Influence of polyethylene glycol on the ligation reaction with calf thymus DNA ligases I and II

High concentrations of the nonspecific macromolecule polyethylene glycol 6000 (PEG 6000) enabled DNA ligases I and II from calf thymus to catalyze intermolecular blunt-end ligation of duplex DNA. Intermolecular cohesive-end ligation with these enzymes was markedly stimulated in the presence of 10-16% (w/v) PEG 6000. The effect of PEG 6000 (4-16%) on the sealing of single-stranded breaks in duplex DNA with DNA ligases I and II was not appreciably stimulatory but rather inhibitory. PEG 6000 (15%) enhanced more twofold the rate of DNA ligase II-AMP complex formation, but moderately suppressed the rate of formation of DNA ligase 1-AMP complex. Polyamines and KCl inhibited blunt-end and cohesive-end ligations with DNA ligases I and II in the presence of PEG 6000.     

Bacteriophage T4 and human type I DNA ligases relax DNA under joining conditions.

Both bacteriophage T4 and human type I DNA ligases in the presence of a mixture of ATP, AMP and PPi altered the topological properties of a supercoiled substrate by a step-wise reaction eventually leading to a population of fully relaxed, covalently closed products. In the presence of only AMP and PPi DNA products containing nicks with 3'OH/5'P termini accumulated in the presence of bacteriophage T4 DNA ligase, suggesting reversal of the entire joining reaction, but not in the presence of human DNA ligase I. Both DNA ligases became deoxyadenylylated in the presence of dATP, but the joining reaction did not proceed to completion. However, with both enzymes the full relaxing reaction took place in the presence of dAMP alone and in the presence of a mixture of dATP, dAMP and PPi. In no case could the joining reaction be reversed by dAMP and PPi. Related experiments with modified derivatives of deoxyribonucleoside 5'-triphosphates and PPi gave results in accord with these observations. The AMP dependent DNA relaxation catalysed by DNA ligases was insensitive to the presence of exonuclease III. These results indicate that controlled relaxation of the substrate by both DNA ligases occurs as a separate reaction rather than by simple reversal of the joining reaction. These findings support the hypothesis that in vivo the DNA topoisomerising ligases relax their substrate at the replication fork both during and separately from ligation of a pre-existing nick.     

Immunochemical analysis of molecular forms of mammalian DNA ligases I and II

Using specific antibodies against calf thymus DNA ligases I and II (EC 6.5.1.1), we have investigated the polypeptide structures of DNA ligases I and II present in the impure enzyme preparations, and estimated the polypeptides of DNA ligases I and II present in vivo. Immunoblot analysis of DNA ligase I after sodium dodecyl sulfate-polyacrylamide gel electrophoresis revealed a 130-kDa polypeptide as a major one in the enzyme preparations from calf thymus throughout the purification. In addition to the 130-kDa polypeptide, a 200-kDa polypeptide was detected in the enzyme preparations at the earlier steps of the purification, and a 90-kDa polypeptide was observed as a minor one in the enzyme preparations at the later steps of the purification. The polypeptides with molecular weight of 130 000 and 90 000 were detected by SDS-polyacrylamide gel electrophoresis of DNA ligase I-[3H]AMP complex. These results suggest that a 200-kDa polypeptide of DNA ligase I present in vivo is degraded to a 130-kDa polypeptide and then to a 90-kDa polypeptide during the isolation and purification procedures. On the other hand, the monospecific antibody against calf thymus DNA ligase II cross-reacted with only a 68 kDa polypeptide in the enzyme preparations throughout the purification, suggesting that the 68-kDa polypeptide is a single form of calf thymus DNA ligase II present in vivo as well as in vitro.     

Delayed DNA joining at 3' mismatches by human DNA ligases.

