Purification, Modification during Purification, and Characterization of a Deoxyribonuclease from Pseudomonas aeruginosa*

SUMMARY The only major deoxyribonuclease forming acid-soluble fragments from radioactively labeled DNA was isolated from cell-free extracts of Pseudomonas aeruginosa. The enzyme was purified 325-fold, yielding two bands on polyacrylamide gel electrophoresis, the lesser of which was a distinct ribonuclease activity. Extensive modification of the deoxyribonuclease during purification to produce multiple species could be avoided by the use of phenyhnethylsulfonyl fluoride and the destruction of endogenous nucleic acids. Marked instability of the enzyme at all stages of purification was overcome by the use of 30% glycerol and Z-mercaptoethanol; the purified preparation had a pH optimum of 7.8. The enzyme closely resembled exonuclease III of Escherichia coli in several respects; the products were 5’-mononucleotides and inorganic phosphate; phosphate release was dependent on the presence of 3’-phosphoryl end groups in DNA; hydrolysis was initiated at the 3’-terminus; native DNA was degraded three to four times faster than heat-denatured DNA. Unlike exonuclease III the entire substrate could be hydrolyzed pro-viding

The enzyme was purified 325-fold, yielding two bands on polyacrylamide gel electrophoresis, the lesser of which was a distinct ribonuclease activity. Extensive modification of the deoxyribonuclease during purification to produce multiple species could be avoided by the use of phenyhnethylsulfonyl fluoride and the destruction of endogenous nucleic acids. Marked instability of the enzyme at all stages of purification was overcome by the use of 30% glycerol and Z-mercaptoethanol; the purified preparation had a pH optimum of 7.8. The enzyme closely resembled exonuclease III of Escherichia coli in several respects; the products were 5'-mononucleotides and inorganic phosphate; phosphate release was dependent on the presence of 3'-phosphoryl end groups in DNA; hydrolysis was initiated at the 3'-terminus; native DNA was degraded three to four times faster than heat-denatured DNA. Unlike exonuclease III the entire substrate could be hydrolyzed providing there were several readditions of enzyme during the incubation.
It was optimally stimulated by 2.5 mM magnesium and to a lesser extent by manganese. The P. aeruginosa exonuclease was inhibited by ionic strength above 0.05 M Tris because of a reduced binding affinity for the DNA substrate; both the exonuclease and phosphatase functions demonstrated almost identical inhibition by p-chloromercuribenzoate and ethylenediaminetetraacetate. The molecular weight was estimated to be 42,500 by gel filtration.
There has been no adequate description of deoxyribonuclease activity in the genus Pseudomonas and in particular for Pseudo-monas aeruginosa.
An extracellular enzyme was detected by Streitfield et al. (1) and partly described by Guschlbauer and Halleck (2). Except for the preliminary investigation by the latter two authors no attention has been paid to intracellular enzymes. The work of Guschlbauer and Halleck (2) suggested some curious properties of P. aeruginosa DNases including a requirement for both citrate and Mgz+ ions for maximal activity in a partly purified preparation.
The intracellular enzymes are of interest, in our opinion, because the low levels of total DNase activity present in the strain we investigated may facilitate the detection of other DNase functions present in small quantities but of special significance, e.g. restriction enzymes.
Holloway (3) has demonstrated an active restriction mechanism in P. aeruginosa. The organisms are also extremely sensitive to ionizing radiation (4) and thus repair and excision enzymes may be of interest.
Furthermore, P. aeruginosa is a member of an ubiquitous group of microorganisms which because of their widespread nature are of more environmental interest than the much more restricted but better characterized Escherichia coli. P. aeruginosa MAC 264 differs from E. coli in several aspects of DNA metabolism including the failure to incorporate exogenous thymidine, and preliminary experiments suggested differences in the DNases present.
