Intrinsic instability of the essential cell division protein FtsL of Bacillus subtilis and a role for DivIB protein in FtsL turnover

Molecular Microbiology (2000) 36(2), 278±289

Intrinsic instability of the essential cell division proteinFtsL of Bacillus subtilis and a role for DivIB protein inFtsL turnover

Richard A. Daniel and Jeff Errington*

Sir William Dunn School of Pathology,

University of Oxford, Oxford OX1 3RE, UK.

Summary

Cell division in most eubacteria is driven by an

assembly of about eight conserved division proteins.

These proteins form a ring structure that constricts in

parallel with the formation of the division septum.

Here, we show that one of the division proteins, FtsL,

is highly unstable. We also show that the protein is

targeted to the ring structure and that targeting

occurs in concert with the recruitment of several

other membrane-associated division proteins. FtsL

stability is further reduced in the absence of DivIB

protein (probably homologous to E. coli FtsQ) at high

temperature, suggesting that DivIB is involved in the

control of FtsL turnover. The reduced stability of FtsL

may explain the temperature dependence of divIB

mutants, because their phenotype can be suppressed

by overexpression of FtsL. The results provide new

insights into the roles of the FtsL and DivIB proteins

in bacterial cell division.

Introduction

Cell division in non-filamentous eubacteria is a tightly

regulated cyclic process that usually occurs after each

round of DNA replication and doubling in cell mass.

Division of the cell into two daughters involves co-

ordinated ingrowth of the cell envelope layers, comprising

the cytoplasmic membrane, wall (peptidoglycan) and, in

Gram-negative bacteria, the outer membrane (Bramhill,

1997; Lutkenhaus and Addinall, 1997).

About eight genes with varying degrees of conservation

have been shown to be required specifically for cell

division in bacteria, mainly through studies of Escherichia

coli and Bacillus subtilis. Two conserved genes, ftsA and

ftsZ, encode soluble proteins that act inside the cytoplas-

mic membrane. FtsZ is a prokaryotic form of tubulin,

which polymerizes in the form of a ring at the site of

division as an early event in the process (Bi and

Lutkenhaus, 1991; Wang and Lutkenhaus, 1993; Erickson

et al., 1996; LoÈwe and Amos, 1998). FtsA and various

transmembrane division proteins are recruited to the Z

ring in a sequence that is beginning to be elucidated (for a

recent summary, see Ghigo et al., 1999). The functions of

the transmembrane division proteins are not yet under-

stood, except for a septum-specific penicillin-binding

protein (PBP3 of E. coli; PBP 2B of B. subtilis), which is

presumably required for the final steps of peptidoglycan

synthesis in the septum (Ghuysen, 1991; Pogliano et al.,

1997; Daniel et al., 2000). FtsW (YlaO of B. subtilis) is an

integral membrane protein and is probably associated in

some way with the functioning of the PBP (Ikeda et al.,

1989; Boyle et al., 1997; Khattar et al., 1997).

The remaining division proteins all have a single

predicted membrane span with their major soluble

domains outside the cytoplasmic membrane (except

ZipA, which is inside). ZipA and FtsN of E. coli (Dai

et al., 1993; Addinall et al., 1997; Hale and De Boer, 1997;

1999; Liu et al., 1999) do not seem to have homologues in

B. subtilis, and B. subtilis has at least one unique division

protein called DivIC (Levin and Losick, 1994; Katis et al.,

1997). Two proteins, FtsL and DivIB/FtsQ, display limited

sequence similarity between E. coli and B. subtilis, but are

likely to be homologous based on, first, the conserved

chromosome position of their respective genes within

clusters of genes that are clearly homologous, secondly,

the similar size and charge distributions of the proteins

and, thirdly, the equivalence of their function based on the

cell division phenotype that arises in the absence of the

functional protein (Guzman et al., 1992; Ueki et al., 1992;

Harry et al., 1993; Harry and Wake, 1997; Daniel et al.,

1998; Chen et al., 1999). However, DivIB may differ from

FtsQ in being dispensable at lower temperatures (Beall

and Lutkenhaus, 1989; Harry et al., 1993).

The ftsL gene of B. subtilis lies immediately upstream of

the pbpB gene encoding PBP 2B. Previously, we showed

that depletion of FtsL by repression of its gene results in a

rapid arrest of cell division (Daniel et al., 1998). Its product

is a small (121 amino acids) protein with three domains:

a short (35 amino acids) cytoplasmic N-terminal domain, a

hydrophobic single membrane-spanning domain and a

longer (64 amino acids) extracellular domain.TheC-terminal

domain of FtsL is predicted to be capable of forming a

Q 2000 Blackwell Science Ltd

Received 19 January, 2000; accepted 26 January, 2000. Forcorrespondence. E-mail [email protected]; Tel. (144) 1865275561; Fax (144) 1865 275556.

Control of cell division in B. subtilis 279

Q 2000 Blackwell Science Ltd, Molecular Microbiology, 36, 278±289

coiled-coil structure (Lupas et al., 1991), which is typical

of proteins that form oligomeric structures (Lupas, 1996).

DivIC has a similar structure, raising the possibility that

the two proteins could form a mixed oligomeric complex

(Levin and Losick, 1994). In support of this idea, depletion

of FtsL results in a rapid loss of DivIC protein, showing

that DivIC stability requires FtsL (Daniel et al., 1998).

In this paper, we describe the construction and use of

epitope-tagged derivatives of FtsL to examine the

localization and stability of FtsL protein under various

conditions. We show that localization at the division site

requires (directly or indirectly) all of the known bitopic

division proteins, DivIB, DivIC and PBP 2B. We show that

FtsL is an intrinsically unstable protein and that its basic

level of stability is dependent on DivIB at higher

temperatures. Furthermore, we show that the requirement

for DivIB at higher growth temperatures can be largely

suppressed by overexpression of FtsL, indicating that the

only important role for DivIB in cell division is in

maintaining the stability of FtsL.

Results

FtsL is targeted to cell division sites

We initially attempted to determine the localization of FtsL

by immunofluorescence microscopy (IFM) by raising a

polyclonal antiserum. Unfortunately, this proved to be

problematical. The gene was extremely toxic when

expressed even at low levels in E. coli. Also, the protein

was highly unstable (as it also proved to be in B. subtilis;

see below). Eventually, an antiserum that reacted well

with FtsL was obtained, but its specificity was not great

enough for IFM. To overcome this problem, we con-

structed strains in which an HA epitope tag was fused to

the N- or C-terminus of FtsL. The fusions were put under

the control of the xylose-inducible Pxyl promoter and

inserted into a prophage vector or at the amyE locus. Both

N- and C-terminal epitope-tagged constructions gave

functional hybrid proteins, as judged by the ability of

strains containing them to grow and divide normally in the

absence of any other FtsL protein in the cell. Western

blotting showed that strains with the tagged proteins

contained a single protein that reacted with the anti-HA

monoclonal antiserum, which had the expected mobility

for a protein of about the size of FtsL and was produced

only in the presence of xylose (see below). Like the wild-

type FtsL, the fusion proteins were unstable and only

detected in lysates made in the presence of proteinase

inhibitors.

