Báo cáo khoa học: Focal localization of MukBEF condensin on the chromosome requires the flexible linker region of MukF docx

Focal localization of MukBEF condensin on the
chromosome requires the flexible linker region of MukF
Ho-Chul Shin
1
, Jae-Hong Lim
2
, Jae-Sung Woo
1
and Byung-Ha Oh
1
1 Center for Biomolecular Recognition and Division of Molecular and Life Science, Pohang University of Science and Technology, Korea
2 Beanline Division, Pohang Accelerator Laboratory, Korea
Introduction
Without a true mitotic spindle apparatus, two copies of
the bacterial chromosome are dynamically and effi-
ciently separated during replication and are faithfully
segregated into two daughter cells [1–3]. This essential
process requires chromosome condensation [4–7] which
is mediated by a number of factors including chromo-
some-compacting protein complexes, termed condensins
[8–10]. Bacterial condensin complexes are composed of
a homodimeric protein, belonging to the structural
maintenance of chromosomes (SMC) family, and two
non-SMC subunits [11]. SMC family proteins serve as
the common subunit in the protein complexes involved
in the chromosome maintenance processes, including
chromosome condensation, sister chromatid cohesion,
DNA double-strand break repair and gene-dosage com-
pensation [12–14]. In these SMC-based complexes, the
non-SMC proteins are heterogeneous and likely to
modulate the function of the SMC subunit.

The MukB–MukE–MukF complex, referred to as
MukBEF, is the bacterial condensin found in c-prote-
Keywords
chromosome condensation; condensing;
kleisin complex; MukBEF complex; SMC
protein
Correspondence
B H. Oh, Department of Life Sciences,
Center for Biomolecular Recognition and
Division of Molecular and Life Science,
Pohang University of Science and
Technology, Pohang, Kyungbuk 790-784,
Korea
Fax: +82 54 279 2199
Tel: +82 54 279 2289
E-mail:
(Received 13 April 2009, revised 5 July
2009, accepted 8 July 2009)
doi:10.1111/j.1742-4658.2009.07206.x
Condensin complexes are the key mediators of chromosome condensation.
The MukB–MukE–MukF complex is a bacterial condensin, in which the
MukB subunit forms a V-shaped dimeric structure with two ATPase head
domains. MukE and MukF together form a tight complex, which binds to
the MukB head via the C-terminal winged-helix domain (C-WHD) of
MukF. One of the two bound C-WHDs of MukF is forced to detach from
two ATP-bound, engaged MukB heads, and this detachment reaction
depends on the MukF flexible linker preceding the C-WHD. Whereas
MukB is known to focally localize at particular positions in cells by an
unknown mechanism, mukE-ormukF-null mutation causes MukB to
become dispersed in cells. Here, we report that mutations in MukF causing

a defect in the detachment reaction interfere with the focal localization of
MukB, and that the dispersed distribution of MukB in cells correlates
directly with defects in cell growth and division. The data strongly suggest
that the MukB–MukE–MukF condensin forms huge clusters through the
ATP-dependent detachment reaction, and this cluster formation is critical
for chromosome condensation by this machinery. We also show that the
MukF flexible linker is involved in the dimerization and ATPase activity of
the MukB head.
Structured digital abstract
l
MINT-7216106: mukBhd (uniprotkb:P22523), mukF (uniprotkb:P60293) and mukE (uni-
protkb:
P22524) physically interact (MI:0915)byblue native page (MI:0276)
Abbreviations
C-WHD, C-terminal winged-helix domain; GFP, green fluorescence protein; SMC, structural maintenance of chromosomes.
FEBS Journal 276 (2009) 5101–5110 ª 2009 The Authors Journal compilation ª 2009 FEBS 5101
obacter family members including Escherichia coli. The
SMC subunit MukB (molecular mass 170 kDa) con-
sists of globular domains at the N- and C-termini con-
nected by a long antiparallel coiled-coil ( 780
residues) folded at a hinge domain in the middle of the
molecule [15]. The bipartite N- and C-terminal
domains together form an ATP-binding cassette-like
ATPase domain (also called a head domain) at the dis-
tal end, whereas the hinge domain provides the dimer-
ization interface, resulting in the characteristic
V-shaped, two-armed dimeric structure [16,17]. The
two head domains are probably catalytically inactive,
unless they engage each other to provide a critical cat-
alytic residue opposing a ATP-binding site. ATP

hydrolysis at the composite active sites separates the
two engaged head domains [18]. This molecular archi-
tecture and the ATPase cycle appear to be shared by
all SMC proteins. MukE and MukF are two non-
SMC subunits that together form an elongated frame
structure in which the C-terminal winged-helix domain
(C-WHD) of MukF is linked by a preceding flexible
linker segment [19–21].
Null mutations of mukB, mukE or mukF cause
defects in normal cell growth, such as temperature sen-
sitivity, elongated cell formation and anucleate cell
production [22], indicating that MukE and MukF are
as important as MukB for chromosome condensation.
Although MukB normally colocalizes with the oriC
region to form distinctive foci at the 1 ⁄ 4 and 3 ⁄ 4 posi-
tions of cells occupied by nucleoids [23,24], it is dis-
persed throughout cells of the mukE-ormukF-null
strain and fails to associate stably on the chromosome
[24,25]. Likewise, the SMC subunit of Bacillus subtilis
condensin also forms discrete foci in cells, but engage-
ment-defective SMC mutants fail to do so [26]. How
the SMC proteins form discrete foci in cells remains
elusive, and it is also unclear whether their failure to
localize focally is the direct cause of defective cell
growth.
Recent structural analyses suggest that the presence
of two C-WHDs of MukF on the dimerized MukB
head is sterically unfavorable, and the MukF flexible
linker can compete with the C-WHD for binding the
MukB head at the same surface [21]. Consistently,