Thus the present work was undertaken to define the basic DNase system in a typical strain of P. aeruginosa and provide the background for subsequent investigations of the problems referred to above. This report deals with the purification and partial characterization of the single quantitatively significant DNase present and with the severe modification problems which occurred during that purification. Issue of February 25, 1972 L. E. Bryan, W. E. Raxxell, and J. N. Campbell 1237 per mg) heated to 80" before use, and DNase (2000 units per mg) from Worthington Biochemical Corp. ; micrococcal nuclease (1.5 x lo3 units per mg) from Miles Laboratories; p-nitrophenyl thymidine 3'-phosphate from Ray10 Chemicals, Edmonton, Alberta; %labeled phosphoric acid and y-labeled ATP from New England Nuclear; polyethyleneimine cellulose thin layer plates from J. T. Baker Chemical Co.; all other reagents were obtained from Sigma and were reagent grade. dlicrococcus sodonensis nuclease (electrophoretically pure and free of 3'.nucleotidase activity) and polynucleotide kinase were the kind gifts, respectively, of Miss Cecily Mills, Department of Microbiology, and Dr. V. Petkau, Department of Biochemistry, University of Alberta.
All bacterial strains were obtained from t,he Department of Microbiology, University of A1bert.a.

Growth of P. aeruginosa
The organism was grown in IO-liter New Brunswick Micro-Ferm fermentors (air flow, 4 to 8 liters per min; stirring, 200 to 400 rpm) at 30". The medium was identical with that of von Tigerstrom and Razzell (5) except that 0.5% glucose replaced ethanol.
Synthetic medium, used only in growth studies, was that of Norris and Campbell (6) except that MgS04.7H20 was reduced to 500 mg per liter.

Enzyme Assays
Standard DNase Assu~--[~H]DS,~ was prepared with JY'. coli 5275 thy-as described by Weissbach and Korn (7) and the DNA isolated by the method of Marmur (8). Unlabeled DNA was prepared in an identical manner with an E. coli strain and a medium containing no thymine.
The assay was modified somewhat from that of Weissbach and Korn (7). Reaction mixtures were reduced to micro proportions.
Each contained in a volume of 0.1 ml, 50 pg per ml of t3H]DNA (2.8 x lo3 dpm per nmole), 0.05 M Tris-HCl, pH 7.8, 2.5 mM MgC12, 0.01 M 2-mercaptoethanol, and sufficient enzyme to produce 6 to 32 '% solubilization of the DNA substrate. Assays were terminated with 100 ~1 of cold 6% perchloric acid; carrier was bovine serum albumin (50 ~1 of 10 mg per ml). The supernatant obtained after centrifugation for 1 min in a Beckman Microfuge (200 ~1) was counted in 5 ml of Bray's scintillation fluid (9) containing 10 ~1 of 5 x KOH at 4" in a Nuclear-Chicago Mark I liquid scintillation counter (used for all scintillation counting).
One unit of activity is defined as the production of 1 nmole of acid-soluble nucleotide equivalent of DNA in 12 min.
RNase Assay-Each reaction mixture contained in a volume of 0.1 ml, 1 mg per ml of yeast RNA, 0.1 M Tris-HCl, pH 7.8, 5 mM EDTA, and 0.01 M 2-mercaptoethanol. The procedure was otherwise that described by Nestle and Roberts (10) except that centrifugation was as described for the DNase assay. One unit is the production of 1 absorbance unit at 260 nm of acidsoluble material in 1 min.
DNA Phosphatase Assay-The assay was performed by the method of Richardson and Kornberg (11) using 3'.phosphoryl [32P]DNA in a reaction mixture identical with that of the standard DNase assay. ,4 unit is defined as the production of 1 nmole of phosphate in 12 min. Assays of DNase activity with [32P]DNA required a correction for the phosphate released by the phosphatase function inherent in the P. aeruginosa DNase. This was done by subtracting the 32P measured by the phosphatase assay from the total 32P solubilized in the DNase assay in identical time periods.
Proteolysis Assays-Casein hydrolysis was measured by the method of Kunitz (12). In order to overcome endogenous changes in materials absorbing at 280 nm in crude extracts, preparations to be assayed were in some cases adjusted to pH 3 with 1 N HCl at 25" for 5 min and then assayed. A unit of activity is the production of one AZ80 unit of acid-soluble material in 20 min. Hydrolysis of benzoyl n-arginine ethyl ester was assayed by the method of Schwert and Takenada (13) with a Gilford recording spectrophotometer, model 2400. One unit is the formation of 1 pmole of benzoyl n-arginine in 1 min.
Other Assays-Assays were performed as previously described for $I. sodonensis nuclease (14), phosphomonoesterase activity (15) modified to contain 1 mM p-nitrophenyl phosphate and 2.5 mM MgC12, and phosphodiesterase I and II activity (16). In the latter case the assays were modified in some instances to meet the requirements of the P. aeruginosa exonuclease. Tris-HCl (0.05 M) at pH 7.5 or 9.3 was used to assay phosphodiesterase I and, in addition to standard assay conditions for phosphodiesterase II, Tris-HCl (0.05 M), pH 7.5, was used to assay for that enzymatic function.
