UvA-DARE is a service provided by the library of the University of Amsterdam (http://dare.uva.nl)

UvA-DARE (Digital Academic Repository)

Epitope mapping of Escherichia coli cell division protein FtsZ with monoclonal antibodies

Voskuil, J.L.A.; Westerbeek, C.A.M.; Wu, C-G.; Kolk, A.H.J.; Nanninga, N.

Published in:Journal of Bacteriology

DOI:10.1128/jb.176.7.1886-1893.1994

Link to publication

Citation for published version (APA):Voskuil, J. L. A., Westerbeek, C. A. M., Wu, C-G., Kolk, A. H. J., & Nanninga, N. (1994). Epitope mapping ofEscherichia coli cell division protein FtsZ with monoclonal antibodies. Journal of Bacteriology, 176(7), 1886-1893. https://doi.org/10.1128/jb.176.7.1886-1893.1994

General rightsIt is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s),other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons).

Disclaimer/Complaints regulationsIf you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, statingyour reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Askthe Library: https://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam,The Netherlands. You will be contacted as soon as possible.

Download date: 04 Oct 2020

JOURNAL OF BACrERIOLOGY, Apr. 1994, p. 1886-18930021-9193/94/$04.00+0Copyright © 1994, American Society for Microbiology

Epitope Mapping of Escherichia coli Cell Division ProteinFtsZ with Monoclonal Antibodies

JAN L. A. VOSKUIL,1 CARLA A. M. WESTERBEEK,1 CHUANGING WU,1tAREND H. J. KOLK,2 AND NANNE NANNINGA`*

Section of Molecular Cytology, Institute for Molecular Cell Biology, BioCentrum, University ofAmsterdam,1018 TVAmsterdam,j and N. H. Swellengrebel Laboratory of Tropical Hygiene, Royal Tropical

Institute, 1105 AZ Amsterdam,2 The Netherlands

Received 20 September 1993/Accepted 3 January 1994

A fusion between lacZ andftsZ of Escherichia coli was constructed to obtain a 0-galactosidase-FtsZ fusionprotein. This fusion protein was used to raise antibodies against cell division protein FtsZ. Six monoclonalantibodies were obtained, and they reacted with FtsZ from cytoplasm and membrane fractions. The epitopesin FtsZ were localized by studying the reactions of the monoclonal antibodies with fusion proteins truncatedat the carboxy terminus and with fragments that were obtained by CNBr cleavage of purified FtsZ. Fivedifferent epitopes were defined. Epitopes I and III reacted with the same monoclonal antibody, without showingapparent amino acid homology. Epitope II was defined by monoclonal antibodies that cross-reacted with an

unknown cytoplasmic 50-kDa protein not related to FtsZ. Epitopes IV and V were recognized by differentmonoclonal antibodies. All monoclonal antibodies reacted strongly under native conditions, so it is likely thatthe five epitopes are situated on the surface of native FtsZ. By using these data and computer analysis, a

provisional model of FtsZ is proposed. The FtsZ protein is considered to be globular, with a hydrophobicpocket containing GTP-binding elements. Epitopes I and II are situated on each side of the hydrophobicpocket. Because the carboxy terminus contains epitope V, the carboxy terminus of FtsZ is likely oriented towardthe protein's surface.

Once the chromosomal DNA has doubled and the twonucleoids are separating in Escherichia coli, the cell undergoesa constriction in between the two nucleoids. During thisprocess, there is a local increase in peptidoglycan synthesis,partially at the cost of lateral peptidoglycan synthesis else-where (41). Cell division proteins, such as FtsZ, FtsQ, FtsA,and penicillin-binding protein 3 (PBP3 encoded by ftsI), par-ticipate in the constriction process (5). The ftsQ(Ts), ftsA(Ts),and ftsl(Ts) mutants fail to continue the constriction process,while the ftsZ84(Ts) mutant does not start new constrictions atrestrictive temperatures (33). In these mutants, nucleoid sep-aration is not impaired (2, 42). These observations have led tothe conclusions that the FtsZ protein is involved in theinitiation of the constriction process and that the FtsQ, FtsA,and PBP3 proteins are involved in its continuation (5).At the initiation of the constriction process, FtsZ forms a

ring-like structure at the cytoplasmic side of the invaginatinginner membrane (4). The location of this ring is complementedby a ring-like peptidoglycan-synthesizing activity in theperiplasm (41). The GTPase activity of FtsZ might be involvedin the assembly of the FtsZ ring (11, 24, 28). The partitioningof the nucleoids is also likely to be related to this process: whenthe expression of ftsZ is decreased artificially, nucleoid sepa-ration is impaired (34). Although there is evidence that theexpression of ftsZ is regulated (1, 23, 31, 37, 39) and twostudies demonstrated its dependency on the cell cycle (13, 15),division does not require a fixed number of FtsZ molecules