although two MukF C-WHDs can bind two disen-
gaged MukB heads, one is forced to detach from two
ATP-sandwiched, engaged MukB heads, and this
detachment reaction depends on the MukF flexible lin-
ker [21].
In order to investigate the functional importance
of the detachment reaction, we generated a series of
MukF variants containing mutations in its flexible
linker and ⁄ or the following a helix, and analyzed the
mutational effects using biochemical and cell-based
experiments. We show that mutations causing a
defect in the detachment reaction result in defective
localization of MukB, and that a dispersed distribu-
tion of MukB directly correlates with defects in cell
growth and division. In addition, we describe an
unexpected finding that this region of MukF is
involved in the engagement and ATPase activity of
the MukB head.
Results
New mutations on the flexible linker region of
MukF
In our previous study, six or seven glycine substitu-
tions in the flexible linker region of MukF, designated
as LM1, LM2 and LM3 (Fig. 1), were shown to cause
temperature-sensitive growth of the mukFEB-null
E. coli strain supplied with a plasmid containing the
muk operon [21]. In this study, we designed 11 new
substitutions, as shown in Fig. 1. Five of them, desig-
nated L1–L5, refer to single alanine substitutions of
five conserved hydrophobic residues on the MukF

flexible linker. One, designated HM, contains three
consecutive glycine substitutions on the first a helix of
the C-WHD following the flexible linker. These subs-
titutions are expected to relieve the putative steric
clash between the two C-WHDs on engaged MukB
heads. The other five mutations, designated LH1–LH5,
refer to alanine substitutions in the linker combined
with HM.
Fig. 1. Names of the mutations introduced into MukF. Aligned are
the phylogenetically dispersed sequences of the flexible linker and
the following a helix (a1) of MukF C-WHD: Hd, Haemophilus duc-
reyi; Va, Vibrio angustum ; Ah, Aeromonas hydrophila; Ec, Escheri-
chia coli. The secondary structural elements in this segment are
shown based on the reported structure (PDB entry: 3EUK). The
black and gray columns indicate conserved residues. The positions
of the glycine or alanine substitutions (red letters) introduced into
E. coli MukF are shown. LM1–LM3 are the same mutations
reported earlier [21].
Involvement of MukF in the localization of MukBEF H C. Shin et al.
5102 FEBS Journal 276 (2009) 5101–5110 ª 2009 The Authors Journal compilation ª 2009 FEBS
Mutations on the flexible linker region of MukF
affect the cell growth and cellular localization of
MukB–GFP
For the cell-based phenotypic assay, we employed the
KAT1 E. coli strain which expresses MukB as a fusion
protein with green fluorescence protein (GFP) at the
C-terminus [24]. From this strain, designated the wild-
type strain, we generated 14 derivative strains harbor-
ing LM1, LM2, LM3, or one of the 11 new mutations
on the chromosomal mukF gene by use of homologous

recombination (Table S1). We readily noted that the
new mutations were less stringent than LM1, LM2
and LM3. This is because mutant strains carrying any
new mutations grew at 37 °C, whereas the strain carry-
ing LM1 grew, but very slowly, and those carrying
LM2 or LM3 were not able to grow at this tempera-
ture [21]. We then compared growth of the 14 mutant
strains by measuring the sizes of single colonies on
agar plates incubated at 37 or 42 °C. In this assay,
each set of mutations decreased the growth rate to
varying degrees (Fig. 2A). At 37 °C, cells carrying L2,
L3 and L5 grew almost normally compared with wild-
type cells. However, L1, L4, HM and LH5 decreased
the growth rate by 20%, LH2, LH3 and LH4
decreased the growth rate by 40%, 60% and 80%,
respectively, and LH1 decreased the growth rate by
85%. The reduction in growth rate by LH1 was similar
to that by LM1. At 42 °C, the growth defect caused
by the mutations became more apparent. HM and
LH5 decreased the growth rate by 85%, and LH1–
LH4 prevented cells from forming a colony at this
temperature (Fig. 2A).
In parallel, we prepared cells of each strain by liquid
culture at 30 °C and observed living cells under fluo-
rescence microscopy (Fig. 2B). The wild-type strain
clearly exhibited MukB–GFP foci within the cells
(Fig. 2B), as reported previously [24]. By contrast,
LM1, LM2 and LM3 each caused dispersion of
MukB–GFP throughout the cells and a high occur-
rence of anucleate or elongated cells (Fig. 2B). The