DNA polymerase was assayed essentially as described by Richardson et al. (17) by means of [3H]TTP (specific activity, 1.5 x lo4 dpm per nmole) with a concentration of 0.2 mM for each of the four deoxynucleoside triphosphates.
Filters were washed (18)) dried, and counted in a toluene-based scintillation fluid (Omnifluor, New England Nuclear). Micrococcal nuclease was assayed exactly as the DNase assay except that the reaction mixture contained 0.01 M CaC12, 0.05 M glycine, pH 9.3, in place of MgCls and Tris-HCl.
The Pi was separated by paper chromatography of the stopped assay and the 32P (detected by scanning with a Nuclear-Chicago Actigraph III) spots cut out and counted in 10 ml of Bray's scintillation fluid (9). The alkaline phosphatase contained no detectable DNase activity.
Specific activity was 5.2 x lo4 dpm per nmole.
The final concentration of all DNA preparations was 250 iug per ml except for the 32P-labeled 5'-phosphoryl end groups in [311]DNA which contained 125 pg per ml.

Chromatography
Chromatography JTaterials-DEAE-cellulose was washed in a solution of 1 M NaCl and 0.1 M NaOH and subsequently equilibrated with Buffer A (0.03 M Tris-HCI, pH 7.5, 0.01 M 2-mercaptoethanol, and 1 mM EDTA) containing 30% glycerol (v/v) and 5 mlr phenylmethylsulfonyl fluoride. Phenylmethylsulfonyl fluoride was deleted in those cases where the effect of proteolytic modification on the elution profile of DEAE-cellulose was being examined.
It was also deleted in the second DEAE-cellulose column in the purification sequence.
Hydroxylapatite was prepared by washing the commercial product in 0.2 M potassium phosphate, pH 7.5, containing 0.01 1hl2-mercaptoethanol.
It was packed as a thin slurry at a flow rate of 4 ml per hour in a column (2.5 X 20 cm). The column was equilibrated with 30% glycerol in 0.01 M potassium phosphate and 0.01 M 2-mercaptoethanol. Sephadex G-200 and G-75 gels were hydrated as suggested by Pharmacia in Buffer A and deaerated by suction. Columns, specified in the figure legends, were packed in the respective buffer minus glycerol and after packing were re-equilibrated with 307' glycerol in the respective buffer. Dextran blue was used to determine the void volumes and NaCl the total volume of Sephadex columns.
Paper Chromatography-Paper chromatography for separation of nucleotides and phosphate was performed for 7 hours in a descending system (23). Spots were located by ultraviolet light or V radioactivity (Nuclear-Chicago Actigraph III). RF values were Pi,0.95,dC1IP,0.62,TMP,0.47,dGMP,0.37,and dAMP,0.19. Polyethyleneimine Cellulose Thin IJayer Chromatography-This was performed according to Randerath and Randerath (24) USing 1 s acetic acid and 0.3 in LiCl.
Spots were located by ultraviolet light (standards) and by radioactivity of 0.5.inch slices counted in Omnifluor. DI~Alkellulose-7 x Urea Chromatography-This was carried out by the method of Tomlinson and Tener (25) using DEAEcellulose columns (1 x 22 cm) equilibrated with 0.02 M Tris-HCl, pH 7.8, and 7 M urea (urea-buffer).
The eluting buffer gradient consisted of 500 ml each of urea buffer and 0.4 M NaCl in that biiffer; fractions were 4.5 ml collected at flow rate of 25 ml per hour. Radioactivity was determined on 200 ~1 of selected samples in 5 ml of Bray's scintillation fluid (9). The contents of tubes examined for acid-insoluble materials were dialyzed against 10 volumes of 0.02 M Tris-HCl, pH 7.8, three times for 4 hours and concentrated IO-fold by pervaporation. The change in absorbance at 260 nm produced by cold 3% perchloric acid (final concentration) was determined on 0.2.ml identical samples ex-cept that one had been incubated at 37" with 1 pg of pancreatic DNase for 30 min.