Strains containing the tagged FtsL proteins were

examined by IFM. Similar results were obtained with

both fusions, so only data for the FtsL±HA fusion (i.e.

fusion of the HA tag to the C-terminus of FtsL) have been

shown. Figure 1A and B shows a typical field of cells of

strain 812. Under the growth and fixation conditions used,

mature septa were just visible as invaginations by phase-

contrast microscopy (black arrowheads in Fig. 1A). (B.

subtilis sister cells tend to remain connected together in

chains for some time after the division septum has been

formed; Paulton, 1971.) Prominent localizations of FtsL

were present in most of the cells (Fig. 1B). In the majority

of cases, the staining was in the form of bands or pairs of

spots, which were invariably located at about mid-cell,

where ongoing or incipient cell division would be expected

(e.g. cells labelled a and b). In general, such cells did not

have a visible constriction of the cell wall at the equivalent

position in the phase-contrast images. In other cells,

shorter bands of fluorescence were observed (e.g. cells c,

d and e). Such cells sometimes showed a slight

invagination, indicating that division had progressed

further (e.g. cell e). A few cells did not show significant

staining (e.g. cell f). Quantification of the staining

according to the cell length (Fig. 1C) showed that the

cells with no localization of FtsL tended to be shorter (and

therefore younger). As the cell length increased, cells

tended to show the `two spots' or `band' pattern. Finally, at

the upper end of the cell length range, cells with a `single

spot' of FtsL predominated, apparently representing cells

that had completed or almost completed septation. Similar

results have since been obtained with a fusion of FtsL to

green fluorescent protein (J. Sievers and J. Errington,

submitted).

To confirm that the mid-cell bands and spots of FtsL

represented division sites, samples were co-stained for

FtsZ, which is at the top of the hierarchy of division protein

assembly (see Introduction). Figure 1D±F shows that, as

expected, FtsZ and FtsL co-localized in the majority of

cells, whether the staining pattern was of the band or

central spot pattern. A few cells (9%) contained an FtsZ

band with no equivalent FtsL staining (arrowed in Fig. 1F).

Although this would be consistent with FtsZ assembling

before FtsL, we cannot exclude the possibility that this

was caused by less efficient staining of the FtsZ protein,

as the conditions used to detect FtsL are not ideal for the

detection of FtsZ.

Similar Western and protein localization results were

obtained with strain 818, in which the epitope-tagged FtsL

was expressed from its normal promoter, although the

signals obtained were weaker (not shown).

Dependence of FtsL localization on other cell division

genes

To investigate the dependence of FtsL localization on the

other known cell division proteins, conditional strains were

constructed in which the epitope-tagged FtsL could be

viewed in the absence of a functional division protein. In

280 R. A. Daniel and J. Errington

Q 2000 Blackwell Science Ltd, Molecular Microbiology, 36, 278±289

the case of FtsZ, DivIB and DivIC, temperature-sensitive

mutations were used. The localization of FtsL±HA was

then determined at both permissive (308C) and non-

permissive (488C) temperatures. Figure 2A shows a

phase-contrast image of a field of wild-type cells grown

at 488C with the corresponding immunofluorescence

image for FtsL. Localization was normal, showing that

the higher temperature does not affect targeting of FtsL.

FtsL localization was also normal in each of the

temperature-sensitive mutants when incubated at the

permissive temperature (data not shown). However, after

40 min incubation at the non-permissive temperature, no

significant localizations of FtsL were detected (. 100

filaments counted) in any of the mutants. Examples of the

typical filaments are shown in Fig. 2B±D. We then

examined the effect of depletion of PBP 2B, a penicillin-

binding protein specific for synthesis of septal wall

material, by removal of IPTG from cells of strain 812.

Again, FtsL localization was normal in the presence of

inducer (IPTG) (Fig. 1A and B) but, in the filamentous

cells made in the absence of inducer, localization was

inhibited, although many cells showed staining at about

mid-filament (Fig. 2E). A similar pattern of staining was

shown by DivIB and DivIC proteins when PBP 2B was

depleted (Daniel et al., 2000). The mid-cell bands seem to

correspond to arrested or slowly forming division septa

that were presumably initiated just as PBP 2B was

becoming limiting. We conclude that stable association

of FtsL with the division site requires most, if not all, other

division proteins.

FtsL is lost rapidly when expression is blocked

It has been shown previously that repression of ftsL by

removal of xylose in cells bearing a Pxyl-ftsL fusion results

in a rapid arrest of cell division (Daniel et al., 1998).

Repression of the epitope-tagged FtsL derivatives (strains

811 and 812) produced a similar phenotype (not shown).

Fig. 1. Localization of FtsL at cell division sites.A and B. Phase-contrast (A) and immunofluorescence (B) images of cells of strain 812 (Pxyl-HA-ftsL) grown in PAB containing xylose (and IPTG tomaintain expression of the inducible pbpB gene in this strain). Cells with characteristic staining are labelled a±e (see text).C. Length distribution of cells prepared as in (B). Cells were scored according to their FtsL staining: black, no FtsL band; white, two dots or bandpattern; grey, central spot.D±F. Co-localization of HA-FtsL and FtsZ.D. Localization of HA-FtsL, false coloured in green.E. Localization of FtsZ, false coloured in red.F. Overlay of the green and red images. Small arrows show examples of FtsL and FtsZ co-localization. The arrowhead marks an FtsZ localizationwith no corresponding FtsL signal. Scale bar � 3 mm.

Control of cell division in B. subtilis 281

Q 2000 Blackwell Science Ltd, Molecular Microbiology, 36, 278±289

Western blot analysis showed that the tagged FtsL

proteins were rapidly degraded, being almost undetect-

able 10 min after the removal of xylose from the culture

medium (Fig. 3B and C, lanes 1±4). To confirm that the

loss of the tagged FtsL was caused by the removal of the

inducer from the culture, xylose was added back to the

Fig. 2. Dependence of FtsL targeting on the function of other divisionproteins. The cells were grown in PAB containing xylose at 488C for40 min (A±D) or at 378C in the absence of IPTG for 60 min (E).Samples were stained for the HA tag on FtsL, and correspondingphase-contrast images are shown on the left.A. Wild-type control (strain 812 grown in the presence of IPTG).B. ftsZts (strain 819).C. divIBts (strain 820).D. divICts (strain 821).E. Pspac-pbpB (strain 812). Localizations of FtsL are marked withwhite arrows in (E). Scale bar � 3 mm.

