JOURNAL OF BACrOLOGY, July 1974, p. 303-324 Vol. 119, No.

1
Copyright) 1974 American Society for Microbiology Printed in U.S.A.

Septum Formation in Escherichia coli: Characterization of

Septal Structure and the Effects of Antibiotics on Cell Division
I. D. J. BURDETT' AND R. G. E. MURRAY
Department of Bacteriology and Immunology, University of Western Ontario, London 72, Ontario, Canada
Received for publication 21 March 1974

Septa can be demonstrated in sections of Escherichia coli strains B and B/r

after fixation with acrolein and glutaraldehyde. The septum consists of an
ingrowth of the cytoplasmic membrane and the mucopeptide layer; the outer
membrane is excluded from the septum until the cells begin to separate.
Mesosomes have also been observed. The septum is highly labile and, except in
the chain-forming strains, E. coli D22 env A and CRT 97, not easily preserved by
standard procedures. The labile nature of the septum may be due to the presence
of autolysin(s) located at the presumptive division site. Blocking division by addi-
tion of ampicillin (2 to 5 Ag/ml) to cells of E. coli B/r produces a bulge at the
middle of the cells; bulge formation is stopped by addition of chloramphenicol.
Cephalosporins also induce bulge formation but may stop cell elongation as well
as division. Bulge formation, due to the presumed action of an autolysin(s), may
be an initial step in the septation sequence when the mucopeptide is modified to
allow construction of the septum. In a nonseptate filament-forming strain, PAT
84, which ceases to divide at 42 C, bulge formation only occurs in the presence
of ampicillin at the time of a shift-down at 30 C or at 42 C in the presence of NaCl
(0.25 to 0.34 M). Experiments with chloramphenicol suggest that the filaments
are fully compartmentalized but fail to divide owing to the inactivation, rather
than loss of synthesis, of an autolysin at 42 C.

