Vol. 171, No. 6

Negative Control of Cell Division by mreB, a Gene That Functionsin Determining the Rod Shape of Escherichia coli Cells

MASAAKI WACHI AND MICHIO MATSUHASHI*

Institute of Applied Microbiology, The University of Tokyo, Bunkyo-ku, Tokyo, Japan

Received 5 December 1988/Accepted 4 March 1989

Exponentially growing Escherichia coli cells containing additional copies of the shape-determining gene mreBwere found to be elongated, whereas mreB mutant cells were spherical and overproduced penicillin-bindingprotein 3, a septum peptidoglycan synthetase. The effect of the mreB gene on expression offtsl, the structuralgene for penicillin-binding protein 3, was examined by using an ftsl-lacZ fusion gene on a plasmid. Formationof ,I-galactosidase from the fusion gene was significantly increased in mreBl29 mutant cells, and itsoverproduction was suppressed to a normal level by the presence of a plasmid containing the mreB gene. Theseresults indicate a negative mechanism of control of cell division by this morphology gene and suggest that thegene functions in determining whether division or elongation of the cells occurs.

Two groups of genes forming clusters mrd (12) at 14.5 minand mre (13) at 71 min on the Escherichia coli chromosomemap are known to function in determining the shape of E.coli cells. Mutants of genes of both clusters are spherical andshow altered sensitivities to amidinopenicillin mecillinam.Products of the genes in the mrd cluster are penicillin-binding protein (PBP) 2, which is a peptidoglycan synthetaseand a target of mecillinam, and the RodA protein, whichprobably functions together with PBP 2 (5). These proteinsare thought to function in a very early step of cell elongation.The region of the mre cluster contains the mreB gene, whichcodes for a 37-kilodalton protein, and other genes thatencode at least three other proteins (3). We previouslyreported that a deletion encompassing these four codingframes causes an increase in the amount ofPBP 1B, which isthought to be a rather general peptidoglycan synthetase, andPBP 3, a septum peptidoglycan synthetase (13).The shape-determining genes are thought to have a nega-

tive influence on cell division and a positive function inelongation of the cell into the correct rod shape. This notionis supported by the recent observation of Begg and others (1;A. Takasuga, K. J. Begg, W. D. Donachie, T. Ohta, and H.Matsuzawa, Seikagaku 60:716, 1988), who reported suppres-

sion of a mutation offtsl by a point mutation in the codingframe of the RodA protein. This paper provides evidence fornegative control of cell division by the mreB gene.

MATERIALS AND METHODSBacterial strains and media. The properties of E. coli K-12

strain PA340 (F- argHl thr-J leuB6 gdh-J hisG-I gItB31thi-l lacYl gal-6 xyl-7 ara-14 mtl-2 malAl rpsL9 tonA2) andits mre derivatives PA340-129 (the same as PA340 but gItB+mreB129) and PA340-678 (the same as PA340 but gltB+A&mre-678) were described in a previous report (13). StrainMC1061 [araD139 A(ara leu)7697 AlacX74 galU galK hsrhsm+ rpsL], used for construction of plasmid pFIM1, was

obtained from H. Matsuzawa, University of Tokyo, Tokyo,Japan. Cells were grown in Lennox broth (6) supplementedwith 20 mg of thymine and 100 jig of lipoic acid per liter(L'-Lip broth). This broth was used to.grow both originalmre lip mutants and their mre derivatives. Cells containingplasmids were grown in L'-Lip broth supplemented with 50

* Corresponding author.

mg of aminobenzylpenicillin per liter, 25 mg of kanamycinper liter, or both.

Plasmid construction. Plasmid pMEL6 (Fig. 1) was con-

structed by ligation of the 2.1-kilobase (kb) HinclI fragmentof pMEL1 (13) into low-copy-number plasmid pLG339 (11).The 1.5-kb SmaI-HincII fragment (8; Fig. 2) including thefirst 267 codons of ftsI was inserted into the SmaI site ofpMC1403 (2; Fig. 2) to make an in-frame joint between theN-terminal part of ftsI and the C-terminal part of lacZ.Transformants of MC1061 forming blue colonies on an

