Vol. 128, No. 1 Printed in U.S.A.

JOURNAL OF BACTERIOLOGY, Oct. 1976, p. 401-412 Copyright © 1976 American Society for Microbiology

Septum Formation-Defective Mutant of Escherichia coli STAFFAN NORMARK,* LENA NORLANDER, THOMAS GRUNDSTROM, GUNNAR D. BLOOM, PAUL BOQUET, AND GERARD FRELAT Departments of Microbiology and Histology, University of Ume&, S-901 87 Umed, Sweden,* and DOpartement de Biologie, Service de Biochimie, Centre d'ktudes Nucleaires de Saclay, 91-Gif-sur-Yvette, France Received for publication 24 June 1976

Mutants of Escherichia coli defective in septum initiation, as well as in septum formation, were obtained spontaneously, without mutagenic treatment, by selection of rifampin-tolerant mutants of an antibiotic-permeable strain carrying the envA mutation. The disturbed phenotype was in all mutants aggrevated at low incubation temperatures. One allele, sefAl, was studied in detail. Septum initiation, as well as septum formation, was promoted by high cell densities or by the addition of low concentrations of certain antibiotics, e.g., rifampin and chloramphenicol, to low-density cultures. The observed rifampin dependence was studied in detail. These experiments indicated that a very modest shift-down situation suppressed the phenotype and enabled constrictions to proceed to cell separation. The rifampin sensitivity of the partially purified deoxyribonucleic acid-dependent ribonucleic acid polymerase was not affected by the sefAl allele, which is located close to proA and is thus distinct from envA. Growth parameters during the shift to 25°C were followed in a transductant carrying the sefAl allele. At the nonpermissive temperature, about one constriction per cell was formed. This constriction was characteristically blunt and did not lead to cell separation. At the time of formation of these frozen constrictions, clear zones representing a separation of wall from cytoplasmic membrane appeared. These polar tips did not inhibit expansion of the cell envelope. The phenotype of cells carrying the sefAl allele suggests a disturbed relationship among protoplasm expansion, envelope growth, and septum formation. It is thought that the blunt constrictions observed are caused by an inability of the two septal peptidoglycan layers to fuse during an early stage of septation. In most gram-negative bacteria, cell division is accomplished by an ingrowth of the cytoplasmic membrane and the peptidoglycan layer of the cell wall. These structures may form a partial or complete septum before cell separation occurs (5, 6, 21). In Escherichia coli B and B/r, complete septa are formed, whereas such structures are rarely observed in E. coli K-12 (5, 6, 17). It is therefore still a matter of controversy whether the latter strain divides by constriction or by septation. Mutants of E. coli K-12 containing the envA or envC allele form complete septa. These septa are composed of cytoplasmic membrane and two layers of probably fused peptidoglycan (17, 19, 23). It is possible that also in wild-type E. coli K-12 opposed peptidoglycan layers in partial septa are fused, but that they cannot be demonstrated due to the activity of autolysins (5, 6). Both conditional and nonconditional bacterial mutants showing a defective cell division without measurable impairment of deoxyribonucleic acid (DNA) synthesis have been isolated (1, 4, 9, 13, 20, 22, 25-28, 30). These mu-

tants fall tentatively into three groups: (i) septum initiation-defective mutants, (ii) septum formation-defective mutants, and (iii) septum separation-defective mutants. Most of the conditional cell division mutants isolated thus far belong to the first group. At restrictive temperatures they all form nonseptated filaments. In this group we would expect to find lesions in functions related to the preparation or triggering of the septum. At the present time, only two conditional E. coli K-12 mutants of the second type (30) have been described which at the nonpermissive temperature retain visible constrictions. The mutated alleles seemed to affect septum assembly. However, it was not concluded whether the frozen constrictions observed represented partial or complete septa. The nonconditional envA and envC mutations lead to a deficiency in septum separation, and mutants containing these alleles belong to group iii (17, 23). Conditional E. coli mutants in this group have not yet been reported. The envA mutation also mediates an increased penetration of drugs through the outer 401

402

J. BACTERIOL.

NORMARK ET AL.

membrane (11, 15, 19). Suppression of septum formation and physiological induction of filaments also diminish high permeability (16, 19). Therefore, we assumed that mutations affecting septum formation might be found among antibiotic-resistant mutants of a strain carrying the envA mutation. In this paper we describe the properties of septum formation-defective mutants isolated as rifampin-tolerant clones of strain D22 (envA).

