THE JOURNAL OF BIOLOQICAL CHBUIBTRY
   Vol. 239, No. 1. January 1964
      Pnnled in U.S.A.

Enzymatic Synthesis of Deoxyribonucleic Acid

XV. PURIFICATION AND PROPERTIES OF A POLYMERASE FROM BACILLUS SUBTILIS*

             TUNEKO OUZAKI t AND ARTHUR KORNBERQ

From the Department of Biochemistry, Stanford University School of Medicine, Palo Alto, California

             (Received for publication, July 18, 1963)

Nuclease activity persists in the most purified deosyribonucleic
acid polymerase preparations from normal and bacteriophage
T2-infected Escherichia coli (1,2).  It has therefore been difficult
to determine whether such a nucleolytic activity is an invariable
component of bacterial polymerase.  An opportunity to explore
a source of polymerase which was promising in this respect was
suggested by a survey of a variety of microorganisms for nuclease
activity.' Bacillus subtilis was identified in a group of organisms
which have low levels of nuclease activity; since extracts of this
organism have substantial polymerase levels, the polymerase to
nuclease ratio compared to that in E. coli is relatively high.
Purification of DNA polymerase from B. subtilis therefore might
yield a nuclease-free enzyme and would also provide a poly-
merase to compare with the E. coli enzyme for specificity of
primers, for relative effectiveness of various base analogues of
the deoxynucleoside triphosphates, for synthesis of deosyadenyl-
ate-deoxythymidylate copolymer de muo, and for the ability
to incorporate a ribonucleotide into a DNA polymer.

This report describes the preparation of a polymerase from
B. subtilis which has little or no nuclease activity.  Studies of
the properties of the enzyme reveal fundamental similarities to
the E. coli polymerase.  Quantitative comparisons of base ana-
logue incorporation support the view that base pairing rather
than enzyme specificity determines the incorporation of bases
into DNA.

EXPERIMENTAL PROCEDURE

Materials

Unlabeled deoxyribonucleoside triphosphates were purchased
from the California Corporation for Biochemical Research.
=P-Labeled deoxyribonucleotides, labeled in the ester phosphate,
were prepared as described previously (3).  Deoxythymidine-
2-14C, purchased from the New England Nuclear Corporation,
was enzymatically phosphorylated to dTTP ("C-dTTP) (4).
I4C-Labeled dCTP was prepared from DNA obtained from a
Chromatiurn species grown on 14302 as the sole carbon source
(5).  Cytidine triphosphate labeled with 32P in the ester phos-
phate was prepared by enzymatic phosphorylation of 32P-labeled
cytidine 5'-phosphate, prepared according to Hurwitz (6).  The
procedure for preparation of the base analogues of deoxynucleo-
side triphosphates was previously described (7).

S2P-Labeled DNA (10 pc per pmole of phosphate) was isolated
* This work was supported by grants from the National Insti-

t Present address, Department of Chemistry, Faculty of Sci-

1 E. A. Pratt and I. R. Lehman, unpublished observations.

tutes of Health, United States Public Health Service.

ence, Nagoya University, Nagoya, Japan.

from E. coli as described by Lehman (8).  Calf thymus DNA
was isolated according to Kay, Simmons, and Dounce (9), and
B. subti2is (SB 19) DNA by the method of Marmur (10).  The
dAT copolymer2 was prepared by synthesis de mo with the
hydroxylapatite fraction of the E. coli DNA polymerase (11).
S*P-dGdC (-*pC-) and "C-dAT (-p*T-) polymers were syn-
thesized in a primed reaction with E. coli polymerase (11, 12).
Heatdenatured DNA was obtained by heating DNA in 0.005
M KCI at 100" for 10 minutes and cooling quickly in an ice bath.
DNA terminally labeled at its 3'-hydroxyl end with 14C-deosy-
thymidylate was prepared with E. coli DN-4 polymerme in a
"limited reaction" (5).

Concentrations of the polynucleotides are expressed as equiv-
alents of nucleotide phosphorus.
The hydroxylapatite fraction of E. coli DNA polymerase was
prepared as described elsewhere (1).  E. coli endonuclease had
a specific activity of 6,360 units per mg of protein (carboxymethyl
cellulose fraction) (13); DNA phosphatase, purified from E. coli,
had a specific activity of 76,500 units per mg of protein (phos-
phocellulose pervaporate fraction) (14).  Crystalline pancreatic
DNase was obtained from Worthington Biochemical Corpora-
tion.  Micrococcal DNase purified from Micrococcus pyogenes
was a gift from Dr. C. A. Dekker.

DEAE-cellulose was purchased from Brown Company; What-
man phosphocellulose (P-70) from W. and R. Ralson, Ltd.;
Hypatite C (hydroxylapatite) from Clarkson Chemical Com-
pany, Inc.; Dextran 500 from Pharmacia; polyethylene glycol
(Carbowax so00) from Union Carbide Chemical Company;
antibiotic medium 3 from Difco Laboratories; and Superbrite
glass beads from Minnesota Mining and Manufacturing Com-
pany.

ililethods

Assay of B. subtilis DNA Polymerase-The my measures
the conversion of a 14C- or S*P-labeled deoxynucleoside triphos-
phate into an acid-insoluble product.  The standard incubation
mixture (0.3 ml) contained 20 pmoles of Tris-maleate-KOH
buffer, pH 8.2, 2 pmoles of MgCl2, 0.3 pmole of 2-mercapto-
ethanol, 6 mpmoles of dAT copolymer, 10 mpmoles each  of
dATP and 14C-dTTP (2 x lo6 c.p.m. per pmole), and 0.01 to
0.1 unit of enzyme.  In some cases, 40 mpmoles of calf thymus
DNA replaced the dAT polymer with 10 mpmoles each of dGTP
and dCTP in addition to dATP and I'C-dTTP.  Enzyme dilu-

* The abbreviations used are: dAT copolymer, copolymer of
deoxyadenylate and deoxythymidylate; dGdC, polymer consist-
ing of homopolymers of polydeoxyguanylate-polydeoxycytidylate;
the prefix (k'' denotes "ribo."

259

260

I. Cell extract,. ................
11. Phase separation.. ..........
111. Ammonium sulfate. .........
IV. DEAE-cellulose. ............
V. Phosphocelluloseb. ..........
VI. Hydroxylapatite (VI-2) .....