Attempts have been made to follow the
growth pattern of individual cells of Escherichia
coli (for reviews, see 7, 17), but it has not been
possible to differentiate the principal stages.
This is due, in part, to the apparent absence of
distinct septa in many strains under the usual
conditions of fixation (49). Instead, the dividing
cell merely seems to constrict at the middle,
and all wall layers, including the cytoplasmic
membrane, infold together (49). Only in certain
chain-forming strains (30, 34), some thermosen-
sitive mutants (22), and in E. coli B held at 45 C
(49) has it been possible to demonstrate septa
routinely.
Although a relationship exists between chro-
mosome replication and cell division in E. coli
(5-7, 15, 20, 53), very little is known about the
conditions which initiate septation, i.e., entry of
the cells into the D period (5). Genetic charac-
terization of conditional thermosensitive mu-
tants (39) indicate that at least seven genes (fts
A-fts G) are involved in septation. Complex
though the analysis of the septation process is
likely to be, it is important to stress that
I
Present address: National Institute for Medical Research, calized hydrolysis of bonds at specific areas of
The Ridgeway, London NW7 1AA, England.
303
septation must essentially involve some modifi-
cation of both the cell wall and the cytoplasmic
membrane. This fact has been amply demon-
strated in studies of Streptococcus faecalis, for
example (17). In this organism it has been
possible to show, by electron microscopy, that
wall growth occurs at a specific region of the cell
and that the development of the septum follows
a definite course. As in bacilli (8), the nascent
cross wall is closely associated with the cyto-
plasmic membrane and with mesosomes.
Growth of the cell wall is likely to be modu-
lated by autolysins (in the case of S. faecalis, an
N-acetylmuramidase) whose primary target of
attack is the mucopeptide (peptidoglycan, mu-
rein) moiety of the cell wall. The mucopeptide
itself consists of covalently bonded polysaccha-
ride strands cross-linked by peptide side chains
(see ref. 3). Cumulative evidence from studies of
bacilli (9, 10, 11, 17, 41) and of S. faecalis (17)
suggest that autolysins may fulfill a number of
specific roles in wall growth, such as: (i) cleav-
age of bonds in mucopeptide to allow for inser-
tion of newly synthesized constituents; (ii) lo-
the wall to allow for alteration in cell shape
304 BURDETT AND MURRAY J. BACTERIOL.
accompanying specific events of the cell cycle, (6.0); KH2PO4 (3.0); NaCl (3.0); MgSO4 (0.25) and
such as septum formation; and (iii) a role in cell glucose (1.0). Stock solutions of MgSO4 and glucose
separation. Autolysins may therefore be in- were autoclaved separately. In some experiments this
volved in the initiation of wall synthesis as well medium was supplemented with Casamino Acids and
as localized remodelling of the wall. In S. tryptophan to final concentrations of 0.2%7 and 20
faecalis, autolysis proceeds from the newly syn- ,g/ml, respectively. (iii) LB medium (ref. 2): tryp-
tone (Difco) (10.0); yeast extract (5.0); NaCl (10.0);
thesized tip of the growing cross wall, whereas glucose (1.0). (iv) Complete medium (ref'. 22): Beef
the peripheral wall is less susceptible to auto- extract (10.0); bactotryptone (10.0) and NaCl (10.0).
lytic attack (16). In the case of E. coli PAT 84, the NaCl was omitted
In E. coli, the evidence of autolysin action is from the medium or the level was reduced to 5.0
less direct, but it has been shown that two g/liter (see below). In all experiments, complete
growth processes, septation and elongation, can medium was supplemented with thymine (20 gg/ml,
be differentiated by the response of the cells to final concentration).
varing doses of penicillin (46). Low doses of Growth of the cultures was monitored by following
the absorbance at 660 nm (A6,,) in a Beckman DU
penicillin (10 to 50 U/ml) block division and spectrophotometer, using cells of 1-cm path length.
give rise to a bulge located at the middle of the Cell numbers were estimated with a model B Coulter
cell at presumptive cross wall sites. Elongation counter (Coulter Electronics, Hialeah, Fla.) using an
of the cells can continue under these conditions, allerture of 30 ,um. Samples (1.0 ml) were removed at
but ceases on the addition of high concentra- intervals from the culture and placed into an equal
tions (100 U/ml or more) of penicillin which volume of 0.5%c formalin (in 0.9%, NaCl) before count-
totally inhibit mucopeptide synthesis. The for- ing or simply diluted into 0.9% NaCl and counted
mation of a bulge is thought to be due to the directly. All fixations were performed with exponen-
continued action of hydrolytic enzymes (autoly- tially growing populations having an A,,,, of 0.2 to
sins) acting on mucopeptide after inhibition of 0.35.
Electron microscopy. (i) Fixation: (a) The stan-
the transpeptidation reaction (46) with penicil- dard OSO4 fixation schedule of Ryter and Kellen-
lin. berger (RK; ref. 45) was used. Samples (0.9 ml) were
The reasons for the common occurrence of removed from the culture and pipetted into the
"constrictive" modes of division is at present fixative (1.0 ml), centrifuged at 5,000 x g for 5 min,
obscure, and one object of this paper is to enrobed in 2% agar, and left for about 16 h in the 1%
re-examine the nature of this phenomenon. The OS04 solution. After being washed with 0.5% uranyl
results indicate that septa can be preserved in acetate dissolved in RK buffer, the pellets were de-
some strains of E. coli, such as B and B/r, and hydrated in ethanol or acetone and embedded in Epon
that the labile nature of the septum may be due or (b) Vestopal by standard procedures.
Most samples were fixed with acrolein and
to the presence of autolysin(s) located at the site glutaraldehyde by adding 1/10 volume of fixative,
of the cross wall. Studies of a thermosensitive either directly to the culture flask or by removing a
mutant, PAT 84, suggest that the autolysin(s) sample of cells. After centrifugation at 5,000 x g for 5
may be involved in modifying the mucopeptide min, the cells were resuspended in 1 to 2 ml of the
layer at an early stage of septation. stock fixative solution containing acrolein (5%), glu-
(This paper was presented, in part, at the taraldehyde (0.25%), and 0.05 M sodium cacodylate
72nd Annual Meeting of the American Society buffer, pH 7.5, for 3 to 6 h at room temperature.(I. D.
for Microbiology, Philadelphia, 1972.) J. Burdett and R. G. E. Murray, submitted for
publication). The pellets, previously enrobed in agar
and washed overnight in buffer, were postfixed with
MATERIALS AND METHODS 1% OSO4 in 0.05 M cacodylate buffer, pH 7.5, for 1 h at
room temperature. Occasionally, some samples were
Bacterial strains and growth conditions. The postfixed with 1%7c OS04 by the RK technique (see
strains of E. coli examined in the present study are above).
shown in Table 1. The thermosensitive mutants CRT Glutaraldehyde and acrolein were obtained as elec-
97, CRT 257, and PAT 84 were grown at 30 C in a tron microscopy grade products, sealed under N2,
reciprocating shaker-water bath; the restrictive tem- from Polysciences Inc. (Warrington, Pa.).
perature chosen was 42 C. All other strains were (c) Our most successful preparations were ob-
grown in water baths at 37 C. Flasks five times the tained by using acrolein and glutaraldehyde (b,
volume of the culture were used throughout. above), but a wide range of fixatives (singly or in com-
The following media were made in glass distilled bination) have also been used. These include acetal-
water, and the concentrations of the ingredients are dehyde, crotonaldehyde, picric acid, and tannic acid,
shown in grams per liter: (i) M9 medium (ref. 29): in concentrations 0.1 to 5%Q and in the presence of
NH4Cl (1.0); Na2HPO4 (7.0); KH2PO4 (3.0) supple- acrolein or glutaraldehyde or both. These fixatives
mented with MgSO4 7H2O and glucose to final con- have also included varying species and concentrations
centrations of 5 x 10-4 M and 0.1%, respectively. (ii) of ions (Li, Zn, Pb, Mg, Mn, Ca, La, and Ni) and in-
Minimal medium (ref. 15): NH4Cl (2.0); Na2HPO4 hibitors (azide, cyanide). Pretreatment of the cells at
VOL. 119, 1974 SEPTA IN E. COLI
0 C with 0.01 M tris(hydroxymethyl)aminomethane
buffer, pH 8.0-9.0, containing ethylenediaminetetra-
acetic acid at a concentration of 10-1 to 10' M be-
fore fixation, was also used.
All sections were cut with a diamond knife on a
Reichert ultramicrotome, picked up on 200-mesh
copper grids covered with Formvar/carbon, and
stained with 1% uranyl magnesium acetate (or uranyl
acetate) for 1 to 2 min, and then with lead citrate for
2 min.
(ii) Freeze-etching: Samples were concentrated by
centrifugation (5,000 x g for 5 min), and portions of
the pellet were deposited on 3-mm copper disks and
immediately frozen in Freon 22. When a cryoprotec-
tive agent (20% glycerol) was used, the bacterial
sample was washed by one cycle of centrifugation and
then frozen in Freon. Preparations were stored in
liquid nitrogen (for no longer than 1 week) prior to
freeze-etching. The method used was essentially that
of Moore (25), using a Balzer's apparatus (model
BA510 M, Balzers AG, Lichtenstein). Samples were
etched for 1.5 min after fracturing and then shadowed
with platinum and carbon. The replicas were washed,
successively, with distilled water and undiluted
H2SO4 (1 h) and distilled water. Following a further
rinse in Javex (1 h) and distilled water, the replicas
were mounted on 200-mesh copper grids bearing a
Formvar film. The micrographs of freeze-etched ma-
terial have not been reversed; i.e., shadows are white.
Micrographs were taken on 35 mm fine-grain posi-
tive film by using a Philips EM200 operated at 60 kV.
Inhibitors and antibiotics. Chloramphenicol
(CAM) was obtained from Sigma Chemical Co. (St.
Louis, Mo.) or from Calbiochem (San Diego, Calif.).
Nalidixic acid was obtained as a gift from Winthrop
Labs. (Aurora, Ont.) and prepared as a stock solution
(1 mg/ml) in 0.1 N NaOH. Ampicillin (Ayerst Labs.,
Montreal, Que.) and sodium cephalothin ("Keflin")
and cephaloridine ("Loridine") from Eli Lilly and Co.
(Canada), Toronto, Ont. were received as gifts and
used from stocks (1 mg/ml in water). All inhibitors
and antibiotics were freshly prepared or refrigerated
overnight before use.
RESULTS
Fine structure of septa. When fixed by the
RK technique, dividing cells of E. coli B, B/r,
and CRT 257 all showed constrictive divisions
(Fig. 1A, B). The invaginated area contained
the cytoplasmic membrane, the mucopeptide
layer, and also the outer membrane (for identi-
fication of layers, see ref. 13, 26, 27, 32); these
layers were usually closely apposed, even at the
apex of the V-shaped constriction. This appar-
ently simultaneous invagination of the wall and
membrane at the site of division will be referred
to as "constriction" in the subsequent discus-
sion. The appearance of the constricted cell
(Fig. 1A, B) suggested that the infolded wall
and membrane did not merely invaginate from
the mid-point of the cell but also from an area
305
TABLE 1. Strains of E. coli examined
UWOa
Strain collection Medium Source and reference
no.
B 301 M9 Laboratory collection
B/r 828 MM C. E. Helmstetter (15)
D21 env A 839 LB H. Boman (30)
D22 env A 840 LB H. Boman (30)
CRT 97 825 CM F. Jacob (22)
CRT 257 826 CM F. Jacob (22)
PAT 84 942 CM F. Jacob (18)
a UWO, University of Western Ontario.
(about 0.05 to 0.1 ,um wide) each side of the
middle, forming a deep trough. Constrictive
divisions were also seen after prefixation with
glutaraldehyde (Fig. 1C) and indeed with a
wide range of other treatments. These treat-
ments included prefixation with acetaldehyde,
crotonaldehyde, picric acid, tannic acid, and
combinations of these fixatives. Various ions
(e.g., Mg, Ca, Zn, or Ni, in varying concentra-
tions) or pretreatment of the cells with
ethylenediaminetetraacetic acid had little mod-
ifying effect on the appearance of constrictive
division. Even freeze-etching of unfixed cells
only revealed constrictive division.
The only fixative providing consistent dem-
onstration of septa in the common strains of E.
coli was the acrolein/glutaraldehyde mixture
described in Materials and Methods. It had the
further advantage in that cell form and internal
structures were preserved with persuasive regu-
larity. Dividing cells of E. coli B and B/r, when
fixed in this manner (see Fig. 2A-D), contained
a distinct septum composed of the cytoplasmic
membrane and the mucopeptide layer. In all
cases, the outer membrane did not appear to
take part in the initial stages of septation.
Vesicles and sheets of membrane, presumably
formed by blebbing of the outer membrane,
were observed at the site of septation (Fig. 2A,
B). E. coli B could also be fixed successfully
solely with 5% acrolein (in 0.05 M cacodylate
buffer, pH 7.5), but the outer membrane of E.
coli B/r appeared to be damaged by this treat-
ment, forming many small vesicles along the
cell. Mesosome-like membranes were also occa-
sionally seen (Fig. 3B) in E. coli B and B/r,
particularly if the basal glucose medium was
supplemented with Casamino Acids and trypto-
phan. Under the latter conditions, some cells
also contained a septum apparently composed
solely of the cytoplasmic membrane (Fig. 3D,
4A); asymmetric divisions, involving the in-
growth of the septum from one .side of the cell,
were also noted. E. coli B completed septa
 .1-

..4
%

1.

.-
.V,
A.

w .s
3*

.44
15. !'to. .

A F *C ;,,@ s

mp

B .1-I
1
.-
.

if
:-4-,A&4mjm-.
... ..- t.
..