cytidine 5'-triphosphate, guanosine 5'-triphosphate, and uridine 5'-triphosphate, each at 0.2 M; potassium phosphate [to inhibit the polynucleotide phosphorylase], 0.4 mM) was added to the preincubated extracts. The activity of [3H]uridine 5'-triphosphate was 33,000 dpm/nmol. Incubation was for 15 min at 37°C; the reactions were terminated by the addition of 5% trichloroacetic acid. The precipitates were chilled, collected on membrane filters (type HA 0.25, Millipore Corp., Bedford, Mass.), washed with 5% trichloroacetic acid, dried, and counted in toluene-based scintillaMATERIALS AND METHODS tion fluid by using an Intertechnique (Dover, N.J.) Bacterial strains. The source and genotype of scintillation counter. The activity of the extracts each bacterial strain are listed in Table 1. The was dependent upon added template. Electron microscopy. The preparation of cells for strains are all derivatives of E. coli K-12. Media and growth conditions. The complete me- electron microscopy was as described previously dium used was ML broth containing (per liter): (17). Transduction and mating procedure. Transducyeast extract, 5 g; tryptone, 10 g; NaCl, 5 g; pH 7.3. The minimal medium was medium E (29) supple- tion with phage Plbt was by the method of Erikssonmented with 0.2% glucose, 1 ,ug of thiamine per ml, Grennberg (10). Transfer of episome F104 was performed as deand 50 ,ug of the required L-amino acid per ml. Solid media contained 1.5% agar. Growth was scribed previously (15). The interrupted-mating prorecorded with a Zeiss spectrophotometer at a wave- cedure was carried out in ML broth as described length of 450 nm. Temperature shift experiments earlier (18). Determinations of cell numbers, polar tips, and were performed by dilution into flasks containing constrictions. Samples of cells were chilled immedipreincubated medium. Materials. D-Ampicillin (a-aminobenzylpenicil- ately and fixed in 0.05% formaldehyde. Cell number lin) was kindly provided by AB Astra, Sodertalje, was determined with a Petroff-Hausser counter. A Sweden. Nalidixic acid was obtained from Winthrop chain of cells, as well as a filament, was always Ltd., Surbiton on Thames, England. Rifampin was counted as one cell. The frequency of polar tips and kindly provided by Lepetit, Milano, Italy. Chloram- constrictions was determined by direct observation phenicol was purchased from Erco, Stockholm, Swe- under a Zeiss phase-contrast microscope. den. RNA polymerase assays. A 5-g amount of cells taken in the late exponential growth phase was disrupted in an Eaton press, and the ribonucleic acid (RNA) polymerase was purified according to the initial steps of the Burgess procedure (7), except that buffer A was used instead of buffer C in the diethylaminoethyl-cellulose purification step. The fraction obtained after diethylaminoethyl-cellulose chromatography was assayed for the ability to incorporate [3H]uridine 5'-triphosphate into trichloroacetic acid-precipitable material. Samples of enzyme (10 ,l) were preincubated with rifampin for 20 min at 0°C. Thereafter, a complete assay mixture (230 ,l) (containing: the salt mixture of Burgess [7]; calf thymus DNA, 20 ,ug/ml; adenosine 5'-triphosphate,

RESULTS

In strain D22 (envA), reversion to wild-type rifampin tolerance is obtained at an incidence of about 5 x 10-1 per viable D22 clone. The majority of these mutants have regained the wild-type properties of parental strain D21 and probably represent true revertants in the envA locus. However, at an incidence of about 10-8 per viable D22 clone, spontaneous rifampintolerant mutants may be isolated which, in addition, show a marked dependence upon this drug. A large number of such apparently rifampin-

TABLE 1. Strains of E. coli K-12 Strain

D21 D22 D2201

D2213 D2101 Gllal HfrH F104/AB2463

Source

Boman et al. (3) Normark et al. (18) Rifampin-resistant mutant of strain D22 Rifampin-resistant mutant of strain D22 sefAl transductant of strain D21 Boman et al. (3) Hayes (12) Low (14)

Mating type

Genotype

FFF-

proA trp his str ampAl proA trp his str ampAl envAl proA trp his str ampAl envAl sefAl

F-

proA trp his str ampAl envAl sefA13 trp his strA ampAl sefAl ilv metB ampAl

FHfr Hfr

F'

proA+largE3 his-4 thr-1 leu-6proA2 recA13

403

SEPTUM FORMATION-DEFECTIVE MUTANT

VOL. 128, 1976

dependent mutants have been isolated; the strain studied most intensively has been D2201. The phenotype ofthis strain depends upon both the envA mutation and a new mutation in a locus denoted sefA. Certain characteristics of double mutant D2201 are listed in Table 2. This strain, when growing in ML broth or minimal glucose medium, forms filaments containing visible constrictions. During prolonged exponential growth at low densities, the mass increase of the cells declines and the filaments eventually lose their ability to form colonies. This lethal, unbalanced growth is considerably more marked at 25°C than at 370C. In a liquid medium, high cell densities and/or the addition of low concentrations of chloramphenicol and rifampin to low-density suspensions suppressed the lethal, unbalanced growth of strain D2201 (Table 2). The population effect on strain D2201 was pronounced. When the absorbance at 450 nm (A,,O) of the culture was kept above 0.4, no inhibition of growth was observed. The cells divided, and in the stationary phase the culture was comprised of virtually normal-sized cells. Strain D2201 (envA sefAl) exhibited similar behavior on solid ML medium. The ability to form colonies was exceedingly low. Colonies appeared predominately in areas where bacteria were deposited more heavily. Addition of rifampin or chloramphenicol to the plates markedly increased the colony-forming ability. The rifampin dependence of strain D2201 (envA sefAl) growing on plates is illustrated in Fig. 1. The following agents tested had no suppressive effect on the lethal, unbalanced growth of strain D2201: tetracycline, phenethyl alcohol, streptolydigin, nalidixic acid, ampicillin, and gentian violet. The molecular action of rifampin is known (8); its suppressive effect on the phenotype of strain D2201 was therefore analyzed in detail. RNA polymerase of strains D22 (envA) and D2201 (envA sefAl). The resistance level of