   TABLE I
Purification of enzyme

nirlml

25.8
9.24
30.6
2.53
0.44
0.09,

Enzymatic Synthesis of DNA. XV

1.4
3.1
5.4
24.3

m
569
(1060)d

Vol. 239, No. 1

0.15


0.41
2.45

20.8
19
(100)d

Fraction and step

I I Polymerase activity"

I I

units X

10-
42.7
33
19.8
14.6

5.8

-c

0 The unit of activity here, as noted in "Methods," is bmed
on the incorporation of the labeled nucleotide only.  Therefore,
in terms of total nucleotide incorporated, the values with dAT
primer are actually twice this amount and those with DNA
primer approximately 4 times that amount.

* This step was performed many times on a smaller scale, and
these values are calculated for the large scale procedure.

See the text.

d Assays in the presence of hydroxylapatite Peak a (0.082 pg

of protein).

t.ions were made in 0.05 M Tris bder, pH 7.5, containing 0.01
M 2-mercaptoethanol, bovine serum albumin (1 mg per ml),
and 0.1 M ammonium sulfate.
After 30 minutes of incubation at 37", the reaction was stopped
by the addition of cold 7 % perchloric acid, and the acid-insoluble
fraction was isolated either by centrifugation (15) or by filtration
(16) as described before. Radioactivity measurements were
made in a windowless gas flow counter (Nuclear-Chicago) or in
a Packard Tri-Carb scintillation counter.
One unit of enzyme is defined as the amount catalyzing the
incorporation of 10 mpmoles of labeled deoxynucleotide into
t8he acid-insoluble product.  Specific activity is expres,sed as
units per mg of protein.

DNase or dATase activities were determined by measurement
of the conversion of a*P-DNA or IC-dAT copolymer (2-W-
t.hymine) to acid-soluble products as described by Lehman (8).
The 0.3-ml incubation mixture contained 25 mpmoles of 32P-
DNA or 6 mpmoles of "C-dAT polymer, 20 pmoles of Tris-
maleate-KOH buffer, pH 8.2, 2 pmoles of MgClz, 0.3 pmole of
2-mercaptoethanol, and 0.01 to 0.2 unit of enzyme.  DNA phos-
phatase activity was assayed by the method described elsewhere
(14).  The 0.3-ml incubation mixture contained 50 mpmoles of
3'-phosphoryl-terminated "P-DKA (micrococcal endonuclease-
treated 'ZP-DNA), 20 pmoles of potassium phosphate buffer,
pH 7.0, 3 pmoles of MgC12, 0.3 pmole of 2-mercaptoethanol, and
0.01 to 0.3 unit of enzyme.  After 30 minutes of incubation at
37", the acid-soluble fraction was isolated and treated with Norit,
and the resulting acid-soluble, Norit-nonadsorbable aZP was
counted.

Treatment of Priws wilh Nuclemes-E. coli DNA phos-
phatase treatment: The incubation mixture (1.0 ml) consisted
of 240 mpmoles of dAT or 500 mpmoles of calf thymus DNA,
50 pmoles of Tris-maleate-KOH buffer, pH 7.0, 10 pmoles of
MgClz, and 13 units of enzyme.  After a 30-minute incubation

at 37", aliquots were used in the polymerase reaction without
inactivating the enzyme.  E. coli endonuclease treatment: The
reaction mixture (1.0 ml) contained 240 mpmoles of dAT or
500 mpmoles of calf thymus DNA, 50 pmoles of Tris-HC1 buffer,
pH 7.8,lO pmoles of MgC12, and 0.2 unit of E. coli endonuclease.
After a 75-minute incubation at 37", 20 mpmoles of soluble RNA
were added to the reaction mixture to inhibit the endonuclease.
Micrococcal nuclease treatment: The incubation mixture (1.0
ml) contained 240 mpmoles of dAT or 500 mpmoles of calf
thymus DNA, 50 pmoles of Tris-HC1 buffer, pH 8.6, 10 pmoles
of CaC12, and 0.4 unit of micrococcal nuclease.  After incubation
for 60 minutes at 37", the reaction mixture was chilled and di-
alyzed against 3 liters of 0.05 M KCI for 15 hours to remove
CaClp.  Further treatment of the micrococcal digest with E.
coli DNA phosphatase was carried out on the dialyzed sample.

Determination of nucleotide sequence in the synthetic polymer
was carried out by chemical hydrolysis in the diphenylamine-
formic acid system described by Burton and Petersen (17).
Protein was determined by the method of Lowry et al. (18).
Samples with interfering materials were first precipitated with
cold 5 % trichloroacetic acid.  Optical measurements were made
with the Zeiss PMQ I1 spectrophotometer.

RESULTS

Purtficdion of Enzym

All procedures were carried out at &5O unless otherwise in-
dicated.
Growth of Bacteria-Wild-type B. subtilis (SB 19) was grown
in 90 liters of broth (Difco antibiotic medium 3) with vigorous
aeration at 37" in a fermentor and was harvested at the end of
logarithmic phase.  The yield of packed wet cells was approxi-
mately 3.3 g per liter of culture; the cell paste was frozen and
stored at -20" until used.

Preparation of Cell Extract-In a 5-liter Waring Blendor
equipped with a cooling jacket were mixed 300 g of cell paste,
900 g of glass beads (Superbrite, average diameter 200 p), and
130 nd of 0.05 M glycylglycine buffer, pH 7.0, containing 0.002
M EDTA and 0.002 M glutathione, for 35 minutes at a rheostat
setting of 80 volts.  The speed was then reduced by two-thirds,
270 ml of the same buffer were added, and mixing was continued
for 10 minutes.  The glass beads were allowed to settle for 10
minutes, the supernatant fluid was decanted, and the beads
were washed with 800 ml of the same buffer.  The supernatant
fluids were combined and centrifuged for 30 minutes at 10,OOO X

g.  The resulting supernatant fluid (1,150 ml) was collected
(Fract<ion I, Table I).  The temperature was maintained at
or below 7" throughout the procedure.