Fig. 3. Western blot analysis of epitope-tagged FtsL in the presenceand absence of various functional cell division proteins.A. Detection of FtsL-HA (HA tag fused to the C-terminus of FtsL)when expressed from the natural ftsL promoter (strain 818) or fromPxyl (strain 812). Cultures were grown in PAB containing xylose, andsamples of total protein were separated by SDS±PAGE and probedwith anti-HA monoclonal antibody. Samples were taken from earlyexponential growth (lanes 1, 4 and 7), mid-exponential (lanes 2, 5 and8) and early stationary phase (lanes 3, 6 and 9).B and C. Detection of FtsL when tagged at its C-terminus (FtsL-HA)(B; strain 812) or its N-terminus (HA-FtsL) (E; strain 811), and therapid disappearance of the proteins after repression of geneexpression. Strains were grown in PAB containing xylose, thencentrifuged and resuspended in xylose-free medium. Samples weretaken immediately after resuspension (lane 1) and at 5 min intervalsthereafter (lanes 2±4). Xylose was then added back to the cultures,and further samples were taken after 10 and 30 min (lanes 5 and 6).Lane M contained a HA-tagged molecular weight marker (Roche;16.5 kDa marker shown). Control samples were obtained by returningan aliquot of the centrifuged cells to medium containing xylose, thensampling after 0, 5 and 20 min (lanes 7±9).D±G. Stability of FtsL-HA in a wild-type division background (D; strain818) and in the absence of functional FtsZ (D; strain 822), DivIB (F,strain 824) or DivIC (G, strain 825). Exponentially growing cultures ofthe strains were shifted from 308C to 488C, and samples were takenimmediately (time 0) and at 10 min intervals thereafter. Proteinamounts loaded were from a constant volume of culture.

282 R. A. Daniel and J. Errington

Q 2000 Blackwell Science Ltd, Molecular Microbiology, 36, 278±289

depleted cultures and the tagged proteins reappeared

(Fig. 3B and C, lanes 5 and 6). Interestingly, when cells

grown in the presence of xylose were centrifuged and

resuspended back into xylose-containing medium, a

transient reduction in FtsL signal was reproducibly

detected (Fig. 3B and C, lanes 7±9), suggesting that

FtsL is destabilized when growth of the culture is

disturbed by centrifugation.

Figure 3A shows that FtsL was overexpressed when

placed under the control of the Pxyl promoter compared

with its natural promoter, as mentioned above. This

experiment also indicated that the expression of FtsL

from its natural promoter showed some growth phase

dependence similar to the expression pattern previously

determined by lacZ fusion analysis (Daniel et al., 1996).

The abundance of FtsL-HA increased during exponential

growth (Fig. 2A, lanes 1 and 2), and then appeared to drop

slightly on entry into stationary phase (Fig. 2A, lane 3).

Increased thermal instability of FtsL in the absence of

DivIB

Strains containing conditional alleles of various division

genes were used to test the stability of FtsL under

Fig. 4. Suppression of the temperature-sensitive phenotype of a divIB null mutant by overproduction of FtsL.A±D. Effect of incubation at 498C on culture growth (OD600). Wild-type (A, strain 168), divIB (B, strain 826), Pxyl-ftsL (C, strain 816) and divIB,Pxyl-ftsL (D, strain 827) strains were grown in PAB in the presence (filled symbols) and absence (open symbols) of xylose.E±H. Effect of incubation at 468C on cell division. Various strains were grown in PAB with or without xylose and shifted to 468C. Typical phase-contrast images of cultures taken 60 min after temperature shift up are shown to the right, and quantitative cell length distributions to the left.E and F. divIB Pxyl-ftsL mutant (strain 827) grown in the absence (E) and presence (F) of xylose.G and H. divIB mutant (G, strain 826) and Pxyl-ftsL strain (H, strain 816) grown in the presence of xylose. In the histograms, grey columns representthe underestimated lengths of filamentous cells that extended beyond the microscopic field, and filled columns represent complete cells. Septawithin chains of cells are marked with arrows on the phase-contrast images. Scale bar � 3 mm.

Control of cell division in B. subtilis 283

Q 2000 Blackwell Science Ltd, Molecular Microbiology, 36, 278±289

different conditions of division arrest. Western blot

analysis of total protein extracts from the various mutants,

in the non-permissive condition, showed that the stability

of FtsL-HA was reduced by any block to cell division, but

the degree by which it was affected depended upon the

division defect. The ftsZ ts-1 temperature-sensitive allele

had relatively little effect on the stability of FtsL (Fig. 3E),

compared with the wild-type strain (Fig. 3D). In the last

sample (40 min), the reduction in the FtsL signal was

probably caused by some cell lysis. In the divIC355

conditional strain, a more severe destabilization of FtsL

was evident (Fig. 3G). However, although FtsL decayed

gradually, it was still readily detectable, even in late

samples. In other experiments, greater amounts of FtsL

were often detected, but they often disappeared abruptly,

probably in association with cell death and or lysis (data

not shown). In contrast, the divIB tms-1 culture showed an

almost total loss of FtsL within 10 min of being moved to

488C (Fig. 3F). Even in the first sample taken immediately

after the temperature shift, FtsL was clearly present in

reduced amounts. Similar results were obtained with a

null mutation of divIB and, in several repeats of these

experiments, FtsL was always undetectable after the first

sample (not shown).

Suppression of the filamentous phenotype of a divIB null

mutant by overexpression of ftsL

DivIB is an unusual cell division gene in that a

temperature-sensitive phenotype can be generated by a

null mutation in the gene (Beall and Lutkenhaus, 1989;

Harry et al., 1993). On the basis of the results described

above, it was possible that the temperature dependence

of cell division caused by the loss of divIB resulted

predominantly from destabilization of FtsL. Thus, the role

of DivIB in division could be indirect and exerted via

stabilization of FtsL at higher temperatures. If this was

true, overexpression of FtsL might suppress this division

defect. To test this, we introduced a Pxyl-ftsL construction

into cells also bearing ftsL at its normal position, with and

without a divIB disruption. Preliminary experiments with

this strain showed that xylose did indeed improve the

viability of the divIB mutant at 488C. Figure 4A±D shows

the growth of various strains after shifting to 488C in the

presence and absence of xylose. The wild-type strain

showed approximately exponential growth during the

course of the experiment, and the presence or absence

of xylose had little or no effect. The divIB null mutant

(Fig. 4B) showed a similar rate of growth for the first

60 min after the temperature shift, but then growth was

arrested and the culture started to lyse. As expected,

xylose had no significant effect on the pattern of growth

and lysis. Figure 4C shows that the introduction of the

Pxyl-ftsL construct into an otherwise wild-type (divIB1)

strain had no effect on growth. However, in the presence

of xylose, this construction rescued the growth arrest

phenotype caused by the divIB null mutation (Fig. 4D).

Thus, it appeared that overexpression of ftsL could indeed

compensate for the absence of DivIB.

Microscopic examination of the cells showed that

rescue of the growth phenotype was accompanied by a

substantial increase in the septation frequency of the

cells. Figure 4E±H shows images of typical fields of cells

of the crucial cultures, together with a quantification of cell

length. (These experiments were carried out at a slightly

lower temperature, 468C, which reduced the amount of

cell lysis, making the cell length measurements more

reliable.) In the absence of xylose, the divIB mutant

showed massively elongated cell filaments, as expected

for a block in division (Fig. 4E). A similar phenotype was

seen in the derivative additionally carrying Pxyl-ftsL but in

the absence of xylose (Fig. 4F). However, in the presence

of xylose (Fig. 4G), the same strain showed a greatly

increased frequency of septation, giving a cell length

distribution much closer to that of the equivalent divIB1

strain under similar conditions (Fig. 4H). Although there

may be other more subtle defects in this strain, these

results suggest that the major part of the division defect in

the divIB mutant results from FtsL degradation.