Use of Phenylmethylsuljonyl Fluoride
Phenylmethylsulfonyl fluoride was prepared as a 0.04 M or 0.01 M solution in 95yo ethanol and diluted 1 part to 19 parts of cell-free extract or the appropriate buffer to obtain a final concentration of 2 or 5 mM. Under these conditions there was no problem with solubility of the phenylmethylsulfonyl fluoride.

Gel Electrophoresis
This was performed by the method of Davis (26). Gels were stained with 1% iZmido schware for 1 hour. Electrophoresis performed for the recovery of DNase or RNase was carried out at 4" with all buffers previously cooled to that temperature.
No sample gel u-as used; instead, 0.2 ml of 40% sucrose was used as the first step in gel formation as described by Davis (26). After polymerization the sucrose was removed and samples containing 30% glycerol were applied to the tubes which had been positioned in the electrophoresis apparatus (containing buffers). At the termination of the run, gels were placed on glass (4") overlying measured paper and sliced into 15 equal sections.

Deozyribose and Ribose
These sugars were determined by the methods of Steele et al. (27) and Mejbaum (28), respectively. dlolecular TVeight Estimation Molecular weight was estimated by the method of A1ndrews (29) except that 30% glycerol was present in the equilibration and eluting buffers (0.05 M potassium phosphate, pH 7.4, 0.01 M 2-mercaptoethanol) and in the samples applied to the columns. Column specifications of the G-75 and G-200 columns, respectively, were 2.5 x 37.5 cm and 2.5 x 35 cm, void volumes, 63 ml, and total volumes, 184 and 166 ml. Fractions of 1 ml were collect,ed at a flow rate of 6 ml per hour. Sample (0.5.ml volumes) protein concentration leas 4 mg per ml except for the P. aeruginosa exonuclease which was 0.25 mg per ml. The latter was detected in fractions by the DNase assay and the standards by absorbance at 280 nm.

Preparation
9~" Cell-jree K&act-Cells were collected by continuous centrifugation at 2" and resuspended at 1 g wet weight per 5 ml of Buffer A. The cell suspension (300-ml volumes) was subjected to ultrasonic treatment (Bronwill Biosonik III, with a a-inch probe, maximal power) for 6 min. The temperature during this procedure did not exceed 12". The sonic extract was centrifuged at 4" at 100,000 x g for 90 min. The supernatant solution was collected.
Xtabilization during Purification-The DNase activity proved to be very labile after the cell-free extract stage. Further fractionation resulted in loss of 50 to 95% of the activity after another one or more steps.  Table I illustrates the effect on stability of various glycerol concentrations.
Thirty per cent was selected because of the similar stabilizing effect to 407, but with a less marked increment in viscosity.
An even more marked effect of glycerol, however, was to stabilize the enzyme while on the various chromatography columns in the purification sequence.
The recovery from the initial DEAE-cellulose column was less than 5% in the absence of 30% glycerol.
The autoactivation step (see below) in the absence of glycerol also led to a marked decline in stability both to storage at 4" and -20" and to further fractionation.

Enzyme
Mod$cation during Puri$cntion-It was observed during the early purification sequences that DNase activity existed in multiple forms. Such multiplicity was seen during either gel filtration or ion exchange chromatography. Two factors appeared responsible for these results; one was association with nucleic acid fragments and the other proteolytic attack by an endogenous protease.
Association with Nucleic Acid-The use of gel filtration demonstrated that the DNase activity of the cell-free extract eluted in at least two positions on Sephadex G-200. Fig. 1 demonstrates that the bulk of DNase activity eluted in the void volume of a Sephadex G-200 column and about 15y0 in the internal volume.
As shown, essentially all of the detectable DNase activity could be made to elute in the internal volume by incubating the cell-free extract for 90 to 150 min at 37" with 2.5 or 5 mM MgCl% and 20 pg per ml of pancreatic RNase (autoactivation). Phenylmethylsulfonyl fluoride was included in the incubated mixture in order to inhibit proteolysis (see "Proteolysis during Purification" below) ; thus, the observed shift in elution pattern is attributed to the destruction of associations between the enzyme and nucleic acid molecules.
The elution profile obtained with the incubated enzyme corresponded closely to that obtained for the purified enzyme.
Data obtained from autoactivation studies revealed several significant points.
The cations successful in increasing total DNase activity were those which activate the DNase (see cation requirements).
