JOURNAL OF BACTERIOLOGY,0021-9193/00/$04.0010

Mar. 2000, p. 1419–1422 Vol. 182, No. 5

Copyright © 2000, American Society for Microbiology. All Rights Reserved.

Effects of Calcium and Calcium Chelators on Growth andMorphology of Escherichia coli L-Form NC-7

T. ONODA,1 J. ENOKIZONO,1 H. KAYA,1 A. OSHIMA,1* P. FREESTONE,2 AND V. NORRIS3

Department of Biological Science, Faculty of Life and Environmental Science, Shimane University, Matsue 690, Japan1;Department of Microbiology and Immunology, University of Leicester, Leicester LE1 9HN, United Kingdom2; and

IFR ‘Systemes Integres’, Laboratoire de Microbiologie, Faculte des Sciences et Techniques de Rouen,F76821 Mont Saint Aignan Cedex, France3

Received 19 July 1999/Accepted 18 November 1999

Growth of a wall-less, L-form of Escherichia coli specifically requires calcium, and in its absence, cells ceaseddividing, became spherical, swelled, developed large vacuoles, and eventually lysed. The key cell divisionprotein, FtsZ, was present in the L-form at a concentration five times less than that in the parental strain. Oneinterpretation of these results is that the L-form possesses an enzoskeleton partly regulated by calcium.

Numerous roles for calcium in bacteria are now becomingapparent (10, 12, 19). We have proposed that calcium has arole as a “general reset” in cells (12) and that it participates inthe regulation of the putative bacterial “enzoskeleton” (14).This enzoskeleton would comprise proteins such as the tubu-lin-like protein, FtsZ, which is the key player in cell divisionand which in vitro has a calcium-stimulated polymerization andGTPase activity (24). L-forms are wall-less derivatives of bac-teria that grow and divide despite their lack of a normal pep-tidoglycan sacculus (7, 8, 15, 18, 22). This means that themorphology and progress through the cell cycle of L-formsmust result from forces acting via some structure other thanthe sacculus. Membrane domains have been considered can-didates for such structures (7), and these probably result fromthe coupled transcription, translation, and insertion (transer-tion) of proteins into and through membranes (1), processesthat generate sufficient force to hold L-forms together (13).The L-form NC-7, which is a derivative of an Escherichia coliK-12 strain (16), possesses a secondary calcium/proton anti-porter (17) and reveals a general inhibition of growth followingaddition of the calcium chelator EGTA (15 [but also see ref-erence 23]). NC-7 is therefore an ideal model system for ex-ploring the hypothesis of an enzoskeleton controlled by cal-cium.

Effects of divalent ions on growth. First, the identity of theL-forms derived from E. coli (16) was confirmed by PCR am-plification of ftsZ, hisS, and orf80 to obtain products of theexpected Mr (20), sequencing of 210 bp of the glnA gene(cloned at random), and N-terminal sequencing of Dps, YfiD,and the E1 component of the pyruvate dehydrogenase complex(4). Then, to study the effects of divalent ions, cells werepreincubated in modified Na-Davis medium plus 1 mM EGTAand 2 mM ionophore A23187 (to equilibrate internal and ex-ternal calcium levels) for 3 h, harvested in the exponentialphase of growth by centrifugation (1,000 3 g for 10 min),washed once with growth medium, and resuspended in modi-fied Na-Davis medium containing either 0.2 or 0.5 mMBAPTA [1,2-bis(2-aminophenoxy)ethane-N,N,N9,N9-tetraace-tic acid]. Modified Na-Davis medium contains 0.7 g ofK2HPO4, 0.2 g of KH2PO4, 1 g of (NH4)2SO4, 0.5 g of sodium

succinate, 2 g of peptone, 1 g of yeast extract, 2 g of glucose,0.34 M NaCl, 105 U of penicillin G, and 50 mM (each) FeCl2,ZnCl2, MgCl2, CaCl2, and MnCl2. Cells were agitated at 45rpm and 32°C, and plastic tubes were used throughout. Fol-lowing preincubation, cells were transferred to fresh mediumcontaining 0.5 mM metal chelator BAPTA and 50 mM con-centrations of Fe, Zn, Mg, and Mn. Under these conditions,growth was inhibited, and only addition of a higher concentra-tion of 1.2 mM calcium was sufficient to restore growth (Table1). Similar results were obtained in other experiments using 0.2mM BAPTA and a 1 mM concentration of single ion species aswell as in experiments with iron, cobalt, copper, and zinc chlo-ride (data not shown). Addition of only 50 mM calcium, suffi-cient to release a trace element, was not sufficient to restorenormal growth (data not shown). It should be noted that pre-incubation of cells in both A23187 and EGTA is essential ifgrowth inhibition by EGTA is to be reversed specifically bycalcium (rather than divalent ions in general). These experi-ments strongly indicate that calcium is required for the growthof L-forms.

