Purification and Properties of a Phosphohydrolase from Enterobacter aerogenes*

SUMMARY A phosphohydrolase from Enferobacfer aerogenes which hydrolyzes phosphate mono-and diesters has been purified approximately SO-fold to apparent homogeneity and crystal-lized. The enzyme is produced when the bacteria utilize phosphate diesters as sole phosphorus source. From sedimentation equilibrium experiments the molecular weight of the native enzyme is 173,000; from sodium dodecyl sulfate polyacrylamide gel electrophoresis the subunit molecular weight is 29,000, indicating that the enzyme is hexameric. The hydrolytic activity of the enzyme using both mono-and diesters is maximal at pH 5; the K, of the enzyme for bis-p-nitrophenyl phosphate is constant from pH 5 to 8.5 whereas that for p-nitrophenyl phosphate increases about 40.fold as the pH increases over the same range. The phosphodiesterase activity is not inhibited by chelating agents but is inhibited by several divalent metal ions. 31P NMR spectroscopy was used to identify the hydrolysis products of glycoside cyclic phosphates. The enzyme-catalyzed hydrolysis of methyl /3-n-ribofuranoside cyclic 3 : 5phosphate yields exclusively the S-phosphate whereas that of adenosine 3’:s’.monophosphate

A phosphohydrolase from Enferobacfer aerogenes which hydrolyzes phosphate mono-and diesters has been purified approximately SO-fold to apparent homogeneity and crystallized. The enzyme is produced when the bacteria utilize phosphate diesters as sole phosphorus source. From sedimentation equilibrium experiments the molecular weight of the native enzyme is 173,000; from sodium dodecyl sulfate polyacrylamide gel electrophoresis the subunit molecular weight is 29,000, indicating that the enzyme is hexameric. The hydrolytic activity of the enzyme using both monoand diesters is maximal at pH 5; the K, of the enzyme for bis-p-nitrophenyl phosphate is constant from pH 5 to 8.5 whereas that for p-nitrophenyl phosphate increases about 40.fold as the pH increases over the same range. The phosphodiesterase activity is not inhibited by chelating agents but is inhibited by several divalent metal ions. 31P NMR spectroscopy was used to identify the hydrolysis products of glycoside cyclic phosphates. The enzyme-catalyzed hydrolysis of methyl /3-n-ribofuranoside cyclic 3 : 5phosphate yields exclusively the S-phosphate whereas that of adenosine 3':s'.monophosphate yields a 4: 1 mixture of 3'. and 5'. AMP.
The enthalpies of hydrolysis of cyclic 3':5'-and 2':3'-nucleotides were reported recently (1,2). Whereas the enthalpies of hydrolysis of cyclic 2':3'-nucleotides were in accord with their structure and reactivity, the values reported for the cyclic 3':5'nucleotides were unexpectedly exothermic, suggesting that their hydrolysis is accompanied by the relief of significant strain. In

Materials
Barium dimethyl phosphate was prepared by the hydrolysis of trimethyl phosphate (Aldrich) with a small excess of barium hydroxide, followed by recrystallization from aqueous ethanol. The barium salt was converted to the sodium salt by passage through an Amberlite In-120 (Na+) column.
The preparation of cyclic alkyl and glycoside phosphodiesters and their enzymatic hydrolysis products is described in the following article (5).
Disodium p-nitrophenyl phosphate was purchased from Aldrich, sodium bis-p-nitrophenyl phosphate from Sigma, and nucleotides from P-L Biochemicals and Calbiochem.
Tris base and Tris-HCI were obtained from Sigma. Pipes' was from Calbiochem.2 DEAE-cellulose (Whatman DE52) was obtained from Reeve-Angel. Senhadex G-200 from Pharmacia, and Bio-Gel HT hydroxylapatitefrom Bio-Rad Laboratories. Protamine sulfate (salmine) and enzyme grade ammonium sulfate were from Schwarz/Mann. Alkaline phosphatase from Escherichia coli was obtained from either Sigma or Worthington.
Proteins obtained from the usual commercial sources served as molecular weight standards in sodium dodecyl sulfate polyacrylamide gel electrophoresis. All other chemicals were the best grade commercially available.
2 On the basis of sodium elemental analyses (Galbraith Laboratories, Knoxville, Tenn.), the two lots of Pipes used in this study were 30 and 36y0 disodium salt, contrary to the supplier's description as "monosodium" salt. The elemental composition of carbon, hydrogen, nitrogen, and sulfur was that predicted from the sodium analyses, indicating that the Calbiochem product is a pure mixture of mono-and disodium salts. Analytically pure Pipes diacid was obtained from the mixture of salts by acidification followed by crystallization of the product from hot water. Heat Treatment--The slightly turbid supernatant (271 ml) was rapidly heated to 60" (less than 3 min) in a 90" water bath and then maintained between 60 and 63" for 10 min. The suspension was rapidly cooled to 15" and then centrifuged at 34,800 X g for 25 min. The supernatant was centrifuged again at 34,800 x g for 15 min to remove all particulate matter. A detailed description of this purification step is presented in the text. The linear NaCl elution gradient was applied at Fraction 50. Phosphodiesterase activity (O-O); absorbance at 280 nm ( l -l ) .