The incubation with 5 mM MgC12 resulted in an increase of acid-soluble deoxyribose and material absorbing at 260 nm; only those couditions producing the former resulted iu a DNase increase. The  The enzyme activity becomes less stable to storage, to chromatography, and also to proteolytic modification (see below). After destruction of the endogenous DNA by the autoactivation step, addition of calf thymus DNA at a concentration of 0.2 mg per ml produced a 20 to 40% increase in enzyme stability.
The association of the enzyme with nucleic acid had to be defined clearly in order to evaluate the second apparent cause of modification of enzyme behavior, that is, proteolytic digestion. Proteolysis during Purification-The association between enzyme and nucleic acid was not the only mechanism causing heterogeneity in purification procedures.
Ion exchange chromatography on DEAE-cellulose yielded multiple peaks of DNase activity, and the autoactivation procedure which eliminated variations on Sephadex only further complicated that effect. Even storage of the cell-free extract at 4" for several days led to greater heterogeneity in ion exchange, presumably by the same mechanism which functioned in a shorter time at the autoactivation temperature of 37". Fig. 2A shows the results of a gradient elution of enzyme and protein from a cell-free extract stored for 7 days at 4" prior to autoactivation as noted. Eight different peaks eluted from the DEAE-cellulose column, all but one small initial peak eluting at higher ionic strength than the unmodified enzyme in Fig. 2C. One of the possible causes of this distribution of enzyme was proteolytic alteration which had been suggested by the occasional observation of a peak of DNase eluted from the G-200 column even later than that shown in Fig. 1 (in spite of the use of an ionic strength of 0.1 or greater in eluting buffers to prevent retardation by electrostatic effects). Table III shows that low levels of endogenous proteolysis associated with cell-free extract could be demonstrated by the hydrolysis of casein and benzoyl 1-arginine ethyl ester. It was possible to inhibit the proteolytic activity by using the serine esterase inhibitor phenylmethylsulfonyl fluoride (30) at high concentrations (2 to 5 mM) and with incubation at 37". Inhibition of endogenous proteolytic activity was negligible at 4". The effects of including phenylmethylsulfonyl fluoride in the cell-free extract during the autoactivation with MgClz and pancreatic RNase as well as in all buffers of DEAE-cellulose chromatography are shown in Fig. 2B. There are five fewer peaks eluted after incubation with phenylmethylsulfonyl fluoride as opposed to incubation without phenylmethylsulfonyl fluoride of the same preparation aged for 7 days at 4". When phenylmethylsulfonyl fluoride was added immediately after thawing of cell-free extract stored overnight at -20" only a single DNase peak eluted (Fig. 2C). It thus seems that two enzyme peaks in L. E. Bryan, W. E. Raxxell, ancl J. N. Campbell 1241 addition to the original peak are generated by storage at 4" and four more on further incubation at 37" in the absence of phenylmethylsulfonyl fluoride. The avoidance of storage of extracts at 4" and the use of the incubation step to eliminate nucleic acid in the presence of phenylmethylsulfonyl fluoride to prevent endogenous proteolysis allowed subsequent purification of the DNase activity as a single eluting peak in all subsequent purification steps. The possibility that the additional peaks eluted from the DEAE-cellulose might be other enzymes was considered highly unlikely because of the treatments which lead to their appearance and because the fractions demonstrated a marked similarity in catalytic requirements (3-to 4-fold preference for native versus denatured DNA, optimal DNase activity with Mg++, complete inhibition by 5 mM EDTA, and almost identical pH profiles from pH 6 to 9.5).
Purification of Enzyme-The results of a typical purification sequence are shown in Table IV (carried out at 4" except as noted). dutoactivation of Cell-free &&&--Cell-free extract was prepared and frozen at -20" overnight.
Pancreatic RNase, while not necessary to produce activation was added at a concentration of 20 pg per ml to hydrolyze contaminating RNA. The preparation was incubated at 37" for 90 min. Monitoring of DNase assays, acid solubility at 260 nm, and hydrolysis of benzoyl r,-arginine ethyl ester was carried out at 0, 30, 60, and 90 min to ensure that the maximal increase in DNase activity, greater than 75% proteolytic inhibition by 30 min and the maximal increase in acidsoluble materials had occurred.
A preliminary trial on 5 ml of each batch was performed to assess these points.
