Cell, Vol. 117, 915–925, June 25, 2004, Copyright 2004 by Cell Press
Coordination of Cell Division and
Chromosome Segregation by a Nucleoid
Occlusion Protein in Bacillus subtilis
Lutkenhaus, 2001; Shih et al., 2003). The net result of
this oscillation appears to be that the inhibitor spends
most of its time, and is therefore most active, near the
cell poles, leaving the mid-cell region available for divi-
sion. In B. subtilis and related Gram-positive bacteria, a
different topological specificity factor, DivIVA, achieves
the same result by the apparently simpler approach of
directly recruiting the inhibitor to the cell poles and re-
taining them there in a static fashion (Cha and Stewart,
1997; Edwards and Errington, 1997; Marston et al.,
1998). In both E. coli and B. subtilis, mutants that lack
MinCD function have a phenotype in which division oc-
curs either at the normal mid-cell position or at sites
small spherical anucleate “minicells.” The positioning of
division in Min?mutants raises the important question
as to whether there is an additional underlying mecha-
nism that can prevent division at sites between the pole
and mid-cell. A range of physiological and genetical
experiments suggest that this other hypothetical ef-
fector of division-site selection could be the nucleoid
(reviewed by Harry, 2001; Margolin, 2001). The nucleoid
is the bacterial equivalent of the nucleus and comprises
a highly compacted and organized chromosome but
with no surrounding nuclear membrane. During DNA
replication and segregation, the nucleoids occupy posi-
tions corresponding to the forbidden intermediate sites
between the pole and mid-cell that are predicted on the
basis of the Min mutant phenotype. Thus, combination
of nucleoid occlusion and the Min system could provide
an effective dual system for directing division to the
E. coli (Marston et al., 1998; Yu and Margolin, 1999). The
“nucleoid occlusion” effect, first postulated by Wol-
dringh and coworkers more than a decade ago (Mulder
because it could not only contribute to positioning divi-
sion at mid-cell but might also provide a checkpoint-like
In Woldringh’s view, the nucleoid occlusion effect is
mediated by molecular crowding effects arising from
the activity of the nucleoid in terms of the linked tran-
scription, translation, and translocation of secreted pro-
teins in the vicinity of the nucleoid (Woldringh, 2002).
An alternative view would be that one or more specific
proteins associated with the nucleoid acts directly to
inhibit assembly of the division machinery around the
nucleoid. Unfortunately, genes specifically affected in
tions in the mukB gene of E. coli and its equivalent, smc,
in B. subtilis, allow guillotining of the nucleoid in some
cells (e.g., Niki et al., 1991; Moriya et al. 1998), these
mutations almost certainly affect nucleoid occlusion in-
directly by their profound effects on the condensation
or organization of nucleoid structure. An alternative
model for division site selection supposes that the mid-
of the division machinery (e.g., Foley et al., 1989; Re-
Ling Juan Wu and Jeff Errington*
Sir William Dunn School of Pathology
University of Oxford
Oxford OX1 3RE
A range of genetical and physiological experiments
have established that diverse bacterial cells possess
a function called nucleoid occlusion, which acts to
prevent cell division in the vicinity of the nucleoid. We
have identified a specific effector of nucleoid occlu-
sion in Bacillus subtilis, Noc (YyaA), as an inhibitor of
division thatis also a nonspecificDNA binding protein.
Under various conditions in which the cell cycle is
perturbed, Noc prevents the division machinery from
assembling in the vicinity of the nucleoid. Unexpect-
edly,cells lackingboth Nocand theMin system(which
prevents division close to the cell poles) are blocked
for division, apparently because they establish multi-
ple nonproductive accumulations of division proteins.
The results help to explain how B. subtilis specifies
the division site under a range of conditions and how
it avoids catastrophic breakage of the chromosome
by division through the nucleoid.
Proper timing and positioning of the division plane is a
crucial aspect of the cell cycle throughout biology. In
higher organisms, it is generally assumed that the posi-
tion of the interphase nucleus is used as a topological
signal that determines the site and orientation of the
division plane. In most cases, however, little is known
about the molecular basis for division site specification.
In fission yeast, several genes required for division site
positioning have been identified, including mid1, plo1,
and pom1, but details of how their gene products work
in this context remain hazy (Rajagopalan et al., 2003).
In the rod-shaped bacteria E. coli and Bacillus subtilis,
division occurs with exquisite precision at mid-cell, and
the timing of division is regulated so that it follows the
replication and segregation of sister chromosomes.
However, despite decades of study, the mechanisms
underlying division timing and positioning in these or-
ized components of this process are the widely con-
served MinC and MinD proteins, which form a complex
that acts as an inhibitor of division (de Boer et al., 1990,
1992; Hu et al., 1999). The activity of MinCD is restricted
factor (de Boer et al., 1989). In many bacteria, this factor
is a protein called MinE, which works by driving a re-
markable pole-to-pole oscillation of the MinCD inhibitor
(Fu et al., 2001; Hale et al., 2001; Hu et al., 2002; Hu and
gamey et al., 2000). Again, the molecular mechanisms
that might underlie this kind of model remain ill defined.
About eight widely conserved proteins are known to
be required for cell division in bacteria (reviewed re-
cently: Errington et al., 2003). The key player in the ma-
chinery is a tubulin homolog, FtsZ, that polymerizes to
form a ring-like structure, the Z ring, at the site of cell
division. The other division proteins, most of which are
integral membrane proteins, are recruited to the Z ring
in a hierarchy that is quite well characterized in both
by a number of proteins. The MinCD protein mentioned
above is an inhibitor of Z ring formation (Hu et al., 1999).
In E. coli, FtsA and ZipA have partially redundant func-
tions, at least one of which needs to be present to allow
formation of the Z ring (Pichoff and Lutkenhaus, 2002).