new mutations affected the foci formation of MukB–
GFP to varying degrees. In order to assess the
foci-forming ability of each strain, we examined the
microscopic images and counted the number of cells
containing MukB–GFP foci. With L2, L3 or L5, foci
were found in  90% of the total cells counted, com-
parable with the wild-type strain. With L1, L4, LH2,
HM or LH5, 50% to  80% of the cells had foci.
With LH3,  40% of the cells had foci. LH1 and LH4
greatly hampered the formation of the foci and most
of the cells did not display any foci (Figs 2B and S1).
In addition, we observed that 6.0% of cells harboring
LH1 and 2.5% of cells harboring LH4 were anucleate.
The observed defects caused by these two mutations
were comparable with that caused by LM1, but were
less severe than those caused by LM2 and LM3.
Mutations on the flexible linker region of MukF
affect the detachment of MukF C-WHD from
engaged MukB heads
For a biochemical assay of the detachment of MukF
C-WHD from the MukB head, we employed
the MukE–MukF(M+C)–MukBhd complex. MukF
(M+C) denotes a MukF fragment (residue 292–443)
spanning the middle region and the C-WHD.
MukF(M+C) retains high binding affinity for MukE
and is monomeric because of the absence of the N-ter-
minal dimerization domain in full-length MukF.
MukBhd denotes the MukB head, which is monomeric
because of a lack of the coiled-coil and the hinge
domain in full-length MukB. Therefore, the MukE–

MukF(M+C)–MukBhd triple complex is also mono-
meric. AMPPNP-mediated engagement of the head
domains results in the dissociation of one copy of
MukE–MukF(M+C) and thus formation of the asym-
metric MukE–MukF(M+C)–(MukBhd:AMPPNP)
2
dimer, which can be observed on a native polyacryl-
amide gel [21]. We have previously shown that
the triple complexes carrying LM1, LM2 or LM3 form
the symmetric (MukE–MukF(M+C)–MukBhd:AMP-
PNP)
2
dimer as a minor product in addition to the
asymmetric dimer [21]. We generated 11 variants of
MukE–MukF(M+C)–MukBhd containing one of the
11 new sets of mutations in MukF(M+C). These com-
plexes were reacted with AMPPNP and visualized on a
native gel. With L1–L5, the symmetric dimer was
undetectable. However, with HM and LH1–LH5, for-
mation of the symmetric dimer was observed (Fig. 3A;
the band labeled as ‘sym’), apparently showing that
they affect the detachment of the MukF C-WHD from
engaged MukB heads. The band intensities of the sym-
metric dimer varied, which should reflect its stability.
According to the band intensity, LH1 is most defec-
tive, followed by LH4 and LH3, and then LH2. HM
and LH5 are the least defective of the six mutations,
because the symmetric dimer band was barely detect-
able (Fig. 3A).
Mutations in the flexible linker region of MukF

affect the ATPase activity of the MukB head
While we were probing the mutational effects on the
detachment reaction we unexpectedly noted that many
of the mutations decreased the dimerization rate of the
H C. Shin et al. Involvement of MukF in the localization of MukBEF
FEBS Journal 276 (2009) 5101–5110 ª 2009 The Authors Journal compilation ª 2009 FEBS 5103
Nomarski MukB-GFP DNA Nomarski MukB-GFP DNA
LM1
WT
LM2
LM3
L1
LH1
HM
L2 LH2
L3
LH3
L4 LH4
L5 LH5
100
20
40
60
80
Percentage of cells
Cells with no foci Cells with foci
WT L3 L2 L5 LH5 L1 HM L4 LH2 LH3 LH4 LH1
0
B
37ºC

42ºC
100
120
Colony area (%)
20
40
60
80
WT LM1 LM2 LM3
0
HM L1 L2 L3 L4 L5 LH1 LH2 LH3 LH4 LH5
A
Fig. 2. Mutational effects on cell growth
and localization of MukB–GFP. (A) Cell
growth. The wild-type KAT1 strain and its
derivative strains harboring each of the indi-
cated mutations on the chromosome were
spread on the LB agar plate and incubated
at 37 or 42 °C for 21 h. Four representative
plates incubated at 37 °C are shown at the
top. The size of 10 well-isolated single colo-
nies was measured and averaged. The aver-
aged sizes scaled to that of the wild-type
strain are shown together with the error
bars for standard deviations. (B) Localization
of MukB–GFP. The cells grown at 30 °C
were mixed with Hoechst 33342 and
observed under fluorescence microscopy.
Representative images are shown in three
different schemes for each strain. Abnor-

mally elongated cells are observable with
LM2 and LM3, whereas noticeably elon-
gated cells are observable with LM1 and
LH1. The mutations on the linker and a1
(LH1–LH5) caused greater dispersion of
MukB–GFP compared with the correspond-
ing mutation on the linker only (L1–L5). The
observed defect caused by HM is similar to
that caused by LH5. The microscopic
images of  200 cells of each strain were
inspected to count the number of cells with
and without MukB–GFP foci (see Fig. S1 for
details). The result is summarized as a bar
graph shown at the bottom.
Involvement of MukF in the localization of MukBEF H C. Shin et al.
5104 FEBS Journal 276 (2009) 5101–5110 ª 2009 The Authors Journal compilation ª 2009 FEBS
triple complex. That is, triple complexes at high con-
centrations exhibited different band intensities to the
engaged complexes, unless the dimerzation reaction
was carried out fully (Fig. 3A). Because head domain
engagement is a prerequisite for the ATPase activity of
the MukB head, we examined whether the muta-
tions also affect the catalytic activity of MukE–
MukF(M+C)–MukBhd. Indeed, the mutant triple
complexes exhibited varying degrees of ATPase activity
(Fig. 3B,C). L5 reduced the ATPase activity by 10%
from that of wild-type, whereas HM and LH5 reduced
the activity by 40%. However, it decreased > 50%
with other mutations, and > 85% with L1 and L4.
ATPase activity was generally proportional to the

