Gene 6 Exonuclease of Bacteriophage T7 PURIFICATION AND PROPERTIES OF THE EXZYME*

An exonuclease activity was detectable in crude extracts from Escherichia coli 1200 cells (Endo I-, su-) infected with T7 phage bearing amber mutations in gene 3 (T7 endonuclease I) and gene 5 (T7 DNA polymerase). The activity was not detectable if the infecting phage also contained an amber mutation in gene 6. That gene 6 is the structural gene for this enzyme was shown by the fact that phage bearing a temperature-sensitive mutation in gene 6 induced an exonuclease activity which was more heat labile than the wild type enzyme. The enzyme has been purified 500-fold, and the purified preparations are essentially free of ribonuclease, endonuclease, DNA polymerase, and exonuclease I activity although they do contain measurable 3′-phosphatase activity, likely the reflection of residual E. coli exonuclease III. The enzyme has an absolute requirement for divalent cations and a sulfhydryl reagent and is stimulated greatly by potassium ions. The enzyme shows a marked preference for duplex DNA and liberates 5′-mononucleotides as the sole acid-soluble product. The gene 6 exonuclease is involved in the degradation of cellular DNA to acid-soluble products after T7 phage infection.

The degradation of the host DNA following infection with bacteriophage T7 occurs in three stages (1). The DNA is first released from a membrane-like structure (1, 2), and this process requires the action of gene 1 (l), the structural gene for a T7specific RNA polymerase (3). The release from the membrane is thought to be catalyzed by a class 2A protein (2) which is in turn synthesized under the control of the T7 RNA polymerase.
The released DNA is then cleaved endonucleolytically to yield fragments with a molecular weight of about 1 x 106. This stage is catalyzed by T7-induced endonuclease I (4-Q, the product of gene 3 (7). These fragments of DNA are then converted directly * This work was supported by the Medical Research Council of Canada and the S. W. Stedman Foundation.
$ Scholar of the Medical Research Council of Canada.
to acid-soluble nucleotidex, a process which requires the action of gene 6 (1). It has been suggested that gene 6 codes for an exonuclease involved in the degradation of cellular DNA (1, 7). This report shows that indeed gene 6 is the structural gene for an exonuclease which liberates 5'-mononucleotides as the sole acid-soluble product. The purification of the enzyme and requirements of the reaction are described in this paper. In the accompanying paper we describe studies of the mechanism of action of the enzyme. medium (1) was used for preparing DNA of both bacterial and phage origin.
[32P]Orthophosphate (carrier free) was from Atomic Energy of Canada, Ottawa, Ontario.
Othm Materials-Streptomycin sulfate was purchased from Chas. Pfizer, Toronto, Canada, and dithiothreitol was obtained from Calbiochem, Los Angeles, California.
Crystalline bovine serum albumin was purchased from Pentex Incorporated, Kankakee, Illinois.
Other materials were those described in previous papers (1, 4).