Filho and Losick, 2002) have been identified as likely
negative and positive regulators of FtsZ polymerization
or Z ring formation. However, how the functions of all
of these proteins are integrated to control the timing
and spatial localization of Z ring formation is not yet
clear. In particular, none of these proteins appears to
be capable of coupling Z ring formation to the position
of the nucleoid.
We have now identified a B. subtilis protein, YyaA,
that associates nonspecifically with the chromosome
and is required to block cell division in the vicinity of
the DNA, finally providing a specific protein with a role
in nucleoid occlusion. Under various conditions, YyaA
contributes to correct placement of the division site,
reducing the risk of chromosome damage during di-
(dependent on an inducer, IPTG) was introduced into a
yyaA null mutant background. This strain grew normally
in the presence of inducer (data not shown). Removal
of the inducer to deplete MinD in the presence of YyaA
resulted, as expected, in the production of minicells
(Figure 1C). However, when MinD was depleted in a
yyaA mutant, division was arrested and the cells grew
into long filaments (Figure 1D). Similar results were ob-
tained when YyaA was depleted in a minD background
(data not shown).These results showed that absence of
both the Min system and the YyaA protein results in a
severe defect in cell division.
yyaA minD Double Mutants Fail to Concentrate
FtsZ at Potential Division Sites
To test at what level the block in cell division was oc-
curring, we examined the effect of loss of YyaA and
MinD on localization of cell division proteins. FtsZ is a
tubulin homolog that lies at the top of the hierarchy of
cell length was normal and FtsZ-GFP was targeted cor-
rectly to bands located at nascent and future division
sites, as in wild-type cells (not shown). In yyaA?cells
depleted for MinD, again discrete bright fluorescent
bands were detected, some of which were adjacent to
each other due to minicell formation but all were cor-
rectly interpolated between segregating nucleoids (Fig-
ure 2A). However, when MinD was depleted for 3–4 gen-
erations at 30?C in the ?yyaA background, bright
transverse bands of FtsZ were formed only rarely in the
elongating filaments; instead, the protein was distrib-
uted much more widely in multiple weak bands (Figure
2B). Similar results were obtained when a nonessential
but early assembling (ftsZ-dependent) component of
the division machinery, EzrA (Levin et al., 1999), was
examined. Thus, single bright transverse bands of EzrA-
of YyaA (Figure 2C), but these were replaced by scat-
mutant (Figure 2D).
Cell Division Is Inhibited in yyaA minD
In genetic crosses designed to construct strains with
systems in chromosome segregation in B. subtilis (Wu
certain combinations of mutations seriously compro-
mised cell viability. A series of genetic crosses estab-
lished that the effect was actually due to simultaneous
loss of the Min system (divIVA, minC, minD, or ezrA)
and the product of a poorly characterized gene called
yyaA (see Supplemental Table S1 at http://www.cell.
com/cgi/content/full/117/7/915/DC1). yyaA lies just up-
stream of the soj-spo0J locus and encodes a product
that is about 40% identical to Spo0J over much of its
length. We previously reported that yyaA mutants do
not have an obvious growth, segregation, or division
phenotype, though overexpression impairs sporulation
(Sievers et al., 2002).
Prolonged low temperature incubation (30?C or less)
of transformation plates bearing potential double mu-
tants of the poorly viable class, e.g., minD yyaA, eventu-
ally resulted in colonies that contained both incoming
and resident markers (Figure 1A; compare with the
abundant uniform colonies in the control transformation
in Figure 1B). To examine the mutant phenotype under
more controlled conditions, a repressible allele of minD
YyaA Colocalizes with the Nucleoid and Inhibits
Division when Overproduced
YyaA was shown previously to be a relatively non-
orescence microscopy suggested that it colocalized
with the nucleoid (Sievers et al., 2002). Unfortunately,
the preparation of specimens for immunofluorescence
microscopy does not preserve the normal morphology
of the nucleoid well. To assess the localization of YyaA
in living cells, a GFP fusion was constructed. The gfp-
yyaA fusion construct was functional because, when
present as the only expressed copy of yyaA in a minD
mutant background, the cell length was typical of that
of a min mutant, and the colonial growth rate was nor-
mal. As shown in Figures 3A–3C, the GFP-YyaA fusion
colocalized withthe nucleoid, in confirmationof the pre-
vious immunofluorescence data (Sievers et al., 2002).
Interestingly, the GFP-YyaA signal did not always cover
the whole nucleoid and a gap was often evident near
the middle of longer nucleoids (arrows in Figure 3C).
In the light of the effects of yyaA mutation on division
(in combination with min), it was important to look
Nucleoid Occlusion Protein in B. subtilis
Figure 1. Severe Cell Division Defect of yyaA minD Double Mutants and Rescue by Overproduction of FtsZ
(A and B) Small colony phenotype and poor recovery of yyaA minD double mutants in a transformation cross. A minD mutant (strain 1901)
was transformed with equal concentrations of DNA from strains carrying a tetracycline resistance cassette inserted into the yyaA gene (A) or
into an intergenic site (yxeDC; B). The treated recipient cells were plated in the presence of tetracycline and incubated for 1 day at 37?C then
1 day at 30?C. The control (B) showed large numbers of uniformly sized colonies. The yyaA-transformed cells gave only a few large colonies
(arrows) that usually had lost the resident minD mutation. Small colonies (arrowheads) retained both resident (minD) and incoming (yyaA)
mutations but appeared only after incubation at 30?C and were extremely underrepresented.
(C and D) Filamentous phenotype arising from depletion of minD in a ?yyaA background (D; strain 1284) but not in a yyaA?background (C;
strain 1998). Strains were grown at 30?C and 2 hr after removal of the inducer (IPTG), cells stained with the membrane dye FM5-95 were
examined directly by fluorescence microscopy. Arrows in (C) point to examples of minicells.
