ArticlePDF Available

DNA Nucleotide Sequence Restricted by the RI Endonuclease

Authors:
  • CompleGen

Abstract and Figures

The sequence of DNA base pairs adjacent to the phosphodiester bonds cleaved by the RI restriction endonuclease in unmodified DNA from coliphage lambda has been determined. The 5'-terminal nucleotide labeled with (32)P and oligonucleotides up to the heptamer were analyzed from a pancreatic DNase digest. The following sequence of nucleotides adjacent to the RI break made in lambda DNA was deduced from these data and from the 3'-dinucleotide sequence and nearest-neighbor analysis obtained from repair synthesis with the DNA polymerase of Rous sarcoma virus [Formula: see text] The RI endonuclease cleavage of the phosphodiester bonds (indicated by arrows) generates 5'-phosphoryls and short cohesive termini of four nucleotides, (p)A(p)A(p)T(p)T. The most striking feature of the sequence is its symmetry.
Content may be subject to copyright.
Proc.
Nat.
Acad.
Sci.
USA
Vol.
69,
No.
11,
pp.
3448-3452,
November
1972
DNA
Nucleotide
Sequence
Restricted
by
the
RI
Endonuclease
(symmetry/cohesive
ends/RSV
DNA
polymerase/DNA-protein
recognition/R-factor
product/coliphage
X)
JOE
HEDGPETH,
HOWARD
M.
GOODMAN,
AND
HERBERT
W.
BOYER
Departments
of
Microbiology
and
Biochemistry,
University
of
California
at
San
Francisco
Medical
Center,
San
Francisco,
Calif.
94122
Communicated
by
Paul
Berg,
September
18,
1972
ABSTRACT
The
sequence
of
DNA
base
pairs
adjacent
to
the
phosphodiester
bonds
cleaved
by
the
RI
restriction
endonuclease
in
unmodified
DNA
from
coliphage
X
has
been
determined.
The
5'-terminal
nucleotide
labeled
with
32p
and
oligonucleotides
up
to
the
heptamer
were
analyzed
from
a
pancreatic
DNase
digest.
The
following
sequence
of
nucleotides
adjacent
to
the
RI
break
made
in
X
DNA
was
deduced
from
these
data
and
from
the
3'-dinucleotide
se-
quence
and
nearest-neighbor
analysis
obtained
from
repair
synthesis
with
the
DNA
polymerase
of
Rous
sarcoma
virus
5'...
.
A/TpG
I
pApAPTPTPCPT/A
....
3'
3'....T/ApCpTpTpApA0t
GpA/T
....
5'
The
RI
endonuclease
cleavage
of
the
phosphodiester
bonds
(indicated
by
arrows)
generates
5'-phosphoryls
and
short
cohesive
termini
of
four
nucleotides,
pApApTpT.
The
most
striking
feature
of
the
sequence
is
its
symmetry.
The
restriction
and
modification
enzymes
of
several
host
specificities
(1-7),
including
the
endonuclease
and
methylase
of
the
host-specificity
(RI)
controlled
by
the
fi+
R-factor
(Yoshimori,
Roulland-Dussoix,
Aldridge
&
Boyer;
Yoshimori,
Roulland-Dussoix,
Goodman
&
Boyer,
to
be
published)
have
been
characterized.
The
restriction
endonuclease
of
a
given host
specificity
interacts
with
a
specific
sequence
of
DNA
base-pairs
and
produces
a
double-strand
break
in
the
molecule
(1-6),
while
the
modification
methylase
methylates
a
base
in
each
strand
of
this
sequence
(refs.
3
and
7;
Yoshi-
mori,
Roulland-Dussoix,
Goodman
&
Boyer,
to
be
published).
Methylation
of
a
base
in
either
strand
is
sufficient
to
prevent
the
endonuclease
from
attacking
the
sequence
(3).
Therefore,
in
molecular
terms,
a
given
DNA
restriction
and
modification
host-specificity
can
be
defined
by
the
sequence
of
base
pairs
recognized
by
the
restriction
endonuclease
and
modification
methylase.
In
this
paper,
we
present
an
analysis
of
the
se-
quence
of
base
pairs
adjacent
to
the
phosphodiester
bond
cleavages
made
in
bacteriophage
X
DNA
by
the
RI
restric-
tion
endonuclease.
MATERIALS
AND
METHODS
Enzyme.
The RI
restriction
endonuclease,
purified
as
will
be
described
by
Yoshimori,
Roulland-Dussiox,
Aldridge,
and
Boyer
(to
be
published),
was
free
of
detectable
nonspecific
exo-
or
endonuclease
activities
and
migrated
as
one
band
in
native
or
Na
dodecyl
S04-polyacrylamide
gels.
