Tue JOURNAL oF BIOLOGICAL CHEMISTRY
Vol. 248, No. 3, Issue of February 10, pp. 627-638, 1968
Printed in U.S.A.

Enzymatic Synthesis of Deoxyribonucleic Acid

XXV. PURIFICATION AND PROPERTIES OF DEOXYRIBONUCLEIC ACID POLYMERASE INDUCED

BY INFECTION WITH PHAGE 7T4+ *

(Received for publication, August 25, 1967)

Merunran Gouiian,{ Zottan J. Lucas,§ anp ArtHuR KornBere
From the Department of Biochemistry, Stanford University School of Medicine, Palo Alto, California 94304

SUMMARY

DNA polymerase induced in Escherichia coli by phage T4
am N82 has been prepared in a highly purified state. The
molecular weight of about 112,000 for this enzyme compares
with a value of 109,000 for the host (Z. coli) DNA polymerase.
The amino acid compositions of the two enzymes, however,
are distinctive, especially the half-cystine value of 15 residues
for the phage enzyme and 3 for the bacterial enzyme. The
purified phage enzyme contains an exonuclease activity
which is physically inseparable from the polymerase activity
but is suppressed or obscured by conditions favorable for
replication. The polymerase requires a single stranded
template, with a free 3’-hydroxyl terminus, which is replicated
to an extent that approaches but never exceeds the DNA
input. The product is a helical structure in which the newly
synthesized strand is covalently linked from its 5’-terminus
to the 3’-terminus of the template. Various features of the
replication process and the product suggest a mechanism in
which the 3’-terminus of the template first loops back upon
itself and then serves as the priming end for replication of the
remainder of the template. The evidence indicates that the
phage polymerase, unlike the E. coli enzyme, is unable to

initiate new strands or utilize a fully helical DNA as template. -

 

The increase in rate of deoxyribonucleic acid synthesis as-
sociated with infection by virulent coliphages (1) was shown, in
the case of phages T2, T4, and T5, to be accompanied by the

* Supported in part by grants from the National Institutes of
Health, United States Public Health Service, and the National
Science Foundation. Preliminary reports of this work have
appeared (Lucas, Z. J., Fed. Proc., 24, 286 (1965); Goutian, M.,
Fed. Proc., 26, 396 (1967)).

t Special Fellow, National Institutes of Health. Present ad-
dress, Department of Medicine, University of Chicago, Chicago,
Illinois 60637.

§ American Cancer Society Fellow. Present address, Depart-
ment of Surgery, Stanford University School of Medicine, Palo
Alto, California 94304.

appearance of a new DNA polymerase (2-4). Recent work
with conditional lethal mutants has justified the earlier pre-
sumptions that this enzyme, unique to phage infection, is the
product of a phage gene and is essential for synthesis of phage
DNA (5, 6).

In the present study, we have examined the properties which
distinguish the phage-induced polymerase from the host enzyme:
the requirement for a single stranded template and the limited
extent of replication. Such studies, we anticipate, will extend
our understanding of DNA synthesis in general and phage DNA
replication in particular.

A procedure is presented for preparation of highly purified
phage T4 polymerase, followed by a description of the physical
and enzymatic properties of the enzyme. Studies of the nature
of the DNA synthesized suggest a mechanism for the action of
phage T4 polymerase in vitro.

EXPERIMENTAL PROCEDURE

Materials

Unlabeled deoxyribonucleoside triphosphates were purchased
from Calbiochem or Pabst and purified by chromatography on
DEAE-Sephadex (A-25) with a triethylammonium bicarbonate
gradient. a-*P-Labeled deoxyribonucleoside triphosphates were
prepared as described previously (7).

3‘H-DNA from M13 phage was isolated as described by Mitra
et al. (8). The linear form of M13 DNA was prepared by limited
exposure to Neurospora endonuclease (0.05 unit per ymole! of
DNA for 20 min). Under conditions described by Linn and
Lehman (9), approximately two-thirds of the DNA circles were
cleaved, and the linear forms were obtained by fractionation in
an alkaline sucrose gradient (8). *%H-labeled T7 DNA and T4
DNA were prepared by the procedure of Richardson, Inman, and
Kornberg (10); *H-labeled T7 DNA was digested with micro-
coccal nuclease by the method of Richardson, Schildkraut, and
Kornberg (11). A single stranded form of T7 DNA was prepared
by digestion with A-exonuclease to the limit of 50% acid solubil-
ity as described by Little (12), followed by alkaline denaturation
at 50°. 3H-Labeled \ DNA was a gift from Dr. Walter Doerfler.

1 Molarity of DNA refers to concentration of nucleotide resi-
dues.

627
628

Salmon sperm and calf thymus DNAs were isolated by the
method of Kay, Simmons, and Dounce (13); partial pancreatic
digestion of this DNA was carried out by the procedure of
Oleson and Koerner (14). Preparation of Bacillus subtilis DNA
was based on the method of Massie and Zimm (15). °H-dAT
copolymer? was made in a primed incubation with 5H-dATP by
the method of Schachman et al. (16). Poly dC was isolated by
CsSQO, density gradient sedimentation (17) from dI:dC that had
been prepared in a primed incubation (17). The poly dC was
heated at pH 3 for 5 min at 95° to reduce contamination with
small fragments of dI. Poly dA, made with calf thymus terminal
deoxynucleotidyltransferase (18), was a gift from Dr. F. J.
Bollum. Unlabeled and “P-labeled Escherichia coli DNAs were
isolated according to the method of Lehman (19).

The following enzymes were used: exonuclease III, prepared
by the method of Richardson and Kornberg (20); exonuclease I
(hydroxylapatite fraction), Lehman and Nussbaum (21); d-
exonuclease, Little, Lehman, and Kaiser (22); Neurospora
endonuclease, Linn and Lehman (23); E. coli alkaline phosphatase,
Worthington Biochemical Corporation; 5’-nucleotidase, Lehman,
Roussos, and Pratt (24); and micrococcal nuclease, Cunningham,
Catlin, and Privat de Garilhe (25).

DEAE-cellulose and phosphocellulose were products of Bio-
Rad. Hydroxylapatite and Sephadex G-25 were purchased
from Clarkson and Pharmacia, respectively. Streptomycin
sulfate was the gift of Merck, Sharp and Dohme.

Methods

Assay of Polymerase—The assay measures the conversion of
2P.labeled deoxyribonucleoside triphosphates to an acid-in-
soluble product. The incubation mixture (0.3 ml) contained 67
mM Tris buffer (pH 8.8), 6.7 mma MgCh, 10 mm 2-mercapto-
ethanol, 16.6 mm ammonium sulfate, 6.7 um EDTA, 0.33 mu
sodium fluoride, 167 wg of bovine plasma albumin per ml, 0.2
mm alkali-denatured salmon sperm DNA, 0.033 mm concentra-
tion each of dCTP, dATP, dTTP, and dGTP (one of which was
labeled in the @ position with “P, 1 to 5 x 108 epm per pmole),
and 0.04 to 0.40 unit of enzyme. Enzyme dilutions were made
in 0.05 m Tris (pH 7.5), 0.1 M ammonium sulfate, bovine plasma
albumin (1 mg per ml), and 0.01 m 2-mercaptoethanol, After a
30-min incubation at 37°, the reaction was stopped by addition
of 0.1 ml of a carrier solution containing 0.1 m sodium pyrophos-
phate, 0.1 m EDTA, and 100 yg of calf thymus DNA per ml.
Acid-insoluble material was collected and washed on glass
filters (26), and radioactivity was measured in a liquid scintilla-
tion counter. Sodium fluoride, which served to inhibit deoxy-
cytidine triphosphatase activity, was omitted in assays of
fractions beyond the phosphocellulose step in the purification
procedure. A unit of enzyme activity‘ is defined as the amount
catalyzing incorporation of 10 mymoles of total nucleotide into
an acid-insoluble product during the period of incubation. The
assay was linear in the range of 0.04 to 0.40 unit of enzyme.

2 The abbreviations used are: dAT copolymer, copolymer of
deoxyadenylate and deoxythymidylate; poly dC, polydeoxycy-
tidylate; poly dA, polydeoxyadenylate; dI:dC, homopolymers
of the deoxyribonucleotides of hypoxanthine and cytosine, hy-
drogen-bonded together.