(E–J) Partial rescue of a yyaA minD double mutant by overproduction of FtsZ.
FtsZ was overproduced by adding both IPTG (1 mM) and xylose (0.5%) to cultures containing inducible copies of wild-type and gfp fusion
derivatives of ftsZ. The strains used carried mutations in minD (strain 1296; E–G), or minD and yyaA (strain 1297; H–J). The panels show typical
fields of cells stained to visualize the cell membrane (FM5-95; E and H), the DNA (DAPI; F and I), and a false color merge with membrane in
red and DNA in green (G and J). Arrows point to septa that appear to have bisected a nucleoid, and arrowheads to septa that formed anucleate
cells (mainly minicells in the minD single mutant and a range of cell sizes in the double mutant).
closely at the phenotype of yyaA single mutants. In our
previous characterization of yyaA, we reported that
overexpression of yyaA inhibited sporulation (Sievers et
increase in cell length. The previous experiments were
done with multicopy plasmids, which provide a variable
gene dosage, and for which controls are difficult be-
cause the vector plasmids with no inserts tend to have
effects on the growth and cell cycle progression of the
host cells. To overexpress yyaA under more controlled
Figure 2. Double Defect in yyaA and minD
Localization of GFP-FtsZ (A and B) or EzrA-
GFP (C and D) in MinD-depleted cells with
(A, strain 1291; C, strain 1286) or without (B,
strain 1292; D, strain 1287) a functional yyaA
gene. The panels show GFP fluorescence
(left), DAPI (DNA; middle), and a false color
merge of the two channels (right) with GFP
in greenand DAPI inred. Arrows pointto GFP
bands that overlap the nucleoids. Scale bar,
conditions, we introduced into otherwise wild-type cells
a second copy of yyaA under the control of the strong
IPTG-inducible Pspac(hy) promoter (Quisel et al., 2001),
which gave about an 8-fold increase in transcription, as
judged from microarray data (data not show). Figures
strain with no overexpressed yyaA, but that in the over-
(Figures 3F and 3G). Thus, overexpression of YyaA re-
sulted in a partial inhibition or delay in cell division.
Western blot analysis showed that FtsZ levels were un-
affected by the overexpression (not shown). However,
immunofluorescence microscopy revealed a reduction
in FtsZ bands in the elongated cells (1 band per 13.2
?m) compared with the control cells (1 band per 6.3
?m), showing that the inhibition of division occurred at
the level of Z ring assembly.
yyaA Disruption Allows Division
through the Nucleoid
The above and previous experiments established that
YyaA is a relatively nonspecific DNA binding protein and
Nucleoid Occlusion Protein in B. subtilis
Figure 3. Localization of GFP-YyaA over the
Nucleoid and Weak Filamentous Phenotype
upon YyaA Overproduction
(A–C) Strain 1288 with a gfp-yyaA fusion
growing exponentially in CH medium con-
taining 0.3% xylose was examined by fluo-
resence microscopy. Panels show the GFP
signal (A), DAPI (B), and a merge of the two
images (C) with GFP false colored green and
DAPI in red. The arrows in (C) point to central
regions of the longer nucleoids where the
GFP signal appears to be reduced relative to
that of the DNA.
(D–G) Isogenic strains with the strong Pspac(hy)
promoter inserted into the amyE locus with
no additional gene (D and E; strain 1289) or
upstream of a copy of the yyaA gene (F and
trast (D and F) and membrane fluorescence
(FM5-95; E and G). Scale bar, 2 ?m.
properties were suggestive of a role in nucleoid occlu-
sion.Insupport ofthisidea,wenoticedthat intheearlier
experiments with yyaA minD double mutants, the aber-
rant accumulations of FtsZ or EzrA protein frequently
overlapped the nucleoid (arrows in Figures 2B and 2D),
in contrast to the discrete bands located precisely be-
tween nucleoids in yyaA?cells (Figures 2A and 2C). To
test directly for a role in nucleoid occlusion, we used a
well-characterized systembased on themanipulation of
germinating spores. Spores contain a single replicated
chromosome, and the first cell cycle following germina-
tion and outgrowth of the spore can be followed under
well-characterized and relatively synchronous condi-
tions. The timing of various events is well known and
the effects of addition of various inhibitors of DNA repli-
cation have been extensively documented. In particular,
Wake and coworkers have shown that if the round of
DNA replication is blocked at the initiation stage, e.g.,
by early addition of the replication elongation inhibitor
HPUra, the cells elongate at the normal rate but division
is delayed. Importantly, when division does eventually
occur, it is asymmetric and positioned to one side of
the centrally located nucleoid (e.g., Harry et al., 1999;
Rowland etal., 1997), presumablyas a resultof nucleoid
occlusion. We therefore tested whether yyaA mutants
are affected in division septum positioning under these
conditions. Wild-type and ?yyaA spores were germi-
nated and cells were allowed to outgrow. After a time
interval sufficient to complete one or two rounds of cell
division, samples of the cultures were examined micro-
scopically to determine the numbers of septa, and their
position relative to the nucleoid andthe cell poles. In the
absence of HPUra, there was no detectable difference
between the wild-type and the ?yyaA cells; all of the
outgrowing cells divided at the expected size and at
added to preventDNA replication at a veryearly stage in
outgrowth, wild-type outgrowing spores showed mainly
acentral division septa (Figures 4E–4H). Thus, only 19%
of the cell (Figure 4H). In contrast, most of the yyaA
septa (85%) were close to mid-cell (Figures 4I–4L). The
of cells) than in wild-type cells (32%). Importantly, the
septa of the yyaA mutant cells frequently had DNA on
both sides of them (28% of all cells), indicating that the
nucleoid had been bisected, whereas in the wild-type
only 5.4% of cells had DNA visible on both sides of a
septum. In the yyaA mutant, cells with central septa
Figure 4. Impaired Nucleoid Occlusion in the Absence of YyaA
(A–L) Spores of wild-type B. subtilis (A, B, and E–H; strain 168) or an isogenic strain with a disruption of yyaA (C, D, and I–L; strain 1282) were
germinated and allowed to outgrow with (E–L; for 342 min) or without (A–D; for 210 min) HPUra inhibition of DNA replication. (A), (C), (E), and
(I) show FM5-95 (membrane) stain; (B), (D), (F), and (J) show DAPI (DNA); (G) and (K) are false color merges of the membrane (green) and DNA
(red) images to their left. Arrows in (E) and (F) point to asymmetric septa formed outside the nucleoids, and in (I) and (J), arrows point to mid-
cell septa that have bisected a nucleoid. Arrowheads in (K) point to highly asymmetrical nucleoids that lie to one side of a mid-cell septum.