strain D2201 (envA sefAl) towards rifampin is about 3 ,ug/ml as compared with 0.01 ug/ml for parental strain D22 (envA). Therefore it was necessary to exclude the possibility that the sefAl allele enhanced the rifampin tolerance of the DNA-dependent RNA polymerase. This enzyme, therefore, was partially purified from strain D22 (envA) and its sefAl derivative and tested for sensitivity to rifampin (Fig. 2). The inactivation curves were identical. Moreover, in the double mutant, D2201 (envA sefAl), we found no evidence for stimulation ofthe enzyme activity by low concentrations of rifampin. It thus appears that the sefAl allele in strain 100

0

c

E

0

0

0l

>10

0

0.01

I

0.1

100

10

(jg/ml) FIG. 1. Effect of rifampin on efficiency of colony formation. Overnight cultures of strains D21 (wild type), D22 (envA), and D2201 (envA sefAl) were inoculated into fresh ML broth and grown for three generations. The cultures were diluted in 0.15 M NaCl and spread on plates containing different concentrations of rifampin. The number of colonies that developed within 48 h ofincubation was counted and compared with the total cell number put on the plates. A similar graph was obtained by replacing rifampin with chloramphenicol. However, in the latter case, D2201 colonies were very small. Symbols: 0, D21 (wild type); A, D22 (envA); 0, D2201 (envA sefAl). Rifampicin

TABLE 2. Phenotypic characteristics of mutant D2201 (envA sefAl) Strain

Genotype

Division abnormality

/ e7,

Dz2

Wild type envA

D2201

envA sefAl 2T/Z ZZZZZZ3

D21

a

b

Outer membrane permeabilitya

Lethal, unbalanced growth

Division recoveryb

Rif

CM

PEA

High

ML broth

Minimal medium

Normal

No

No

Increased

No

No

No

No

Yes

Yes

Normal

Yes

Yes

Yes

Yes

No

Yes

Outer-membrane permeability was measured by the uptake of gentian violet (11). Abbreviations: Rif, rifampin; CM, chloramphenicol; PEA, phenethylalcohol (16).

density

404

J. BACTERIOL.

NORMARK ET AL.

X

a 70

\

._

70-

*

40

00

Z20 so30 0

10

0

0m

05

0.1

0.5

1.0

5.0

Rifampicin concentration ( W,n-)

FIG. 2. Effect of rifampin on RNA polymerase activity in vitro. The enzyme was partially purified as described in the text. Symbols: 0, strain D22 (envA); *, strain D2201 (envA sefAl); E, strain D22011 envA sefAl rifr).

peared virtually normal in the stationary phase. The growth of such small cells was followed in dilute ML suspensions with or without the addition of rifampin (Fig. 3). In the presence of rifampin, growth resumed at a stage later than that in the untreated culture. After 60 min of incubation, the growth rates for treated and untreated cells were virtually the same. Whereas growth continued in the treated culture at an approximately constant rate, it gradually declined among the rifampin-deprived cells. Cells in the latter culture were able to constrict. However, this ability decreased during incubation. After 3 h of incubation, the frequency of visible constrictions diminished without an accompanying increase in cell number. During the entire experiment, D2201 cells were able to undergo, on an average, one division per cell. In the rifampintreated culture, constrictions appeared slightly later than those in the control culture. However, at least part of these constrictions were able to proceed, within a short time span, to cell 0.2

9 0.2 a

D2201 does not affect the rifampin-binding site on the RNA polymerase. However, it could not be excluded that the suppressive effect of rifampin on strain D2201 was due to some reaction distinct from an interaction with the RNA polymerase. A rifampinresistant mutant (D22011) of strain D2201 was therefore isolated, which contains a totally rifampin-resistant RNA polymerase (Fig. 2). The colony-forming ability of strain D22011 on ML plates was increased markedly by a low concentration of chloramphenicol (1 ug/ml) but not with rifampin. Thus, the rescue effect of rifampin upon strain D2201 (envA sefAl) is due to an interaction with the RNA polymerase. Moreover, the increased rifampin tolerance of the double mutant as compared with that of the envA strain, D22, is most likely due to permea-

bility differences. Growth and division of strain D2201 (envA sefAl) and the effect of rifampin. In the presence of low concentrations of rifampin, D2201 cells reached a constant cell length even in very dilute suspensions. Removal of the drug caused an almost immediate reduction in the rate of cell separation. If the rifampin-deprived cells were kept at a low density, long filaments emerged and growth was eventually inhibited (19). Cells of strain D2201 grown in ML medium at higher densities underwent division and ap-

D

4

..