3 All enzyme incubations described in this paper were conducted
at 37° unless stated otherwise.

‘This definition of a polymerase unit differs from that used
previously (2) in that it is based on total rather than labeled nu-
cleotide.

Enzymatic Synthesis of DNA. XXV

Vol. 243, No. 3

DNA Synthesis—Incubations for extensive synthesis were the
same as for assay except that the concentration of each deoxy-
ribonucleoside triphosphate was increased to 0.3 mM, and the
amounts of enzyme and template’ were varied. Aliquots for
assay of the extent of reaction were added to 0.1 ml of carrier
solution, and acid-insoluble radioactivity was determined as in
the assay procedure.

Nuclease Assays—Assay of the nuclease associated with T4
polymerase was ordinarily carried out by the method described
by Oleson and Koerner (14); the substrate was salmon sperm
DNA partially digested by pancreatic DNase. Appearance of
acid-soluble material was followed spectrophotometrically.
One unit refers to that amount of enzyme releasing 10 mumoles
of acid-soluble nucleotide in 20 min at 30°. In those assays in
which nuclease activity toward ®P-labeled E. coli DNA was
determined, the conditions were the same except that incubation
was conducted at 37° instead of 30°, and the appearance of acid-
soluble radioactivity was measured.

Other Methods—Amino acids were analyzed by the methods of
Moore and Stein (27), modified for greater sensitivity with a
long path flow cell and range card (28). Analyses were carried
out in a Spinco model 120B amino acid analyzer with Spinco
AAI5 and PA35 resins in the 58- and 10-em columns, respectively.

Equilibrium centrifugation experiments were carried out in a
Spinco model E analytical ultracentrifuge equipped with Rayleigh
interference optics.

Sucrose gradient sedimentation was run in a 5 to 20% gradient
according to the procedure of Martin and Ames (29). The
material was collected directly on Whatman No. 83MM paper
discs, which were then washed with 5% trichloracetic acid and
ethanol before radioactivity measurements. Sedimentation
coefficients were determined from an internal DNA reference
marker,

Polyacrylamide gel electrophoresis employed 7.5% gels
(0.5 X 5 cm) and the conditions for electrophoresis and elution
described by Falaschi and Kornberg (30).

Alkaline denaturation of DNA was ordinarily carried out for
5 min in 0.15 n NaOH at room temperature, followed by chilling
and neutralization with a predetermined amount of HCL.

Protein was determined by the method of Lowry e¢ al. (31)
after precipitation with cold trichloracetic acid. The standard
was bovine plasma albumin (Armour).

RESULTS

Purification of Enzyme

Except as noted, all steps were carried out at 0-4°. Centrif-
ugations were done at 15,000 x g for 15 mm. The purification
procedure is summarized in Table I.

Preparation of Extract—E. coli B cells were grown from 1%
inocula in 100-liter quantities, at 37°, with aeration. The
medium contained the following, per liter: KzHPQ,, 13.2 g;
KH,PO,, 3.26 g; Difco Casamino acids, 10 g; p1L-tryptophan,
0.1 g; cysteine HCl, 0.13 g; (NH4)SO,, 2.6 g; MgSO,, 0.25 g;
glucose, 10 g; Fe(NH,)2(SOu.)2, 0.4 mg. At an optical density
(590 my) of 1.0, the culture was infected with 3 x 10! T4 am
N82 per liter (multiplicity, 4). After 75 min of further incuba-

5 The DNA required for activity of the enzyme will be referred
to as template, although it serves a primer function as well.

6 The T4 mutant was kindly provided by Dr. Robert Edgar.
This amber mutant was used, with the nonpermissive host #.
Issue of February 10, 1968

tion, the culture was chilled and harvested by centrifugation,
and the cell paste was stored at ~20° until the next step.

Pooled cell paste from two 100-liter cultures, totaling 552 g,
was blended with glass beads in glycylglycine buffer (0.05 m, pH
7.0)-2 mm EDTA-2 mm reduced glutathione as described pre-
viously (26), and after centrifugation yielded 2300 ml of extract
(Fraction I, Table 1).

Streptomycin Precipitation—Yo Fraction 1 were added, with
stirring, 5100 ml of 2 mm reduced glutathione-2 mm EDTA,
followed by 2200 ml of 5% streptomycin sulfate during a 30-min
interval. After additional stirring for 20 min, the precipitate
was collected by centrifugation and suspended in 1290 ml of po-
tassium phosphate buffer, 0.05 M, pH 7.0, containing 2 mm re-
duced glutathione (Fraction IH, Table I).

Autolysis—Fraction II was made 3 mm in MgCl, and incubated
at 37° until 95% of the ultraviolet-absorbing material at 260 my
had become acid-soluble (26) (approximately 90 min). ‘The autol-
ysate was then rapidly chilled to 0°, and the supernatant (Frac-
tion III, Table I) was separated by centrifugation.

Ammonium Sulfate Fractionation—To each liter of Fraction III
were added 218 g of ammonium sulfate with stirring during a
30-min interval. After an additional 30-min stirring period, the
precipitate was removed by centrifugation. Additional am-
monium sulfate, 78 g per liter of supernatant, was added with the
same procedure. The precipitate was collected and dissolved in
120 ml of potassium phosphate buffer, 0.05 mM, pH 7.4, containing
2mm EDTA and 2 mm reduced glutathione. This solution was
passed through a column of Sephadex G-25 (20 cm? x 50 em)
equilibrated with potassium phosphate buffer, 0.05 m, pH 6.5, and
0.01 m 2-mercaptoethanol, and followed by the same buffer.
Active fractions were pooled (Fraction IV, Table I).

Phosphocellulose Chromatography—Fraction IV was applied to
a column of phosphocellulose (10 cm? x 20 cm) that had been
equilibrated with potassium phosphate buffer, 0.05 um, pH 6.5,
and 0.01 mM 2-mercaptoethanol. This was followed by a linear
gradient of 3 liters of potassium phosphate buffer, pH 6.5, with
limits of 0.05 u and 0.4 mM, containing 0.01 m 2-mercaptoethanol.
The flow rate was 2.5 ml per min, and 40-ml fractions were col-
lected. The peak of enzyme activity emerged after passage of
approximately 2 liters of the gradient solution. Fractions with
specific activity of greater than 12,000 were pooled (Fraction V,
Table I).

DEAE-cellulose Chromatography—The volume of Fraction V
was reduced from 220 ml to approximately 20 ml by dialysis
against solid sucrose; fresh changes of dry sucrose were applied
as necessary to achieve the volume reduction. The sample was
then dialyzed against potassium phosphate buffer, 0.02 m, pH
7.4, and 0.01 mM 2-mercaptoethanol and applied to a column of
DEAE-cellulose (1.6 em? x 25 cm) that had been equilibrated
against the same buffer. A linear gradient of 800 ml of potas-
sium phosphate buffer, pH 6.3, was applied, with limits of 0.02
M and 0.15 M, containing 0.01 m 2-mercaptoethanol. The flow
rate was 0.6 ml per min, and 5-ml fractions were collected. The
peak of protein and polymerase activity appeared after passage
of approximately 300 ml of gradient solution. The active frac-
tions were pooled, concentrated by vacuum ultrafiltration (20),
and dialyzed against 0.1 m Tris-HCl buffer, pH 7.5, and 1 mm
reduced glutathione (Fraction VI, Table I). Fraction VI,

 

coli B, because specific activity of crude extracts is approximately
5 to 10 times the level obtained with wild type phage T4 or T2
(2, 3).

M. Goulian, Z. J. Lucas, and A. Kornberg

 

 

629
TaBLe I
Purification of enzyme
Fraction Units Protein Specie vena
x 10-8 mg/ml units/mg %
T. Extract. ............... 30 16.2 78 100
Il. Streptomycin........... 38 14.8 200 | 127
III. Autolysis............... 26 3.1 370 87
IV. Ammonium sulfate..... 20 29.5 530 67
V. Phosphocellulose....... ll 0.33 | 15,300 37
VI. DEAE-cellulose. ....... 7.3 0.67 | 33,000 24
VII. Hydroxylapatite........ 1.22 0.86 | 31,000 8

 

@ Only 10.6 mg of the 22 mg of Fraction VI were purified on
hydroxylapatite; the yield for this step is corrected to apply to
the total amount of Fraction VI, but includes the 16% loss in
activity of the latter after storage in liquid nitrogen.