Repair synthesis catalysed by DNA polymerase beta at 1 nt gaps occurs in the main pathway of mammalian base excision repair. DNA polymerase beta has no exonucleolytic proof-reading ability, and exhibits high error frequency during DNA synthesis. Consequently, continuous correction of endogenous DNA damage by short-patch repair synthesis might lead to a high spontaneous mutation rate, unless subsequent steps in the repair pathway allow for selective removal of incorporation errors. We show here that both human DNA ligase I and III discriminate strongly between a correctly paired versus a mispaired residue at the 3' position of a nick in DNA, when assayed in the presence of physiological concentrations of KCl. The resulting delay in joining after misincorporation by DNA polymerase beta during gap filling could allow for removal of the mismatched terminal residue by a distinct 3' exonuclease.     

Mammalian DNA ligases. Catalytic domain and size of DNA ligase I.

DNA ligase I is the major DNA ligase activity in proliferating mammalian cells. The protein has been purified to apparent homogeneity from calf thymus. It has a monomeric structure and a blocked N-terminal residue. DNA ligase I is a 125-kDa polypeptide as estimated by sodium dodecyl sulfate-gel electrophoresis and by gel chromatography under denaturing conditions, whereas hydrodynamic measurements indicate that the enzyme is an asymmetric 98-kDa protein. Immunoblotting with rabbit polyclonal antibodies to the enzyme revealed a single polypeptide of 125 kDa in freshly prepared crude cell extracts of calf thymus. Limited digestion of the purified DNA ligase I with several reagent proteolytic enzymes generated a relatively protease-resistant 85-kDa fragment. This domain retained full catalytic activity. Similar results were obtained with partially purified human DNA ligase I. The active large fragment represents the C-terminal part of the intact protein, and contains an epitope conserved between mammalian DNA ligase I and yeast and vaccinia virus DNA ligases. The function of the N-terminal region of DNA ligase I is unknown.     

Mammalian DNA ligases. Biosynthesis and intracellular localization of DNA ligase I

Mammalian DNA ligase I is presumed to act in DNA replication. Rabbit antibodies against the homogeneous enzyme from calf thymus inhibited DNA ligase I activity and consistently recognized a single polypeptide of 125 kDa when cells from an established bovine kidney cell line (MDBK) were lysed rapidly by a variety of procedures and subjected to immunoblotting analysis. After biosynthetic labeling of MDBK cells with [35S]methionine, immunoprecipitation experiments revealed a polypeptide of 125 kDa that did not appear when purified calf thymus DNA ligase I was used in competition. A 125-kDa polypeptide was adenylated when immunoprecipitated protein from MDBK cells was incubated with [alpha-32P]ATP. Thus, the apparent molecular mass of the initial translation product is identical or nearly so to that of the purified enzyme. The half-life of the protein is 7 h as determined by pulse-chase experiments in asynchronous MDBK cells. Immunocytochemistry and indirect immunofluorescence experiments showed that DNA ligase I is localized to cell nuclei.     

Expression of DNA ligases I and II during oogenesis and early development of Xenopus laevis.

We have analyzed the expression of DNA ligase I protein during oogenesis and early development of Xenopus laevis. The protein is already present in stage I oocytes and then accumulates throughout oogenesis to reach a steady state level by stage VI. It remains at this level at least until tadpole stage. In stage VI oocytes DNA ligase I protein is almost exclusively localized in the germinal vesicle. We have partially purified a DNA ligase II activity from stage VI oocytes, unfertilized eggs, and stage 8 embryos. An 80-kDa polypeptide can be specifically adenylated in all three purified extracts. It is not recognized by antibodies directed against DNA ligase I and is active on oligo(dT)-poly(rA) substrate. It could therefore represent DNA ligase II protein. The presence of both DNA ligases I and II in oocytes and embryos is inconsistent with the DNA ligase model that had been previously proposed for amphibia.     

Mismatch and blunt to protruding-end joining by DNA ligases.

