JOURNAL OF BACTERIOLOGY, Jan., 1965Copyright © 1965 American Society for Microbiology

Vol. 89, No. 1Printed in U.S.A.

Time-Lapse Photomicrography of Cell Growthand Division in Escherichia coli

HEINER HOFFMAN AND MICHAEL E. FRANK

Department of Microbiology, New York University, College of Dentistry, New York, New York

Received for publication 26 August 1964

ABSTRACT

HOFFMAN, HEINER (New York University, New York, N.Y.), AND MICHAEL

E. FRANK. Time-lapse photomicrography of cell growth and division in Escherichia coli.J. Bacteriol. 89:212-216. 1965.-Photomicrographs at 15-see intervals of cells growingat 37 C disclosed that in a cell with a generation time of 21.0 miin the processes of fur-rowing, cross-wall formation, and cell separation are completed within 2.5 min after thedivision furrow first becomes clearly visible. Among a large number of cultivationsexamined, only a few cells late in one microculture at 43.5 C failed to separate once thecross wall was completed. Measurements of cell lengths during a 5-min period, extendingfrom just before to just after division, showed that elongation of the cell is a discon-tinuous process, although the growth rate over the 5-min period is exponential. At thetime of cell division, it appears that the synthesis of cell-wall material is diverted en-

tirely into formation of the cross wall.

The introduction of phase-contrast microscopygreatly improved the possibilities for the directcytological investigation of single-cell growth ofliving bacteria in relation to cytokinesis. Never-theless, very few studies of this problem havebeen carried out. Among the better known earlierattempts is that of Mason and Powelson (1956),who were primarily interested in nuclear divi-sion. They took photomicrographs at 2-min inter-vals, with bright phase-contrast optics, of micro-cultures on a gelatin medium wAith a refractiveindex designed to disclose greater internal detailthan seemed possible with agar. Although well-defined "nuclear figures" were found in Esch-erichia coli cells, the cell wall could not be seen,and Mason and Powelson were unable to establishwhen the transverse septum completely separatedthe sister cells. MIore recently, Schaechter et al.(1962) employed a modified version of the tech-nique in a study which timed the period requiredfor division of bacterial nuclei (nuclear interdivi-sion time). Photographs were taken at intervalsranging between 1 and 3 min. From measure-ments of cell lengths, these authors concludedthat E. coli cells increased in length at all timesduring the division cycle without noticeablediscontinuities.

In the present investigation, we have re-ex-amined the problem of cell growth in relation tocytokinesis in E. coli by means of dark phaseoptics rather than bright. The study was carriedout primarily on the basis of one micr-ocolonywhich was photographed at 15-see intervals and

a second which was photographed at 3-min inter-vals. Analysis of the photomicrographs revealedsome previously undescribed aspects of the timerequired for certain stages in the division process,but failed to confirm some of the observations ofSchaechter et al. (1962) in regard to cell growth.

MATERIALS AND METHODS

The orgainism used was rough-phase E. coliATCC 8677. The Fleminig microculture technique,in which a dry film of smeared broth culture on acover slip is overlaid with agar, was used forgrowing the microcoloniies. A 37 C "rapid-se-quence" cultivationi was photographed at 15-seciintervals; the record of this cultivation has beenused in an earlier paper (1lhoffman and Frank,1964b), which describes some details of the tech-niques used (see also Hoffman and Frank, 1961,1963a). The second cultivation, photographed at3-min intervals, was incubated at 43.5 C in a Zeissheatiing stage rather thani in the microscopeinicubator box used for the first culture. Measure-meiits of cell lenigth were made with a draftsman'scaliper to the closest 0.5 mm from photographicprints at 7,OOOX for the rapid-sequenice cultivationand at 3,750X for the 43.5 C cultivation.

RESULTS

Rapid-sequence cultivation at 37 C. The centercell in Fig. 1 is in the fourth generation, and had ageneration time of 21.0 min. TNenty cells werepresent in the microcolony at this time, and thecenter cell in Fig. 1 was well in the middle of them.The first visual indication that the cell was under-

212

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

25

Nov

embe

r 20

21 b

y 79

.165

.97.

43.

