Single-molecule imaging of DNA curtains reveals
mechanisms of KOPS sequence targeting by
the DNA translocase FtsK
Ja Yil Leea,1, Ilya J. Finkelsteina,1, Estelle Crozatb,2, David J. Sherrattb, and Eric C. Greenea,c,3
aDepartment of Biochemistry and Molecular Biophysics andcHoward Hughes Medical Institute, Columbia University, New York, NY 10032; andbDepartment
of Biochemistry, University of Oxford, Oxford OX1 3QU, United Kingdom
Edited by Nancy E. Kleckner, Harvard University, Cambridge, MA, and approved March 15, 2012 (received for review February 1, 2012)
FtsK is a hexameric DNA translocase that participates in the final
stages of bacterial chromosome segregation. Here we investigate
the DNA-binding and translocation activities of FtsK in real time by
imaging fluorescently tagged proteins on nanofabricated curtains
of DNA. We show that FtsK preferentially loads at 8-bp KOPS (FtsK
Orienting Polar Sequences) sites and that loading is enhanced in
the presence of ADP. We also demonstrate that FtsK locates KOPS
through a mechanism that does not involve extensive 1D diffusion
at the scale of our resolution. Upon addition of ATP, KOPS-bound
FtsK translocates in the direction dictated by KOPS polarity, and
once FtsK has begun translocating it does not rerecognize KOPS
from either direction. However, FtsK can abruptly change direc-
tions while translocating along DNA independent of KOPS, sug-
gesting that the ability to reorient on DNA does not arise from
DNA sequence-specific effects. Taken together, our data support
a model in which FtsK locates KOPS through random collisions,
preferentially engages KOPS in the ADP-bound state, translocates
in the direction dictated by the polar orientation of KOPS, and is
incapable of recognizing KOPS while translocating along DNA.
single molecule|DNA curtains|ASCE translocase|target search|
chromosome dimers that can arise after homologous recombi-
nation (1, 2). Failure to resolve these dimers prevents chromo-
some segregation and leads to cell death. In Escherichia coli the
tyrosine recombinase XerCD promotes dimer resolution at the
28-bp dif site within the replication termination region. FtsK
stimulates XerCD at dif, and similar mechanisms for chromo-
some dimer resolution are found in other bacteria (3, 4).
E. coli FtsK has an N-terminal integral membrane domain
responsible for anchoring the protein to the septum and a C-
terminal RecA-like motor domain, which are separated by a long
(≈600 amino acids) proline/glutamine-rich linker region (1, 2).
The motor domain has three subdomains, called α, β, and γ.
FtsKαβ is a RecA-like ATPase that forms a homohexamer,
which encircles DNA and couples the energy derived from ATP
hydrolysis to translocation along DNA (5, 6). FtsKγ is a winged
helix domain that binds to KOPS (FtsK Orienting Polar Sequen-
ces), an 8-mer DNA sequence that is overrepresented in the E.
coli genome and is preferentially oriented toward dif (7, 8). During
chromosome segregation, FtsK is guided toward the terminus
region by KOPS, and upon reaching dif, FtsKγ activates XerCD.
Skewed sequences similar to KOPS are present in most bacteria,
as are homologs of FtsK and XerCD (3, 4).
In vitro work has focused on the C-terminal FtsKαβγ motor
domain, referred to as FtsK50C, or similar constructs (5–16).
FtsK50Cisa DNA-dependentATPase,which canactivateXerCD-
dif recombination (6). Single-molecule studies have shown that
65 pN (8, 12, 13, 17). In addition, several studies have investigated
how FtsK is guided toward dif through interactions with KOPS
tsK is a membrane-bound DNA translocase that localizes to
the division septum in bacteria and is essential for unlinking
(7–9, 14, 18). Early reports suggested that FtsK could recognize
KOPS during translocation, enabling it to reorient while trans-
locating (7, 8, 14, 19). However, other studies have suggested that
KOPS acts only as a loading site for FtsK (9, 13).
Here we establish a DNA curtain assay that enables real-time
visualization of FtsK activities. We directly visualize FtsK as it
searches for and engages KOPS sites. Our data support a model
in which FtsK locates KOPS through a mechanism that can be
ascribed to a random 3D target search, with no evidence of long-
distance 1D diffusion within our resolution limits, with the most
efficient KOPS binding occurring in the presence of ADP. FtsK
is also efficiently targeted to KOPS when ATP hydrolysis is
prevented through the use of adenylyl imidodiphosphate (AMP-
PNP), ATPγS, ATPase-deficient mutant proteins, or omission of
MgCl2. However, ATP hydrolysis suppresses KOPS recognition,
causing FtsK to bind nonspecific sites, which suggests the exis-
tence of ATP hydrolysis-mediated allosteric communication be-
tween the FtsKαβ motor and the KOPS-binding FtsKγ domain.
