phase of the reaction. Stopped-flow turnover experiments show that kcat is 830 s-1 compared with 420 s-1 found in standard steady-state experiments. Mutations at the si-side of the flavin prosthetic group that cause severe inhibition by NADH were studied. Tyr16 was replaced by phenylalanine and serine, which causes the loss of two intersubunit H-bonds. [F16]E3 shows only 5.7% of wild-type activity in the standard assay procedure, but analyzed by stopped-flow the activity is 70% of the wild-type enzyme. The NADH-->Cl2Ind (dichloroindophenol) activity was normal or slightly increased. The inhibition by NADH is competitive with respect to NAD+, Ki = 50 microM. Spectral analysis show that electrons readily pass over from the disulfide to the FAD, indicating an increase in the redox potential of the flavin. It is concluded that subunit interaction plays an important role in the protection of the enzyme against over-reduction by decreasing the redox potential of the flavin. The interaction of wild-type or mutant enzymes with the core component of the pyruvate (E2p) or oxoglutarate (E2o) dehydrogenase multienzyme complex relieves the inhibition to a large extent. In the mutant enzymes, the mechanism of inhibition changes from competitive to the mixed-type inhibition observed for the wild-type enzyme. The stabilizing effect of E2 on [F16]E3 was used as an assay to analyze the stoichiometry of interaction of E3 with E2p as well as E2o. 1 mol E2p monomer was sufficient to saturate 1 mol E3 dimer with a Kd of about 1 nM. Similarly, 1 mol E2o saturated the E3 dimer with a Kd of 30 nM. From these experiments it is concluded that the E3-binding domain of E2 interacts with the subunit interface of E3 near the dyad axis, thus preventing sterically the interaction with a second molecule of the binding domain. This mode of interaction, which causes asymmetry in the complex, explains the stabilization against over-reduction by tightening the subunit interaction. Subgene cloning of the E2p component of the pyruvate dehydrogenase complex is described in order to obtain a complex between the lipoamide dehydrogenase component (E3) and the binding domain of E2p. A unique restriction site in the DNA encoding the flexible linker between the third lipoyl domain and the binding domain combined with timed digestion with exonuclease Bal31 was used to create a set of deletion mutants in the N-terminal region of the binding-catalytic didomain, fused to six N-terminal amino acids from beta-galactosidase. The expressed proteins, selected for E2p activity, were analyzed for binding of E3 and E1p. The shortest fusion protein containing a functional binding domain was expressed and purified. [F16]E3 was combined with this fusion protein in a stoichiometric ratio and the resulting complex was subjected to limited proteolysis to remove the catalytic domain. The resulting [F16]E3-binding domain preparation was purified to homogeneity.

Yeast checkpoint genes in DNA damage processing: implications for repair and arrest

Yeast checkpoint control genes were found to affect processing of DNA damage as well as cell cycle arrest. An assay that measures DNA damage processing in vivo showed that the checkpoint genes RAD17, RAD24, and MEC3 activated an exonuclease that degrades DNA. The degradation is probably a direct consequence of checkpoint protein function, because RAD17 encodes a putative 3'-5' DNA exonuclease. Another checkpoint gene, RAD9, had a different role: It inhibited the degradation by RAD17, RAD24, and MEC3. A model of how processing of DNA damage may be linked to both DNA repair and cell cycle arrest is proposed.

The 3'-5' exonuclease site of DNA polymerase III from gram-positive bacteria: definition of a novel motif structure

The primary structure of the 3'-5' exonuclease (Exo) site of the Gram+ bacterial DNA polymerase III (Pol III) was examined by site-directed mutagenesis of Bacillus subtilis Pol III (BsPol III). It was found to differ significantly from the conventional three-motif substructure established for the Exo site of DNA polymerase I of Escherichia coli (EcPol I) and the majority of other DNA polymerase-exonucleases. Motifs I and II were conventionally organized and anchored functionally by the predicted carboxylate residues. However, the conventional downstream motif, motif III, was replaced by motif III epsilon, a novel 55-amino-acid (aa) segment incorporating three essential aa (His565, Asp533 and Asp570) which are strictly conserved in three Gram+ Pol III and in the Ec Exo epsilon (epsilon). Despite its unique substructure, the Gram+ Pol III-specific Exo site was conventionally independent of Pol, the site of 2'-deoxyribonucleoside 5-triphosphate (dNTP) binding and polymerization. The entire Exo site, including motif III epsilon, could be deleted without profoundly affecting the enzyme's capacity to polymerize dNTPs. Conversely, Pol and all other sequences downstream of the Exo site could be deleted with little apparent effect on Exo activity. Whether the three essential aa within the unique motif III epsilon substructure participate in the conventional two-metal-ion mechanism elucidated for the model Exo site of EcPol I, remains to be established.