 

 
 
   

 

T T T 1 T T T
at
Bost Pea po
E EC2
> 28
= Be!
5 oo
~1200 [= 5¢ 60
9 Z£E >
a oS 3
o og +
$ IE g.
O as 3
2 2 Polymerase .
< 800+ Q +1405
roy a” >
‘ 7
9
€ 400{- 20
=
a

 

 

_ 40
Fraction number

Fic. 1. Hydroxylapatite chromatography of Fraction VI.
The DEAE-cellulose fraction (Fraction VI) (10.6 mg) was eluted
from hydroxylapatite as described in the text. Each fraction
contained 20 ml. Nuclease assays were performed by the method
of Oleson and Koerner (14).

stored in liquid nitrogen, retained 84% of the original activity
after 10 months.

Hydroxylapatite Chromatography’—Fraction VI (15.8 ml),
from liquid nitrogen storage, was diluted with an equal volume
of 0.02 w 2-mercaptoethanol and adjusted to pH 6.5 with 0.05 m
KH,PO, containing 0.01 m 2-mercaptoethanol. The diluted
fraction was applied to a column of hydroxylapatite (4.5 cm? x
15 cm) that had been washed with 2500 ml of 0.05 m potassium
phosphate buffer, pH 6.5, and 0.01 m 2-mercaptoethanol. The
column was washed with 100 ml of the same buffer, and then a
linear gradient of 1500 ml of potassium phosphate, pH 6.5, with
limits of 0.05 m and 0.3 M, containing 0.01 m 2-mercaptoethanol,
was applied. The flow rate was 0.8 ml per min. Protein and
enzyme activity appeared in a single peak midway in the gradient
(Fig. 1). Tubes with greater than 300 units per ml were pooled,

7 Although the hydroxylapatite step resulted in a large loss of
activity without increase in specific activity, it was included
because it removed traces of residual endonuclease activity still
detectable in Fraction VI.
630

of

 

0.4

Absorbancy
°
ot

9
w

 

 

 

1 t !
260

| 1

 

Zao
Wave length, mu

240

Fig. 2. Ultraviolet absorption spectrum of the enzyme. The
spectrum of Fraction VII was determined in a Zeiss PMQ II
spectrophotometer. Light path was 1 cm; protein concentration
was 0.23 mg per ml. The solvent for the neutral spectrum was
9 mm NH.HCOs, pH 8.0, and for the alkaline spectrum it was 9
mm NH.HCO;-0.1 n NaOH.

concentrated by dialysis against solid sucrose, and dialyzed ex-
haustively against 50% glycerol-0.2 mM potassium phosphate,
pH 6.5, containing 0.01 m 2-mercaptoethanol. This material
(Fraction VII) had a protein concentration of 0.86 mg per ml and
when stored at —20° retained full activity for 10 months. Ex-
cept where stated otherwise, this fraction was used for the ex-
periments described below.

Physical Properties

Ultraviolet Spectrum—The hydroxylapatite fraction (Fraction
VII) exhibited a characteristic protein absorption spectrum at
neutral and alkaline pH (Fig. 2). The absorption at 280 mu of a
solution containing 1 mg per ml (in 0.01 m NH,HCO,, pH 8;
l-em light path) was 1.40 when using a protein concentration
calculated from the amino acid analysis and corrected for light
scattering (32). The ratio of absorbance at 280 mp to that at
260 my was 1.71 after a similar correction.

Amino Acid Composition—The amino acid composition,
summarized in Table II, shows several differences from that of
the E. coli polymerase. The T4 enzyme has 15 half-cystine
residues, whereas the EF. coli enzyme has 3,’ a point of particular
interest in view of the great sensitivity of the phage-induced
enzyme to sulfhydryl] inactivation, to be cited below. Other
| notable differences are higher serine and isoleucine contents in

8 T.M. Jovin, P. T. Englund, and L. L. Bertsch, personal com-
munication.

Enzymatic Synthesis of DNA. XXV

Vol. 248, No. 3

the phage enzyme, and lower values for alanine, leucine, and
phenylalanine.

Molecular Weight—Equilibrium centrifugation was carried
out by the high speed procedure of Yphantis (85). The molecu-
lar weight obtained by this method emphasizes low molecular
weight components in a multicomponent system and does not
represent an average molecular weight. However, the plot of
log of concentration with respect to the square of distance from
the center of rotation gave a straight line and thus showed no
evidence for more than one component in the region of the cell
that could be analyzed (Fig. 3). The slopes derived from two
separate experiments gave molecular weights of 110,000 and
114,000, based on a partial specific volume of 0.734 calculated
from the amino acid composition (36). This molecular weight
may be compared with a value of 109,000 for the Z. colt po-
lymerase (82).

Homogeneity—A high degree of purity of the hydroxylapatite
fraction is indicated by three findings: the relatively constant
specific activity determined across the protein peak in the
hydroxylapatite chromatogram (Fig. 1), a single protein band
obtained on electrophoresis of 20 yg of Fraction VII in poly-
acrylamide gel (analysis of 15 wg of Fraction VI under the same
conditions revealed several minor components), and the linear
function (of log of concentration with respect to distance squared)
in the Yphantis equilibrium sedimentations (Fig. 3).

Associated Nuclease Activity

The presence of endonuclease activity was tested for by in-
cubating *H-labeled M13 DNA or *H-labeled T7 DNA with 5

TasLe II
Amino acid composition
Duplicate samples were hydrolyzed in 6 n HCl with a crystal
of phenol (32), after exhaustion of air, for 24 or 78 hours at
110°. Half-cystine was determined as cysteic acid after oxidation
with performic acid (33).

 

 

 

Amino acid Content
moles/112,000 g
Lysine... 0... eee ee eee 85
Histidine. 2.0.0.0... cece cee ees 16
Arginine... 0.0... cece cece ee eee 46
Half-cystine... 00.0... eee eee 15
Aspartic acid. 2.0.0.0... eee eee 113
Threonine®............... 2220 eee eee eee 29
Serine’. oo... eee eee eee 65
Glutamic acid*...............0..-.0005- 113
Proline... 0.0... cece eee 40
Glycine... 0... cece eee e cee ene 61
Alanine... 0.0.0.0... ccc cece cence eee 55
Valine’. 00... eee eeeeeee 55
Methionine. ........ 0... cece eee eee 35
Tsoleucine®........ 0.0... cee cece ee eens 86
Leucine.......... 0.00 59
Tyrosine...... 0.00. ccc cee eee eee eee 47
Phenylalanine. ............. 0.0 cee eee 12
Tryptophan’. ......... 0... eee eee eee 12

 

* Includes amide.

> Extrapolated to zero time of hydrolysis.
* Value after 78 hours of hydrolysis.

4 Calculated from spectral analysis (34).
Issue of February 10, 1968

 

 

 

 

T } T q T T T T T T i} T
L. ™ 4
o”
13h ve”
© wv
oF Pad -
a
° eo
i eo”
L. eo" 4
eo”
1Ok o* |
eo”
I | I l 1 1 | j ] { I |
490 49.5 50.0
r7fom2]

Fig. 3. Sedimentation equilibrium analysis of the enzyme.
A 2.3-mm column was analyzed in a double sector cell with sap-
phire windows and with Rayleigh interference optics. Rotor
speed was 19,160 rpm; protein concentration, 0.50 mg per ml;
temperature, 6°; buffer, 0.1 m KCI1-0.05 m potassium phosphate
(pH 6.8)-0.01 m 2-mercaptoethanol. Abscissa, square of distance
from center of rotation (r); ordinate, logarithm of vertical fringe
displacement (c) (proportional to protein concentration). The
positions of the meniscus and bottom of sample were at 47.77
cm? and 50.25 em?, respectively.

Tasie IIT
Production of 6'-mononucleotides as result of nuclease
aclivily in purified enzyme fractions

Heat-denatured #P-labeled E. coli DNA (120 mumoles) was
incubated with 11 units of T4 polymerase until 6% and 18% of
the JJNA was made acid-soluble. *P; released by treatment of
the acid-soluble fraction of the digested DNA with either bac-
terial alkaline phosphatase or 5’-nucleotidase was measured as
Norit-nonadsorbable material.

