Genetic Dissection of the Biochemical Activities of Recbcd Enzyme

RecBCD enzyme of Escherichia coli is required for the major pathway of homologous recombination following conjugation. The enzyme has an ATP-dependent DNA unwinding activity, ATP-dependent single-stranded (ss) and double-stranded (ds) DNA exonuclease activities, and an activity that makes a ss DNA endonucleolytic cut near Chi sites. We have isolated and characterized ten mutations that reduced recombination proficiency and inactivated some, but not all, activities of RecBCD enzyme. One class of mutants had weak ds DNA exonuclease activity and lacked Chi-dependent DNA cleavage activity, a second class lacked only Chi-dependent DNA cleavage activity, and a third class retained all activities tested. The properties of these mutants indicate that the DNA unwinding and ss DNA exonuclease activities of the RecBCD enzyme are not sufficient for recombination. Furthermore, they suggest that the Chi-dependent DNA cleavage activity or another, as yet unidentified activity or both are required for recombination. The roles of the RecBCD enzymatic activities in recombination and exclusion of foreign DNA are discussed in light of the properties of these and other recBCD mutations.     

Activity of single-stranded DNA endonucleases in mung bean is associated with cell division

A single-strand-specific endonuclease from mung bean sprouts is widely usedin molecular biology. However, the biological role of this enzyme is unknown. We studied the spatial and temporal activity of single-stranded DNA endonucleases in mung bean seedling by following enzyme activity that linearizes supercoiled plasmid DNA, a characteristic of this type of enzyme. The formation of a linear molecule from supercoiled DNA was found to occur in two distinguishable steps. The first, which involves introducing a nick into the supercoiled DNA and relaxing it, is very rapid and complete within a few seconds. The second step of cleaving the opposite strand to generate a unit-length linear duplex DNA is a relatively slow process. Analysis of the DNA cleavage sites showed the nuclease preferentially cuts supercoiled DNA at an AT-rich region. Varying levels of nuclease activity could be detected in different tissues of the mung bean seedling. The highest activity was in the root tip and was correlated with histone H1 kinase activity. This implies a link between nuclease activity and cell division. Induction of cell division in mung bean hypocotyls with auxin promoted formation of root primordia and considerably increased the activity of single-stranded DNA endonucleases. The nuclease activity and histone H1 kinase activity were reduced in mung bean cuttings treated with hydroxyurea, but not in cuttings treated with oryzalin. The potential function of single-stranded DNA endonucleases is discussed.     

A mammalian cell variant in which 3-aminobenzamide does not potentiate the cytotoxicity of dimethyl sulphate

Variants of mouse leukaemia L1210 cells have been isolated in which cytotoxicity to dimethyl sulphate is not fully potentiated by ADP-ribosyl transferase inhibitor 3-aminobenzamide, as occurs in normal L1210 cells. These variants were selected after mutagenesis by growing the cells in dimethyl sulphate and 3-aminobenzamide. The characterisation of one of these variants is described. Variant 3 cells repair low doses of DNA damage in the presence of ADP-ribosyl transferase inhibitors. The Vmax of the ADP-ribosyl transferase enzyme in these cells is only increased 35% compared to normal wild-type L1210 cells. The basal DNA ligase I activity is increased 66% above wild-type whereas DNA ligase II activity appears to be unchanged. The most striking observation, however, is that the DNA ligase II activity is not increased after dimethyl sulphate treatment as occurs in wild-type L1210 cells. It seems that by increasing DNA ligase I levels these cells can survive DNA damage in the presence of 3-aminobenzamide. This variant (mutant) provides genetic evidence for our previously published hypothesis that (ADP-ribose)n biosynthesis is required for efficient DNA repair after DNA damage by monofunctional alkylating agents, because ADP-ribosyl transferase activity regulates DNA ligase activity. This variant is the first mammalian cell reported in which DNA ligase activity is altered, as far as we are aware. In yeast, a DNA ligase mutant has a cell division cycle (cdc) phenotype. Presumably, DNA ligase is essential for DNA synthesis, repair and recombination. The present variant provides further evidence that in mammalian cells, DNA ligase II activity is related to ADP-ribosyl transferase activity.     