We also monitor FtsK as it initiates translocation from KOPS
and demonstrate that KOPS binding dictates the initial direction
of FtsK movement. Subsequent encounters with KOPS have no
influence on FtsK. Together our data help build a more com-
plete picture of how FtsK interacts with KOPS and how these
interactions are regulated by nucleotide cofactors.
Single-Molecule DNA Curtain Assay for Real-Time Imaging of FtsK
Activity. We developed a single-molecule fluorescence assay for
visualizing the DNA-binding and translocation activities of FtsK
using nanofabricated DNA curtains (Fig. 1A and Fig. S1) (20,
21). We used a λ-phage DNA (48,571 bp) bearing two naturally
occurring single KOPS sites (1xKOPS), along with a sequence
comprising three overlapping KOPS sites (3xKOPS) inserted
within the phage genome (Fig. 1B and SI Materials and Methods).
Rather than using FtsK50C, which is prone to aggregation, we
used a linked trimer of FtsK, which lacks the N-terminal 50
amino acids that seem to be responsible for aggregation of
FtsK50C(22). Two FtsK trimers assemble to form a hexamer that
retains the in vitro and in vivo activities of FtsK50C, as previously
described (22). Unless stated otherwise we refer to the linked
FtsK trimer as FtsK, and all protein concentrations refer to the
Author contributions: J.Y.L., I.J.F., and E.C.G. designed research; J.Y.L. and I.J.F. performed
research; J.Y.L., I.J.F., E.C., and D.J.S. contributed new reagents/analytic tools; J.Y.L., I.J.F.,
D.J.S., and E.C.G. analyzed data; and J.Y.L., I.J.F., D.J.S., and E.C.G. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
1J.Y.L. and I.J.F. contributed equally to this work.
2Present address: Institut de Génétique Microbienne, Université Paris Sud 11, 91405
3To whom correspondence should be addressed. E-mail: firstname.lastname@example.org.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
| April 24, 2012
| vol. 109
| no. 17
concentration of trimer. Use of the trimer is expected to bypass
assembly steps otherwise required to make a hexameric motor
from monomeric subunits (Discussion). FtsK was expressed with
an N-terminal 14-amino-acid tag that was biotinylated in vivo
(Fig. 1C and Fig. S2). For imaging, FtsK was labeled by mixing
with a 20-fold molar excess of streptavidin-conjugated quantum
dots (QDs; SI Materials and Methods), allowing the proteins to be
visualized while bound to the DNA curtains (Fig. 1D). Bulk
assays confirmed that QD-tagged FtsK was active for ATP hy-
drolysis activity (Fig. S3).
Visualizing ATP-Dependent Translocation of FtsK. FtsK translocated
on the DNA in the presence of ATP (Fig. 2A). There was no
evidence of translocation in reactions with ATPγS (Fig. 2B),
AMP-PNP, ADP, no nucleotide, or in reactions using ATPase-
deficient mutants (see below). Trajectories of single FtsK motors
revealed abrupt changes in direction, as previously reported (7, 8,
13, 22), although these changes were not associated with KOPS
(Fig. 2A and Figs. S4 and S5; see below). The velocities were the
same in both directions (Fig. 2C), and the velocities of individual
translocases were correlated in the forward (+) and reverse (−)
directions (Fig. 2A), with a Pearson’s linear correlation of r =
0.99 (Fig. 2D). FtsK displayed a mean (±SD) velocity of 4.66 ±
1.30 kb s−1at 20 °C and 1 mM ATP (Fig. 2E), and the velocity
scaled with ATP concentration, as expected (Fig. 2F) (12, 17).
These results verified that the movement of FtsK was ATP-
FtsK Is the Fastest Characterized DNA Translocase. Previous studies
have shown that FtsK can translocate at a mean velocity of ≈5 kb
s−1at 20–25 °C (11–13, 17, 22). However, bulk ATPase assays
revealed a peak for FtsK DNA-dependent ATP hydrolysis activity
above this temperature range (Fig. S6), suggesting that FtsK might
be able to translocate even faster than has been reported. There-
fore, we tested whether ≈5 kb s−1observed at 20 °C reflected the
maximum velocity of FtsK, or whether the translocase could travel
even faster. Remarkably, the velocity of FtsK translocation in-
creased to 17.5 ± 3.5 kb s−1at 37 °C (Fig. 2G). The optimal tem-
perature for ATP hydrolysis by FtsK (43–47 °C) is outside the
accessible range of our current instrumentation, suggesting that
FtsK might translocate even faster than the observed maximum.
