Download Mechanism, and Role in Recombination Type-1

Downloaded from symposium.cshlp.org on December 20, 2012 - Published by Cold Spring Harbor Laboratory
Press
Escherichia coli Type-1 Topoisomerases: Identification,
Mechanism, and Role in Recombination
F. Dean, M.A. Krasnow, R. Otter, et al.
Cold Spring Harb Symp Quant Biol 1983 47: 769-777
Access the most recent version at doi:10.1101/SQB.1983.047.01.088
References
This article cites 45 articles, 19 of which can be accessed free at:
http://symposium.cshlp.org/content/47/769.refs.html
Article cited in:
http://symposium.cshlp.org/content/47/769#related-urls
Email alerting
service
Receive free email alerts when new articles cite this article - sign up in
the box at the top right corner of the article or click here
To subscribe to Cold Spring Harbor Symposia on Quantitative Biology go to:
http://symposium.cshlp.org/subscriptions
Copyright © 1983 Cold Spring Harbor Laboratory Press
Downloaded from symposium.cshlp.org on December 20, 2012 - Published by Cold Spring Harbor Laboratory
Press
Escherichia coli Type-I Topoisomerases: Identification, Mechanism,
and Role in Recombination
F. DEAN,* M.A. KRASNOW,* R. OTTER,* M.M. MATZUK,? S.J. SPENGLER,t AND N.R. COZZARELLI?
*Department of Biochemistry, University of Chicago, Chicago, Illinois 60637; ?Department of
Molecular Biology, University of California, Berkeley, California 94720
Topoisomerases catalyze the concerted breakage and
reunion of DNA strands. The energy of the broken
phosphodiester bond is conserved in a covalent linkage
of the enzyme to DNA, allowing the break to be resealed in the absence of an external energy source.
Topoisomerases carefully control the breakage and reunion process, never leaving the DNA while it is broken. These enzymes carry out diverse reactions including supercoiling, catenation, knotting, recombination,
and aspects of DNA replication. Type-2 topoisomerases
make a transient double-stranded break and pass duplex
DNA through the break. This changes the topological
linking number of DNA (Lk) in steps of two, as explained by the sign-inversion mechanism (Brown and
Cozzarelli 1979; Liu et al. 1980; Gellert 1981). Type-1
topoisomerases, on the other hand, break only one
strand at a time and alter the Lk in steps of one (Champoux 1978; Wang and Liu 1979; Cozzarelli 1980b).
In this paper we focus on the type-1 topoisomerases of
Escherichia coli, enzymes that have recently received
increased attention. Five topics are considered. First, in
the analysis of the electrophoretic mobility of negatively
supercoiled topoisomers, we discovered that adjacent
DNA bands on a gel need not differ by one in linking
number. Second, we identified a new E. coli topoisomerase, topoisomerase III. Third, the mechanism of
catenation of duplex DNA rings by E. coli topoisomerase I has been elucidated. Fourth, we describe the
involvement of a type- 1 topoisomerase, the resolvase of
Tn3, in site-specific recombination. In catenation and
recombination by type-I topoisomerases, duplex DNA
is topologically rearranged via the breakage and rejoining of DNA a single strand at a time. Fifth, we consider
what cellular processes might be particularly suited for
the involvement of type-l, but not type-2, topoisomerases.
composite acrylamide-agarose gel (Peacock and Dingman 1968). Nearly all of the topoisomers from fully
relaxed to native are resolved well by a single gel, and
thus a small change in topoisomer distribution can be
discerned.
Using this high-resolution system, we discovered that
neighboring electrophoretic bands do not always differ
in Lk by one. As the Lk of pAO3 DNA is decreased,
topoisomer mobility initially increases, as expected.
However, as the number of titratable negative supercoils increases beyond seven, there is a reduction in
electrophoretic mobility (Fig. 1). The reduction is not a
general consequence of an increase in superhelical density beyond a critical point, because as greater numbers
of positive supercoils are introduced, electrophoretic
mobility continues to rise (Fig. 1). We suggest instead
that increased underwinding stabilizes a DNA structure
such as a cruciform (Lilley 1980; Panayotatos and Wells
1981). Since extrusion of a DNA segment into a cruciform reduces the length of the DNA in the circle while
I
i
[
I
Positive
20
I
I
.,tl"
j-
15
N e g o t i v e ~
/,,
,o5 /~',' f"
o
I
0
Assay of Topoisomerases and the Relationship of
Linking Number to Electrophoretic Mobility
I
25
t
2
I
4
I
6
ILko-Lkl
I
8
I
I0
I
12
Figure t. Electrophoretic mobility of plasmid pAO3 topoisomers.
Samples of the 1683-bp pAO3 DNA (Oka et al. 1979) were
prepared containing various amounts of positive (0) and negative
(9 supercoils. After electrophoresis in a composite acrylamideagarose gel (Peacockand Dingman 1968), the DNA was visualized
by ethidium bromide staining. The absolute value of the difference
in the Lk between relaxed and topoisomer DNAs (ILko-Lkl) is
plotted vs. electrophoretic mobility in arbitrary units, with a
mobility of zero corresponding to the position of nicked DNA.
Topoisomer assignments were confirmed using two-dimensional
gel electrophoresis in which the second dimension was in the
presence of the intercalating dye, chloroquine (Shure et al. 1977).