1-

-41"W.-

b~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
tW b AA

FIG. 1. Examples of constrictive division in sections of E. coli Bir (A), E. coli CRT 257 (B), and E. coli B (C);
both A and B were fixed by RK technique; C was fixed with glutaraldehyde-osmium. Note deep cleft formed by
invaginated layers of closely apposed wall and membrane layers; cytoplasmic membrane (cm), outer membrane
(om), and mucopeptide (mp). Unless otherwise stated, magnification bar equals 0.1 ,um.
FIG. 2. Septa in sections of E. coli B in M9 medium (A, C) and B/r in glucose minimal medium (B) or
glucose-Casamino Acids (D); fixed with acrolein-glutaraldehyde. Abbreviations as in Fig. 1. Unlabeled arrows
show mucopeptide portion of septum; note blebs or sheets of outer membrane (A, B) and double lamellae of
mucopeptide in septum.
306
 .A

A E; 5^ A

X.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~JF
c.
IIF:~~ _ : tom

4~~~~~4W.

B~~~~~~

B 4CF. - . t ,.

C
D46. -
;' a X

FIG. 2
307
 1"

U'

FIG. 3
308
VOL. 119, 1974 SEPTA IN E. COLI
before any distinct ingrowth of the outer mem-
brane rounded the profiles of the newly formed
poles of the cells thus formed. However, E. coli
B/r (Fig. 3A, 3D) cells were seen in which the
rounding of the poles accompanied, or closely
followed, the ingrowth of the septum, but the
mucopeptide portion of the septum still pre-
ceded the invagination of the outer membrane.
Similar results were obtained from examination
of sections of E. coli D21 env A, the parent
strain of the chain-forming mutant D22 env A
discussed below.
The initial stages of septation were the most
difficult to preserve. Also, relatively few cells
could be detected in batch cultures which
showed this stage. The basic structure of the
septum (e.g., Fig. 2B, 2C) showed that it was
composed of an invaginated portion of the
cytoplasmic membrane containing two distinct,
but separate, lamellae of mucopeptide. Study of
synchronous cultures of E. coli B and B/r
(Burdett and Murray, submitted for publica-
tion) has fully confirmed this picture.
Our most successful preparations have been
of cells prefixed at the growth temperature with
a one-tenth volume of fixative for 1 to 3 min
and then resuspended at room temperature (23
to 25 C) in full-strength acrolein/glutaralde-
hyde. During the prefixation step, the fixative
was simply added to the flasks, allowing the
culture to shake meanwhile. If the pH of the
fixative was less than 7.0, septa were seen
infrequently, and the cytoplasm was aggregated
into clumps of amorphous material. Attempts
to confirm the presence of septa in E. coli B or
B/r, previously fixed with acrolein/glutaralde-
hyde and then freeze-etched, have been only
partially successful, owing to technical difficul-
ties in obtaining clean replicas. However, pre-
fixation of E. coli B with OsO, by exposing a
lawn of cells on nutrient agar to vapors of the
fixative immediately before freezing, suggest
that septa are present in many dividing cells (R.
G. E. Murray, unpublished data).
In E. coli D22 env A and E. coli CRT 97 (Fig.
5C, 5D), prominent septa were seen without
difficulty by using any of the standard fixatives,
and Fig. 4C-4E show a simulated division
sequence. Both these organisms tended to grow
in chains, with 2 to 3 cells per chain (Fig. 4B;
refs. 22, 30). Although septa were easily pre-
served in sectioned material of E. coli D22 env
FIG. 3. Septation in E. coli Bir in glucose-Casamino Acids (A, C, D) and E. coli B (B). Septum is formed by
ingrowth of cytoplasmic membrane and mucopeptide (mp; A) or by cytoplasmic membrane across the cell (D).
An intermediate layer (i; A) between mucopeptide and outer membrane is excluded from the septum.
Unlabeled arrows show mesosome-like structures (B, C); nucleoid (n). Acrolein-glutaraldehyde fixation was
used. Bar in Fig. 3B equals 0.5 Am.
309
A, very few septa were seen in unfixed, freeze-
etched cells. In all cases of septation (Fig.
4C-4E), only the mucopeptide portion of the
wall, plus the cytoplasmic membrane or, more
rarely, the cytoplasmic membrane alone,
formed the septum. As in E. coli B or B/r, the
mucopeptide in the septum was composed of
two lamellae (Fig. 4C-4D). In a few cells of E.
coli D22 env A (Fig. 4E), the septum was
apparently composed of a single dense layer of
mucopeptide, about twice the thickness (4 to 6
nm) of the mucopeptide at the peripheral wall.
Splitting of the septum in these cells apparently
started at the center (Fig. 4D). Attempts to
stain this gap within the septum with phos-
photungstic acid or to see whether extracellular
tracers (e.g., lanthanum or ruthenium red)
would penetrate the septal arms, have been
unsuccessful. Mesosome-like structures were
also seen occasionally, associated with the ini-
tial stages of septation or with the completed
septum (Fig. 5B). Both in thin sections and in
freeze-etched material (Fig. 5A, 5B), the outer
membrane was closely apposed to the outer wall
components, although, as in E. coli B and B/r, it
was excluded from the septum until the onset
of cell separation.
Antibiotics and septation. Because low
doses of penicillin can block division selectively
(see Introduction), we have examined further
the relationship between septum formation and
lysis of E. coli. In these experiments, ampicillin,
rather than penicillin, was used, because one
strain (E. coli PAT 84; see below) was found to
be susceptible to ampicillin but indifferently so
to penicillin (see also ref. 39).
When ampicillin (2 to 5,ug/ml, final concen-
tration) was added to cultures of E. coli B/r
growing in glucose minimal medium, absorb-
ance (A,,0) continued to rise for about 1.5 h
(Fig. 6A). As monitored by phase-contrast mi-
croscopy and decrease in absorbance (Fig. 6A),
the cells rapidly lysed after this time. A promi-
nent bulge was observed in cells treated with
ampicillin (Fig. 7B, 7C) as previously shown for
penicillin-treated cultures of E. coli (46). The
bulge was found at or near the middle of the
cells (Fig. 7B, C) or, more rarely, towards one
pole. All layers of the wall plus the cytoplasmic
membrane were found at the site of the bulge.
In lysing cells (Fig. 7G), dense granules were
observed in the cytoplasm, and, on rupture of
310 BURDETT AND MURRAY J. BACTERIOL.

B ~~ i -
_Ajtpz

D '~~" ;~-W

A~~~~~~~~~~~~~A

E 'i t 4
FIG. 4. Septation in E. coli Bir in glucose-Casamino Acids (A) and E. coli D22 env A (B-E). Septum is
formed by cytoplasmic membrane (A) or by growth of the membrane and mucopeptide layer (C-E). Note
double lamellae of mucopeptide (arrows; D) or single structure (E); chain formation in cells of D22 env A is
shown in B. Acrolein-glutaraldehyde fixation was used. Bar in Fig. 4B equals 1 um.
VOL. 119, 1974 SEPTA IN E. COLI 311

II2
1.

M.om
4

,?,P.