A nuclear DNA ligase activity from immature chicken erythrocytes, and to a lesser extent T4-induced DNA ligase, can join cohesive-ends (3 and 5-nucleotides long) having one of the mismatches, A/A, T/T, C/C, G/G, at the middle position. The rate of ligation depends on the length and stability of the mispaired intermediate (G/G, T/T greater than A/A, C/C). When the non-complementary overhanging-ends are short (i.e. 1-nucleotide) both ligases catalyze the joining of the single-stranded protruding-end with a blunt-end. This reaction occurs at low but significant rates compared to blunt-end ligation. The chicken ligase has lower flush-end joining activity than T4 DNA ligase, but it is more permissive since it joins C/C or A/A mismatched-ends, whereas the prokaryotic ligase does not. Possible biological implications of the reactions are discussed. We have also found that BstEII easily cleaves at sites harboring a C/C or a G/G mismatch at the center of its recognition sequence, whereas AvaII (T/T or A/A), HinfI (G/G) and DdeI (G/G) do not.     

Purification and properties of two DNA ligases from human placenta

Two DNA ligase activities have been separated, purified, and characterized. The resolution of the two enzymes from crude extracts was initially achieved through Polymin P precipitation. The ligation activity precipitating with the nucleic acids on treatment with Polymin P is designated as DNA ligase I, and an activity remaining in the supernatant fraction, as DNA ligase II. DNA ligase I and II are ATP and Mg2+-dependent enzymes with pH optima of 7.8 and 8.0 and isoelectric points of 6.9 and 7.6, respectively. The purified I and II DNA ligase activities have molecular weights of 83,000 and 89,000, respectively. Both activities are inhibited by dATP and inorganic pyrophosphate. However, in the presence of optimum rATP levels, dATP stimulates DNA ligase II activity, whereas DNA ligase I is inhibited under the same conditions. Both activities are DNA specific and ligation follows reaction steps similar to those described for the Escherichia coli DNA ligase.     

Differential inhibition of class I and class II 5-enolpyruvylshikimate-3-phosphate synthases by tetrahedral reaction intermediate analogues

The shikimate pathway enzyme 5-enolpyruvylshikimate-3-phosphate synthase (EPSP synthase or EPSPS) is best known as the target of the herbicide glyphosate. EPSPS is also considered an attractive target for the development of novel antibiotics since the pathogenicity of many microorganisms depends on the functionality of the shikimate pathway. Here, we have investigated the inhibitory potency of stable fluorinated or phosphonate-based analogues of the tetrahedral reaction intermediate (TI) in a parallel study utilizing class I (glyphosate-sensitive) and class II (glyphosate-tolerant) EPSPS. The (R)-difluoromethyl and (R)-phosphonate analogues of the TI are the most potent inhibitors of EPSPS described to date. However, we found that class II EPSPS are up to 400 times less sensitive to inhibition by these TI analogues. X-ray crystallographic data revealed that the conformational changes of active site residues observed upon inhibitor binding to the representative class I EPSPS from Escherichia coli do not occur in the prototypical class II enzyme from Agrobacterium sp. strain CP4. It appears that because the active sites of class II EPSPS do not possess the flexibility to accommodate these TI analogues, the analogues themselves undergo conformational changes, resulting in less favorable inhibitory properties. Since pathogenic microorganisms such as Staphylococcus aureus utilize class II EPSPS, we conclude that the rational design of novel EPSPS inhibitors with potential as broad-spectrum antibiotics should be based on the active site structures of class II EPSP synthases.     

DNA ligases in the repair and replication of DNA

DNA ligases are critical enzymes of DNA metabolism. The reaction they catalyse (the joining of nicked DNA) is required in DNA replication and in DNA repair pathways that require the re-synthesis of DNA.Most organisms express DNA ligases powered by ATP, but eubacteria appear to be unique in having ligases driven by NAD(+). Interestingly, despite protein sequence and biochemical differences between the two classes of ligase, the structure of the adenylation domain is remarkably similar. Higher organisms express a variety of different ligases, which appear to be targetted to specific functions. DNA ligase I is required for Okazaki fragment joining and some repair pathways; DNA ligase II appears to be a degradation product of ligase III; DNA ligase III has several isoforms, which are involved in repair and recombination and DNA ligase IV is necessary for V(D)J recombination and non-homologous end-joining. Sequence and structural analysis of DNA ligases has shown that these enzymes are built around a common catalytic core, which is likely to be similar in three-dimensional structure to that of T7-bacteriophage ligase. The differences between the various ligases are likely to be mediated by regions outside of this common core, the structures of which are not known. Therefore, the determination of these structures, along with the structures of ligases bound to substrate DNAs and partner proteins ought to be seen as a priority.     