Topoisomerases are usually assayed by the relaxation
of supercoils in a plasmid, as measured by the production of topoisomers with reduced electrophoretic mobility (Crick et al. 1979). However, individual topoisomers
of native plasmids in the size range usually employed
(4-6 kb) are not well resolved by agarose gel electrophoresis, and only a large reduction in supertwist
density can be detected. For a sensitive assay of relaxation, we have used the 1683-bp plasmid, pAO3, and a
769
Downloaded from symposium.cshlp.org on December 20, 2012 - Published by Cold Spring Harbor Laboratory
Press
770
D E A N ET AL.
Table 1. E. coli Topoisomerases
Name
I (o~)
Structure
II (gyrase)
a2~2
II'
cl2~2
topA
gyrA, gyrB
gyrA, gyrB?
~1
c~2
tnpR
III
Resolvase
c~1
Gene
?
Protomer
molecular
weight
( • 10-3)
Type
Relaxes
positive
supertwists
101
105; 95
105; 50
75
1
2
2
1
no
no
yes
?
+
+
-
+
-
21
1
?
-
-
Inhibited by
Nal Novo
The a and Bdesignationsfor protomercompositiondo not indicategeneticrelatedness(shownin a
separatecolumn). The structuresindicatedfor topoisomeraseHI and resolvaseare likely,but not as
rigorouslydemonstratedas for the otherenzymes.OnlyE. coli strainsthatharborTn3, 3,6,or related
transposonshave resolvase.Excludedfrom the table are phage-encodedtopoisomerases.Nal and
Novoindicatenalidixicacidand novobiocin,respectively.
the Lk remains constant, the negative superhelical density of the DNA in the circle drops. The DNA is then
less compact and migrates more slowly than a topoisomer that is less underwound.
A New E. coli Topoisomerase
The properties of the three previously identified E.
coli topoisomerases are summarized in Table 1. Topo-
isomerase I (co protein) is a 101,000-dalton polypeptide
encoded by topA (Wang and Liu 1979; Sternglanz et al.
1981; Trucksis and Depew 1981). It is a type-1 enzyme
that relaxes negatively supercoiled DNA. Topoisomerase II (DNA gyrase) is a type-2 enzyme that introduces negative supercoils into DNA at the expense of
ATP and slowly removes them in the absence of ATP
(Cozzarelli 1980a). The subunit encoded by gyrA dictates the sensitivity of gyrase to nalidixic acid. The
gyrB-coded subunit is responsible for ATP binding and
is the target for novobiocin. Topoisomerase I I ' (Brown
et al. 1979; Gellert et al. 1979) contains the gyrA
subunit and is therefore sensitive to nalidixic acid. It
also contains a subunit that is homologous to only part
of the B subunit and cannot introduce negative supercoils. It is a type-2 topoisomerase that is unique among
E. coli topoisomerases in its ability to relax both
negatively and positively supercoiled DNAs (Brown et
al. 1979).
We have discovered a new topoisomerase, topoisomerase III. The enzyme is a type-1 topoisomerase
that relaxes negatively supercoiled DNA. Electrophoresis of highly purified material in the presence of SDS
reveals predominantly a pair of polypeptides, 77,000
and 75,000 daltons, that comprise 80% of the preparation. The ratio of the two varies in separate prepara-
tions. They are the only polypeptides coincident with
activity when the enzyme is analyzed by velocity
sedimentation. The identity of their partial proteolysis
patterns (Cleveland et al. 1977) is good evidence that
they are encoded by a single gene and differ only in processing. The sedimentation value of 5.0S is consistent
with a molecular weight of about 65,000 (Martin and
Ames 1961), and thus topoisomerase III very likely consists of a single polypeptide of about 75,000 daltons.
Several characteristics distinguish topoisomerase III
from topoisomerase I. First, antibodies to topoisomerase I (Trucksis and Depew 1981) do not inhibit topoisomerase III. Second, topoisomerase III is present in a
topA deletion strain, DM800 (Sternglanz et al. 1981),
that has no detectable topoisomerase I activity. DM800
is a convenient strain for the purification of topoisomerase HI because the two enzymes behave similarly
in several purification steps. Third, the partial proteolysis patterns of the topoisomerases are entirely different. Fourth, the two enzymes have different sequence
specificity. Incubation of topoisomerase I or III with
DNA in the absence of Mg §247followed by addition of
SDS, resulted in cleavage of the DNA and attachment of
the enzyme to the resulting broken end (Liu and Wang
1979). The cleavage sites were mapped using singlestranded [5'-32p]DNA. Display of the cleavage products by polyacrylamide gel electrophoresis showed no
common cut sites (Fig. 2). Thirty topoisomerase I
cleavage sites on single-stranded DNA have been sequenced (Tse et al. 1980; F. Dean and N. Cozzarelli,
unpubl.). All have a cytosine residue four nucleotides to
the 5 ' side of the break (Fig. 3). The breaks produced
by topoisomerase III lack cytosine in this position.
Topoisomerase III is clearly distinguished from topoisomerases II and I I ' on the basis that it is type 1, insen-
Figure 2. Single-strandedDNA cleavage specificity of topoisomerases I and III. The substrate was a
687-base fragment of single-stranded DNA prepared from q~X DNA. Reactions (10 #1) contained
30% glycerol, 20 nu4 Tris-HCl (pH 7.6), 20 mM
KCI, 0.5 nua dithiothreitol, 50 #g/ml of serum
albumin, [5 '-a2p]DNA, and 15 ng of topoisomerase
I or 6 ng of topoisomerase III. After 25 min at
52~ SDS was added to 1%. Samples were analyzed by polyacrylamide (12.5%) gel electrophoresis (Maxam and Gilbert 1980).