T*I
I

D
FIG. 5. Freeze-etch (with glycerol) micrographs of septa (s) and mesosome-like membranes (m) in E. coli
D22 env A (A, B); arrows at top left indicate direction of shadowing. Septation in E. coli CRT97showing partial
(C) and complete (D) septa; RK fixation was used. Abbreviations as in Fig. 1. Bar in Fig. 5A equals 0.5 ,m.
312
.4
I A
0 , r- I i' D
",b
".5
0. 3-
2.1 E over 1 to 3 h. The env A mutants, both
2 3 2 3
TIME (HOURS)
FIG. 6. Effects of ampicillin, cephalothin and
CAM on growth of E. coli Bir in glucose minimal
medium (A). Control (no additions, 0), with ampicil-
lin (5 gg/ml, 0) or cephalothin (5 jg/ml, 0) added at
45 min. (B) Control (no additions, 0), with ampicillin
(5 gg/ml, 0), with CAM (2 jig/ml, V, or 40 tig/ml,
A), with cepthalothin (5ug/ml) and CAM (40 ,g/ml,
*); with ampicillin (5 ,ug/ml) and CAM (2 ,g/ml,
0). Monitored by change in absorbance at A,,,0,.
the wall, the edges of the outer membrane and
cytoplasmic membrane appeared to be fused
(Fig. 7G).
The addition of cephalothin (or cephalori-
dine) in low concentrations (1 to 5 ,ug/ml, final
concentration) resulted in lysis of the cells after
approximately one-half of a generation time,
i.e., about 25 min (Fig. 6A). Very few distinct
bulges were noted in cells treated with very low
doses of cephalothin (0.1 to 1.0 ,g/ml). Instead,
the septal area merely appeared as a furrow
(Fig. 7E). Cells treated with higher concentra-
tions of cephalothin (5,ug/ml or greater) showed
more extreme central enlargement and bulge
formation prior to lysis. Control cultures (no
additions) were septate, as described above.
The concentrations of ampicillin and cephalo-
thin used in these studies stopped division
immediately (unpublished data).
Bulge formation was suppressed by the addi-
tion of chloramphenicol if added (at a final
concentration of 2 to 40 ,ug/ml) at the same time
as ampicillin (5 Lg/ml) or cephalothin (5 jg/
ml). The absorbance of the cultures also re-
mained level (Fig. 6B). In thin section, the outer
membrane of cells treated with ampicillin and
low levels of CAM (2 ,ug/ml, final concentration)
was often extended and displaced (Fig. 7F);
BURDETT AND MURRAY J. BACTERIOL.
multiple layers of the outer membrane were also
seen (Fig. 7F). A slow increase in the A,,,8 was
noted for cultures treated solely with CAM (2
,ug/ml; Fig. 6B), whereas increase in absorbance
ceased after approximately 30 min in cultures
exposed to CAM at a final concentration of 40
,ug/ml (see Fig. 6B). On the other hand, if CAM
(2 to 4 og/ml) was added to a culture of E. coli
B/r previously treated with ampicillin (5.0 ,g/
ml), lysis could not be prevented unless the
CAM was added within 0.1 to 0.20 of a genera-
tion time. Control cultures (no additions)
showed no evidence of bulge formation (Fig.
7A).
Ampicillin and cephalothin in low concentra-
tions also induced bulge formation in E. coli
CRT 97, although lysis of the cultures extended
4 producers of penicillinase (30), continued to
grow and divide in the presence of the ampicil-
lin and cephalothin.
Studies on E. coli PAT 84. Some of the basic
properties of this mutant have been described
(18, 38). When grown at 30 C, the cells divided
normally (Fig. 10A), but on a shift to 42 C
(restrictive temperature) division ceased and
the cells began to elongate (Fig. 8). There was
no apparent increase in the growth rate follow-
ing the shift to 42 C, as measured by increase of
absorbance (Fig. 9). As far as we could deter-
mine, the cessation of division at 42 C was
abrupt, suggesting that even divisions in prog-
ress were halted. The filaments at 42 C con-
tained nucleoids spaced at intervals, as seen by
electron microscopy (Fig. 10B) or by a nucleoid
staining technique (18). When the cultures were
held at 42 C for 2 to 2.5 h and then returned to
30 C, the filaments became compartmental-
ized, and a rapid increase in cell numbers
followed, sequentially but not synchronously,
after a period of 15 to 25 min (Fig. 8). Prolonged
culture of PAT 84 at 42 C resulted in loss of
viability and decrease in cell numbers (Fig. 8).
Division at 42 C and effects of CAM. Divi-
sion of filaments at 42 C occurred if NaCl was
added to a final concentration of 0.25 to 0.34 M
if cells were initially grown in complete medium
either without NaCl or at low levels (0.08 M;
i.e., 5 g/liter). An increase of cell numbers, after
a lag of 15 to 20 min, occurred after addition of
NaCl (Fig. 11); in no case was the number of
cells as great as in a control culture left at 30 C.
Also, if NaCl (0.25 to 0.34 M) was added at the
time of a shift to 42 C, the cells continued to
divide (Fig. 11).
If CAM (40 to 100 ,g/ml, final concentration)
was added to cells at 42 C and at the time of a
VOL. 119, 1974 SEPTA IN E. COLI
shift-down to 30 C, a proportion of the cells
divided (Fig. 12). Both the A,,,, and cell num-
bers showed only slight increases if CAM was
added 10 min (or longer) before the shift-down
(Fig. 12, 13). If CAM (40 ug/ml) and NaCl (0.25
M) were added to cultures at 42 C at the same
time, then a limited increase in cell numbers
also was detectable (Fig. 11).
Bulge formation and lysis. We tested PAT
84 further to determine whether the nonseptate
filaments would form bulges if cultures at 42 C
were treated with ampicillin. The response of
the cells was found to be related to the NaCl
content of the medium. If NaCl was omitted
from the medium, no bulges were formed, but in
media containing 0.08 M NaCl low concentra-
tions of ampicillin (2.5 ,g/ml) added at the time
of a shift to 42 C induced bulge formation in
very few cells (Fig. 10C). These cells may have
possessed partially completed septa or were
otherwise "committed" to septation at the time
of the shift-up. Intracellular membranes were
also seen in these cells (Fig. 10D). The concen-
tration of ampicillin used (2.5 ,g/ml) induced
bulge formation in cultures at 30 C; these
cultures, like those at 42 C containing ampicil-
lin (2.5 ,gg/ml) lysed after about 1 to 1.5 h (Fig.
14A). Very low concentrations of ampicillin (1
,ug/ml) permitted growth of the cells at 42 C
without lysis (Fig. 14A).
If NaCl (0.25 M, final concentration) was
added to a culture of PAT 84 growing in the
presence of ampicillin (1 Ag/ml, added at time
of shift-up), lysis of the culture ensued after a
delay of about 20 min (Fig. 15A); bulges were
also observed (Fig. 10E). Similarly, if ampicillin
(2.5 Ag/ml) was added to a culture on a shift-up
to 42 C and the flask was returned to 30 C, lysis
also occurred (Fig. 14B). Conversely, if ampicil-
lin (2.5 ,ug/ml), lysed after about 1 to 1.5 h (Fig.
shift-down to 30 C, lysis occurred (Fig. 14B).
These experiments, like those described below,
used complete medium containing 0.08 M
NaCl. In most cases, lysis and bulge formation
were found to occur at one or two sites along the
filaments or occasionally at the poles of the cell
(Fig. 16A). Completely lysed cells, viewed in
thin sections, showed a profile where no obvious
layer of mucopeptide was found in the vicinity
of the bulge or at the presumed septal area (Fig.
16B, C; 17A); the remainder of the wall ap-
peared to be intact. It should be noted that
high- concentrations of ampicillin (50 ug/ml),
when added to cells at 42 C, promoted rapid
lysis (Fig. 14A). Our initial studies using peni-
cillin G produced variable results in experi-
ments of this type.
313
With the exception of the experiment de-
scribed below, we have not been able to preserve
a septum in PAT 84 grown at 30 or 42 C (cf. Fig.
10A). Freeze-etch studies also showed constric-
tive divisions, but suggested that differences in
the distribution of particles on the cytoplasmic
membrane existed between cells grown at 30
and 42 C. A netlike array of particles on the
convex membrane face was seen on replicas of
cells grown at 30 C, whereas cultures at 42 C
appeared to have randomly grouped particles
(Fig. 17C-D). The surface of freeze-etched cells,
transferred to 42 C in the presence of ampicillin
(2.5 jug/ml) and then shifted down to 30 C,
appeared to possess fewer particles (Fig. 17E)
than the control samples at 30 or at 42 C.
When cells of PAT 84 were diluted into
nalidixic acid (10 ,g/ml, final concentration)
and then shifted to 42 C, elongation occurred. If
ampicillin (1 gg/ml, final concentration) was
also added, then bulge formation, as noted
above, did not occur. Increase in A,,,, (Fig. 15B)
continued to increase in both cases but levelled
off after about 1 to 1.5 h at 42 C. When NaCl
(0.25 M, final concentration) was added to
cultures containing nalidixic acid and ampicil-
lin at 42 C, lysis eventually followed (Fig. 15B).
Examination of the cells under the light micro-
scope suggested that the division site was, in
some cells, displaced towards a pole of the
filament. In thin section, the septum was ob-
served to consist solely of the cytoplasmic
membrane, and intracellular membranes, per-
haps erratic incursions of the cytoplasmic mem-
brane, were also present (Fig. 17B). Rupture of
the cells was evident at sites of contact of these
membranes with the wall (Fig. 17B).
DISCUSSION
Constrictive division in E. coli is an artifact;
division of E. coli B and B/r is accomplished by
septum formation and, in common with the
majority of bacteria, the septum is formed by an
ingrowth of the cytoplasmic membrane plus the
mucopeptide of the wall. Septa of similar struc-
ture were also observed in E. coli D22 env A and
in E. coli CRT 97. Our current model of the
structure of the septum is shown in Fig. 18. The
outer membrane does not appear to participate
in septation until the cells begin to separate,
but preparatory activity is indicated by the
presence of a bleb, or groups of vesicles, at the
site of division. This activity was particularly
marked in fixed and sectioned cells of E. coli B
and B/r, although it was not noted in freeze-
etched material or in sections of E. coli D22 env
A or CRT 97. The mucopeptide portion of the
_iar'|."zs,;6^{wt,S.
A