extent of dimerization of the triple complex at a given
time, which was estimated by quantifying the band
intensities of the engaged complexes (bands sym and b)
on the native gel (Fig. 3C). As a control experiment,
we introduced an alanine substitution for Asp335, a
linker residue (Fig. 1) which does not interact with the
MukB head in the structure of the asymmetric dimer
[21]. This D335A mutation did not affect the ATPase
activity of the triple complex (not shown), indicating
that the reduced ATPase activities observed with the
other mutations do not arise from nonspecific muta-
tional defects.
1L LH1 L2 LH2 L3 LH3 L4 L5 LH4
LH5
WT HM
AMPPNP – + – + – + – + – + – + – + – + – + – + – + – +
A

b
sym
MukBhd
a
c
2 X
(a) (sym)
+
(b) (c)
MukF (M + C)
MukE
2 AMPPNP ( )

+
B
% Intensity of the dimer band (s)
% ATPase activity
C
120
6
7
L1
L2
L3
WT
WT
L5
40
60
80
100
2
3
4
5
L4
L5
LH1
LH2
LH3
LH4
LH5
HM

L2
LH2
wt HM L1 LH1 L2 LH2 L
3
LH
3
L4 LH4 L
5
LH
5

0
20
0 10 20 30 40
0
1
ATPase activity
Fig. 3. Mutational effects on the dimerization and the ATPase activity of MukE–MukF(M+C)–MukBhd. (A) AMPPNP-mediated dimerization.
The triple complexes (166 l
M; band a) containing the indicated mutations were reacted with 10 mM AMPPNP and visualized on a native gel.
Three different reaction products were observed: the symmetric (MukE–MukF(M+C)–MukBhd:AMPPNP)
2
dimer (band sym), the asymmetric
MukE–MukF(M+C)–(MukBhd:AMPPNP)
2
dimer (band b) and detached MukE–MukF(M+C) (band c). The symmetric dimer is observable with
HM and LH1–LH5 (highlighted by red arrows). The procedures for the identification of the protein bands have been reported [21]. The sche-
matic drawing illustrates the conversion reaction. (B) ATPase activity. Time-course ATPase assay was performed with the triple complexes
containing the indicated mutations. The activity is expressed as the number of ATP molecules hydrolyzed per one MukBhd monomer. (C)
Summary diagram. The intensities of band b (plus band sym when observed) on the native gel shown in (A) were quantified and scaled to

that of the wild-type complex. The initial rates of ATP hydrolysis were deduced from the time-course assay in (B). Average values of tripli-
cate measurements were used and scaled to that of the wild-type complex.
H C. Shin et al. Involvement of MukF in the localization of MukBEF
FEBS Journal 276 (2009) 5101–5110 ª 2009 The Authors Journal compilation ª 2009 FEBS 5105
Correlation between the detachment reaction,
localization of MukB–GFP and cell growth
We combined the observed phenotypic effects of the
new mutations into a tabular form (Table 1). This
shows that the extent of the dimerization of the
MukE–MukF(M+C)–MukBhd complexes correlates
with their ATPase activity, albeit roughly. However, it
is unrelated to how well they form the symmetric
dimer. Good correlation was found between the micro-
scopic observation of the individual cells and the cell
growth on agar plates. Strains carrying L2, L3 or L5
displayed bright MukB–GFP foci and grew as well as
the wild-type strain at 37 °C. In strains carrying L1,
L4, HM, LH2 or LH5, foci formation was compro-
mised and cell growth was retarded noticeably at
37 °C and significantly at 42 °C (Table 1). In strains
carrying the other mutations (LH1, LH3, LH4), the
foci were undetectable in the majority of cells (Fig. 2B)
and cell growth was severely retarded at 37 °C and
completely prevented at 42 °C. The band intensity of
the symmetric dimer on the native gel (Fig. 3A) corre-
lates with the defect in the foci formation of MukB–
GFP, because LH1, LH3 and LH4 affected the forma-
tion of MukB–GFP foci more severely than did LH2
(Table 1). Moreover, LM1, LM2 and LM3, for which
the symmetric dimer was clearly observed [21], did not

display any MukB–GFP foci and accompanied severe
defects in cell growth and division (Fig. 2). These
observations indicate that the detachment reaction is
required for the formation of MukB foci, and that this
focal localization of MukB on the chromosome is a
requisite for normal cell growth and division.
Because the mutations on the linker region of
MukF also affect the ATPase activity of MukE–
MukF(M+C)–MukBhd, the observed physiological
defects may simply have arisen from the decreased AT-
Pase activity of MukB in cells. Comparison of the
observed phenotypes reveals that this was not the case.
L1 and L4 decreased the ATPase activity more signifi-
cantly than LH1–LH4 did (Table 1). However, L1 and
L4 only mildly affected cell growth ⁄ division and the
foci formation of MukB–GFP, whereas LH1, LH3
and LH4 caused clearly notable defects in the cell-
based tests. Furthermore, the negative effect of HM,
LH2 and LH5 on the ATPase activity was much less
than that of L1 and L4, but their effects on cell growth
and the foci formation of MukB–GFP were similar
(Table 1). Moreover, LM1 decreased ATPase activity
by only 22% (not shown), but the other phenotypic
defects caused by this mutation were more serious than
those caused by any other of the 11 new mutations.
These observations suggest that the defective detach-
ment reaction, rather than the reduced ATPase activ-
ity, is responsible for the phenotypic defects caused by
the mutations. The observed symmetric dimer is only a
fractional portion compared with the asymmetric