DEAE-cellulose
Chromatography-The bright yellow supernatant was applied to a column (3.6 x 13 cm) of Whatman DE52 equilibrated in 0.01 M Tris-HCl, pH 7.6. After sample application, the column was washed for 6 hours with the Tris buffer containing 0.15 M NaCI. A 4.0-liter linear gradient of 0.15 to 0.55 M NaCl in the Tris buffer was then applied for elution of the phosphohydrolase.
The flow rate was maintained at about 200 ml/hour, and fractions containing approximately 30 ml were collected throughout the procedure. Fractions in the region of Number 90 contain the enzyme (Fig.  2). Fractions with a specific activity greater than -2.0 were combined and concentrated with an Amicon ultrafilter employing a PM-10 membrane. Ammonium Sulfate Precipitation-The concentrated enzyme solution (24.8 ml) was brought to 60% saturation in ammonium sulfate by the addition of 9.59 g of the solid enzyme grade salt. The suspension was stirred for 30 min before centrifugation at 34,800 x g for 20 min. The precipitate was dissolved in about 4 ml of 0.01 M potassium phosphate, pH 6.8 (equimolar mixture of the mono-and dibasic salts), and centrifuged for 34,800 x g for 20 min. The enzyme eluted after about 260 ml had passed through the column. Fractions with a specific activity greater than 11 were combined and concentrated with an Amicon ultrafilter employing a PM-10 membrane.
Hydroxylapatite Chromatography-The concentrated enzyme solution (18.5 ml) was applied to a column (2.6 x 7.5 cm) of Bio-Gel HT hydroxylapatite equilibrated in the 0.01 M potassium phosphate buffer. After sample application, the column was washed for 1 hour with the phosphate buffer prior to application of a 400-ml linear gradient of 0.01 to 0.10 M potassium phosphate, pH 6.8. The enzyme was eluted at a flow rate of 22 ml/hour.
The activity eluted after about 160 ml of the gradient had passed through the column. Fractions containing enzyme activity were combined and concentrated with the Amicon ultrafilter.
Crystallization-The concentrated enzyme solution (18.7 ml with 2 mg/ml) was clarified by centrifugation and then carefully adjusted to 35% saturation in ammonium sulfate by addition of the solid enzyme grade salt. This amount of ammonium sulfate was such that the solution became slightly turbid upon standing for several minutes. The suspension was set aside in a refrigerator where crystallization proceeds slowly (Fig. 3).
Criteria of Purity-The purity of the phosphohydrolase was assessed with polyacrylamide gel electrophoresis. Under native (8) and denaturing (7) conditions, the enzyme migrated as a single species. (Although ultracentrifugal methods are less sensitive to impurities, neither sedimentation velocity nor equilibrium experiments gave any evidence of heterogeneity.) That the phosphomonoesterase and phosphodiesterase activities which accompany the purified protein (see "Substrate Specificity") are catalyzed by the single protein species detectable in the Hedrick and Smith electrophoresis system was demonstrated as follows. Gels run in triplicate were separately stained for protein and incubated in p-nitrophenol phosphate and bis-p-nitrophenyl phosphate solutions. Single bands with identical mobility were detected by each staining procedure. Results were similar whether electrophoresis was conducted in 6,8, or 10% gels. Molecular Weight-From sedimentation equilibrium centrifugation and assuming a partial specific volume of 0.73, based on the amino acid composition (Table II), the molecular weight,  TABLE   II   TABLE   III   Amino  acid composition   Substrate   specificity Protein samples were hydrolyzed in 6 N HCl for 24, 48, and 72 hours in the presence of phenol to prevent destruction of tyrosine (lo), with norleucine added to serve as an internal standard. Threonine and serine values were obtained by extrapolation to zero time, and half-cystine was determined after performic acid oxidation (11). Amide ammonia and tryptophan were not deter-Rates measured either by inorganic phosphate production in the presence of excess alkaline phosphatase or by production of p-nitrophenolate in 0.1 M Tris-HCl, pH 8.0, at 30". of undissociated enzyme was determined to be 173,000. When the phosphohydrolase was subjected to electrophoresis with appropriate molecular weight standards in polyacrylamide gels containing sodium dodecyl sulfate (7), a single species of molecular weight 29,000 was observed. These data indicate that the enzyme is a hexamer of subunits with identical molecular weight.
,%&&ion Coeficient-The extinction coefficient, E& estimated as described under "Methods," is 15.1. Substrate Specificity--In addition to the diesters used to follow the purification procedure (trimethylene and bis-p-nitrophenyl phosphates), the enzyme catalyzes the hydrolysis of other phosphate esters (Table III).