C
i.

tA''.''e''
D
_
;J.. ':

eSfk¢*
,e
""e+;

*'2beve
>'+'.;' ;

'v-
t *'s,"'
*> -r
::zZ
%Arv1 "

_7
>. .e,-

v.,Za,
:,

Ms. ;'

-
-

fi

*#-X
*.

,.
S
.

.s,
I -.

-.4
1.
A

,:,
?'-w.
... ..
kq.!-
, r

ii 1.
"
i ".

rS-
yALS *-:L. 40c,_

E 1. ..!.!

-.A

, Esb *Mt;S,?

F
*5kLe FIG. 7
314
VOL. 119, 1974 SEPTA IN E. COLI
10
9
8
C)
x
-j
2: 3
-j
LLJ
C)
0 1 2 3 4 5 6 7
TIME (HOURS
FIG. 8. Effect of temperature shift on division of E.
coli PAT 84. Cells at 30 C (0) were shifted to 42 C
(0), and division ceased abruptly. A portion of this
sample was returned to 30 C (0), and increase in cell
number was followed with the Coulter counter.
septum in E. coli is essentially a double struc-
ture, formed by two lamellae separated by a gap
(Fig. 2D, 5C). Electron microscopy of synchro-
nised culture of E. coli B and B/r (Burdett and
Murray, submitted or publication) have fully
confirmed this observation. Similar structures
were reported in cells of E. coli B held at 45 C
(49) and are visible in micrographs of E. coli
15T (50). The nature of the material present in
the "gap" (Fig. 18) is not known. As we have
suggested elsewhere (Burdett and Murray, sub-
mitted for publication), it is possible that at
least a portion of the lipoprotein (4) covalently
linked to mucopeptide may be located here,
because there is continuity between the site of
this component of the peripheral wall and the
central portion of the septum (Fig. 18). Septa of
Spirillum serpens (49) are formed of a single
dense line, and this organism has been reported
to lack the covalently bound lipoprotein (24).
Mesosomes of the type reported here have
also been observed by others in E. coli (33, 44)
and in Pseudomonas aeruginosa (19); meso-
FIG. 7. Sections of E. coli Bir in glucose minimal medium (A) and treated for I h with ampicillin (5 jg/ml)
showing bulges (arrows, B, C, G); CAM (2.g/ml, D); cephalothin (1 ,ug/ml, E; note furrow [f] at division site);
ampicillin (5 ug/ml) plus CAM (2 jg/ml, F). Arrows in F show two layers of outer membrane; arrows in G show
apparent fusion between outer membrane and cytoplasmic membrane; nucleoid (n). Acrolein-glutaraldehyde
fixation was used. All bars equal 0.5 Mm, except 7F and G (0.1 Am).
315
somes associated with the septum have been
reported in a gram-variable coccus (12). The
latter organism, like some cells of E. coli B/r
(Fig. 3D, 4A), appears to initiate septation by
forming a membrane across the cell prior to
ingrowth of the wall.
Septa in E. coli are difficult to preserve. Even
when the cells were rapidly frozen in Freon,
prior to freeze-etching, it was very seldom
possible to demonstrate septa. The necessity to
concentrate the suspension of cells prior to
freezing may well provide conditions that would
trigger the disruption of the septum, or disrup-
tion may occur during the freezing process
through concentration of solutes. This disrup-
tion is hard to prevent with fixatives. However,
it is by no means clear that disruption of the
septum will necessarily result in constrictive
divisions. This is illustrated in Fig. 19, which
emphasizes two features: (i) even in constrictive
cells, the mucopeptide layer appears continu-
ous; (ii) the wall adjacent to the nascent septum
(arrow 1) might also infold to yield the cleft seen
in constrictive cells (arrows 2, 3). Thus, "col-
lapse" of the septum may occur when the
material in the "gap" (Fig. 19) is broken down
and the septum can no longer maintain its
integrity. Enzymatic attack at this location
could conceivably occur through proteases or
enzymes breaking the linkage of the lipoprotein
to mucopeptide. Some cells of E. coli D22 env A
were observed to possess a septum consisting of
a single dense lamella of mucopeptide (see also
ref. 30). This may indicate that this mutant is
deficient in the synthesis of "gap" substances.
1 00
050
C)
to
to 0.1 0
0.0 5
0.01
0 1 2 3 4 5 6 7
TIME (HOURS )
FIG. 9. Growth of E. coli PAT 84 following a shift
from 30 to 42 C and then back to 30 C, measured by
change in absorbance at As*o.
 1- I-C -, .. 1, .1.
't .

..o. 11 . .-

jt

4. 4I..

A
- ^.*
*,,.
4..0.
4..',0

'a
4
WI..,.I.
B
I 1.
m"W
3..

4...
-.1F.. -V".xarli
I-t,

*-.1

t..
>o s.:, :....

D
FIG. 10
316
VOL. 119, 1974 SEPTA IN E. COLI 317
lF suggest that the center of the cell is a site of
9 F t _ incorporation of newly synthesized mucopep-
8 e /S tide (43). This site is also likely to be the
7 /)^ - location of hydrolytic enzymes (46). Both peni-
6 cillin and ampicillin block division by inhibi-
W///0 _ tion of the transpeptidation reaction (21) nor-
5 mally cross-linking D-alanine to diaminopimelic
2// acid on adjacent peptide chains (51). There is
4- 0 some evidence showing that the transpeptidase
is at least one of the penicillin-sensitive targets
r X / involved in septation (14). Although penicillin
3 and ampicillin both block transpeptidation, it is
likely that ampicillin may enter cells of E. coli
more easily than penicillin (51). The effects of
cephalosporins, which may block both septation
2t tf and elongation, are difficult to understand (for
discussion, see ref. 23, 31). As in Staphylococcus
/ 1 2 3 aureus (40), lysis may be prevented by addition
of CAM in the presence of ampicillin, possibly
due to the inhibition of synthesis of hydrolytic
l 10
9
0 1 2 3 4 5 8
TIME (HOURS
FIG. 11. Effect of NaCI on division of E. coli PAT 7
84. Control (in complete medium + 0.08 M NaCI, 0); 0
o
cells at 42 C in same medium (arrow 1, A) or with _
NaCI (to 0.25 M) at time of shift to 42 C (0); addition 5
of NaCI (to 0.25 M) to cells at 42 C (arrow 2, 0); with
NaCI (to 0.25 M) plus CAM (40 ,g/ml) at 42 C (arrow 4