CELL GROWTH AND DIVISION IN E. COLI

going division was the presence of a well-defineddark body in the center of the cell. The origin of

1.oo0 this central body could not be traced with cer-tainty in the earlier photomicrographs. The ini-tial phase of division itself was marked by theappearance of a slight division furrow (Fig. 1),

2 which gradually developed. Within 1.5 min (Fig.2), it was so fully developed that the cell was

1.50' noticeably constricted at the center. Within thenext minute (Fig. 6), the dark central body ap-parently split into two parts separated by anarrow transverse band which was not as dark

3 as the adjoining central bodies. The time atwhich this band appeared was the point which we

75/ arbitrarily took as the completion of division.1 75 Thus, division was completed for the described

cell within 2.5 min after the slight indication offurrowing in Fig. 1. The proximal ends of the

4 newly formed daughters then gradually roundedoff and became more clearly defined (Fig. 7 and

2.00' 8). The two granules resulting from the splittingof the central body in the mother remained aspolar bodies in the daughters. The daughter cellsthen began to move apart, and were well sepa-

5 rated (Fig. 10) within 2.75 min after the trans-verse wall was formed. However, a strand of what

2.25/ may be capsular material still extended betweenthe proximal ends of the daughters.

Caliper measurements from time-lapse photo-micrographs (Fig. 1 to 10) of the lengths of the

6 cell and its 2 daughters at 15-sec intervals duringthe period from shortly before to shortly after

250 ' division indicated that there are intervals duringthis period when growth in length does not occur.The longest period of such apparent growthsuspension in the cell was found in an interval

7 extending from 0.75 min before completion ofdivision to 0.75 min after. Similar observations

3.00 were obtained from measurements of an addi-tional dozen cells in this microcolony. Neverthe-less, plotting the length of the central cell in each15-sec photographic frame over a 5-min period,

8 including the time required for cell division, indi-cated that growth in the direction of the longi-

as50 tudinal axis of the cell was exponential for the5-min period as a whole (Fig. 11).

M31icroculture at 43.5 C photographed at 3-minintervals. This microcolony was described as C61

9 in an earlier report (Hoffman and Frank, 1963a).With the exception of heat-induced filaments,

4.501 the cells in this microcolony appeared quitenormal in their growth and division behaviorthrough the greater part of the cultivation. Inthe ninth generation, however, cell growth slowed

K s''"w sF -- .. w RF __|l- 't_ w>_ >

i.. ...w_is, a>.__w

Wa_ ^J ^

_F ... wj._K_ -___ s. _ ._ .

__ s. |_.. .__-_si._s

xi__ 5

10

5.00 1

FIG. 1 to 10. Microculture at 37 C, dark phase-contrast microscopy, photographed at 15-sec inter-vals. 7,OOOX. Selected representative frames over a5-min period to show division of the centrally locatedcell (fourth generation).

213VOL. 89, 1965

Ak.D

ownl

oade

d fr

om h

ttps:

//jou

rnal

s.as

m.o

rg/jo

urna

l/jb

on 2

5 N

ovem

ber

2021

by

79.1

65.9

7.43

.

HOFFMAN AND FRANK

35

34

33

E

E 32-c

3C

J 30

0 290

Cell

Division

I/

28

27

26

25 . I0 I 2 3 4 5

MinutesFIG. 11. Growth of the centrally located cell and

its daughters over the 5-min period shown in Fig. 1to 10. Cell lengths are plotted in millimeters as readdirectly from the enlarged prints of the photomicro-graphs.

down considerably or stopped completely, al-though division continued. In most cases, thedaughter cells in a sibling pair easily separatedfrom each other after division (Fig. 12 to 21), theprocess apparently taking about as much timeas in the early phases of cultivation. A few cellswere found, however, which developed a cross

septum (Fig. 14), but the resulting daughters didnot separate from each other during the following21-min period to the end of the photomicro-graphic record (Fig. 15 to 21).

DISCUSSION

The quality of the photomicrographs obtainedis satisfying for the purposes of the present study,but the amount of detail of the intracytoplasmicstructures visible in the photomicrographs was

not as great as had been hoped for in view of thehigh resolution of Kodak high-contrast copyfilm and the very fine quality of the optical lensesused. One of the reasons for this result is theprobable occurrence of some slight movement of

FIG. 12 to 21. Microculture at 43.5 C, darkphase-contrast microscopy, photographed at 3-minintervals. 3,750X. Cells are in the ninth generation;encircled cell divided in Fig. 14, but the daughtersare still closely attached 21 min later (Fig. 21) atthe end of the photomicrographic record; cell enclosedby square exhibits normal separation followingdivision.