The Escherichia coli primosome can translocate actively in either direction along a DNA strand

The primosome is a mobile multiprotein DNA replication-priming apparatus that requires seven Escherichia coli proteins (replication factor Y (protein n'), proteins n and n", and the products of the dnaB, dnaC, dnaT, and dnaG genes) for assembly at a specific site (termed a primosome assembly site) on single-stranded DNA binding protein-coated single-stranded DNA. Two of the protein components of the primosome have intrinsic DNA helicase activity. The DNA B protein acts in the 5'----3' direction, whereas factor Y acts in the 3'----5' direction. The primosome complex has DNA helicase activity when present at a replication fork in conjunction with the DNA polymerase III holoenzyme. In this report, evidence is presented that the multiprotein primosome per se can act as a DNA helicase in the absence of the DNA polymerase III holoenzyme. The primosome DNA helicase activity can be manifested in either direction along the DNA strand. The directionality of the primosome DNA helicase activity is modulated by the concentration and type of nucleoside triphosphate present in the reaction mixture. This DNA helicase activity requires all the preprimosomal proteins (the primosomal proteins minus the dnaG-encoded primase). Preprimosome complexes must assemble at a primosome assembly site in order to be loaded onto the single-stranded DNA and act subsequently as a DNA helicase. The 5'----3' primosome DNA helicase activity requires a 3' single-stranded tail on the fragment to be displaced, while the 3'----5' activity does not require a 5' single-stranded tail on the fragment to be displaced. Multienzyme preprimosomes moving in either direction are capable of associating with the primase to form complete primosomes that can synthesize RNA primers.     

Specific inhibition of human DNA ligase adenylation by a distamycin derivative possessing antitumor activity.

The antiviral distamycin A and its phenyl mustard derivative FCE24517 possessing antitumor activity were tested for their ability to inhibit macromolecular synthesis in three human and one murine cell line. While distamycin A was poorly active in these systems, FCE24517 inhibited DNA synthesis efficiently, RNA synthesis to a lower extent and had little effect on protein synthesis. These findings suggest that the in vivo activity of FCE24517 derives from the specific inhibition of DNA synthesis. When the two drugs were tested on several enzymes involved in human DNA metabolism a strikingly similar pattern of inhibition appeared, with distamycin A being the more potent. Both drugs showed: A), no inhibitory activity against thymidine kinase and DNA primase; B), low activity against DNA topoisomerases I and II and the 3'-5' exonuclease associated with the DNA polymerase epsilon; C), high activity against DNA polymerases alpha and epsilon, uracil-DNA glycosylase and the joining activity of the replicative DNA ligase; D), the highest inhibitory activity against the AMP-dependent DNA relaxing activity of DNA ligase. The strong in vitro inhibition of several DNA enzymatic activities, including DNA ligase, do not match with the in vivo activities of the two drugs. However a unique difference was observed: only FCE24517 inhibited the DNA-independent reaction of adenylation of human DNA ligase while the adenylation reaction of T4 and E. coli DNA ligase was unaffected by either drug. It is still unclear whether these properties are relevant for modulating the killing activity of FCE24517 against proliferating cells both in culture and in vivo. Nevertheless FCE24517 is the first known molecule capable of interacting directly and specifically with human DNA ligase.     

DNA-PKcs regulates a single-stranded DNA endonuclease activity of Artemis

Human nuclease Artemis belongs to the metallo-beta-lactamase protein family. It acquires double-stranded DNA endonuclease activity in the presence of DNA-PKcs. This double-stranded DNA endonuclease activity is critical for opening DNA hairpins in V(D)J recombination and is thought to be important for processing overhangs during the nonhomologous DNA end joining (NHEJ) process. Here we show that purified human Artemis exhibits single-stranded DNA endonuclease activity. This activity is proportional to the amount of highly purified Artemis from a gel filtration column. The activity is stimulated by DNA-PKcs and modulated by purified antibodies raised against Artemis. Moreover, the divalent cation-dependence and sequence-dependence of this single-stranded endonuclease activity is the same as the double-stranded DNA endonuclease activity of Artemis:DNA-PKcs. These findings further expand the range of DNA substrates upon which Artemis and Artemis:DNA-PKcs can act. The findings are discussed in the context of NHEJ.     