Functional redundancy between DNA ligases I and III in DNA replication in vertebrate cells

In eukaryotes, the three families of ATP-dependent DNA ligases are associated with specific functions in DNA metabolism. DNA ligase I (LigI) catalyzes Okazaki-fragment ligation at the replication fork and nucleotide excision repair (NER). DNA ligase IV (LigIV) mediates repair of DNA double strand breaks (DSB) via the canonical non-homologous end-joining (NHEJ) pathway. The evolutionary younger DNA ligase III (LigIII) is restricted to higher eukaryotes and has been associated with base excision (BER) and single strand break repair (SSBR). Here, using conditional knockout strategies for LIG3 and concomitant inactivation of the LIG1 and LIG4 genes, we show that in DT40 cells LigIII efficiently supports semi-conservative DNA replication. Our observations demonstrate a high functional versatility for the evolutionary new LigIII in DNA replication and mitochondrial metabolism, and suggest the presence of an alternative pathway for Okazaki fragment ligation.     

Double-stranded DNA cleavage/religation reaction of eukaryotic topoisomerase II: evidence for a nicked DNA intermediate

The DNA cleavage reaction of eukaryotic topoisomerase II produces nicked DNA along with linear nucleic acid products. Therefore, relationships between the enzyme's DNA nicking and double-stranded cleavage reactions were determined. This was accomplished by altering the pH at which assays were performed. At pH 5.0 Drosophila melanogaster topoisomerase II generated predominantly (greater than 90%) single-stranded breaks in duplex DNA. With increasing pH, less single-stranded and more double-stranded cleavage was observed, regardless of the buffer or the divalent cation employed. As has been shown for double-stranded DNA cleavage, topoisomerase II was covalently bound to nicked DNA products, and enzyme-mediated single-stranded cleavage was salt reversible. Moreover, sites of single-stranded DNA breaks were identical with those mapped for double-stranded breaks. To further characterize the enzyme's cleavage mechanism, electron microscopy studies were performed. These experiments revealed that separate polypeptide chains were complexed with both ends of linear DNA molecules generated during cleavage reactions. Finally, by use of a novel religation assay [Osheroff, N., & Zechiedrich, E. L. (1987) Biochemistry 26, 4303-4309], it was shown that nicked DNA is an obligatory kinetic intermediate in the topoisomerase II mediated reunion of double-stranded breaks. Under the conditions employed, the apparent first-order rate constant for the religation of the first break was approximately 6-fold faster than that for the religation of the second break. The above results indicate that topoisomerase II carries out double-stranded DNA cleavage/religation by making two sequential single-stranded breaks in the nucleic acid backbone, each of which is mediated by a separate subunit of the homodimeric enzyme.     

Characterization of the reaction product of the oriT nicking reaction catalyzed by Escherichia coli DNA helicase I.

DNA helicase I, encoded on the Escherichia coli F plasmid, catalyzes a site- and strand-specific nicking reaction within the F plasmid origin of transfer (oriT) to initiate conjugative DNA strand transfer. The product of the nicking reaction contains a single phosphodiester bond interruption as determined by single-nucleotide resolution mapping of both sides of the nick site. This analysis has demonstrated that the nick is located at precisely the same site previously shown to be nicked in vivo (T. L. Thompson, M. B. Centola, and R. C. Deonier, J. Mol. Biol. 207:505-512, 1989). In addition, studies with two oriT point mutants have confirmed the specificity of the in vitro reaction. Characterization of the nicked DNA product has revealed a modified 5' end and a 3' OH available for extension by E. coli DNA polymerase I. Precipitation of nicked DNA with cold KCl in the presence of sodium dodecyl sulfate suggests the existence of protein covalently attached to the nicked DNA molecule. The covalent nature of this interaction has been directly demonstrated by transfer of radiolabeled phosphate from DNA to protein. On the basis of these results, we propose that helicase I becomes covalently bound to the 5' end of the nicked DNA strand as part of the reaction mechanism for phosphodiester bond cleavage. A model is presented to suggest how helicase I could nick the F plasmid at oriT and subsequently unwind the duplex DNA to provide single-stranded DNA for strand transfer during bacterial conjugation.     