L'-Lip agar plate containing 20 mg of 5-bromo-4-chloro-3-indolyl-p-D-galactopyranoside per liter were selected toobtain the correctly oriented plasmid pFIM1 (Fig. 2).Enzyme assays. Cells were cultured at 30°C until the A660

reached 0.5, collected at 4°C, and disrupted in 50 mMsodium phosphate buffer (pH 7.4) by sonication at 10 kHz.P-Galactosidase activity was determined by measuring deg-radation of o-nitrophenyl-,B-D-galactopyranoside as de-scribed by Miller (7). The reaction was carried out in 1 ml ofa mixture of 50 mM sodium phosphate (pH 7.4), 0.5 mg ofo-nitrophenyl-,-D-galactopyranoside, and enzyme at 30°Cand stopped by addition of 1 ml of 1 M sodium carbonate.One unit of 3-galactosidase activity was defined as thatcausing an increase in A410 of 0.001/min at 30°C. ,-Lacta-mase activity was determined by measuring degradation ofnitrocefin as described by O'Callaghan et al. (9). The reac-

tion mixture consisted of 50 mM sodium phosphate (pH 7.4),5 pLg of nitrocefin, and enzyme in a total volume of 0.1 ml.One unit of 1-lactamase activity was defined as that causingan increase in A490 of 1.0/min at room temperature.PBPs were detected fluorographically by binding of

[3H]benzylpenicillin, followed by separation on sodium do-decyl sulfate-polyacrylamide gels as described previously(10, 12).

Microscopic observation of cells. Cells were fixed in 3%formaldehyde, spread on 0.5% agar (L'-Lip broth), andobserved under a phase-contrast microscope with dark-fieldillumination.

Radioactive materials and reagents. [benzyl-4-3H]benzyl-penicillin (N-ethylpiperidinium salt, 65.2 mCi/mg, dissolvedin acetone) was a generous gift from P. Cassidy, MerckSharp & Dohme, Rahway, N.J. Nitrocefin was kindly pro-vided by G. W. Ross, Glaxo Group Research Ltd., Green-

3123

JOURNAL OF BACTERIOLOGY, June 1989, p. 3123-31270021-9193/89/063123-05$02.00/0Copyright © 1989, American Society for Microbiology

on March 16, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

3124 WACHI AND MATSUHASHI

0 1 2 3 4 5 6L I I I

kb

S/H H H H K H S/HI I , I I I

Kmre

KFIG. 1. Construction of plasmid pMEL6. The restriction map

and gene alignment of the mre region are shown at the top.Abbreviations: Km, kanamycin resistance gene; Tc, tetracyclineresistance gene; H, HincII; K, KpnI; S, SaIl.

E

EO 1 2 3 4 10

c U

a)

0 0.1

108

0.010 143 1

Time (hr)FIG. 3. Growth of strain PA340 carrying plasmid pLG339 or

pMEL6. Cells were cultured in L'-Lip broth containing 25 mg ofkanamycin per liter at 37°C with shaking. A660 [0, PA340(pLG339);0, PA340(pMEL6)] was measured in an ADS photometer (FujiKogyo Co., Tokyo, Japan), and cell numbers [O, PA340(pLG339);*, PA340(pMEL6)] were measured microscopically.

ford, England. Other reagents and enzymes for DNA recom-bination experiments were commercial products.

RESULTS

Filamentous cell growth caused by additional copies of themreB gene. The low-copy-number plasmid pMEL6 (Fig. 1),

o 1 2 3 4 5 6 7 kb which carried a 2.1-kb chromosomal fragment containingI I I I I I I I mreB as the sole open reading frame, was introduced into the

El P Sm P H K P HIlIl mreB+ strain PA340 by transformation, and the shape of theI I It% | ' ' transformed cells was monitored throughout growth. PA340

ftsI El SrnB containing the vector plasmid pLG339 was used as a control.SmB_ a The absorbances of both PA340(pLG339) and PA340

SimaI (pMEL6) increased exponentially in the same way for theHi n c p first 2.5 h at 37°C and then slowed (Fig. 3). In the early toIllC//Ap lacZ mid-exponential phase of growth, there were considerably

SmaI fewer cells that carried additional copies of the mreB geneISinaI | pMC1403 [strain PA340(pMEL6)] than did not carry this gene on the

plasmid [strain PA340(pLG339)]. Microscopic examinationM6,Bi la } showed that some of the cells carrying pMEL6 were large

p(H/Sm)Ba lac and filamentous (Fig. 4A). Giemsa staining showed thatthese cells were multinucleated (data not shown). In the late