Downloaded from symposium.cshlp.org on December 20, 2012 - Published by Cold Spring Harbor Laboratory
Press
E. COLI TYPE-1 TOPOISOMERASES
cut| site
5'
-6 -5 -4-3 -2 -I ~, I 2 3 4
A
8
5
0
713
G
5
16
0
6
C
8
130
9
T
9
8
08
9
II
9
li
7
7
:>
5
6
6
9
2
I0
2
6
8
6
8
912
9
5
8
4
3j
3'
[ nicked
/o.A
S
=
Figure 3. Topoisomerase I cleavage sequences for single-stranded
DNA. The incidence of nucleotides around 30 topolsomerase I
cleavage sites (Tse et al. 1980; F. Dean and N. Cozzarelli, unpubl.) are compiled. The common attribute is the presence of a
cytosine residue four nucleotides to the 5' side of the cut site.
Z
,!
O(
sitive to nalidixic acid and ATP, and much smaller.
Topoisomerase III is therefore a new topoisomerase; its
properties are summarized in Table 1. It has approximately one quarter of the supercoil relaxation activity of
topoisomerase I and is the second most active E. coil
topoisomerase.
Mechanism of Type-I Topoisomerases
The discovery of catenation and knotting of intact
duplex DNA rings provided strong evidence for the
sign-inversion model for type-2 topoisomerases because
these reactions require the passage of duplex DNA
segments through each other (Brown and Cozzarelli
1979; Liu et al. 1980). The simple connection of these
reactions to type-2 topoisomerases was unsettled by the
discovery that type-1 enzymes could also catenate and
knot DNA (Tse and Wang 1980; Brown and Cozzarelli
1981). The resolution of this paradox comes from a consideration of DNA substrate specificity. Catenation requires a donor circle and a recipient circle that is transiently broken by the enzyme. For type-1 enzymes, the
recipient must be nicked, whereas the donor can be
supercoiled, nicked, or relaxed (Tse and Wang 1980;
Brown and Cozzarelli 1981). Neither specificity is expected for topoisomerase I, which binds preferentially
to negatively supercoiled DNA (Liu and Wang 1979).
However, a simple scheme explains these requirements
(see Fig. 4). We postulate that a type-1 topoisomerase
introduces a transient single-stranded break opposite the
771
0
intact
,
0
30
Total~, final
,
I0
Figure 5. Binding of topoisomerase I to DNA. ColEI DNA was
treated with DNase I in the presence of ethidium bromide to introduce approximately one nick per molecule (Greenfield et al.
1975). Both the nicked DNA ( 9 and intact native ColE1 DNA
(O) were linearized with EcoRI restriction enzyme and end-labeled
using [7-32p]ATP and polynucleotide kinase. The 10-/zl reactions
containing 30% glycerol, 20 mM Tris-HC1 (pH 7.6), 20 mM KC1,
0.5 mM dithiothreitol, 50 #g/ml of serum albumin, 6.8 fmoles of
DNA, and topoisomerase I (~0) were incubated at 37~ for 25 min.
Reactions were diluted 100-fold in binding buffer (Fennewald et al.
1981) and applied dropwise to a nitrocellulose filter (BA85;
Schleicher and Schuell). The filters were washed with 1 ml of binding buffer, dried, and counted.
nick in the recipient ring. The donor ring passes through
the resultant double-strand interruption; resealing the
nick forms a catenane. We demonstrate below that the
postulated operation of a type-1 enzyme opposite a nick
is correct.
Topoisomerase I has a higher affinity for singly
nicked plasmid DNA than for intact DNA (Fig. 5).
Fivefold more enzyme is required to bind 1 fmole of intact DNA to a nitrocellulose filter than is required to
bind 1 fmole of nicked molecules. Thus, a single nick in
a 6.6-kb DNA creates an opportunity for a novel interaction with topoisomerase I.
The binding to a nick cannot be a kinetic dead end
because it directs cleavage to the region. An 1108-bp
fragment of ~bX174 DNA was prepared with 32p at the
5' end of the viral ( + ) strand and a single nick in the
9
\
bindopposite
nick
Figure 4. Model for catenation by type-1 topoisomerases. The enzyme, depicted as a C-shaped structure, binds to a nick in a recipient DNA ring. The
strand opposite the nick is cleaved, with the 5' end
becoming covalently bound to the enzyme and the
3' end remaining bound noncovalently. A segment
from a donor DNA ring passes through the break
created by the enzyme. The transient nick in the
recipient DNA circle is resealed, and the DNA is
released from the enzyme.
passDNA~
throughbreak
( ~
sealbreak
Q
) ~
Downloaded from symposium.cshlp.org on December 20, 2012 - Published by Cold Spring Harbor Laboratory
Press
772
D E A N ET AL.
