438
Biochemical Society Transactions (2010) Volume 38, part 2
The role of ATP in the reactions of type II DNA
topoisomerases
Andrew D. Bates*1 and Anthony Maxwell†
*School of Biological Sciences, University of Liverpool, Crown Street, Liverpool L69 7ZB, U.K. and †Department of Biological Chemistry, John Innes Centre,
Colney, Norwich NR4 7UH, U.K.
Abstract
Type II DNA topoisomerases catalyse changes in DNA topology in reactions coupled to the hydrolysis of ATP.
In the case of DNA gyrase, which can introduce supercoils into DNA, the requirement for free energy is
clear. However, the non-supercoiling type II enzymes carry out reactions that are apparently energetically
favourable, so their requirement for ATP hydrolysis is not so obvious. It has been shown that many of these
enzymes (the type IIA family) can simplify the topology of their DNA substrates to a level beyond that
expected at equilibrium. Although this seems to explain their usage of ATP, we show that the free energies
involved in topology simplification are very small (<0.2% of that available from ATP) and we argue that
topology simplification may simply be an evolutionary relic.
DNA topology and DNA topoisomerases
One of the consequences of the double-helix structure of
DNA is that the two strands are intertwined and must
be separated for the code to be ‘read’ during transcription
and replication. Since DNA molecules are either closedcircular or periodically tethered, e.g. by attachment to
proteins, strand separation leads to overwinding, or positive
supercoiling, of the DNA. Supercoiling is a common
topological feature of DNA, along with knotting and
catenation (the interlinking of two or more DNA circles)
[1]. Supercoils, knots and catenanes occur naturally as a
consequence of various biological processes. For example,
DNA replication can lead to positive supercoiling, and fully
replicated daughter chromosomes are normally catenated.
Recombination reactions can also lead to knots and
catenanes. Although most cells maintain DNA in a negatively
supercoiled state {underwound, corresponding to a reduction
in the linking number (Lk) of a closed-circular DNA from
that of the unsupercoiled (relaxed) circle [1]}, the other
topological forms, positive supercoiling, knots and catenanes,
are undesirable. DNA topoisomerases are the group of
enzymes that have evolved to resolve these topological
problems, through mechanisms involving transient breakage
and rejoining of the DNA [2,3].
Topoisomerase reactions are of two types: type I enzyme
reactions involve single-stranded breaks in DNA, whereas
the reactions of type II enzymes proceed through doublestranded breaks. In all cases, transient phosphodiester intermediates are formed between active-site tyrosine residues
on the enzyme and the DNA backbone. Type II enzymes
are dimeric and form a 5 -phosphodiester bond with each
Key words: energy coupling, gyrase, strand passage, supercoiling, topoisomerase.
Abbreviations used: G segment, gate DNA segment; Lk, linking number; T segment, transported
DNA segment; topo, topoisomerase.
1
To whom correspondence should be addressed (email [email protected]).
C The
C 2010 Biochemical Society
Authors Journal compilation DNA strand, creating a double-stranded break in a section of
DNA known as the ‘G’ (or ‘gate’) segment, through which
a second double-stranded segment, the ‘T’ (or ‘transported’)
segment, is passed. Intramolecular strand passage results in a
change of supercoiling, specifically a change of ±2 in the
Lk of a DNA topoisomer (a closed-circular molecule of
a given linking number), whereas intermolecular reactions
lead to catenation or decatenation. Type II enzymes are
classified into two subtypes, IIA and IIB, on the basis of
structural and evolutionary considerations [4]. A general
mechanism for type IIA DNA topoisomerases is shown in
Figure 1.
Topology simplification by type II
topoisomerases
All type II enzymes require ATP to carry out their
topoisomerase reactions. In the case of DNA gyrase, the
archetypal bacterial type II enzyme that introduces negative
supercoils, the requirement for ATP is clear: supercoiling is an
endergonic reaction and there is an approximate equivalence
between the free energy of ATP hydrolysis and the extent
of supercoiling that can be catalysed by gyrase (reviewed in
[5]). However, all other type II enzymes also hydrolyse ATP,
even though they perform reactions that are not obviously
energy-requiring, e.g. DNA relaxation, decatenation. This
remained something of a puzzle until 1997, when Rybenkov
et al. [6] showed that these enzymes could actively simplify
the topology of DNA, producing steady-state levels of
supercoiling, knotting and catenation that are lower than
those expected at equilibrium under the same conditions.