* Corresponding author. Mailing address: Institute for MolecularCell Biology, University of Amsterdam, Plantage Muidergracht 14,1018 TV Amsterdam, The Netherlands. Phone: (20) 5255187. Fax:(20) 5256271.

t Present address: Department of Experimental Internal Medicine,Academic Medical Centre, 1105 AZ Amsterdam, The Netherlands.

(35). On the other hand, overexpression of ftsZ leads todivisions at the cell poles, and either overexpression offtsZ tostill higher levels (40) or depletion of ftsZ (9, 27) causes

complete inhibition of the constriction process anywhere in thecell. The ratio of the levels of FtsA to FtsZ is essential forproper regulation of the constriction process, and not thelevels of these proteins individually (10, 12).We would like to distinguish the different domains in FtsZ

that participate in the interaction with other cell divisionproteins, including the ones mentioned above. To obtain dataabout the different domains of FtsZ, including the GTP-binding site, we decided to produce monoclonal antibodies(MAbs) against FtsZ. We report the localization of fivedifferent antigenic determinants (epitopes), which might beavailable for interactions with other (cell division) proteins. Atentative model of FtsZ is presented.

MATERIALS AND METHODS

Bacterial strains and growth conditions. E. coli strains(Table 1) were grown at 37°C in Luria broth (LB), unlessmentioned otherwise. Strains LMC1102 and LMC1104 were

induced for 90 min, starting at an optical density (at 600 nm) of0.2 by adding isopropyl-,B-thiogalactoside (IPTG) to a finalconcentration of 2 mM. Strain VIP2 (27) was grown in minimalsalts medium (33) at 30°C in the presence of 0.1% CasaminoAcids.

Cell fractionation. Exponentially growing cells were har-vested at an optical density at 600 nm of 0.9 and washed with1 mM phenylmethylsulfonyl fluoride (PMSF) in phosphate-buffered saline (PBS). The cell pellet was resuspended in 4 mlof PBS-PMSF, and cells were disrupted at 0°C by sonication.The cells were sonicated for 30 s, and then cooled for 30 s. Thisprocess was continued until the suspension was transparent.Crude material was removed by centrifugation in a microcen-

1886

Vol. 176, No. 7

EPITOPE MAPPING OF FtsZ 1887

TABLE 1. Bacterial strains and plasmids

Strain or Genotype or description Source or referenceplasmid

E. coli strainsMC4100 F- araD139 A(argF-lac)U169 7

deoCi flbB5301 ptsF25 rbsRreLA1 rpsLS50

LMC500 MC4100 lysA Our collectionLMC824 POP2136 harboring pEX2LMC1000 LMC500 F' lacIqL8LMC1101 POP2136 harboring JV100LMC1102 MC1061 harboring pJPB100LMC1104 LMC1000 harboring pJPB100MC1061 araDJ39 A(ara-leu)7697 A(lac)X74

galU galK strAVIP2 A(lacX174) araD139 A(ara-leu) 27

7697 galU galK hsr hsm+rpsL(str), harboring pLAR10

POP2136 F- araD139 A(lac)U169 thiA rpsL 38relA flaB cI857(Ts)

PlasmidspZAQ TetrftsQ ftsA ftsZ 40pJPB100 Ampr p(lacZ)-ftsZ E. Mulder and

T. P. BouchepLAR10 CamrftsZ' rep(Ts) 27pEX2 Ampr PR promoter of phage 32

lambda and cro-lacZ genefusion

pJV100 EcoRI-NaeI fragment offtsZ, This worksubcloned into pEX2

pJV106L pJV100 derivativepJV109E pJV100 derivativepJV124D pJV100 derivativepJV127B pJV100 derivative

trifuge at 18,000 x g for 5 min at 4°C. The supernatant was

spun at 120,000 x g for 30 min at 4°C. The membrane pelletwas washed once with water to remove remaining cytoplasmelements and resuspended in 1 ml of PBS-PMSF. Membranefractions and cytoplasm fractions were stored at - 20°C.