ADP Stimulates Loading of FtsK at KOPS. We sought to determine
whether FtsK was targeted to KOPS during initial binding to the
DNA. In the presence of ADP, FtsK (10 pM) displayed a pref-
erence for the KOPS sites present in the λ-phage substrate (Fig.
3A). Binding was detected at both 1xKOPS and at 3xKOPS, and
removing the 3xKOPS site eliminated the central peak ascribed
to 3xKOPS binding but did not affect binding to the other two
naturally occurring KOPS sites in the DNA substrate (Fig. S7).
In the absence of nucleotides the DNA-binding activity of
FtsK was reduced, with no binding detected at 10 pM FtsK.
However, when the concentration of FtsK was increased to 100
pM, our assays revealed KOPS-specific association of FtsK in the
absence of nucleotide cofactor (Fig. 3B). A preference for KOPS
DNA curtain assay, (B) λ phage DNA bearing the two native 1xKOPS sites and
the engineered 3xKOPS site, and (C) the covalently linked FtsK trimer with an
N-terminal biotin tag. (D) Images of a double-tethered YOYO1-stained DNA
curtain (green) bound by QD-tagged FtsK (magenta). The DNA, FtsK, and
corresponding overlay are shown at Top, Middle, and Bottom, respectively.
Schematic representation of (A) nanofabricated double-tethered
tagged FtsK (magenta) along a single DNA molecule (unlabeled). Transient
gaps in the magenta signal correspond to QD blinking. The location of the
KOPS sites is illustrated schematically on the left, and arrowheads indicate
the arbitrarily assigned (+) and (-) designations for translocation direction.
(B) Example of FtsK bound to DNA in the presence of ATPγS. (C) Velocity
distribution with data segregated into (+) and (−) directions. (D) Scatter plot
showing the relationship between (+) and (−) direction velocities for in-
dividual molecules of translocating FtsK. The red line illustrates a fit to the
data and yields a slope of 1. (E) Velocity distribution histogram comprising
the combined (+) and (−) velocity data sets, revealing a mean velocity of V =
4.66 ± 1.3 kb s−1. (F) Translocation velocities at varying concentrations of
ATP. (G) Velocity distributions for data collected at 37 °C and 5 mM ATP,
revealing a mean velocity of V = 17.5 ± 3.5 kb s−1.
(A) Kymogram illustrating ATP-dependent translocation of QD-
| www.pnas.org/cgi/doi/10.1073/pnas.1201613109 Lee et al.
was also observed when ADP was replaced with AMP-PNP (Fig.
3C) or ATPγS (Fig. S8). These results show that under our
in vitro conditions FtsK has an intrinsic preference for binding
KOPS and that nucleotide cofactors increase the affinity of FtsK
for DNA. These results imply that although FtsKγ is responsible
for recognizing and binding KOPS, additional binding stability is
provided by FtsKαβ when in the nucleotide-bound state.
ATP Hydrolysis Suppresses KOPS Specificity. To determine the in-
fluence of ATP hydrolysis on DNA-binding activity we asked
whether FtsK was targeted to KOPS in the presence of ATP. FtsK
actively translocates in reactions with ATP (Fig. 2), therefore
binding distributions built from data collected in the presence of
ATP reflect only the initial binding locations (Fig. S9 and SI
Materials and Methods). Surprisingly, KOPS-specific binding was
eliminated in the presence of ATP (Fig. 3D). To determine
presence of ATP or reflected an effect of actual ATP hydrolysis,
we measured the binding distribution of FtsK in reactions con-
taining ATP, where 1 mM MgCl2was replaced with 1 mM EDTA.
OmissionofMgCl2restoredKOPSbindingin reactions withATP,
indicating that the loss of KOPS specificity could be attributed to
ATP hydrolysis (Fig. 3E). We also assessed KOPS binding in a 3:1
mixture of ATP to ADP (Fig. 3F), as well as an order-of-addition
experiment in which FtsK was preincubated with ADP and then
chased with ATP before DNA binding (Fig. S8B). Neither con-
dition revealed KOPS binding, although there was very modest
and then loaded onto the DNA in a 3:1 ATP:ADP mixture. We
conclude that KOPS recognition by FtsK was suppressed even at
physiologically relevant ratios of ADP to ATP.