 

Release of #P;

 

 

 

Enzyme 4a
6% acid-soluble | 18% acid-soluble
DNA DNA
| % %
Bacterial alkaline phosphatase...... | 103 99
5’-Nucleotidase............ccee ce eee | 103 101

 

units of enzyme per mumole of DNA for 6 hours under conditions
for synthesis, except that one deoxyribonucleoside triphosphate
was omitted; the DNA was then analyzed by sucrose gradient
sedimentation at alkaline pH. In studies with both DNA
samples, the sedimentation profiles were unchanged as a result of
incubation with the enzyme, showing the absence of single strand
scissions and indicating the absence of endonuclease activity.

Exonuclease activity was detectable by the action of the en-
zyme on “P-labeled #. coli DNA. Approximately 3.0 uumoles
of *P-labeled FE. colt DNA per unit of Fraction VII were made
acid-soluble in 30 min. This rate was increased 3-fold by prior
heat denaturation of the DNA. With a DNA substrate partially
digested with pancreatic deoxyribonuclease, the nuclease ac-
tivity was about 50 times that observed with native **P-labeled
E. coli DNA. Digestion of dAT copolymer by the enzyme pro-
ceeded at about the same rate as for partially digested DNA.
The exonucleolytic nature of the nuclease activity was indicated
by the appearance of 5’-mononucleotides as the exclusive product

M. Goulian, Z. J. Lucas, and A. Kornberg

631

following partial digestion (Table III). 5’-Mononucleotides
were identified by susceptibility to the action of 5’-nucleotidase.

Nuclease activity as reflected by loss of template was not
evident under the conditions of synthesis of DNA (Fig. 4A) but
was detectable to a moderate extent with dAT template (Fig.
4B). In the absence of deoxyribonucleoside triphosphates, the
nuclease activity of the enzyme was greatly enhanced (Fig. 4,
A and B). The maximum inhibitory effect of triphosphates on
the nuclease activity required the complete complement neces-
sary for replication of the template, but smaller effects were ob-
tained with less than a full complement (Table IV). Triphos-
phates not involved in replication, for example, dGTP and
dCTP with dAT as template, produced no effect (Table IV).

The close association of nuclease and polymerase activities is
evident from the constant ratio of these activities in fractions

 

 

 

 

  
 
   

 

| A. M13
4§- $H-templare 4
1
I
|
3
. 32D. product
— ~
ol f
I
|
|
1 4
|
rs By. . .
EP ao Htemplane (riphosphares omitted}
& 1 ;
2 B. dAT
E a
1°f
fy 5
H- Template
1
“it
\ °
1
L. e 7
A
e
‘ ™—. 326 product
/ en
° t
ary \
\ °y-template [riphesphates omitted)
je ~~

 

 

60 120 180
Time in minutes

Fig. 4. Effect of the enzyme on M13 DNA and dAT, in the
presence and absence of deoxyribonucleoside triphosphates. In
A, incubations contained 2 mumoles of linear *H-labeled M13
DNA and 7 units of Fraction VII in 0.4 ml; in B, incubations
contained 8 mgmoles of *H-dAT and 30 units of Fraction VII
in 1.0 ml. Conditions for DNA synthesis were those described
in “Methods,” except that deoxyribonucleoside triphosphates
were omitted where indicated. Abscissa, time of incubation;
ordinate, acid-precipitable DNA per ml of incubation mixture.
632

from hydroxylapatite chromatography (Fig. 1) and in material
eluted from polyacrylamide gel (Fig. 5). The ratios of polym-
erase to nuclease activity (measured as in Fig. 1) in pooled
fractions from phosphocellulose chromatography (Fraction V),

TaBLe IV

Effect of deoxyribonucleoside triphosphates on
nuclease activity of enzyme

The incubation conditions were those used for assay of poly-
merase activity as in ‘“Methods,’’ except for omission or addition
of deoxyribonucleoside triphosphates as indicated. The experi-
ment with #H-labeled T7 DNA used 9 mumoles of DNA and 24
units of enzyme in a 0.10-ml volume; for the experiment with
3H-dAT, 1.7 mumoles of dAT and 4 units of enzyme were contained
in 0.021 ml. Each deoxyribonucleoside triphosphate was present
at a concentration of 0.2 mm for the experiment with *H-labeled
T7 DNA, and 0.1 mm with 3H-dAT. The activities refer to rela-
tive amount of acid-soluble nucleotide after 15 min at 37°; 100%
activity represents the digestion of 0.9 mgmole of *H-labeled T7
DNA and 1.2 mygmoles of *H-dAT.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Substrate and additions Nuclease activity
%
8H-T7 DNA
No additions.................000 0000s 100
dGTP. 0... eee cee eee ee 51
dGTP, dATP........... cece eee eee 19
dGTP, dATP, dCTP................... 4
dGTP, dATP, dCTP, dTTP............ <1
sH-dAT
No additions............. 00.0.0. e eee eee 100
GGTP, dCTP. ....... 02. eee eee 89
GATP, dTTP................... 0. eee 9
T ] T

60f- m7 \protein 4
©
E
c
3 ‘a “+
9 Polymerase
Q
5
540+ |
2
5
é
5
v
&
=
&

 

 

 

10 20 30
Fraction number

Fig. 5. Polymerase and nuclease activities in fractions ob-
tained by polyacrylamide gel electrophoresis. Electrophoresis
was carried out on 17 yg of Fraction VII; 1.25-mm segments of the
gel were subsequently eluted. Recovery of polymerase activity
was 30%. Values for polymerase and nuclease refer to the amount
per segment and are plotted according to location in the gel;
the area under the nuclease plot is indicated by hatching. The
protein band in an identical gel, run in parallel and stained for
protein, is shown at the top in alignment with the abscissa.

Enzymatic Synthesis. of DNA. XXV

Vol. 243, No. 3

DEAE-cellulose chromatography (Fraction VI), hydroxylapatite
chromatography (Fraction VII), and polyacrylamide gel elec-
trophoresis (Fig. 5) were 1.84, 1.80, 1.80, and 1.92 (units per
unit), respectively. An additional indication of the insepara-
bility of the nuclease and polymerase activities was the similarity
of their rates of heat inactivation (Fig. 6).

Requirements for Activity
pH Optimum—Maximum activity was obtained over the pH
range of about 8 to 9 (Fig. 7).
Divalent Metal—There was no detectable polymerase activity
in the absence of divalent cation. Maximum activity was ob-
served at 6 mm Mgtt+. Mnt~ at an optimal concentration of

0.1 mm sustained a rate about one-fourth of that obtained with
6 mm Mg*t,

100!

 

80

60

Nn b
oO o

% activity remaining
3

o

o

 

4 _

! 1 | { { L |
o 40 30 40

 

 

 

20
Minutes at 42°

Fia. 6. Heat inactivation of polymerase and nuclease. Frac-
tion VII, 8 ug per ml, was incubated at 42° in 0.05 m Tris-HCl
(pH 7.5), 0.1 mM (NH4)2SQO,, and 1 mg of bovine plasma albumin
per ml. At the times indicated, aliquots were chilled and made
0.01 m in 2-mereaptoethanol. Assays for nuclease (14) and poly-
merase were conducted at 30°.

 

 

 

 

100;-—> T 1 1 ! Te7® i
80-
€ e
&
2 sok \
3 . \
> Na glycinate ‘.
~ 40
2
e
5 |
a .
20 Tris-maleate ° 1
eo.
— ! i L l L {
6.0 7.0 8.0 9.0 10.0
pH

Fig. 7. pH-activity curve for the enzyme. The standard
assay was used, with substitution of the indicated buffers at
0.05 m; 0.056 unit of enzyme was added per tube. pH was meas-
ured at room temperature. The activity with sodium glycinate,
pH 9.0, was arbitrarily taken as 100%.
Issue of February 10, 1968

 

 

  
 

 

 

 

 

 

J 1 ] T TB. subtilis! T
100 -
80 4
M13
c
°
gc 60-7 4
o a Salmon spern)
Oo
2
= 4
2
T7
A 40 _y
20 4
j 1 { | {
100 150 200

Time in minutes

Fig. 8. Extent of replication of various DNA templates by the
enzyme. Incubation conditions for synthesis were those de-
scribed in ‘“Methods.’”’ The incubations contained the following:
2.9 mumoles of alkali-denatured *H-labeled M13 DNA (linear
form) and 8 units of enzyme in 0.23 ml; 1.8 mgmoles of *H-labeled
T7 DNA (treated with \-nuclease and then denatured) and 6
units of enzyme in 0.1 ml; 0.9: myumole of alkali-denatured salmon
sperm DNA and 4 units of enzyme in 0.1 ml; or 5.8 mumoles of
alkali-denatured B. subtilis DNA and 150 units of enzyme in 0.4
ml.