Construction and expression of the fusion gene lacZ-ftsZ. ADNA fragment containing ftsZ was obtained by digestion ofpZAQ (40) with EcoRI and NaeI. This 1,235-bp fragment wassubcloned into an EcoRI-Smal-digested pEX2 vector (32).The resulting construct pJV100 was transformed into strainPOP2136 (38) by the method of Chung and Miller (8). Thetransformant was grown at 30°C, and expression of the lacZ-ftsZ fusion gene was induced by incubation at 42°C for 30 minand then at 37°C for 1 h. To isolate the 150-kDa fusion protein,a 100-ml LB culture of the induced transformant was centri-fuged and the pellet was resuspended in Laemmli samplebuffer (21). The cells were then lysed by exposure to 100°C ina water bath for 5 min. Debris was removed by centrifugationat 18,000 x g, and the lysate was loaded onto a preparativesodium dodecyl sulfate-5.8% polyacrylamide gel. After elec-trophoresis, the 150-kDa band was electroeluted overnight at 3W and dialyzed against PBS for 6 h at 4°C.

Immunization. At day 0, BALB/c mouse AA1 was immu-nized by injection with 62.5 tLg of the 150-kDa fusion proteinmixed with an equal volume of incomplete Freund's adjuvant.On day 13, the mouse was injected with 125 jig of fusionprotein without adjuvant. Half of the material was injectedsubcutaneously, and the other half was injected intraperitone-ally. On day 24, a blood sample was taken from the mouse tomonitor the presence of FtsZ-specific antibodies. After 2

months (day 60), the mouse was injected intraperitoneally with16.2 mg of cytoplasm fraction of the IPTG-induced LMC1104to activate B lymphocytes that produce FtsZ-specific antibod-ies. Three days later, antiserum was obtained, and lymphocyteswere isolated from the spleen for fusion with NS1 myelomacells by standard techniques (19). The resulting hybridomaswere grown in microtiter plates as described before (19).

For a second immunization, FtsZ protein was purified fromIPTG-induced LMC1102 as previously described (11, 28). The40-kDa protein was isolated by electroelution and injectedtogether with incomplete Freund's adjuvant into rabbitKAS287 every week. Half of the material was injected subcu-taneously, and the other half was injected intermuscularly.Nine days after the fourth injection, KAS287 antiserum wasobtained.

Enzyme-linked immunosorbent assay (ELISA). Polystyrenemicrotiter plates with high binding capacity (Greiner, Nurtin-gen, Germany) were coated with 0.50 ,ug of protein per well ofthe cytoplasm fraction from IPTG-induced LMC1104 andincubated overnight at 4°C. Nonspecific binding sites wereblocked by incubation with 1% skim milk powder at 150 ,ul perwell for 1 h at 37°C. After the plates had been washed threetimes with 0.05% Tween 20 in PBS, incubations with thehybridoma culture supematants were performed in the pres-ence or absence of ND buffer (1% Nonidet P-40 and 0.5%sodium desoxycholate) for 1 h at 37°C. A third set of microtiterplates was coated with purified 0.25 ,ug per well of ,B-galacto-sidase (Sigma Chemical Co., St. Louis, Mo.) and similarlyincubated without ND buffer. Subsequent labeling was per-formed with anti-mouse peroxidase conjugate (Pasteur,Marnes-Lacoquette, France). Tetramethylbenzidine (Merck,Darmstadt, Germany) was used as a substrate for colorimetricanalysis (36).

Selection of MAbs. Hybridomas were selected on the basis ofa strong positive reaction in ELISA with the cytoplasm fractionof IPTG-induced LMC1104 and a negative reaction with,3-galactosidase. A second selection was carried out on thebasis of a positive reaction with the 40-kDa band and noreaction with ,-galactosidase in Western blot (immunoblot)analysis as described previously (18).Hybridomas showing positive reactions of the desired spec-

ificity were cloned by the limiting dilution technique at adensity of 1 cell per well and were recloned twice at a densityof 0.3 cell per well. Selected cultures were grown in bulk andinjected intraperitoneally into pristane-primed BALB/c micein order to obtain ascitic fluid. The culture supernatants of theselected hybridomas were used for characterization of theMAbs.

