Holliday junction resolving enzymes of archaeal viruses SIRV1 and SIRV2.

In the final stages of genetic recombination, Holliday junction resolving enzymes transform the four-way DNA intermediate into two duplex DNA molecules by introducing pairs of staggered nicks flanking the junction. This fundamental process is apparently common to cells from all three domains of life. Two cellular resolving enzymes from extremely thermophilic representatives of both kingdoms of the domain Archaea, the euryarchaeon Pyrococcus furiosus and the crenarchaeon Sulfolobus solfataricus, have been described recently. Here we report for the first time the isolation, purification and characterization of Holliday junction cleaving enzymes (Hjc) from two archaeal viruses. Both viruses, SIRV1 and SIRV2, infect Sulfolobus islandicus. Their Hjcs both consist of 121 amino acid residues (aa) differing only by 18 aa. Both proteins bind selectively to synthetic Holliday-structure analogues with an apparent dissociation constant of 25 nM. In the presence of Mg(2+) the enzymes produce identical cleavage patterns near the junction. While S. islandicus shows optimal growth at about 80 degrees C, the nucleolytic activities of recombinant SIRV2 Hjc was highest between 45 degrees C and 70 degrees C. Based on their specificity for four-way DNA structures the enzymes may play a general role in genetic recombination, DNA repair and the resolution of replicative intermediates.

Conformational flexibility in T4 endonuclease VII revealed by crystallography: implications for substrate binding and cleavage.

The structure of the N62D mutant of the junction-resolving endonuclease VII (EndoVII) from phage T4 has been refined at 1.3 A, and a second wild-type crystal form solved and refined at 2.8 A resolution. Comparison of the mutant with the wild-type protein structure in two different crystal environments reveals considerable conformational flexibility at the dimer level affecting the substrate-binding cleft, the dimerization interface and the orientation of the C-terminal domains. The opening of the DNA-binding cleft, the orientation of the C-terminal domains relative to the central dimerization domain as well as the relative positioning of helices in the dimerization interface appear to be sensitive to the crystal packing environment. The highly unexpected rearrangement within the extended hydrophobic interface does change the contact surface area but keeps the number of hydrophobic contacts about the same and will therefore not require significant energy input. The conformational flexibility most likely is of functional significance for the broad substrate specificity of EndoVII. Binding of sulphate ions in the mutant structure and their positions relative to the active-site metal ions and residues known to be essential for catalysis allows us to propose a possible catalytic mechanism. A comparison with the active-site geometries of other magnesium-dependent nucleases, among them the homing endonuclease I-PpoI and Serratia endonuclease, shows common features, suggesting related catalytic mechanisms.

Endonuclease VII has two DNA-binding sites each composed from one N- and one C-terminus provided by different subunits of the protein dimer.

Endonuclease VII (endo VII) is a Holliday structure-resolving enzyme of bacteriophage T4. Its activity depends on dimerization, DNA binding and hydrolysis of two phosphodiester bonds flanking the Holliday junction. We analysed the DNA-binding activity of truncated monomeric and covalently linked dimeric endo VII proteins. We show that both ends of endo VII are involved in DNA binding. In particular, the C-terminus of one subunit interacts with the N-terminus of the other subunit, constituting one DNA-binding site; the other two termini form the second binding site of the dimer. One binding site is sufficient to bind cruciform DNA. The concerted mechanism involving termini from different subunits ensures that only dimers bind to Holliday structures, thus providing two catalytic centres which introduce two cleavages in opposite strands. This is a precondition for precise resolution of Holliday structures.

Localization and characterization of the dimerization domain of holliday structure resolving endonuclease VII of phage T4.

Endonuclease VII (Endo VII) is a Holliday structure resolving enzyme of bacteriophage T4. Its nucleolytic activity depends on subactivities, which in order of execution are: (i) dimerization, (ii) binding to DNA, (iii) and cleavage of DNA. In an effort to assign these subfunctions to the primary sequence of the protein, a series of spontaneous point mutations deficient in DNA cleavage was isolated. Some of these mutations affected the dimerization of Endo VII. Compared with wild-type protein, which dimerizes completely in solution, more than 95% of one of the mutant proteins (W87R) remained in the monomeric state. Only the dimeric fraction of this protein bound to DNA. The dimerization domain of Endo VII was mapped by truncating the gene from both ends and analysing the dimerization ability of the purified peptides by crosslinking with glutaraldehyde. The dimerization domain was thus determined to reside between amino acid residues 55 and 105. Computer analyses predicted two alpha-helices (H2 and H3) in this section of the protein. As demonstrated by heterodimer formation, two copies of helix H3, but only one copy of helix H2, are required for dimerization. Helical wheel analyses revealed that both helices expose a hydrophobic face along their axes, suggesting that hydrophobic interaction between helices H3 mediate formation of Endo VII dimers, while helices H2 stabilize them.

Improved large-scale preparation of phage T4 endonuclease VII overexpressed in E. coli.