First DEAE-cellulose Chromatography-From the autoactivation step, 2.3 g of protein in a loo-ml volume was added to a DEAE-cellulose column (2.5 x 60 cm) prepared as described under 'Methods," and washed in with 100 ml of Buffer B (Buffer A with 2 mM phenylmethylsulfonyl fluoride and 30% glycerol). A linear gradient of 550 ml each of Buffer B and Buffer B containing 0.5 M NaCl was applied and 20 ml fractions collected (flow rate, 30 to 40 ml per hour).
All tubes were examined for absorbance at 280 nm and DNase activity. Linearity of the gradient was confirmed by examining samples of every third tube with a radiometer conductivity meter. The results of a typical column are seen in Fig. 2C. The enzyme eluted at 0.07 M NaCl.
Precipitation of Inactive Material at pH 5..2-Active fractions from the above step were combined and concentrated to 0.2 of original volume using an Amicon UM-IO filter at 40 psi. Sodium acetate (5.0 M, pH 5.2) was added to produce a final concentration of 0.05 M in the DEAE-cellulose concentrate.
The pH was adjusted to 5.2 in a stepwise manner with 2 N acetic acid. This procedure was carried out with constant stirring and continuous pH monitoring on a Beckman Expandomatic pH meter. When the pH reached 5.5, lo-min stirring periods were allowed between acetic acid additions.
After reaching pH 5.2 the mixture was stirred for 20 min and centrifuged at 27,000 x g for 30 min in the cold. The pH of the supernatant solution was adjusted to 7.0 with 2 M Tris (free base) before the next step.
Ammonium Sulfate Fractionation-Solid ammonium sulfate was added gradually in the proportion of 313 g per liter. After being stirred in an ice bath for 20 min, the preparation was cen- trifuged at 27,000 x g for 20 min. To the supernatant solution a further 70 g per liter of (NH&S04 was added with stirring to produce saturation in the glycerol buffer system (about 60% saturation in water).
The mixture was centrifuged at 27,000 x g for 20 min, and the supernatant solution was dialyzed against 100 volumes of buffer containing 30% glycerol, 0.01 M 2-mercaptoethanol, and 0.01 M potassium phosphate, pH 7.5, for 16 hours. The preparation was concentrated by UM-10 ultrafiltration to 30 ml.
Hydroxylapatite Chromatography-From the ultrafiltrate 200 mg of protein in a 15-ml volume was added to a hydroxylapatite column and washed in with 80 ml of 30% glycerol in 0.01 M potassium phosphate and 0.01 M 2-mercaptoethanol (Buffer C). The column size was 2.5 X 20 cm. A linear gradient of 300 ml each of Buffer C and of 0.2 ~5 potassium phosphate (pH 7.5) with 300/, glycerol and 0.01 M 2-mercaptoethanol was applied, and 12.ml fractions were collected.
Every second tube was examined for DNase and absorbance at 280 nm, and gradient linearity was confirmed as described.
The active fractions which eluted at 0.05 Y phosphate were collected, combined, and concentrated to 3 ml by UM-10 ultrafiltration.
Second DEAE-cellulose Chromatography-The 3 ml of concentrate containing 30 mg of protein was diluted to 18 ml in 30% glycerol containing 0.01 M 2-mercaptoethanol. That preparation was applied to a column (1.5 x 25 cm) of DEAE-cellulose and washed in with 15 ml of Buffer C. A linear gradient of 75 ml each of Buffer C and Buffer C modified to contain 0.2 M potassium phosphate was applied, and l-ml fractions were collected at a flow rate of 25 ml per hour.
Absorbance at 280 nm, DNase, and molarity were examined as before.
DNase activity eluted at 0.06 M phosphate and was concentrated to 1 ml by UM-10 ultrafiltration.
Sephadex G-75 Filtration-G-75 gel filtration was carried out with 30% glycerol in 0.01 M 2-mercaptoethanol and 0.05 M potassium phosphate, pH 7.5. The column used was as specified in Fig. 3. It had a void volume of 66 ml and a total volume of 184 ml. The DNase eluted consistently in the fractions seen in Fig. 3.
Purity-The purified enzyme was contaminated with a single additional band detectable on polyacrylamide gel electrophoresis at pH 8.3 or pH 4.0 as well as by assay evidence of RNase activity.