Template—Like the T2 (2) and T5 polymerases (4), the T4
polymerase required DNA template in a single stranded form.
In general, the highest rates were seen with denaturated salmon
sperm DNA, but rates varied-widely depending upon source and
specific preparation. For example, the rates with denatured
DNA from E. coli, calf thymus, and T4 phage were 4, 1.5, and
1%, respectively, of the rate with denatured salmon sperm DNA.
It is assumed that these differences are primarily a reflection of
the content of ends available for initiation of synthesis. The
extent of synthesis also varied widely; it never exceeded 1 full
template equivalent, although in some cases this was closely ap-
proached (Fig. 8).

Native DNA had little or no template activity. For example,
the amount of acid-insoluble material formed with native T7
DNA after 4 hours of incubation was less than 1% of the tem-
plate, compared to 35% for the same preparation after denatura-
tion with alkali. Unfractionated, circular, single stranded M13
phage DNA was replicated approximately 3% in 3 hours of in-
cubation, with 95% of the observed synthesis occurring in the
first 20 min of incubation. From the kinetics of replication and
the augmentation that occurs with ring scission (see experiments
below with linear M13, Figs. 4, 8 and 12), it is assumed that this
low level of synthesis with M13 DNA results from replication of
the small amount of linear forms contaminating such prepara-
tions. It appears, therefore, that in addition to being single
stranded, DNA must have a free end to serve as a template for
T4 polymerase.

The synthetic copolymer dAT (16), which is a good template
for the FE. coli polymerase although double stranded, also serves
as template for the T4 enzyme. Rates with two different prep-

M. Goulian, Z. J. Lucas, and A. Kornberg

633

arations of dAT were 5 and 9% of that with denatured salmon
sperm DNA; a value of 11% was reported for the polymerase
induced by T2 infection (4). The maximum extent of synthesis
with dAT as template reached 58% of the input, after which
there was progressive loss of the synthesized product (Fig. 4).

Two different preparations of poly dC were incubated with
dGTP under synthetic conditions; the extent of replication of
one sample was less that 5%, and of the other less than 1%.
Replication of poly dA with dTTP was less than 0.5%.

Native DNA made partially single stranded by digestion with
exonuclease IIT (37) is capable of supporting synthesis with the
T4 polymerase. The extent of synthesis was equivalent to the
extent of prior digestion by exonuclease III (Fig. 9). Inasmuch as
the enzyme was unable to initiate synthesis from the 5’-end (see
below), it is assumed that the mechanism is the same as the “re-
pair” reaction described for FE. coli DNA polymerase; namely,
there is extension at the 3’-hydroxyl terminus to fill in the de-
nuded portions of the molecule (10). Repair-type synthesis
carried out by the phage enzyme stopped abruptly upon com-
pletion of repair, as would be expected from its inability to act on
fully helical DNA.

Inhibition by 3'-Phosphoryl Termini—A micrococcal nuclease
digest of DNA exerted an inhibitory effect upon the activity of |
the T4 polymerase. For example, addition of T7 DNA treated
with micrococcal nuclease to untreated, denatured T7 DNA,
in a ratio of 1:10, reduced the rate of synthesis to 42% of that
obtained with the same amount of untreated, denatured T7 DNA
alone. The corresponding value for a 1:1 mixture was 5%.
This is interpreted, as with 1. colt polymerase, as inhibition of
the enzyme by 3’-phosphory! termini (11). The inhibition was
completely reversed by removal of the 3’-phosphates with exo-
nuclease III (DNA phosphatase-exonuclease) (11).

Other Requirements—In common with the other purified DNA
polymerases, the T4 polymerase requires all four deoxyribonu-

 

50 ) TO T

ase
&
]
J

olymer'

a
6
T

a

y P

iy
lo

1
i]

% replication b
iS
T
J

 

 

il I J \ t
10 20 30 30 50
Prior degradation by exonuclease mM, %

 

Fig. 9. Repair of partially single stranded DNA. E. coli
DNA (59 mumoles) that had been subjected to graded extents of
digestion with exonuclease III (10) was incubated for 150 min
with 10 units of the enzyme in a volume of 0.3 ml and under stand-
ard conditions for synthesis. Extent of replication is plotted
against the extent of prior digestion. The figures in both cases
are based on the amount of DNA present before digestion with
exonuclease III.
634

TaBLe V
Effect of salt concentration on rate of synthesis
All incubations were conducted under the assay conditions as
in ‘Methods,’ except for omission of ammonium sulfate. NH,Cl
was added at the concentrations noted. All incubations con-
tained 0.34 unit of Fraction VII. The highest activity, with
0.05 m NHC, represented 3.0 mymole of acid-insoluble nucleo-
tide in 30 min and was taken as 100%.

 

Concentration of NHiCl | Relative activity

 

x %
0.00 63
0.02 75
0.05 100
0.10 83

0.20 | 3

 

«Tn addition to 0.067 m Tris-HCl, the mixture contained 0.01
M NaCl which was included with the DNA.

cleoside triphosphates, dATP, dGTP, dTTP, dCTP, or their
analogues, for synthesis of DNA. The concentration of each
deoxyribonucleoside triphosphate ordinarily used for assays was
0.033 mm; however, enzyme activity was augmented 38-fold by
increasing this to 0.33 mm. Omission of one, two, or three tri-
phosphates reduced the activity to 0.63, 0.23, and 0.10%, re-
spectively, of that obtained with all four present. Unlike the
E. coli polymerase (38), replacement of Mgt* by Mn*+ did not
permit substitution of a ribonucleoside triphosphate for a deoxy-
ribonucleoside triphosphate. Under the usual assay conditions
with either 6 mm Mgt or 0.2 mm Mnt*, replacement of dCTP
by CTP or of dATP by ATP resulted in less than 1% of the
normal activity.

A total salt concentration of approximately 0.10 m was nec-

essary for optimal activity (Table V); however, if this was in-
creased by the presence of 0.2 m NH,Cl to a total salt concen-
tration near 0.3 M, a level required for optimal activity with the
phage T5 polymerase (4), there was 97% inhibition (Table V).
The stimulatory and inhibitory effects were identical with NH,Cl,
NaCl, and KCl.
. Sulfhydry! reagents, either reduced glutathione or 2-mercapto-
ethanol], were included throughout the purification procedure,
during storage of enzyme, and in all incubations except where
noted. When the level of 2-mercaptoethanol in the incubation
mixture was reduced from 10 to 1 mm, the rate of synthesis
dropped 63%, while with 2-mercaptoethanol omitted entirely it
dropped by 81%. In the presence of p-hydroxymercuribenzoate
at a concentration of 1074 m, the rate was less than 2% of nor-
mal; the level of enzyme was 0.02 wg per ml, and bovine plasma
albumin (0.15 mg per ml) was present.

Nature of Product

Hybrid Nature of Template-Product Complexz—With a density
label, bromouracil, in the product, the product-template complex
exhibited a broad range of hybrid density values and product to
template ratios (Fig. 10). Most of the material closely ap-
proached but did not exceed the density corresponding to 50%
bromouracil and a product to template ratio of 1 (Fig. 10). The
over-all ratio in this experiment, determined directly on the in-
cubation mixture, was 0.71.

Sedimentation of the product-template hybrid was slower than
of template alone in a neutral sucrose gradient containing 1 M

Enzymatic Synthesis of DNA. XXV

Vol. 248, No. 3

sodium chloride, and the lower sedimentation rate could be as-
cribed to the conversion to a helical structure with increased
molecular weight. Thus, a linear M13 DNA fraction of 23 S,
corresponding to a molecular weight of 1.2 x 10° (40), was repli-
cated to an extent of 92%, yielding a product-template complex
of 158. The molecular weight calculated for a helical molecule
of the latter sedimentation value is 2.6 x 108 (40).