Hphl
Ava11"
HaeTr
Figure 6. Uniquely nicked ~X DNA. ~X replicative form-I DNA
linearized at the unique Avail site was mixed with a threefold molar
excess of ~X viral ( + ) single-stranded circles. The DNA was
denatured by dialysis against 95% formamide and 10 mM EDTA
(pH 8.3) for 1 hr at 22"C and renatured by dialysis first against
50% formamide, 0.2 MTris-HCl (pH 8.0), and 20 rnM EDTA for
3 hr and then against 0.1 MNaCl.and 10 rnM EDTA (pH 8.0) for 2
hr (Clayton et al. 1970). The resulting nicked duplex DNA circles
were purified by agarose gel electrophoresis. Purified nicked
circles were digested with HphI restriction nuclease and endlabeled using h,-32P]ATP and polynucleotide kinase. The DNA
was then cut with HaelI nuclease, and the 1108-bp fragment containing the Avail restriction enzyme site was isolated.
complementary ( - ) strand 70 bases from the labeled
end (Fig. 6). Cleavages made in the viral strand by
topoisomerase I were displayed by gel electrophoresis
(see Fig. 7). The exact position of the break was determined by comparison with the mobility of standards
produced by chemical cleavage (Maxam and Gilbert
1980) and by restriction enzyme digestion. Only one
cleavage was seen and it was three nucleotides from the
nick. When the substrate was denatured prior to treatment with topoisomerase I, at least 20 cleavage fragments were produced, including the one seen with the
nicked duplex fragment. Unnicked duplex DNA was not
cleaved under these conditions.
Under more vigorous cleavage conditions, other
breaks are made by topoisomerase I (Fig. 8), but all still
cluster in the area of the nick. The nick does not have a
rigid directing effect on cleavage, and thus the two
breaks do not possess a unique stagger as they do with
gyrase cleavage (Morrison and Cozzarelli 1979). In
every case a cytosine residue is four nucleotides to the
5 ' side of the break. Since the same sequence specificity
is observed with single-stranded DNA and nicked DNA,
we conclude that the enzyme recognizes a similar conformation in each case. The strong binding affinity of
topoisomerase I for negatively supercoiled DNA is likely also due to its preference for an unwound region.
We have suggested a model for relaxation of negative
supercoils by topoisomerase I (Brown and Cozzarelli
1981) that is consistent with the data on binding and
catenation (see Fig. 9). After the unpairing o f the
strands by enzyme binding, one strand can be passed
through a transient break in the other. This will change
the Lk in steps of one. The cleaved strand and the passing strand are not base-paired in this one-strand signinversion mechanism. This is necessary because catenation is unaffected by sequence complementarity. In fact,
topoisomerase I has little specificity for the DNA segment that is passed through a break; it allows a single
strand in the process of relaxation, but it can pass
nicked, relaxed, or supercoiled duplex DNA in catenation.
Figure 7. Specificity of DNA cleavage by topoisomerase I. A
5'-end-labeled DNA fragment with a site-specific nick (Fig. 6)
was incubated in the presence ( + ) or absence ( - ) of enzyme (E)
as described in Fig. 5. Cleavage was induced by the addition of
SDS to 1% (Liu and Wang 1979). Sequencing ladders were created
by chemical cleavages (Maxam and Gilbert 1980) at the residues
indicated at the top of the first three lanes. The two bands in the
- E lane are from restriction by Avail and Hphl enzymes and provide 65-bp and 70-bp length standards. The additional band created
by topoisomerase I has the 3'-terminal sequence pApCpCpTpT.
5'-A- C-c-
T- T!G-C-T-
- T - G pG - A - A -
C-G-A-
G-C~-T-A_
C-G-
A-T-T-
A~- A _
T-5'
Figure 8. Topoisomerase I cleavage sites near a specific nick in
DNA. The substrate for cleavage by topoisomerase I is shown in
Fig. 6. Thin arrows indicate the cleavage sites, and the heavy arrow indicates the preferred site. There is a cytosine residue four
nucleotides to the 5' side of each cut.
Downloaded from symposium.cshlp.org on December 20, 2012 - Published by Cold Spring Harbor Laboratory
Press
E. COLI TYPE-1 TOPOISOMERASES
773
k
nick )
~/
tnp R)
~~ ligote~~
Figure 9. A one-strand sign-inversion mechanism for type-1
topoisomerases. Topoisomerase I binds the DNA such that the
double helix is unwoundand a node is formed by the crossingof
the two singlestrands. The enzymetransientlynicks one strand, remains bound to each broken end, and passes the other strand
through the break. Sealingthe break results in a change in linking
fixed at +l.
Role of a Type-1 Topoisomerase in Recombination
A more complex case of controlled breakage and reunion occurs in recombination where the segments
flanking a pair of double-strand breaks are rearranged
before rejoining. There is evidence that a topoisomerase
can be responsible for this central event in recombination. The bacteriophage X Int protein is necessary for the
site-specific recombination that mediates the integration
and excision of k DNA from the E. coli chromosome.
Int is a true topoisomerase since it removes supercoils
from circular plasmid DNA (Kikachi and Nash 1979).
We have begun a study of the breakage and reunion
reaction of a site-specific recombination involved in
DNA transposition. This is part of our continuing study
RSFI050
I
,
F ~ r e 10. Role of Tn3 genes in transposition. The transposition
of Tm~(heavyarrow) from one replicon(
) m another ( ~,~ )
is shown, t~A and mpR are the only required Tn3 genes, tnpA is
needed for formation of a cointegrate structure in which the two
replicons are joined by Tm~ bridges, and tnpR is necessary for
resolution of the two replicons. The final products are catenated
here, as in vitro, but they are found separated in vivo.