Sm tsl exponential to early stationary phase, cell division becameEl lacZ greater than cell growth (Fig. 3), and the cells became short

rods (Fig. 4B). These cells appeared to be somewhat swol-len. A small population of cells appeared fairly spread out

pF IM 1 and irregular in shape. Under the same culture conditions,p Ap cells of the strain that did not contain the mreB gene on the

I// plasmid [PA340(pLG339)] grew as normal rods throughout

acY// the culture period (Fig. 4C and D). Similar results wereobtained at 30°C. Curing of the plasmid resulted in recoveryof normal rod-shaped growth. Introduction of pMEL1 (3),

FIG. 2. Construction of plasmid pFIM1. The restriction map of which contained a larger chromosomal fragment than didthe 7.4-kb EcoRI-HindIII fragment containing the ftsl gene is shown pMEL6, had a similar effect (data not shown).at the top. Abbreviations: Ap, ampicillin resistance gene; Ba, Effect of the mreB gene on expression of theftsl-lacZ fusionBamHI; El, EcoRI; H, HincIl; HIII, HindIII; K, KpnI; P, PstI; Sm, gene. We previously observed simultaneous overproductionSmaI. of PBPs 1B and 3 in Amre-678 mutant cells, which had a 5-kb

J. BACTERIOL.

on March 16, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

CELL DIVISION CONTROL IN E. COLI BY mreB 3125

A. .IIhEEm.-mp.

FIG. 4. Change in cell shape of strain PA340 carrying plasmid pMEL6. Cells were cultured at 37°C, and cell shape was observed undera dark-field phase-contrast microscope in the exponential phase of growth [A, PA340(pMEL6); C, PA340(pLG339)] and in the stationaryphase of growth [B, PA340(pMEL6); D, PA340(pLG339)]. Bar, 10 R,m.

deletion of the chromosome encompassing mreB and threecoding frames for the 40-, 22-, and 51-kilodalton proteins inthe mre region (13). We subsequently confirmed that over-

production ofPBP 3 was mainly due to mutation of the mreBgene; the mreB129 mutant cells showed an increased level ofPBP 3 but normal levels of all other PBPs (Fig. 5).To examine the possibility that the mreB gene repressed

the production of PBP 3, we introduced theftsI-lacZ proteinfusion plasmid pFIM1 (Fig. 2) into both the mreB+ strainPA340 and the mre mutants PA340-129 (mreB129) andPA340-678 (Amre-678) and measured the formation of ,B-galactosidase from the fusion gene. The fts-I-acZ fusiongene was expressed constitutively in mreB+ cells, whereasits expression was increased 2.6- to 3.6-fold in the mreB129and Amre-678 mutant cells (Table 1). Overproduction inmreB129 cells was reduced to the level in wild-type cells bythe coexistence of pMEL6 containing the mreB gene. Incontrast, overproduction in Amre-678 deletion mutant cellscould not be suppressed by plasmid pMEL6, which con-

tained the mreB gene but not the entire mre region; suppres-

sion to the wild-type level required the larger plasmidpMEL1, which covered the entire mre region complement-ing the Amre-678 mutation.

Table 1 also shows the level of production of P-lactamasefrom plasmid pFIM1, which was measured as an indicator ofthe copy number of plasmid pFIM1 in the cells tested.Enzyme activities were almost the same in all cells exam-ined.

DISCUSSION

This study provides a new concept of the control of celldivision in E. coli by the shape-determining gene mreB.Expression of additional copies of the mreB gene resulted infilamentous cells, indicating inhibition of cell division. Fur-thermore, in mreB mutant cells, which are spherical, boththe amount of septum peptidoglycan synthetase PBP 3 andthe rate of expression of the cell division geneftsI, encodingthis protein, were higher than in wild-type cells. The mreBgene is known to be responsible for formation of the normal

VOL. 171, 1989

-' "-'

on March 16, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

3126 WACHI AND MATSUHASHI

aalII I I | | | | +5| 6|

FIG ...Ovrrduto of PB 3 Xyteme uain

1_s231 __ _ _

-b ___ _ KI__ _ I_ L_

___-9 i----

densitogram of a fluorogram is shown. (a) PA340; (b) PA340-129(mreB 129); (c) PA340-678 (Amre-6 78).