Covalent Attachment of Product and Template—In an alkaline
sucrose gradient, the template and product. were found to be
associated and to have a faster sedimentation rate than the tem-
plate alone (Fig. 11). Inasmuch as no secondary structure re-
mains at this alkaline pH, template and product are judged to be
linked through a covalent bond.

Attachment of Product to 3’-Hydrozryl of Template—Digestion of
the product-template complex with exonuclease III, an enzyme
that attacks at the 3’-hydroxyl termini of double stranded DNA
(87), quantitatively removed product material with minimal loss
of template (Fig. 124A). In contrast, \-exonuclease, an enzyme
that attacks only at the 5’-phosphory! termini of double stranded
molecules (12), removed only template material (Fig. 128).

 

I T T

| e=1203 | P=4.753 | P= 1.703
100 % BU 50% BU 0% BU

wR

mu moles

 

 

 

 

Fraction number

Fig, 10. Neutral CsCl density gradient centrifugation of the
product-template complex. Alkali-denatured *H-labeled \ DNA
was replicated under synthesis conditions except for the replace-
ment of dTTP by bromodeoxyuridine triphosphate. Incubation
volume was 1.0 ml, with 54 mumoles of template and 300 units
of enzyme. After 6 hours of incubation, the mixture was made
0.01 m in EDTA and heated at 60° for 10 min to inactivate the
enzyme. Density gradient centrifugation was carried out in
8.5 ml of CsCl, p = 1.745, in 0.01 m Tris, pH 8.0, and 1 mm EDTA.
The No. 50 rotor was run in the Spinco model L ultracentrifuge at
45,000 rpm for 80 hours at 25°. After measurement of specific
gravity, the fractions were precipitated with acid and collected
on glass filters as for the polymerase assay. Parallel centrifuga-
tions were carried out with a portion of the incubation mixture
and an internal marker of #H-labeled 4 DNA. Positions are
indicated for the densities of native \ DNA containing 100, 50,
and 0% bromouracil (BU). The value of p for \ DNA agrees with
data of Baldwin and Shooter (39) for E. coli DNA, which has the
same base composition as \ DNA. The p values for bromouracil-
containing \ DNA were calculated by using a value of 0.10 for
the density difference between 0 and 100% bromouracil estimated
from the difference in density between dAT and the copolymer
of deoxyadenylate and bromodeoxyuridylate (39).
Issue of February 10, 1968

This evidence for masking of the 3’-terminus of the template and
the 5’-terminus of the product supports the tentative conclusion
that a phosphodiester bridge between these termini links tem-
plate and product.

Nondenaturability—The covalent linkage between product and
template constitutes, in effect, a form of terminal cross-linkage
between the two strands of the hybrid. Since cross-linked DNA
renatures with great facility (41), it might be expected that the
structure produced with T4 polymerase would have similar
properties. This was tested by sedimenting the template-
product hybrid in a neutral sucrose gradient after alkaline de-
naturation. Under the conditions used, denatured DNA, be-
cause of its collapsed structure, sediments much more rapidly
than helical DNA of the same molecular weight (40). The
“denatured” hybrid, sedimenting with a value of 15 S, was not
distinguishable from the untreated product-template complex
(Fig. 13), thus indicating a facile reassumption of secondary
structure. This result was confirmed by the resistance of alkali-
treated hybrid to exonuclease I, an enzyme active only on single
stranded DNA (19). Less than 30% of the product and template
were digested by the latter enzyme after the alkaline treatment
in an experiment in which the control, alkali-denatured T7 DNA,
underwent 97% digestion. These experiments also provide
additional support for covalent attachment between product and
template.

Proposed Model

Several aspects of the action of T4 polymerase, when con-
sidered together, suggest a “hairpin” or “loop” model for replica-
tion of single stranded DNA, as illustrated in Fig. 144. These
aspects are as follows: (a) requirement for a linear, single stranded
template, or partially single stranded as in the case of DNA
treated with exonuclease III; (6) inhibition by 3’-phosphoryl
termini; (¢) hybrid nature of the product-template complex; (d)
covalent attachment between the 5’-end of the product and the
3’-end of the template; (e) facile renaturation of the product

 

J I 4 i T T T

Unreplicated
template
{gS

0.12 --
22S

mumotes
oO
3°
ao

0.04

 

 

 

Fraction number

Fic. 11. Alkaline sucrose gradient sedimentation of the prod-
uct-template complex. Linear $H-labeled M13 DNA (0.7 mumole)
and 2.8 units of enzyme were incubated in 0.07 ml for 3 hours
under conditions for synthesis. After addition of 3 ymoles of
EDTA and 0.8 umole of sodium pyrophosphate, the mixture was
centrifuged in a sucrose gradient containing 0.2 m NaOH, 0.8 m
NaCl, and 1 mm EDTA at 37,600 rpm (Spinco SW 39 rotor) for 8
hours at 10°. A duplicate experiment with an internal marker
in the sedimentation analysis was used to determine the sedi-
mentation constants.

M. Goulian, Z. J. Lucas, and A. Kornberg

635

 

T ( ’

t
A.  Exonuclease HI B. =AExonuclease | Bon. I

 

 

 

100<- 3u_ 0 |
\—e H-template P-productr
|

80 *, ~
Pry |
cot |
2 sol 4 l
pl
e |i |
= 4olf - 5 I
4 o—. H-template |

20 |} \ -}t
\ P-product
Ee es wes ee) |

 

 

 

° 20 40 60 0 20 40

Time in minutes

Fie. 12. Selective susceptibility of the product-template
complex to action of specific exonucleases. Linear *H-labeled
M13 DNA (0.45 mpmole) was incubated in 0.10 ml with 2.0 units
of enzyme for 3 hours under conditions for synthesis. Polymerase
was inactivated by heating at 65° for 20 min. A, exonuclease
IIT (20 units) was added and aliquots, taken at the times indicated,
were assayed for acid-insoluble *H-template and “P-product.
B, d-exonuclease (0.6 unit) was used in place of exonuclease III.
In a second part of this experiment, exonuclease I (Exon. I) (5
units per mumole of DNA) was added at 60 min as indicated by
the arrow.

 

T T T T T '
Denotured Template
complex S25] 4505 fs yes

020

0.05

 

 

 

Fraction number

Fig. 13. Neutral sucrose gradient sedimentation of alkali-
treated product-template complex. *H-Labeled M13 DNA (0.4
mymole) and 1.6 units of enzyme were incubated for 3 hours under
conditions for synthesis. The product was denatured with alkali
as described in ‘‘Methods,” a marker of *H-labeled M13 DNA
was added, and the mixture was centrifuged in a sucrose gradient
containing 1 M NaCl, 0.01 m Tris (pH 7.8), and 1 mm EDTA at
24,500 rpm for 9} hours at 10°. The sedimentation constant of
the template (23 S) was determined in a separate analysis with the
same marker. The s value of a fully denatured product-template
complex in which the template was replicated to an extent of 0.8
was calculated as 32S by an empirical formula (40).

template hybrid; and (f) limitation to one or less than one replica-
tion of the template.

According to the model (Fig. 14), initiation occurs at the 3’-
hydroxyl terminus of the single stranded template, which has
looped back, a looping which may be facilitated by fortuitous
base pair homology between the terminal residues and regions
636

Cc. AExonuclease
‘ 1) * Inno)
3 3!
* 5’
22 2,
se 3 —_—___)

&
3.