Removal of Nucleic Acids by Phase Partition3 (19)-To Frac-

8 A simpler purification procedure based on a protamine precipi-
tation of polymerase followed by chromatography on DEAE-cellu-
lose was abandoned for three reasons.  The yields with the prota-
mine step were variable, the capacity of DEAE-cellulose for the
protamine fraction was low, and, most significant, the polymerase
activity was distributed in two separate peaks on the DEAE-cel-
Iulose chromatogram.  One of the peaks was eluted with 0.2 M
phosphate buffer, pH 7.4 (20 to 50$&), and the other with 0.32 M
buffer (50 to 80%). When the latter peak was rechromatographed
on DEAE-cellulose, over 70% of the activity was then eluted with
0.2 M buffer. We suspect that nucleic acid persisting in the prota-
mine fraction was responsible for the heterogeneous character of
polymerase activity on the DEAE-cellulose chromatogram.  The
phase separation method appeared to give a more complete and
reproducible removal of nucleic acid from the polymerase.

January 1964 T. Okazaki and A. Kornberg 261

tion I were added 132 ml of 20% (w/w) Dextran 500 solution
and 370 ml of 30% (w/w) polyethylene glycol to give final con-
centrations (by weight) of 1.6% and 6.4% for dextran and poly-
ethylene glycol, respectively (ZO).'  NaCl (387 g) was added
gradually to the solution (final concentration, about 4 M), and
the mixture was stirred for 2 hours at 0" and then centrifuged
at 500 X Q for 10 minutes.  The resulting 1480 ml of clear top
phase (polyethylene glycol -i- protein) were recovered to give
Fraction 11, and the turbid bottom phase (280 ml, dextran +
nucleic acids) was discarded.

Ammonium Sulfate--To remove the NaCI, Fraction I1 was
dialyzed against 40 liters of 0.3 M potassium phosphate buffer,
pH 7.4, containing 0.002 M EDTA and 0.01 M 2-mercaptoethanol,
for 20 hours; the volume increased to 2,400 ml and 94% of the
activity was recovered.  Solid ammonium sulfate (480 g) was
added gradually to the dialyzed fraction with continuous stirring,
and the solution was transferred to a 4-liter separatory funnel.
After 7 hours at 4", a lower phase containing the enzyme sep-
arated from the upper polyethylene glycol phase.  To 2,100 ml
of the lower phase were added 192 g of ammonium sulfate, and
the resulting precipitate was removed by centrifugation at 10,000
x g for 10 minutes.  Another 264 g of ammonium sulfate were
added to the supernatant fluid, and the precipitate was recovered
by centrifugation and dissolved in 120 ml of 0.3 M potassium
phosphate buffer, pH 7.4, containing 0.002 M EDTA and 0.002
M glutathione (Fraction 111).  Fraction I11 was stored at - 15"
until used.

DEAE-cellulose Chromatag aphy-A column of DEAE-cellu-
lose (8.3 cm* x 12 cm) was prepared and equilibrated with 0.02
M K2HP04 containing 0.01 M 2-mercaptoethanol.  Fraction I11
(70 ml) was dialyzed against 3 liters of 0.3 M potassium phosphate
buffer, pH 7.4, containing 0.002 M EDTA and 0.01 IU 2-mercap-
toethanol, for 4 hours, and the dialysate (75 ml) was diluted to
450 ml with 0.01 M 2-mercaptoethanol before being adsorbed
to the column at the rate of 4.5 ml per minute.  The column
was then washed with 200 ml of 0.05 M potassium phosphate
buffer, pH 7.4, containing 0.002 M 2-mercaptoethanol and 0.002
M EDTA.  A linear gradient of elution was applied with 0.05 M
and 0.5 M potassium phosphate, pH 7.4, as limiting concentra-
tions.  The total volume of the gradient was 2OOO ml, and 0.01
M 2-mercaptoethanol and 0.002 M EDTA were present through-
out the gradient.  The flow rate was 4.5 ml per minute, and
40-ml fractions were collected.  Of the activity applied to the
column, 85% was eluted between 6.4 and 12 resin bed volumes
of effluent.  Peak fractions having specific activities of 15 to
41 were combined (360 ml) and concentrated to one-third this
volume by dialysis against 30% polyethylene glycol in 0.3 M
potassium phosphate bder, pH 7.4, containing 0.002 M EDTA
and 0.002 M glutathione.  The DEAE-cellulose step was carried
out with the remaining 50 ml of Fraction 111 on a smaller colum
(8.3 cm* x 10 cm).  The peak fractions recovered from the
two columns were combined to yield 240 ml of Fraction IV (74%
recovery from Fraction 111).  Fraction IV was stored at - 15"
until used.

PhosphOceUulose Chromatography-Fraction IV (80 ml), di-
alyzed against 8 liters of 0.05 M potassium phosphate buffer,
pH 6.8 (containing 0.002 M EDTA and 0.01 M 2-mercaptoetha-
nol), for 12 hours and diluted to 200 ml with 0.01 M 2-mercapto-
ethanol, was passed through a phosphocellulose column (15

4 We gratefully acknowledge the advice and help of Dr. Per-Ake
Albertsson with this procedure.

cm2 X 10 cm) previously equilibrated with 0.02 M potassium
phosphate buffer, pH 6.5, containing 0.002 M EDTA and 0.01
M 2-mercaptoethanol.  The column was washed with 300 ml of
0.05 M potassium phosphate buffer, pH 6.8, containing 0.002 M
EDTA and 0.01 M 2-mercaptoethanol.  The column was eluted
with a linear gradient of potassium phosphate buffer, pH 6.8,
from 0.05 M to 0.40 M.  The buffer contained 0.01 M 2-mercapto-
ethanol and 0.002 M EDTA, and 500 ml of each buffer were used.
Fractions of 10 ml were collected at 3-minute intervals.  The
enzyme was eluted between 2.2 to 3.7 resin bed volumes of
effluent in a zone containing 15% of the initial protein.  Frac-
tions having specific activities of 172 to 302 were combined (55
ml, 51 % recovery), concentrated by dialysis for 4 hours against
1 liter of 30% polyethylene glycol solution in the buffer used for
washing the column, and then dialyzed against 2 liters of this
buffer for 12 hours (Fraction V, 21 ml, 40% recovery from Frac-
tion IV).  It was used immediately for the next step of purifica-
tion.  In dilute solution at 0", Fraction V lost half of its activity
in 24 hours (see footnote b, Table I).