Purified
RNA-
directed
DNA
polymerase
of
Rous
Sarcoma
virus,
free
of
detectable
exo-
or
endonucleolytic
activities
(8,
9),
was
a
generous
gift
from
Dr.
A.
J.
Faras.
Polynucleotide
kinase
(10)
was
purified
as
described
elsewhere.
Escherichia
coli
alkaline
phosphatase
rechromatographed
as
described
else-
where
(11),
pancreatic
DNase,
micrococcal
nuclease,
venom
phosphodiesterase,
and
spleen
phosphodiesterase
were
ob-
tained
from
Worthington
Biochemical
Co.
Phage
X
DNA.
A
X
(CI857SUSS7)
lysogen
of
E.
coli
1100
was
used
as
a
source
of
X
phage.
Tryptone
broth
cultures
of
ly-
sogens
(2
X
108
cell
per
ml)
were
heated
to
420
for
0.5
hr
(isotope
was
added
here
for
labeled
DNA)
and
incubated
at
370
for
3
hr.
The
cells
were
collected
by
low-speed
centrifuga-
tion
and
resuspended
in
5-10
ml
of
50
mM
Tris-HCl
(pH
7.5)-10
mM
MgCl2.
0.5
ml
of
CHC13
was
added
and
the
sus-
pension
was
incubated
at
370
for
15
min.
After
low-speed
centrifugation,
the
viscosity
of
the
supernatant
was
reduced
by
addition
of
pancreatic
DNase
(10-50
Mug).
The
supernatant
was
centrifuged
at
79,000
X
g
at
40,
for
1
hr.
The
pellet
was
resuspended
in
50
mM
Tris
HCl
(pH
7.5)
and
layered
on
a
CsCl
step-gradient
of
the
following
composition:
1.7,
1.5,
1.3,
1.2
g/cm3
CsCl
and
a
top
layer
of
36%
sucrose,
all
in
50
mM
Trist
HCl
(pH
7.5).
The
layered
gradient
was
centri-
fuged
for
2
hr
at
41,000
X
g
and
100.
The
phage,
which
banded
at
1.5
g/cm3
CsCl,
was
aspirated
and
recentrifuged
to
equi-
librium
in
CsCl.
The
purified
phage
was
dialyzed
and
ex-
tracted
with
phenol.
After
ether
extraction
of
the
phenol,
the
DNA
preparation
was
dialyzed
against
50
mM
Tris-
HCl
(pH
7.5)-0.1
mM
EDTA
(TE
buffer).
Usually
90%
of
the
labeled
DNA
prepared
this
way
sedimented
as
one
band
in
alkaline
sucrose
gradients.
However,
terminal
labeling
data
(see
Results)
indicated
that
the
unlabeled
DNA
used
here
had
2-3
internal
nicks
per
X
molecule.
DNA
concentrations
were
determined
spectrophotometrically;
6.6
A260
=
1
mM
DNA
nucleotide
(12).
Terminal
Labeling
of
DNA.
RI
endonucleolytic
digestions
of
unmodified
X
DNA
were
performed
in
100
mM
Tris-
HCl
(pH
7.5)-5
mM
MgCl2
for
30
min
at
37°.
After
exhaustive
endonucleolytic
digestion
the
DNA
was
extracted
with
phenol
and
dialyzed
against
10
mM
Tris-HC1
(pH
7.5)-i
mM
EDTA.
Terminal
phosphoryl
groups
were
removed
by
alkaline
phosphatase
at
370
(11),
and
the
DNA
was
ex-
tracted
with
phenol
and
dialyzed
against
TE
buffer.
The
5'-
terminal
hydroxyls
were
phosphorylated
with
polynucleotide
kinase
(10),
and
[y_32P
]rATP
[synthesized
by
a
modification
of
the
procedure
of
Glynn
and
Chappell
(13)
and
purified
as
described
by
Wehrli
et
al.
(14),
2-3
Ci/mmol].
The
terminally-
3448
Abbreviations:
RSV,
Rous
sarcoma
virus;
nucleosides
without
a
prefix
(A,
T,
G,
C)
are
deoxynucleosides.
Nucleotide
Sequence
of
RI
Restriction
Site
3449
labeled
[32P]DNA
was
extracted
with
phenol,
dialyzed
against
TE
buffer-0.5
M
NaCl,
and
precipitated
with
ethanol.
Generation
and
Purification
of
Terminally
Labeled
Oligonu-
cleotides.
The
terminally
labeled
[5'-32P]DNA
was
mixed
with
uniformly
labeled
X
[33P]DNA
and
digested
with
pancreatic
DNase
and
venom
phosphodiesterase
(15).