3

D. AExonuciease + Polymerase
> 5’

4. CoO)
3

B. Exonuclease II Polym. — AExo.
, >
1 ’ 2. 3s
3

° 5 5, 3 5/
3! )

A, Polymerase

-_

Exo. Polunt’

Fic. 14. Schematic representation of a proposed model for the
action of T4 polymerase, and effects of nucleases on the product-
template complex. A, replication of single stranded template
by polymerase. 1 represents a single strand of template DNA
(light line) in which the 3’-terminus loops back to permit initia-
tion by the polymerase (2). With a small amount of synthesis
(2) (heavy line), the loop is fixed in that position and synthesis
(“repair’’) continues to the 5’-terminus (8). 3B, effect of exo-
nuclease III on the isolated product-template complex shown in
A, $. C, effect of A-exonuclease action on the isolated product-
template complex of A, 8. D, action of d-exonuclease (AEzo.)
on product-template complex in the continued presence of poly-
merase (Polym.). The sequence of events in the experiment of
Fig. 15 is shown in terms of the model. Digestion of primer (1)
is followed by formation of a new loop and a wave of new synthesis
‘‘chasing’’ the wave of digestion (2, 3). When the loop region has
been replicated (4), A-exonuclease can resume action, with repeti-
tion of the process, eventually making product material available
to this nuclease (6).

along the same strand. This configuration now resembles a
DNA molecule partially digested by exonuclease III, which is
known to be an effective template. Chain growth occurs by the
same mechanism as in the case of the /. coli polymerase (42),
with attack by the termina] 3’-hydroxy! of the template upon the
incoming deoxynucleoside triphosphate, the selection of which is
determined by base pairing with the corresponding nucleotide in
the template. After replication has proceeded for a short dis-
tance the loop, otherwise unstable, is irreversibly fixed. Further
replication, as in repair of DNA made partially single stranded
with exonuclease III, continues to the 5’-terminus of the tem-
plate, where it stops because the DNA is now fully helical at the
3’-growing end. The failure of homopolymers as templates is
consistent with the absence of the postulated base pairing pre-
requisite for looping. The inhibition by 3’-phosphoryl termini
is expected if initiation occurs only at the 3’-hydroxyl terminus.
Facile renaturation of the product-template complex is readily
explained by the extensive complementarity within the hybrid
molecule.

The effects of specific exonucleases on the structure proposed
in this model are shown schematically in Fig. 14, Band C. Exo-

Enzymatic Synthesis of DNA. XXV

Vol. 243, No. 3

nuclease III attacks at the 3’-terminus of the helical molecule
(37) and removes only product material (Figs. 14B and 124),
whereas \-exonuclease, which acts on the 5’-terminus (12), re-
moves only the template (Figs. 14C and 12B). Since both en-
zymes are inactive with single stranded DNA, the complementary
portion of the hybrid is in each case spared from attack. Con-
version of the product to single stranded form, as in the experi-
ment with d-exonuclease, makes the product susceptible to
exonuclease I action (Figs. 14C and 12B) and should also permit
resumption of polymerase action (Fig. 14D). The latter pre-
diction led to the experiment described in Fig. 15. Addition of
d-exonuclease to a replication mixture containing fully replicated
DNA and active polymerase was followed by a burst of synthesis
paralleling the digestion of template. Thus the digestion by )-
exonuclease at the 5’-terminus of template enables the 3’-ter-
minus of the product to loop back and the polymerase to start
synthesis again (Fig. 14D). The subsequent loss of template
(Fig. 15) can be explained as the result of still another round of
digestion, which becomes possible when the nuclease and poly-
merase reach the site of the original loop (Fig. 14D).
Heterogeneity in product to template ratios was observed in
density gradients run with CsCl or with sucrose at alkaline pH
(Fig. 11), indicating that the extent of replication was not the
same for all template molecules. According to the mechanism
proposed, this suggests that the loops are of varying size. The
close approach to a full replication of all molecules of some DNAs
(Fig. 8), or of a fraction of the molecules of other DNAs (as re-
vealed by density studies in Fig. 10), indicates that the portion of

 

y T T

3y-template
Oe g

Pa
SS e
hy
4

32P-producr
ae ed

\
\

we
T

Te
/

mumoles/m|.

~
T

e
Og

 

AExonuclease
J L 1
60 120 180 240
Time in minutes

 

oO

O ee ee

Fig. 15. Effect of A-exonuclease on the product-template com-
plex in the presence of T4 polymerase. The experiment was
carried out as in Fig. 12, except that polymerase was not inacti-
vated prior to addition of \-exonuclease; no exonuclease I] was
added. Figures on the ordinate represent acid-precipitable *H-
template and “P-product during the period of replication and
after addition of \-exonuclease.
Issue of February 10, 1968

template remaining in the form of a single stranded loop in the
final product-template complex can be very small.

DISCUSSION

The replication by T4 polymerase of a single stranded DNA
template leads to a product which is in a complex with the tem-
plate. In this complex, the template strand is covalently at-
tached by its 3’-terminus to the 5’-terminus of the product.
According to our current notions, as presented in Fig. 14, the en-
zyme produces this structure by extension of the single stranded
DNA at the 3’-terminus, which has looped back, thereby per-
mitting the strand to act as template. Subsequent replication
to the 5’-terminus of the template then resembles the “repair”
reaction, as in the replacement of the denuded portions of a DNA
treated with exonuclease IIT (10), a reaction carried out by both
the T4 and #. coli polymerase enzymes. The loop mechanism
as proposed here for the replication of single stranded DNA by
phage enzyme may also apply to the replication of such DNA by
the E. colt polymerase. The principal difference in replicative
functions between the phage and bacterial enzymes then appears
to be the ability of the bacterial enzyme, and inability of the
phage enzyme, to initiate new strands and to utilize fully helical
DNA as template (43).

The DNA polymerase activity purified from calf thymus re-
sembles the T4 polymerase in its requirement for denatured DNA
and limitation to one replication of template (44). The evidence
for rapid renaturation of the template-product complex produced
by the thymus polymerase, and for the covalent attachment be-
tween template and product (44, 45), suggests that a loop mecha-
nism may also apply to the thymus enzyme. The DNA poly-
merase activities from a variety of other animal tissues have, for
the most part, shown a similar preference for single stranded
template and limited replication (46). It seems likely, therefore,
that these enzymes may all have the same mechanism, but more
experimental data will be required to establish this.

Two other purified phage polymerases, induced by infections
with T2 (2) and T5 (4), also require single stranded DNA and
achieve a replication limited to 1 template equivalent. The
similarities in properties between the purified T2 and T4 poly-
merases, in addition to the extensive areas in which the T2 and
T4 bacteriophages themselves resemble each other (47), make
it likely that the mechanism described for the T4 polymerase also
applies to the T2 enzyme. The comparison between the T4 and
T5 enzymes reveals certain dissimilarities at this time. The
product of T5 polymerase action is readily separated by denatura-
tion from its template (48), and the T5 enzyme is able to utilize
®X174 DNA (49). If further work firmly establishes these
observations, then the T5 enzyme will be distinctive as compared
with the T4 enzyme in its capacity to initiate new strands in the
replication of a single stranded template.

Another point of comparison between the T4 and T5 poly-
merases is their response to the ionic strength of the replication
medium. The T4 enzyme activity is inhibited 97% by 0.2 m
NH.Cl, a salt concentration required for optimal activity with
the T5 enzyme (4). This fact suggests that, if the T5 polymerase
were to act by the loop mechanism proposed for the T4 enzyme,
the salt effects would be primarily upon the enzymes rather than
on the DNA templates.

Recent studies of the enzyme activities of extracts of nonper-
missive bacteria infected with mutants of T4 have identified the
structural gene for DNA polymerase (5). The fact that mutants

M. Goulain, Z. J. Lucas, and A. Kornberg

637

in the same gene are unable to induce synthesis of phage DNA
(50) indicates that the enzyme does serve an essential function in
DNA synthesis. The gene-enzyme-function relationship is
strengthened by the demonstration of temperature-dependent
activity for the polymerase purified from temperature-sensitive
mutants in the same gene (5). Some mutants in the polymerase
gene have the additional property of inducing a high mutation
rate in other genes (51). Although it would seem reasonable to
ascribe the latter to errors in replication resulting directly from
structural alterations in the polymerase enzyme, this has not yet
been firmly established.

The genetic evidence for T4 polymerase as an integral com-
ponent of DNA replication in vive draws sharper attention to the
paradoxical failure of the polymerase to replicate helical DNA
in vitro. Proper expression of the enzyme function may require
a different form of the enzyme present in vivo, or a specific form
of template DNA (52), or additional factors, which may or may
not be enzymatic in character. Unsuccessful experiments not
described in this report have attempted to correct the deficiency
in vitro by using “replicative forms” of T4 DNA extracted from
infected cells (52), and by supplementation with the purified host
cell polymerase or with crude extracts of infected cells. It seems
clear that complexities of replication of the T4 chromosome are
responsible for the current lack of success in reconstituting the
entire event in vitro.