Methods
Preparation of Phage Stocks-Stocks of amber mutants were grown on the permissive host E. coli BBW/l. An overnight culture was diluted 50-fold in LB medium and grown in a gyratory shaker at 30" until the cell concentration reached 5 X lo8 per ml. The appropriate T7 phage was then added at a multiplicity of infection of 0.1, and the culture was shaken vigorously until lysis occurred about l+ hours later.
Chloroform was added, and the culture was centrifuged at 16,000 x g for 10 min. Titers of 5 x lOlo to 2 x 10" per ml were obtained.
Reversion frequencies were never greater than 1 x lop5 for single T7 mutants, whereas multiple mutants had reversion frequencies of less than 1 x 10-s. Phage PM 2 was grown and purified by the method of Espejo and Canelo (10). Preparation of Nucleic Acids-Phage DNA's were prepared by phenol extraction (11, 12). E. coli DNA was labeled with [3H]thymidine in the following way. An overnight culture of E. coli B was diluted 50-fold into 500 ml of GCA medium containing 250 pg per ml of deoxyadenosine, 0.4 pg per ml of thymidine, and 0.5 MCi per ml of [3H]thymidine (23 Ci per mmole). The cells were grown for 7 hours at 37", harvested by centrifugation at 16,000 x g for 10 min and washed in 100 ml of GCA medium.  Littauer and Eisenberg (14). Residual contamination by DNA (0.1%) was removed by treatment with pancreatic DNase.
Growth of Phuge-infected Cells-The procedure is generally the same as already described (4). LBP medium (100 liters) was inoculated with 3 liters of an overnight culture of E. coli 1200. The cells were grown to a density of 1 x log per ml at 30" and infected with T7 am 29, 28 (genes 3 + 5) at a multiplicity of infection of 10. After 15 min cooling was begun with liquid nitrogen and 10 min later the cells were collected in a continuous flow centrifuge.
The yield of cells was 242 g. Small scale infections were done at 30" in 2-liter flasks. The percentage of cells infected was always greater than 98% as determined by viable cell titer.
However, it was found that with triple mutants it was necessary to use a multiplicity of 20 phage per cell in order to infect greater than 98yo of the cells.
Assay of Exonuclease-The assay measures the release of acidsoluble radioactivity from E. coli [3H]DNA in the presence of the exonuclease.
The reaction mixture (0.1 ml) contained 5 Imoles of Tris-chloride buffer, pH 8.1, 0.5 pmole of MgC12, 0.1 pmole of dithiothreitol, 2.0 pmoles of KCl, 10 nmoles of E. coli [aH]DNA (1000 to 2000 cpm per nmole), and 0.1 to 4.0 units of enzyme. After incubation for 15 mm at 37" the reaction was terminated by addition of 300 pg of bovine serum albumin and 0.01 ml of 100% trichloroacetic acid (w/v). After centrifugation, the supernatant was counted in Bray's solution in a liquid scintillation counter.
The blank values were less than 1% of the radioactivity.
One unit is the release of 1.0 nmole of acid-soluble nucleotide after 15 mm at 37". The activity was linearly related to enzyme concentration in the range of 0.1 to 4.0 units. Enzyme fractions were diluted in a solution containing 0.05 M Tris-chloride buffer, pH 7.5, 0.001 M dithiothreitol, and 100 pg per ml of crystalline bovine serum albumin.
A stock solution of dithiothreitol was prepared each day since it was found to deteriorate if stored frozen for prolonged periods of time.
Potassium chloride was omitted from assays of crude extracts as it was not required for activity in such fractions.
Different DNA preparations gave up to 4-fold variation in activities in this assay possibly because of variation in the chain length of the DNA.
Other Methods-Analytical band centrifugation of PM 2 DNA was performed in 2.5 M CsCl containing 0.1 N NaOH.
Greater than 90% of the DNA was in the supercoiled form with an uncorrected sedimentation coefficient of 62 S.

RESULTS
Identi$cation of T7 gene 6 as Structural Gene for Exonuclease E. coli 1200 was infected with various T7 phages and the exonuclease activity in crude extracts was determined (Table I). Following infection with T7 am 3 + 5, the exonuclease activity rose greater than 12-fold, whereas a rise of 5-fold was found when the cells were infected with T7 am 3 + 6 phage. No rise in activity was detected when T7 am 3 + 5 + 6 phage were used. We conclude that both gene 6 and gene 5 control nuclease activities. Since gene 5 is said to code for a phage-induced DNA polymerase (7), the nuclease activity seen after infection with T7 am 3 + 6 phage is probably a polymerase-associated exonuclease. The activity induced by wild type phage was greater than that in-Issue of January 10, 1972 C. Kerr  FIG. 1. Heat inactivation of gene 6 exonuclease activity in extracts from E. coli 1200 cells infected with am 3 + 5 or am 3 + 5 + 6 ts T7 phage. E. coli 1200 was infected with either am 3 + 5 (*---*) or am 3 + 5 + 6 ts (0-0 ) T7 phage and harvested after 15 min at 30'. Aliquots of crude extracts were incubated for varying times at 37" and chilled on ice. Exonuclease activity was determined by means of the standard assay. One hundred per cent activity (units per mg) for am 3 + 5 phage was 105, about 17 times more than uninfected cells, while 100% activity for am 3 + 5 + 6 ts phage was 20, about 4 times more than uninfected cells.
duced by T7 am 3 + 5 phage probably in part due to the presence of T7 endonuclease I, the product of gene 3 (7). Although this enzyme makes no acid-soluble nucleotides from native DNA (4, 307 FIG. 2. Appearance of gene 6 exonuclease activity after infection of E. coli 1200 with T7 am 3 + 5 phage. E. coli 1200 was grown in LBP at 30' to a concentration of 1 X 109 cells per ml. T7 am 3 + 5 phage were added at a multiplicity of infection of 10 and at appropriate times, samples were poured over frozen 0.15 M NaCl solution. Where indicated 100 fig per ml of chlorampheni-co1 (CA) were added prior to addition of t,he phage and the infection was terminated 15 min later. Crude extracts were prepared and exonuclease activity was measured by means of the standard assay. O-O, no CA; A---A, CA added. TABLE II PuriJication of l"r gene 6 exonuclease from am S + 5 T'7 phage-infected E. coli 1BOOa 5) it does introduce single and double strand breaks which might provide sites for the action of exonuclease(s).