Effect of EGTA on morphology. L-forms were grown inNaPY medium containing per liter: 10 g of peptone, 5 g ofyeast extract, 2 g of glucose, 0.34 M NaCl, and 105 U ofpenicillin G. NaPY medium contains 0.12 mM Ca21, 12.7 mMFe21, 8.1 mM Zn21, and 1.5 mM Mn21, as determined by flamespectrophotometry. The pH in the medium was adjusted to 7.2with NaOH. Phase-contrast microscopy revealed that cells aredifferent shapes and sizes (Fig. 1A). Some images may simplycorrespond to deformations that have no functional role, whileothers must also correspond to cell division, which is oftenasymmetric and involves budding (Fig. 1A). On transfer tomedium containing 1 mM EGTA, most cells became spherical,and the average volume increased 1.5 times during incubation(Fig. 1B). This increase is not due to fusion of cells, since thenumbers of CFU did not reveal a concomitant decrease duringthis period (data not shown) and since fusing and dividing cellswere not observed. Up to 6 h, these morphological changeswere completely reversible, and the 1 mM EGTA-treated,spherical cells reverted to the polymorphic form 1 to 2 h afteraddition of 2 mM calcium (Fig. 1C). In the absence of calcium,however, continued incubation in 1 mM EGTA led to theformation of what appeared to be vacuoles inside cells, re-duced viability, and, finally, lysis (Fig. 1D).

The structure of the L-forms. To investigate further thestructure of the L-forms and the morphological changes result-

* Corresponding author. Mailing address: Department of BiologicalScience, Faculty of Life and Environmental Science, Shimane Univer-sity, Matsue 690-0823, Japan. Phone: 81 852 32 6443. Fax: 81 852 326449. E-mail: oshima@life.shimane-u.ac.jp.

1419

on July 10, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

ing from EGTA addition, L-forms growing exponentially inNaPY medium containing 1 mM calcium were harvested, pre-incubated in NaPY medium containing 1 mM EGTA and 2mM A23187 for 1 h, and then transferred to fresh medium

containing 1 mM EGTA. Scanning electron microscopy wasperformed on freeze-fractured cells (Fig. 2A). At the timesindicated, cells were harvested by centrifugation at 5,000 3 gfor 5 min and washed once with 0.067 M phosphate buffercontaining 0.75 M KCl. Washed cells were resuspended in asmall amount of 0.067 M phosphate buffer containing 0.75 MKCl, transferred to a paper filter (Whatmann 3MM; 5 by 5mm) and fixed by the osmium-tannic acid-osmium method(21). Specimens were dehydrated with ethanol in increasingconcentrations, dried in a critical point dryer (Hitachi HCP-2),coated with Pt-Pd in an ion spatter device (Hitachi H102), andanalyzed by scanning electron microscopy (Hitachi S-800). In-tracellular structures were observed by a combination of thechitosan embedding and the osmium dimethyl sulfoxide-os-mium methods (5). In cells at the start of the experiment, acoralline structure with a dense, granular surface filled thecytoplasm (Fig. 2A, panel 1). Cell division appeared to occur ina variety of symmetrical and asymmetrical ways. Buds of some-times very different sizes were observed separated by long

FIG. 1. Phase-contrast micrographs of the L-form treated with EGTA. Cells growing exponentially in NaPY medium containing 1 mM calcium were harvested andtransferred to fresh medium with the following additions: none (A), 1 mM EGTA for 6 h (B), 1 mM EGTA for 6 h followed by 2 mM calcium for 1.5 h (C), or 1 mMEGTA for 36 h (D). The bar represents 5 mm.

TABLE 1. Effects of divalent ions on L-form growth yieldsa

Ion addition(1.2 mM)

L-form growth yield(OD600)

None...................................................................................... 0CaCl2..................................................................................... 0.21FeCl2 ..................................................................................... 0.01ZnCl2..................................................................................... 0.01MgCl2.................................................................................... 0.00MnCl2.................................................................................... 0.01

a Cells were precultured in modified Na-Davis medium containing 1 mMEGTA and 2 mM A23187 and then transferred to fresh medium (containing 0.5mM BAPTA, 1.2 mM the ion shown, and 50 mM the other divalent ions for 24 h).Optical densities at 600 nm (OD600) were determined.