The nuclease activity of the highly purified T4 enzyme re-
mains closely associated with polymerase activity under all of the
conditions used in these experiments, similar to the result ob-
tained by Short and Koerner in studies on partially purified T2
polymerase (53). A marked inhibition of nuclease activity oc-
curs in the presence of a full complement of deoxyribonucleoside
triphosphates. The inhibitory effect of the triphosphates ap-
pears to be related to their ability to support polymerase activity;
thus with DNA as template the full inhibitory effect required
that all four triphosphates be present, whereas with dAT as
template dGTP and dCTP had no effect. Apparently the nu-
clease activity of the enzyme remains suppressed or is more than
compensated for as long as the requirements for polymerase ac-
tivity are met. The association between nuclease and polymer-
ase in the T4 enzyme is reminiscent of the close relationship be-
tween #. coli polymerase and exonuclease II (26). For neither
polymerase are the functions of the nuclease known, but it does
invite the speculation that the nucleases may function in correc-
tion of errors made during replication. Perhaps errors by
polymerase in base selection result in structures which are rela-
tively unsuitable for additional polymerase action and are rela-
tively more susceptible to excision of the mismatched base by
the nuclease.

REFERENCES

. Cowen, 8. 8., Fed. Proc., 20, 641 (1961).
2. AposHiaN, H. V., AnD KoRNBERG, A., J. Biol. Chem., 237, 519
(1962).

3. Wisere, J.S., Dirgsen, M.-L., Epstein, R. H., Luria, 8S. E.,
AND BucHANAN, J. M., Proc. Nat. Acad. Sct. U. S. A., 48,
293 (1962).

. Orr, C. W. M., Herriorr, 8. T., anp Bessman, M. J., J.
Biol. Chem., 240, 4652 (1965).

. DE Waarp, A., Paun, A V., AND Lexman, I. R., Proc. Nat.
Acad. Sei. U. 8. A., 64, 1241 (1965).

. Warner, H: R., anp Barnes, J. E., Virology, 28, 100 (1966).

. Lexman, I. R., Bessman, M. J., Sums, E.S., anp Kornsere,
A., J. Biol. Chem., 233, 163 (1958).

—

oo >

Im
638

8.

9.
10.

Il.

12.
13.

14.
15.
16.

17.
18.

19.
20.

21.
22,

23.
24,

25.
26.

27.

28.
29.

30.

Mirra, 8., Reicuarp, P., Inman, R. B., Bertscu, L. L., anp
Kornzere, A., J. Mol. Biol., 24, 429 (1967).

Linn, 8., anp Leaman, I. R., J. Biol. Chem., 240, 1287 (1965).

Ricwarpson, C. C., Inman, R. B., anp Kornuera, A., J. Mol.
Biol., 9, 46 (1964).

Ricwarpson, C. C., ScurupKravr, C. L., anp Kornperg, A.,
Cold Spring Harbor Symp. Quant. Biol., 28, 9 (1963).

Lirriz, J. W., J. Biol. Chem., 242, 679 (1967).

Kay, E. R. L., Simmons, N. S., anp Dounce, A. L., J. Amer.
Chem. Soc., T4, 1724 (1952).

OLEson, Q. E., anp Korrner, J. F., J. Biol. Chem., 289, 2935
(1964).

Massiz, H. R., ann Zim, B. H., Proc. Nat. Acad. Sci. U.S. A.,
54, 1641 (1965).

ScHacHMAN, H. K., ApuEr, J., Rapping, C. M., Leuman, I. R.,
AND Kornsera, A., J. Biol. Chem., 235, 3242 (1960).

Inman, R. B., anp BaLpwin, R. L., J. Mol. Biol., 8, 452 (1964).

Botium, F. J., Groznicmr, E., anp YONEDA, M., Proc. Nat.
Acad. Sct. U. S. A., 51, 853 (1964).

Leuman, I. R., J. Biol. Chem., 285, 1479 (1960).

Ricwarpson, C. C., anp Kornsera, A., J. Biol. Chem.,
242 (1964).

Leuman, I. R., anp Nusspaum, A. L., J. Biol. Chem., 239,
2628 (1964).

LirrLe, J. W., Lenman, I. R., anp Kaiszr, A. D., J. Biol.
Chem., 242, 672 (1967).

Linn, S., anp Lenman, I. R., J. Biol. Chem., 240, 1204 (1965).

Leuman, I. R., Roussos, G. G., aNnp Pratt, E. A., J. Biol.
Chem., 287, 819 (1962).

CunninanaM, L., Caruin, B. W., anp Privat DE GARILHE,
M., J. Amer. Chem. Soc., 78, 4642 (1956).

Ricwarpson, C. C., Scuiupkraut, C. L., Aposuian, H. V.,
AND Kornpere, A., J. Biol. Chem., 239, 222 (1964).

Moore, §., anp Stein, W. H., in 8. P. Cotowicx anp N. O.
Kaptan (Editors), Methods in enzymology, Vol. VI, Aca-
demic Press, New York, 1963, p. 819.

Frucutsr, R.G., anp Crestriecp, A. M., J. Biol. Chem., 240,
3875 (1965).

Martin, R. G., ann Amgs, B. N., J. Biol. Chem., 2386, 1372
(1961).

Fauascut, A., AnD Kornaere, A., J. Biol. Chem., 241, 1478
(1966).

239,

Enzymatic Synthesis of DNA. XXV

31.
32.

33.
34,

35.
36.
37.
38.

39.

40.
41.

42,
43.

48.
49.

51.
52.
. SHort, E.C., Jr..

Vol. 243, No. 3

Lowry, O. H., Roszprouen, N.J., Farr, A. L.,
R. J., J. Biol. Chem., 193, 265 (1951).

Saneemr, F,, anp Taompson, E. O. P., Biochim. Biophys. Acta,
71, 468 (1963).

Moors, &., J. Biol. Chem., 288, 235 (1963).

Beaven, G. H., anp Hotmay, E. R., Advance. Protein Chem.,
7, 319 (1952).

Yeruantis, D. A., Biochemistry, 8, 297 (1964).

Coun, E. J., ano Epsauu, J. T., Proteins, amino acids, and
peptides as ions and dipolar ions, Reinhold Publishing Cor-
poration, New York, 1943, p. 370.

Ricuarpson, C. C., Leuman, I. R., anp Kornsere, A, J.
Biol. Chem., 239, 251 (1964).

Bure, P., Fancurer, H., anp CHAMBERLIN, M., in H. VocEt,
V. Bryson, anp J. O. Lampren (Editors), Symposium on
informational macromolecules, Academic Press, New York,
1963, p. 467.

Batpwin, R. L., anp SHooter, E. M., J. Mol. Biol.
(1963).

Srupigr, F. W., J. Mol. Biol., 11, 373 (1965).

Becxer, E. F., ZIMMERMAN, B. K., anp GEIDUSCHEE, E. P.,
J. Mol. Biol., 8, 377 (1964).

Kornegera, A., Enzymatic synthesis of DNA, John Wiley ana
Sons, Inc., New York, 1961.

ScuitpERAvT, C. L., Ricnarpson, C. C.,
J. Mol. Biol., 9, 24 (1964).

AND RANDALL,

7, dll

AND KoRNBERG, A. Fi

. Bouttum, F. J., Cold Spring Harbor Symp. Quant. Biol., 28, an

(1963).

. Bouuum, F. J., Progr. Nucleic Acid Res.,1, 1 (1963).
46.
47.

Keir, H. M., Progr. Nucleic Acid Res., 4, 81 (1965).

STENT, G. S, Molecular biology of bacterial viruses, W. Hs
Freeman and Company, San Francisco, 1963.

Bressman, M. J., aND Orr, C. W. M., Fed. 'Proc., 24, 286 (1965).

Srevart, C. D., AND BESSMAN, M. J., Fed. ’Proc., 26, 396
(1967).

. Epstein, R. H., Boute, A., Sreinperc, C. M., KELLEN-

BERGER, E., Boy DE LA Tour, E. , CHEVALLEY, R. , EpgaR,
R.&., Susman, M., DENHARDT, G. H., anp LizLavsts, A,
Cold Spring Harbor Symp. Quant. Biol., 28, 375 (1963).
Spzyer, J. F., Biochem. Biophys. Res. Commun., 21, 6 (1965).
Frankeu, F. R., J. Mol. Biol., 18, 127 (1966).
AND Korrner, J. F., Proc. Nat. Acad. Sct.
U.S. A., 64, 595 (1965).