(Fennewald et al. 1981) of the mechanism of transposition of the E. coli transposon, Tn3 (Heffron 1982).
There are two steps in the transposition of Tn3 from one
replicon to another (Fig. 10) (Shapiro 1979). The first is
the complex event in which the transposon is duplicated
and the replicons are fused via directly repeated
transposon bridges. Cointegrate formation requires the
product of a single Tn3 gene, tnpA, and unidentified
host factors. The second step is a site-specific recombination between the transposons that regenerates the
two replicons, each now containing Tn3. The second
reaction has been shown for the related transposon, -y~i,
to require only the product of the tnpR gene, resolvase
(Reed 1981; Reed and Grindley 1981).
The tnpA- and tnpR-gene products of Tn3 were cloned
under the control of the ), P, promoter and cI repressor
as illustrated in Figure 11. This allows the induction of
ill
tnpA
pMC1426
tnpR
,L
?npA
to
~
to
pk8
N PL
,
,
, [
tnpR
I
blo
,
bla
~.
b/o
Clo [
Figure 11. Cloning of mpA and tnpR genes of Tn3 under control of the bacteriophage k promoter, PL. Parental plasmidsare illustrated
linearizedat the unique EcoRI restrictionenzymesite. Vector pX8 (Reed 1981) was digested with C/aI nucleasefor tnpA insertionor C/aI
and PvuI restrictionenzymesfor mpR insertion. In the resultantplasmids, thin lines indicatethe vector sequencesand the boxed areas are
the cloned sequences. (I-Z) tnpR; ( 1 ) mpA.
Downloaded from symposium.cshlp.org on December 20, 2012 - Published by Cold Spring Harbor Laboratory
Press
774
D E A N E T AL.
o
§
~
-
§
-
+
-
§
-
r e s o l v a s e
NmimJ
-BamHI
minutes
0 2 5 15 60240
lille,
+BomHI
Figure 13. Specificity of relaxation by resolvase. Relaxation by
Tn3 resolvase was assayed using pBR322, which has no res site or
derivatives containing one or two res sites. Arrows indicate the
number and relative orientation of the sites. Reaction conditions
were as described in Fig. 12.
~7
V
~4iBm
;
Figure 12. Time course of DNA recombination and relaxation by
resolvase, pRR51 (1.8 #g). a 5750-bp plasmid containing two
directly repeated copies of the resolution site (Reed 1981), was incubated with 50 ng of purified Trd resolvase at 37~ in a 192-/zl
reaction mixture containing 50 mM NaCI, 20 mM Tris-HCl (pH
7.45), 10 mM MgCI2, and 1 mM dithiothreitol. Samples (16 td)
were removed at the times indicated and either quenched with
0.4% SDS to monitor relaxation (top) or first treated with BamHl
nuclease to monitor recombination (bottom). The products were
electrophoresed through 1% agarose gels. The two resolution sites
in pRR51 are shown as open and closed arrows for ease of
visualization, and the unique BamHI site is indicated by an open
triangle. BamHI linearizes only the smaller of the rings of the product catenane because there are no BamHl sites in the larger
daughter ring.
large quantities of these proteins, and hundreds of
milligrams of resolvase have been prepared (M.A.
Krasnow et al., unpubl.).
The Tn3 resolvase, like its 3,(5 counterpart (Reed
1981), effectively converts a plasmid containing two
directly repeated sites for recombination, called res, into a catenane of the two product rings (Fig. 12). Scission
at the unique B a m H I nuclease site disrupts the catenane
but leaves one product ring circular. Since this ring is
nearly fully supercoiled, resolvase maintains close control of the broken ends during the recombination.
Resolvase was shown to be a topoisomerase by its
removal of supercoils from the substrate (Fig. 12). It
changes the L k in steps of one and is therefore a type-1
topoisomerase; a comparison with other E. coli topoisomerases is shown in Table 1. The time course for the
relaxation and recombination by resolvase was the same
(Fig. 12). A more important correspondence in the two
processes is that like recombination, relaxation required
res sites in the same orientation (Fig. 13). A plasmid
with no sites, one site, or two sites in head-to-head
orientation was not relaxed at all. Three conclusions can
be drawn from this experiment: First, the topoisomerase
activity must be intrinsic to the breakage and reunion
that marks recombination, because both have the same
unique substrate requirements. Relaxation of D N A by
Int did not show any sequence preference (Kikuchi and
Nash 1979), and therefore the role of its topoisomerase
activity in recombination was inferential. Second,
resolvase must have at least two active sites, both of
which have to be occupied for either to break and rejoin
DNA. Third, the two res sites must be brought together
in a fixed orientation so that the direct and inverted configuration can be distinguished.
A mechanism for this distinction was suggested by an
analysis of the interlinking complexity of the catenated
product circles. Using ribbon models for DNA, it can be
shown that intramolecular recombination can convert
the supercoils of the substrate into intertwining of the
product-catenated rings, if the supercoils are plec-
Downloaded from symposium.cshlp.org on December 20, 2012 - Published by Cold Spring Harbor Laboratory
Press
TYPE-1 T O P O I S O M E R A S E S
E. COLI
SYNAPSI~"
4
RECOMBINATION )'
'9
tJ
Figure 14. Intramolecular recombination of a supercoiled
substrate can lead passively to catenation of the product rings. The
two resolution sites are shown as open and closed arrows. Synapsis
is the parallel alignment of the resolution sites. Upon recombination, the interwound supercoils trapped between the recombination
sites can be directly converted to interlinkings of the daughters. If
n is the number of supercoils and M e is the twist lost during
breakage and reunion, then (n--Me)~2 interlockings of product
catenanes can result.
tonemic rather than toroidal (Fig. 14) (Mizuuchi et al.