4

0.1

0.I

Q05

a05

10 50

0.9

0.8 20

0.7 /

10

0.6

'V. 'I

E .

£

I,

5

*

II

N

0/

Q3

/

la

2

0.4

N

4' I.1

I

0.5

*

0

ijg.§

0.2 0

A.. ..h

0.1 0

60

120 Time

180

240

of incubation

300

360

420

(mimr

FIG. 3. Cell division and formation of constrictions in strain D2201 (envA sefAl) during growth in ML medium with or without rifampin (2 pg/ml). The mutant was grown overnight in ML plus rifampin (2 mg/ml). The small cells were inoculated into two flasks, one containing ML and the other containing ML plus rifampin. The A450, cell number, and number of constrictions per cell were followed. The relative cell number was obtained by considering the initial cell number as 1 and calculating the cell number as if the culture was not diluted. Open symbols, ML medium; filled symbols, ML plus rifampin.

VOL. 128, 1976

SEPTUM FORMATION-DEFECTIVE MUTANT

separation. Even in the presence of rifampin, the increase in cell number was lagging behind the mass increase, leading to progressively longer cells. A balanced situation was obtained after 5 h of exponential growth. It was of interest to study in detail the effect of rifampin upon septum formation. A growth experiment was performed, therefore, in glucose minimal medium. Small, stationary-phase D2201 cells were inoculated into fresh medium (Fig. 4). After one mass doubling, the culture was divided and rifampin (0.5 j,g/ml) was added to one portion. Rifampin had no effect upon the initiation of the first generation of constrictions in the population. In the absence of the drug, these constrictions did not lead to cell separation. Instead, a second generation of constrictions was formed before an increase in cell number was noticed. The addition of rifampin, however, enabled the first generation of constrictions to proceed to cell separation. Therefore, it is suggested that rifampin suppresses the phenotype mediated by the envA sefAl mutations at the level of septum formation. Genetical analysis of the sefAl and sefA13 alleles. Strain D2201 contains two mutated genes that affect cellular division: envA and sefA, respectively. A priori it was not possible to predict the phenotype of the sefAl allele alone. Strain D2201 (envA sefAl) forms colonies on ML plates at 25°C at a very low efficiency. We assumed that this cold sensitivity was due to the sefAl allele. To determine the approximate location of the sefAl allele, different F' factors were introduced into strain D2201, with scoring for coldresistant offspring. Such clones were obtained by the F' factor F104 carrying the thr-proA region of the E. coli chromosome (14). Furthermore, in interrupted-mating experiments with strains HfrH and Gllal as the donors, the allele of strain D2201 determining cold sensitivity was located 1 to 3 min anticlockwise from proA (Fig. 5). By transduction analysis it was shown that strain D2201 still contained the envA mutation at 2.1 min (14) (Table 3, cross 1). It was expected, therefore, that Sef+ transductants of strain D2201 would exhibit the EnvA phenotype. Such transductants were obtained in a FIG. 4. Effect of rifampin on septum formation and cell division in strain D2201 (envA sefAl) during growth in minimal glucose medium. Strain D2201 (envA sefAl) is able to form colonies on minimal glucose plates. The cells in these colonies are small. Such small cells were inoculated into fresh minimal glucose medium. After approximately one mass doubling, the culture was divided and rifampin

405

I 9

a

.

-

II a

I

I

Ic0

I Time of Incubation (min) (0.5 ,g/ml) was added to one part. The low rifampin concentration used was due to the fact that D2201 cells growing in minimal glucose medium are more rifampin sensitive than those growing in broth. Different growth parameters were followed. It should be stressed that prolonged exponential growth (more than 9 h) causes lethal, unbalanced growth of D2201 cells in minimal glucose medium. Open symbols: Minimal glucose medium; filled symbols, minimal glucose medium plus 0.5 lig of rifampin per ml.

J. BACTERIOL.

NORMARK ET AL.

