Isolation of exonuclease VIII: The enzyme associated with the sbcA indirect suppressor

recB and/or recC deficiency in Escherichia coli K-12 is indirectly suppressed by the presence of sbcAmutations. sbcA strains contain an increased level of an ATP-independent nuclease. Genetic and enzymatic tests indicate that this activity is not exonuclease III, exonuclease V (recB-recC nuclease), DNA polymerase 1, or lambda exonuclease. This new enzyme (exonuclease VII1) has been purified 750-fold and shows a striking preference for double-stranded DNA over heat-denatured DNA. It does not act endonucleolytically on closed circular, single-stranded DNA as exonuclease V does. It also lacks a 3'-phosphatase function. Analysis on sodium dodecyl sulfate-polyacrylamide gels indicates that exonuclease VIII is not present in unsuppressed (sbcA +)strains. It is thought that sbcA determines some type of control function; the structural gene for exonuclease VIII is denoted by recE. The observation that recBand/or recC mutants carried out residual levels of wild-type genetic recombination (1, 2) indicated that Escherichia coli might possess a pathway of genetic recombination that normally was repressed and did not involve the recB and recC gene products. Analysis of Rec+ revertants from recBand/or recC strains showed that indirect suppression did occur from alterations at two distinct loci. These strains remained genotypically recB and/or recC deficient but became recombination proficient. Barbour et al. showed that strains carrying sbcA mutations contained increased levels of an ATP-independent DNase (3). A second class of indirect suppressor, called sbcB, involved the loss of exonuclease I, an enzyme specific for single-stranded DNA (4). A more detailed analysis of strains carrying recBand sbcBmutations led to the demonstration of an alternate pathway for carrying out genetic recombination in E. coli (5). Because of difficulties encountered in mapping the sbcA locus, attention was focused on the sbcB type of indirect suppressor. However, the striking difference in the apparent mode of action of these two loci suggested the importance of a complete characterization of the ATP-independent nuclease that was apparently associated with sbcA alleles. Goldmark and Linn have shown that the recB-recC gene products form a single enzyme (exonuclease V), which, in the presence of ATP, exonucleolytically digests singleand double-stranded DNA (6). In addition, exonuclease V acts as an ATP-stimulated, single-stranded endonuclease and a DNA-dependent ATPase (6). If the ATP-independent nuclease found in sbcA strains Abbreviations: sbc, suppressor of recB and recC; NaDodSO4, sodium dodecyl sulfate. t Present address: Department of Biochemistry, University of Georgia, Athens, Ga. 30602. 3593 substitutes for exonuclease V, an analysis of its substrate specificity will provide an important clue into the function of the recB-recC gene products in vivo. In this communication we show by genetic and enzymatic tests that the nuclease present in sbcA strains is not exonuclease III, exonuclease V, DNA polymerase I, or lambda exonuclease. Electrophoretic analysis of purified fractions suggests that this enzyme, which we call exonuclease VIII, does not exist in unsuppressed strains. In its partially purified form the enzyme preferentially digests double-stranded DNA over heat-denatured, single-stranded DNA. It is suggested that sbcA is not the structural gene for exonuclease VIII, but determines some type of control function. The structural gene has been denoted by recE (7). MATERIALS AND METHODS Strains. JC7693 and JC7722 are isogenic derivatives of a Su parent (JC4693), which has been described elsewhere (8). Bolth strains are genotypically recB21, sbcB15, supD+, trp-, and 1eu(the trp and leu mutations are caused by amber codoums). In addition, JC7693 contains sbcA9, which was obtained by mutagenesis with ethyl methane sulfonate (3). JC77722 is the unmutagenized parent. The presence of recB21 and sbcB15 in both strains was tested genetically and enzymatically prior to exonuclease VIII purifications. BW9091 carries the xthAl mutation inactivating exonuclease III (9). Materials. Reagents were obtained from the following sources: carrier-free [32P]H3PO4, [methyl-3H]thymidine, "Ultra Pure" (NH4)2S04, and streptomycin sulfate, Schwarz/ Mann; salmon sperm DNA, Calbiochem Corp.; calf-thymus DNA, Sigma Chemical Co.; egg-white lysozyme and micrococcal nuclease, Worthington Biochemical Corp.; Brij-58, Atlas Chemical Industries; Munktell 410 cellulose, Bio-Rad Corp.; and DE-52, Reeve Angel. Antiserum to E. coli DNA polymerase I was the gift of Dr. A. Kornberg. Antiserum to lambda exonuclease was obtained from Dr. John Little, and purified lambda exonuclease from Dr. A. D. Kaiser. E. coli B DNA was labeled with 32PO4 in M-70 supplemented medium (10) and extracted by the method of Marmur (11). 3H-labeled E. coli DNA was prepared by the method of Mahler (12). 3H-labeled M13 DNA was made by the procedure of Forsheit and Ray (13). 32P-labeled nicked DNA for exonuclease III phosphatase assays was prepared as described by Richardson and Kornberg (14). 32p_ or 3H-labeled DNA (150 gM) was heated for 10 min at 100° and quickly chilled to produce heat-denatured DNA. 3594 Biochemistry: Kushner et al. TABLE 1 . Purification of exonuclease VIII from JC7722 (sbcA +) and JC7693 (sbcA9) *