FtsK Walker A and Walker B Mutants Are Targeted to KOPS. We next
assayed the FtsK Walker A (K997A) and Walker B (D1121A)
mutants using trimeric FtsK wherein all of the subunits were
mutated (22). Neither mutant could translocate on DNA (Fig.
4A), although both were targeted to KOPS in the presence of
ADP (Fig. 4 B and C, Upper). In contrast to WT FtsK, KOPS
recognition by FtsK Walker mutants was not suppressed by ATP
(Fig. 4 B and C). These results suggest that KOPS loading occurs
when FtsK is unable to hydrolyze ATP. Therefore, if FtsK loads
containing 100 mM NaCl. For D and F, the reported distribution corresponds to
the first detected binding position of FtsK molecules before their translocation
away from the initial binding sites (Fig. S9 and SI Materials and Methods). The
relative locations of the KOPS sites are illustrated schematically in A.
(A–F) Binding distribution histograms for QD-tagged FtsK in the pres-
grams (Upper) for the FtsK 3xK997A Walker A mutant and the Walker B
3xD1121A mutant. (B) Binding distributions for the K997A FtsK Walker A
mutants, as indicated. (C) Binding distributions of D1121A mutants, as in-
dicated. All reactions used 10 pM mutated FtsK trimer and contained 100
mM NaCl. The relative number and location of the mutant and WT subunits
within the FtsK hexamers are schematically illustrated in B and C.
KOPS recognition by ATPase defective mutants of FtsK. (A) Kymo-
Lee et al. PNAS
| April 24, 2012
| vol. 109
| no. 17
at KOPS in vivo, then mechanisms may exist to suppress ATP
hydrolysis until the hexamer is fully assembled at KOPS.
Next we tested whether KOPS recognition was disrupted when
a subset of ATPase active sites were mutated. We took advan-
tage of the trimeric FtsK constructs to generate hexamers with
either two or four mutated subunits (22), as illustrated in Fig. 4 B
and C. These experiments revealed that the KOPS binding
specificity of the Walker mutants in the presence of ATP was
related to the number of mutant subunits within the hexamer:
FtsK bearing six mutant subunits bound specifically to KOPS,
and hexamers with four mutant subunits showed intermediate
KOPS specificity, whereas hexamers with just two mutated sub-
units displayed greatly reduced KOPS specificity, albeit not as
reduced as the WT enzyme (Fig. 4 B and C). We conclude that
the ability of FtsK to bind KOPS is enhanced when ATP hy-
drolysis is prevented and that the extent of this effect scales with
the number of mutated subunits within the hexamer.
Real-Time Visualization of FtsK Binding to KOPS. We next asked how
FtsK was targeted to specific sites by visualizing QD-tagged FtsK
in real time as it searched for and engaged KOPS. As shown in
Fig. 5A, FtsK could bind DNA at either nonspecific sites or at the
KOPS sites in the presence of ADP. When FtsK encountered
KOPS it remained bound with a lifetime of ≈1,700 s, with ≈30%
of the proteins remaining bound to KOPS beyond the duration of
the measurements (Fig. 5 A and B). When FtsK engaged non-
specific sites it exhibited a half-life of 10 ± 0.6 s before dissoci-
ating (Fig. 5 A and C). There was no evidence for 1D diffusion of
FtsK along the DNA within our current spatial (±30-nm) reso-
lution limits, although we cannot rule out the possibility that FtsK
may undergo 1D diffusion below our current resolution limits.
KOPS-Bound FtsK Translocates in a Defined Direction. To determine
whether KOPS conferred orientation specificity with respect to
prebound to KOPS in the presence of ADP. DNA translocation
then was initiated by flushing 20 mM ATP into the sample
chamber; the high concentration of ATP was used to mitigate any
dilution effects arising from flow heterogeneity within the micro-
of 40) and 87.5% of those bound to the 1xKOPS sites (n = 35 of
40) (Fig. 6A). These results confirm that the orientation of KOPS
defines the initial translocation direction of KOPS-bound FtsK.
FtsK Does Not Recognize KOPS While Translocating on DNA. To de-
termine whether KOPS influenced the behavior of FtsK during
translocation we followed the trajectories of fluorescently tagged
FtsKas they encounteredKOPS. We found no evidencethat FtsK
changed directions or paused upon encountering KOPS in either
orientation (Fig. 6 B and C). These findings show that trans-
locating FtsK does not reorient upon encountering KOPS, which
is in agreement with recent bulk biochemical data (9). We con-
clude that translocating FtsK is not functionally affected by
time at which KOPS is able to influence the behavior of FtsK is
upon initial DNA binding.