Binding, hydration, and decarboxylation of the reaction intermediate glutaconyl-coenzyme A by human glutaryl-CoA dehydrogenase.

Glutaconyl-coenzyme A (CoA) is the presumed enzyme-bound intermediate in the oxidative decarboxylation of glutaryl-CoA that is catalyzed by glutaryl-CoA dehydrogenase. We demonstrated glutaconyl-CoA bound to glutaryl-CoA dehydrogenase after anaerobic reduction of the dehydrogenase with glutaryl-CoA. Glutaryl-CoA dehydrogenase also has intrinsic enoyl-CoA hydratase activity, a property of other members of the acyl-CoA dehydrogenase family. The enzyme rapidly hydrates glutaconyl-CoA at pH 7.6 with a k(cat) of 2.7 s(-1). The k(cat) in the overall oxidation-decarboxylation reaction at pH 7.6 is about 9 s(-1). The binding of glutaconyl-CoA was quantitatively assessed from the K(m) in the hydratase reaction, 3 microM, and the K(i), 1.0 microM, as a competitive inhibitor of the dehydrogenase. These values compare with K(m) and K(i) of 4.0 and 12.9 microM, respectively, for crotonyl-CoA. Glu370 is the general base catalyst in the dehydrogenase that abstracts an alpha-proton of the substrate to initiate the catalytic pathway. The mutant dehydrogenase, Glu370Gln, is inactive in the dehydrogenation and the hydratase reactions. However, this mutant dehydrogenase decarboxylates glutaconyl-CoA to crotonyl-CoA without oxidation-reduction reactions of the dehydrogenase flavin. Addition of glutaconyl-CoA to this mutant dehydrogenase results in a rapid, transient increase in long-wavelength absorbance (lambda(max) approximately 725 nm), and crotonyl-CoA is found as the sole product. We propose that this 725 nm-absorbing species is the delocalized crotonyl-CoA anion that follows decarboxylation and that the decay is the result of slow protonation of the anion in the absence of the general acid catalyst, Glu370(H(+)). In the absence of detectable oxidation-reduction, the data indicate that oxidation-reduction of the dehydrogenase flavin is not essential for decarboxylation of glutaconyl-CoA.     

Optimised ligation of oligonucleotides by thermal ligases: comparison of Thermus scotoductus and Rhodothermus marinus DNA ligases to other thermophilic ligases

We describe the characterisation of four thermostable NAD⁺-dependent DNA ligases, from Thermus thermophilus (Tth), Thermus scotoductus (Ts), Rhodothermus marinus (Rm) and Thermus aquaticus (Taq), by an assay which measures ligation rate and mismatch discrimination. Complete libraries of octa-, nona- and decanucleotides were used as substrates. The assay comprised the polymerisation of oligonucleotides initiated from a 17 base ‘primer’, using M13mp18 ssDNA as template. Polymers of ligation products were analysed by polyacrylamide gel electrophoresis. Under optimum conditions, the enzymes produced polymers ranging from 8 to 16 additions; there was variation between enzymes and the length of the oligonucleotides had a strong effect. The optimal total oligonucleotide concentration for each library was ~4 nmol. We compared the rates of ligation between the four ligases using an octanucleotide library as substrate. By this criterion, the Ts and Rm ligases are far more active compared to the more commonly available thermostable ligases.     