12 Aw__

O2

15 '_

I.

17 _212s,

24i19_27'__._:..

20 130_/ s

21 __ F m33 ~ ~

214 J. BACTEBRIOL.

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

25

Nov

embe

r 20

21 b

y 79

.165

.97.

43.

CELL GROWTH AND DIVISION IN E. COLI

the cellular organelles during the relatively longfilm exposure time of 5 to 6 see. An additionaland more significant reason involves the differ-ence in refractive index between the cytoplasmand the enclosed structures. No attempt wasmade in the present study to use the technique ofMason and Powelson (1956) to increase contrastby adding gelatin to the mounting medium, sinceMason and Powelson had found that the cell wallin E. coli was not revealed in this way, and sincethey also had been unable to establish the time ofappearance of the transverse cell wall. Difficultieswith the technique have been reported by othersalso (Adams, 1963; Brieger, 1963).The cytological details of division visible in the

light microscope have interested a number ofrecent investigators concerned with bacterialgeneration times. Powell (1958) distinguishedthree cell generation times, the intervals betweensuccessive divisions of nuclei, cytoplasm, andcell wall. He utilized cell-wall division as the endpoint in determining generation times. Quesnel(1963) took the time at which light appearedbetween formative sister pairs as the extremityof the period required for completion of cell-walldivision. Our own observations indicate that thispoint can be determined with greater certaintythan any other aspect of bacterial cell divisionvisible in the light microscope.The photomicrographic record of cell division

which was obtained from the rapid-sequencecultivation is apparently the first time-lapse studyto be presented with short intervals subjected toframe-by-frame analysis. From an extensivesearch of the literature, we have found only asingle report (Bisset, 1939) on the length of timerequired by a rod form to pass through the divi-sion process itself. Bisset (1939), observingBacillus mycoides with an oil immersion lens,found that the time required for the completionof division after the appearance of an extremelyfine line across the organism was 5 to 15 min.The generation times of these cells were not given.The central body, which precedes the optical

appearance of the transverse cell wall, may wellbe the intracytoplasmic membranous organellewhich Imaeda and Ogura (1963) had associatedwith the formation of cell wall during divisionof mycobacteria. Up to the present, however, itmust be acknowledged that attempts to demon-strate these organelles in E. coli through electronmicroscopy have not been entirely successful.Studies by Conti and G-ettner (1962) and Ed-wards and Stevens (1963) failed to demonstratethem, although Kaye and Chapman (1963)found some rare examples in control cells andmore prominent and numerous examples in cellstreated with colistin sulfate.

Only a single microcolony was uncovered, theculture at 43.5 C, in which we found that theappearance of the transverse cell wall was notpromptly followed by a separation of the daughtercells. Even in this case, only a few cells failed toseparate, and this was late in the cultivationwhen the colony was obviously exhibiting verylittle growth. No cases were found, in this care-fully examined microcolony of three- or four-member chains or filaments with cross walls,which would have indicated further divisionswithout separation. These points are emphasizedsince, on the basis of cell-wall staining, it has beenclaimed (Bisset, 1955) that rough-phase eubac-teria are normally four-celled.

Particularly interesting were the findings onfluctuations of growth in cell length, especially inrelation to cell division, since the observationsand discussions by previous investigators of theseproblems did not encourage the experimentalapproach taken. Schaechter et al. (1962) hadfound from their measurements of individualE. coli cells, made from projections of photo-micrographs taken at intervals ranging between1 and 3 min during growth, that E. coli cellsincreased in length at all times during the divisioncycle, without noticeable discontinuities. Thediscrepancy with our results may perhaps be dueto the longer interval between photomicrographsin the study by Schaechter et al. Furthermore,Mitchison (1961) rejected rod-shaped bacteriain his study of cell growth of single bacteria onthe grounds that, since bacterial rods are usuallymultinucleate, there is opportunity for variationdepending upon the degree to which the cellularunits are out of phase with each other in theirnuclear or cell cycles. Our data do not seem tohave reflected this possibility. In addition to theconsistent agreement of the described centralcell with the measurements of the other dozencells in the rapid-sequence recorded microcolony,we have also established (Hoffman and Frank,1964a) that microcolonies cultivated by thetechniques developed in our laboratory exhibit ahigh degree of synchrony of division throughoutthe course of the cultivation.