dimer in the in vitro reaction (Fig. 3A), and thus the
effect of the mutations on the detachment reaction
appears insignificant. We suggest that the effect is
significant in vivo, because a mutated linker may
frequently fail to displace MukF C-WHD from MukB
Table 1. Summary of the phenotypic observations. WT, wild-type.
Mutation
In vitro assays Microscopic observations Cell growth
ATPase activity
a
Dimer formation
b
Symmetric dimer
c
MukB foci
d
Anucleate cells 37 °C42°C
WT +++++ +++ ) ++ None +++++ ++++
L1 + + ) + None ++++ +
L2 +++ ++ ) ++ None +++++ +++
L3 ++ + ) ++ None +++++ ++
L4 + + ) + None ++++ +
L5 +++++ ++++ ) ++ None +++++ ++++
HM ++++ ++++ (+) + None ++++ +
LH1 ++ ++ ++++ ) Observed + )
LH2 +++ +++ + + None +++ )
LH3 ++ ++ ++ + None ++ )
LH4 ++ ++ +++ ) Observed + )
LH5 ++++ ++++ (+) + None ++++ +
a

The activities were divided into five levels. Each ‘+’ sign corresponds to 20% of the ATPase activity of the wild-type complex.
b
The extent
of dimer formation: ‘++++’, 100–120% of the band intensity of the asymmetric dimer of the wild-type complex; ‘+++’, 80–100%; ‘++’, 60–
80%; ‘+’, 40–60%.
c
‘)’: Symmetric dimer was undetectable. The number of + signs is in accordance with the band intensity of the sym-
metric dimer on the native gel shown in Fig. 3A. Parentheses indicate that the bands were observable but only faintly.
d
Percentage of cells
exhibiting MukB–GFP foci: ‘++’, 90% of the total cells counted; ‘+’, 40–80%; ‘)’, < 40%.
Involvement of MukF in the localization of MukBEF H C. Shin et al.
5106 FEBS Journal 276 (2009) 5101–5110 ª 2009 The Authors Journal compilation ª 2009 FEBS
heads transiently dimerized by ATP [21]. This provides
an explanation as to why L1–L5 resulted in physiologi-
cal defects, even though the symmetric dimer was not
observed with these mutations in vitro.
Discussion
Because MukB is a part of the MukBEF condensin, the
distinctive foci of MukB–GFP in wild-type cells indicate
that this condensin tends to form huge clusters on chro-
mosome at the 1 ⁄ 4 and 3 ⁄ 4 positions of cells. Cluster
formation is driven by ATP binding and hydrolysis,
because a single amino acid substitution at the ATP-
binding site on the MukB head (K40I, S1366R,
E1407Q) resulted in dispersion of the GFP signal
(Fig. 4), as observed for the engagement-defective
B. subtilis SMC [26]. We showed that mutations on the
MukF flexible linker may result in the failure of MukB–
GFP to form discrete foci, demonstrating that the intact

linker is required for the focal localization of MukB.
Importantly, mutations that caused a higher dispersion
of MukB–GFP resulted in greater retardation of cell
growth accompanied by a more frequent occurrence
of abnormal cells, indicating that clustering of the
MukBEF condensin is functionally significant. Because
the phenotypic defects observed at the cellular level cor-
relate well with the defect in the detachment of MukF
C-WHD from dimerized MukB heads, formation of
MukBEF clusters is likely to require the detachment
reaction, which is coupled to ATP binding by the MukB
head. What is the nature of the huge cluster of
MukBEF? MukB and MukEF form various heteroge-
neous closed ring-like structures, and these condensin
rings may be opened and closed as a result of ATP
binding and hydrolysis by MukB heads [21]. It is tempt-
ing to speculate that the MukBEF clusters, observed as
the distinctive MukB foci, may involve concatenation of
MukBEF condensin rings. Very recent studies have
shown that ParB ⁄ SpoOJ recruits the SMC–ScpA–ScpB
condensin to the origin regions in B. subtilis [27,28].
Although a cross-linking experiment indicated a direct
interaction between SMC and ParB ⁄ SpoOJ [28], this is
unlikely to be the sole driving force, considering that
the point mutations on the ATP-binding site of MukB,
which would not affect such interaction in E. coli,
resulted in the dispersion of MukB–GFP (Fig. 4).
In the AMPPNP-mediated dimerization reaction, the
asymmetric dimer is energetically more stable than the
symmetric dimer, because the latter was unobservable