Fraction
A 4-fold increase in exonuclease activity was detected following infection with T7 am 3 + 5 + 6 ts, a phage bearing a temperaturesensitive mutation in gene 6 ( Table I). The activity induced by this phage was more sensitive to thermal inactivation than that i' ~~~~e~t~~a~e~. ';t;;;:' ' induced by T7 am 3 + 5 phage (Fig. 1) Purification E. coli cells were infected with T7 am 3 + 5 phage, and exonuclease activity was determined at varying times after infection at 30" (Fig. 2). The activity was barely detectable after 5 min, was readily detectable after 10 min, and reached a maximum after 20 min. The maximum activity was increased 20-fold over that found in uninfected cells. The rise in activity could be prevented by addition of chloramphenicol (100 pg per ml) at the time of infection.
These kinetics of induction suggest that the gene 6 exonuclease belongs to the late class of proteins as defined by Studier and Maize1 (15) or the class 2B proteins as defined by Morrison and Malamy (2).
Puri$cation of Gene 6 Exonuclease from T7 am S + 6 Phageinfected E. coli 1200 A summary of the purification procedure is given in Table 11.
Since it was found that the exonuclease activity was readily inactivated by too vigorous sonication, the following procedure was devised for disrupting the cells. The cell suspension was divided into two equal volumes and sonication was done in 50-ml polypropylene tubes. The Q-inch diameter probe of a Branson model 140 D sonicator (Heat Systems Ultrasonics, Plainview, New York) was used, and sonication was carried out for 15-set intervals interspersed by cooling in an ice-salt bath. The output setting was set at 50% of maximum and the total sonication time was 60 sec. Small quantities of cells (less than 1 g) were disrupted in the same manner except the micro-tip was used. The sonically disrupted cells were centrifuged at 40,000 rpm in the IEC A-237 rotor for 30 min at 5". The optical density at 260 nm of the supernatant, was adjusted to 150 by addition of Buffer B. This fraction was the crude extract (70 ml).

Streptomycin Precipitation and Nucleic Acid Digestion
To 70 ml of crude extract were added 17.5 ml of a 5% solution of streptomycin sulfate. The mixture was incubated at 0" for 10 min and centrifuged at 16,000 x g for 10 min (In this and succeeding steps, all buffers contained this concentration of dithiothreitol.) A solution of 1.0 M MgClz (0.35 ml) was added to bring the magnesium concentration to 0.005 M and the mixture was incubated at 30". After 2 hours 18% of the ultraviolet-absorbing material at 260 nm had become acid-soluble as measured by the procedure of Richardson et al. (16). To the autolysate was added 0.525 mmole of EDTA and 17 pg per ml of pancreatic RNase. After 30 min at 30", 96% of the ultraviolet-absorbing material at 260 nm had become acid soluble.
The mixture was chilled on ice and centrifuged at 16,000 x g for 10 min. The supernatant (70 ml) was subjected to DEAE-cellulose chromatography.
The column was washed with 30 ml of equilibrating buffer and eluted with a 306ml linear gradient of 0.1 M to 0.4 M ammonium sulfate in equilibrating buffer. The flow rate was maintained at 1 ml per min, and fractions of 15 ml were collected.
The exonuclease activity was eluted between 0.18 and 0.22.~ ammonium sulfate (Fractions 6 to 8) with an average purification of about 5-fold.