3, U).0
However, the tendency to grow in chains may
also indicate an impairment of the separation -KZ
xJ
mechanism, a reflection perhaps of an altered \
autolysin. In B. subtilis (9, 10) and S. faecalis -
(48), isolated autolysins and also lysozyme (10)
have been used to induce fragmentation of
O 2 1'*
chains of cells. Lysozyme disrupts the septum of 42 30
the D22 env A mutant (30). Chain formation in
Bacillus subtilis and S. faecalis appears to
result from a lowered cell content of autolysin
rather than from changes in the type of autoly- I
sin or alterations in wall structure. It is possible 3

that constrictive divisions arise from two 0 1 2 3 4 5 6 7

sources, related to the stage of septation TIME (HOURS)
reached by individual cells: (i) rupture of par- FIG.PAT12.84Effect of CAM (40 gg/ml) on division of E.
tially
tially complete
complete septa
septa byby iassolution the coli
dissolutionofof the Control returned
culture returned to C30 after
to 30 C (nogrowth at 42 C.
additions, 0);
"gap" substance, and (ii) autolysis of septum plus CAM at time of transfer to 30 C (E); with CAM
"initials" by hydrolytic enzymes associated added S min (A) or 10 min (A) prior to transfer to 30 C;
with the early ingrowth of the septum. control left at 30 C (0) or shifted to 42 C without
Recent autoradiographic studies of E. coli W7 CAM (0).
FIG. 10. Sections of cells of E. coli PAT 84 grown at 30 C (A), at 42 C (B), or at 42 C in the presence of
ampicillin (2.5,ug/ml for 45 min; C, D; arrows in D show an intracellular membrane). A portion of the filaments
is shown in B and C. Cells grown in complete medium plus 0.08 M NaCI at 30 C and shifted to 42 C in the
presence of ampicillin (2.5 ug/ml), to which was added further NaCI to a final concentration of 0.25 M, show
bulges (arrow: E); nucleoid (n). Acrolein-glutaraldehyde fixation was used. Bars equal 0.1 um (D), 0.2 gm (E),
0.5,m (A), and 1,gm (B, C).
318 BURDETT AND MURRAY J. BACTERIOL.
enzymes. Alternatively, if transport of the ac-
tive hydrolase(s) from the cytoplasm requires
protein synthesis, bulge formation might still be
prevented by addition of CAM. In bacilli,
protein synthesis must be inhibited before addi-
tion of wall inhibitors (vancomycin, D-cycloser-
ine) in order to arrest lysis (42).
Bulge formation in E. coli occurs at, or close
to, the sites of division. The potential for the
C)
to
expression of bulge formation appears to be
(D
correlated with the deoxyribonucleic acid
(DNA) replication cycle and it has been shown
that the maximum occurs at or near the time
coincident with the end of a round of replication
(Burdett and Murray, submitted for publica-
tion; 20). If bulge formation is due to the action
of mucopeptide hydrolases, then there would
appear to be a significant increase in hydrolytic
activity close to the point where the cells are
entering the septation sequence, i.e., the D
period (5). An essential prelude to septation is
likely to be a structural modification and, since
-45 -30 -15 0 +415+30 +45 +60 +75 +90 +105 both septation and bulge formation involved
TIME (MINS.) primarily the mucopeptide, the likely candidate
FIG. 13. Absorbance of cultures- (A..,) of E. coli
PAT84 treated with CAM (40 ug/ml) at various times .0 !r Ti-
,
o-l.G
0. 9 9
XA B 9
prior to shift-down from 42 to 30 C. 0. 8 -_,8
I 7

3 9__ 0 .9 0.- -0.6

0h-a A 0.6 -
, - 0.5
-.10
I -g Co6
O. S.
0X
,

0.4 _
-
ID
c o0II -0.3 w

0
"I

< 0.23
K - 0.2

1 iI

0.1 In
2 3 4 2 3 4 0 2 3 2 3
TIME (H0URS) TIME (HOURS
FIG. 14. Effect of ampicillin on growth of cultures FIG. 15. Effect of NaCI on lysis of E. coli PAT84 at
of E. coli PAT 84 monitored by change in absorbance 42 C measured by change in abosrbance (A*..). Con-
(A ...). (A) Control at 30 C (no additions, 0) with trol at 42 C (no additions, 0) or with NaCI (to 0.25 M)
ampicillin (2.5 ug/ml, U); cells at 42 C (no additions, from time of shift up to 42 C (A); cells at 42 C in
0) or with ampicillin (1 jg/ml, A), 2.5 jg/ml (V), or presence of ampicillin (1 ug/ml) added at time of
50 ug/ml (0). Arrows indicate time of transfer to 42 C. shift-up to 42 C (0) and treated by addition of NaCI
(B) Control (no additions, 0) shifted from 30 to 42 C (to 0.25 M). (B) Portions of same culture as in (A)
and back to 30 C; cells shifted to 42 C plus ampicillin treated with nalidixic acid (10 jg/ml) at shift to 42 C
(2.5 jg/ml, at time of shift to 42 C), and then returned (0) or with nalidixi'c acid and ampicillin (1 jig/ml,
to 30 C (0) or ampicillin (2.5 ug/ml) added at time of V); NaCI was added (arrow) to 0.25 M to a portion of
shift down to 30 C (A). the latter culture.
FIG. 16. Sections of cells of E. coli PAT 84 grown at 42 C in presence of ampicillin (2.5 jig/ml) and then
shifted down to 30 C. (A) Spheroplast emerging from a filament (arrow); (C) portions of lysed cells showing
bulge formation. The mucopeptide layer (arrows, B) appears to be absent from the site of cell lysis (C). Acrolein-
glutaraldehyde fixation was used. Bar in Fig. 16A equals 0.5 ,um.
 -
;k$
S;N &
W..'
..
I..