Discussion

Interdependent assembly pathway for the membrane-

associated division proteins of B. subtilis

We used epitope-tagged derivatives of ftsL to examine

the subcellular localization of its protein in B. subtilis. As

for all previously examined proteins specifically required

for cell division, FtsL was found to target to the division

site. This was demonstrated with both N- and C-terminally

tagged proteins, expressed from either the natural

promoter or Pxyl. As expected, the protein co-localized

almost completely with FtsZ protein, which is generally

thought to be the first protein to target to the division site

(Addinall et al., 1996; Ghigo et al., 1999). The pattern of

FtsL staining changed during cell cycle progression from

no localization, in newborn (short) cells, through a band to

a centre spot in the longest cells, as described previously

for proteins DivIB and DivIC (Harry and Wake, 1997; Katis

et al., 1997). The band pattern presumably corresponds

to cells that have assembled the division machinery

around the circumference of the cell but not yet under-

taken significant septal invagination. The centre spot

pattern presumably arises by removal of FtsL during

septum invagination, or by it remaining associated with

the leading edge of the septum as the annulus constricts.

FtsL was generally not associated with mature cell poles,

consistent with the protein being released from the

284 R. A. Daniel and J. Errington

Q 2000 Blackwell Science Ltd, Molecular Microbiology, 36, 278±289

septum during or soon after its formation. Other proteins,

notably DivIVA and PBP 2B, seem to remain associated

with the mature pole to a greater degree (Marston et al.,

1998; Daniel et al., 2000).

Introduction of the epitope-tagged FtsL derivatives into

strains bearing conditional alleles of other cell division

genes allowed us to test the dependence of FtsL targeting

on other division proteins. As expected, recruitment of

FtsL to potential division sites was abolished in the

absence of the FtsZ protein. It was also abolished or

substantially reduced by mutations affecting all the other

division genes that we tested, including, to our surprise,

the pbpB gene (Fig. 2). In E. coli, the likely homologue of

FtsL (see Introduction) was recently shown to target to the

division site independently of pbpB (ftsI), although the

frequency of targeting was reduced (Ghigo et al., 1999).

However, we showed recently that depletion of B. subtilis

PBP 2B has an unusual effect: the cells elongate, and

assembly of late (DivIB and DivIC) but not early (FtsZ)

proteins at potential division sites is arrested, indicating a

requirement for PBP 2B for targeting of the late proteins

(Daniel et al., 2000). At about the mid-point of each

filament, structures corresponding to arrested or greatly

slowed invaginating septa form, presumably containing

residual amounts of PBP 2B, and these structures seem

to contain all the division proteins, including DivIB, DivIC

and now FtsL. The similar targeting of FtsL to the mid-

filament structures but not elsewhere in elongated PBP

2B-depleted filaments leads us to conclude that targeting

of FtsL to the division site is dependent on PBP 2B. The

reduced frequency of targeting of FtsL in PBP3ts filaments

of E. coli (Ghigo et al., 1999) could reflect a similar effect.

The apparent interdependence of assembly of mem-

brane-associated division proteins in B. subtilis contrasts

with the linear pathway described recently for E. coli (Ghigo

et al., 1999). In principle, several different kinds of mechan-

ism could produce an interdependent pathway of assembly.

For example, the proteins might assemble at the Z ring in a

defined order, but the complete or partial structures formed

might be too unstable in the absence of any one of the

factors for targeting to be detected. Alternatively, the

proteins might interact with each other in a defined order

away from the Z ring and only be recruited to the ring when

this preassembly step is complete. In either case, it is

formally possible that there is no strict order of assembly of

the components, with several different subpathways

operating in parallel, all leading to the assembled or

preassembled structure. Clearly, experiments to investi-

gate pairwise interactions between these proteins are now

needed to be able to resolve these questions.

A function for DivIB (FtsQ) in regulation of FtsL turnover

A broad range of experimental observations all pointed to

FtsL being an extremely unstable protein. The protein

proved to be very difficult to purify, and an antiserum

that readily detected FtsL fusion proteins overexpressed

in E. coli did not detect the native protein in B. subtilis cell

extracts (data not shown). The epitope-tagged proteins

provided the most sensitive means of monitoring the

protein. Western blot analysis of these fusions showed

that the proteins disappear rapidly after removal of an

inducer of transcription of the gene (Fig. 3B and C). The

protein also disappeared transiently from extracts of cells

that had been centrifuged. Such instability almost

certainly explains our previous observations that repres-

sion of ftsL transcription leads to a rapid arrest in cell

division (Daniel et al., 1998).

The rapid turn off and turn on of division when ftsL

expression is artificially modulated (Daniel et al., 1998;

Feucht et al., 1999) suggests that it plays a pivotal role in

regulation of division. In general, it seems that unstable

proteins are frequently involved in regulation (reviewed by

Gottesman and Maurizi, 1992). There are several ways in

which such regulation could operate. We favour the notion

that FtsL is stabilized by interaction with one or more

proteins during assembly of the division machinery. As

constriction begins, FtsL protein is released, and its

degradation could help to prevent immediate reassembly

of the machinery and formation of a second septum.

However, we cannot yet exclude other possibilities. FtsL

might only be synthesized transiently during cell cycle

progression, in which case instability would result in its

rapid elimination after synthesis stops. Periodic accumu-

lation of FtsL could then help to drive the division cycle,

rather like the cyclins of eukaryotic cells (Roberts, 1999).

However, the ability of cells with ftsL under the control of

an ectopic promoter (e.g. Pxyl) to divide shows that, even if

it exists, such regulation is not essential. It is possible that

modulation of FtsL synthesis or stability could be used as a

regulatory check point to block cell division in response to

certain stimuli. B. subtilis appears not to have a homologue

of the SulA protein that blocks cell division when the SOS

response is induced by DNA damage (Huisman and D'Ari,

1981; Mukherjee et al., 1998), so it will be interesting to test

whether FtsL could play a role in this kind of regulation.

Finally, it is possible that FtsL instability is simply a

mechanism for controlling its stoichiometry relative to other

components of the division machinery, in which case the

degradation is used to degrade excess subunits that are

not assembled at the division site.

Whatever the function of instability, DivIB seems to be a

crucial factor in FtsL turnover. Unlike the results obtained

with other division mutants, in the absence of DivIB, at

higher temperatures, FtsL disappeared completely and

extremely rapidly. Thus, DivIB plays some role in

stabilizing FtsL at higher growth temperatures. It seems

likely that FtsL binds to or otherwise associates with DivIB

Control of cell division in B. subtilis 285

Q 2000 Blackwell Science Ltd, Molecular Microbiology, 36, 278±289

and that it is stabilised by this interaction. This stabiliza-

tion is apparently independent of whether the two proteins

have been recruited to the division ring because, in

filaments of an ftsZts mutant, targeting to division sites

was blocked, but FtsL remained fairly stable. The

intermediate level of instability seen in a divICts mutant

suggests that DivIC also influences FtsL stability, but

most probably by a more indirect route.