Phosphocellulose Chromatography
A phosphocellulose (Whatman P-11) column (2 x 8 cm) was equilibrated with 0.02 M potassium phosphate buffer, pH 6.5, containing 10ea M EDTA (equilibrating buffer). The DEAEcellulose eluate was dialyzed against 2 liters of the same buffer for 1 hour. The buffer was changed and dialysis was continued for an additional hour.
The dialyzed DEAE-cellulose eluate (50 ml) was applied to the column under gravity, and the column was washed with 25 ml of equilibrating buffer. The column was eluted with a linear 250-ml gradient of 0.05 to 0.5 M ammonium sulfate in the same buffer at a flow rate of 2 ml per min. Fractions of 12.5 ml were collected, and the exonuclease was eluted at 0.13 to 0.15 M ammonium sulfate (Fractions 3 and 4). The recovery of exonuclease in these fractions was about 40% of the activity applied with a 5-fold purification.
The over-all recovery was 24y0 with a 500.fold purification.
Since the enzyme was extremely unstable at this point the fraction was made 40y0 with glycerol, and 1 mg per ml of crystalline bovine serum albumin was added. The enzyme was stored at -20" and has been stable for 6 months.

Comments on PuriJication Procedure
With the exception of the phosphocellulose eluate, all fractions were stable for short periods of time when stored at 0" (24 to 48 hours).
The crude extract and the streptomycin autolysate retained about 60% of the original activity after 3 weeks whereas the DEAE-cellulose eluate lost 70% of its original activity in t.his time.
The stability of crude extracts was considerably enhanced by the use of dithiothreitol.
The activity was much less stable in the presence of 2-mercaptoethanol.
All experiments described were performed with the phosphocellulose eluate.

Presence of Other Enzymatic Activities in Purified Enzyme
Ribonuclease-When 10 nmoles of E. coli [32P]RNA (19,000 cpm per nmole) were incubated with 10 units of exonuclease for 15 min under standard assay conditions, there was no detectable release of acid-soluble material.
If the incubation was prolonged for 5 hours, 0.66 nmole of acid-soluble nucleotide was released. However, the crystalline bovine serum albumin contained in the final enzyme preparation was found to contain an equivalent amount of ribonuclease when assayed by itself. Thus it is probable that the exonuclease contains no ribonuclease activity.
In any case the RNase detected in the final enzyme preparations constitutes less than 0.5% of its activity toward native DNA.
Endonuclease-When 10 nmoles of the closed circular duplex DNA of phage PM 2 were incubated with 5 units of gene 6 exonuclease for 30 min at 37" there was no conversion of the supercoiled form to the open circular form as analyzed by band centrifugation in alkaline cesium chloride.
(The reaction mixtures (0.05 ml) had the same composition as the standard assay except reduced glutathione (lo-* M) was substituted for dithiothreitol. The latter reagent introduced approximately twice as many single strand breaks into PM 2 DNA as reduced glutathione.)' Prolongation of the incubation time to 60 min resulted in a detectable conversion of the supercoiled form to the open circular form of the DNA indicating the introduction of single strand breaks. That this was in part due to the presence of a sulfhydryl reagent (glutathione) was shown by the fact that if the sulfhydryl reagent was omitted from the reaction there was no detectable breakage of the DNA by 5 units of exonuclease.
Thus although the final enzyme preparations may be contaminated by a sulfhydrylrequiring endonuclease, we estimate that the exonuclease introduced less than one endonucleolytic break per 10,000 phosphodiester bonds broken exonucleolytically.
Exonuclease I-When the gene 6 exonuclease was incubated with 10 nmoles of heat-denatured E. coli DNA the activity was 0.5% of that obtained with native DNA.
Thus the enzyme has a marked preference for native DNA.
We are unable to determine whether the residual activity toward denatured DNA reflects contamination with E. coli exonuclease I, an enzyme highly specific for single stranded DNA (18), or is an inherent property of the gene 6 exonuclease.
E. coli DNA Polymerase-l'en units of gene 6 exonuclease caused no detectable incorporation of [3H]dATP (20,000 cpm per nmole) into acid-insoluble polydeoxyadenylatethymidylate copolymer as measured in the standard assay for E. co& DNA polymerase (16). (This amount of exonuclease inhibited authentic DNA polymerase about 74%.) The ratio of exonucleolytic activity to DNA polymerase activity was greater than 1000 and we thus conclude that the gene 6 exonuclease preparations are not significantly contaminated by the 3'-or 5'exonuclease activities of E. coli DNA polymerase.
E. coli Exonuclease III-The presence of exonuclease III contamination was detected by measuring the 3'-phosphatase activity (13) of t'he preparations using 3'-phosphoryl-[32P]DNA (average chain length, 70 nucleotides as determined by sensitivity of the 32P to bacterial alkaline phosphatase).
One unit of gene 6 exonuclease liberated 0.012 nmole of Norit nonadsorbable 32P in 15 min indicating the presence of a 3'-phosphatase activity in the final preparations.
That this activity was not attributable to t,he gene 6 exonuclease was shown by the fact that no 3'-phosphatase activity was induced by the various T7 phages tested (Table  III).
Thus, after infection with T7 am 3 + 5 phage the exonuclease activity increased lo-fold relative to the 3'-phosphatase activity.
1 The introduction of single strand breaks into duplex DNA by reducing reagents has been observed by Bode (17).  This level was not appreciable since we show in the accompanying paper that the gene 6 exonuclease begins its attack at the 5'-terminus of DNA whereas exonuclease III starts its attack at the 3'-terminus (19).