The
digest
was
adsorbed
to
a
DEAE-Sephadex
A-25
column
(0.9
X
20
cm)
equilibrated
with
5
mM
Tris
HCl
(pH
7.5)-7
M
urea,
and
maintained
at
650
(16).
Oligonucleotides
of
increasing
chain
length
were
eluted
with a
linear
gradient
of
(0-0.4
M)
NaCl,
desalted
on
small
DEAE-Sephadex
columns,
and
evaporated
to
dryness
by
flash
evaporation.
RSV
DNA
Polymerase
Reactions.
RI-digested
and
un-
digested
X
DNA
were
used
as
primer-templates
for
the
RSV
DNA
polymerase
with
various
combinations
of
[a-32P]deoxy-
ribonucleotide
triphosphates
(8-20
Ci/mmol,
ICN
Corp.)
as
substrates.
A
typical
reaction
contained
0.1
M
Tris-
HCl
(pH
8.0),
2
mM
2-mercaptoethanol,
10
mM
MgCl2,
10
,M
NTP,
and
0.0002
units
of
RSV
DNA
polymerase
per
ug
of
DNA.
The
enzyme
specific
activity
was
0.1
unit/Mug
of
enzyme
pro-
tein
(8).
Incubations
were
for
2
hr
at
37°.
An
aliquot
of
the
reaction
was
precipitated
with
perchloric
acid
to
determine
the
total
incorporation.
The
remainder
was
extracted
with
phenol,
and
dialyzed
for
2
days
against
at
least
three
changes
of
4
liters
of
0.3
M
sodium
acetate-10
mM
Tris-
HCl
(pH
7.5).
The
DNA
was
precipitated
with
ethanol
and
resuspended
in
a
minimal
volume
of
water.
The
labeled
DNA
was
completely
digested
for
nearest-neighbor
analysis
[40
units
of
micro-
coccal
nuclease
in
10
mM
sodium
borate
buffer
(pH
8.6)-20
mM
CaCl2
for
30
min
at
370,
then with
0.2
units
of
spleen
phosphodiesterase,
50
mM
Tris
*
HCl
(pH
7.5),
10
mM
MgC12,
5
mM
rAMP].
Electrophoresis.
Oligonucleotides
were
separated
by
two-
dimensional
electrophoresis
on
DEAE-paper
(Whatman
DE-
81);
the
first
dimension
was
pyridine-acetate
(pH
3.5)
and
the
second
dimension
was
formate-acetate
(pH
1.9)
(17,
18).
At
times,
7%
formic
acid
was
used
for
the
second
dimension.
Mononucleotides
were
separated
by
electrophoresis
on
What-
man
3MM
paper
in
pyridine-acetate
(pH
3.5).
Appropriate
absorbance
markers
were
added
where
necessary,
the
pyridine
was
removed
from
the
paper
in
an
NH3
atmosphere,
and
the
marker
was
visualized
under
a
UV
lamp.
All
the
other
micro-
techniques
for
nucleotide
sequence
analysis
are
essentially
the
same
as
described
elsewhere
(17,
18).
Radioautography
and
Radioactive
Counting.
Radioactive
spots
on
electropherograms
were
located
by
placing
three
sheets
of
x-ray
film
(Kodak
RPR
14)
over
the
electrophoresis
paper.
Spots
containing
32p
and
33P,
or
spots
containing
33P
alone,
were
located
by
exposure
of
the
first
film
adjacent
to
the
paper.
Spots
containing
32p
could
be
distinguished
from
spots-
containing
only
33P
by
exposures
on
the
third
film,
since
a
given
amount
of
33P
was
only
1-2%
as
effective
at
exposing
the
third
film
as
was
an
identical
amount
of
32p.
Radioactivity
was
quantitated
by
counting
in
a
Nuclear-Chicago
gas-flow
counter,
with
or
without
appropriate
screens,
or
by
differen-
RESULTS
The
5'-terminal
mononucleotide
at
the
RI
endonuclease
break
Preliminary
experiments
established
that
X
DNA
treated
with
endonuclease
RI
and
alkaline
phosphatase
before
phosphoryl-
ation
with
polynucleotide
kinase
incorporated
six
to
twelve
more
molecules
of
82p
per
X
genome
than
did
X
DNA
treated
only
with
phosphatase.
DNA
treated
only
with
endonuclease
RI
accepted
about
the
same
amount
of
32p
as
did
untreated
DNA.
We
take
these
results
to
indicate
that
the
RI
endo-
nuclease
generates
5'
phosphoryls
and
three
to
six
double-
strand
breaks
per
X
genome.
Two
other
independent
estimates
indicate
about
five
sites
per
X
genome,
the
number
we
have
used
for
subsequent
calculations
(R.