Discovery of the role of DivIB in the protection of FtsL

from degradation raised the interesting possibility that

degradation of FtsL was responsible for the long-studied

division phenotype of divIB mutants. divIB is an unusual

division gene in that null mutations have a temperature-

sensitive phenotype (Beall and Lutkenhaus, 1989; Harry

et al., 1993). Usually, thermosensitivity arises from an

amino acid substitution that reduces the thermal stability

of the protein directly. The thermosensitivity of divIB null

mutants suggests that the normal function of DivIB is to

protect cells from the deleterious effects of some

intrinsically thermosensitive process. In the light of the

experiments illustrated in Fig. 3D±H, it was possible that

the intrinsic thermoinstability of FtsL, presumably an

important feature of its function, was the process on

which DivIB acted. If so, overexpression of FtsL might at

least partially suppress a divIB null mutation and, indeed,

we showed that this was the case. This striking result

suggests that, at high temperature, the only essential role

of DivIB in division is the indirect one of stabilizing FtsL.

Although DivIB is not well conserved in terms of primary

amino acid sequence, its size, membrane topology,

division phenotype and chromosome position all suggest

that it is a homologue of FtsQ of E. coli (see Introduction).

However, published work on ftsQ null mutants all points to

the gene being essential (Carson et al., 1991; Guzman

et al., 1997), although the question of whether they are

viable at low temperature may not have been tested.

Nevertheless, our results now suggest that DivIB/FtsQ

may have a chaperone-like role, protecting FtsL from

degradation, which in turn probably plays a role in

regulating the assembly of the membrane-associated

division proteins of B. subtilis. A major challenge now is to

elucidate the nature of the interactions between these

proteins that leads to their tightly regulated assembly at

the site of septum formation.

Experimental procedures

Bacterial strains and plasmids

The B. subtilis strains used are listed in Table 1.

General methods

B. subtilis strains were grown in Difco antibiotic medium 3

(PAB) supplemented where required with xylose (0.5%) and/or IPTG (0.5 mM). Escherichia coli strains were grown at378C in 2� YT (Sambrook et al., 1989), supplemented withampicillin (100 mg ml21) and/or kanamycin (25 mg ml21), asnecessary.

B. subtilis strains were transformed according to themethod of Anagnostopoulos and Spizizen (1961) as modifiedby Jenkinson (1983). Selection for B. subtilis transformantswas carried out on nutrient agar (Oxoid), supplemented withchloramphenicol (5 mg ml21), kanamycin (5 mg ml21) and/orspectinomycin (50 mg ml21), with xylose (0.5%) and IPTG(0.5 mM), as necessary.

Images of cell fields were acquired with a cooled CCDcamera (System 3000; Digital Pixel Advanced ImagingSystems) and analysed using Object-Image software(Vischer et al., 1994). The figures were then prepared usingAdobe Photoshop.

Expression and purification of FtsL in E. coli

Plasmid pQE FtsL was constructed by cloning a 394 bppolymerase chain reaction (PCR) product generated usingthe two oligonucleotides 6740 (5 0-CAGGGATCCAGCAATTTAGCTTACCAACC-3 0) and 6741 (5 0-CTTAGGTACCCTGCT-CCTCTATTC-3 0) into the pQE30 expression vector (Qiagen),using BamHI and KpnI restriction sites. The resulting plasmidencoded the whole ftsL gene fused to a His tag leadersequence.

Induction of this plasmid in E. coli strain NM554 containingplasmid pREP4 (Qiagen) produced a small yield of protein ofthe expected mobility. The protein was insoluble, so it waspartially purified by centrifugation and then fractionation on aBio-Rad Prep cell system. The eluted protein was concen-trated with a Centriplus concentrator (Amicon) and dialysedagainst phosphate-buffered saline (PBS). Rabbit polyclonalantiserum was raised against this purified protein by standardmethods (Harlow and Lane, 1988).

Construction of epitope-tagged fusion derivatives of FtsL

To fuse the HA epitope tag to the N-terminus of FtsL, thecoding sequence of ftsL was amplified using oligonucleotides6740 (5 0-CAGGGATCCAGCAATTTAGCTTACCAACC-3 0) and8964 (5 0-GGCATTTGGTACCTTCCTGTATGTTTTTCAC-3 0).The PCR product was digested with BamHI, and the DNAends were blunted. The resulting DNA was ligated to NaeI-digested plasmid pEPI 1, and the reaction products wereamplified with oligonucleotides A17 (5 0-AGGCTCTAGAA-AAGGAGGTGATGAAATGCCCAAAGGCCTCGCGAGTACTTACC-3 0) and 8964 (see above), so providing a ribosomebinding site and start codon. The products of this secondround of PCR were digested with XbaI and BstEII and clonedinto pRD99 to give plasmid pRD132. The C-terminal epitope-tagged derivative was made by a similar procedure. Thesame primary PCR product described above was digestedwith KpnI and blunted. The DNA was ligated to NruI-digestedpEPI 1, and the products were amplified by PCR witholigonucleotides 6740 (see above) and A18 (5 0-ACCGCT-GCAGTCATCAACCCGTTAACCCGGGCCGGC-3 0), whichintroduced a stop codon at the end of the epitope coding

286 R. A. Daniel and J. Errington

Q 2000 Blackwell Science Ltd, Molecular Microbiology, 36, 278±289

sequence. The products of this reaction were digested withBstEII and PstI and cloned into pRD99 to give pRD133.

To introduce these epitope-FtsL fusions into B. subtilis,three methods were used. First, the cassettes encoding thecat gene, Pxyl and the ftsL fusions from pRD132 and pRD133were excised using EcoRI and PstI, and the DNA ends wereblunted. These DNAs were then ligated to SmaI-digestedf105J106 DNA. This concatameric DNA was transformedinto a f105J125 lysogen of strain 168, with selection forchloramphenicol resistance. The transformants were pooled,and phage lysates were prepared by thermoinduction(Errington, 1990). The resultant lysates were then used toinfect 168, selecting for transduction of the chloramphenicolmarker. Six lysogens were purified for each type of FtsL

fusion, and an insertion into the natural ftsL gene wasintroduced by transformation with chromosomal DNA fromstrain 799, selecting for both chloramphenicol and kanamycinresistance, with xylose and IPTG both present. Transform-ants from each cross were screened for dependence on bothxylose (driving expression of the ftsL fusion gene) and IPTG(driving expression of the essential pbpB gene downstreamof ftsL) and, finally, for the xylose-dependent expression ofthe epitope-tagged protein by Western blot analysis. Trans-formants thus obtained were then checked for the correctgenetic construction by PCR, yielding strains 811 (N-terminaltag) and 812 (C-terminal tag).

Secondly, the cassettes from pRD132 and pRD133 wereisolated as EcoRI, PstI fragments and cloned into pRD154,

Table 1. Bacterial strains and plasmids.