Isolation from barley embryo of endonuclease specific for apurinic sites in DNA

The endonuclease activity specific for apurinic sites in DNA was detected in barley embryos. The enzyme was partially purified. It reveals high activity on partially depurinated DNA but low or nil activity on intact and alkylated DNA. The method used for the detection of enzyme activity was based on the changes in the sedimentation velocity of substrate DNA in neutral sucrose gradients with 80 % formamide.     

The proofreading 3′→5′ exonuclease activity of DNA polymerases: a kinetic barrier to translesion DNA synthesis

The 3′→5′ exonuclease activity intrinsic to several DNA polymerases plays a primary role in genetic stability; it acts as a first line of defense in correcting DNA polymerase errors. A mismatched basepair at the primer terminus is the preferred substrate for the exonuclease activity over a correct basepair. The efficiency of the exonuclease as a proofreading activity for mispairs containing a DNA lesion varies, however, being dependent upon both the DNA polymerase/exonuclease and the type of DNA lesion. The exonuclease activities intrinsic to the T4 polymerase (family B) and DNA polymerase γ (family A) proofread DNA mispairs opposite endogenous DNA lesions, including alkylation, oxidation, and abasic adducts. However, the exonuclease of the Klenow polymerase cannot discriminate between correct and incorrect bases opposite alkylation and oxidative lesions. DNA damage alters the dynamics of the intramolecular partitioning of DNA substrates between the 3′→5′ exonuclease and polymerase activities. Enzymatic idling at lesions occurs when an exonuclease activity efficiently removes the same base that is preferentially incorporated by the DNA polymerase activity. Thus, the exonuclease activity can also act as a kinetic barrier to translesion synthesis (TLS) by preventing the stable incorporation of bases opposite DNA lesions. Understanding the downstream consequences of exonuclease activity at DNA lesions is necessary for elucidating the mechanisms of translesion synthesis and damage-induced cytotoxicity.     

The proofreading 3'-->5' exonuclease activity of DNA polymerases: a kinetic barrier to translesion DNA synthesis

The 3'-->5' exonuclease activity intrinsic to several DNA polymerases plays a primary role in genetic stability; it acts as a first line of defense in correcting DNA polymerase errors. A mismatched basepair at the primer terminus is the preferred substrate for the exonuclease activity over a correct basepair. The efficiency of the exonuclease as a proofreading activity for mispairs containing a DNA lesion varies, however, being dependent upon both the DNA polymerase/exonuclease and the type of DNA lesion. The exonuclease activities intrinsic to the T4 polymerase (family B) and DNA polymerase gamma (family A) proofread DNA mispairs opposite endogenous DNA lesions, including alkylation, oxidation, and abasic adducts. However, the exonuclease of the Klenow polymerase cannot discriminate between correct and incorrect bases opposite alkylation and oxidative lesions. DNA damage alters the dynamics of the intramolecular partitioning of DNA substrates between the 3'-->5' exonuclease and polymerase activities. Enzymatic idling at lesions occurs when an exonuclease activity efficiently removes the same base that is preferentially incorporated by the DNA polymerase activity. Thus, the exonuclease activity can also act as a kinetic barrier to translesion synthesis (TLS) by preventing the stable incorporation of bases opposite DNA lesions. Understanding the downstream consequences of exonuclease activity at DNA lesions is necessary for elucidating the mechanisms of translesion synthesis and damage-induced cytotoxicity.     

Identification of DNA polymerase delta in CV-1 cells: studies implicating both DNA polymerase delta and DNA polymerase alpha in DNA replication

DNA polymerases delta and alpha were purified from CV-1 cells, and their sensitivities to the inhibitors aphidicolin, (p-n-butylphenyl)deoxyguanosine triphosphate (BuPdGTP), and monoclonal antibodies directed against DNA polymerase alpha were determined. The effects of these inhibitors on DNA replication in permeabilized CV-1 cells were studied to investigate the potential roles of polymerases delta and alpha in DNA replication. Aphidicolin was shown to be a more potent inhibitor of DNA replication than of DNA polymerase alpha or delta activity. Inhibition of DNA replication by various concentrations of BuPdGTP was intermediate between inhibition of purified polymerase alpha or delta activity. Concentrations of BuPdGTP which totally abolished DNA polymerase alpha activity were much less effective in reducing DNA replication, as well as the activity of DNA polymerase delta. Monoclonal antibodies which specifically inhibited polymerase alpha activity reduced, but did not abolish, DNA replication in permeable cells. BuPdGTP, as well as anti-polymerase alpha antibodies, inhibited DNA replication in a nonlinear manner as a function of time. Depending upon the initial or final rates of inhibition of replication by BuPdGTP and anti-alpha antibodies, as little as 50%, or as much as 80%, of the replication activity can be attributed to polymerase alpha. The remaining replication activity (20-50%) is tentatively attributed to polymerase delta, because it was aphidicolin sensitive and resistant to both anti-polymerase alpha antibodies and low concentrations of BuPdGTP. A concentration of BuPdGTP which abolished polymerase alpha activity reduced, but did not abolish, both the synthesis and maturation of nascent DNA fragments.(ABSTRACT TRUNCATED AT 250 WORDS)     