Davis,
personal
com-
munication;
Yoshimori,
Roulland-Dussoix,
Goodman,
and
Boyer,
to
be
published).
To
determine
the
mononucleotide
at
the
5'
terminus,
an
aliquot
of
DNA
that
had
been
treated
with
endonuclease
RI
and
alkaline
phosphatase
was
phos-
phorylated
at
the
5'-terminus
with
32p
and
digested
to
5'-
mononucleotides
by
successive
treatment
with
pancreatic
DNase
and
venom
phosphodiesterase.
Analysis
of
the
mono-
nucleotides
after
separation
by
paper
electrophoresis
demon-
strated
that
the
amount
of
32p
occurring
in
pA
increased
6-
to
7-fold
when
the
DNA
had
been
treated
with
endonuclease
RI
before
dephosphorylation
(see
Table
1).
Only
pA
showed
a
significant
increase
in
32p,
and
the
increase
did
not
occur
in
DNA
treated
only
with
endonuclease
RI. This
result
indicates
that
the
5'-mononucleotide
at
each
strand
of
an
endonuclease
RI
break
is
pA.
Although
the
increase
(7-8
molecules
per
genome)
in
total
32p
incorporated
into
endonuclease
RI-phosphatase-treated
DNA
(compared
to
DNA
treated
with
phosphatase
only)
appears
to
be
accounted
for
by
the
increase
in
pA,
we
cannot
say
with
certainty
that
this
accounts
for
all
of
the
RI-created
termini.
We
were
unable
to
convincingly
demonstrate
in-
corporation
of
32p
above
background
into
the
normal
X
termini.
These
are
probably
obscured
by
the
presence
of
two
to
three
presumably
internal
random
breaks
per
X
molecule,
which
evidently
occurred
during
purification
and/or
storage
(see
Table
1,
Exp.
II).
Therefore,
we
cannot
exclude
the
possibility
that
one
in
ten
breaks
had
a
different
5'-terminus.
However,
these
data,
along
with
the
RSV
DNA
polymerase
TABLE
1.
The
5'-terminal
nucleotide
at
the
Rl
endonuclease
break*
32p
atoms
incor-
Treatment
porated/
[32P]
cpm
Exp.
before
X
no.
phosphorylation
genome
pC
pA
pG
pT
I
None
170
470
524
213
Endonuclease
RI
160
1056
415
209
Endonuclease
RI
231
7854
690
314
+
phosphatase
II
Phosphatase
(a)
3.6 402
780
518
630
(b)
4.8
290
625
427
416
Endonuclease
RI
(a)
12.4
4
55
4165
657
474
+
phosphatase
(b)
12.2
389
4016
509
528
tial
counting
in
a
liquid
scintillation
counter.
*
See
note
added
in
proof.
Proc.
Nat.
Acad.
Sci.
USA
69
(1972)
3450
Biochemistry:
Hedgpeth
et
al.
TABLE
2.
The
5'-terminal
sequence
at
the
RI
endonuclease
break
Ter-
Base
compo-
minal
sition
of
32p.
nucleo-
%
of
containing
tide
Chain
input
oligonu-
Value
from
Deduced
length
cpm
cleotide
of
M*
M
sequence
Mono-
6.6
Al
pA
Di-
15.2
A2
-
pApA
Tri-
19.7
A2,
T
1.2
T
pApApT
Tetra-
24.7
A2,
T2
2
T
pApApTpT
Penta-
14.3
A2,
T2,
C
0.03
C
pApApTpTpC
Hexa-
8.9
A2,
T3,
C
2.3
T
pApApTpTpCpT
A3,
T2,
C
0.3
A
pApApTpTpCpA
Hepta-
5.1
-
*
The
value
of
M
is
characteristic
of
the
nucleotide
removed
(see
text),
and
the
range
of
M-
values
for
each
nucleotide
are
C
-(0.08-0.1),
A
(0.3-0.5),
G
(1.4-3),
and
T
(1.4-3).
experiments
discussed
below,
make
it
highly
likely
that
we
have
accounted
for
nearly
all
(nine
out
of
ten)
or
probably
all
of
the
5'-termini
at
the
RI
endonuclease
breaks
in
X
DNA.
The
5'-terminal
sequence
RI
endonuclease-treated
X
DNA
labeled
at
the
5'-termini
with
82p
(10)
was
added
to
X
DNA
uniformly
labeled
with
33P,
and
the
mixture
was
digested
exhaustively
with
pancreatic
DNase
and
briefly
with
venom
phosphodiesterase
to
generate
5'-
phosphoryl-terminated
oligonucleotides
of
average
chain
length
three
to
five
(Table
2
and
ref.