Strain Relevant features Source/construction

B. subtilis168 trpC2 Laboratory stockts1 trp-3 ftsZ ts-1 R.G. Wake, Department of Biochemistry, University of

Sydney, AustraliaMB76 divIB tms-12 Laboratory stockSU347 divIC355 Katis et al. (1997)799 trpC2 D ftsL799 (lacI neo PspacpbpB) f105J505

( 0yllA yllB yllC ftsL pbpB)Daniel et al. (1996)

811 trpC2 D ftsL799 (lacI neo PspacpbpB) f105J507 (cat Pxyl -HA-ftsL) See Experimental procedures812 trpC2 D ftsL799 (lacI neo PspacpbpB) f105J508 (cat Pxyl-FtsL-HA) See Experimental procedures813 trpC2 VamyE813 (cat Pxyl -ftsL-HA) D ftsL799 (lacI neo PspacpbpB) See Experimental procedures814 trpC2 VamyE814 (cat Pxyl-HA-ftsL) D ftsL799 (lacI neo PspacpbpB) See Experimental procedures816 trpC2 VamyE816 (cat Pxyl -ftsL) See Experimental procedures817 trpC2 VamyE816 (cat Pxyl -ftsL) DdivIB826 (spc) 826!816a

818 trpC2 ftsL::pRD175 (ftsLHA cat Pxyl -0ftsL pbpB) pRD175!168

819 trpC2 ftsZts1 VamyE813 (cat Pxyl -0ftsL-HA) D ftsL799

(lacI neo Pspac-pbpB)See Experimental procedures

820 trpC2 divIBtms-12 VamyE813 (cat Pxyl -ftsL-HA) D ftsL799(lacI neo Pspac-pbpB)

See Experimental procedures

821 trpC2 divIC355 VamyE813 (cat Pxyl-ftsL-HA) D ftsL799(lacI neo Pspac-pbpB)

See Experimental procedures

822 trpC2 chr::pRD175 (ftsL-HA cat Pxyl -0ftsL pbpB) ftsZts-1 See Experimental procedures

824 trpC2 chr::pRD175 (ftsL-HA cat Pxyl -0ftsL pbpB) divIBtms-12 See Experimental procedures

825 trpC2 chr::pRD175 (ftsL-HA cat Pxyl -0ftsL pbpB) divIC355 See Experimental procedures

826 trpC2 DdivIB1306 (spc) 1306!1681306 trpC2 DdivIB1306 (spc) VamyE (gpr 0- 0 lacZ cat) Feucht et al. (1999)

Phagesé105J125 ind 125 cts-23 sal-104 mcs-23 D (DI:1t) Thornewell et al. (1993)é105J505 ind 125 cts-23 sal-104 mcs-23 D (DI:1t) 0yllA yllB yllC ftsL pbpB Daniel et al. (1996)é105J506 ind 125 cts-23 sal-104 mcs-23 D (DI:1t) cat Pxyl-ftsL Daniel et al. (1998)é105J507 ind 125 cts-23 sal-104 mcs-23 D (DI:1t) cat Pxyl -HA-ftsL See Experimental proceduresé105J508 ind 125 cts-23 sal-104 mcs-23 D (DI:1t) cat Pxyl-ftsL-HA See Experimental procedures

PlasmidspEPI 1 pUC19 with HA epitope cassette A. Driks, Department of Microbiology and Immunology,

Loyola University, Chicago, USApMLK83 amyE integration vector neo Karow and Piggot (1995)pRD99 cat Pxyl -ftsL cassette in pUC13 Daniel et al. (1998)pRD132 bla cat Pxyl-HA-ftsL See Experimental procedurespRD133 bla cat Pxyl -ftsL-HA See Experimental procedurespRD154 amyE insertion vector NotI deletion of pMLK83pRD175 bla cat Pxyl -

0ftsL-HA See Experimental procedurespRD176 bla cat Pxyl-HA-ftsL See Experimental procedurespRD177 bla cat Pxyl -ftsL-HA See Experimental procedurespREP4 ori p15A aphA lacIq QiagenpQE30 Protein overproduction vector QiagenpQE FtsL bla lacI 6xHis-ftsL See Experimental procedures

a.`X'!`Y' indicates that strain Y was transformed with DNA from source X.

Control of cell division in B. subtilis 287

Q 2000 Blackwell Science Ltd, Molecular Microbiology, 36, 278±289

similarly cut, to give plasmids pRD176 and pRD177respectively. pRD176 and pRD177 were transformed into168, with selection for chloramphenicol resistance andscreening for an Amy2 phenotype. The resulting intermediatestrains were transformed with DNA from strain 799 tointroduce the ftsL disruption, as described above, givingstrains 814 and 813 respectively.

Finally, plasmid pRD133 was digested with XbaI andBstEII. The DNA ends were blunted and treated with ligase togive deleted plasmid pRD175. This plasmid was transformedinto 168 with selection for chloramphenicol resistance to givestrain 818. Single cross-over insertion of this plasmid at theftsL locus resulted in the addition of the C-terminal epitopetag coding sequence to ftsL (expressed from its naturalpromoter) and placed the downstream pbpB gene under thecontrol of Pxyl.

Strains with the HA-tagged ftsL derivatives at the amyElocus, a chromosomal knock-out of the ftsL gene and varioustemperature-sensitive division mutations were constructed asfollows. Each temperature-sensitive mutant strain was trans-formed with plasmid pRD177 DNA, selecting for chloram-phenicol resistance and checking for an Amy2 phenotype at308C. Genomic DNA from one such transformant was mixedwith a small amount of genomic DNA from strain 813 andtransformed into 168 with simultaneous selection for resis-tance to kanamycin and chloramphenicol in the presence ofboth xylose and IPTG, at 308C. The transformants werescreened for dependence on xylose (for ftsL expression) andIPTG (for pbpB expression) and a temperature-sensitivephenotype at 488C (having acquired the ts mutation bycongression). This generated strains 819, 820 and 821.

To make strains to test for the stability of FtsL±HA in theabsence of other functional division proteins, strains bearing tsmutations in ftsZ (ts1), divIB (tms-12) and divIC (SU347) weretransformed with chromosomal DNA from 818, selecting forchloramphenicol resistance. Chromosomal DNA from transfor-mants of each type (checked for a temperature-sensitivephenotype) was transformed into 168 with selection forchloramphenicol resistance. Transformants were screened fora temperature-sensitive phenotype (arising by congression ofthe division mutation). This yielded strains 822, 824 and 825.

Depletion of FtsL

Strains 811 and 812 were grown overnight in PAB with xylose(0.5%) and IPTG (1 mM). These cultures were diluted 10-foldin the same medium, then moved to 378C, grown for 1 h anddiluted again to an OD600 of 0.1. The exponentially growingcultures were incubated until an OD600 of 0.5 was reached.The cells were then centrifuged, washed with prewarmedxylose-free medium and resuspended to the original volumein xylose-free medium. Each culture was split into twoportions, to one of which xylose was added, and the cultureswere returned to 378C. This point was defined as T0. Sampleswere then take from this point at various times for proteinanalysis and cell morphology.