Stimulation of mouse DNA primase-catalyzed oligoribonucleotide synthesis by mouse DNA helicase B.

Many prokaryotic and viral DNA helicases involved in DNA replication stimulate their cognate DNA primase activity. To assess the stimulation of DNA primase activity by mammalian DNA helicases, we analyzed the synthesis of oligoribonucleotides by mouse DNA polymerase alpha-primase complex on single-stranded circular M13 DNA in the presence of mouse DNA helicase B. DNA helicase B was purified by sequential chromatography through eight columns. When the purified DNA helicase B was applied to a Mono Q column, the stimulatory activity for DNA primase-catalyzed oligoribonucleotide synthesis and DNA helicase and DNA-dependent ATPase activities of DNA helicase B were co-eluted from the column. The synthesis of oligoribonucleotides 5-10 nt in length was markedly stimulated by DNA helicase B. The synthesis of longer species of oligoribonucleotides, which were synthesized at a low level in the absence of DNA helicase B, was inhibited by DNA helicase B. The stimulatory effect of DNA helicase B was marked at low template concentrations and little or no effect was observed at high concentrations. The mouse single-stranded DNA binding protein, replication protein A (RP-A), inhibited the primase activity of the DNA polymerase alpha-primase complex and DNA helicase B partially reversed the inhibition caused by RP-A.     

Aphidicolin inhibits DNA polymerizing activity but not nucleolytic activity of Escherichia coli DNA polymerase II.

We have purified the DNA polymerase II of Escherichia coli from the recombinant strain carrying the plasmid which encodes the polB gene. We confirmed that the purified protein, of molecular weight 90,000, possesses a 3'----5' exonuclease activity in addition to DNA polymerizing activity in a single polypeptide. Its DNA polymerizing activity was sensitive to the drug aphidicoline, which is a specific and direct inhibitor of the alpha-like DNA polymerases including eukaryotic replicative DNA polymerases. Aphidicolin had no detectable effect on the 3'----5' exonuclease activity. The inhibition by aphidicolin on the polymerizing activity of polymerase II was competitive with respect to dNTP and uncompetitive with respect to template DNA. This mode of action is the same as that on eukaryotic DNA polymerase alpha. The apparent Ki value calculated from Lineweaver-Burk plots was 55.6 microM.     

Human DNA polymerase lambda possesses terminal deoxyribonucleotidyl transferase activity and can elongate RNA primers: implications for novel functions

DNA polymerase lambda is a novel enzyme of the family X of DNA polymerases. The recent demonstration of an intrinsic 5'-deoxyribose-5'-phosphate lyase activity, a template/primer dependent polymerase activity, a distributive manner of DNA synthesis and sequence similarity to DNA polymerase beta suggested a novel beta-like enzyme. All these properties support a role of DNA polymerase lambda in base excision repair. On the other hand, the biochemical properties of the polymerisation activity of DNA polymerase lambda are still largely unknown. Here we give evidence that human DNA polymerase lambda has an intrinsic terminal deoxyribonucleotidyl transferase activity that preferentially adds pyrimidines onto 3'OH ends of DNA oligonucleotides. Furthermore, human DNA polymerase lambda efficiently elongates an RNA primer hybridized to a DNA template. These two novel properties of human DNA polymerase lambda might suggest additional roles for this enzyme in DNA replication and repair processes.     

cAMP regulation of cell differentiation in Dictyostelium discoideum and the role of the cAMP receptor