or only slightly observed. If a mutation destabilizes
the asymmetric dimer, it should increase the rate of the
reverse conversion of the asymmetric dimer into the
symmetric dimer and then into the starting monomeric
MukE–MukF(M+C)–MukBhd complex. This thermo-
dynamic consideration provides an explanation for why
the mutations on the MukF flexible linker affect the
dimerization rate of the triple complex. These mutations
also affected the ATPase activity of the triple complex.
We previously suggested that the ATPase cycle of
MukBEF involves: (a) formation of a symmetric dimer
of MukB heads to which two MukF C-WHDs are
attached, (b) the subsequent formation of an asym-
metric dimer to which the MukF flexible linker is bound
by replacing MukF C-WHD on the opposing MukB
head, and (c) ATP hydrolysis accompanying the dissoci-
ation of the two MukB heads [21]. Although the linker
is involved in formation of the asymmetric dimer in this
reaction path, it would not be able to affect symmetric
dimer formation, which is a physical association step
between two MukB heads. Therefore, the reduction in
the ATPase activity by the mutations on the linker
further supports that ATP is not hydrolyzed as soon as
the two head domains engage each other. However, the
observation seems to suggest that the ATPase activity
Nomarski MukB-GFP DNA
WT
K40I
S1366R
E1407Q

Fig. 4. Point mutations at the active site of MukB head cause dis-
persion of MukB–GFP and the elongation of cells. Mutant KAT1
cells harboring each of the indicated mutations on the chromo-
somal MukB gene were grown at 30 °C, stained with Hoechst
33342 and observed under fluorescence microscopy. The K40I,
S1366R and E1407Q mutations are expected to abolish ATP bind-
ing, head domain engagement and disengagement of dimerized
MukB heads, respectively, according to the defects caused by the
corresponding mutations in B. subtilis SMC [26]. All the three
mutant strains exhibit dispersion of MukB–GFP throughout the cells
that are abnormally elongated.
H C. Shin et al. Involvement of MukF in the localization of MukBEF
FEBS Journal 276 (2009) 5101–5110 ª 2009 The Authors Journal compilation ª 2009 FEBS 5107
of the symmetrically engaged MukB dimer with two
bound ATP molecules is catalytically less active than
the asymmetrically engaged dimer. Although further
investigation is needed to explain our observation, the
suggested lower ATPase activity of the symmetrically
engaged dimer may ensure that the MukF flexible linker
has time to competitively displace MukF C-WHD from
the opposing MukB head before the hydrolysis of
bound ATP takes place during the course of the cata-
lytic cycle of MukBEF.
In conclusion, this study shows a strong correlation
between normal cell growth and the focal localization
of MukB on the chromosome, both of which are
affected by mutations in the MukF linker region which
result in a defect in the detachment of MukF C-WHD
from engaged MukB heads. The commonly observed
dispersion of MukB in mutant cells containing detach-

ment-defective MukF or ATPase-defective MukB sug-
gests that formation of huge clusters of the MukBEF
condensin requires the ATP-dependent detachment
reaction as an important part of its functional mecha-
nism. It will be of great interest to elucidate the nature
of the driving forces for the concentration of the con-
densin complexes at the particular cell positions.
Materials and methods
Construction of mutant KAT1 strains
The E. coli KAT1 strain which produces a MukB–GFPuv4
fusion protein [24] was a kind gift of S. Adachi (University
of Nottingham, UK). All derivative strains of KAT1 were
constructed using Counter-Selection BAC Modification Kit
(Gene Bridges, Heidelberg, Germany). In brief, we first
modified the wild-type rpsL allele of KAT1 into rpsL150 to
confer resistance to streptomycin; a synthetic single strand
oligo 5¢-ACGGCATACTTTACGCAGCGCGGAGTTCG
GTTTTCTAGGAGTGGTAGTATATACACGAGTACAT
ACGCCA-3¢ was introduced into the recombination-com-
petent KAT1 cells via electroporation and the successful
transformants were obtained on streptomycin-containing
agar plates. Next, we replaced a gene fragment spanning
 400 bp around the site of mutation on bacterial chromo-
some with the rpsL-neo counter-selection ⁄ selection cassette.
This selectable marker gene fragment was designed to have
50 bp homology arms at both ends which delimiting the
region of homologous recombination, and was electroporat-
ed into streptomycin-resistant KAT1 cells. Neomycin-resis-
tant, streptomycin-sensitive transformants were selected at
20 °C. Finally, we cleared the selectable marker gene inte-

grated on chromosome by substitution with the original
DNA fragment but this time containing appropriate muta-
tion on it. Colonies that survived streptomycin selection on
agar plates were further examined and the mutations were
confirmed by DNA sequencing.
Fluorescence microscopy
Wild-type and mutant KAT1 cells were grown at 30 °Cin
LB medium with 50 lgÆmL
)1
streptomycin and 15 lgÆmL
)1
chloramphenicol. When D
600
reached between 0.7 and 1.2,
cell aliquots were supplemented with the Hoechst 33342
(Invitrogen, Carlsbad, CA, USA) dye (5 lm) and the incuba-
tion was continued for 15 min. A 150 lL aliquot of cell
suspension was dropped on a poly-(l-lysine)-coated cover
slip. After 5 min incubation, the cover slip was rinsed six
times in NaCl ⁄ P
i
, placed on a 50 lL NaCl ⁄ P
i
-spotted
microscope slide, and observed using an Axiovert 200M
fluorescence microscope (Carl Zeiss).
Size measurement of colonies on agar plate
Wild-type and mutant KAT1 cells were grown twice at
18 °C in LB media containing streptomycin and chloram-
phenicol from D