For example, bacterial topo (topoisomerase) IV, a close
homologue of gyrase, can produce a narrower distribution
of relaxed topoisomers than topo I (an ATP-independent
enzyme) (Figure 2). This is only possible with an input of
energy, hence the requirement for ATP.
Biochem. Soc. Trans. (2010) 38, 438–442; doi:10.1042/BST0380438
Machines on Genes: Enzymes That Make, Break and Move DNA and RNA
Figure 1 General mechanism of type IIA topoisomerases
Figure 2 Topoisomer distributions demonstrating the
1 The enzyme binds to and bends the G segment (green). 2 A T
segment (red) binds between the N-terminal domains (yellow). 2–3
topology-simplification effect
Calculated topoisomer mole fractions (corresponding to relative
The N-terminal domains dimerize on binding two ATP molecules (*),
allowing the opening of the DNA gate and passage of the T segment. 4
The T segment is dissociated from the bottom of the enzyme, leading to
intensities on an agarose gel) are plotted against the nominal linking
difference (Lk) of the topoisomer (i) for an equilibrated relaxed
distribution (䊏) and a narrowed steady-state distribution (R =1.8), as
unidirectional strand-passage. 4–1 Hydrolysis of the ATP molecules and
loss of the products dissociates the N-terminal domains and resets the
enzyme. Blue and orange represent other domains of the enzymes; the
produced by a type II topoisomerase+ATP (䉬). The continuous lines
represent the Gaussian distributions. The circle size used in the calculation
is that of pBR322 (4361 bp). Lk = 0 corresponds to the most intense
C-terminal domains are not shown. Adapted with permission from [5].
c 2007 American Chemical Society.
topoisomer in the distribution. In this case, the true centre of the
distribution (Lk ◦ ) lies between the Lk = 0 and Lk = −1 topoisomers,
and is the same for both distributions, indicated by the triangle on the
x-axis. Inset: real topoisomer distributions separated on an agarose gel,
formed by topo I (T I) and topo IV+ATP (T IV). Contaminating open-circular
(oc) and linear (lin) DNAs are indicated.
The topology-simplification phenomenon has been shown
to occur for all type IIA topoisomerases tested [6,7], but a
type IIB enzyme, topo VI from Methanosarcina mazei, was
found not to exhibit the effect [7,8]. This suggests that the
ability to perform topology simplification may be due to a
structural difference between type IIA and B enzymes.
One possibility is that topology simplification is dependent on the C-terminal domains of type IIA topoisomerases
(not present in type IIBs), which have been implicated
in substrate selection [9,10]; for example, the C-terminal
domain of the ParC subunit of topo IV has been shown
to be responsible for the enzyme’s preference for relaxing
positive over negative supercoils [9]. However, deletion of
this domain from a number of type IIA topoisomerases did
not influence topology simplification [7], and the type IIA
enzyme from Paramecium bursaria Chlorella virus 1, which
naturally lacks the C-terminal domain, also carries out
topology simplification.
Given that topology simplification requires an input of
energy, we would expect that varying the free energy of ATP
hydrolysis would alter the extent of the effect. However,
using a range of ATP concentrations and ADP/ATP ratios up
to 10:1, no change in the topology-simplification effect was
seen [7]. This result contrasts with experiments carried out
with the gyrase supercoiling reaction, which showed that the
level of supercoiled product was dependent on the ADP/ATP
ratio [11]. The most likely interpretation is that we could not
reduce the free energy of hydrolysis sufficiently to influence
the reaction (see below).
One model that has been suggested for topology
simplification by type IIA topoisomerases is that ATP
hydrolysis might drive the enzyme along a loop of DNA,
trapping a putative T segment in a smaller and smaller
loop, facilitating its transport [6]. This is perhaps inherently
unlikely, as it contradicts our knowledge of the role of ATP in
dimerizing the N-terminal domains of the enzyme (Figure 1).
However, we have tested this model in an experiment in which
the putative tracking of topo IV is prevented by ‘roadblocks’
consisting of inactive mutant EcoRI enzymes. Topology
simplification is unaffected, suggesting that tracking does not
play a role [7].
We also examined the dependence of topology simplification on DNA circle size by measuring the effect with topo
IV and DNA circles from ∼2–9 kb [7]. Certain models of
the mechanism of topology simplification would suggest that
the effect would be enhanced with small circles (see below).