Synthesis and biological activities of a phosphorodithioate analog of 2',5'-oligoadenylate

To enhance the resistance of 2-5A (pppA2'p5'A2'p5'A) to degradation by exo- and endonucleases, a phosphorodithioate analog was synthesized using a solid-phase phosphite triester approach with N6-benzoyl-5'-O-dimethoxytrityl-3'-O-t-butyldimethylsilyladenosine 2'-[S-(beta-thiobenzoylethyl)-pyrrolidinophosphorothioamidit e]. 5'-Monophosphorylation was accomplished with 2-[2-(4,4'-dimethoxytrityloxy)-ethylsulfonyl]ethyl-(2-cyanoe thyl)-(N,N- diisopropyl)-phosphoramidite. The resulting product, p5'A2'(s2p)- 5'A2'(s2p)5'A, was approximately 10-fold less effective as an activator of purified human recombinant 2-5A-dependent RNase than was 2-5A itself. This loss of activation ability was related directly to the loss of binding ability of the phosphorodiothioate analog. As predicted, p5'A2'(s2p)5'A2' (s2p)5'A was stable to snake venom phosphodiesterase and the nucleolytic activities of both human lymphoblastoid CEM cell extracts and human serum, under conditions that led to facile degradation of parent 2-5A. This nuclease stability permitted the observation of the CEM cell extracts and human serum phosphatase activity which led to 5'-dephosphorylation of p5'A2'(s2p)5'A2'(s2p)5'A.

Removal of anti-human immunodeficiency virus 2',3'-dideoxynucleoside monophosphates from DNA by a novel human cytosolic 3'-->5' exonuclease

A 3'-->5' exonuclease has been highly purified from the cytosol of human acute lymphoblastic leukemia H9 cells. The apparent molecular weight of this enzyme was approximately 50,000, as indicated by its sedimentation in glycerol gradients. The exonuclease did not copurify with DNA polymerase activity, required MgCl2 for its exonucleolytic activity, and was inhibited by KCl above 60 mM. The enzyme was active on single-stranded DNA, DNA duplexes and DNA/RNA duplexes, and it was efficient at removing 3'-terminal mispairs from DNA. The products of the exonucleolytic reaction were deoxynucleoside 5'-monophosphates. The behavior of the exonuclease was examined on DNA terminated at the 3' end with a variety of dideoxynucleosides that are potent against human immunodeficiency virus type 1. The exonuclease has a broad substrate specificity; however, the rate of the enzymatic reaction varied among the D dideoxynucleosides tested (ddAMP = ddCMP > d4TMP > AZTMP). Similarly, the enzyme was examined for its reactivity with DNA terminated by either the D or L enantiomers of ddC, SddC or FddC. The removal of analogs with the native D configuration was at least 6-fold more rapid than that of the L-compounds, and the type of structural modification had an impact on the rate at which the D enantiomers were removed (SddCMP > ddCMP > FddCMP). The monophosphate forms of AZT, D4T, L-FddC and L-ddC were potent inhibitors of the exonuclease at micromolar concentrations, while D-ddCMP partially inhibited the enzyme at millimolar concentrations. Based on its physical and enzymatic properties, this exonuclease represents a novel enzyme that may have an important role in determining the relative potencies of dideoxynucleosides against human immunodeficiency virus type 1.

An aphidicolin-resistant mutant of Chinese hamster ovary cell with altered DNA polymerase and 3' exonuclease activities

