Cell, Vol. 88, 577–579, March 7, 1997, Copyright 1997 by Cell Press
Bacterial Chromosome Segregation:
Is There a Mitotic Apparatus?
Robert T. Wheeler and Lucy Shapiro
Department of Developmental Biology
Beckman Center
Stanford University School of Medicine
Stanford, California 94305
A deceptively simple, yet long standing, puzzle in bacterial cell division is how newly replicated chromosomes
are equally distributed between two progeny cells. For
many years, beginning with the seminal Jacob, Brenner,
and Cuzin paper (1963), microbial geneticists have reasoned that the cell membrane is likely to play a role in
this process by providing a relatively stable support
structure for the replicating chromosomes. Pieces of
this puzzle have indeed implicated the membrane in the
regulation of DNA replication initiation, but a membraneassociated mitotic apparatus that is able to position
chromosomes three-dimensionally and implement directed chromosome movement has yet to be described.
The critical questions that must be answered include:
Are chromosomes spatially oriented with respect to the
cell pole and the division site? Are they pulled towards
the cell poles by an active mechanism, or divided as an
inescapable consequence of cell growth and septation?
These lead to the larger question: Will years of study
of bacterial ‘mitosis’ uncover a bacterial cytoskeleton
and its accompanying molecular motors as is the case
for eukaryotic mitosis? Visualization of subcellular protein location in bacterial cells is now providing the means
to approach these questions directly. Two papers in this
issue of Cell address the first question by providing
direct evidence that the newly replicated chromosomes
are aligned with their origins oriented toward the poles
of the predivisional cell (Webb et al., 1997; Mohl and
Gober, 1997; Figure 1).
The search for a general mechanism for plasmid and
chromosome segregation has been heavily influenced
by the replicon theory of Jacob et al. (1963), which was
an extension of a model used to explain F plasmid replication and transmission in Escherichia coli. This theory
Minireview
proposes that membrane attachment of the plasmid or
chromosome during replication and cell division, coupled with directed growth at the midcell, could explain
DNA equipartitioning to progeny cells. Since that time,
it has become clear that growth of membrane and cell
wall occurs heterogeneously over the entire cell surface,
which is inconsistent with this model as first presented
(Woldringh et al., 1987). However, the analysis of plasmid partitioning has made a significant contribution to
the identification of factors that influence chromosome
segregation. Notably, chromosomally-encoded homologs to the parA and parB families of plasmid-encoded
genes required for active plasmid partitioning are located near the origin of replication in E. coli, Pseudomonas putida, Caulobacter crescentus, and Bacillus subtilis (Ogasawara and Yoshokawa, 1992; Mohl and Gober,
1997). The ParA protein from phage P1 provides a useful
model for the ParA family of plasmid and chromosomally
encoded proteins: phage P1 ParA is a transcriptional
repressor with ATPase activity which is stimulated in
vitro by the ParB protein (Davis et al., 1992). Phage P1
ParB similarly serves as a model for the family of ParB
proteins: it is a DNA-binding protein which binds to parS,
a centromere-like site within the parABS operon (Funnell
and Gagnier, 1994). For several plasmid and phage systems, both ParA and ParB homologs are required in
trans for active plasmid partitioning, leading to the proposal that ParA and ParB cooperate with host proteins
to bind parS and align newly-replicated plasmids, ensuring their equipartition during division (Firshein and Kim,
1997; Williams and Thomas, 1992). Significantly, genetic
characterization of these genes in B. subtilis and C.
crescentus also implicates the par genes in chromosome segregation (Ireton et al., 1994; Mohl and Gober,
1997).
E. coli and B. subtilis segregate progeny chromosomes gradually throughout the cell cycle. However, if
protein synthesis is inhibited under appropriate conditions and then released from this inhibition, the two
chromosomes can be seen to segregate rapidly (Wake
and Errington, 1995). This provides additional evidence
Figure 1. A Model of Bacterial Chromosome
Segregation
Cell
578
that chromosome segregation is not solely achieved
through chromosome-membrane attachment coupled
to membrane and cell wall synthesis, but might also
require dedicated machinery to direct chromosome
movement to the progeny cells.
The identity of such a driving force or tension responsible for bringing bacterial chromosomes steadily (or
rapidly) apart remains a central and unanswered question. This question was initially approached by genetic
screens for conditional segregation mutants in E. coli.
The identification of these mutants has led to the elegant
characterization of the complex topological problems
associated with decatenation and dimer resolution during the separation of newly replicated chromosomes
(Adams et al., 1992; Drlica, 1992). Despite this success,
initial genetic screens for chromosome segregation mutants yielded neither molecules linking the chromosome
to the cell membrane during segregation nor those that
could provide the force required to actively separate
the two progeny chromosomes. However, a ripple of
excitement was generated when a screen for E. coli
mutants defective in chromosome but not plasmid partitioning revealed the involvement of MukB, a large ATPbinding protein, in chromosome segregation. MukB,
with its globular head and coiled-coil tail, was proposed
to be a novel bacterial molecular motor (Hiraga et al.,
1991). Hiraga suggested that MukB binds to the chromosome, and may contribute to chromosome separation. It could be imagined that MukB binds to the chromosome and a membrane complex at the cell pole,
providing the tension to pull apart decatenated chromosomes from midcell to their polar destinations in the
predivisional cell. Although the role of MukB in chromosome segregation remains unclear, the results of Webb
et al. (1997) and Mohl and Gober (1997) provide us with
the insight that a polar membrane–chromosome complex may be a component of a segregation apparatus.