1420 NOTES J. BACTERIOL.

on July 10, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

necks in which no clear septum was visible (Fig. 2B). In somedividing cells, the first stage of budding could be seen, and thisoften appeared to involve a future daughter about 1 mm indiameter forming from a parental cell about 3 mm in diameter(Fig. 2B). There are many small spherical objects around 300nm in diameter that are probably the lysed remains of mem-branes (Fig. 2B). After 12 h in EGTA medium (Fig. 2A, panels2, 3, and 4), the structure of the cells was different, and assuggested by light microscopy (Fig. 1D), large vacuoles hadformed. These vacuoles, which were up to 5 mm in diameter atthe 12-h stage, were much larger than those that were some-times seen in the cells grown in the presence of free calcium,and there were often several of them in each cell. The forma-tion of vacuoles (6, 8) and structures resembling microtubules(3) have been reported in L-forms of E. coli and other bacteriaas well as paracrystalline inclusion bodies adjacent to the mem-brane and “stiff, nontubular cores” (6). While such cytoplasmiccores were not observed in this study, a network of filaments,possibly adjacent to the membrane, did appear to be present insome cells at the start of the experiment (data not shown). Thepolymerization of FtsZ into a ring-like structure associatedwith the cytoplasmic membrane is considered the key step incell division in bacteria. In vitro, this polymerization can bestimulated by calcium. FtsZ is an evident component of anenzoskeleton, and we speculated that a substantial increase inthe level of FtsZ in the L-form might confer a structural sta-bility that would be dependent on calcium. The L-form and itsparental strain were therefore grown under identical condi-tions, and immunoblot experiments were performed at a rangeof protein concentrations using a 1:4,000 dilution of anti-FtsZpolyclonal antibodies (generously given by Miguel Vicente), a1:4,000 dilution of antirabbit horseradish peroxidase-conju-gated secondary antibody (Sigma), and enhanced chemilumi-nescence (Amersham) with typical exposure times of 1 min.L-form extracts were not centrifuged after sonication, sincesignificant amounts of FtsZ are associated with L-form mem-

brane. Surprisingly, densitometry revealed that FtsZ levelswere fivefold lower per unit of protein in the L-form than inthe parental strain (data not shown).

It has been proposed that the peptidoglycan sacculus hasbeen replaced in L-forms by the macromolecular componentsof the cytoplasm that can act as structural components (7).Indeed, L-forms offer the possibility of revealing an enzoskel-eton that is masked in wild-type bacteria by the sacculus. Thisenzoskeleton would comprise equilibrium and nonequilibriumhyperstructures, some of which would be regulated by calcium(9). In the latter case, hyperstructures are assemblies of pro-teins, membranes, and nucleic acids, each responsible for aparticular function such as sugar transport or cell division (11),as others have also proposed (2). It is therefore conceivablethat the absence of a normal wall in the L-form leads to ageneral reduction in expression in the 2-min cluster where ftsZlies, perhaps via a reduction in transertion. The consequentlylow level of FtsZ may be too low for division to occur efficientlyin the L-form, as evidenced perhaps by its varied patterns ofcell division.

We thank Nanne Nanninga for FtsZ protein, Miguel Vicente forantibodies to FtsZ, Kathryn Lilley and Janette Maley for technicalassistance, and Susan Grant and Istvan Toth for encouragement.

We also thank the BBSRC and the EU for support.

REFERENCES

1. Binenbaum, Z., A. H. Parola, A. Zaritsky, and I. Fishov. 1999. Transcription-and translation-dependent changes in membrane dynamics in bacteria: test-ing the transertion model for domain formation. Mol. Microbiol. 32:1173–1182.

2. Buddelmeijer, N., M. E. G. Aarsman, A. H. J. Kolk, M. Vicente, and N.Nanninga. 1998. Localization of cell division protein FtsQ by immunofluo-rescence microscopy in dividing and nondividing cells of Escherichia coli. J.Bacteriol. 180:6107–6116.

3. Eda, T., Y. Kanda, and S. Kimura. 1976. Membrane structures in stableL-forms of Escherichia coli. J. Bacteriol. 127:1564–1567.

4. Freestone, P., S. Grant, M. Trinei, T. Onoda, and V. Norris. 1998. Protein

FIG. 2. Scanning electron micrographs of the L-form. (A) Freeze-fractured cells after growth in NaPY medium (panel 1) and after a 12-h incubation in NaPYmedium containing 1 mM EGTA (panels 2, 3, and 4). (B) Cells after growth in NaPY medium. Each bar represents 1 mm.

VOL. 182, 2000 NOTES 1421

on July 10, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

phosphorylation in Escherichia coli L-form NC-7. Microbiology 144:3289–3295.