A comparison was made of partially purified DNA polymerases alpha, delta, epsilon and from normal Chinese hamster ovary cells and mutant cells (JK3-1-2A) resistant to aphidicolin, araA, and araC. In vitro the pol alpha from the mutant cells (1) was resistant to aphidicolin and araCTP but was sensitive to araATP, (2) showed a 1.6 to 2.6-fold reduced specific activity, and (3) was more sensitive to carbonyldiphosphonate, DMSO and SJK 287-38 anti-pol alpha antibody inhibition, but was less sensitive to alkylphenyl nucleotide analogs BuPdGTP and BuAdATP. On the other hand, pol delta and pol epsilon of the mutant cells did not show increased aphidicolin-resistance but differed from the wild type enzymes with regard to their 3' exonuclease activity. The 3' exonuclease/DNA polymerase activity ratio was increased 6-fold for pol delta and 3.3-fold for pol epsilon for enzymes from the mutant cells in comparison to wild type values. It is suggested that these altered properties of the DNA polymerases of the alpha-family are responsible for in vivo aphidicolin resistance of the mutant cells. The higher 3' exonuclease activity may explain the observed antimutator phenotype of this cell line. In view of the proficient 3' exonuclease activities of the DNA pol delta and epsilon, the present aphr mutant is unique among all mammalian DNA polymerase mutants.

Rolipram increases cyclic GMP content in L-arginine-treated cultured bovine aortic endothelial cells

Cultured bovine aortic endothelial cells only contain two cyclic nucleotide phosphodiesterases isoforms: PDE II (cyclic GMP stimulated) and PDE IV (rolipram sensitive). The effects of cilostamide or rolipram alone or together, on cyclic AMP and cyclic GMP levels, were measured in indomethacin-treated endothelial cells alone or in the presence of nitric oxide (NO) modulators. In all conditions, cyclic AMP levels were potently increased (8-13-fold) only when PDE II and PDE IV inhibitors were given together. Cyclic GMP levels were not modified by these PDE inhibitors in control and NG-nitro-L-arginine-methyl ester-treated cells. But surprisingly, in L-arginine-treated cells, cyclic GMP content was increased by 42% by rolipram alone, and combination of rolipram with cilostamide resulted in a further increase in cyclic GMP content (to 153% compared to control cells). These results suggest that in presence of the NO synthase substrate (L-arginine), an increase in cyclic AMP level may upregulate the L-arginine/NO/cyclic GMP pathway.

Use of monoclonal antibodies in the functional characterization of the Saccharomyces cerevisiae Sep1 protein

The Saccharomyces cerevisiae strand-exchange protein 1 (Sep1 also known as Xrn1, Kem1, Rar5, Stp beta/DST2) has been demonstrated to mediate the formation of hybrid DNA from model substrates of linear double-stranded and circular single-stranded DNA in vitro. To delineate the mechanism by which Sep1 acts in the strand-exchange reaction, we analyzed mouse anti-Sep1 monoclonal antibodies for inhibition of the Sep1 in vitro activity. Of 12 class-G immunoglobulins tested, four were found to consistently inhibit the Sep1-mediated strand-exchange reaction. The inhibiting antibodies were tested for inhibition of a variety of Sep1-catalyzed DNA reactions including exonuclease activity on double-stranded and single-stranded DNA, renaturation of complementary single-stranded DNA and condensation of DNA into large aggregates. All four inhibiting antibodies had no effect on the exonuclease activity of Sep1. Three antibodies specifically blocked DNA aggregation. In addition, one antibody inhibited renaturation of complementary single-stranded DNA. This inhibition pattern underlines the importance of condensation of DNA into large aggregates in conjunction with double-stranded DNA exonuclease activity for the in vitro homologous pairing activity of Sep1. The implications of these data for the interpretation of proteins which promote homologous pairing of DNA are discussed, in particular in light of the reannealing activity of the p53 human tumor-suppressor protein.

Incision activity of human apurinic endonuclease (Ape) at abasic site analogs in DNA

The major apurinic/apyrimidinic (AP) endonuclease of human cells, the Ape protein, incises DNA adjacent to abasic sites to initiate DNA repair and counteract the cytotoxic and mutagenic effects of AP sites. Here we address the determinants of Ape AP endonuclease activity using duplex DNA substrates that contain synthetic analogs of AP sites: tetrahydrofuranyl (F), propanediol (P), ethanediol (E), or 2-(aminobutyl)-1,3-propanediol (Q). The last of these, a branched abasic structure, was a poor substrate for which Ape had kcat > 1000-fold lower than for F. In contrast, the specificity constant (kcat/Km) for E or P of Ape purified from HeLa cells was only 5-8-fold lower than for F. Positioning a phosphorothioate ester immediately 5' to F inhibited Ape incision activity 20-fold (Rp isomer) or > 10,000-fold (Sp isomer). Although Ape did not have detectable endonuclease activity toward single-stranded substrates or unmodified double-stranded DNA, the enzyme displayed a low level of 3'-exonuclease activity for duplex DNA (< 0.03% of its AP endonuclease activity), which was influenced by the reaction conditions. The base positioned opposite F did not dramatically affect the cleavage efficiency of Ape, but an F:F arrangement was cleaved at approximately one-third of the efficiency of F:C. A 3'-mismatch diminished P and E cleavage only slightly and F not at all. A 5'-mismatch reduced the Ape cleavage rate 4-10-fold for F and approximately 100-fold for P and E. A series of substrates with F at different positions along the oligonucleotide showed that Ape requires > or = 4 base pairs 5' to the abasic site and > or = 3 base pairs on the 3'-side. The implications of these results for substrate recognition by Ape are discussed.