Webb et al. (1997) provide a live, graphic view of chromosomal origin movement in B. subtilis by following the
cellular position of the tagged chromosomal origin of
replication in dividing cells. They exploited a technique
developed in the Murray and Belmont laboratories (Robinett et al., 1997; Straight et al., 1997) in which a tandem
array of lacO sequences is integrated at a chromosomal
location and multiple Green Fluorescent Protein-LacI
(GFP-LacI) fusion proteins that bind to the array are
visualized with fluorescence microscopy. Webb et al.
(1997) integrated the array at either the terminus or origin
regions of the B. subtilis chromosome and visualized
the array by producing the GFP-LacI fusion in the cells.
They found that in dividing cells the chromosomal origin
preferentially localizes near the cell poles whereas the
terminus region is preferentially at midcell.
Complementing the work done in B. subtilis, Mohl
and Gober (1997) provide independent evidence for the
anchoring of chromosomal origin regions to the cell
poles. They isolated the C. crescentus ParA and ParB
homologs and demonstrated that both were essential
for viability. Further, they find that the ParB protein binds
parS, a DNA sequence within the parAB operon, which
is located only 80 kB from the replication origin. Overexpression of C. crescentus ParA or ParB caused aberrant
chromosome partitioning, providing evidence that these
proteins, like their plasmid homologs, are involved in
segregation of newly replicated DNA.
Their most remarkable results, however, are that the
C. crescentus ParA and ParB proteins were colocalized
to the cell poles in late predivisional cells. Given that
other ParA and ParB homologs are membrane associated (Firshein and Kim, 1997), and that the C. crescentus
ParB homolog binds a specific region of chromosomal
DNA near the origin, the colocalization of these proteins
to the poles of post-replication predivisional cells indicates that they are plausible components in a complex
that tethers the chromosomal origin to the cell pole.
Because the polar immunolocalization was only seen
late in the cell cycle, their data also suggests that ParA/
ParB-mediated chromosomal anchoring is regulated by
cell cycle cues. Of course, a notable caveat to these
conclusions is that it has not been shown that the chromosomal site bound by ParB is required for segregation.
Nevertheless, these results provide a valuable link between plasmid partitioning and chromosomal segregation that merits further study. Thus, complementary lines
of evidence from two different bacteria converge in indicating the existence of a mitotic-like apparatus for positioning the replication origin regions near opposite ends
of the cell and pointing to the possible involvement of
Par proteins in this process.
Taken together with previous work on E. coli chromosome and plasmid segregation, these papers suggest
a working model for the mechanics and molecular players in segregation (Figure 1). The origin of replication
(O) is likely to be membrane bound early in the cell cycle,
contributing to the control of DNA replication initiation
(Lu et al., 1994; Herrick et al., 1994). The site(s) of originmembrane attachment at this time in the cell cycle are
not known. Later, the newly replicated origin regions
migrate to bipolar locations. This parallels the localization of ParA and ParB homologs to one, then both, poles.
The mechanics of this chromosome alignment and polar
attachment are unknown. Decatenation follows polar
alignment and plays an essential role in the act of chromosome separation. Although the last region of the
chromosome to be replicated, the termination region
(T), is situated at midcell, it is not known if mid-cell
septum formation plays an active role in the segregation
process.
The analysis of bacteria that undergo asymmetric cell
division has provided a window into the dynamics of
chromosome segregation which has proven difficult to
obtain from bacteria that carry out symmetric divisions.
B. subtilis not only undergoes binary fission, but like C.
crescentus it is also capable of undergoing asymmetric
division at the onset of sporulation. This asymmetric
division in B. subtilis generates a small, forespore cell
and a large, mother cell. A pivotal observation on the
mechanisms of chromosome segregation has come
from analysis of the initial stages of sporulation in B.
subtilis. The isolation of mutants with aberrant nucleoid
segregation led to the identification of SpoIIIE, a putative
membrane DNA translocase (Wake and Errington, 1995).
This protein is responsible for threading the terminal
two-thirds of the progeny chromosome destined for the
forespore through the nearly complete septum into the
Minireview
579
small forespore compartment during early sporogenesis. This work established SpoIIIE as a genuine segregation molecule, playing a role in moving the B. subtilis
chromosome during the division of the forespore and
mother cell. Although SpoIIIE plays an essential role in
sporulation, it also plays a role during vegetative growth,
required when DNA replication is inhibited and septation
occurs before chromosome segregation has been completed.
Experiments with a spoIIIE mutant led to the surprising
conclusion that the same one-third of the chromosome
(the first third to be replicated) was trapped in the forespore in each mutant cell (Wake and Errington, 1995).