The RNase was considered to be a contaminant for sev- bands which did not contain DNase (pH 8.3) ; in addition, the RNase, unlike the DNase, was stimulated by 5 mM EDTA to produce a 2.fold rise in activity and showed no significant loss of activity when acidified to pH 3.0 prior to assay. The RNase is not the pancreatic RNase added during the DNase purification, however, as a trial purification in t.he absence of added RNase contained RNase in the final purification product. Finally, it can be seen in Fig. 3 that the RNase and DNase activities elute with separate profiles from a Sephadex G-75 column; the two R?;ase peaks presumably represent the added pancreatic RNase and the Pseudomonas RNase.

Stability of Purijied
Enzyme-The enzyme was stable to freezing at -20" in the cell-free extract in the absence of glycerol. was included.
In the presence of glycerol 95 to 100yO activity could be maintained for a minimum of 8 weeks at -20" for all purification stages. If the purified enzyme was dialyzed free of glycerol and frozen at -20" in 2.5 mM potassium phosphate, pH 7, with or without 2-mercaptoethanol, 60% of activity was lost in a single freeze. The half-life at 37" in the same buffer with 0.01 M 2-mercaptoethanol was 7.5 min. The purified enzyme could be stored at 4" in 30% glycerol in buffer for 14 days with a 5 to 15% loss in activity.
Sulfhydryl reagents were essential to maintain activity in the 37" assay. 2-Mercaptoethanol (0.01 M) or 5 to 10 mM dithiothreitol produced maximal activity and maintained a linear reaction (with time) for at least 12 to 18 min. DNase activity was proportional to the protein concentration per assay over a range of 0.5 to 2.25 pg (6 to 32% substrate aolubilization under usual assay conditions). Mg2+ (2.5 mar; cations were used as chlorides) produced optimal activation of the DNase; Mnz+ and Co2+ at the same concentration were 65%; and 23%, respectively, as effective.
In the absence of added cations activity was 10% of maximal.

SpeciJicity-The
DNase demonstrated an approximate 3-to 4-fold preference for native DNA in initial reaction rates. All of the native and 65y0 of the denatured DSA substrate could be solubilized with five readditions of 1.2 units of DNase to a standard assay incubated at 37' for 45 hours; a single addition solubilized a maximum of 48%. The results seen in Fig. 4 suggest the maximum of 487, was due to both enz)matic inactivation and a relative resistance to further hydrolysis of the partially digested substrate.
The effect of modification of the substrate DNA on apparent R, and V,,,,, values is shown in Table V. The I<, does not   significantly  change for native, denatured, 3'.phosphoryl, or 3'.hydroxyl DNA. The TT1,,ax is, however, more than doubled by the introduction of increased numbers of 3'-phosphoryl or hydroxyl groups.
The very close relationship of these two T',,,,, values indicates that the increase in V,,,,, is not dependent on the 3'-phosphoryl group but rather on the presence of increased raphy (see "Methods") allowed the identification of each peak as giren in Fig. 2.
The experiment and results documented in Fig. 6 establish that the mononucleotide products are susceptible to the 5'nucleotidase function of JJ. sodonensi s nuclease described by Berry and Campbell (14) and are, thus, 5'.mononucleotides.
The DTu'ase was also found to have phosphatase activity which could be demonstrated by the phosphatase assay. That assay was shown by paper chromatography (see "Methods") and Sephadex G-10 and G-15 gel filtration (ionic strength of eluting buffers >0.2) to be specific for Pi. Such activity was dependent on the presence of 3'-phosphoryl end groups in DNA (Table VI). The failure of the enzyme to demonstrate phosphatase activity on unmodified DNA or DNA modified to produce 5'-phosphoryl termini as illustrated in This effect, which is commonly observed with other DNases (23, 32, 33) is seen from the data of Table V to be due to a reduced binding affinity for the substrate.
The K, with 0.07 RI NaCl in the assay is raised by 4-fold whereas the V max is identical with that with no NaCl present. Estimation-Estimat,ion of the molecular 15 min or longer; the pH optimum is 7.5 (Tris-HCl); optimal weight using Sephadex G-75 and G-200 was 41,500 and 43,500, activation is obtained by 1 mM magnesium, the relative activity respectively, (Fig. 8) with an average of 42,500. shown with manganese (1 InM) and cobalt is identical with that Presence of Other DNase Activity in P. aeruginosa-The only obtained with the exonuclease (65 and 2370, respectively), and the activity with no cations is 7%.
It would appear the exonuclease and phosphatase functions are components of the same enzyme.