We have established a system for studying FtsK in vitro using
nanofabricated DNA curtains and QD-tagged FtsK, providing
a powerful tool for studying the properties of hexameric motor
proteins. Our assays enable us to directly visualize DNA binding
approaches.Collectively,our resultssupport a modelwhereinFtsK
locates KOPS through random collisions and preferentially binds
KOPS while in the ADP-bound state or when ATP hydrolysis is
prohibited. Once assembled at KOPS, FtsK translocates in the di-
rection dictated by KOPS but is not functionally affected by
encounters with KOPS during translocation.
showing the initial association of FtsK with the λ-phage substrate in a re-
action containing 1 mM ADP. Examples of nonspecific and KOPS-specific
binding are highlighted and shown along with the corresponding particle
tracking data. (B) Binding distribution lifetimes of FtsK bound to KOPS. (C)
Binding distribution lifetimes of FtsK bound to nonspecific sites. Specific and
nonspecific binding data were collected at 1 mM ADP, and black lines cor-
respond to single exponential fits to the data.
Real-time visualization of KOPS targeting by FtsK. (A) Kymogram
Examples of tracking data showing that FtsK leaves KOPS in the direction
dictated by the orientation of KOPS. The relative orientation of each is
depicted with arrows, and the location of each KOPS site is indicated with
a dashed orange line. Fifteen seconds of each trajectory are shown. (B and C)
Histograms showing (B) positions at which FtsK changed direction or (C)
paused. Pauses were defined as five consecutive frames with no detectable
in B and C to avoid biasing the data due to FtsK collisions with the barriers.
KOPS dictates translocation direction only upon initial loading. (A)
| www.pnas.org/cgi/doi/10.1073/pnas.1201613109Lee et al.
Model for KOPS-Specific Loading of FtsK in the Absence of ATP
Hydrolysis. Our data suggest that FtsK is preferentially loaded at
This finding contrasts with bulk biochemical assays, which have
suggested that FtsK loads at KOPS when ATP is used as the nu-
cleotide cofactor (9). One explanation for this discrepancy is that
prior bulk assays used relatively short DNA substrates with a cor-
30 bp up to 1 KOPS per 336 bp. In contrast, there is only 1 KOPS
site per ≈12 kb in theE.coli chromosome,andthis lower density is
closely reflected by the 1 KOPS per ≈10 kb used in our single-
molecule assays. It ispossible that FtsK morereadily discriminates
KOPS is present at relatively high density.
Our finding that FtsK preferentially loads at KOPS only in the
absence of ATP hydrolysis leads to the hypothesis that in vivo
FtsK may be loaded onto KOPS under conditions in which ATP
hydrolysis is prevented until the hexamer is completely assem-
bled, thus ensuring initial suppression of nonspecific DNA-
binding activity, so that FtsK can translocate in the proper di-
rection with respect to the chromosomal location of dif. Given
the abundance of ATP in vivo, and the finding that KOPS rec-
ognition is suppressed even in ADP/ATP mixtures, we speculate
that cooperative stepwise assembly of FtsK hexamers from mo-
nomeric subunits at KOPS may be used as a mechanism for
regulating the initiation of translocation, as previously suggested
(5, 6, 9, 23). Such a model would presume that complete hex-
amer assembly from monomers is inefficient on nonspecific
DNA and that ATP hydrolysis is inhibited until the hexamer is
fully assembled at KOPS. Our use of the covalently linked trimer
bypasses such an assembly driven mechanism for enhancing
KOPS specificity by allowing more efficient binding to non-
specific DNA in the presence of ATP, thus revealing the oth-
erwise unanticipated influence of ATP on KOPS recognition.
Evidence for Allosteric Communication Between FtsKγ and FtsKαβ.
The possibility of allosteric communication between FtsKγ and
FtsKαβ has been previously proposed, based on the close prox-
imity of the ATP-binding pocket to the presumed interface be-
tween the different FtsK domains (18). The ATPase active sites
are located in FtsKαβ, whereas KOPS is recognized by FtsKγ;
thus our finding that ATP hydrolysis abrogates KOPS specificity
is consistent with the hypothesis that FtsKαβ is allosterically
coupled to FtsKγ, such that it no longer recognizes KOPS when
FtsKαβ is hydrolyzing ATP. Further evidence for communication
between FtsKαβ and FtsKγ is provided by our finding that there
was a ≈10-fold increase in the affinity of FtsK for DNA, without
loss of KOPS specificity, in the presence of nucleotides relative
to reactions lacking nucleotide cofactor.