This result suggested that the B. subtilis chromosomes
are oriented within the cell, with the origins closest to the
cell poles. It also raised questions: What does SpoIIIE
recognize to allow the translocation of the terminal twothirds of the forespore chromosome to the forespore
compartment? What mechanism places the origin in
the forespore compartment prior to septation? Are the
newly replicated vegetative chromosomes also oriented
with their origins facing the cell poles?
The experiments described by Webb et al. (1997) in
this issue directly demonstrate that the origin region of
the forespore chromosome is oriented toward the cell
pole. They found that prior to asymmetric division the
two chromosomes have their origin regions located near
opposite ends of the cell and their termini near the midcell position. The observed preferential orientation of
the chromosomes confirms the predictions based on
the segregation defects in the spoIIIE mutant. Moreover,
this establishes that polar anchoring of the replication
origin region is a general feature of symmetric and asymmetric division.
To obtain clues about the molecular identity of the
players in chromosome alignment in B. subtilis, mutants
of genes encoding Soj, the ParA homolog, and SpoOJ,
the ParB homolog, were examined using the GFP-LacI
constructions. Mutants of spoOJ alone are blocked at
entry into sporulation, but a soj spoOJ double mutant
is capable of sporulating with high efficiency (Wake and
Errington, 1995; Sharpe and Errington, 1996). Surprisingly, Webb et al. (1997) observed soj-spoOJ double
mutant sporangia with more than two chromosomes.
The chromosomes in sporulating cells of soj-spoOJ double mutants appeared to be randomly oriented, seemingly failing to align with the origins facing the poles
and the termini at midcell. Their experiments do not,
however, clarify whether misalignment was a direct result of Soj and SpoOJ absence or was solely due to the
presence of multiple chromosomes.
The work of Webb et al. (1997) and Mohl and Gober
(1997) emphasizes the importance of the ParA and ParB
families of partitioning proteins in chromosome segregation. Further, the impact of their visualization of specific regions of the chromosome is particularly significant when viewed in the light of the long and tortuous
search for molecular factors involved in bacterial chromosome segregation. The task now is to identify the
molecules that interact with the Par proteins and with
the chromosome itself to bring about the alignment of
the chromosomes and their movement to the two poles
of the predivisional cell.
Selected Reading
Adams, D.E., Shekhtman, E.M., Zechiedrich, E.L., Schmid, M.B., and
Cozzarelli, N.R. (1992). Cell 71, 277–288.
Davis, T.L., Martin, K.A., and Austin, S.J. (1992). Mol. Microbiol. 6,
1141–1147.
Drlica, K. (1992). Mol. Microbiol. 6, 425–433.
Firshein, W., and Kim, P. (1997). Mol. Microbiol. 23, 1–10.
Funnell, B.E., and Gagnier, L. (1994). Biochimie 76, 924–932.
Herrick, J., Kern, R., Guha, S., Landoulsi, A., Fayet, O., Malki, A.,
and Kohiyama, M. (1994). EMBO J. 13, 4695–4703.
Hiraga, S., Niki, H., Imamura, R., Ogura, T., Yamanaka, K., Feng, J.,
Ezaki, B., and Jaffe, A. (1991). Res. Microbiol. 142, 189–194.
Ireton, K., Gunther, N.W., and Grossman, A.D. (1994). J. Bacteriol.
176, 5320–5329.
Jacob, F., Brenner, S., and Cuzin, F. (1963). Cold Spring Harbor
Symp. Quant. Biol. 28, 329–348.
Lu, M., Campbell, J.L., Boye, E., and Kleckner, N. (1994). Cell 77,
413–426.
Mohl, D.A., and Gober, J.W. (1997). Cell 88, this issue.
Ogasawara, N., and Yoshokawa, H. (1992). Mol. Microbiol. 6,
629–634.
Robinett, C.C., Straight, A.F., Li, G., Willhelm, C., Sudlow, G., Murray,
A., and Belmont, A.S. (1996). J. Cell Biol. 135, 1685–1700.
Sharpe, M.E., and Errington, J. (1996). Mol. Microbiol. 21, 501–509.
Straight, A.F., Belmont, A.S., Robinett, C.C., and Murray, A.W. (1996).
Curr. Biol. 6, 1599–1608.
Wake, R.G., and Errington, J. (1995). Annu. Rev. Genetics 29, 41–67.
Webb, C.D., Teleman, A., Gordon, S., Straight, A., Belmont, A., Lin,
D.C., Grossman, A.D., Wright, A., and Losick, R. (1997). Cell 88, this
issue.
Williams, D.R., and Thomas, C.M. (1992). J. Gen. Microbiol. 138,
1–16.
Woldringh, C.L., Huls, P., Pas, E., Brakenholf, G.L., and Nanninga,
N. (1987). J. Gen. Microbiol. 133, 575–586.
Note Added in Proof
It has recently been reported that the B. subtilis SpoOJ protein is
associated with the origin-proximal third of the chromosome and
is positioned toward the cell pole prior to the initiation of cell division
(Lin, D.C.-H., Levin, P.A., and Grossman, A.D. [1997]. Proc. Natl.
Acad. Sci. USA, in press).