Class and subclass determination of MAbs. Ascitic fluidsamples were diluted 1:1,000 in PBS. The immunoglobulin (Ig)subclasses of the MAbs were determined by ELISA withsubclass-specific antisera (Serotec, Bicester, England).

Epitope mapping. Carboxy-terminal truncations of the ,B-ga-lactosidase-FtsZ fusion protein were generated with exonucle-ase III and S1 nuclease (Double Stranded Nested Deletion Kit;Pharmacia, Uppsala, Sweden). The plasmid pJV100 was di-gested with BamHI and PstI. Incubations with exonuclease IIIat 30°C at various time intervals and subsequent incubationwith S1 nuclease at 30°C for 30 min resulted in truncatedversions of this DNA construct. Since the 3' overhang of thePstI site is resistant to exonuclease III, we assumed that theenzyme digested only into the ftsZ gene. The truncated plas-mids were ligated and transformed into POP2136. Afterinduction, we obtained transformants producing fusion pro-teins truncated at the carboxy terminus. To determine the

VOL. 176, 1994

1888 VOSKUIL ET AL.

A -%$'

205 -

IUe,@

Ak

116 -

,p Gal-Fts Z

13 Gal

B

94

67

43-

97.4

30-

661 2 1 2

FIG. 1. Detection of 3-galactosidase, 3-galactosidase-FtsZ fusionprotein, and FtsZ. (A) Sodium dodecyl sulfate-5.8% polyacrylamidegel stained with Coomassie brilliant blue. Lane 1, total cell lysate ofLMC824 induced to overexpress 1-galactosidase; lane 2, total celllysate of LMC1101 induced to overexpress the lacZ-ftsZ fusion con-struct (pJV100). a Gal, 1-galactosidase. (B) Western blot ofLMC1104, induced to overexpress ftsZ. Lane 1, 16 p.g of protein ofcytoplasm; lane 2, 18 jig of protein of membrane fraction. FtsZ waslabeled with antiserum obtained from mouse AA1 after injections withthe ,B-galactosidase-FtsZ fusion protein. To the left of the gels are thepositions of molecular size markers (in kilodaltons).

position of each truncation, sequence analysis was performedby the method of Sanger et al. (29). A set of six transformantswas selected for epitope mapping.

In addition to the procedure described above, FtsZ proteinwas purified as previously described (11, 28) and cleaved byCNBr (26). The truncated fusion proteins were separated asdescribed by Laemmli (21), and the CNBr fragments wereseparated as described by Schagger and Von Jagow (30). Thefragments and the truncated fusion proteins were analyzed forreaction with the MAbs in Western blots.

Chemicals. Restriction enzymes and buffers were obtainedfrom Promega (Madison, Wis.) and Boehringer GmbH(Mannheim, Germany). The components of LB were obtainedfrom Difco (Detroit, Mich.) and peroxidase conjugates anddetection chemicals for Western blot analysis were obtainedfrom Sigma. All other chemicals were obtained from BDH(Poole, England) or Serva (Heidelberg, Germany).

RESULTSPolyclonal antibodies against a 13-galactosidase-FtsZ fusion

protein. We studied the migration of the ,B-galactosidase-FtsZfusion protein on a 5.8% polyacrylamide denaturing gel system(Fig. 1A). The ,B-galactosidase protein migrated as a 116-kDaprotein (lane 1). The fusion protein migrated as a 150-kDaprotein (lane 2), which proved to be large enough to separateit from most other E. coli proteins. The 150-kDa band wasisolated by electroelution and used for immunization of amouse (AA1). Antibodies thus raised against the fusion pro-tein reacted with a 40-kDa band in both cytoplasm andmembrane preparations from IPTG-induced LMC1104 thatoverexpressed FtsZ (Fig. 1B). When the same amount ofproteins from wild-type cells was analyzed, a 40-kDa proteinwas just visible (data not shown). This result implies that the40-kDa protein shown in Fig. lB is overexpressed and repre-sents FtsZ.