Hydroxylapatite Chromatography-A column of hydroxylapa-
tite (0.9 cm2 x 10 cm) was prepared and washed with 0.02 M
potassium phosphate buffer, pH 6.8, containing 0.01 I 2-mer-
captoethanol. Fraction V (19 ml) was applied to the column at a
rate of 0.4 ml per minute, and the adsorbent was washed with
10 ml of 0.05 M potassium phosphate buffer, pH 7.4, containing
0.01 M 2-mercaptoethanol.  The enzyme was then eluted with a
linear gradient of potassium phosphate buffer at pH 7.4, the
limits of which were 0.05 M and 0.25 M.  The buffers contained
0.01 M 2-mercaptoethanol and 80 ml of each buffer were used;
fractions of 2 ml were collected every 5 minutes.  The chromato-
gram is shown in Fig. 1.

A peak of nuclease activity was detected between 2.2 to 6.6
resin bed volumes of effluent (Peak a) whereas polymerase ac-
tivity occurred between 6.6 to 13.3 resin bed volumes (Peak b)
with only a trace of nuclease activity.  The polymerase peak
was divided into four portions as shown in Fig. 1.  Each portion
was concentrated as described previously in 0.2 M potassium
phosphate buffer, pH 7.4, containing 0.002 M glutathione and
0.002 M EDTA, to give a protein concentration of about 0.1 mg
per ml (Fractions VI-1 to VI-4).

Fraction VI failed to show enzyme proportionality to the
amount of protein added (Fig. 2), and the recovery at this step
was only 10 to 25%.  Addition of Peak a material to the poly-
merase reaction mixture stimulated the activity of Fraction VI
several fold, as shown in Fig. 2, improved enzyme proportionality,
and increased the rekovery of activity (about 50% from Frac-
tion V).

Nuclease Actwities in Enzyme Preparatbm

The nuclease and polymerase activities of various fractions are
compared in Table 11.  Values for the most purified E. coli DNA
polymerase are also provided for reference.  The polymerase to
nuclease activity ratio in the crude extract (Fraction I) of B.
subtilis was higher to begin with than that of the extensively
purified E. coli polymerase.  In Fraction V, the nuclease ac-
tivity was relatively greater on native DNA than on heat-de-
natured DNA or dAT polymer.  However, chromatography on
the hydroxylapatite column separated this nuclease activity
(Peak a, Fig. 1) from the polymerase fraction, and only a trace
of nuclease activity could be detected in Fraction VI.6

Assays of this fraction for nuclease based on the destruction

262 Enzymatic Synthesis of DNA. XV Vol. 239, No. 1

Fraction number

FIG. 1. Chromatography of Fraction V on hydroxylapatite.
Elution was achieved with a linear gradient of potassium phos-
phate buffer ranging from 0.05 to 0.25 M at pH 7.4.  The eluate
fractions were assayed for protein, DNase, and polymerase (dAT-

primed) as described in "Methods."  Nuclease fractions are la-
beled as a, and polymerase fractions as b.  The contents of tubes
35 to 40, 41 to 45, 46 to 50, and 51 to 60 were pooled and concen-
trated separately (Fractions VI-1, -2, -3, and -4, respectively).

The pH optimum of the DNase activity in Fractions V and
VI and Peak Q was 8.2 in Tris-maleate-KOH buffer, and MnCll
(0.2 pmole) could be used to replace MgCls (2 pmoles).  All three
fractions showed DNA phosphatase activity (cleavage of inor-
ganic phosphate from 3'-phosphoryl ends of DNA molecules)
(14) in constant ratio to DNase activity; the ratios for Fractions
V and VI and Peak a were 0.26, 0.22, and 0.19, respectively.
The DNase in Peak a can apparently initiate its attack at the
3`-hydroxyl ends of DNA.  This was shown with DNA termi-
nally labeled at its 3`-hydroxyl terminus with IC-deoxythymidy-
late as a substrate (5).  When 85% of the radioactivity was
rendered acid-soluble, only 3 '% of the nucleotides in the DNA had
been released, as judged by ultraviolet absorption of the acid-
soluble fraction.  This is a result similar to that obtained with
venom diesterase (5), an exonuclease known to initiate its step-
wise hydrolysis from the 3`-hydroxyl end of DNA (21).

The stimulatory effect of Peak a on polymerase activity is
most probably due, as in the E. coli studies (22), to removal of
inhibitory 3'-phosphoryl ends from DNA chains and to the pro-
duction of new 3`-hydroxyl ends in their place.  The same stimu-

of B. subtilis transforming factor activity were generously per-
formed by Dr. Walter Bodmer.  With 1.88pg of this fraction incu-
bated with 6.2 -moles of DNA for 60 minutes at 37", 90% of the
transforming factor activity waa retained.

latory effect was obtained when the purified E. coli DNA phos-
phatase (14) was used in place of Peak a.  Since the E. coli DNA
phosphatase-exonuclease has the dual capacity to act as a DNA
phosphatase and exonuclease (cleaves the 3'-phosphoryl linkage
of mono- or diesterified phosphate residues of a DNA chain), it
seems likely that the Peak a enzyme of B. mbfi2ig is also a single
protein with bot,h functions.

Properties of Enzyme

pH Optimum-With Fraction V, the maximal rate of synthesis
in the dAT-primed reaction occurred around pH 8.2 in Tris-
maleate-KOH buffer (Fig. 3).  The same results were obtained
with Fraction VI and Tris-maleate-KOH or glycine buffer.

Requiremenls for Reaction--A DNA primer, a divalent metal,
and the appropriate deoxynucleoside triphosphates were essen-
tial for the reaction (Table 111).  Under the assay conditions
and at the level of dAT copolymer used, MnCL was almost as
effective as MgClz at low concentrations (2 x lO-' M) but was
very inhibitory at a higher concentration (1 x 10-8 M) (Fig. 4).