Using PCR, we cloned T4 gene 49, which encodes the endonuclease VII, and the inactive mutant gene 49 amE727 into vector pET-11a. In combination with Escherichia coli host strain BL21 (DE3), this system provided excellent repression of the expression of the highly toxic protein before induction with IPTG. After induction, the proteins were made in high quantities while remaining soluble. Dilution of the crude lysate at 1:10,000 continued to show a highly specific activity in the case of the wild-type enzyme. The protein was purified to homogeneity with a recovery of 33% using two chromatography steps. The yield was 20 times higher and the specific activity 500 times higher than that obtained by using the previously published protocol.

The rad21 gene product of Schizosaccharomyces pombe is a nuclear, cell cycle-regulated phosphoprotein.

The rad21 gene of Schizosaccharomyces pombe is involved in the repair of double-strand breaks in DNA and is essential for mitotic growth (Birkenbihl, R. P., and Subramani, S. (1992) Nucleic Acids Res. 20, 6605-6611). We show that the Rad21 protein migrates with an aberrantly slow mobility, has a thrombin cleavage site, and is multiply phosphorylated mainly at serine residues. The expression of the rad21 mRNA and the Rad21 protein is cell cycle-regulated, with the peak of mRNA and protein expression occurring near the G1 to S transition. Following translation of the protein, hypophosphorylated forms of the protein appear. However, the most phosphorylated form of Rad21 appears only later in the cell cycle (in S to G2). Analysis of the radiosensitive mutant rad21-45 revealed that the mutant protein is permanently hypophosphorylated. The Rad21 protein is nuclear during the cell cycle. The nuclear localization signal was identified in the C-terminal third of the protein. Upon repression of the Rad21 protein expressed from the repressible nmt1 promoter, the unphosphorylated and hypophosphorylated forms of Rad21 disappeared first. When the concentration of the most highly phosphorylated form of Rad21 sank under a critical level, the cells underwent aberrant mitoses. They exhibited loss of proper nuclear organization and abnormal septation.

Cloning and characterization of rad21 an essential gene of Schizosaccharomyces pombe involved in DNA double-strand-break repair.

Analysis of the Schizosaccharomyces pombe chromosomes by pulsed field gel electrophoresis showed that the fission yeast has a very efficient DNA double-strand-break (dsb) repair system, which properly restores the three chromosomes after they are degraded by gamma-irradiation. The radiation-sensitive mutant rad21-45 is deficient in this repair pathway but is capable of cell-cycle arrest in G2 following DNA damage. We cloned the rad21 gene by complementing the radiation sensitivity of the rad21-45 mutant. The plasmid-borne gene completely reestablished the DNA dsb repair pathway. The rad21 gene was localized to chromosome III by hybridization. The transcript is 2.5 kb long and expressed at a moderate level. The 1884-bp open reading frame encodes a 628 amino acid, very acidic peptide with a calculated molecular mass of 67,854 D. The rad21 gene shows no significant homology to other known nucleotide or peptide sequences. The inability of the mutant to perform efficient DNA repair is caused by a single base substitution, which changes wild-type isoleucine67 into threonine in the mutant. Deletion of the genomic rad21 gene showed that it is essential for mitotic growth of S.pombe.

Completion of the IS map in E. coli: IS186 positions on the E. coli K12 chromosome.

Insertion sites of the transposable element IS186 were physically mapped in the genome of E. coli K12 strain BHB2600. This strain maintains four IS186 copies of which three, assigned to 0.3, 14.1 and 51.8 map min., share common map positions with the three IS186 copies in strains W3110 and HB101. The fourth, unique IS copy in BHB2600 maps at 49.3 min. The IS186 data complete the BHB2600 map for all chromosomal sites of known K12-associated IS types.

Complete maps of IS1, IS2, IS3, IS4, IS5, IS30 and IS150 locations in Escherichia coli K12.

In this paper complete distribution maps are presented of the seven IS elements 1, 2, 3, 4, 5, 30 and 150. These maps were obtained during the construction of an almost complete restriction map of the Escherichia coli genome of K12 strain BHB2600. The positions of IS elements were correlated to this map. The distribution of integration sites of all IS types is nonrandom. Besides a large gap from 79 min to 96 min, there is a pronounced IS cluster at 6 min and another at 97 min, map locations that have low gene incidences on the classical map. One cluster coincides with a region of IS induced rearrangements. The IS distribution pattern was compared to patterns of strains W3110 and HB101.

Cosmid-derived map of E. coli strain BHB2600 in comparison to the map of strain W3110.

A physical map for the genome of E. coli K12 strain BHB2600 was constructed by use of 570 cloned DNA elements (CDEs) withdrawn from a cosmid library. Dot blot hybridisation was applied to establish contig interrelations with subsequent fine mapping achieved by analysis of EcoR1 restriction patterns on Southern blots. The derived map covers nearly 95% of the E. coli genome resulting in 12 minor gaps. It may be compared to the almost complete map for strain W3110 of Kohara et al. (1). Except for one tiny gap (lpp,36.5') remaining gaps in BHB2600 do not coincide with those in W3110 so that both maps complement each other establishing an essentially complete clone represented map. Besides numerous minute differences (site and fragment gains and losses) both strains harbour at differing positions extended rearrangements flanked by mutually inverted repetitive elements, in our case insertion elements (IS1 and IS5).