Purification and properties of wild-type and exonuclease-deficient DNA polymerase II from Escherichia coli

Wild-type DNA polymerase II (pol II) and an exonuclease-deficient pol II mutant (D155A/E157A) have been overexpressed and purified in high yield from Escherichia coli. Wild-type pol II exhibits a high proofreading 3'-exonuclease to polymerase ratio, similar in magnitude to that observed for bacteriophage T4 DNA polymerase. While copying a 250-nucleotide region of the lacZ alpha gene, the fidelity of wild-type pol II is high, with error rates for single-base substitution and frameshift errors being < or = 10(-6). In contrast, the pol II exonuclease-deficient mutant generated a variety of base substitution and single base frameshift errors, as well as deletions between both perfect and imperfect directly repeated sequences separated by a few to hundreds of nucleotides. Error rates for the pol II exonuclease-deficient mutant were from > or = 13- to > or = 240-fold higher than for wild-type pol II, depending on the type of error considered. These data suggest that from 90 to > 99% of base substitutions, frameshifts, and large deletions are efficiently proofread by the enzyme. The results of these experiments together with recent in vivo studies suggest an important role for pol II in the fidelity of DNA synthesis in cells.

Exonucleolytic processing of small nucleolar RNAs from pre-mRNA introns

Many small nucleolar RNAs (snoRNAs) in vertebrates are encoded within introns of protein genes. We have reported previously that two isoforms of human U17 snoRNA are encoded in introns of the cell-cycle regulatory gene, RCC1. We have now investigated the mechanism of processing of U17 RNAs and of another intron-encoded snoRNA, U19. Experiments in which the processing of intronic RNA substrates was tested in HeLa cell extracts suggest that exonucleases rather than endonucleases are involved in the excision of U17 and U19 RNAs: (1) Cutoff products that would be expected from endonucleolytic cleavages were not detected; (2) capping or circularization of substrates inhibited formation of snoRNAs; and (3) U17 RNA was faithfully processed from a substrate carrying unrelated flanking sequences. To study in vivo processing the coding regions of snoRNAs were inserted into intron 2 of the human beta-globin gene. Expression of resulting pre-mRNAs in simian COS cells resulted in formation of correctly processed snoRNAs and of the spliced globin mRNA, demonstrating that snoRNAs can be excised from a nonhost intron and that their sequences contain all the signals essential for accurate processing. When the U17 sequence was placed in a beta-globin exon, no formation of U17 RNA took place, and when two U17 RNA-coding regions were placed in a single intron, doublet U17 RNA molecules accumulated. The results support a model according to which 5'-->3' and 3'-->5' exonucleases are involved in maturation of U17 and U19 RNAs and that excised and debranched introns are the substrates of the processing reaction.

Characterization of Epstein-Barr virus DNase and its interaction with the major DNA binding protein

Bacterially expressed Epstein-Barr virus (EBV) DNase was purified to 98% purity and used as the source for characterization of the enzyme activities. Complete digestion of DNA by EBV DNase yielded 5'-monophosphate nucleosides as the final products. During the logarithmic phase of the reaction, EBV DNase acted processively on dsDNA but distributively on ssDNA. Both 5' to 3' and 3' to 5' exonuclease activities were present, although the former was shown to be 10-fold stronger. No significant discrepancy was seen in the liberation of end-labeled nucleotides by DNase when substrates with 5'-protruding, blunt, or 3'-protruding ends were used. EBV DNase was demonstrated also to have an endonuclease activity using supercoiled plasmid DNA as substrate. Two preferential dsDNA cleavage sites were mapped on pBS-TR, a pBlueScript vector containing one copy of the EBV terminal repeat; both are in vector sequences. Finally, an N-terminally truncated EBV major DNA binding protein, but not EA-D, was shown to inhibit EBV DNase activity. This inhibitory effect may due to direct protein-protein interactions between EBV DNase and the major DNA binding protein. The biological significance of these characteristics is discussed.