Incorporation of Base Analogues-With the E. coli and T2
polymerases, it has been shown that analogues of the naturally
occurring bases could serve as substitutes in a manner governed
by the base pairing of adenine to thymine and of guanine to cy-
tosine, as proposed by Watson and Crick for the double helical

January 1964 T. Okazaki and A. Kornberg 263

mpmolss rvclootide m p+9hate
   relmcd/mg protar
13.7O 43.7

3.1 234   28.7 827


6.2 119    3.2 346


    62    6.4 153

structure of DNA (23).  The same replacements, in the form of
deoxynucleoside triphosphates, were found with Fraction V of
B. subtilis polymerase.  Thus, uracil and its 5-bromo and 5-
fluor0 derivatives could replace thymine; hydroxymethyl-, 5-
methyl-, 5-bromo-, and 5-fluorocytsine could replace cytosine;
and hypoxanthine could replace guanine.  Xanthine was not
incorporated, and N-methyl-5-fluorocytsine gave very little in-
corporation when used in place of dCTP (Table IV).  The values
obtained are virtually indistinguishable from those obtained with
the E. coli polymerase (hydroxylapatite fraction) tested in paral-
lel experiments.

2.03


5.14

I 1

2.94

37

I

/

29.6
(148) b

3,328
(6,680)

AL~ OC Craction SZI-

6.3 685

31.9 4,400

3'-Phosphoryl-terminated DNA

FIQ. 2. Proportionality of the polymerase activity of Fraction
VI-2.  The standard dAT-primed assay was used with and with-
out addition of 0.082 pg of Peak a of the hydroxylapatite chroma-
togram.

  850
(4,250) *
10,650
(21,300)b

-P1 Tris-Malea te- KOH

P- K; phosphate

6.0 700 8.0 9.0

PH

FIQ. 3. Effect of pH and buffer on polymeraae activity.  Activ-
ity was measured in the routine dAT-primed assay except that
the buffer indicated replaced the usual buffer.  The pH of each
buffer (0.05 M) was determined at room temperature.  To each
assay was added 0.88 pg of Fraction V, and the value in Tris-
maleate-KOH at pH 8.2 was set at 100%.

Atkxnpt to Impor& a Ribonucleoside Triphosphate-As de-
scribed by Berg, Fancher, and Chamberlin (X), E. coli DNA
polymerase, in the presence of Mn*, incorporates ribonucleoside
triphosphates in place of their deoxy counterparts.  However,

TABLE I1

Comparison of nuclease and polymerase activities in various entm fractions

Asaay conditions for the B. subtilis enzymes are described in "Methods."  E. coli polymerase assay conditions were like those
used for B. subtilis, except that potassium phosphate buffer (pH 7.4, 0.06 M) was used.  Nuclease activities in E. coli polymerase
fraction were assayed at pH 9.2 in glycine buffer.

Primer or substrate

Polymerase, Fraction

I 1 v 1 VI-2

Heated DNA
Native DNA

dAT

Nuclease, Fraction I Polymcrasc to nuclease ratio, Fraction

0.58


1.47

,, When MgClr was replaced by CaClz (0.2 pmole per 0.3 ml of assay mixture), this value was 1,780 and less than 1.9 with Fraction V.
b Values assayed in the presence of 0.082 pg of hydroxylapatite Peak a; this amount of Peak a had no detectable nuclease activity

by itself.

264 Enqmatic Synthesis of DNA. XV VoI. 239, No. 1

System

, Complete.. ......................

..................

..................

Minus MgClp, plus MnC12b.. .........
Minus dATP ...... .........
Minus dATP, plus ...........
Minus dATP, plus
Minus dGTP ...... ............
Minus dCTP. ........
Minus dATP, dGTP, dCTP.. ....

... ... .

W-dTMP incorporated
UT primer D$~~~,,r

     #S@lOlrr

0.435      0.461
<0.005"    <o.m
<0.005     <0.005

0.394 0.395
0.012 0.012
0.013 0.010
0.007

0.013
0.020
0.019

B. subtilis DNA polymerase fails to utilize a ribonucleoside tri-
phosphate when examined under the conditions described in
Table V.  With rCTS2P in place of dCT*zP, no incorporation was
observed either with Mg++ or with Mn++.  rATP and rGTP

also failed to replace dATP or dGTP, respectively, in supporting
the incorporation of WdTTP.  The value for incorporation with
the four deoxynucleoside triphosphates wm 0.31 mpmole; in the
absence of dATP or dGTP or upon replacement of dATP by
rATP or dGTP by rGTP, there was no detectable synthesis, Le.
less than 0.01 mpmole.

Eflects of VarWu.9 Treatments of Primers on Prim.ing Actwdies-
In order to study the influence of end groups of primer molecules
on polymerase reactions, dAT polymer and calf thymus DNA
were treated with the following nucleases: (a) E. coli DNA phos-
phatase-exonuclease, an enzyme known to specif~cally remove
3'-phosphoryl end groups from native DNA; (a) E. coli endo-
nuclease and pancreatic DNase, which cleave phosphodiester
bonds of DNA to produce 3'-hydroxyl and 5'-phosphoryl termini
throughout the molecule; (c) micrococcal endonuclease, which
cleaves phosphodiester bonds of DNA , producing 3'-phosphoryl
and 5`-hydroxyl termini; and (d) subsequent treatment of the
micrococcal endonuclease digestion product with E. coli DNA
phosphatase to remove 3'-phosphoryl end groups.  Priming ac-
tivities of these DNAs for Fraction VI-2 and highly purified E.
coli polymerase were determined (Table VI).  With both dAT
polymer and thymus DNA, removal of the 3'-phosphate by DNA
phosphatase resulted in an increase of priming activities for B.
subti1i.s as well m for E. coli polymerase.  The treatment with
the endonucleases which produce 3`-hydroxyl and 5'-phosphoryl
ends also stimulated priming activities of both primers, whereas
treatment with micrococcal nuclease resulted in marked reduc-
tion of priming activities.  It should be noted that although the

FIG. 4. The influence of metal ion concentration on the rate of polymerase activity. The standard dAT-primed assay was used,
except that the metal ion concentration was varied, as shown.  To each assay was added 0.88rg of Fraction V enzyme, and the value
with 2 pmolea of MgCls per assay mixture (0.66 X 10-* M) was set at 100%.

T. Okazaki and A. Kornberg

6
2
5
3
4

265

36
33
66

107 (61)
118 (127)
12 (10)
82 (55)
98 (129)

37
57

latter treatment does not change the total number of 3'-hydroxyl
and 5'-phosphoryl end groups, there was a marked decrease in
priming activity.  This priming activity could be restored by the
enzymatic removal of 3'-phosphoryl end groups from the primer
molecule by the DNA phosphatase.