A requirement for communication between FtsKγ and FtsKαβ
may arise from the need for FtsK to translocate away from
KOPS. As previously reported, interactions between FtsKγ and
KOPS would have to be disrupted before the motor could begin
translocating, otherwise FtsKγ could act as a “brake” preventing
movement away from KOPS (18). Allosteric interactions be-
tween FtsKαβ and FtsKγ upon assembly of an ATPase active
hexamer at KOPS could be used to coordinate disruption of
FtsKγ–KOPS interactions with translocation away from KOPS.
Crystal structures reveal that three adjacent FtsKγ monomers
interact cooperatively with a single KOPS site by assembling into
a screw-like configuration that follows the pitch of the DNA (18).
Surprisingly, the helical structure formed by FtsKγ is not in
register with the FtsKαβ hexameric ring. One possibility is that
ATP hydrolysis by FtsKαβ alters the angular register of FtsKγ,
such that it is no longer in an optimal configuration for co-
operative binding to KOPS. Such a mechanism would allow
translocation away from KOPS without FtsKγ impeding forward
progression and would also provide an explanation for why FtsK
does not recognize KOPS as it translocates along DNA.
FtsKαβγ Is Targeted to KOPS Through a Mechanism Involving Random
3D Collisions. The ability of proteins to locate specific targets
among a vast excess of nonspecific DNA is a fundamental theme
in biology, yet basic principles governing these search mecha-
nisms remain poorly understood (24–26). The advent of single-
molecule imaging technologies (26), along with new NMR
methodologies (27, 28), have fueled new interest in the target
search problem. However, to date no published single-molecule
study has directly revealed the so-called “search-to-capture”
mechanism for any site-specific DNA-binding protein (29). At
a microscopic scale, QD-FtsK seems to locate KOPS through 3D
collisions and exhibits no evidence of 1D diffusion along the
DNA within our resolution limits, although we emphasize that
we cannot yet address whether FtsK slides short distances below
our current resolution limits (±30 nm, equivalent to ≈230 bp)
during the target search process. In addition, we cannot formally
rule out the possibility that FtsK might search for target sites by
intersegmental transfer because the DNA in our assays cannot
form the loops necessary for this mode of target searching.
However, because FtsK can find KOPS in our assay, we conclude
that intersegmental transfer cannot be an obligatory component
of the FtsK target search processes.
Target search processes involving site-specific DNA-binding
proteins are typically thought of within the context of proteins
that can undergo free diffusion. However, FtsK is normally an-
chored to the division septum through its N-terminal integral
membrane domain, and chromosomes undergoing segregation
are threaded through this septum as they pass into new daughter
cells. This geometry would seemingly preclude FtsK from un-
dergoing a free diffusive search throughout the bacterial ge-
nome, and therefore we envision that FtsK searches for KOPS
through a mechanism involving random sampling of the DNA as
the genome is being pulled through the division septum. The dif
activation zone of the chromosome encompasses a ≈400-kb
segment within the terminus region and contains a total of 34
KOPS sites in the correct orientation to direct FtsK toward dif
(30). This large number of sites could help ensure that FtsK
engages KOPS. Thus, the organization of the division septum,
along with the locally high protein and DNA concentrations
expected within the septum, and the high density of KOPS sites
within the dif activation zone could enhance KOPS-specific tar-
geting in vivo.
Pauses and Reversals During FtsK Translocation Are Independent of
KOPS. Our results show that FtsK initially translocates in the
direction dictated by the polarity of KOPS. The motor can pause
or change directions while translocating on DNA, but it does not
do so in response to KOPS. We do not yet understand this
phenomenon, although there are at least two reasonable expla-
nations: (i) FtsK might be bound to the DNA as double hex-
amers of opposing polarity, with one hexamer in an “on” state
capable of ATP-driven translocation and the other in an “off”
state—coordinated switching between “on” and “off” states would
then give rise to apparent changes in direction; or (ii) a single
FtsK hexamer may spontaneously dissociate from the DNA but
then quickly reassemble in the opposite orientation, which would
give rise to a change in the translocation direction. The strong
correlation found between the forward and reverse translocation
directions for individually tracked proteins (Fig. 2D) implies that
the changes in translocation direction arise from the behavior of
single FtsK hexamers, rather than from the action of two dif-
ferent FtsK hexamers. The mechanism of FtsK reversal will be
the subject of future studies; however, it is important to consider
the potential impacts of changes in FtsK translocation direction
for an in vivo scenario. For instance, it would be difficult for FtsK
Lee et al.PNAS
| April 24, 2012
| vol. 109
| no. 17
to maintain polarity in vivo if the motor domains spontaneously
changed directions in the absence of a rectifying mechanism.