Directional cloning of an oligonucleotide fragment into a single restriction site

Oligonucleotide fragments can be directionally subcloned into vectors at a single restriction site. By using T4 DNA polymerase exonuclease activity to treat vector DNA, single-stranded ends can be generated. The oligonucleotide sequences are designed to have sequence complementary to these single-stranded ends. Through the homologous annealing of oligonucleotides to the treated vector ends, the successfully subcloned molecules forms a circular recombinant DNA that is ready for transformation. There is no sequence restriction at the ends of the DNA fragment. All restriction site ends are accessible to this method. This approach for oligonucleotide fragment insertion and together with our previously described general method of exonuclease induced DNA subcloning provide convenient methods for the construction of recombinant DNA.

RPA involvement in the damage-recognition and incision steps of nucleotide excision repair

Human replication protein (RPA) functions in DNA replication, homologous recombination and nucleotide excision repair. This multisubunit single-stranded DNA-binding protein may be required to make unique protein-protein contacts because heterologous single-stranded binding proteins cannot substitute for RPA in these diverse DNA transactions. We report here that, by using affinity chromatography and immunoprecipitation, we found that human RPA bound specifically and directly to two excision repair proteins, the xeroderma pigmentosum damage-recognition protein XPA (refs 8, 9) and the endonuclease XPG (refs 10-13). Although it had been suggested that RPA might function before the DNA synthesis repair stage, our finding that a complex of RPA and XPA showed a striking cooperativity in binding to DNA lesions indicates that RPA may function at the very earliest stage of excision repair. In addition, by binding XPG, RPA may target this endonuclease to damaged DNA.

Evidence for giant linear plasmids in the ascomycete Podospora anserina

In the extrachromosomal mutant AL2 of the ascomycete Podospora anserina longevity is correlated with the presence of the linear mitochondrial plasmid pAL2-1. In addition to this autonomous genetic element, two types of closely related pAL2-1-homologous molecules were detected in the high-molecular-weight mitochondrial DNA (mtDNA). One of these molecules is of linear and the other of circular structure. Both molecules contain pAL2-1 sequences which appear to be integrated at the same site in the mtDNA. Sequence analysis of a DNA fragment cloned from one of these molecules revealed that it contains an almost full-length copy of pAL2-1. At the site of plasmid integration a 15-nucleotide AT-spacer and long inverted mtDNA sequences were identified. Finally, two giant linear plasmid-like DNAs of about 50 kbp and 70 kbp were detected in pulsed-field gels of mutant AL2. These molecules are composed of mtDNA and pAL2-1-specific sequences and may result from the integration of mtDNA sequences into linear plasmid pAL2-1.

Mammalian endo-exonuclease activity and its level in various radiation sensitive cell lines

The levels of endo-exonuclease in various mammalian cell lines were examined. While the expression of the endo-exonuclease during cell growth behaved exactly the same as the pattern observed in lower eukaryotes, the amount of activity was found to be reduced in the radiosensitive Chinese hamster ovary (CHO) xrs-5 and various human AT, AT-5 and NE-1 cells when compared to the radionormal CHO K1 and human HeLa cell lines. The reduced endo-exonuclease activity in these cells was due to a decreased amount of protein as demonstrated with the immuno-blot method. The results presented here suggest that endo-exonuclease may be one of the many proteins whose expression is regulated by genes coding for xrs-5 in CHO and AT in humans.

Characterization of two nuclear mammalian homologous DNA-pairing activities that do not require associated exonuclease activity

We have developed an assay to study homologous DNA-pairing activities in mammalian nuclear extracts. This assay is derived from the POM blot assay, described earlier, which was specific for RecA activity in bacterial crude extracts. In the present work, proteins from mammalian nuclear extracts were resolved by electrophoresis on SDS/polyacrylamide gels and then electrotransferred onto a nitrocellulose membrane coated with circular single-stranded DNA (ssDNA). The blot obtained was incubated with a labeled homologous double-stranded DNA (dsDNA). Homologous pairing between the ssDNA and the labeled dsDNA was detected by autoradiography as a radioactive spot on the membrane. In nuclear extracts from mammalian cells, we found two major polypeptides of 100 and 75 kDa, able to promote the formation of stable plectonemic joints. Joint molecule formation required at least one homologous end on the dsDNA, but either end of the dsDNA could be recruited to initiate the reaction. For each polypeptide, the reaction required divalent cations such as Mg2+, Ca2+, or Mn2+. Although ATP was not necessary, ADP was inhibitory in each case. Unlike most of the known eukaryotic DNA-pairing proteins, both activities identified here were able to promote the formation of joint molecules without requiring an associated exonuclease activity. In addition, these two proteins were detected in cell lines from different tissues and from different mammalian species (human, mouse, and hamster).