C The
C 2010 Biochemical Society
Authors Journal compilation 439
440
Biochemical Society Transactions (2010) Volume 38, part 2
However, we found little or no dependence of topology
simplification on circle size in this range.
Energetics of topology simplification
Although it is clear that the topology-simplification effect
must require an input of free energy, since the DNA
topoisomer concentrations are being perturbed to a steady
state away from equilibrium, it is important to determine the
amount of energy required. We have calculated the free energy
involved in the effect, using a simulation of the topoisomer
concentrations both at equilibrium and under the steadystate conditions of the topology-simplification reaction. This
is equivalent to a calculation using experimental data, but
allows us to investigate the effect of circle size and other
parameters very easily [7].
Investigations over many years ([12]; reviewed in [1])
have shown that relaxed mixtures of topoisomers formed by
the action of non-ATP-dependent topoisomerases, such as
eukaryotic topo I, have the form of a Gaussian or normal
distribution, indicating that the free energy of supercoiled
topoisomers (Gsc ) has a quadratic dependence on the level
of supercoiling (measured as Lk, the change in Lk from
the completely relaxed situation). This reflects the fact that
supercoiling is an elastic process. In other words:
2
Gsc = K · Lk
where K is an elastic constant. For plasmid DNA circles
above approx. 3000 bp (those normally used as substrates
in topoisomerase reactions), the value of K is inversely
proportional to circle size, N (it is harder to introduce a
supercoil into a smaller circle). So, the value of NK is a
constant for given solution conditions. Under standard in
vitro conditions [12,13]:
NK ≈ 1100 RT bp
where N is the circle size in bp, R is the gas constant, and T
is the absolute temperature.
We can now calculate the distribution of topoisomers. The
relative concentration of a topoisomer, i, in an equilibrium
Gaussian distribution, [i]rel eq , is given by:
K
[i]rel eq = Ae− RT Lki
2
from the equation for a normal distribution with variance
T/2K [1]. A is a constant corresponding to the peak height
(concentration) of the distribution. We can convert this into
a mole fraction of topoisomer i, Pi , by dividing by the total
topoisomer concentration:
Pi
eq
[i]rel eq
= [i]rel eq
i
The topology-simplification reaction of type II topoisomerases results in a steady-state distribution of topoisomers
that is narrower, but still well described by a Gaussian with
a reduced variance. The magnitude of the effect can be
measured by the ratio of the equilibrium to the steady-state
C The
C 2010 Biochemical Society
Authors Journal compilation variance; this value has been labelled R [6] (here called R to
distinguish it from the gas constant). So, we can model the
narrower steady-state (ss) distribution in the same way, but
including the R factor:
[i]rel ss = Ae−
R K
RT
Lki 2
R is approx. 1.8 in our experiments with a range of type IIA
topoisomerases [7]. We can now plot a simulated pair of
distributions, one equilibrated by a topo I, and one in steady
state, as produced by a type II topoisomerase such as topo IV
in the presence of ATP (Figure 2).
To calculate the free energy difference between these
two distributions, we need to sum the changes due to the
change in concentration of each topoisomer (some of which
increase in concentration while some decrease; Figure 2). For
any of the topoisomers, the free energy, Gi is:
Gi = Gi◦ + RT ln[i]
where Gi◦ is the standard free energy of the topoisomer
under the relevant conditions. This is the same for both
distributions, of course, so the molar free energy change for
a given topoisomer is:
Gi = RT ln
[i]ss
[i]eq
This value depends only on the ratio of the concentrations
(or mole fractions). These are molar free energy changes for
each topoisomer, so we need to multiply by the mole fraction
of each topoisomer (Pi ss ), and sum over all the different
topoisomers (in practice, only those with a significant mole
fraction matter) to give the overall free energy change for the
topology simplification reaction:
[i]ss
Pi ss · RT ln
Gss =
[i]eq
i
When we do this calculation based on the simulated
distributions in Figure 2, we find the free energy change has
a surprisingly small value of 0.19 kJ · mol−1 , and this value
is independent of circle size above 3000 bp. This compares
with the free energy available from the hydrolysis of two
molecules of ATP of 100–150 kJ · mol−1 (depending on the
concentrations of the nucleotides; see above), meaning that
the reaction is extremely inefficient in energetic terms; only
approx. 0.2% of the free energy available is transduced into
the product DNA. In contrast, DNA gyrase can produce
a supercoiled distribution of topoisomers with a free energy
of supercoiling of >500 kJ · mol−1 [7]; this is greater than the
free energy available from ATP in one step, and represents
multiple rounds of reaction by gyrase. Similarly, the steadystate distributions of the other type II enzymes presumably
represent the result of multiple rounds of reaction. We
analysed data for the unknotting and decatenation activities
of topo IV [6] in the same way [7], and the excess free energy
of the steady-state distributions formed by topo IV is even
smaller, <0.1 kJ · mol−1 for unknotting, and approx. 0.01
kJ · mol−1 for decatenation.