406

cross with strain Gllal as the donor, selecting ML broth plates at a very low efficiency. Like for Pro+. Among 68 Pro+ transductants, 8.5% those of the double mutant, these colonies apshowed the same antibiotic sensitivity as peared in areas where bacteria had been deposstrains carrying the envA allele (Table 3, cross ited more densely. When an ML culture of 2). No transductants with the wild-type pheno- D2101 (sefAl) cells was shifted to 25°C, cell type were obtained. To construct an envA+ separation was inhibited when the density of sefAl strain, transduction was performed with the culture was kept low (Fig. 6). However, Pro+ transductants of strains D2201 (donor) and constrictions unable to proceed to cell separaD21 (recipient). Among 182 transductants tion were formed. After incubation at 25°C for 4 tested, 3 were found to be cold sensitive (Table h, the temperature was increased to 37°C. The 3, cross 3). One such transductant, D2101 growth rate increased without a measurable (envA+ sefAl), was chosen for further studies. lag period and was paralleled by an increase in By the introduction of F104, cold-resistant cell number. However, no burst in division acoffspring have been obtained for all rifampin- tivity was observed. Instead, the number of dependent, cold-sensitive mutants isolated thus cells longer than 20 um increased slightly after far. Cotransduction analyses of one independ- the shift. An increase in the mean cell mass ently isolated allele, sefA13, also showed link- was also observed during the first 30 min. After age to proA (Table 3, crosses 4 and 5). This fact, the shift, the cell population consisted of filatogether with the low incidence of mutants ments, many of which showed polar constric(10-8/viable D22 clone), suggests that all muta- tions, as well as small cells, whose frequency tions isolated by this procedure only affect one increased during incubation. cistron that is clearly distinct from envA. When chloramphenicol (100 ug/ml) was Septum formation and cell separation in added at the time of the shift to 37°C, a slight strain D2101 (sefAW) after temperature shift. increase in cell number was observed. This Strain D2101 (sefAl), like strain D2201 (envA increase correlated with a decrease in the aversefAl), at low cell densities formed filaments age number of constrictions per cell, suggesting with visible constrictions. However, in ML that a fraction of constrictions induced at 25°C broth at 37°C, no lethal phase was observed. At was able to go to completion at 37°C despite 250C, strain D2101 (sefAl) formed colonies on inhibition of protein synthesis. Cells of strain D2101 (sefAl) were able to induce at least one constriction after a shift to 200! 25°C. It could be argued that these constrictions i were initiated before the shift. Therefore, nali~ ~~~~~~~~~~.// HI i2O,, too 2 ,p dixic acid, at a concentration of 20 ug/ml, was I I added to a culture of D2101 (sefAl) cells immeIi diately after a shift to 25°C (Fig. 7). This treatment caused a considerable decrease in the forTr, "mOlng (nin) mation of visible constrictions, suggesting that FIG. 5. Two separate interrupted-mating experi- strain D2101 (sefAl) was able to initiate, but ments in which strains HfrH (Hayes) and Gllal were not complete, septa at the nonpermissive dencrossed with strain D2201 (envA sefAl). Selection for sity and temperature. cold-resistant recombinants was done at room temWhen D2101 cells were shifted to 256C, a perature on ML broth plates containing streptomycin (100 pg/ml). proA+ recombinants were obtained on clear zone was observed at the time the first minimal medium plates containing streptomycin constriction was formed in the cell population. (100 ,ug/ml). These clear zones will be referred to as polar