Inhibition of Leishmania donovani promastigote DNA topoisomerase I and human monocyte DNA topoisomerases I and II by antimonial drugs and classical antitopoisomerase agents

We have compared the inhibitor sensitivities of DNA topoisomerase I (TOPI) from Leishmania donovani promastigotes and TOPs I and II of human monocytes using pentavalent and trivalent antimonials (SbV, SbIII) and classical TOP inhibitors. Bis-benzimidazoles (Hoechst-33258 and -33342) were potent inhibitors of both parasite and human TOPI, but Hoechst-33342 was markedly less cytotoxic to promastigotes than to monocytes in vitro. Leishmania donovani was also considerably less sensitive than monocytes to camptothecin, both at enzyme and cellular levels. Sodium stibogluconate (SSG) was the only antimonial to inhibit TOPI, exhibiting a significant (P < 0.05) 3-fold greater potency against the L. donovani enzyme but showed low cytotoxicities against intact promastigotes. The SbV meglumine antimoniate failed to inhibit TOPI and showed negligible cytotoxicities, whereas SbIII drugs were lethal to parasites and monocytes yet poor inhibitors of TOPI. Monocyte TOPII was inhibited by bis-benzimidazoles and insensitive to antimonials and camptothecin. The disparity between the high leishmanicidal activity and low anti-TOPI potency of SbIII indicates that in vivo targeting of L. donovani TOPI by the reductive pathway of antimonial activation is improbable. Nevertheless, the potent direct inhibition of TOPI by SSG and the differential interactions of camptothecin with L. donovani and human TOPI support the possibility of developing parasite-specific derivatives.     

An intermediate in the phage lambda site-specific recombination reaction is revealed by phosphorothioate substitution in DNA.

It has been proposed that phage lambda site-specific recombination proceeds via two independent strand exchanges: the first exchange forming a Holliday-structure which is then converted into complete recombinant products by the second strand exchange. If this hypothesis is correct, one should be able to trap the putative Holliday intermediate by preventing the second strand exchange. In this paper, we show that substitution of phosphorothioate for phosphate in one strand of a recombination site is an effective way to block recombination while permitting the accumulation of a novel structure. This effect is seen only when phosphorothioate is positioned at a point of potential cleavage by Int recombinase, demonstrating that the inhibition of strand exchange is highly specific. Analysis of the novel structure that accumulates in these reactions proves that it contains a Holliday joint. Holliday-structures can also be detected in unblocked recombinations but are present at very low levels. The characteristics of Holliday-structure formation that we describe substantiate the proposed recombination pathway.     

WRN protects against topo I but not topo II inhibitors by preventing DNA break formation

The Werner syndrome helicase/3′-exonuclease (WRN) is a major component of the DNA repair and replication machinery. To analyze whether WRN is involved in the repair of topoisomerase-induced DNA damage we utilized U2-OS cells, in which WRN is stably down-regulated (wrn-kd), and the corresponding wild-type cells (wrn-wt). We show that cells not expressing WRN are hypersensitive to the toxic effect of the topoisomerase I inhibitor topotecan, but not to the topoisomerase II inhibitor etoposide. This was shown by mass survival assays, colony formation and induction of apoptosis. Upon topotecan treatment WRN deficient cells showed enhanced DNA replication inhibition and S-phase arrest, whereas after treatment with etoposide they showed the same cell cycle response as the wild-type. A considerable difference between WRN and wild-type cells was observed for DNA single- and double-strand break formation in response to topotecan. Topotecan induced DNA single-strand breaks 6 h after treatment. In both wrn-wt and wrn-kd cells these breaks were repaired at similar kinetics. However, in wrn-kd but not wrn-wt cells they were converted into DNA double-strand breaks (DSBs) at high frequency, as shown by neutral comet assay and phosphorylation of H2AX. Our data provide evidence that WRN is involved in the repair of topoisomerase I, but not topoisomerase II-induced DNA damage, most likely via preventing the conversion of DNA single-strand breaks into DSBs during the resolution of stalled replication forks at topo I–DNA complexes. We suggest that the WRN status of tumor cells impacts anticancer therapy with topoisomerase I, but not topoisomerase II inhibitors.