15).
After
separation
according
to
chain
length
(16),
the
oligonucleotide
mixtures
of
identical
chain
length
were
separated
according
to
base
composition
and
sequence
by
two-dimensional
high-voltage
electrophoresis
(17,
18).
Spots
containing
32p
and/or
a3P
were
located
by
radioautography,
and
both
32p
and
33P
were
quan-
titated.
After
elution
from
the
paper,
digestion
with
venom
phosphodiesterase,
and
separation
of
the
5'-mononucleotides,
the
base
composition
of
each
oligonucleotide
spot
was
estab-
lished
by
determination
of
the
relative
amount
of
33P
in
each
mononucleotide
(82P
was
only
in
pA).
The
sequences
of
the
di-
through
hexanucleotide
32P-containing
oligonucleotides
de-
duced
from
these
data
are
shown
in
Table
2.
Two
other
means
were
also
used
to
confirm
these
data:
(i)
The
position
of
an
oligonucleotide
in
these
two-dimensional
electropherograms
is
extremely
characteristic,
and
the
positions
of
each
of
the
oligonucleotides
deduced
here
was
in
agreement
with
previous
maps
(17,
18).
(ii)
The
isolated
32P-containing
tetra-
to
hepta-
nucleotides
and,
separately,
a
partial
venom
phosphodiester-
ase
digest
of
the
pentanucleotide
were
re-run
on
DEAE-paper
at
pH
1.9.
The
ratio
of
the
distance
between
an
oligonucleo-
tide
and
its
first
degradation
product
(X)
to
the
distance
of
the
oligonucleotide
from
the
origin
(Y)
gives
a
value
(M
=
X/Y)
that
is
characteristic
of
the
3'-terminal
nucleotide
that
has
been
removed
(17,
18).
The
M
values
listed
in
Table
2
agree
with
the
deduced
sequences.
A
single
32P-labeled
spot
predominated
in
each
size
class
from
the
di-
through
the
pentanucleotide.
In
each
case,
this
spot
contained
at
least
60%
of
the
recovered
82p,
no
other
spot
contained
more
than
10%.
Since
most
of
the
32p
occurred
in
one
spot,
the
same
pentanucleotide
sequence
probably
occurs
at
each
5'-terminus
generated
by
the
endonuclease.
Two
spots
labeled
with
32p,
each
containing
only
A,
C,
and
T,
were
detected
in
the
hexanucleotide
fingerprint.
The
amount
of
38P
recovered
in
these
spots
was
insufficient
to
accurately
determine
the
relative
proportion
of
each
of
the
nucleotides,
but
the
hexanucleotide
base
sequences
were
de-
termined
from
their
electrophoretic
mobilities
and
M
values
on
DEAE-paper
at
pH
1.9
(18).
These
data
indicate
that
only
A
or
T
was
found
at
the
sixth
position
internal
from
the
5'-
terminus.
The
heptanucleotide
fraction
gave
at
least
three
'2P-containing
spots
after
two-dimensional
electrophoresis,
two
of
which
contained
G
in
addition
to
the
other
nucleotides.
Although
eight
oligonucleotide
spots
are
to
be
expected
if
the
seventh
position
is
completely
degenerate,
we
probably
would
not
have
been
able
to
detect
them
all
because
as
the
chain
length
(n)
of
an
oligonucleotide
increases,
it
becomes
more
difficult
to
separate
the
various
isomers
due
to
their
increase
in
number
4n
and
the
decrease
in
relative
electrophoretic
mobility
differences.
We
estimate
that
a
maximum
of
five
heptanucleotides
could
be
resolved
under
our
conditions.
RSV
DNA
polymerase
experiments
The
5'-terminal
sequence
described
above
can
be
arranged
in
four
ways
(Fig.
1,
a-d),
depending
on
the
topology
of
the
phosphodiester-bond
cleavages.
The
finding
of
Mertz
and
Davis
(19)
that
the
RI
endonuclease
makes
a
short
cohesive
end
suggested
that
if
the
break
consisted
of
a
3'-hydroxyl
end
and
a
protruding
5'-single-strand
end
(possibilities
la
or
lb),
it
would
serve
as
a
primer-template
for
a
DNA
polymerase.
The
recent
characterization
of
the
DNA
polymerases
from
RNA
tumor
viruses
(e.g.,
Rous
Sarcoma
Virus)
showed
that
they
catalyze
repair-like
reactions
on
such
templates,
and
are
free
of
the
two
exonuclease
activities
associated
with
E.
coli
DNA
polymerase
I
(8,
9).