Ar

? * ~,.,w,Vo,,

I

~~~~~ /

O

a

. D2 It

l Gn

a

o

to

o

20

10

20

lo

of

TABLE 3. Transduction experiments with the envAl sefAl and sefA13 alleles, using phage PlbtP Cross

Donor

Recipient

Selected phenotype

No. of transduc- Cotransduction between altants tested

leles

leu+-envAl (22.5) 79 Leu+ E64-11 D2201 proA+-sefA+ (8.5) 68 Pro+ D2201 Gllal proA +-sefAI (1.6) D22012 D21 182 Pro+ 4 102 proA+-sefA+ (18) Gllal D2213 Pro+ 5 proA+-sefA13 (1.1) 275 D21 Pro+ D22131 a Scoring for envA transductants was done as described previously (15). Inability to form colonies at 25°C on ML broth plates was taken as evidence for the presence of the sefAl and sefA13 alleles. Pro+ transductants were scored for on minimal medium plates containing tryptophan and histidine. Strains D22012 and D22131 were proA+ transductants of strains D2201 and D2213, respectively. 1 2 3

VOL. 128, 1976

SEPTUM FORMATION-DEFECTIVE MUTANT

407

strains was not restricted to 25°C but could also be observed at 37°C (19). Morphological characterization of strains

I-

D2201 (envA sefAl) and D2101 (sefAl). Strain D22 (envA), when grown in broth, forms chains in which individual cells are separated by a complete septum, i.e., cytoplasmic membrane and two layers of probably fused peptidoglycan (17). In the double mutant D2201 (envA sefAl), constrictions, but no complete septa, were observed. In contrast to constrictions observed in wild-type E. coli, those of strains D2201 and D2101 were often characteristically blunt (Fig.

0

1:

v

U)

Tim (min) FIG. 6. Double-temperature-shift experiment with strain D2101 (sefAl). Strain D2101 was grown in ML broth at 37°C. The culture was shifted to 25°C by dilution. After 4 h at 25°C, the culture was diluted with prewarmed ML medium with or without chloramphenicol (100 pg/ml). The A450, cell number, and frequency of constrictions per cell were then determined. The relative mean cell mass was given as the ratio (A4501cell count) x 10-9. Symbols: *, ML broth culture followed during the entire experiment; El, part of the culture remaining at 37°C; 0, ML containing 100 Ag of chloramphenicol per ml; A, number of cells longer than 20 pm.

tips. They were always located in one ofthe two poles. During prolonged incubation at 250C, the frequency of tip formation declined, suggesting that the observed separation of protoplasm from the wall was a reversible process. D2101 cells treated with nalidixic acid after one mass doubling exhibited polar tips at a much lower frequency as compared with the control cells Tirme after addition of nalidixic acid (min) (Fig. 7). However, polar tips eventually ap7. FIG. Effect of nalidixic acid on polar tip formapeared in the nalidixic acid-treated culture, but tion and constriction ability of strain D2101 (sefAl) only after inhibition of mass increase. When after a shift to 25°C. Strain was grown for four septation at 25°C was inhibited by low concen- generations at 37°C in MLD2101 broth. At zero time, the trations of ampicillin, no inhibition of polar tip cells were shifted to 250C by dilution. To one part, formation was observed. nalidixic acid (20 jig/ml) was added. The A450, cell Polar tips were noticed also in double mutant number, frequency of polar tips, and number of conD2201 (envA sefAl). Tip formation in both strictions per cell were followed.

408

NORMARK ET AL.

J. BACTERIOL.

8 and 9). This was most pronounced in centrally located constrictions. In many cases it was not possible to delimit a constriction region from the remaining part of the envelope (Fig. 10). Polar constrictions (Fig. 8 and 11), probably younger, greatly resembled those commonly observed in the wild-type strain. Polar tips were observed in strains D2201 (envA sefAl) and D2101 (sefAl) at both 37 and 2500. These polar tips were caused by a separation of the cytoplasmic membrane and the protoplasm from the peptidoglycan layer and the outer membrane. In these tips no adhesion sites between the wall and cytoplasmic membrane could be discerned (Fig. 8, 10, and 12). An additional, characteristic feature of strain D2201 (envA sefAl), as well as of strain D2101 (sefAl), was aggregates of cytoplasmic material. This material, in most instances, was located either between segregated chromosomes at the cellular poles or at constriction sites (Fig. 9 and 10). The aggregates we're considerably less abundant in D2201 cells grown in the presence of rifampin (2 pug/ml). The biochemical nature of this material is not known at the present time. DISCUSSION By the procedure used, i.e., selection for rifampin tolerance, only one phenotypic class of septum formation-defective mutants was obtained. The incidence of mutants (10-8/viable D22 clone) and the genetic data suggest that only one locus may be involved. Interruptedmating experiments and transduction analysis indicate a location for sefAl and sefA13 close to 4 min on the linkage map (2). Except for ftsC, no mutations affecting cellular division have been localized to this region ofthe chromosome (22). The possible relationship between ftsC and the sefAl allele studied here has not been tested by complementation analysis.