Spontaneous reversals would likely be deleterious during chro-
mosome segregation, suggesting that mechanisms may be in
place to limit their occurrence once FtsK was properly loaded at
KOPS. One possibility is that the isolated motor domains do not
fully recapitulate the functional attributes of a full-length FtsK
hemaxer anchored within the confines of the division septum,
and that the organization and geometry of the septum may itself
have an important regulatory impact on the behavior of FtsK.
Alternatively, there may be unrecognized protein factors that
interact with FtsK to rectify its motion during translocation.
Future work will be essential to understand how these higher
levels of cellular organization and/or interactions with other
proteins may affect the activities of FtsK.
Materials and Methods
DNA Curtains. Double-tethered curtains were made as previously described
(20, 21). In brief, after deposition of the bilayer, the sample chamber was
flushed with buffer A [10 mM Tris·HCl (pH 7.8) and 100 mM NaCl] containing
30 μg mL−1of anti-DIG Fab (Roche catalog no. 1214667001) and then incu-
bated for 20 min to allow the Fab fragments to adsorb to the pentagons.
The chamber was then flushed with buffer B [40 mM Tris·HCl (pH 8.0), 50
mM NaCl, 0.2 mg/mL BSA, and 2 mM MgCl2) for 5 min. Streptavidin (0.02 mg
mL−1) in buffer B was then injected into the sample chamber and incubated
for 20 min. After rinsing with additional buffer B, λ-DNA (15–20 pM) labeled
at one end with biotin and at the other end with DIG (SI Materials and
Methods and Table S1) was injected, incubated for 10 min, and unbound
DNA was removed by flushing with buffer at 0.3 mL min−1. Application of
flow aligned the DNA molecules along the diffusion barriers and stretched
the molecules so the free ends could attach to the pentagons.
Single-Molecule Assays. FtsK wasexpressed and purified as previously described
(22). Biotinylated FtsK was conjugated to streptavidin-labeled QDs (Qdot 705;
Invitrogen) in reaction mixes containing 0.4 μM streptavidin-QD and 20nM FtsK
(trimer) in buffer B. Reactions were incubated on ice for 10 min. Before use, the
QD-tagged FtsK was diluted to a final concentration of 1–100pM (as indicated),
into reaction buffer [40 mM Tris·HCl (pH 8.0), 2 mM DTT, 2 mM MgCl2, 0.2 mg
ml−1BSA, 50 or 100 mM NaCl (as indicated), 0.3 nM YOYO-1, gloxy, and
1.6% (wt/vol) glucose, along with the indicated nucleotide conditions]. All
reactions also contained ≈1 mM free biotin. Nucleotide cofactors were included
Data Analysis. Position distribution measurementswere made by fittingtheQD
signals from DNA-bound proteins to 2D Gaussian functions, as previously de-
scribed (31). Error bars on the distribution measurements represent 70% con-
fidence intervals obtained through bootstrap analysis of the distribution data.
ACKNOWLEDGMENTS. This work was supported by National Institutes of
Health (NIH) Grant GM074739 (to E.C.G.), NIH Grants F32GM80864 and
1K99GM097177 (to I.J.F), a National Science Foundation Presidential Early
Career Award for Scientists and Engineers (PECASE) Award (to E.C.G.), and
Wellcome Trust Program Grant WT083469 (to D.J.S.). J.Y.L. was supported in
part by Korea Research Foundation Grant KRF-2008-357-C00048 funded
by the Korean Government. E.C. was supported in part by a European
Molecular Biology Organization (EMBO) short-term fellowship.
1. Sherratt DJ, Arciszewska LK, Crozat E, Graham JE, Grainge I (2010) The Escherichia coli
DNA translocase FtsK. Biochem Soc Trans 38:395–398.
2. Barre FX (2007) FtsK and SpoIIIE: The tale of the conserved tails. Mol Microbiol 66:
3. Carnoy C, Roten CA (2009) The dif/Xer recombination systems in proteobacteria. PLoS
4. Kono N, Arakawa K, Tomita M (2011) Comprehensive prediction of chromosome di-
mer resolution sites in bacterial genomes. BMC Genomics 12:19.
5. Massey TH, Mercogliano CP, Yates J, Sherratt DJ, Löwe J (2006) Double-stranded DNA
translocation: Structure and mechanism of hexameric FtsK. Mol Cell 23:457–469.