Fluctuations in cell length very similar to thosenoted in the present study may be found inTable 1 of an earlier report (Hoffman and Frank,1963b), which indicates that cell elongation issuspended during the pinching off of an endgranule. Here also, the cell quickly catches up tothe length of its sister after the granule has beenseparated from the cell.The observation that the cell length does not

increase during the period that the cross wallapparently is being formed indicates that the

VOL. 89, 1965 215

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

25

Nov

embe

r 20

21 b

y 79

.165

.97.

43.

HOFFMAN AND FRANK

resources for synthesizing cell-wall material arediverted entirely into formation of the cross wall.

ACKNOWLEDGMENTS

This investigation was supported by PublicHealth Service grants DE 01462-02 and DE01462-03 from the National Institute of DentalResearch.

LITERATURE CITED

ADAMS, J. N. 1963. Nuclear morphogenesis duringthe developmental cycles of some members ofthe genus Nocardia. J. Gen. Microbiol. 33:429-435.

BISSET, K. A. 1939. The mode of growth of bac-terial colonies. J. Pathol. Bacteriol. 48:427-435.

BISSET, K. A. 1955. The cytology and life-historyof bacteria, 2nd ed., p. 86. E. & S. LivingstoneLtd., Edinburgh.

BRIEGER, E. M. 1963. Structure and ultrastructureof microorganisms, p. 77. Academic Press,Inc., New York.

CONTI, S. F., AND M. E. GETTNER. 1962. Electronmicroscopy of cellular division in Escherichiacoli. J. Bacteriol. 83:544-550.

EDWARDS, M. R., AND R . W. STEVENS. 1963. Finestructure of Listeria monocytogenes. J. Bacteriol.86:414-428.

HOFFMAN, H., AND M. E. FRANK. 1961. Form andinternal structure of cellular aggregations inearly Escherichia coli microcultures. J. Gen.Microbiol. 25:353-364.

HOFFMAN, H., AND M. E. FRANK. 1963a. Tempera-ture limits, genealogical origin, developmentalcourse, and ultimate fate of heat-induced

filaments in Escherichia coli microcultures. J.Bacteriol. 85:1221-1234.

HOFFMAN, H., AND M. E. FRANK. 1963b. Time-lapse photomicrography of the formation of afree spherical granule in an Escherichia colicell end. J. Bacteriol. 86:1075-1078.

HOFFMAN, H., AND M. E. FRANK. 1964a. Celldivision and generation times in an Escherichiacoli microcolony. Bacteriol. Proc., p. 33.

HOFFMAN, H., AND M. E. FRANK. 1964b. "Ger-mination tube" growth in Escherichia colimicrocultures. J. Bacteriol. 88:1151-1154.

IMAEDA, T., AND M. OGURA. 1963. Formation ofintracytoplasmic membrane system of myco-bacteria related to cell division. J. Bacteriol.85:150-163.

KAYE, J. J., AND G. B. CHAPMAN. 1963. Cytologicalaspects of antimicrobial antibiosis. III. Cyto-logically distinguishable stages in antibioticaction of colistin sulfate on Escherichia coli.J. Bacteriol. 86:536-543.

MASON, D. J., AND D. M. POWELSON. 1956. Nucleardivision as observed in live bacteria by a newtechnique. J. Bacteriol. 71:474-479.

MITCHISON, J. M. 1961. The growth of singlecells. III. Streptococcus faecalis. Exp. Cell Res.22:208-225.

POWELL, E. 0. 1958. An outline of the pattern ofbacterial generation times. J. Gen. Microbiol.18:382-417.

QUESNEL, L. B. 1963. A genealogical study ofclonal development of Escherichia coli. J. Appl.Bacteriol. 26:127-151.

SCHAECHTER, M., J. P. WILLIAMSON, J. R. HOOD,JR., AND A. L. KOCH. 1962. Growth, cell andnuclear divisions in some bacteria. J. Gen.Microbiol. 29:421-434.

216 J. BACTERIOL.

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

25

Nov

embe

r 20

21 b

y 79

.165

.97.

43.