The mutant strain, D2201 (envA sefAl), as well as the transductant, D2101 (sefAl), may constrict but, unlike the envA mutant, no complete septa are formed. In strain D22 (envA), the peptidoglycan from each daughter cell appears to be fused in the septum (17). We believe that, in both the wild type and in strains carrying the envA allele, septal peptidoglycan is fused, but that in the latter strains a septum peptidoglycan-splitting activity is lower than normal (H. Wolf-Watz and S. Normark, manuscript in preparation). In strains D2201 (envA sefAl) and D2101 (sefAl), blunt constrictions were observed. During growth, the frequency of formation of these constrictions declined in the absence of cell separation (Fig. 3), suggesting that they were altered during growth, and eventually they were impossible to observe. It is possible that, in cells containing the sefAl allele, the two septal peptidoglycan layers are not fused during the early stage of invaginationt, thdrefore, septum material would be inserted parallel rather than vertical to the length axis, creating a blunt constriction later transformed to a shallow girdle. We propose that high cell densities or additions of low concentrations of drugs like rifampin cause a very modest shift-down situation, enabling the septum peptidoglycan layers to fuse. This, in turn, induces partial septa that are able to proceed to cell separation. This idea is supported by the fact that rifampin added very late in the cell cycle promotes cell separation of induced constrictions. Moreover, partial septa were seen, although rarely, in cells of strain D2201 (envA sefAl) grown in the presence of rifampin. The polar tips observed represented a separation of cell mass and cytoplasmic membrane from the peptidoglycan layer and outer membrane. The polar tips appeared regularly when cells carrying the sefAl allele formed constrictions. They did not inhibit expansion of the rod.

'I_

FIG. 8. Electron micrograph ofstrain D2101 (sefAl) grown in ML medium at 37C. Note the characteristically blunt central invagination (double arrows). An early stage of constriction is also seen (single arrow). It does not differ markedly in appearance from those of wild-type strains. A typical polar tip is seen at the right bacterial pole.

VOL. 128, 1976

SEPTUM FORMATION-DEFECTIVE MUTANT

409

FIG. 9. Blunt constriction at late stage in strain D2201 (envA sefAl) grown in ML medium at 25°C. No tendency towards separation of the envelope layers is seen. No septal structure whatever is present. Note the presence ofelectron-dense, fairly homogeneous material that extends through the cytoplasmic bridge between the cells and infringes on the ribosomal mass.

Furthermore, the separation of protoplasm from the wall was a reversible process. If the cell envelope grows at a linear rate, with doublings in rate in relation to cell division (24, 31), then tip formation may be the result of an increase in tube expansion not compensated for by the exponential expansion of protoplasm. Septum material inserted parallel rather than vertical to the length axis may be the direct cause of polar tip formation. Nalidixic acid probably inhibits septum formation by blocking formation of new growth zones (31). The fact that nalidixic acid prevented polar tip forma-

tion in growing cells supports this hypothesis. The sefAl allele affected not only septum formation, but also septum initiation. Prolonged exponential growth in dilute suspensions therefore resulted in filaments containing one or two shallow constrictions. The interpretation of this finding depends on whether the septum represents a unique structure or whether it solely represents processed surplus envelope material (25, 31). In the latter case, one mutation could very well affect both septum initiation and septum formation. In the former case, on the other hand, one would ex-

t

f1/I t I

I,

I

.--bk

A_

FIG. 10. Survey electron micrograph of strain D2201 (envA sefAl) grown in ML medium at 25°C. Central cell regions are diminished in width but do not exhibit any clear constrictions. Unilateral polar tips are common, as are aggregates ofdense, homogeneous material (arrows), the latter, as a rule, associated with the cytoplasmic membrane.

FIG. 11. Cell of strain D2101 grown in ML medium at 37°C. Two markedly different division regions are depicted. To the right (single arrow) is seen a shallow constriction with no tendency towards septum formation. To the left (double arrows) an apparently normal division site is present. 410

SEPTUM FORMATION-DEFECTIVE MUTANT

VOL. 128, 1976

.,

411

A. t.

I a

4

*i. 4,-

I

.vf X

NNW. a

)

k.

.

4.4

.z. d*4'SW .

-.

j t

s.*

.4,.

S.

.4.I

im; -AF---.

:

*I~

-

FIG. 12. High-magnification electron micrograph of a polar tip in strain D2101 (sefAl) grown in ML medium at 37°C. Note the complete separation between the cytoplasmic membrane and the peptidoglycan layer of the cell wall. The space between these structures contains a fairly electron-dense, finely flocculant material of unknown nature.

pect to find mutants in which each potential division site gives rise to a frozen constriction. The isolation of such mutants has not yet been reported. The primary defect in sefAl -containing cells is unknown. The list of hypothetical mechanisms for formation of constrictions rather than septa is long. It is possible that the aggregates of cytoplasmic material observed in cells carrying the sefAl allele are related to the deficiency in septum formation. A characterization of this material, therefore, may cast some light on the gene product coded for by sefA.

general de la Recherche Scientifique et Technique and the Swedish Natural Science Research Council (Dnr B3373003).

LITERATURE CITED 1. Allen, J. S., C. C. Filip, R. A. Gustafsson, R. G. Allen. and J. R. Walker. 1974. Regulation of bacterial cell

2. 3.