1980). This results because the DNA segments that are
to be separated by recombination are interwound in the
substrate. This passive model for interlinking predicts
that the catenated products will be largely relaxed and
highly intertwined. Neither is observed. The supercoil
density of the resolved products is nearly that of the
substrate, and the catenanes appear simple and
homogeneous, intertwined only once. Since it is a
reasonable assumption that the substrate supercoils are
plectonemic (Camerini-Otero and Felsenfeld 1978), we
suggest that resolvase removes the intertwining prior to
recombination and actively introduces a single product
interlock. This can be accomplished by a tracking
mechanism for resolution (Fig. 15). One subunit of a
dimeric enzyme initially binds nonspecifically to the
DNA. The enzyme then slides bidirectionally along the
DNA to a r e s site where it remains bound; the second
subunit then repeats the search for a second r e s site.
When both sites are bound, recombination and/or relax-
NON-)'I
SPECIFIC
BINDING
WALK TO
SITE
NON-
)
BINDING
/
TO),
2ndSITE
WALK
775
ation results. The search for the second r e s site will
disentangle the supercoils such that substrate supercoiling will not affect the number of intertwinings of the
product rings. Interlinking would instead be dictated by
the mechanism of the enzyme, and the products need not
be catenated at all. The tracking between the r e s sites
also explains how the enzyme distinguishes the orientation of the r e s sites and thereby operates only when they
are in the same orientation. If instead the second subunit
released DNA and rebound in either orientation, then it
would not distinguish the relative orientation of r e s
sites.
Reactions of Type-I Topoisomerases
Type-2 topoisomerases catalyze the passage of duplex
DNA segments through each other and can therefore
reversibly catenate and knot intact duplex DNA rings.
Because type-1 enzymes make only transient singlestranded breaks, they can (with the exception of sitespecific recombination enzymes) only knot and catenate
rings that have a preexisting break. They seem, however, particularly suited to a number of reactions (Fig.
16). At the termination of replication of circular DNA,
there are topological problems (Gefter 1975) that can be
solved by conversion of the incomplete rings to catenanes (Sundin and Varshavsky 1980), and these catenanes could be separated by type-1 enzymes (Fig. 16a).
The replication of duplex DNA in which the two strands
are covalently joined by a terminal hairpin, such as vaccinia DNA (Geshelin and Berns 1974; Baroudy et al.
1982), could be via single-strand catenanes in which the
hairpins are intertwined (Fig. 16b). Separation requires
a topoisomerase that operates on single strands. Recombination generally proceeds through a Holliday-structure intermediate (Dressier and Potter 1982) in which
duplex DNA segments are joined covalently through
two of the four strands (Fig. 16c). Both the formation
and resolution of such structures could be by a type-1
topoisomerase such as resolvase. Another process im-
) (
RECOMBINE)
Figure 15. Tracking model for resolvase-mediated recombination. A resolvase dimer (stippled area) binds DNA nonspecifically and
translocates randomly along the DNA until reaching ares site (arrow) where it binds tightly and specifically. The second subunit then binds
nonspecificallyto adjacent DNA, trapping a loop of DNA. One-dimensionaldiffusion of the second subunit with respect to the DNA causes
the loop to expand and contract; resolvase serves as a loose annulus. This segregates the supercoils and prevents net interwinding of the
DNA segments between the sites. When the second subunit reaches the second res site in the proper orientation, it is tightly bound, and the
enzyme changes conformation, activating breakage and reunion. Strand exchange is directed by resolvase with a specific geometry,
resulting in singly interlinked recombinants.
Downloaded from symposium.cshlp.org on December 20, 2012 - Published by Cold Spring Harbor Laboratory
Press
776
DEAN ET AL.
(a) Decatenate interrupted circles
(b) Decatenate single-strand loops
-- 9
-- 9
(c) Recombination via Holliday structure
(3
(d) Strand transfer
9169
(e) Rolling circle
nick
replicate
join
Figure 16. Reactionsof type-1 topoisomerases. The 3' and 5' ends of DNA are indicatedby arrowheads and closed circles, respectively.
Parallel lines indicatebase-pairedstrands. (a) Interruptedduplexrings such as those in SV40 DNA replicationare decatenated; (b) terminal
hairpins as may be formed in vacciniaDNA replicationare resolved. (c) Hollidaystructure is formed at a homologousregion betweentwo
circular DNA moleculesand subsequentlyconvertedto a dimeric recombinationproduct; (d) a single strand of DNA is transferred from a
gapped duplex ring to an intact duplex ring to form a D loop; (e) replicationof q~Xand fd viral DNAs. A topoisomerase, cistron-A, or
gene-2 product (1) nicks the replicative form I, but closure is delayed until a single-strandedcircular product is synthesized.