The positions of the septum in about 150 dividing cells of the wild-type (H) and mutant (L) spores were measured relative to the nearest cell
pole and scored as a percentage of total cell length. Note that the bright green ovoid objects are spore coats or nongerminated spores.
(M–P) Isogenic strains containing an IPTG-inducible dnaA gene and either wild-type yyaA (M and N; strain PL10) or a deletion of yyaA (O and
P; strain 1295) were grown in the presence of IPTG, then the inducer was withdrawn to allow ongoing rounds of DNA replication to be
completed but no new rounds to be started. Cells were imaged after a time sufficient to allow approximately 2 doublings in cell length. Shown
are phase contrast (M and O) and combined phase contrast and DAPI images (N and P). Nucleoids are visible as white areas against the dark
cell background. Arrows point to asymmetrical septa that lie distant to the nucleoid and arrowheads to medial septa that have bisected a
nucleoid. Scale bar, 2 ?m.
rically, lying mostly or completely on one side of the
septum (arrowheads in Figure 4K). We assume that this
tion of the nucleoid after septation (Sharpe and Er-
rington, 1995). This experiment showed that yyaA plays
an important role in preventing division at mid-cell when
occlusion, we examined the effects of depleting cells
for the key initiator protein for DNA replication, DnaA,
Nucleoid Occlusion Protein in B. subtilis
in exponentially growing cells. The cultures induced to
express dnaA had a wild-type phenotype, and yyaA?
and ?yyaA strains were indistinguishablein terms of cell
lengthand cellcycleprogression(data notshown).After
about two generations of growth in the absence of in-
ducer, DnaA-depleted cells of the yyaA?strain were
elongated and tended to contain a single nucleoid, cor-
responding to a single chromosome (Figures 4M and
4N). Wheredivision septawere visible,they werealways
located asymmetrically, away from the nucleoid (arrows).
In contrast, the DnaA-depleted cells of the yyaA mutant
were shorter (Figures 4O and 4P), as if division was not
inhibited, and division frequently occurred right through
the nucleoidat mid-cell(arrowheads inFigure 4P).Thus,
under two quite different experimental conditions, yyaA
mutant cells apparently lacked the nucleoid occlusion
function that normally prevents cell division from oc-
curring through the nucleoid.
shown in Figure 5I, the yyaA?cells tended to have a
single bright Z band at mid-cell, and in longer cells,
further weak Z bands were evident at approximately 1/4
and 3/4 positions, as expected. In contrast, the yyaA
mutant cells tended to have helical or multibanded sys-
tems occupying the medial portion of the cell but not
in the polar regions (compare with the poles of cells of
a min mutant in Figure 2B). These results were consis-
tent with the fact that the Min system is capable of
protecting an extended polar zone of the cell from FtsZ
polymerization but YyaA becomes increasingly impor-
tant as the cell elongates.
Division through the Nucleoid in yyaA minD
Double Mutants Overproducing FtsZ
On the basis of the above results, it seemed likely that
the filamentous phenotype of yyaA minD double mu-
tants was due to delocalized or uncoordinated polymer-
ization of FtsZ. If so, it might be possible to partially
rescue division by overproduction of FtsZ. We therefore
built a strain with two inducible copies of ftsZ (one also
fused to gfp) in which this could be tested. The experi-
duction of FtsZ did indeed lead to a partial restoration
of division in cells impaired in yyaA and minD function.
Figures 1E to 1G show a control experiment with a minD
single mutant overproducing FtsZ. The overproduction
had little effect on the cells, which were of relatively
normal length and produced anucleate minicells at a
substantial frequency (arrowheads), as expected. When
the overproduction was done in a yyaA minD back-
mally seen in the double mutant (e.g., Figure 1D) were
resolved into much shorter units (Figure 1H), due to a
substantially increased rate of division. However, unlike
the equivalent minD single mutant (Figures 1E–1G), in
which division occurred neatly in the nucleoid free
spaces, many cell divisions were clearly asymmetrical
and occurred over or through nucleoids (arrows). Many
of the divisions again produced anucleate “minicells”
(arrowheads), but these showed an unusually variable
length compared with the tiny spherical minicells pro-
duced by the minD single mutant.