(2'-5')-Oligo-3'-deoxynucleotides: selective binding to single-stranded RNA but not DNA

Oligodeoxynucleotides with (2'-5') internucleotide linkages have been synthesized on a solid support via standard cyanoethyl phosphoramidite chemistry. This simple change in the oligonucleotide bond connectivity led to unique properties. UV melting temperature experiments indicate that the (2'-5')-oligo-3'-deoxyadenylates, (2'-5')-3'-dA8 and (2'-5')-3'-dA8(s) phosphorothioate, hybridize selectively to single-stranded RNA but not DNA. The complex (2'-5')-3'-dA8:poly (U) (Tm = 32 degrees C) was nearly as stable as the natural (3'-5')-2'-dA8 and poly (U) (Tm = 33 degrees C) in 130 mM NaCl, and 10 mM phosphate buffer (pH 7.5). However, no association was observed upon mixing (2'-5')-3'-dA8 and poly (dT). The (2'-5') linkages also confer greater resistance to exo- and endonucleolytic degradation compared with (3'-5')-linked oligomers. The rate of degradation of (2'-5')-3'-dA8 was almost four times less than that of (3'-5')-2'-dA8 in cell culture medium containing 10% heat-inactivated fetal calf serum. An increase in stability for (2'-5')-3'-dA8 against endonuclease activity was observed in both cytoplasmic and nuclear extracts. The nucleic acid selectivity of (2'-5')-oligo-3'-deoxynucleotides may represent an important design feature to improve the efficacy of antisense oligonucleotides.

Synthesis and evaluation of oligodeoxynucleotides containing acyclic nucleosides: introduction of three novel analogues and a summary

Novel flexible oligodeoxynucleotides containing (S)-1-(2,3-dihydroxypropyl)thymine or 2',3'-seco-thymidine nucleoside analogues were synthesized on an automated DNA-synthesizer. Oligodeoxynucleotides with one, two or three acyclic nucleosides incorporated in the middle or in the ends of 17-mers have been evaluated. 3'-End-modified oligomers were significantly stabilized towards 3'-exonucleolytic degradation compared to unmodified analogues and showed acceptable hybridization properties as measured by UV experiments. For oligodeoxynucleotide analogues containing the three novel acyclic monomers in the middle, a more pronounced reduction in duplex stability was observed. All oligodeoxynucleotides containing acyclic nucleoside analogues made so far are evaluated with respect to stability towards 3'-exonucleolytic degradation and hybridization properties.

The fidelity of the human leading and lagging strand DNA replication apparatus with 8-oxodeoxyguanosine triphosphate

A product of oxidative metabolism, 8-oxodeoxyguanosine triphosphate (8-O-dGTP), readily pairs with adenine during DNA replication, ultimately causing A.T-->C.G transversions. This study utilized 8-O-dGTP as a probe to examine the fidelity of the leading and lagging strand replication apparatus in extracts of HeLa cells. Simian virus (SV) 40 T antigen-dependent DNA replication reactions were performed with two M13mp2 vectors with the SV40 origin located on opposite sides of the lacZ alpha sequence used to score replication errors. The presence of 8-O-dGTP at equimolar concentration with each of the 4 normal dNTPs resulted in a > 46-fold increase in error rate for A.T-->C.G transversion over that observed in the absence of 8-O-dGTP. A similar average error rate was observed on the (+) and (-) strands in both vectors, suggesting that the fidelity of replication by leading and lagging strand replication proteins is similar for the dA.8-O-dGMP mispair. Replication fidelity in the presence of 8-O-dGTP was reduced on both strands when an inhibitor of exonucleolytic proofreading (dGMP) was added to the reaction. These data suggest that the majority of dA.8-O-dGMP mispairs are proofread by both leading and lagging strand replication proteins.