plicated in recombination is the transfer of a singlestranded segment from one duplex to another (Radding
1978) (Fig. 16d). Single-strand passage by a type-1 enzyme would solve the topological problems of such an
intermolecular helix. This is an interpretation of the
facilitation of recA-protein-mediated strand transfer by
E. coli topoisomerase I (Cunningham et al. 1981). This
winding problem also occurs in the annealing of complementary single-stranded circles and this reaction can
be carried out with type-1 enzymes (Champoux 1977;
Kirkegaard and Wang 1978). The ~X cistron-A protein
and the related gene-2 protein of filamentous phages
(Kornberg 1980) are examples of still another topoisomerase-mediated process (Fig. 16e). Here, the singlestrand breakage and reunion reactions are temporarily
interrupted by synthesis of a single-stranded DNA circle. These enzymes relax supercoils and are therefore
topoisomerases (Geider and Meyer 1979; Langeveld et
al. 1980). Thus, type-1 topoisomerases promote a variety of DNA manipulations that allow facile participation
of single strands of DNA in cellular processes.
ACKNOWLEDGMENTS
This work was supported by National Institutes of
Health research grants GM-21397 and GM-22729 from
the National Institute of General Medical Sciences.
F.D., M.A.K., and S.J.S. were supported by National
Institutes of Health National Research Service Awards;
F.D., by grant HD-07136 from the National Institutes
of Child Health and Development; M.A.K., by grant
GM-07281 from the National Institutes of General
Medical Sciences; and S.J.S., by grant CA-09273 from
the National Cancer Institute.
REFERENCES
BAROtJDV, B.M., S. VENKATESAN,and B. Moss. 1982. Incompletely base-paired flip-flop terminal loops link the two
DNA strands of the vaccinia virus genome into one uninterrupted polynucleotide chain. Cell 28:315.
BROWN, P.O. and N.R. COZZARELLI. 1979. A sign inversion
mechanism for enzymatic supercoiling of DNA. Science
206: 1081.
--.
1981. Catenation and knotting of duplex DNA by type 1
topoisomerases: A mechanistic parallel with type 2 topoisomerases. Proc. Natl. Acad. Sci. 78: 843:.
BROWN, P.O., C.L. PEEBLES, and N.R. COZZARELLI. 1979.
A topoisomerase from Escherichia coli related to DNA
gyrase. Proc. Natl. Acad. Sci. 76: 6110.
CAMERIN1-OTERO,R.D. and G. FELSENFELD.1978. A simple
model of DNA superhelices in solution. Proc. Natl. Acad.
Sci. 75: 1708.
CHAMr'OUX,J.J. 1977. Renaturation of complementary singlestranded DNA circles: Complete rewinding facilitated by
the DNA untwisting enzyme. Proc. Natl. Acad. Sci.
74: 5328.
--.
1978. Proteins that affect DNA conformation. Annu.
Rev. Biochem. 47: 449.
CLAYTON, D.A., R.W. DAvis, and J. VINO~RAD. 1970.
Homology and structural relationships between the dimeric
and monomeric circular forms of mitochondrial DNA from
human leukemic leukocytes. J. Mol. Biol. 47: 137.
CLEVELAND,D.W., S.G. FISCHER,M.W. KIRSCHNER,and
Downloaded from symposium.cshlp.org on December 20, 2012 - Published by Cold Spring Harbor Laboratory
Press
E. C O L I T Y P E - 1 T O P O I S O M E R A S E S
U.K. LAEMMLI. 1977. Peptide mapping by limited proteolysis in sodium dodecyl sulfate and analysis by gel electrophoresis. J. Biol. Chem. 252:1102.
COZZARELLI, N.R. 1980a. DNA gyrase and the supercoiling
of DNA. Science 207: 953.
91980b. DNA topoisomerases. Cell 22: 327.
CRICK, F.H.C., J.C. WANG, and W.R. BAUER. 1979. IS DNA
really a double helix? J. Mol. Biol. 129: 449.
CUNNINGHAM, R.P., A.M. WO, T. SmnATA, C. DAsGUPTA,
and C.M. RADDING. 1981. Homologous pairing and
topological linkage of DNA molecules by combined action
ofE. coli recA protein and topoisomerase I. Cell 24: 213.
DP-ESSLER, D. and H. POTTER. 1982. Molecular mechanisms
in genetic recombination. Annu. Rev. Biochem. 51: 727.
FENNEWALD, M.A., S.P. GERRARD,J. CHOU, M.J. CASADABAN, and N.R. COZZARELLI. 1981. Purification of the
Tn3 transposase and analysis of its binding to DNA. J.
Biol. Chem. 256: 4687.
GEFTER, M. 1975. DNA replication. Annu. Rev. Biochem.
44: 45.
GLIDER, K. and T.F. MEYER. 1979. Gene-II protein of bacteriophage fd in enzymatic replication of viral duplex
DNA. Cold Spring Harbor Symp. Quant. Biol. 43: 59.
GELLERT, M. 1981. DNA topoisomerases. Annu. Rev.
Biochem. 50: 879.
GELLERT, M., L.M. FISHER, and M.M. O'DEA. 1979. DNA
gyrase: Purification and catalytic properties of a fragment
of gyrase B protein. Proc. Natl. Acad. Sci. 76: 6289.
GESnELIN, P. and K.I. BERNS. 1974. Characterization and
localization of the naturally occurring crosslinks in vaccinia virus DNA. J. Mol. Biol. 88: 785.