Although these experiments again highlighted the
need for YyaA protein to prevent division through the
nucleoid, it was notable that a substantial proportion of
the cell divisions that occurred under these conditions
were located in internucleoid positions, and this also
applied to the positioning of Z rings as visualized by
FtsZ-GFP (not shown). This raises the possibility that
under these conditions, a yyaA-indepependent system
can bias the division machinery away from the bulk of
YyaA-Dependent Positioning of Z Rings
between Nucleoids in Elongated Cells
Several authors have reported that cells blocked late
in division form Z rings at regular intervals, “potential
division sites,” which generally correspond to inter-
nucleoid spaces. In E. coli, the oscillating Min system
filaments (Raskin and de Boer, 1997), but the fixed polar
Min system of B. subtilis should not have this capability
(Harry, 2001; Margolin, 2001). It seemed possible that
YyaA might be involved in directing Z ring assembly to
internucleoid spaces in abnormally long cells. To test
this, we used a strain in which expression of two “late”
cell division genes, ftsL and pbpB, could be simultane-
ously repressed (Daniel et al., 1996), leading to elon-
gated cells in which Z rings would be expected to form
but no division septa. In the initial experiments, EzrA-
tion. As shown in Figures 5A–5D, when the yyaA?cells
were depleted, the cell filaments contained EzrA-GFP
bands positioned precisely in internucleoid spaces.
Some of the bands appeared to correspond to double
ring structures or short helices, but these structures
were always localized in the internucleoid spaces. In
sharp contrast, in the equivalent filaments of the ?yyaA
derivative (Figures 5E–5H), clusters of EzrA-GFP bands,
all helical configuration, were located in broad regions,
which substantially overlapped the nucleoids. The
nucleoids had their normal regular spacing in the cell
but no longer apparently excluded FtsZ polymerization.
Experiments looking directly at FtsZ localization by im-
munofluorescence microscopy in fixed cells gave re-
These results provided further evidence that yyaA is
required for specification of Z ring position indepen-
dently of the Min system.
Interestingly, in the above experiments, it was evident
to the poles of the filaments, consistent with these re-
gions being protected by MinCD. In a further repetition
of the above experiment, we examined cells containing
the cells were just beginning to become elongated. As
A Gene for Nucleoid Occlusion
The nucleoid occlusion model was first promulgated by
Woldringh and coworkers (Mulder and Woldringh, 1989;
Woldringh et al., 1990) to explain their observation that
in E. coli cell division did not occur in the vicinity of
the nucleoid under various conditions in which DNA
Figure 5. Decreased Ability of yyaA Mutants to Direct the Division Machinery to Internucleoid Spaces in Filamentous Cells
Cells bearing an IPTG-inducible promoter driving expression of the essential division genes ftsL and pbpB were grown in the presence of
IPTG, then the inducer was removed to block division. After sufficient time to allow about 2 doublings in cell length, cell filaments were
imaged. In (A)–(H), the strains also contained an ezrA-gfp fusion to monitor FtsZ assembly and either wild-type yyaA (A–D; strain 1293) or a
deletion mutation (E–H; strain 1294). The panels show GFP fluorescence (A and E), DNA (B and F), a false color merge with GFP in green,
DNA in red (C and G), and membrane (Nile Red) (D and H). (I) and (J) show the localization of GFP-FtsZ in yyaA?(I; strain 1298) or ?yyaA
mutant background (J; strain 1299) when FtsL and PbpB were depleted, as for panels (A)–(H). Scale bar, 2 ?m.
ing regions of the filamentous cells, away from the
nucleoid. Similar conclusions had been drawn in earlier
extensive work on B. subtilis, by Wake and coworkers
(see Regamey et al.  and references therein). Out-
growing spores initiate a relatively synchronous round
of DNA replication that can be readily manipulated. In
this system, it was shown that an early block in chromo-
although this inhibitory effect begins to disappear when
the round of DNA replication has progressed beyond
about 60% toward completion. In terms of nucleoid oc-
clusion, the latter observation could be explained by
the switch in configuration of the nucleoid to a bilobed
shape, which occurs at about that time (Sharpe et al.,
1998). The bilobed nucleoid could allow Z ring formation
in the space between the two major lobes of DNA ahead
of the completion of DNA replication. In support of this
notion, in various other conditions in which division has
been observed to occur through the nucleoid, the effect
seems to be associated with a localized reduction of
DNA concentration (see Margolin ). Recent work
on both E. coli and B. subtils has confirmed and ex-
tended the work on division by demonstrating that
nucleoidocclusion ismediatedat somestepin FtsZring
assembly (reviewed by Harry ; Margolin ).
Despite these numerous observations, the molecular
basis for nucleoid occlusion has remained unclear,
mainly because no genes or proteins directly responsi-
ble for this function had been identified (see Introduc-
tion). We have now identified the YyaA protein as an
important effector of nucleoid occlusion in B. subtilis.
We propose to rename this protein Noc, to reflect its
role in nucleoid occlusion.
Several lines of evidence demonstrated impaired
nucleoid occlusion in the noc mutant. When initiation of
of outgrowing spores or by depleting DnaA in batch-
grown cells, division occurred much more frequently in
the noc mutant than in the equivalent wild-type strain.
Crucially, many divisions occurred through the nucleoid
in the noc mutant, whereas the nucleoid is usually
missed in the wild-type (Figure 4). When noc and min
mutations were combined, the filamentous cells fre-
quently showed concentrations of FtsZ (or EzrA) as
spots or bands over the nucleoid, whereas in isogenic
noc?cells, the FtsZ assemblies were almost always in
the internucleoid spaces (Figure 2). The clearest indica-
tion of the function of Noc protein in directing FtsZ
Nucleoid Occlusion Protein in B. subtilis
ments illustratedin Figure 5. Inelongated filaments gen-
erated by a late block in cell division, the Z rings that
normally form accurately in internucleoid spaces no
longer did so in the absence of Noc. Instead, elongated
structures, probably helical in gross architecture, were
Noc is therefore required to protect the nucleoid from
division, and it does so apparently by inhibiting FtsZ
accumulation or polymerization in the vicinity of the
nucleoid;it functions,directlyorindirectly, asanucleoid
that Noc has a critical role as the first line of defense in
protecting the nucleoid from damage by the division
septum. It might also be thought of as a checkpoint
protein that helps to prevent division from occurring
until the chromosomal DNA has properly cleared from
the mid-cell site for division.