Catalytic
Properties-The enzymatic activity for hydrolysis of both bis-p-nitrophenyl and p-nitrophenyl phosphate was maximal at about pH 5, with the ratio of the two maximal velocities at pH 4.9 being about 4 (Fig. 4).
Lineweaver-Burk plots for bis-p-nitrophenyl phosphate hydrolysis were often nonlinear, with Hill plots of the same data having a slope (n) less t,han unity. The apparent K, was calculated from such data by a graphical procedure (12). Phosphodiesters undergo no change in protonation, from pH 5 to 8.5, and the apparent K, for bis-p-nitrophenyl phosphate is indcpendent of pH in this range (Fig. 4); the K, for p-nitrophenyl phosphate which has a pK, of 5.4 (13) increases 40.fold as the pH increases over this same range. This behavior suggests that a phosphoric acid ester is recognized as a substrate on the basis of its single negative charge. Incubation of the enzyme with several chelators3 did not cause any significant inactivation.
Of various divalent metal ions tested, none produced significant stimulation of activity, but several did produce partial inactivation. 4 Al-  on pH. Buffers used were sodium acetate, pH 4.6, sodium cacodylate, pH 4.9, sodium Pipes, pH 5.9 and 6.9, Tris-HCl, pH 8.4. The buffer concentrations were 0.05 M. The assays were performed at 30" by following the production of p-nitrophenolate at 400 nm.
though nucleases and other phosphohydrolases frequently require metal ions for activity, the lack of pronounced activation with added metal ions or inhibition by chelators suggests that this phosphohydrolase is not a metal-dependent enzyme.

Hydrolysis
Products-The possible hydrolysis products of methyl a-u-glucopyranoside cyclic 4:6-phosphate (14), methyl P-n-ribofuranoside cyclic 3 :5-phosphate (5)) and cyclic AMP were distinguished by the splitting pattern in 31l' NMR spectra produced by 31P-H coupling. The signal from the phosphorus atom of esters of a secondary alcohol, which have 1 hydrogen atom on the hydroxyl carbon, is a doublet; the signal from esters of a primary alcohol, which have 2 hydrogen atoms on the hydroxyl carbon, is a triplet. (In glycoside phosphates, the 2 hydrogen atoms on the carbon with a primary hydroxyl group are not equivalent, but in practice the resonance does appear as a triplet.) The methyl riboside cyclic phosphate (Fig. 5) and the methyl glucoside cyclic phosphate (data not shown) were hydrolyzed by the enzyme in Pipes buffer at pH 7.3 to yield products whose spectra were identical with those of authentic methyl /!Ln-ribo- Determination of the enzymatic hydrolysis product of methyl @-n-ribofuranoside cyclic 3:5-phosphate with 31P NMR spectroscopy. The top left and top right spectra arc those of authentic cyclic phosphate and authentic methyl p-n-ribofuranoside 5-phosphate, respectively. The bottom left spectrum is that of cyclic phosphate which was hydrolyzed in barium hydroxide solution. The bottom right spectrum is that of cyclic phosphate which was hydrolyzed with the phosphohydrolase. 5-phosphate (5) and methyl a-n-glucopyranoside 6-phosphate (15)) respectively. Hydrolysis of each glycoside cyclic phosphate with barium hydroxide at 100" yielded a mixture of products, with the predominant products being identified as methyl @-n-ribofuranoside 3-phosphate and methyl a-n-glucopyranoside 4-phosphate, respectively. In proton-decoupled spectra only a single resonance was observed in the spectrum of the enzymatic hydrolysis product of each glycoside; the spectrum of each base-hydrolyzed glycoside cyclic phosphate had two clearly resolvable resonances.
Cyclic AMP (Fig. 6) was hydrolyzed by both the phosphohydrolase and barium hydroxide to yield mixtures of 3'-and 5'-AMP. The composition of a mixture can be evaluated from the intensities of resonances in proton-decoupled spectra. The barium hydroxide hydrolysis of cyclic AMP yielded 3'.AMP and 5'.AMP in a ratio of about 9; the reason for the discrepancy between this value and the literature value of 5 (16) is not known. The enzymatic hydrolysis resulted in a 3'-AMP : 5'-AMP ratio of 4.
Relationship to Other Phosphohydrolases-Enterobacteriaceae are known to produce a cyclic 2':3'-nucleotide phosphodiesterase (17). The enzyme isolated in this study appears to be distinct from the previously reported enzymes since the latter are synthesized in the presence of inorganic phosphate, are released by osmotic shock, have different substrate specificities, and require added metal ions for activity. Also, the phosphohydrolase from Enterobacter aerogenes appears to be the first isolated which is known to cleave simple aliphatic phosphodiesters, although it is not unusual for phosphodiesterases to cleave bis-p-nitrophenyl FIG. 6. Determination of the enzymatic hydrolysis product of cyclic AMP with 31P NMR spectroscopy. The top left and top right spectra are those of authentic cyclic AMP and 3'-AMP, respectively. The bottom left spectra are proton-coupled and -decoupled spectra, respectively, of cyclic AMP hydrolyzed by the phosphohydrolase. The bottom right spectra are proton-coupled and -decoupled spectra, respectively, of cyclic AMP hydrolyzed in barium hydroxide solution. Spinning side bands are marked with arrows.
phosphate or other esters in which a p-nitrophenyl residue is a leaving group (18).