GREENFIELD,L., L. SIMPSON,and D. KAPLAN. 1975. Conversion of closed circular DNA molecules to single-nicked
molecules by digestion with DNase I in the presence of
ethidium bromide. Biochim. Biophys. Acta 4tl7: 365.
HEFFRON, F. 1982. Tn3 and its relatives. In Mobile genetic
elements (ed. J9 Shapiro). Academic Press, New York.
(In press.)
KIKUCm, Y. and H.A. NASn. 1979. Nicking-closing activity
associated with bacteriophage k int gene product. Proc.
Natl. Acad. Sci. 76: 3760.
KIRgEGAARD, K. and J.C. WANG. 1978. Escherichia coli
topoisomerase I catalyzed linking of single-stranded rings
of complementary base sequences. Nucleic Acids Res.
5: 3811.
KORNBERG, A. 1980. DNA replication. W.H. Freeman, San
Francisco.
LANGEVELD, S.A., G.A. VON ACKEL, and P.J. WEISBEEK.
1980. Improved method for the isolation of the A and A*
proteins of bacteriophage r
FEBSLett. 114: 269.
LILLEV, D.M.J. 1980. The inverted repeat as a recognizable
structural feature in supercoiled DNA molecules. Proc.
Natl. Acad. Sci. 77: 6468.
LIu, L.F. and J.C. WANG. 1979. Interaction between DNA
and Escherichia coli DNA topoisomerase I. J. BioL Chem.
254:11082.
LltJ, L.F., C.-C. LIu, and B.M. ALaERTS. 1980. Type II
DNA topoisomerases: Enzymes that can unknot a
topologically knotted DNA molecule via a reversible
double-strand break. Cell 19: 697.
777
MARTIN, R.G. and B.N. AMES. 1961. A method for determining the sedimentation behavior of enzymes: Application to
protein mixtures. J. Biol. Chem. 236: 1372.
MAXAM, A.M. and W. GILBERT. 1980. Sequencing endlabeled DNA with base-specific chemical cleavages9
Methods Enzymol. 65: 499.
MIZUUCHI, K., M. GELLERT, R.A. WEISBERG, and H.A.
NASH. 1980. Catenation and supercoiling in the products of
bactriophage k intergrative recombination in vitro. J. Mol.
Biol. 141: 485.
MORRISON, A. and N.R. COZZARELLI. 19~79. Site-specific
cleavage of DNA by E. coli DNA gyrase. Cell 17: 175.
PANAYOTATOS,N. and R.D. WELLS. 1981. Cruciform structures in supercoiled DNA. Nature 289: 466.
PEACOCK, A.C. and C.W. DING~IAN. 1968. Molecular weight
estimation and separation of ribonucleic acid by electrophoresis in agarose-acrylamide composite gels. Biochemistry 7: 668.
OKA, A., N. NOMURA, M. MORITA, H. SUGISAKI,K. SUGIMOTO, and M. TAKANAMI. 1979. Nucleotide sequence of
small ColEI derivatives: Structure of the regions essential
for autonomous replication and colicin El immunity. Mol.
Gen. Genet. 172: 151.
RADDING, C.M. 1978. Genetic recombination: Strand transfer
and mismatch repair. Annu. Rev. Biochem. 47: 847.
REED, R.R. 1981. Transposon-mediated site-specific recombination: A defined in vitro system. Cell 25: 713.
REED, R.R. and N.D.F. GRINDLEY. 1981. Transposonmediated site-specific recombination in vitro: DNA
cleavage and protein-DNA linkage at the recombination
site. Cell 25:721.
SHAPIRO, J. 1979. Molecular model for the transposition of
bacteriophage mu and other transposable elements. Proc.
Naa. Acad. Sci. 76: 1933.
SHORE, M., D.E. PULLEYBLANK,and J. VINOGRAD. 1977.
The problems of eukaryotic and prokaryotic DNA packaging and in vivo conformation posed by superhelix density
heterogeneity. Nucleic Acids Res. 4:1183.
STERNGLANZ, R., S. DINARDO, K.A. VOELKEL, Y. NISHIMURA, Y. HIROTA, K. BECNERER,L. ZUMSTEIN,and J.C.
WANG. 1981. Mutations in the gene coding for Escherichia
coli DNA topoisomerase I affect transcription and
transposition. Proc. Natl. Acad. Sci. 78: 2747.
SUNDIN, O. and A. VARSHAVSKY. 1980. Terminal stages of
SV40 DNA replication proceed via multiply intertwined
catenated dimers. Cell 21: 103.
TRUCKSlS, M. and R.E. DEPEW. 1981. Identification and
localization of a gene that specifies production of Escherichia coli DNA topoisomerase I. Proc. Natl. Acad. Sci. 78:
2164.
TSE, Y.-C. and J.C9 WANG. 1980. E. coli and M. luteus DNA
topoisomerase I can catalyze catenation or decatenation of
double-stranded DNA rings. Cell 22: 269.
TSE, Y.-C., K. KIRKEGAARD,and J.C9 WANG. 1980. Covalent
bonds between protein and DNA. J. Biol. Chem. 255:
5560.
WANG, J.C. and L.F. LIU. 1979. DNA topoisomerases: Enzymes which catalyze the concerted breaking and rejoining
of DNA backbone bonds. In Molecular genetics (ed. J.H.
Taylor), vol. 3, p. 65. Academic Press, New York.