600
= 0.05 to D
600
= 0.42–0.47, first in
3 mL and then in 20 mL of culture medium. Subsequently,
these cells were mixed with glycerol (10%) and were frozen
in liquid nitrogen and stored at )80 °C. Each cell stock
was spread on a LB agar plate using glass beads and incu-
bated at 37 or 42 °C for 21 h. The images of the plates
were digitized using LAS-3000 (Fuji Film, Tokyo, Japan),
and 10 colonies on each plate were chosen and their areas
were measured using the multi gauge v3.0 program
supplied with the instrument.
Production of MukE–MukF(M+C)–MukBhd
Construction, production and purification of the complex
was as previously described [21]. In brief, the relevant DNA
fragments were amplified by PCR from the E. coli K12 gen-
ome. MukBhd was designed to contain the N- and C-termi-
nal parts (residues 1–242, and 1235–1486) connected by an
artificial SGGSGGS sequence. The three proteins were coex-
pressed from a two-promoter expression vector in the E. coli
BL21 (DE3) RIL strain (Novagen, Merck Chemical Ltd.,
Nottingham, UK) at 22 °C. The wild-type complex was
purified by metal affinity, ion exchange and size-exclusion
chromatography. Mutations on MukF(M+C) were intro-
duced using a standard PCR-based site-directed mutagenesis
method, and confirmed by DNA sequencing. The mutant
complexes were purified in the same way.
Reaction of MukE–MukF(M+C)–MukBhd with
AMPPNP

Each mutant protein complex (166 l gÆlL
)1
) was reacted
with 10 mm AMPPNP (Sigma, St Louis, MO, USA) in a
Involvement of MukF in the localization of MukBEF H C. Shin et al.
5108 FEBS Journal 276 (2009) 5101–5110 ª 2009 The Authors Journal compilation ª 2009 FEBS
buffer containing 10 m m NaCl, 1 mm MgCl
2
and 20 m m
Tris ⁄ HCl (pH 7.5) at 37 °C for 1 h in a final volume of
5 lL. The reaction mixtures were separated on a native gel
and visualized by Coomassie Brilliant Blue staining. The
intensity of protein bands was analyzed using the multi
gauge v3.0 program.
ATPase assay
Protein samples (each at 10 lm) were mixed with 500 lm
Mg
2+
–ATP (Sigma) in a buffer solution containing
42.5 mm Tris ⁄ HCl (pH 7.5), 100 mm NaCl and 1 mm
MgCl
2
at 22 °C. ATPase activity was measured by quanti-
fying released inorganic phosphate using EnzChek
Phosphate Assay Kit (Invitrogen) based on the 2-amino-
6-mercapto-7-methylpurine riboside ⁄ purine nucleoside phos-
phorylase system.
Acknowledgements
We greatly appreciate Dr Shun Adachi for providing
the KAT1 strain. We also thank Dr Joung-Hun Kim

for the use of the fluorescence microscope. This study
was supported by Creative Research Initiatives (Center
for Biomolecular Interaction) of MOST ⁄ KOSEF.
H-CS, J-HL were supported by the Brain Korea 21
Project.
References
1 Strunnikov AV (2006) SMC complexes in bacterial
chromosome condensation and segregation. Plasmid 55,
135–144.
2 Swedlow JR & Hirano T (2003) The making of the
mitotic chromosome: modern insights into classical
questions. Mol Cell 11, 557–569.
3 Webb CD, Teleman A, Gordon S, Straight A, Belmont
A, Lin DC, Grossman AD, Wright A & Losick R
(1997) Bipolar localization of the replication origin
regions of chromosomes in vegetative and sporulating
cells of B. subtilis. Cell 88, 667–674.
4 Ben-Yehuda S, Fujita M, Liu XS, Gorbatyuk B, Skoko
D, Yan J, Marko JF, Liu JS, Eichenberger P, Rudner DZ
et al. (2005) Defining a centromere-like element in Bacil-
lus subtilis by identifying the binding sites for the chro-
mosome-anchoring protein RacA. Mol Cell 17, 773–782.
5 Long SW & Faguy DM (2004) Anucleate and titan cell
phenotypes caused by insertional inactivation of the
structural maintenance of chromosomes (smc) gene in
the archaeon Methanococcus voltae. Mol Microbiol 52 ,
1567–1577.
6 Strunnikov AV (2003) Condensin and biological role
of chromosome condensation. Prog Cell Cycle Res 5,
361–367.