:A '1

.
-

0.r
FIG. 16
319
 B

FIG. 17
320
VOL. 119, 1974 SEPTA IN E. COLI 321
OM(B ) necessary materials are made at 42 C but can-
not be assembled. On a shift-down to 30 C, and
following a lag of 15 to 20 min, the cells divided
at a rate initially faster than that of a control
culture left at 30 C. A proportion of the cells
also divided in the presence of CAM and, as
shown elsewhere (Burdett and Murray, submit-
ted for publication; 6), it is likely that protein
synthesis may not be required in the terminal
stages of division. The cells that actually di-
vided, therefore, may have already entered the
late stages of septation. Does the apparent de-
lay before the emergence of new cells at 30 C
(Fig. 8), or at 42 C in the presence of NaCl (Fig.
11), reflect an assembly of the septation mech-
anism or is some other process taking place as a
FIG. 18. Diagram of septum structure (see text). preliminary step to septation?
Symbols: OM (B), bleb formed by outer membrane When treated with ampicillin (1 to 2.5 Agl
(OM); mucopeptide (MP); cytoplasmic membrane ml) at the time of a shift-up, very few cells
(CM). Mucopeptide lamellae in septum, MP(S); gap actually formed bulges. But when shifted down
is electron-transparent space between lamellae. The to 30 C, or when NaCl (to 0.25 M) was added at
position of the lipoprotein covalently bound to muco- 42 C, the cells lysed after a delay of some 20
peptide is indicated (a). Note that outer membrane is min. This could be interpreted to mean that the
excluded from septum (see text).
essential lesion in the mutant is the lack of
mucopeptide hydrolase activity at 42 C. Elec-
is the cross-linked and covalently bonded net- tron microscopy confirmed that rupture of the
work of the mucopeptide layer. It is claimed cells involved the formation of distinct bulges;
that the rate of mucopeptide synthesis fluctu- discontinuities in the mucopeptide layer at sites
ates throughout the cell cycle of E. coli (20) and of bulge formation were also noted. Experi-
rises at about the start of the D period. One ments by Shannon et al. (47) on a temperature-
possibility, therefore, is that local hydrolase sensitive DNA initiation mutant of Salmonella
activity at the center of the cell modifies the typhimurium also suggest that division sites
existing mucopeptide network to allow for the can exist in filaments and that septation is
insertion of the newly synthesized septum. The preceded by an "initiation" step. PAT 84 there-
visible events of the D period appear to occur in
the latter half of the septation sequence (Bur- 2 3
dett and Murray, submitted for publication),
implying that synthesis and assembly occur A
prior to the initial ingrowth of the cross wall.
This interpretation of events leading to septa-
tion is also supported by our observations on E. 1
coli PAT 84. This mutant, like BUG-6 (35-37), 2 3
retains the capacity to divide after a shift-down
to 30 C. The cells can also divide at 42 C by B
adding high concentrations of NaCl (or other
solutes, including sucrose; see ref. 38). It has FIG. 19. Diagram illustrating a possible mecha-
been suggested that PAT 84 is a missense nism accounting for constrictive divisions. 7he sep-
mutant and that the lesion can be corrected by tum (arrow 1; A) might be pulled apart and collapse
the addition of NaCl (39). into a wide cleft (arrows 2, 3; B). The profiles of a
Our experiments indicate that, although the septate cell (A) and a constricted cell (B) are also
cells cannot normally form septa at 42 C, all the shown. See text for discussion.
FIG. 17. (A) Same as Fig. 16A-C; note mucopeptide layer (arrows) is absent in septal area (arrow a); (B)
section of E. coli PAT 84 treated with nalidixic acid (10 Ag/ml), ampicillin (1 ug/ml), and NaCI (to 0.25 M); see
Fig. 15B. Septum(a) is composed of cytoplasmic membrane; incursions of the membrane (arrow) are also
associated with sites of cell lysis (nucleoid, n). (C-E) Freeze-etched cells (with glycerol) of E. coli PAT 84 at
30 C (C) showing netlike array of particles on convex layer of cytoplasmic membrane, at 42 C (D) where
particles are-more aggregated or as netlike arrays after shift-down to 30 C in presence of ampicillin (2.5 ,Ag/ml;
E). Arrows at top left indicate direction of shadowing. Bar in Fig. 17D equals 0.5 um.
 322 BURDErT AND MURRAY
fore appears to be fully compartmentalized at
42 C, in the sense that DNA synthesis can
continue and the nucleoids can be segregated
(18), but the filaments cannot divide.
For the mutant BUG-6, Reeve et al. (35-37)
have proposed that a protein ("division poten-
tial") is synthesized throughout the cell cycle.
Shortly after the completion of a round of DNA
replication, this protein is assumed to become
involved in the division process. Our studies of
PAT 84 suggest that "division potential" might
be, in fact, an impaired mucopeptide hydro-
lase(s). This implies that PAT 84 (and perhaps
BUG-6) are mutants which are blocked in an
early stage when the wall is modified prior to
the assembly of the septation apparatus. All the
necessary precursors or requirements for divi-
sion (DNA, ribonucleic acid, and protein syn-
thesis) appear to be synthesized during the
preceding C period (Burdett and Murray, sub-
mitted for publication; 6). Both PAT 84 and
BUG-6 can divide in the presence of CAM,
suggesting that the septum or its components
are actually synthesized at 42 C but cannot be
assembled. We are assuming that the mucopep-
tide substrate itself is not modified at the
restrictive temperature. Because division ceases
abruptly on a shift-up to 42 C, it is possible that
the impaired hydrolase(s) is in fact used
throughout septation and cell separation.
Ahmed and Rowbury (1) have performed
similar experiments on a temperature-sensitive
mutant of S. typhimurium. This mutant, unlike
PAT 84, cannot be induced to divide at 42 C by
addition of NaCl. On addition of penicillin (25
U/ml) to cells at 42 C, no bulges were obsetved
in the filaments, implying, as these authors
suggest, the impairment of an autolysin at 42 C.
Other mutants of E. coli (28) can only divide at
42 C provided they are placed initially at 30 C
for a brief period. This mutant may possess an
impairment in the initiation of septation, which
may include an inactivation, or loss of synthe-
sis, of an autolysin. The requirement for a brief
period of protein synthesis at 30 C in tempera-
ture-sensitive E. coli 20 (ref. 28), if division is to
occur in the presence of CAM, suggests that one
or more proteins are not synthesized at 42 C.
Detection and quantitative measurement of the
component proteins are clearly needed to re-
solve these alternatives. Using envelopes of
PAT 84, we have obtained preliminary data, by
means of sodium dodecyl sulfate gel electro-
phoresis, indicating that the components of the
wall and membrane are essentially the same at
30 and 42 C. But the preparations are of such
relative crudity, indicating no more than the
J. BACTERIOL.
presence of a spectrum of proteins, that the
results may be misleading and are therefore not
included in this paper. Other types of lesion
in these mutants are also possible. If, for ex-
ample, the hydrolase(s) were synthesized in a
latent, perhaps cytoplasmic, form, transport
of the active enzyme through the cytoplasmic
membrane might be affected. Whether the dis-
tribution of membrane particles noted in freeze-
etched material (see Fig. 17C-E) reflect con-
formational alteration of proteins in the mem-
brane is not clear (cf. ref. 52).
In summary, then, we envisage the septation
sequence in E. coli to involve: (i) cleavage of the
existing mucopeptide framework by a hydrolytic
enzyme(s) prior to synthesis of the septum, and
(ii) assembly of the septum utilizing the compo-
nents synthesized during the preceding round of
DNA replication. The labile nature of the sep-
tum, especially at the initial stages, may be due
to the close association of the autolysin(s) with
the septation site. The preservation of the
dynamic structural components of cell division
for electron microscopy to provide satisfactory
correlation of structural and biochemical events
in the septation process will require new
methods of fixation that inhibit the enzymes
involved in autolysis. The fixatives we have
used, although effective for E. coli B and B/r,
are demonstrably ineffective for some other
bacteria showing "constrictive division" with
RK fixation, including PAT 84.
ACKNOWLEDGMENTS
We wish to thank H. Boman, C. E. Helmstetter, and F.
Jacob for generously donating strains of E. coli used in this
study. The technical assistance of M. Hall, H. Koppenhoef-
fer, J. Marak, and G. Sanders is much appreciated.
The financial assistance of the Medical Research Council