Noc may act directly because it has the two key prop-
erties potentially needed for a nucleoid occlusion pro-
tein. First, it is a relatively nonspecific DNA binding pro-
tein that localizes almost uniformly over the nucleoid
(Sievers et al., 2002 and Figure 2A). This contrasts with
which colocalizes with oriC regions (Glaser et al., 1997;
Lin et al., 1997). Second, when overproduced, it partially
ably due to Noc localization still being restricted to the
vicinity of the nucleoid. Overproduction may simply in-
crease the concentration of Noc on the nucleoid and
delay the time at which division at mid-cell, between
segregating nucleoids, occurs. It will be interesting to
carry out a finer detail analysis of the distribution of the
protein on the nucleoid (e.g., by chromatin immunopre-
cipitation). Inspection of the images shown in Figures
3A–3C suggests that the protein may be less abundant
to the unreplicated terminus region. This might allow
of replication, as observed in the experiments by the
Wake lab described above.
Several lines of argument support the notion that Noc
acts directly, rather than at the level of transcriptional
regulation of a division inhibitor. First, although Noc
appears to localize all over the nucleoid, preliminary
change in expression in the absence of Noc. In particu-
lar, we found no significant change in the expression of
any known gene involved in cell division or its regulation
(L.J.W., unpublished data). This includes the recently
described yneA gene, which seems to be the B. subtilis
functional equivalent of the SOS-responsive division in-
hibitor, sulA of E. coli (Kawai et al., 2003). Second, the
protein is apparently expressed constitutively and is
present on the nucleoid independent of cell cycle pro-
gression or DNA damage (Sievers et al., 2002; Figure
2A; L.J.W., unpublished data), providing no hint for a
role in a putative signal transduction pathway. Third, the
results shown in Figures 2 and 5 strongly suggest that
the loss of Noc function affects the spatial organization
of Z rings rather than conferring a general inhibition of
polymerization, and the location of the inhibitory effect
in wild-type cells (i.e., over the nucleoid) is where the
Noc protein localizes. Finally, when overproduced, Noc
caused a reduction in cell division (Figure 3) that oc-
curred at the level of FtsZ assembly.
Implications for the Mechanism of Z Ring
Assembly and Positioning
This investigation of noc function was prompted by the
near lethal filamentous phenotype of the noc minD mu-
tant (Figure 1). This phenotype was initially confusing
because we had anticipated that combination of a
nucleoid occlusion mutation with a min mutation would
give rise to uncontrolled division by elimination of two
negative regulators. Why should the loss of two division
inhibition systems result in a block in division? The im-
ages shown in Figure 2 suggest that the failure to divide
occurs at multiple sites, most of which do not lead to
the formation of a productive Z ring structure. We sug-
gest that with multiple minor polymerization centers all
competing for protein, none of these can recruit enough
FtsZ protein to form a productive ring. In support of this
interpretation, we showed that the division frequency
could be substantially recovered by overproduction of
FtsZ (Figures 1E–1J). This provides strong support for
the notion that topological restriction of Z localization
tion of a complete, functional Z ring. Similar conclusions
have been drawn for E. coli based on analysis of the
effects of combining deletion of the min system with an
ftsZ mutation that reduces FtsZ polymerization (Yu and
suggest that condensation of FtsZ into a defined region
could be sufficient to drive the formation of an active
division machine, eliminating the need for a predeter-
mined site for Z ring assembly, as has been postulated
previously (e.g., Cook and Rothfield, 1999; Regamey et
If nucleoid occlusion was important in selecting the
mid-cell site for division, noc mutants should exhibit
some loss of specificity of division site placement, but
this was not apparent in unperturbed cells. We think
that there are at least two explanations for this. First,
there is some evidence for a weak noc-independent
system that biases division to internucleoid spaces, as
evident in the successful divisions that occur in Figures
1D–1F. Second, the Min system may be capable of pre-
venting division at a significant distance from the cell
poles in B. subtilis, rather than acting only to prevent
minicell divisions close to existing cell poles. This was
particularly evident in the experiment shown in Figures
5I and 5J. We previously noted that MinD localization,
though concentrated at the cell poles, appears to ema-
nate out from the poles toward mid-cell (Marston et al.,
1998). In retrospect, it appears to make sense for the
Min system to prevent division from occurring within
about one nucleoid-length equivalent from each cell
pole (Figures 6A and 6B). Under normal conditions, this
could be sufficient to allow division only at mid-cell and
in cells that had reached the appropriate length. In pre-
duction of MinCD increases the cell length distribution
(L.J.W., unpublished results), consistent with an in-
crease in the length of the pole-proximal part of the cell
to grow for 95 min. Cells were then diluted once more into PAB
containing 1 mM IPTG to an OD600nmof 0.05 and grown for a further
65 min before examination by fluorescence microscopy.
Spore Germination, Outgrowth, and Inhibition of DNA
Replication using HPUra
Spore (2 ? 108) germination and outgrowth were performed at 30?C
in S medium (Sharpe et al., 1998) supplemented with 0.2 mg/ml of
alanine and 0.002% Tween as described by Hamoen and Errington
(2003). After 65 min, the DNA polymerase III inhibitor 6-(p-hydroxy-
tration of 200 ?M to block DNA replication.