Immunoblotting

Culture samples (1 ml) were centrifuged and the cell pellet

frozen immediately in liquid nitrogen. Once sampling wascomplete, 200 ml of sample loading buffer (with proteinaseinhibitor Complete added; Roche) was added to the frozenpellet, and the sample was held on ice before sonication(one pulse of 8 mm amplitude for 15 s). The samples werethen heated to 858C for 2 min, and equal amounts ofprotein (adjusted according to the OD of the originalculture) were separated by 14% SDS±PAGE. The proteinswere then transferred to polyvinylidene difluoride (PVDF)membrane (Hybond-P; Amersham) by electroblotting forimmunodetection.

Western blots were treated with PBS, 0.1% Tween-20(PBST), 5% milk powder for 1 h at ambient temperature withshaking, after which monoclonal HA antibody (Roche) wasadded at a dilution of 1:4000, and incubation was continuedfor 1 h. The blots were then washed several times with PBSTand again blocked with PBST containing 5% milk powder for30 min. The secondary antiserum, anti-mouse-horseradishperoxidase (HRP), was added at a dilution of 1:8000, and theblot was incubated with shaking for 1 h before washing anddetection using an ECL kit (Amersham). Dilutions of 1:5000for rabbit polyclonal anti-DivIB (Harry et al., 1993) and anti-DivIC (Katis et al., 1997) sera and a 1:10 000 dilution ofrabbit polyclonal anti-FtsZ serum (A. S. Taylor, unpublished)were also used, with a 1:8000 dilution of anti-rabbit±HRPconjugate (Sigma).

Immunofluorescence microscopy

Immunolocalization was performed essentially as describedpreviously (Daniel et al., 2000). For analysis of localization inthe wild-type strain 812, samples were taken from exponen-tially growing cultures in PAB medium containing xylose andIPTG. For the temperature-sensitive strains and the control(strain 813) (Fig. 2), cultures were grown in PAB at 308Ccontaining both xylose and IPTG. Then, at an OD600 of 0.2, aportion of each culture was shifted to 488C. After 40 min,samples from the cultures at each temperature wereprocessed for immunolocalization. Cells were fixed andpermeabilized essentially as described by Harry and Wake(1997) and Katis et al. (1997), except for FtsZ, in which themicroscope slide was treated with polylysine before lysozymetreatment of the cells. After mounting on multiwell slides, thecells were treated as described by Daniel et al. (2000). Theprimary antibodies were used at the following dilutions:monoclonal mouse anti-HA 12CA5 (Roche), 1:400; polyclo-nal rabbit anti-FtsZ, 1:1000. Secondary antisera comprised1:1000 dilutions of anti-rabbit or anti-mouse (as appropriate)fluorescein isothiocyanate (FITC) conjugate (Sigma) or a1:2500 dilution of anti-rabbit Cy3 conjugate (Sigma) for co-localization of FtsL and FtsZ. All dilutions were made in PBSwith 2% BSA.

Suppression of divIB by overexpression of FtsL

A cat Pxyl -ftsL cassette was excised from pRD99 by dig-estion with EcoRI and PstI and inserted into the amyE locus

by ligation to similarly digested pRD154 at high DNA

concentration. The ligation reaction was transformed into

168 with selection for chloramphenicol resistance. Amy2

288 R. A. Daniel and J. Errington

Q 2000 Blackwell Science Ltd, Molecular Microbiology, 36, 278±289

transformants were isolated, and insertion of cat Pxyl-ftsL atthe amyE locus into one such transformant, designated 816,was confirmed by PCR. The copy of ftsL cloned into theamyE locus of strain 816 was shown to be functional bytransforming this strain with DNA from strain 803, withselection for kanamycin resistance in the presence of bothxylose and IPTG. All these transformants were found to bexylose dependent for growth, confirming that the Pxyl

promoter was expressing a functional copy of the ftsL gene.

Overnight cultures of strains 168, 816, 817 and 826, grownat 308C in PAB, were diluted 10-fold in fresh PAB andincubated for 1 h. Each culture was then used to inoculateduplicate 10 ml aliquots of PAB to give a final OD600 of 0.05.Incubation was then continued for 1 h, after which timexylose (1% final concentration) was added to one set ofaliquots, and glucose (0.25%) was added to the other set.Incubation was continued for 20 min before the cultures weremoved to 488C or 468C. Samples for microscopy were fixed inethanol and held on ice before mounting on polylysine-treated microscope slides as described by Hauser andErrington (1995).

Acknowledgements

This work was supported by grants from the Biotechnology and

Biological Sciences Research Council, the BIOTECH programme

of the EC and Prolysis Ltd. We thank Liz Harry for many helpful

discussions, and Adam Driks for the pEPI 1 plasmid.

References

Addinall, S.G., Bi, E., and Lutkenhaus, J. (1996) FtsZ ringformation in fts mutants. J Bacteriol 178: 3877±3884.

Addinall, S.G., Cao, C., and Lutkenhaus, J. (1997) FtsN, a late

recruit to the septum in Escherichia coli. Mol Microbiol 25:

303±309.Anagnostopoulos, C., and Spizizen, J. (1961) Requirements for

transformation in Bacillus subtilis. J Bacteriol 81: 741±746.

Beall, B., and Lutkenhaus, J. (1989) Nucleotide sequence andinsertional inactivation of a Bacillus subtilis gene that affects

cell division, sporulation, and temperature sensitivity. J

Bacteriol 171: 6821±6834.

Bi, E., and Lutkenhaus, J. (1991) FtsZ ring structure associatedwith division in Escherichia coli. Nature 354: 161±164.

Boyle, D.S., Khattar, M.M., Addinall, S.G., Lutkenhaus, J., and

Donachie, W.D. (1997) ftsW is an essential cell-division gene in

Escherichia coli. Mol Microbiol 24: 1263±1273.Bramhill, D. (1997) Bacterial cell division. Annu Rev Cell Dev Biol

13: 395±424.

Carson, M.J., Barondess, J., and Beckwith, J. (1991) The FtsQprotein of Escherichia coli: membrane topology, abundance,

and cell division phenotypes due to overproduction and

insertion mutations. J Bacteriol 173: 2187±2195.

Chen, J.C., Weiss, D.S., Ghigo, J.-M., and Beckwith, J. (1999)Septal localization of FtsQ, an essential cell division protein in

Escherichia coli. J Bacteriol 181: 521±530.

Dai, K., Xu, Y., and Lutkenhaus, J. (1993) Cloning and

characterisation of ftsN, an essential cell division gene inEscherichia coli isolated as a multicopy suppressor of ftsA12

(Ts). J Bacteriol 175: 3790±3797.

Daniel, R.A., Williams, A.M., and Errington, J. (1996) A complex

four-gene operon containing essential cell division gene pbpBin Bacillus subtilis. J Bacteriol 178: 2343±2350.

Daniel, R.A., Harry, E.J., Katis, V.L., Wake, R.G., and Errington,J. (1998) Characterization of the essential cell division gene

ftsL (yllD) of Bacillus subtilis and its role in the assembly of the

division apparatus. Mol Microbiol 29: 593±604.

Daniel, R.A., Harry, E.J., and Errington, J. (2000) Role of

penicillin-binding protein PBP 2B in assembly and functioningof the division machinery of Bacillus subtilis. Mol Microbiol 35:

299±311.