5. Fukudome, H., and K. Tanaka. 1986. A method for observing intracellularstructures of free cells by scanning electron microscopy. J. Micros. 141:171–178.

6. Gumpert, J. 1983. Ultrastructural characterization of core structures andparacrystalline inclusion bodies in L-form cells of streptomycetes. Z. Allg.Mikrobiol. 23:625–633.

7. Gumpert, J. 1992. Cellular growth without a murein sacculus—the nucleoid-associated compartmentation concept, p. 453–463. In M. A. de Pedro, J.-V.Holtje, and W. Loffelhardt (ed.), Bacterial growth and lysis. Metabolism andstructure of the bacterial sacculus. Plenum Press, New York, N.Y.

8. Lederberg, J., and J. St. Clair. 1958. Protoplasts and L-type growth ofEscherichia coli. J. Bacteriol. 75:143–160.

9. Norris, V., S. Alexandre, Y. Bouligand, D. Cellier, M. Demarty, G. Grehan,G. Gouesbet, J. Guespin, E. Insinna, L. Le Sceller, B. Maheu, C. Monnier,N. Grant, T. Onoda, N. Orange, A. Oshima, L. Picton, H. Polaert, C. Ripoll,M. Thellier, J.-M. Valleton, M.-C. Verdus, J.-C. Vincent, G. White, and P.Wiggins. 1999. Hypothesis: hyperstructures regulate bacterial structure andthe cell cycle. Biochimie 81:915–920.

10. Norris, V., M. Chen, M. Goldberg, J. Voskuil, M. McGurk, and I. B. Holland.1991. Calcium in bacteria: a solution to which problem? Mol. Microbiol.5:775–778.

11. Norris, V., P. Gascuel, J. Guespin-Michel, C. Ripoll, and M. H. Saier, Jr.1999. Metabolite-induced metabolons: the activation of transporter-enzymecomplexes by substrate binding. Mol. Microbiol. 31:1592–1595.

12. Norris, V., S. Grant, P. Freestone, J. Canvin, N. F. Sheikh, I. Toth, M. Trinei,K. Modha, and R. I. Norman. 1996. Calcium signalling in bacteria. J. Bac-teriol. 178:3677–3682.

13. Norris, V., T. Onoda, H. Pollaert, and G. Grehan. 1999. The mechanical

origins of life. Biosystems 49:71–78.14. Norris, V., G. Turnock, and D. Sigee. 1996. The Escherichia coli enzoskel-

eton. Mol. Microbiol. 19:197–204.15. Onoda, T., and A. Oshima. 1988. Effects of Ca21 and a protonophore on

growth of an Escherichia coli L-form. J. Gen. Microbiol. 134:3071–3077.16. Onoda, T., A. Oshima, S. Nakano, and A. Matsuno. 1987. Morphology,

growth and reversion in a stable L-form of Escherichia coli K12. J. Gen.Microbiol. 133:527–534.

17. Onoda, T., H. Shinjou, and A. Oshima. 1989. Cation/proton antiport systemsin Escherichia coli K12, L-form NC-7. Mem. Fac. Sci. Shimane Univ. 20:69–76.

18. Paton, A. M. 1987. L-forms: evolution or revolution? J. App. Bacteriol.63:365–371.

19. Smith, R. J. 1995. Calcium and bacteria. Adv. Microb. Physiol. 37:83–103.20. Sweeney, S. T., P. Freestone, and V. Norris. 1995. Identification of novel

phosphoproteins in Escherichia coli using the gene-protein database. FEMSMicrobiol. Lett. 127:133–138.

21. Takahashi, G. 1978. OsO4-tannin-OsO4 fixation and staining of biologicalspecimens for electron microscopy. J. Electron Microsc. 27:66.

22. Waterhouse, R. N., H. Buhariwalla, D. Bourn, E. J. Rattray, and L. A.Glover. 1996. CCD detection of lux-marked Pseudomonas syringae pv. phase-olicola L-forms associated with Chinese cabbage and the resulting diseaseprotection against Xanthomonas campestris. Lett. App. Microbiol. 22:262–266.

23. Youatt, J. 1994. The toxicity of metal chelate complexes of EGTA precludesthe use of EGTA buffered media for the fungi Allomyces and Achlya. Mi-crobios 79:171–185.

24. Yu, X.-C., and W. Margolin. 1997. Ca21-mediated GTP-dependent assemblyof bacterial cell division protein FtsZ into asters and polymer networks invitro. EMBO J. 16:5455–5463.

1422 NOTES J. BACTERIOL.

on July 10, 2020 by guesthttp://jb.asm

.org/D

ownloaded from