of Canada is gratefully acknowledged.
LITERATURE CITED
1. Ahmed, N., and R. J. Rowbury. 1971. Antibiotics and cell
division in a filament-forming mutant of Salmonella
typhimurium. Microbios 4:181-191.
2. Bertani, G. 1951. Studies on lysogenesis. I. The mode of
phage liberation by lyosgenic Escherichia coli. J. Bac-
teriol. 62:293-300.
3. Braun, V., H. Gnirke, U. Hennig, and K. Rehn. 1973.
Model for the structure of the shape-maintaining layer
of the Escherichia coli cell envelope. J. Bacteriol.
114:1264-1270.
4. Braun, V., and K. Rehn. 1969. Chemical characteriza-
tion, spatial distribution and function of a lipoprotein
(murein-lipoprotein) of the E. coli cell wall. The
specific effect of trypsin on the membrane structure.
Eur. J. Biochem. 10:426-438.
5. Cooper, S., and C. E. Helmstetter. 1968. Chromosome
replication and the division cycle of Escherichia coli
B/r. J. Mol. Biol. 31:519-540.
6. Dix, D. E., and C. E. Helmstetter. 1973. Coupling
between chromosome completion and cell division in
VOL. 119, 1974 SEPTA IN E. COLI
Escherichia coli. J. Bacteriol. 115:786-795.
7. Donachie, W. D., N. C. Jones, and R. Teather. 1973. The
bacterial cell cycle. Symp. Soc. Gen. Microbiol.
23:9-44.
8. Ellar, D. J., D. G. Lundgren, and R. A. Slepecky. 1967.
Fine structure of Bacillus megaterium during synchro-
nous growth. J. Bacteriol. 94:1189-1205.
9. Fan, D. P. 1970. Autolysin(s) of Bacillus subtilis as
dechaining enzyme. J. Bacteriol. 103:494-499.
10. Fan, D. P., M. P. Beckman, and W. P. Cunningham.
1972. Ultrastructural studies on a mutant of Bacillus
subtilis whose growth is inhibited due to insufficient
autolysin production. J. Bacteriol. 109:1247-1257.
11. Forsberg, C., and H. J. Rogers. 1971. Autolytic enzymes
in growth of bacteria. Nature (London) 229:272-273.
12. Gilleland, H. E., Jr., I. L. Roth, and R. G. Eagon. 1971.
Ultrastructure of the cell wall and the mechanism of
cellular division of a gram-variable coccus. Can. J.
Microbiol. 17:421-424.
13. Glauert, A. M., and M. J. Thornley. 1969. The topogra-
phy of the bacterial cell wall. Annu. Rev. Microbiol.
23: 159-198.
14. Hartmann, R., J. V. Holtje, and U. Schwarz. 1972.
Targets of penicillin action in Escherichia coli. Nature
(London) 235:426-429.
15. Helmstetter, C. E., and S. Cooper. 1968. DNA synthesis
during the division cycle of rapidly growing Escherichia
coli B/r. J. Mol. Biol. 31:507-518.
16. Higgins, M. L., H. M. Pooley, and G. D. Shockman. 1970.
Site of initiation of cellular autolysis in Streptococcus
faecalis as seen by electron microscopy. J. Bacteriol.
103:504-512.
17. Higgins, M. L., and G. D. Shockman. 1971. Procaryotic
cell division with respect to walls and membranes.
CRC Crit. Rev. Microbiol. 1:29-72.
18. Hirota, Y., A. Ryter, and F. Jacob. 1968. Thermosensitive
mutants of E. coli affected in the processes of DNA
synthesis and cellular division. Cold Spring Harbor
Symp. Quant. Biol. 33:677-693.
19. Hoffmann, H.-P., S. G. Geftic, H. Heymann, and F. W.
Adair. 1973. Mesosomes in Pseudomonas aeruginosa. J.
Bacteriol. 114:434-438.
20. Hoffmann, B., W. Messer, and U. Schwarz. 1972. Regula-
tion of polar cap formation in the life cycle of Esche-
richia coli. J. Supramol. Struct. 1:29-37.
21. Izaki, K., M. Matsuhashi, and J. L. Strominger. 1968.
Biosynthesis of the peptidoglycan of bacterial cell
walls. XIII. Peptidoglycan transpeptidase and D-ala-
nine carboxypeptidase: penicillin sensitive enzymatic
reaction in strains of Escherichia coli. J. Biol. Chem.
243:3180-3192.
22. Kohiyama, M., D. Cousin, A. Ryter, and F. Jacob. 1966.
Mutants thermosensibles d'Escherichia coli. Ann. Inst.
Pasteur (Paris) 110:465-486.
23. Lorian, V., and L. D. Sabath. 1972. Penicillins and
cephalosporins: differences in morphologic effects on
Proteus mirabilis. J. Infect. Dis. 125:560-564.
24. Martin, H. H., H. D. Heilmann, and H. J. Preusser. 1972.
State of the rigid-layer in cell walls of some gram-nega-
tive bacteria. Arch. Mikrobiol. 83:332-346.
25. Moore, H. 1966. Use of freeze-etching in the study of
biological ultrastructure. Int. Rev. Exp. Pathol.
5:179-216.
26. Murray, R. G. E. 1968. Bacterial cell wall anatomy in
relation to the formation of spheroplasts and proto-
plasts, p. 1-18. In L. B. Guze (ed.), Microbial proto-
plasts, spheroplasts and L-forms. The Williams &
Wilkins Co., Baltimore.
27. Murray, R. G. E., P. Steed, and H. E. Elson. 1965. The
location of mucopeptide in sections of the cell wall of E.
coli and other gram-negative bacteria. Can. J. Micro-
323
biol. 11:547-560.
28. Nagai, K., and G. Tamura. 1972. Mutant of Escherichia
coli with thermosensitive protein in the process of
cellular division. J. Bacteriol. 112:959-966.
29. Nagata, T. 1963. The molecular synchrony and sequential
replication of DNA in Escherichia coli. Proc. Nat.
Acad. Sci. U.S.A. 49:551-559.
30. Normark, S., H. G. Boman, and G. D. Bloom. 1971. Cell
division in a chain-forming env A mutant of Esche-
richia coli K12. Acta Pathol. Microbiol. Scand. B
79:651-664.
31. Perkins, R. L., and M. A. Miller. 1973. Scanning electron
microscopy of morphological alterations in Proteus
mirabilis induced by cephalosporins and semi-syn-
thetic penicillins. Can. J. Microbiol. 19:251-255.
32. de Petris, S. 1967. Ultrastructure of the cell wall of E. coli
and chemical nature of its constituent layers. J. Ultra-
struct. Res. 19:45-83.
33. Pontefract, R. D., G. Bergeron, and F. S. Thatcher. 1969.
Mesosomes in Escherichia coli. J. Bacteriol.
97:367-375.
34. Radolakis, A., P. Thomas, and J. Starka, 1973. Morpho-
logical mutants of Escherichia coli. Isolation and
ultrastructure of a chain-forming env C mutant. J.
Gen. Microbiol. 75:409-416.
35. Reeve, J. N., and D. J. Clark. 1972. Cell division of
Escherichia coli BUG-6: effect of varying the length of
growth at the nonpermissive temperature. J. Bacteriol.
110:117-121.
36. Reeve, J. N., and D. J. Clark. 1972. Cell division of
Escherichia coli BUG-6: effect of varying the tempera-
ture used as the non-permissive growth condition. J.
Bacteriol. 110:122-125.
37. Reeve, J. N., D. Groves, and D. J. Clark. 1970. Regulation
of cell division in Escherichia coli: characterization of
temperature-sensitive division mutants. J. Bacteriol.
104:1052-1064.
38. Ricard, M., and Y. Hirota. 1973. Effet osmotique sur le
phenotype de mutants thermosensibles d'E. coli. Ann.
Inst. Pasteur (Paris) 124A:29-43.
39. Ricard, M., and Y. Hirota. 1973. Process of cellular
division in Escherichia coli: physiological study on
thermosensitive mutants defective in cell division. J.
Bacteriol. 116:314-322.
40. Rogers, H. J. 1967. Killing of staphylococci by penicillins.
Nature (London) 213:31-33.
41. Rogers, H. J. 1970. Bacterial growth and the cell enve-
lope. Bacteriol. Rev. 34:194-214.
42. Rogers, H. J., and C. W. Forsberg. 1971. Role of autoly-
sins in the killing of bacteria by some bactericidal
antibiotics. J. Bacteriol. 108:1235-1243.
43. Ryter, A., Y. Hirota, and U. Schwarz. 1973. Process of
cellular division in Escherichia coli growth pattern of
E. coli murein. J. Mol. Biol. 78:185-195.
44. Ryter, A., and F. Jacob. 1966. Etude morphologique de la
liason du noyau a la membrane chez E. coli et les
protoplases de B. subtilis. Ann. Inst. Pasteur (Paris)
110:801-812.
45. Ryter, A., and E. Kellenberger. 1958. Etude au micro-
scope electronique de plasmas contenant de l'acide
desoxyribonucleique. I. Les nucleoides des bacteries en
croissance active. Z. Naturforsch. B 13:597-605.
46. Schwarz, U., A. Asmus, and H. Frank. 1969. Autolytic
enzymes and cell division of Escherichia coli. J. Mol.
Biol. 41:419-429.
47. Shannon, K. P., B. G. Spratt, and R. J. Rowbury. 1972.
Cell division and the production of cells lacking nuclear
bodies in a mutant of Salmonella typhimurium. Mol.
Gen. Genet. 118:185-197.
48. Soper, J. W., and C. G. Winter. 1973. Role of cell wall
autolysin in chain formation by a mutant strain of
324 BURDETT AND MURRAY
Streptococcus faecalis. Biochim. Biophys. Acta
297:333-342.
49. Steed, P., and R. G. E. Murray. 1966. The cell wall and
cell division of gram-negative bacteria. Can. J. Micro-
biol. 12:263-270.
50. Stevenson, I. L. 1969. The cytological effects of myxin on
Escherichia coli. Can. J. Microbiol. 15:707-711.
51. Strominger, J. L. 1970. Penicillin-sensitive enzymatic
reactions in bacterial cell wall synthesis, p. 179-213.
J. BACTERIOL.
Harvey Lectures, series 64. Academic Press Inc., New
York.
52. Wu, P. C., and A. B. Pardee. 1973. Cell division of
Escherichia coli: control by membrane organization. J.
Bacteriol. 114:603-611.
53. Zusman, D. R., M. Inouye, and A. B. Pardee. 1972. Cell
division in E. coli: evidence for regulation of septation
by effector molecules. J. Mol. Biol. 69:119-136.