DNA polymerases and DNA ligases have been studied during development of the amphibian, axolotl. Three forms of DNA polymerase, I, II, and III, with sedimentation coefficients in sucrose of 9, 6, and 3.1 S, respectively, have been found in the axolotl egg. The activity of these three DNA polymerases is unchanged during early embryonic development. The activity of DNA polymerase III then increases significantly, beginning at the tailbud stage, while the activity of DNA polymerase II increases at the larval stage. DNA polymerase I does not show significant variations during this time. On the basis of their catalytic properties, it appears that DNA polymerases I and II are α-type DNA polymerases whereas DNA polymerase III is a β-type enzyme. Two different DNA ligases are found in the axolotl, one showing a sedimentation coefficient in sucrose of 8.2 S (heavy form) and the other, 6 S (light form). The 6 S enzyme is the major DNA ligase activity found in the egg before and after fertilization. Its activity then decreases during embryonic development. It can be observed again, as the only DNA ligase activity, in some adult tissues. The 8.2 S enzyme appears during the first division cycle of the fertilized egg, is present at all stages of embryonic development, and is absent from the adult tissues tested. Properties of the two DNA ligases at different stages of embryonic development have also been compared.     

DNMT1 is a component of a multiprotein DNA replication complex

DNA methylation is a major determinant of epigenetic inheritance and plays an important role in genome stability. The accurate propagation of DNA methylation patterns with cell division requires that methylation be closely coupled to DNA replication, however the precise molecular determinants of this interaction have not been defined. In the present study, we show that the predominant DNA methyltransferase species in somatic cells, DNMT1, is a component of a multiprotein DNA replication complex termed the DNA synthesome that fully supports semi-conservative DNA replication in a cell-free system. DNMT1 protein and activity were found to co-purify with the human DNA synthesome through a series of subcellular fractionation and chromatography steps, resulting in an enrichment of methyltransferase specific activity from two human cell lines. DNA methyltransferase activity co-eluted with in vitro replication activity and DNA polymerase alpha activity on sucrose density gradients suggesting that DNMT1 is a tightly bound, core component of the replication complex. The synthesome-associated pool of DNA methyltransferase exhibited both maintenance and de novo methyltransferase activity and the ratio of the two was similar to that observed in whole cell lysates and for recombinant DNMT1. These data indicate that interactions within the synthesome complex do not influence the intrinsic preference of DNMT1 for hemimethylated DNA, but suggest that newly replicated DNA may be subject to low level de novo methylation. The data indicate that DNA methylation is tightly coupled to replication through physical interaction of DNMT1 and core components of the replication machinery. The definition of the molecular interactions between DNMT1 and other proteins in the replication complex in normal and neoplastic cells will provide further insight into the regulation of DNA methylation and the mechanisms underlying the alteration of DNA methylation patterns during carcinogenesis.     

DNMT1 is a Component of a Multiprotein DNA Replication Complex

DNA methylation is a major determinant of epigenetic inheritance and plays an important role in genome stability. The accurate propagation of DNA methylation patterns with cell division requires that methylation be closely coupled to DNA replication, however the precise molecular determinants of this interaction have not been defined. In the present study, we show that the predominant DNA methyltransferase species in somatic cells, DNMT1, is a component of a multiprotein DNA replication complex termed the DNA synthesome that fully supports semi-conservative DNA replication in a cell-free system. DNMT1 protein and activity were found to co-purify with the human DNA synthesome through a series of subcellular fractionation and chromatography steps, resulting in an enrichment of methyltransferase specific activity from two human cell lines. DNA methyltransferase activity co-eluted with in vitro replication activity and DNA polymerase a activity on sucrose density gradients suggesting that DNMT1 is a tightly bound, core component of the replication complex. The synthesome-associated pool of DNA methyltransferase exhibited both maintenance and de novo methyltransferase activity and the ratio of the two was similar to that observed in whole cell lysates and for recombinant DNMT1. These data indicate that interactions within the synthesome complex do not influence the intrinsic preference of DNMT1 for hemimethylated DNA, but suggest that newly replicated DNA may be subject to low level de novo methylation. The data indicate that DNA methylation is tightly coupled to replication through physical interaction of DNMT1 and core components of the replication machinery. The definition of the molecular interactions between DNMT1 and other proteins in the replication complex in normal and neoplastic cells will provide further insight into the regulation of DNA methylation and the mechanisms underlying the alteration of DNA methylation patterns during carcinogenesis.     