6. Aussel L, et al. (2002) FtsK Is a DNA motor protein that activates chromosome dimer
resolution by switching the catalytic state of the XerC and XerD recombinases. Cell
7. Bigot S, et al. (2005) KOPS: DNA motifs that control E. coli chromosome segregation
by orienting the FtsK translocase. EMBO J 24:3770–3780.
8. Levy O, et al. (2005) Identification of oligonucleotide sequences that direct the
movement of the Escherichia coli FtsK translocase. Proc Natl Acad Sci USA 102:
9. Graham JE, Sherratt DJ, Szczelkun MD (2010) Sequence-specific assembly of FtsK
hexamers establishes directional translocation on DNA. Proc Natl Acad Sci USA 107:
10. Grainge I, Lesterlin C, Sherratt DJ (2011) Activation of XerCD-dif recombination by the
FtsK DNA translocase. Nucleic Acids Res 39:5140–5148.
11. Pease PJ, et al. (2005) Sequence-directed DNA translocation by purified FtsK. Science
12. Saleh OA, Pérals C, Barre FX, Allemand JF (2004) Fast, DNA-sequence independent
translocation by FtsK in a single-molecule experiment. EMBO J 23:2430–2439.
13. Bigot S, Saleh OA, Cornet F, Allemand JF, Barre FX (2006) Oriented loading of FtsK on
KOPS. Nat Struct Mol Biol 13:1026–1028.
14. Sivanathan V, et al. (2006) The FtsK gamma domain directs oriented DNA trans-
location by interacting with KOPS. Nat Struct Mol Biol 13:965–972.
15. Marquis KA, et al. (2008) SpoIIIE strips proteins off the DNA during chromosome
translocation. Genes Dev 22:1786–1795.
16. Ptacin JL, et al. (2008) Sequence-directed DNA export guides chromosome trans-
location during sporulation in Bacillus subtilis. Nat Struct Mol Biol 15:485–493.
17. Saleh OA, Bigot S, Barre FX, Allemand JF (2005) Analysis of DNA supercoil induction
by FtsK indicates translocation without groove-tracking. Nat Struct Mol Biol 12:
18. Löwe J, et al. (2008) Molecular mechanism of sequence-directed DNA loading and
translocation by FtsK. Mol Cell 31:498–509.
19. Ptacin JL, Nöllmann M, Bustamante C, Cozzarelli NR (2006) Identification of the FtsK
sequence-recognition domain. Nat Struct Mol Biol 13:1023–1025.
20. Gorman J, Fazio T, Wang F, Wind S, Greene EC (2010) Nanofabricated racks of aligned
and anchored DNA substrates for single-molecule imaging. Langmuir 26:1372–1379.
21. Greene EC, Wind S, Fazio T, Gorman J, Visnapuu ML (2010) DNA curtains for high-
throughput single-molecule optical imaging. Methods Enzymol 472:293–315.
22. Crozat E, et al. (2010) Separating speed and ability to displace roadblocks during DNA
translocation by FtsK. EMBO J 29:1423–1433.
23. Kennedy SP, Chevalier F, Barre FX (2008) Delayed activation of Xer recombination at
dif by FtsK during septum assembly in Escherichia coli. Mol Microbiol 68:1018–1028.
24. von Hippel PH, Berg OG (1989) Facilitated target location in biological systems. J Biol
25. Halford SE, Marko JF (2004) How do site-specific DNA-binding proteins find their
targets? Nucleic Acids Res 32:3040–3052.
26. Gorman J, Greene EC (2008) Visualizing one-dimensional diffusion of proteins along
DNA. Nat Struct Mol Biol 15:768–774.
27. Doucleff M, Clore GM (2008) Global jumping and domain-specific intersegment
transfer between DNA cognate sites of the multidomain transcription factor Oct-1.
Proc Natl Acad Sci USA 105:13871–13876.
28. Iwahara J, Zweckstetter M, Clore GM (2006) NMR structural and kinetic character-
ization of a homeodomain diffusing and hopping on nonspecific DNA. Proc Natl Acad
Sci USA 103:15062–15067.
29. Halford SE (2009) An end to 40 years of mistakes in DNA-protein association kinetics?
Biochem Soc Trans 37:343–348.
30. Deghorain M, et al. (2011) A defined terminal region of the E. coli chromosome shows
late segregation and high FtsK activity. PLoS ONE 6:e22164.
31. Gorman J, Plys AJ, Visnapuu ML, Alani E, Greene EC (2010) Visualizing one-di-
mensional diffusion of eukaryotic DNA repair factors along a chromatin lattice. Nat
Struct Mol Biol 17:932–938.
| www.pnas.org/cgi/doi/10.1073/pnas.1201613109Lee et al.