Overexpression of ftsZ/ftsZ-gfp
Strains 1296 and 1297 were grownin CH medium containing 0.4 mM
IPTG at 30?C for 90 to 120 min then diluted into the same medium
to an OD600nmof 0.1. The diluted culture was divided into two, one
was supplemented with 0.5% xylose and an extra amount of IPTG
(to a final concentration of 1 mM) while the concentration of IPTG
of the other remained at 0.4 mM. The cultures were then grown at
37?C and at intervals portions were removed for staining with DAPI
This work was supported by a grant from the Biotechnology and
Biological Sciences Research Council. We thank Alan Grossman,
Richard Daniel, and Joy Rawlins for providing strains and vectors
and Leendert Hamoen for helpful comments on the manuscript.
Figure 6. An Extended Role for the Min System in B. subtilis
The MinCD division inhibitor activity (gray shading) accumulates at
the cell poles and extends out in a diminishing gradient toward
(A and B) Under normal conditions, the Min system is sufficient to
cells that have not yet replicated and segregated their chromosome
(gray ovals) (A), the Min system prevents division throughout the
length of the cell. As the cells grow, and replication and segregation
are completed (B), a Min-free space appears at mid-cell, allowing
division to occur in the correct place (arrow).
(C and D) Under certain abnormal conditions, nucleoid occlusion,
via Noc protein, is needed to prevent inappropriate division (indi-
tion. (D) Elongated cells with a large Min-free zone.
Received: October 28, 2003
Revised: April 23, 2004
Accepted: April 27, 2004
Published: June 24, 2004
Brown, N.C. (1971). Inhibition of bacterial DNA replication by
conservative synthesis in Bacillus subtilis. J. Mol. Biol. 59, 1–16.
Cha, J.-H., and Stewart, G.C. (1997). The divIVA minicell locus of
Bacillus subtilis. J. Bacteriol. 179, 1671–1683.
cation of cell division sites in Escherichia coli. J. Bacteriol. 181,
Daniel, R.A., Williams, A.M., and Errington, J. (1996). A complex
four-gene operon containing essential cell division gene pbpB in
Bacillus subtilis. J. Bacteriol. 178, 2343–2350.
de Boer, P.A.J., Crossley, R.E., and Rothfield, L.I. (1989). A division
locus determine proper placement of the division septum in E. coli.
Cell 56, 641–649.
de Boer, P.A.J., Crossley, R.E., and Rothfield, L.I. (1990). Central
role for the Escherichia coli minC gene product in two different cell
division-inhibition systems. Proc. Natl. Acad. Sci. USA 87, 1129–
de Boer, P.A.J., Crossley, R.E., and Rothfield, L.I. (1992). Roles of
MinC and MinD in the site-specific septation block mediated by the
MinCDE system of Escherichia coli. J. Bacteriol. 174, 63–70.
Edwards, D.H., and Errington, J. (1997). The Bacillus subtilis DivIVA
protein targets to the division septum and controls the site specific-
ity of cell division. Mol. Microbiol. 24, 905–915.
Errington, J., Daniel, R.A., and Scheffers, D.J. (2003). Cytokinesis in
bacteria. Microbiol. Mol. Biol. Rev. 67, 52–65.
Foley, M., Brass, J.M., Birmingham, J., Cook, W.R., Garland, P.B.,
Higgins, C.F., and Rothfield, L.I. (1989). Compartmentalization of
the periplasm at cell division sites in Escherichia coli as shown by
fluorescence photobleaching experiments. Mol. Microbiol. 3, 1329–
Fu, X., Shih, Y.-L., and Rothfield, L.I. (2001). The MinE ring required
for proper placement of the division site is a mobile structure that
cells, the Min system cannot direct proper placement
of the division site in cells that are abnormally long or
gate their replicated chromosomes (Figures 6C and 6D).
Here a nucleoid occlusion system becomes crucial.
Many bacteria do not contain either a Min system or a
recognizable homolog of Noc. It follows that bacteria
may have evolved a range of different systems to regu-
late the correct timing and positioning of their division
Bacterial Strains and General Methods
These are detailed in the Supplemental Data on the Cell website.
Depletion of MinD, FtsL/PbpB, or DnaA by Removal of IPTG
To deplete MinD, FtsL/PbpB, or DnaA, cells grown in either PAB or
CH with 0.5 mM IPTG were diluted (1 in 100) into the same medium
and grown to midexponential phase. Cells were then pelleted,
washed once, and diluted in the medium without IPTG to an OD600nm
of 0.05 or 0.1. The culture was then divided into two and IPTG was
added into one of the cultures and omitted from the other.
Overexpression of YyaA
Strains 1289 and 1290 were grown at 37?C in PAB to exponential
Nucleoid Occlusion Protein in B. subtilis Download full-text
changes its cellular location during the Escherichia coli division
cycle. Proc. Natl. Acad. Sci. USA 98, 980–985.
Glaser, P., Sharpe, M.E., Raether, B., Perego, M., Ohlsen, K., and
Errington, J. (1997). Dynamic, mitotic-like behaviour of a bacterial
protein required for accurate chromosome partitioning. Genes Dev.
Gueiros-Filho, F.J.,and Losick,R. (2002).A widelyconserved bacte-
rial cell division protein that promotes assembly of the tubulin-like
protein FtsZ. Genes Dev. 16, 2544–2556.
tion cycle of the cell division regulator MinE in Escherichia coli.
EMBO J. 20, 1563–1572.
Hamoen, L.W., and Errington, J. (2003). Polar targeting of DivIVA in
Bacillus subtilis is not directly dependent on FtsZ or PBP 2B. J.
Bacteriol. 185, 693–697.
Harry, E.J. (2001). Bacterialcell division: regulating Z-ring formation.
Mol. Microbiol. 40, 795–803.
Harry, E.J., Rodwell, J., and Wake, R.G. (1999). Co-ordinating DNA
replication with cell division in bacteria: a link between the early
stages of a round of replication and mid-cell Z ring assembly. Mol.