These results indicate that the 3'-hydroxyl terminus is neces-
sary for the reaction of both polymerases whereas a 3'-phosphoryl
end is inhibitory.

Heat denaturation of thymus DNA resulted in little change of

MgClz
MnCll
MgClt
MnC1,

                TABLE IV
Replacement of deozynucleoside triphosphates by various base
        analogues in polymerase reaction

Reaction mixtures were those of the standard assay with 40
mpmoles of calf thymus DNA activated with E. coli DNA phos-
phatase (see "Treatment of Primers with Nucleases") and 1.3
pg of Fraction V.  Incubations were carried out for 30 minutes
at 37".  Control values were measured as rates of radioactive
deoxynucleotide incorporation into DNA in the presence of the
four usual nucleoside triphosphates but in the absence of the
analogue.  Values were 0.34 mpmole of 'C-dTMP for Columns
2,3, and 4, and 0.19 mmole of 1'C-dCMP for Column 1.  Omission
of one of the deoxynucleoside triphosphates gave 2 to 6% of the
control values.  Values in parentheses were obtained with E.
coli polymerase (0.28 pg of hydroxylapatite fraction) in potassium

1 phosphate buffer, pH 7.4.

dCTV
dCT"P
rCT'2P
rCTSlp

Analogue used in form of
deoxynucleoside triphosphates

Uracil. ...................
5-Bromouracil ............
5-Fluorouracil ............
5-Hydroxymethylcytosine .
5-Methylcytosine .........
N-Methyl-5-fluorocytosine
5-Fluorocytosine . .........
5-Bromocytosine . .........
Hypoxanthine ............
Xanthine. ................

Deoxynucleoside tri hosphate replaced by
    anafoguc

dTTP IdATPI dCTF' I dGTP

69 (61)
91 (93)
28 (27)

6
0
5
3
3
5
7

~

3
2
4
6
5
3
6
6

4

69 (50)

            TABLE V
Attempt to incorporate a ribonucleoside triphosphate

As in "Methods," assay mixtures contained 20 pmoles of Tris-
maleate-KOH, pH 8.2, 10 mpmoles each of dATP, dTTP, and
dGTP, 40 mpmoles of calf thymus DNA, 10 mpmoles of dCTuP
(2 X 106 c.p.m. per pmole) or 100 mpmoles of rCTazP (2 X 106
c.p.m. per pmole), 2 pmoles of MgCL or 0.05 pmole of MnCl,,
and 2.84 pg of Fraction VI-1 enzyme.  Incubation was performed
for 30 minutes at 37'.

Metal

Substrate

Incorporation of dCMaP
   or rCM*zP

mpmole

0.317
0.314
<0.01=
<0.014

(I Prolonged incubation (120 minutes) or the use of 0.5 pmole of
MnCll and 350 wmoles of calf thymus DNA did not produce
any detectable increase of incorporation of rCMaaP.

                TABLE VI
Comparison of activities of various primers for B. subtilis and
           E. coli polymerases

Pretreatment of dAT copolymer and calf thymus DNA was
carried out as described in "Methods."  The pretreated dAT,
6 mpmoles, was used as primer in the standard incubation mixture,
with 0.024 pg of B. subtilus polymerase (Fraction VI-2) or 0.007
pg of E. coli polymerase (hydroxylapatite fraction).  The pre-
treated calf thymus DNA, 40 moles, was used as primer in the
standard incubation mixture, with 0.47 pg of B. subtilis polymer-
ase or 0.28 pg of E. coti polymerme.  The rates of incorporation
of nucleoside monophosphates for nontreated dAT primer were
11.4 and 53.4 mpmoles per pg of Fraction VI-2 and E. coli polymer-
ase, respectively, and were set at 100%.

I Relative priming activity

Primer and pretreatment

B. subtilis
Fraction VI-2

dAT copolymer

None. ............................
E. coli DNA phosphatase.. .......
E. coli endonuclease.. ............
Pancreatic DNase, ...............
Micrococcal endonuclease. ........
Micrococcal endonuclease fol-

lowed by E. coli DNA phos-
phatase. ......................

None. ............................
Heated at 100" and quickly cooled.
E. coli DNA phosphatase.. .......
E. coli endonuclease. .............
Heat denaturation of E. coli endo-
nuclease-treated DNA. .........
Pancreatic DNase. ...............
Micrococcal endonuclease. ........
Micrococcal endonuclease fol-

lowed by E. coli DNA phos-
phatase. .......................

Thymus DNA

100
312
344
366
35

268

6.7
8.6

31
70


74
28

2.5

49

  E. coli
hydroxylapatite

100
329
250

16

180

2.8
3.6

37
63


73

1.4

45

priming activity compared with native DNA under the condi-
tions used in this experiment.
Net Synthesis of DNA-With Fraction VI-2 and B. subtilzj
native DNA as primer, 2.5 replications (2.5 parts synthesized
DNA per unit of primer) were obtained in 25 hours.  With heat-
denatured DNA as primer, the initial rate of synthesis was simi-
lar to that with native primer; however, the rate decreased after
0.15 replication and in 25 hours of incubation only 0.64 replica-
tion was observed (Fig. 5).  The thymine to cytosine ratio of
the replication product was checked at intervals with two identi-
cal reaction mixtures, one containing "CdTTP and the other
"PdCTP.  The ratio was 1.57 at 0.08 replication and 1.35 be-
tween 0.15 and 1.0 replication; the thymine to cytosine ratio in
B. subtilis is 1.3.

dAT Synthesk, de Novo and Primed-B. subtilis polymerase
carries out extensive replication of dAT polymer when primed
with dAT made by E. coli polymerase.  As shown in Fig. 6,
70% of the added substrate was made acid-precipitable in 3
hours as determined by W-dTMP incorporation (6.2 replica-
tions).  Unlike the case with even the most purified E. coli poly-
merase fractions, this product was not degraded after prolonged
incubation (more than 10 hours).  Without added primer, syn-
thesis of dAT de mvo was observed after a 3- to &hour lag