Machines on Genes: Enzymes That Make, Break and Move DNA and RNA
Figure 3 Geometric models for topology simplification
(A) G-segment bending: a bend in the G segment induced by binding
directs the enzyme towards the inside of the DNA contour (broken
line), leading to selective strand passage of a T segment from the
inside of the contour, and consequent topology simplification (relaxation,
decatenation or unknotting). (B) Hooked juxtapositions: the enzymes
selectively bind to hooked juxtapositions in DNA as G and T segments,
which are more likely to occur in topologically complex molecules (a
catenane in this case, indicated by the broken lines). Strand passage
again leads to simplification.
to the energy released by the hydrolysis of ATP, cast
some circumstantial doubt on the physiological relevance
of the effect. Indeed, if we consider that post-replication
decatenation is probably the primary physiological function
of non-supercoiling type II enzymes [20], it is not clear
that active decatenation will be required. As chromosome
partition pulls apart the catenated daughter molecules, the
equilibrium will in any case be perturbed in the direction
of decatenation [21]. Instead, we propose that topology
simplification is a non-adaptive evolutionary relic, possibly a
consequence of the bending of the G segment by the type IIA
enzymes, which might have evolved from the requirement for
DNA wrapping in the case of DNA gyrase. There is evidence
that bacterial topo IVs evolved from gyrase, although the
situation for eukaryotic topo IIAs is less clear [9,22]. It may
be that the role of ATP hydrolysis in the mechanism of type II
topoisomerases lies elsewhere.
Acknowledgements
We thank Tanya Stuchinskaya for helpful discussions.
Funding
Conclusions
We believe that the available data concerning topology
simplification are most consistent with a model in which
the bending of the DNA G segment by the enzyme
directs the enzyme towards the ‘inside’ of the contour
of a supercoiled or catenated DNA circle (Figure 3A).
The directional strand-passage reaction would then lead
to preferential transport from inside to outside [14]. This
selectivity would drive the reduction of supercoiling or
catenation. X-ray structures of complexes of parts of both
yeast topo II and Streptococcus pneumoniae topo IV with
DNA G segments have revealed strong bending of the
G segment consistent with this model [15,16]. It is not
yet clear whether G-segment bending is a feature of the
type IIB enzymes [8]. A related model envisages that the
enzymes bind preferentially to ‘hooked juxtapositions’ of
DNA strands (Figure 3B), which simulations suggest are
over-represented in knotted and catenated DNA [17–19];
this model would presumably require geometrical selection
of the T segment as well as the G segment. These geometric
models involving DNA bending would suggest that small
circles might show an increased topology simplification
effect. However, experiments in which circles in the size
range ∼2–9 kb were relaxed with topo IV showed little or
no variation in the effect [7]. It is possible, however, that an
enhancement of the effect might only be apparent for circles
<2 kb, and we have shown that measurements of relaxation
in that range are intrinsically problematic [7].
There is no doubt that topology simplification is
a measurable consequence of the action of type IIA
topoisomerases; however, the extremely small free energy
changes induced in the DNA by the enzymes, relative
This work was supported by the Biotechnology and Biological
Sciences Research Council [grant number BB/C517376/1 to A.D.B.
and A.M.].
References
1 Bates, A.D. and Maxwell, A. (2005) DNA Topology, Oxford University
Press, Oxford
2 Champoux, J.J. (2001) DNA topoisomerases: structure, function, and
mechanism. Annu. Rev. Biochem. 70, 369–413
3 Schoeffler, A.J. and Berger, J.M. (2008) DNA topoisomerases: harnessing
and constraining energy to govern chromosome topology. Q. Rev.