ACKNOWLEDGMENTS We wish to express our gratitude to Marianne Borg and Monica Grahn for skillful technical assistance. The stimulating discussions with Pierre Fromageot, Andre Sentenac, Hans G. Boman, and Hans Wolf-Watz are gratefully acknowledged. This work was supported by grants from the Direction

4.

5.

division: genetic and phenotypic analysis of temperature-sensitive, multinucleate, filament-forming mutants ofEscherichia coli. J. Bacteriol. 117:978-986. Bachmann, B. J., K. B. Low, and A. L. Taylor. 1976. Recalibrated linkage map of Escherichia coli K-12. Bacteriol. Rev. 40:116-167. Boman, H. G., K. G. Eriksson-Grennberg, S. Normark, and E. Matsson. 1968. Resistance ofEscherichia coli to penicillins. IV. Genetic study of mutants resistant to D,L-ampicillin concentrations of 100 ,ug/ml. Genet. Res. 12:169-185. Breakefield, X. O., and 0. E. Landman. 1973. Temperature-sensitive divisionless mutant of Bacillus subtilis defective in the initiation of septation. J. Bacteriol. 113:985-998. Burdett, I. D. J., and R. G. E. Murray. 1974. Septum formation in Escherichia coli: characterization of sep-

412

NORMARK ET AL.

tal structure and the effects of antibiotics on cell division. J. Bacteriol. 119:303-324. 6. Burdett, I. D. J., and R. G. E. Murray. 1974. Electron microscope study of septum formation in Escherichia coli strains B and B/r during synchronous growth. J. Bacteriol. 119:1039-1056. 7. Burgess, R. 1969. A new method for the large scale purification of Escherichia coli deoxyribonucleic acid dependent ribonucleic acid polymerase. J. Biol. Chem. 244:6160-6167. 8. Burgess, R. R. 1971. RNA polymerase. Annu. Rev. Biochem. 40:711-740. 9. Ciezla, Z., M. Bagdasarian, W. Szezurkiewiez, M. Przyg6nska, and T. Klopotowski. 1972. Defective cell division in thermosensitive mutants of Salmonella typhimurium. Mol. Gen. Genet. 116:107-125. 10. Eriksson-Grennberg, K. G. 1968. Resistance of Escherichia coli to penicillins. fl. An improved mapping of the ampA gene. Genet. Res. 12:147-156. 11. Gustafsson, P., K. Nordstrom, and S. Normark. 1973. Outer penettation barrier of Escherichia coli K-12: kinetics of the uptake of gentian violet in wild type and envelope mutants. J. Bacteriol. 116:893-900. 12. Hayes, W. 1964. The genetics of bacteria and their viruses. Blackwell Scientific Publishers, Oxford. 13. Hirota, Y., A. Ryter, and F. Jacob. 1968. Thermosensitive mutants of E. coli affected in the process of DNA synthesis and cellular division. Cold Spring Harbor Symp. Quant. Biol. 33:677-693. 14. Low, K. B. 1972. Escherichia coli K-12 F-prime factors, old and new. Bacteriol. Rev. 36:587-607. 15. Normark, S. 1970. Genetics of a chain-forming mutant of Escherichia coli. Transduction and dominance of the envA gene mediating increased penetration to some antibacterial agents. Genet. Res. 16:63-78. 16. Normark, S. 1971. Phenethyl alcohol as a suppressor of the envA phenotype associated with the envA gene in Escherichia coli K-12. J. Bacteriol. 108:51-58. 17. Normark, S., H. G. Boman, and G. D. Bloom. 1971. Cell division in a chain-forming envA mutant ofEscherichia coli K-12. Fine structure of division sites and effects of EDTA, lysozyme and ampicillin. Acta Pathol. Microbiol. Scand. Sect. B 79:651-664. 18. Normark, S., H. G. Boman, and E. Matsson. 1969. Mutants ofEscherichia coli with anomalous cell division and ability to decrease episomially and chromosomally mediated resistance to ampicillin and several other antibiotics. J. Bacteriol. 97:1334-1342.

J. BACTERIOL. 19. Normark, S., and H. Wolf-Watz. 1974. Cell division and permeability of unbalanced envelope mutants of Escherichia coli K-12. Ann. Microbiol. (Paris) 125:211-226. 20. Reeve, J. N., D. J. Groves, and D. J. Clark. 1970. Regulation of cell division in Escherichia coli: characterization of temperature-sensitive division mutants. J. Bacteriol. 104:1052-1064. 21. Reyn, A., A. Birch-Andersen, and M. Berger. 1970. Fine structure and taxonomic position of Neisseria haemolysans (Thjotta and Boe, 1938) or Gimella haemolysans (Berger, 1960). Acta Pathol. Microbiol. Scand. Sect. B 78:375-389. 22. Ricard, M., and Y. Hirota. 1973. Process of cellular division in Eacherichia coli: physiological study on thermosensitive mutants defective in cell division. J. Bacteriol. 116:314-322. 23. Rodolakis, A., P. Thomas, and J. Starka. 1973. Morphological mutants of Escherichia coli. Isolation and ultrastructure of a chain-forming envC mutant. J. Gen. Microbiol. 75:409-416. 24. Sargent, M. G. 1975. Control of cell length in Bacillus subtilis. J. Bacteriol. 123:7-19. 25. Slater, M., and E. Schaechter. 1974. Control of cell division in bacteria. Bacteriol. Rev. 38:199-221. 26. Stone, A. B. 1973. Regulation of cell division in a temperature-sensitive division mutant of Escherichia coli. J. Bacteriol. 116:741-750. 27. Van Alsynte, D., and M. J. Simon. 1971. Division mutants of Bacillus subtilis: isolation and PBS1 transduction of division-specific markers. J. Bacteriol. 108:1366-1379. 28. Van de Putte, P., J. van Dillewin, and A. Rorsch. 1964. The selection of mutants ofEscherichia coli with impaired cell division at elevated temperatures. Mutat. Res. 1:121-128. 29. Vogel, H. J., and D. M. Bonner. 1956. Acetylornithinase ofEscherichia coli: partial purification and some properties. J. Biol. Chem. 218:97-106. 30. Walker, J. R., A. Kovarik, J. S. Allen, and R. A. Gustafsson. 1975. Regulation of bacterial cell division: temperature-sensitive mutants of Escherichia coli that are defective in septum fornation. J. Bacteriol. 123:693-703. 31. Zaritsky, A., and R. H. Pritchard. 1973. Changes in cell size and shape associated with changes in the replication time of the chromosome of Escherichia coli. J. Bacteriol. 114:824-837.