Such is suggested by the relative similarity of the degree of activation by magnesium and other cations, the pH optima, the similar preference for native over denatured DNA, and the remarkably similar sensitivity to inhibition by EDTA and p-chloromercuribenzoate (Table VII). In addition, a nearly constant ratio of phosphatase to exonuclease activity of 0.187 to 0.198 is maintained in the last three purification steps. When the enzymatic activity was eluted from polyacrylamide gels both activities eluted from the same gel slice and none was obtained from the contaminating band.
The activity ratio in the preparation obtained from the gels was 0.190.

Direction of Enzymatic
Attack-The cleavage of phosphate groups from the 3'-terminus of DNA and the release of 5'-mononucleotides suggests that the P. aeruginosa attacks from the 3'.end of the DNA chain. The results documented in Fig. 7 prove that this is true. DNA which was selectively labeled with 32P at the 5'.terminus by the enzyme polynucleotide kinase un dergoes a slow release of the [32P]phospllate relative to the release of 32P from the 3'.end.
Detection of the latter is based on the specificity of the phosphatase assay for Pi and the use of specifically modified DNA to produce a marked predominance of 3'termini.
The amount of [32P]phospllate released in the absence of that modification is less than 10% of t.hat produced in its presence (Table VI).
A 10% correction for a source from other positions was made but no change in the pattern shown in Fig. 7 occurred.
The only products produced in addition to Pi have been shown to be 5'.mononucleotides. Therefore the assay seems specific for 3'.phosphate groups.  other DNase detected was present as a peak which emerged in the wash fraction of the first DEAE-cellulose column, provided the autoactivation step was not carried out. To ensure that other DNases were not failing to elute from the cellulose, column trials were performed identically with that of Fig. 2C except that no autoactivation step was carried out. However, even though the eluting gradient was followed by 1 M NaCl, no additional activities were detected.
The fractions from the initial wash of the column accounted for only 1% of the original DNase measured and were extremely labile.
As a result, only rudimentary characterization was performed in order to distinguish that activity from the exonuclease eluting at 0.07 M NaCl.
The early (wash) DNase could be differentiated in two ways. Uranyl acetate in trichloroacetic acid was about twice as effective a Pseudomonas aemginosa l3Nase Vol. 247,I\To. 4 levels and could be an extracellular enzyme contaminating our preparations.
The absence of other detectable DNases in our system and the detection of an enzyme found in several different bacterial genera strongly suggests that this enzyme is an essential component of the cells' DNA repair, exclusion, or replication systems. The failure to detect other DNases does not exclude their presence as specific additives (e.g. ATE' (38)), or substrate modification (e.g. alkyla,tion (39, 40)) or specialized assay procedures (41) might be necessary to confirm their presence (or absence).
Stabilization, endogenous proteolysis, and association with nucleic acid in the cell extract made this an exceedingly difficult enzyme to purify.
Proteolytic modification of enzymes during their purification to produce diminished stability or multiple molecular forms has been noted in several enzyme systems recently (42-46).
Yeast hexokinase has in particular received considerable attention (43) in this regard. The influence of macromolecules in crude enzyme preparations on chromatography and other purification systems has, however, received very little attention.
The apparent purification of the exonuclease-phosphatase is only 325-fold, but there are reasons to believe that this is an underestimation.
Support for this supposition is advanced by the reduction from several protein bands in gel electrophoretograms in the second to last purification step to only two in the last step and by the marked decrease in intensity of the RNase band relative to that of the exonuclease in the same two steps. The change in specific activity in these same steps, however, was only 1.25 which is not in keeping with the above observatious.
It seems probable that inactivated exonuclease (which may have occurred at any step, even the apparent activation process) was purified through the sequence and diluted the final specific activity.
The effect of ionic strength on the apparent binding affinity for the subst.rate DNA explains the inhibitory effect exhibited on DNase activity and may be the general reason for this widespread finding with other nucleases.
The molecular weight estimation by gel filtration while open to many errors associated with that procedure is the first estimate of the size of enzymes of this group which is available.
The DBase described in this paper differs from that described by Guschlbauer and Halleck (2). The pH optimum of 6 and a combined requirement for citrate as well as Mg"+ were not reproduced in our experimentation.
However, characteristics observed with relatively crude preparations subject to modifying effects of nucleic acids and proteolytic contaminants would not necessarily be expected to be reproducible from one laboratory to another.