rod shape of E. coli cells (3). The inhibition of cell divisionby excessive copies of the mreB gene and enhanced expres-sion of the cell division gene ftsI in mreB mutant cells mayindicate a function of the mreB gene as a regulator fordetermining progression to cell division or elongation. How-ever, mreB is certainly not the only gene involved in thisproposed mechanism. A decrease in the amount of PBP 3 inmreB' cells by additional copies of the mreB gene could notbe proved becaus'e of the very small amount of this protein inthe strain used (Fig. 5).The other spherical cells, Amre-678 mutant cells, which

have a large deletion of the chromosome encompassing themreB genle and at least three other coding frames causingoverproduction of PBPs lB and 3 (3, 13), also increased thelevel of expression of the ftsl-acZ fusion gene; this increasecould be suppressed by plasmid pMEL1, which contains theentire mre re'gion, but not by the smaller pMEL6, whichcontains onily the mreB gene. Therefore, more than one ofthe gene products of the mre region are directly involved inthis proposed division-elongation on-and-off control system,or the effects of mutation and deletion on ftsI expression inthe mre region are indirect as a result of changes in cellmetabolism resulting from these mutations.

TABLE 1. Expression of the fts-I-acZ fusion genein mre mutants

Avg enzyme activity" ± SEStrain (U/mg of protein)

P-Galactosidase P-Lactamase

PA340 10.9 ± 0.6 <1.0PA340(pFIM1) 153.1 ± 38.9 267.9 ± 34.4

PA340-129 14.1 ± 1.0 <1.0PA340-129(pMEL6) 12.7 + 0.3 <1.0PA340-129(pFIM1) 386.0 ± 61.2 306.3 ± 14.6PA340-129(pFIM1, pMEL6) 161.5 ± 0.9 351.2 ± 15.7

PA340-678 14.2 ± 2.0 <1.0PA340-678(pMEL6) 14.4 ± 0.2 <1.0PA340-678(pMEL1) 10.6 ± 1.3 <1.0PA340-678(pFIMl) 531.9 + 83.1 229.4 ± 18.4PA340-678(pFIM1, pMEL6) 599.8 ± 143.6 233.2 ± 16.6PA340-678(pFIM1, pMEL1) 117.3 ± 13.7 332.2 ± 51.6

" Averages of experiments using three independent clones.

It is unknown whetherftsl is the only gene affected by themreB product. It is also unknown at what point the mreBgene product inhibits cell elongation. Repression of the Iongene or inhibition of the Lon protein may also be themechanism. The mreB mutant forms spherical cells andshows altered sensitivity to the antibiotic mecillinam, whichspecifically inhibits the transpeptidase activity ofPBP 2. ThemreB product is probably involved as a positive regulator inthe process of cell elongation. However, since the amount ofPBP 2 is not affected by the mreB mutation, the positivecontrol target of the mreB gene may not be simply the genecoding for PBP 2 (mrdA).As reported previously, the C-terminal portion of the

MreB protein shows high homology with the FtsA protein(3), which is thought to be involved in a late step of celldivision (4). Therefore, these two proteins may have regu-latory, though opposing, functions to ensure that rod-shapedcells begin to either elongate or divide at the correct positionon the cell surface and at the appropriate time in the cellcycle.

Preliminary experiments on fractionation of the MreBprotein by ultracentrifugation followed by sarcosyl extrac-tion showed that the MreB protein was located in thecytoplasmic membrane (data not shown). The MreB proteinprobably functions in this domain of the cell. It may bereasonable to assume that this protein functions as a kind ofsensor that recognizes the topology of the cells, determiningthe correct position of the zone for growth or division on thecell surface according to the length or mass of the cells, thatis, the start of cell elongation at the one-unit rod length (ormass) and division at the two-unit length (or mass) of thecells.

Further work is necessary to determine the effect of themreB gene on a number of other genes and the basesequences and functions of the products of other genesinvolved in the mre region. The expression and effects of themrd genes mrdA (coding for PBP 2) and mrdB (coding for theRodA protein) also require further investigation. Previously,Begg et al. (1) reported that a mutation in the mrdB gene(rodA) suppressed the thermosensitivity and filamentationcaused by a mutation atftsl. Since the rod mutation usuallydoes not appreciably affect the amount of PBP 3 in cells, wepresume that this suppression could be due to a very delicate

J. BACTERIOL.

on March 16, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

CELL DIVISION CONTROL IN E. COLI BY mreB 3127

change in the balance of proteins involved in the division-elongation on-and-off mechanism.