Requirements of T7 Gene 6 Exonuclease
The requirements of the gene 6 exonuclease are shown in  were used as the divalent metal the activity was 5% and lo%, respectively, of that obtained with magnesium ions.
The enzyme had an absolute requirement for a sulfhydryl reagent (dithiothreitol) ( Table IV) ; no activity was detectable in its absence. Potassium ions stimulated the activity about a-fold and the optimal potassium ion concentration was 0.02 M. The enzyme was not stimulated by addition of equivalent concentration of NH&l or NaCl in place of KCl. The pH optimum for the enzyme is 8.1 (Tris-chloride buffer); the activity at pH 7.5 (Tris-chloride buffer) was 60% of that obtained at pH 8. bearing an amber mutation in gene 6 and that the activity inand MnClz were added, the activity was 45'% of that obtained duced by a phage bearing a temperature-sensitive mutation in with the complete system and when KC1 was omitted but MnCl, gene 6 was more heat labile than the wild type enzyme. Since and dithiothreitol were added the activity was 75% of that ob-gene 6 mutants fail to degrade host DNA to acid-soluble fragtained with the complete system. ments, it seems likely that the gene 6 exonuclease is involved in this function in viva. The mechanism of action of the enzyme in Nature of Acid-soluble Product this process will be discussed in the accompanying paper.
When double stranded 32P-labeled E. coli DNA was digested with the enzyme, 43% of the radioactivity became acid-soluble. Acknowledgments-We are indebted to Mr. I. Berzins for Greater than 95% of this label was rendered nonadsorbable to growing phage-infected cells and to Mr. D. Kells for performing Norit by incubation with bacterial alkaline phosphatase or venom the analytical ultracentrifugation.
We thank Miss Marica 5'-nucleotidase (Table VI). The acid-soluble product was also Michael for typing the manuscript.
analyzed chromatographically on DEAE-cellulose paper in 7 M urea (4) and found to chromatograph as mononucleotides. Furthermore, no acid-soluble oligonucleotides could be detected chromatographically when lo'%, 21%, 29%, or 38% of the input radioactivity had been rendered acid soluble. Thus the acidsoluble product is predominantly if not entirely 5'-mononucleotides and we conclude that the enzyme acts exonucleolytically. In this paper we report the purification and partial characterization of a potent exonuclease from E. coli 1200 cells infected with T7 phage bearing amber mutations in genes 3 and 5. This phage was chosen so as to eliminate T7 endonuclease I (4-7) and a T7 DNA polymerase-associated exonuclease.
We have called the enzyme the T7 gene 6 exonuclease.
The purification was facilitated by the findings that the activity was enhanced and stabilized in the presence of dithiothreitol, that the enzyme was inactivated by too vigorous sonication, and that the activity was stimulated by potassium ions. The final preparation was essentially free of ribonuclease, endonuclease, and DNA polymerase activities and acted at least 200 times faster with native DNA as substrate than with denatured DNA. The preparations contained detectable 3'-phosphatase activity which is likely due to some residual contamination with E. coli exonuclease III.