RSV
DNA
polymerase
(8)
was
used
to
study
the
incorpora-
tion
of
various
combinations
of
[a-3"P]-
and
nonradioactive-
nucleoside
triphosphates
into
X
DNA
cleaved
with
RI
endo-
nuclease;
the
products
were
analyzed
by
nearest-neighbor
transfer
experiments
(Table
3).
Increasing
the
time
of
incu-
bation
or
the
amount
of
enzyme
in
the
standard
reaction
did
not
appreciably
change
the
incorporation
nor
the
nearest-
neighbor
data.
Since
significant
incorporation
was
detected
for
RI-treated
X
DNA
above
untreated
X
DNA
controls
(Expts.
1
and
3-6),
the
two
arrangements
of
the
RI
ends
shown
in
Fig.
lc
and
d
are
not
possible
because
they
would
not
serve
as
primer-templates
for
RSV
DNA
polymerase.
It
is
possible
to
detect
incorporation
into
the
RI
cohesive
ends,
even
though
there
are
two
X
cohesive
ends
(twelve
nu-
cleotides
long)
and
ten
RI
cohesive
ends
(four
nucleotides
long)
per
X
DNA
molecule.
The
reason
is
evident
from
inspec-
tion
of
the
nucleotide
sequences
of
the
X
cohesive
ends
(20),
which
indicate
that
in
the
absence
of
GTP
and
CTP
only
one
A
residue
can
be
incorporated
per
X
genome.
Therefore,
for
Expts.
1-6
the
expected
pmol
of
["2P]NTP
incorporated
per
pmol
of
X
DNA
from
the
X
cohesive
ends
would
be
1,
0,
1,
0,
1,
and
4,
respectively,
compared
to
20,
0,
20,
20,
40,
and
40
for
the
RI
cohesive
ends.
This
prediction
is
consistent
with
the
data
in
Table
3
where
the
shorter
RI
cohesive
ends,
which
con-
tain
only
A
and
T
but
which
occur
more
frequently,
accounted
for
85-97%
of
the
incorporation
in
the
absence
of
CTP
and
GTP.
The
incorporation
and
nearest-neighbor
data
of
Table
3
(Expts.
1-5)
confirm
the
facts
that
the
RI
restriction
endo-
nuclease
makes
a
3'-hydroxyl
5'-phosphoryl
break,
forms
Proc.
Nat.
Acad.
Sci.
USA
69
(1972)
Nucleotide
Sequence
of
RI
Restriction
Site
3451
TABLE
3.
Nearest-neighbor
analysis
of
nucleotides
incorporated
into
RI
endonuclease-treated
X
DNA
by
RSV
DNA
polymerase
pmol
of
[32P]NTP
incorporated
per
pmol
of
[32p]_
pmol
X
DNA
NTP
incor-
Products
of
micrococcal
RI-
porated
per
nuclease
&
spleen
Labeled
Unlabeled
treated
pmol
RI
Endt
phosphodiesterase
digestion:
Deduced
ubstrate
substrate
X
DNA
X
DNA
Exp.
Theory
C*
A*
G*
T
sequence
s
1.
ppp*A
14.7
2.6
1.2
2
<0.02
1.2
1.0
<0.02
G½*A*A
2.
ppp*T
1.4
0.81
0.06
0
3.
ppp*A
PPPT
25.2
0.69
2.4
2
<0.02
1.3
1.0
<0.04
G1AA
4.
ppp*T
pppA
15.6
1.3
1.4
2
<0.02
1.1§
<0.02
1.0
G
pApAp*TP*T
5.
pp*A
31.0
1.4
3.0
4
<0.02
1.9
1.0 1.0
G'pAp*ApTpT
PP*T
6.
pp*A
68.7
9.1
5.9
4
<0.02
1.6
1.0
1.1
G
p*Ap*ApTp*T
PPPTs
pppGx
7.
ppp*C
pppA
8.9
10.5
<0
0
-
PPPU
p
Conclusion:
G
pApApTpT
RI
endonuclease-treated
and
untreated
X
DNA
were
used
as
primer-templates
for
the
incorporation
of
[a-'2P]-and/or
nonradioactive-
nucleoside
triphosphates
with
RSV
DNA
polymerase.
The
asterisks
show
the
position
of
the
32P
atom.
The
arrow
indicates
the
position
of
the
RI
endonuclease
break.
t
The
ratio
of
pmoles
was
calculated
on
the
assumption
of
10
RI
ends
per
X
genome
after
subtracting
the
incorporation
into
untreated
X
DNA.
The
value
listed
under
'Theory'
is
deduced
from
the
final
sequence.
t
The
nearest-neighbor
analyses
are
listed
for
the
RI-treated
X
DNA.
§
In
Exp.