ACKNOWLEDGMENTS

This work was supported in part by a Grant-in-Aid for Encour-agement of Young Scientists (no. 62790243 to M.W.), a Grant-in-Aidfor Special Scientific Research (no. 63615505 to M.M.), and a

Grant-in-Aid for Scientific Research (no. 63430020 to M.M.) fromthe Ministry of Education, Science and Culture of Japan. M.W. is a

fellow of the Japan Society for Promotion of Science for JapaneseJunior Scientists.

LITERATURE CITED1. Begg, K. J., B. G. Spratt, and W. D. Donachie. 1986. Interaction

between membrane proteins PBP3 and RodA is required fornormal cell shape and division in Escherichia coli. J. Bacteriol.167:1004-1008.

2. Casadaban, M. J., J. Chou, and S. N. Cohen. 1980. In vitro gene

fusions that join an enzymatically active P-galactosidase seg-

ment to amino-terminal fragments of exogenous proteins: Esch-erichia coli plasmid vectors for the detection and cloning oftranslational initiation signals. J. Bacteriol. 143:971-980.

3. Doi, M., M. Wachi, F. Ishino, S. Tomioka, M. Ito, Y. Sakagami,A. Suzuki, and M. Matsuhashi. 1988. Determinations of theDNA sequence of the mreB gene and of the gene products of themre region that function in formation of the rod shape ofEscherichia coli cells. J. Bacteriol. 170:4619-4624.

4. Donachie, W. D., K. J. Begg, J. F. Lutkenhaus, G. P. C.Salmond, E. Martinez-Salas, and M. Vincente. 1979. Role of theftsA gene product in control of Escherichia coli cell division. J.Bacteriol. 140:388-394.

5. Ishino, F., W. Park, S. Tomioka, S. Tamaki, I. Takase, K.Kunugita, H. Matsuzawa, S. Asoh, T. Ohta, B. G. Spratt, and M.Matsuhashi. 1986. Peptidoglycan synthetic activities in mem-branes of Escherichia coli caused by overproduction of penicil-lin-binding protein 2 and RodA protein. J. Biol. Chem. 261:7024-7031.

6. Lennox, E. S. 1955. Transduction of linked genetic characters ofthe host by bacteriophage P1. Virology 1:190-206.

7. Miller, J. H. 1972. Experiments in molecular genetics, p.352-355. Cold Spring Harbor Laboratory, Cold Spring Harbor,N.Y.

8. Nakamura, M., I. N. Maruyama, M. Soma, J. Kato, H. Suzuki,and Y. Hirota. 1983. On the process of cellular division inEscherichia coli: nucleotide sequence of the gene for penicillin-binding protein 3. Mol. Gen. Genet. 191:1-9.

9. O'Callaghan, C. H., A. Morris, S. M. Kirby, and A. H. Shigler.1972. Novel method for detection of P-lactamases by using achromogenic cephalosporin substrate. Antimicrob. Agents Che-mother. 1:283-288.

10. Spratt, B. G., and A. B. Pardee. 1975. Penicillin-binding proteinsand cell shape in E. coli. Nature (London) 254:516-517.

11. Stoker, N. G., N. F. Fairweather, and B. G. Spratt. 1982.Versatile low-copy-number plasmid vectors for cloning in Esch-erichia coli. Gene 18:335-341.

12. Tamaki, S., H. Matsuzawa, and M. Matsuhashi. 1980. Cluster ofmrdA and mrdB genes responsible for the rod shape andmecillinam sensitivity of Escherichia coli. J. Bacteriol. 141:52-57.

13. Wachi, M., M. Doi, S. Tamaki, W. Park, S. Nakajima-lijima,and M. Matsuhashi. 1987. Mutant isolation and molecular clon-ing of inre genes, which determine cell shape, sensitivity tomecillinam, and amount of penicillin-binding proteins in Esch-erichia coli. J. Bacteriol. 169:4935-4940.

VOL. 171, 1989

on March 16, 2020 by guest

http://jb.asm.org/

Dow

nloaded from