266

250

B
; 200
IJ

L

Q)

b150


0

c
0)


0)
Q


0
t

€

--

a

t

100

h

.-

50

t

c
x
v,

0

Enzymatic Synthesis of DNA. XV Vol. 239, No. 1

1

I I 114

5 10 15 25

Hours

FIG. 5. The course of replication of native and heated DNA.
The reaction mixture contained (in 1.5 ml) 100 pmoles of Tris-
maleate-KOH, pH 8.2,lO pmoles of MgClz, 1.5 pmolee of 2-mercap-
toethanol, 0.125 amole each of dATP, dCTP, dGTP, and W-
dTTP (2 X 10' c.p.m. per pmole), 62 wmoles of B. subtilis DNA,
50 pl of enzyme diluent, and 18.8 pg of Fraction VI-2.  The reac-
tion wm carried out at 37".  At intervals, 50-p1 aliquots were
taken for measurement of acid-precipitable radioactivity.  At the
time shown by the arrow, an additional 2pg of enzyme were added.

period.  The copolymeric nature of the product of synthesis ok
novo (with dATP and l*C-dTTP) was determined by hydrolysis
with diphenylamine-formic acid (12, 17).  Acidic depurination
with subsequent scission of the polynucleotide chain at the de-
purinated sites results in acid solubilization of the residues of the
hydrolyzed polynucleotide.  a*P-DNA isolated from E. coli and
dGdC and dAT polymers made by E. coli polymerase were
also subjected to the same treatment to serve as standards.
IJnder the conditions which result in complete acid solubiliza-
tion of E. coli DNA or dAT copolymer, there is no significant
acid solubilization of pyrimidine-labeled dGdC polymer.  The
dAT polymer synthesized by B. dtdis polymerase was rendered
acid-soluble to the extent of 98% (Table VII), suggesting that
it is a copolymer of deoyxthymidylate and deoxyadenylate units
like that formed de m by E. coli DNA polymerase.

DISCUSSION

The finding that the DNA polymerase from B. subtilis may be
freed of virtually all nuclease activity contrasts with the per-
sistence of such activity in the purified E. coli and T2 phage
polymerase preparations.  It is clear, therefore, that even if an
exonucleolytic activity should prove to be an intrinsic capacity

of a pure DNA polymerase, this is not an essential property of
all DNA-synthesizing enzymes.  On the other hand, it should
not be assumed that any nuclease activity associated with poly-
merase is necessarily unrelated to DNA replication.  For exam-
ple, the DNA phosphatase-exonuclease is found in both the B.
subtilis and E. coli polymerase preparations and separated by
hydrosylapatite chromatography only in the last step of the
purification procedure.  This enzyme, by removing phosphoryl
groups attached to the 3'-hydroxyl termini of DNA chains, con-
verts inhibitory DNA to an active primer.  Such 3'-phosphoryl
groups may occur as a result of endonuclease action or be pro-
duced by shearing forces; for some reason, such groups do occur
in native DNA isolated by usual methods, as may be inferred
from the many fold increase in priming activity of the DNA after
treatment with the DNA phosphatase-exonuclease.

The DNA primer requirements of the B. subtilis polymerase
are remarkably similar to those of the E. coli polymerase.  The
reaction rate with dAT is about 20 times faster than with native
DNA, and the rates with native and single stranded DNA are
about equal.  The inhibitory effect of 3'-phosphoryl ends and
the activating effect of 3'-hydroxyl ends are also .seen with both
enzymes.  By contrast, the T2 phage-induced polymerase (2)
and the enzyme from calf thymus (25) require single stranded
DNA; with the T2 enzyme, dAT is a relatively poor primer and
the production of additional 3'-hydroxyl ends in native DNA
does not activate its priming capacity.  It should be emphasized
that these comparisons are based on rather gross characteristics
of the DNA preparations and that the finer details of the DNA
may reveal important distinctions between the B. aubti2is and
E. coli polymerases.

Studies of the utilization of nucleoside triphosphates by the
several bacterial polymerases reveal differences that reflect on
the mechanism of replication.  The relative rates of reaction
with a number of base analogues of the deoxynucleoside triphos-
phates are remarkably similar with the B. mbtilis, E. coli, and
T2 phage-induced polymerases.  From this it seems likely that
the determination of which bases are selected in assembling the
DNA chain is made by the template rather than by the enzyme.
On the other hand, the capacity of the E. wli polymerase but
not the B. subtilis enzyme to utilize ribonucleoside triphosphates
must be a specific attribute of the enzyme.  These results fortify
the hypothesis that the sequential ordering of the bases is primar-
ily due to directions by the DNA template.

One of the purposes in purifying the DNA polymerase from
B. subtilis was to discover a source of the enzyme which might
prove more amenable to purification than is E. coli.  This was
disappointing, in that the B. subtilis enzyme was far more dS-
cult to obtain free of nucleic acids, wrag less stable, and was re-
covered in smaller yields.  The specific activity of the most
purified B. subtilis preparation was only about one-tenth that of
the best E. coli fractions.

SUMMARY

1. A deoxyribonucleic acid polymerase has been purified from
Bacillus subti& and its properties compared to the polymerase
from Escherichia coli.
2. Unlike the E. coli polymerase, the purified preparation has
barely detectable exonucleolytic activity and, as with the E. wli
enzyme, no measurable endonuclease activity.  Hydroxylapatite
chromatography, the last step in the purification procedure,

January 1964 T. Okamki and A. Kornberg 267

--

C.).rn.

4,278

12,050

2,733

1,525

I I I I

c.).rn.


93

144


29

1,444

0

Unprirned

Primed

5

Hours

FIG. 6. Synthesis of dAT copolymer, de novo and primed.  A
3-ml reaction mixture contained 0.5 pmole each of dATP and W-
dTTP (2 X IO6 c.p.m. per pmole), 200 pmoles of Tris-maleate-KOH
buffer, pH 8.2,20pmoles of MgClI, 3pmoles of 2-mercaptoethanol,
and 0.2 pl of enzyme diluent.  For primed synthesis, 60 mpmoles
of dAT polymer made by E. coli polymerase were added, and for

        TABLE VI1
Demonstration that dA T is a copolymer

dAT polymer was synthesized in a nonprimed reaction from
"C-dTTP and unlabeled dATP with B. subtilis polymerase
(Fraction VI-2) and then hydrolyzed by the method of Burton
and Petersen (17). Each polymer (0.5 ml, 0.2 pmole per ml,
106 c.p.m. per ml) was incubated with 1 ml of freshly prepared
3% diphenylamine in formic acid for 17 hours at 30'.  A 0.3-ml
,aliquot was used for the assay of acid-insoluble radioactivity
 before and after hydrolysis as described previously (12).
 labeled preparations of E. coli DNA and of dGdC and I'C-
labeled dAT polymers made by E. coli polymerase were treated
'in the same manner to serve as controls.