4,
the
ratio
of
A:
to
T*
varied
between
1.1
and
1.9
in
different
experiments.
a
cohesive
end,
and
that
the
sequence
of
the
first
four
nucleo-
tides
at
the
5'
end
are
pApApTpT,
as
already
determined
from
the
polynucleotide
kinase
experiments
described
above.
Fur-
thermore,
the
nucleotide
on
the
3'
side
of
the
break
is
unique
since
[a-32P]ATP
transferred
label
only
to
Gp
(other
than
to
the
A.
known
to
be
incorporated
into
the
cohesive
end).
These
experiments
do
not
distinguish
between
the
two
possible
configurations
depicted
in
Fig.
la
and
b.
However,
in
Expts.
6
and
7
(Table
3),
the
incorporation
data
are
more
compatible
with
those
predicted
from
configuration
(a),
and
the
nearest-neighbor
analysis
of
Exp.
6
provides
additional
support
for
this
configuration.
In
addition,
the
Tm
(560)
determined
for
the
RI
cohesive
termini
is
compatible
with
this
arrangement
(19).
Therefore,
the
combined
data
prove
that
the
topology
of
the
breaks
is
as
shown
in
Fig.
la.
The
3'-terminal
dinucleotide
sequence
The
dinucleotide
at
the
3'-hydroxyl
side
of
the
break
was
analyzed
by
labeling
RI
endonuclease-treated
X
DNA
with
[a-32P]ATP
and
unlabeled
TTP
by
the
use
of
RSV
DNA
poly-
merase.
The
3'-terminally
labeled
DNA
was
exhaustively
digested
with
micrococcal
nuclease,
and
the
82P-labeled
di-
nucleotides
were
isolated
by
column
chromatography
or
paper
electrophoresis.
The
dinucleotides
NpGp,
G*A*
and
ApAp
are to
be
expected
from
the
previous
labeling
data
and
the
known
sequence
(.
GpApATpT).
We
detected
ApA
which
occurs
in
relatively
high
proportion
(8.5%)
in
a
limit
micrococcal
nuclease
digest,
but
did
not
detect
GpAp
because
it
occurs
in
extremely
low
amount
(0.09%).
There
is
a
large
difference
in
occurrence,
especially
between
isomeric
pairs,
of
each
dinucleotide
in
an
exhaustive
micrococcal
nuclease
digest
due
to
the
specificity
of
the
nuclease
(our
unpublished
results).
Of
the
four
possible
dinucleotides
at
the
3'
end
of
the
RI
endo-
nuclease
break,
which
all
occur
in
relatively
high
frequency
in
the
digest
AAGp
(21%),
TpGp
(19%),
CpGp
(9%),
and
GpGp
(2%),
only
ApGp
and
TpGp
were
detected
in
reason-
able
amounts.
The
32p
counts
in
CpGp
and
GpG*
together
accounted
for
less
than
5%
of
the
counts
in
ApG*
plus
TpGp.
The
amounts
of
82p
in
ApG*
and
TpGp
were
almost
equal.
If
the
known
background
incorporation
into
sites
other
than
the
RI
cohesive
ends
is
corrected
for
the
occurrence
of
each
58-----T/A
JpA
A
T
T
C
A/T-----3'
(a)
(b)
(¢)
Cd)
3
----'A/T
C
T T
A
A
G
T/A-----58
5'----AATTCA/T.
T/AGAATT--3'
38-----
T T
A
A
G
T/A
.....
A/T
C
T
T
A
Apt---51
5'
-----A/T
G
A
A
T
T
...
1PIA
A
T
T
C
A/T-----3'
38-----T/A
C
T
T
A
At....T
T
A
A
G
T/A-----5'
58-----T/A
G
A
A
T
4A
A
T
T
C
A/T
----
3'
38-----A/T
C
T
T
A
A
T
A
A
G
TIA-----5'
FIG.
1.
The
four
possible
arrangements
of
the
RI
endonuclease
breaks
are
shown.
The
known
sequence
from
the
polynucleotide
kinase
labeling
is
indicated
by
the
symbols,
the
possible
unknown
by
dots,
and
the
remainder
of
the
double-stranded
DNA
struc-
ture
by
two
parallel
dashed
lines.
Proc.
Nat.
Acad.
Sci.
USA
69
(1972)
3452
Biochemistry:
Hedgpeth
et
al.
dinucleotide
in
the
micrococcal
nuclease
digest
and
sub-
tracted
from
the
82p
found
in
C
G*
and
G G
*
the
difference
is
zero.
Therefore,
of
the
ten
RI
endonuclease
single-strand
breaks
in
X
DNA,
less
than
0.5
have
a
CPG
or
GPG
at
the
3'
side
of
the
break.