A truncated form of DNA topoisomerase IIbeta associates with the mtDNA genome in mammalian mitochondria.

Despite the likely requirement for a DNA topoisomerase II activity during synthesis of mitochondrial DNA in mammals, this activity has been very difficult to identify convincingly. The only DNA topoisomerase II activity conclusively demonstrated to be mitochondrial in origin is that of a type II activity found associated with the mitochondrial, kinetoplast DNA network in trypanosomatid protozoa [Melendy, T., Sheline, C., and Ray, D.S. (1988) Cell 55, 1083-1088; Shapiro, T.A., Klein, V.A., and Englund, P.A. (1989) J. Biol. Chem.264, 4173-4178]. In the present study, we report the discovery of a type DNA topoisomerase II activity in bovine mitochondria. Identified among mtDNA replicative proteins recovered from complexes of mtDNA and protein, the DNA topoisomerase relaxes a negatively, supercoiled DNA template in vitro, in a reaction that requires Mg2+ and ATP. The relaxation activity is inhibited by etoposide and other inhibitors of eucaryotic type II enzymes. The DNA topoisomerase II copurifies with mitochondria and directly associates with mtDNA, as indicated by sensitivity of some mtDNA circles in the isolated complex of mtDNA and protein to cleavage by etoposide. The purified activity can be assigned to a approximately 150-kDa protein, which is recognized by a polyclonal antibody made against the trypanosomal mitochondrial topo II enzyme. Mass spectrometry performed on peptides prepared from the approximately 150-kDa protein demonstrate that this bovine mitochondrial activity is a truncated version of DNA topoisomerase IIbeta, one of two DNA topoisomerase II activities known to exist in mammalian nuclei.     

Age-dependent changes in the DNA-polymerase activity of nuclei and mitochondria of the regenerating rat liver

The DNA polymerase activity in nuclei and mitochondria of adult and old rat liver was determined. No age-dependent changes in the DNA polymerase activity have been found in the intact rat liver. On the contrary, after partial hepatectomy the DNA polymerase activity was higher in nuclei of the adult rat liver and in mitochondria of the old one. These peculiarities of the DNA polymerase activity with ageing may depend on changes in molecular properties of DNA polymerases, their synthesis and intracellular concentration.     

DNA polymerase I and DNA primase complex in yeast

Chromatographic analysis of poly(dT) replication activity in fresh yeast extracts showed that the activities required co-fractionate with the yeast DNA polymerase I. Since poly(dT) replication requires both a primase and a DNA polymerase, the results of the fractionation studies suggest that these two enzymes might exist as a complex in the yeast extract. Sucrose gradient analysis of concentrated purified yeast DNA polymerase I preparations demonstrates that the yeast DNA polymerase I does sediment as a complex with DNA primase activity. Two DNA polymerase I peptides estimated at 78,000 and 140,000 Da were found in the complex that were absent from the primase-free DNA polymerase fraction. Rabbit anti-yeast DNA polymerase I antibody inhibits DNA polymerase I but not DNA primase although rabbit antibodies are shown to remove DNA primase activity from solution by binding to the complex. Mouse monoclonal antibody to yeast DNA polymerase I binds to free yeast DNA polymerase I as well as the complex, but not to the free DNA primase activity. These results suggest that these two activities exist as a complex and reside on the different polypeptides. Replication of poly(dT) and single-stranded circular phage DNA by yeast DNA polymerase I and primase requires ATP and dNTPs. The size of the primer produced is 8 to 9 nucleotides in the presence of dNTPs and somewhat larger in the absence of dNTPs. Aphidicolin, an inhibitor of yeast DNA polymerase I, is not inhibitory to the yeast DNA primase activity. The primase activity is inhibited by adenosine 5'-(3-thio)tri-phosphate but not by alpha-amanitin. The association of yeast DNA polymerase I and yeast DNA primase can be demonstrated directly by isolation of the complex on a column containing yeast DNA polymerase I mouse monoclonal antibody covalently linked to Protein A-Sepharose. Both DNA polymerase I and DNA primase activities are retained by the column and can be eluted with 3.5 M MgCl2. Part of the primase activity can be dissociated from DNA polymerase on the column with 1 M MgCl2 and this free primase activity can be detected as poly(dT) replication activity in the presence of Escherichia coli polymerase I.