Microbiol. 33, 33–40.
Hu, Z., and Lutkenhaus, J. (2001). Topological regulation of cell
division in E. coli. spatiotemporal oscillation of MinD requires stimu-
lation of its ATPase by MinE and phospholipid. Mol. Cell 7, 1337–
Hu, Z., Mukherjee, A., Pichoff, S., and Lutkenhaus, J. (1999). The
MinC component of the division site selection system in Escherichia
coli interacts with FtsZ to prevent polymerization. Proc. Natl. Acad.
Sci. USA 96, 14819–14824.
Hu, Z., Gogol, E.P., and Lutkenhaus, J. (2002). Dynamic assembly
of MinD on phospholipid vesicles regulated by ATP and MinE. Proc.
Natl. Acad. Sci. USA 99, 6761–6766.
Kawai, Y., Moriya, S., and Ogasawara, N. (2003). Identification of a
protein, YneA, responsible for cell division suppression during the
SOS response in Bacillus subtilis. Mol. Microbiol. 47, 1113–1122.
Levin, P.A., Kurtser, I.G., and Grossman, A.D. (1999). Identification
and characterization of a negative regulator of FtsZ ring formation
in Bacillus subtilis. Proc. Natl. Acad. Sci. USA 96, 9642–9647.
tion of a chromosome partition protein in Bacillus subtilis. Proc.
Natl. Acad. Sci. USA 94, 4721–4726.
Margolin, W. (2001). Spatial regulation of cytokinesis in bacteria.
Curr. Opin. Microbiol. 4, 647–652.
Marston, A.L., Thomaides, H.B., Edwards, D.H., Sharpe, M.E., and
Errington, J. (1998). Polar localization of the MinD protein of Bacillus
subtilis and its role in selection of the mid-cell division site. Genes
Dev. 12, 3419–3430.
Moriya, S., Tsujikawa, E., Hassan, A.K.M., Asai, K., Kodama, T.,
and Ogasawara, N. (1998). A Bacillus subtilis gene-encoding protein
homologous to eukaryotic SMC motor protein is necessary for chro-
mosome partition. Mol. Microbiol. 29, 179–187.
influence positioning of division sites in Escherichia coli filaments
forming cells lacking DNA. J. Bacteriol. 171, 4303–4314.
Niki, H., Jaffe, A., Imamura, R., Ogura, T., and Hiraga, S. (1991). The
new gene mukB codes for a 177 kd protein with coiled-coil domains
Pichoff, S., and Lutkenhaus, J. (2002). Unique and overlapping roles
for ZipA and FtsA in septal ring assembly in Escherichia coli. EMBO
J. 21, 685–693.
Quisel, J.D., Burkholder, W.F., and Grossman, A.D. (2001). In vivo
effects of sporulation kinases on mutant Spo0A proteins in Bacillus
subtilis. J. Bacteriol. 183, 6573–6578.
Rajagopalan, S., Wachtler, V., and Balasubramanian, M. (2003). Cy-
tokinesis in fission yeast: a story of rings, rafts and walls. Trends
Genet. 19, 403–408.
Raskin, D.M., and De Boer, P.A.J. (1997). The MinE ring: an FtsZ-
independent cell structure required for selection of the correct divi-
sion site in E. coli. Cell 91, 685–694.
Regamey, A., Harry, E.J., and Wake, R.G. (2000). Mid-cell Z ring
assembly in the absence of entry into the elongation phase of the
round of replication in bacteria: co-ordinating chromosome replica-
tion with cell division. Mol. Microbiol. 38, 423–434.
Rowland, S.L., Katis, V.L., Partridge, S.R., and Wake, R.G. (1997).
DivIB, FtsZ and cell division in Bacillus subtilis. Mol. Microbiol.
Sharpe, M.E., and Errington, J. (1995). Postseptational chromosome
partitioning in bacteria. Proc. Natl. Acad. Sci. USA 92, 8630–8634.
Sharpe, M.E., Hauser, P.M., Sharpe, R.G., and Errington, J. (1998).
Bacillus subtilis cell cycle as studied by fluorescence microscopy:
constancy of the cell length at initiation of DNA replication and
evidence for active nucleoid partitioning. J. Bacteriol. 180, 547–555.
Shih, Y.L., Le, T., and Rothfield, L. (2003). Division site selection in
Escherichia coli involves dynamic redistribution of Min proteins
within coiled structures that extend between the two cell poles.
Proc. Natl. Acad. Sci. USA 100, 7865–7870.
Sievers,J., Raether,B.,Perego, M.,andErrington,J. (2002).Charac-
terization of the parB-like yyaA gene of Bacillus subtilis. J. Bacteriol.
Woldringh, C.L. (2002). The role of co-transcriptional translation and
tion. Mol. Microbiol. 45, 17–29.
Woldringh, C.L., Mulder, E., Valkenburg, J.A., Wientjes, F.B., Zarit-
sky, A., and Nanninga, N. (1990). Role of the nucleoid in the topore-
gulation of division. Res. Microbiol. 141, 39–49.
Wu, L.J., and Errington, J. (2003). RacA and the Soj-Spo0J system
combine to effect polar chromosome segregation in sporulating
Bacillus subtilis. Mol. Microbiol. 49, 1463–1475.
Yu, X.C., and Margolin, W. (1999). FtsZ ring clusters in min and
partition mutants: role of both the Min system and the nucleoid in
regulating FtsZ ring localization. Mol. Microbiol. 32, 315–326.
Yu, X.C., and Margolin, W. (2000). Deletion of the min operon results
in increased thermosensitivity of an ftsZ84 mutant and abnormal
FtsZ ring assembly, placement, and disassembly. J. Bacteriol.