The
corrected
data
reduces
this
number
to
zero,
so
that
there
are
five-
breaks
with
TpG
and
five
breaks
with
APG
at
the
3'-hydroxyl
side
of
the
RI
endonuclease
breaks
in
X
DNA.
The
sequence
is
therefore
only
partially
degenerate
(A
or
T)
at
this
position,
in
agreement
with
the
data
from
the
5'-terminally
labeled
hexanucleotides
and
the
topology
of
the
breaks
(Fig.
la).
DISCUSSION
The
5'-terminal
hexanucleotide
sequences
(pAATTCA
and
PAATTCT)
at
the
RI
endonuclease
breaks
in
X
DNA
were
determined
by
labeling
the
ends
with
82p
by
the
use
of
poly-
nucleotide
kinase,
followed
by
digestion
and
separation
of
the
oligonucleotides.
The
5'-terminal
tetranucleotide
(pAATT)
and
3'-hydroxyl
terminal
dinucleotides
(
pApG
and
. .
.PTPG)
were
determined
from
nearest-neighbor
analysis
of
DNA
labeled
by
RSV
DNA
polymerase-directed
incorpora-
tion
of
[a-32P]nucleoside
triphosphates
into
the
short
cohesive
ends
generated
in
X
DNA
by
RI
endonuclease
cleavage.
Con-
sidering
the
anti-parallel
complementarity
of
double-helical
DNA,
these
two
independent
methods
give
the
same
sequence
and
provide
evidence
for
the
nucleotide
sequence
adjacent
to
the
phosphodiester
bonds
in
X
DNA
cleaved
by
the
RI
endo-
nuclease
(Fig.
la
and
Abstract).
The
sequence
is
unique
for
six
base
pairs,
and
the
outside
bases
can
be
either
A
or
T.
Although
we
could
not
determine
with
absolute
certainty
whether
one,
to
a
maximum
of
two,
out
of
ten
RI
endonu-
clease
breaks
in
X
DNA
had
a
different
sequence
by
either
method
above,
the
results
obtained
by
these
two
methods
to-
gether
make
it
extremely
likely
that
the
sequences
around
all
ten
breaks
in
X
DNA
are
identical.
Also,
the
5'-terminal
hexa-
nucleotide
data
and
3'-hydroxyl
dinucleotide
data
are
inde-
pendent
determinations
of
whether
C
and/or
G
can
occur
adjacent
to
the
six
unique
base
pairs
at
each
site.
As
both
estimates
indicate
less
than
one
out
of
ten
(and
probably
zero)
breaks
has
a
C
or
G
at
this
position,
it
appears
that
in
X
DNA
all
five
sites
cleaved
by
RI
endonuclease
are
degenerate
for
only
A
and
T
at
the
outside
bases.
We
conclude
that
in
X
DNA
G
or
C
does
not
occur
on
either
side
of
the
symmetrical
hexanucleotide.
The
sample
size
(ten)
was
large
enough
to
have
detected
G
or
C
at
the
outside
posi-
tions
if
they
occur
randomly
in
X
DNA
and
with
the
expected
frequencies,
i.e.,
25%
(confidence
level
>95%
by
chi-square
test).
This
conclusion
can
be
interpreted
in
one
of
two
ways:
either
the
specific
sequence
cleaved
by
the
endonuclease
is
a
symmetrical
hexanucleotide,
but
in
X
DNA
this
hexanucleo-
tide
is
not
flanked
by
G
or
C,
or
the
specific
sequence
cleaved
is
eight
nucleotides
long
and
the
enzyme
cannot
discriminate
between
A
and
T,
but
can
distinguish
either
from
G
or
C
at
the
outside
positions.
Although
additional
experiments
are
required
to
decide
between
the
two
interpretations,
we
favor
the
latter.
The
RI
sequence
exhibits
180-degree
rotational
symmetry
for
either
of
the
possibilities
just
discussed,
a
result
also
found
in
the
sequence
of
six
base
pairs
restricted
by
the
Hemophilus
influenzae
restriction
endonuclease
(15).
However,
the
ambi-
guity
of
the
RI
sequence
is
different
in
two
respects
from
that
of
the
H.
influenzae
sequence.
(i)
It
occurs
at
the
outside
of
the
sequence
rather
than
centrally,
and
(ii)
It
is
a
purine-
pyrimidine
(A
to
T)
rather
than
a
purine-purine
(A
to
G)
degeneracy.
It
has
been
suggested
that
the
substrate
sym-
metry
in
these
cases
is
related
to
the
subunit
structure
of
the
enzymes
that
interact
with
the
sequence
(2,
15).
It
is
of
inter-
est
that
the
purified
RI
endonuclease
is
composed
of
two
sub-
units