Erickson, H.P., Taylor, D.W., Taylor, K.A., and Bramhill, D.

(1996) Bacterial cell division protein FtsZ assembles into

protofilament sheets and minirings, structural homologues oftubulin polymers. Proc Natl Acad Sci USA 93: 519±523.

Errington, J. (1990) Gene cloning techniques. In MolecularBiological Methods for Bacillus. Harwood, C.R., and Cutting,

S.M. (eds). Chichester: Wiley, pp. 175±220.

Feucht, A., Daniel, R.A., and Errington, J. (1999) Characteriza-

tion of a morphological checkpoint coupling cell-specific

transcription to septation in Bacillus subtilis. Mol Microbiol 33:1015±1026.

Ghigo, J.-M., Weiss, D.S., Chen, J.C., Yarrow, J.C., and

Beckwith, J. (1999) Localization of FtsL to the Escherichia

coli septal ring. Mol Microbiol 31: 725±737.

Ghuysen, J.-M. (1991) Serine b-lactamases and penicillin-

binding proteins. Annu Rev Microbiol 45: 37±67.

Gottesman, S., and Maurizi, M.R. (1992) Regulation by proteo-

lysis: energy-dependent proteases and their targets. MicrobiolRev 56: 592±621.

Guzman, L.-M., Barondess, J.J., and Beckwith, J. (1992) FtsL, anessential cytoplasmic membrane protein involved in cell

division in Escherichia coli. J Bacteriol 174: 7716±7728.

Guzman, L.M., Weiss, D.S., and Beckwith, J. (1997) Domain-

swapping analysis of FtsI, FtsL, and FtsQ, bitopic membrane

proteins essential for cell division in Escherichia coli. JBacteriol 179: 5094±5103.

Hale, C.A., and De Boer, P.A.J. (1997) Direct binding of FtsZ to

ZipA, an essential component of the septal ring structure that

mediates cell division in E. coli. Cell 88: 175±185.

Hale, C.A., and De Boer, P.A.J. (1999) Recruitment of ZipA to the

septal ring of Escherichia coli is dependent on FtsZ and

independent of FtsA. J Bacteriol 181: 167±176.

Harlow, E., and Lane, D.P. (1988) Antibodies. A LaboratoryManual. Cold Spring Harbor, NY: Cold Spring Harbor

Laboratory Press.

Harry, E.J., and Wake, R.G. (1997) The membrane-bound cell

division protein DivIB is localized to the division site in Bacillus

subtilis. Mol Microbiol 25: 275±283.

Harry, E.J., Stewart, B.J., and Wake, R.G. (1993) Characteriza-

tion of mutations in divIB of Bacillus subtilis and cellularlocalization of the DivIB protein. Mol Microbiol 7: 611±621.

Hauser, P.M., and Errington, J. (1995) Characterization of cellcycle events during the onset of sporulation in Bacillus subtilis.

J Bacteriol 177: 3923±3931.

Huisman, O., and D'Ari, R. (1981) An inducible DNA replication-

cell division coupling mechanism in Escherichia coli. Nature

290: 797±799.

Ikeda, M., Sato, T., Wachi, M., Jung, H.K., Ishino, F., Kobayashi,Y., et al. (1989) Structural similarity among Escherichia coli

FtsW and RodA proteins and Bacillus subtilis SpoVE protein,

which function in cell division, cell elongation, and spore

formation, respectively. J Bacteriol 171: 6375±6378.

Jenkinson, H.F. (1983) Altered arrangement of proteins in the

Control of cell division in B. subtilis 289

Q 2000 Blackwell Science Ltd, Molecular Microbiology, 36, 278±289

spore coat of a germination mutant of Bacillus subtilis. J GenMicrobiol 129: 1945±1958.

Karow, M.L., and Piggot, P. (1995) Construction of gusA

transcriptional fusion vectors for Bacillus subtilis and their

utilization for studies of spore formation. Gene 163: 69±74.Katis, V.L., Harry, E.J., and Wake, R.G. (1997) The Bacillus

subtilis division protein DivIC is a highly abundant membrane-

bound protein that localizes to the division site. Mol Microbiol26: 1047±1055.

Khattar, M.M., Addinall, S.G., Stedul, K.H., Boyle, D.S.,

Lutkenhaus, J., and Donachie, W.D. (1997) Two polypeptide

products of the Escherichia coli cell division gene ftsW and apossible role for FtsW in FtsZ function. J Bacteriol 179: 784±

793.

Levin, P.A., and Losick, R. (1994) Characterization of a cell

division gene from Bacillus subtilis that is required forvegetative and sporulation septum formation. J Bacteriol 176:

1451±1459.

Liu, Z., Mukherjee, A., and Lutkenhaus, J. (1999) Recruitment of

ZipA to the division site by interaction with FtsZ. Mol Microbiol31: 1853±1861.

LoÈwe, J., and Amos, L.A. (1998) Crystal structure of the bacterial

cell-division protein FtsZ. Nature 391: 203±206.Lupas, A. (1996) Coiled coils: new structures and new functions.

Trends Biochem Sci 21: 375±382.

Lupas, A., Van Dyke, M., and Stock, J. (1991) Predicting coiled

coils from protein sequences. Science 252: 1162±1164.Lutkenhaus, J., and Addinall, S.G. (1997) Bacterial cell division

and the Z ring. Annu Rev Biochem 66: 93±116.

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-celldivision site. Genes Dev 12: 3419±3430.

Mukherjee, A., Cao, C., and Lutkenhaus, J. (1998) Inhibition ofFtsZ polymerization by SulA, an inhibitor of septation in

Escherichia coli. Proc Natl Acad Sci USA 95: 2885±2890.

Paulton, R.J.L. (1971) Nuclear and cell division in filamentous

bacteria. Nature New Biol 231: 271±274.

Pogliano, J., Pogliano, K., Weiss, D.S., Losick, R., and Beckwith,

J. (1997) Inactivation of FtsI inhibits constriction of the FtsZ

cytokinetic ring and delays the assembly of FtsZ rings at

potential division sites. Proc Natl Acad Sci USA 94: 559±564.

Roberts, J.M. (1999) Evolving ideas about cyclins. Cell 98: 129±

132.

Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989) MolecularCloning. A Laboratory Manual. Cold Spring Harbor, NY: Cold

Spring Harbor Laboratory Press.

Thornewell, S.J., East, A.K., and Errington, J. (1993) An efficient

expression and secretion system based on Bacillus subtilis

phage f105 and its use for the production of B. cereus b-

lactamase I. Gene 133: 47±53.

Ueki, M., Wachi, M., Jung, H.K., Ishino, F., and Matsuhashi, M.

(1992) Escherichia coli mraR gene involved in cell growth anddivision. J Bacteriol 174: 7841±7843.

Vischer, N.O.E., Huls, P.G., and Woldringh, C.L. (1994) Object-Image: an interactive image analysis program using structured

point collection. Binary 6: 160±166.

Wang, X., and Lutkenhaus, J. (1993) The FtsZ protein of Bacillus

subtilis is localized at the division site and has GTPase activity

that is dependent upon FtsZ concentration. Mol Microbiol 9:

435±442.