Preparation

dAT made by B. subtilis poly-
merase. ......................
E. coli DNA.. ..................
dAT made by E. coli poly-
merase. ......................
dGdC made by E. coli poly-
merase. ......................

Xstribu.
tion of
label

-p*T-

-0

-p*T-

-*pc-

Acid-preeipitoble radi~
activity after hydrolysis

0 time I 17 hrs

-

%


2.2
1.2


1.1

95

a Uniform.

identifies and separates a DNA phosphatase-exonuclease (14).
just as in the case of E. coli.
3. The priming activities of deoxyadenylate-deoxythymi-
dylate copolymer (dAT copolymer) and of native DNA, and the
effect on priming of heating and deoxyribonuclease treatment of
DNA are essentially the same with the purified B. subtilis and
E. coli polymerases.  Replication of native DNA and synthesis

20 I' 30

both primed and unprimed reactions, 35 pg of Fraction VI-2 were
present. Aliquots (30 pl) were taken at intervals, and acid-pre-
cipitable count was determined.  After 20 hours, 1 ml of reaction
mixture was taken from the unprimed reaction for chemical deter-
mination of nucleotide sequence.

de novo of dAT copolymer proceed readily, and in the latter case
to the nearly complete utilization of the deoxynucleoside tri-
phosphate substrates provided. Because of the absence of
nucleases, the synthesized dAT remains intact even after pro-
longed incubation.
4. Various base analogues of the naturally occurring deoxy-
nucleoside triphosphates substitute in the B. subtilis polymerase
system in a fashion quantitatively identical with that determined
in the E. coli system.  These and other data reaffirm that
adenine to thymine and guanine to cytosine base pairing between
template and substrate, rather than enzyme specificity, deter-
mines the sequential polymerization of nucleotides.

REFERENCES

1. RICHARDSON, C. C., SCHILDKRAUT, C. L., APOSHIAN, H. V.,
  AND KORNBERQ, A., J. Biol. Chem., 199, 222 (1964).
2. APOSHIAN, H. V., AND KORNBERO, A., J. Biol. Chem., 297,519

  (1962).
3. LEHMAN, I. R., BESSMAN, M. J., SIMMS, E. S., AND KORNBERG,
  A., J. Biol. Chem., 233, 163 (1958).
4. OKAZAHI, R., AND KORNBERG, A,, J. Biol. Chem., 899, 269
  (1964).

  KORNBERG, A., Proc. Natl. Acad. Sei. U. S., 44, 641 (1958).
6. HURWITZ, J., J. Biol. Chem., 25L. 2351 (1959).
7. BESSMAN,M. J., LEHMAN, I. R., ADLER, J., ZIMMERMAN, S. B.,

  SIMMS, E. S., AND KORNBERG, A., Proc. Natl. Acad. Sei.

8. LEHMAN, I. R., J. Biol. Chem., a36. 1479 (1960).
9. KAY, E. R. L., SIMMONS, N. S., AND DOUNCE, A. L., J. Am.

10. MARMUR, J., J. Molecular Biol., S, 208 (1961).

11. SCBACHMAN, H. K., ADLER, J., RADDING, C.M., LEHMAN, I. R.,
  AND KORNBERG, A., J. Biol. Chem., 236, 3242 (1960).
12. RADDING, C. M., JOSSE, J., AND KORNBERG, A., J. Biol. Chem.,

  197, 2869 (1962).
13. LEHMAN, I. R., Roussos, G. G., AND PRATT, E. A., J. Bid.
  Chem., 297, 819 (1962).

5. ADLER, J., LEHMAN, I. R., BESSMAN, M. J., SIMMS, E. s., AND

u. s., 44, 633 (1958).

Chem. SOC., 74, 1724 (1952).

268 Enzymatic Synthesis of DNA. XV Vol. 239, No. 1

14. RICHARDSON, C. C., AND KORNBERQ, A,, J. Biol. Chem., 139,

15. KORNBERQ, S. R., ZIMMERMAN, 8. B., AND KORNBERQ, A., J.

16. JOSSE, J., AND KORNBERQ, A., J. Bid. Chem., a57, 1968 (1962).

17. BURTON, K., AND PETERSEN, G. B., Biochem. J., 76.17 (1960).

18. LOWRY, 0. H., ROSEBROUQH, N. J., FARR, A. L., AND RANDALL,
19. ALBERTSSON, P.-B., Partition of cell particles and macromole-
20. ALBERTSSON, P.-A., Arch. Biochem. Biophys., Suppl. 1, 264

242 (1964).

Biol. Chem., 136. 1487 (1961).

R. J., J. Biol. Chem., 193. 265 (1951).

cules, John Wiley and Sons, Inc., New York, 1960.

(1962).

21. RAZZELL, W. E., AND KHORANA, H. G., J. Biol. Chem., 234,
  2114 (1959).
22. RICHARDSON, C. C., SCHILDHRAUT, C. L., APOSHIAN, H. V.,
  KORNBERQ, A., BODYER, W., AND LEDERBERQ, J., in H.
  VOQEL (Editor), Symposium on informational macromole-
  cules, Academic Press, he., New York, 1963, p. 13.
23. KORNBERQ, A., Enzymatic synthesis of DNA, Ciba lecfurea in
  microbial biochemistry, John Wiley and Sons, Inc., New
  York, 1961.
24. BERQ, P., FANCHER, H., AND CHAMBERLIN, M., in H. VOQEL
  (Editor) , Symposium on informational macromolecules, Aca-
  demic Press, Inc., New York, 1963, p. 467.

25. BOLLUM, F. J., J. Biol. Chem., 236, 2399 (1960).