Biophys. 41, 41–101
4 Wang, J.C. (1996) DNA topoisomerases. Annu. Rev. Biochem. 65,
635–692
5 Bates, A.D. and Maxwell, A. (2007) Energy coupling in type II
topoisomerases: why do they hydrolyze ATP? Biochemistry 46,
7929–7941
6 Rybenkov, V.V., Ullsperger, C., Vologodskii, A.V. and Cozzarelli, N.R.
(1997) Simplification of DNA topology below equilibrium values by
type II topoisomerases. Science 277, 690–693
7 Stuchinskaya, T., Mitchenall, L.A., Schoeffler, A.J., Corbett, K.D., Berger,
J.M., Bates, A.D. and Maxwell, A. (2009) How do type II topoisomerases
use ATP hydrolysis to simplify DNA topology beyond equilibrium?
Investigating the relaxation reaction of nonsupercoiling type II
topoisomerases. J. Mol. Biol. 385, 1397–1408
8 Corbett, K.D., Benedetti, P. and Berger, J.M. (2007) Holoenzyme
assembly and ATP-mediated conformational dynamics of topoisomerase
VI. Nat. Struct. Mol. Biol. 14, 611–619
9 Corbett, K.D., Schoeffler, A.J., Thomsen, N.D. and Berger, J.M. (2005) The
structural basis for substrate specificity in DNA topoisomerase IV. J. Mol.
Biol. 351, 545–561
10 McClendon, A.K., Gentry, A.C., Dickey, J.S., Brinch, M., Bendsen, S.,
Andersen, A.H. and Osheroff, N. (2008) Bimodal recognition of DNA
geometry by human topoisomerase IIα: preferential relaxation of
positively supercoiled DNA requires elements in the C-terminal domain.
Biochemistry 47, 13169–13178
11 Westerhoff, H.V., O’Dea, M.H., Maxwell, A. and Gellert, M. (1988) DNA
supercoiling by DNA gyrase: a static head analysis. Cell Biophys. 12,
157–181
C The
C 2010 Biochemical Society
Authors Journal compilation 441
442
Biochemical Society Transactions (2010) Volume 38, part 2
12 Depew, R.E. and Wang, J.C. (1975) Conformational fluctuations of DNA
helix. Proc. Natl. Acad. Sci. U.S.A. 72, 4275–4279
13 Rybenkov, V.V., Vologodskii, A.V. and Cozzarelli, N.R. (1997) The effect of
ionic conditions on DNA helical repeat, effective diameter and free
energy of supercoiling. Nucleic Acids Res. 25, 1412–1418
14 Vologodskii, A.V., Zhang, W.T., Rybenkov, V.V., Podtelezhnikov, A.A.,
Subramanian, D., Griffith, J.D. and Cozzarelli, N.R. (2001) Mechanism of
topology simplification by type II DNA topoisomerases. Proc. Natl. Acad.
Sci. U.S.A. 98, 3045–3049
15 Dong, K.C. and Berger, J.M. (2007) Structural basis for gate-DNA
recognition and bending by type IIA topoisomerases. Nature 450,
1201–1204
16 Laponogov, I., Sohi, M.K., Veselkov, D.A., Pan, X.S., Sawhney, R.,
Thompson, A.W., McAuley, K.E., Fisher, L.M. and Sanderson, M.R. (2009)
Structural insight into the quinolone–DNA cleavage complex of type IIA
topoisomerases. Nat. Struct. Mol. Biol. 16, 667–669
17 Buck, G.R. and Zechiedrich, E.L. (2004) DNA disentangling by type-2
topoisomerases. J. Mol. Biol. 340, 933–939
C The
C 2010 Biochemical Society
Authors Journal compilation 18 Liu, Z.R., Mann, J.K., Zechiedrich, E.L. and Chan, H.S. (2006) Topological
information embodied in local juxtaposition geometry provides a
statistical mechanical basis for unknotting by type-2 DNA
topoisomerases. J. Mol. Biol. 361, 268–285
19 Liu, Z.R., Zechiedrich, E.L. and Chan, H.S. (2006) Inferring global
topology from local juxtaposition geometry: interlinking polymer
rings and ramifications for topoisomerase action. Biophys. J. 90,
2344–2355
20 Wang, J.C. (2002) Cellular roles of DNA topoisomerases: a molecular
perspective. Nat. Rev. Mol. Cell Biol. 3, 430–440
21 Holm, C. (1994) Coming undone: how to untangle a chromosome. Cell
77, 955–957
22 Forterre, P., Gribaldo, S., Gadelle, D. and Serre, M.C. (2007) Origin and
evolution of DNA topoisomerases. Biochimie 89, 427–446
Received 10 July 2009
doi:10.1042/BST0380438