7 Volkov A, Mascarenhas J, Andrei-Selmer C, Ulrich HD
& Graumann PL (2003) A prokaryotic condensin ⁄ cohe-
sin-like complex can actively compact chromosomes
from a single position on the nucleoid and binds to
DNA as a ring-like structure. Mol Cell Biol 23, 5638–
5650.
8 Hirano T (2005) Condensins: organizing and segregat-
ing the genome. Curr Biol 15, R265–R275.
9 Britton RA, Lin DC & Grossman AD (1998) Charac-
terization of a prokaryotic SMC protein involved in
chromosome partitioning. Genes Dev 12, 1254–1259.
10 Strunnikov AV, Hogan E & Koshland D (1995) SMC2,
a Saccharomyces cerevisiae gene essential for chromo-
some segregation and condensation, defines a subgroup
within the SMC family. Genes Dev 9, 587–599.
11 Losada A & Hirano T (2005) Dynamic molecular link-
ers of the genome: the first decade of SMC proteins.
Genes Dev 19, 1269–1287.
12 Hirano T (1999) SMC-mediated chromosome mechan-
ics: a conserved scheme from bacteria to vertebrates?
Genes Dev 13, 11–19, doi: 10.1101/gad.13.1.11.
13 Strom L & Sjogren C (2007) Chromosome segregation
and double-strand break repair – a complex connection.
Curr Opin Cell Biol 19, 344–349.
14 Strunnikov AV & Jessberger R (1999) Structural main-
tenance of chromosomes (SMC) proteins: conserved
molecular properties for multiple biological functions.
Eur J Biochem 263 , 6–13.
15 Niki H, Jaffe A, Imamura R, Ogura T & Hiraga S
(1991) The new gene mukB codes for a 177 kd protein

with coiled-coil domains involved in chromosome parti-
tioning of E. coli. EMBO J 10, 183–193.
16 Niki H, Imamura R, Kitaoka M, Yamanaka K, Ogura
T & Hiraga S (1992) E. coli MukB protein involved in
chromosome partition forms a homodimer with a
rod-and-hinge structure having DNA binding and
ATP ⁄ GTP binding activities. EMBO J 11, 5101–5109.
17 Melby TE, Ciampaglio CN, Briscoe G & Erickson HP
(1998) The symmetrical structure of structural mainte-
nance of chromosomes (SMC) and MukB proteins:
long, antiparallel coiled coils, folded at a flexible hinge.
J Cell Biol 142 , 1595–1604.
18 Hirano T (2006) At the heart of the chromosome: SMC
proteins in action. Nat Rev Mol Cell Biol 7, 311–322.
19 Fennell-Fezzie R, Gradia SD, Akey D & Berger JM
(2005) The MukF subunit of Escherichia coli condensin:
architecture and functional relationship to kleisins.
EMBO J 24, 1921–1930.
20 Yamazoe M, Onogi T, Sunako Y, Niki H, Yamanaka
K, Ichimura T & Hiraga S (1999) Complex formation
of MukB, MukE and MukF proteins involved in chro-
mosome partitioning in Escherichia coli. EMBO
J 18, 5873–5884.
21 Woo JS, Lim JH, Shin HC, Suh MK, Ku B, Lee KH,
Joo K, Robinson H, Lee J, Park SY et al. (2009) Struc-
H C. Shin et al. Involvement of MukF in the localization of MukBEF
FEBS Journal 276 (2009) 5101–5110 ª 2009 The Authors Journal compilation ª 2009 FEBS 5109
tural studies of a bacterial condensin complex reveal
ATP-dependent disruption of intersubunit interactions.
Cell 136, 85–96.

22 Yamanaka K, Ogura T, Niki H & Hiraga S (1996)
Identification of two new genes, mukE and mukF,
involved in chromosome partitioning in Escherichia coli.
Mol Gen Genet 250, 241–251.
23 Danilova O, Reyes-Lamothe R, Pinskaya M, Sherratt
D & Possoz C (2007) MukB colocalizes with the oriC
region and is required for organization of the two
Escherichia coli chromosome arms into separate cell
halves. Mol Microbiol 65, 1485–1492.
24 Ohsumi K, Yamazoe M & Hiraga S (2001) Different
localization of SeqA-bound nascent DNA clusters and
MukF–MukE–MukB complex in Escherichia coli cells.
Mol Microbiol 40, 835–845.
25 She W, Wang Q, Mordukhova EA & Rybenkov VV
(2007) MukEF is required for stable association of
MukB with the chromosome. J Bacteriol 189, 7062–
7068.
26 Mascarenhas J, Volkov AV, Rinn C, Schiener J,
Guckenberger R & Graumann PL (2005) Dynamic
assembly, localization and proteolysis of the Bacil-
lus subtilis SMC complex. BMC Cell Biol 6, 28.
27 Sullivan NL, Marquis KA & Rudner DZ (2009)
Recruitment of SMC by ParB–parS organizes the origin
region and promotes efficient chromosome segregation.
Cell 137, 697–707.
28 Gruber S & Errington J (2009) Recruitment of condensin
to replication origin regions by ParB ⁄ SpoOJ promotes
chromosome segregation in B. subtilis. Cell 137, 685–696.
Supporting information
The following supplementary material is available:

Fig. S1. Counting the number of cells exhibiting
MukB–GFP foci.
Table S1. Strains used in this study.
This supplementary material can be found in the
online article.
Please note: As a service to our authors and readers,
this journal provides supporting information supplied
by the authors. Such materials are peer-reviewed and
may be re-organized for online delivery, but are not
copy-edited or typeset. Technical support issues arising
from supporting information (other than missing files)
should be addressed to the authors.
Involvement of MukF in the localization of MukBEF H C. Shin et al.
5110 FEBS Journal 276 (2009) 5101–5110 ª 2009 The Authors Journal compilation ª 2009 FEBS