Selection of MAbs. The selection of hybridomas resulted inMAbs F168-4, F168-8, F168-11, F168-12, F168-13, and F168-18, which in ELISA showed strong positive reactions withIPTG-induced LMC1104 in the presence or absence of NDbuffer (extinction values of 2.0 and higher) and showed noreaction with 1-galactosidase. In Western blots, the six MAbsreacted with a 40-kDa band, which corresponds to the size ofFtsZ, and no reaction with 3-galactosidase was found (resultsnot shown).

Epitope mapping. Reactivity of MAbs with the ,3-galactosi-dase-FtsZ fusion protein and the four fusion proteins trun-cated at the carboxy terminus was determined by Western blotanalysis (Fig. 2). All MAbs reacted with the total fusion proteinJV. MAb F168-4 reacted with nearly all truncated fusionproteins, except for protein 27B. MAbs F168-8, F168-13, andF168-18 reacted with the truncated proteins 6L and 9E, but notwith the truncated proteins 24D and 27B. MAb F168-11

5,

Fts Z

EcoRI

- COOHFusion proteins(last amino acid)

Reactivity withmonoclonal antibodies F168-

4 8 13 18 11 12

3,

JV(Asp 383)

6L(Gln 332)

9E(Asn 158)

24D(Gly 103)

27B(Ala 80)

13-Gal

+ + + + + +

+ + + + + _

+ + + + _-

+__

FIG. 2. Truncated versions of the lacZ-ftsZ fusion and recognition of its gene products by the MAbs in Western blots. ,B-Gal, 13-galactosidase.

protein H2N -

D3-Galactosidase

DNA

J. BACTERIOL.

. QO ^- r-- -7

EPITOPE MAPPING OF FtsZ 1889

TABLE 2. Reactivities of CNBr fragments of purified FtsZa

Reactivity of fragment with:Fragment MOL wt

MJV8 KAS287 F168-4 F168-8 F168-11 F168-12 F168-13 F168-18

A 7,065 + + + - - - -B 6,595 - - - + - - + +C 6,291 + + + - - - -D 4,318 + - - - - + -E 3,885 - - - - - - -

a Five fragments of FtsZ were detected (+) or not detected (-) by antisera and MAbs in Western blots.

reacted only with the truncated protein 6L, whereas MAbF168-12 did not react with any of the truncated fusion proteins(Fig. 2).

In addition, the reactivities of the MAbs with CNBr frag-ments from purified FtsZ were determined in Western blots(Table 2). MAb F168-4 reacted with fragment A as well as withfragment C. These fragments are situated at different positionsin the FtsZ protein. MAbs F168-8, F168-13, and F168-18recognized fragment B. MAb F168-11 did not react with any ofthe analyzed fragments. MAb F168-12 reacted with the car-boxy-terminal fragment D, which was consistent with theresults from the truncated fusion proteins (Fig. 2).The last amino acids of the truncated fusion proteins were

determined by DNA sequence analysis of the correspondingplasmids. The results of reactions of the MAbs with thetruncated fusion proteins and with the CNBr-cleaved FtsZfragments were combined, and five different epitopes weredefined (Table 3 and Fig. 3). Epitope I, which was recognizedby MAb F168-4, is situated in an area between amino acids 80and 97 on the left flank of the GTP-binding sequence (G box[Fig. 3]). Epitope II reacted with MAbs F168-8, F168-13, andF168-18 and is situated on the right flank of the G box betweenamino acids 144 and 158. Epitope III was the second epitopethat reacted with MAb F168-4 and is situated in CNBrfragment C. Comparison of the amino acid sequences ofepitopes I and III by computer analysis (FastP program) didnot reveal any homology. Comparison of these epitopes byHydrophobic Cluster Analysis (DNAid+ program) also didnot reveal any homology, indicating that the epitopes aredifferent in amino acid composition and structural conforma-tion. Epitope IV is defined by MAb F168-11 and was stillpresent on fusion protein 6L. Since CNBr fragment C did notreact with this antibody, the epitope reacting with MAbF168-11 must be situated between amino acids 303 and 332.Epitope V reacted with MAb F168-12 and is situated in CNBrfragment D between amino acids 337 and 383.

Class and subclass determination ofMAbs. The types of thedifferent MAbs were determined by ELISA (Table 3). Five

TABLE 3. Characterization of MAbs against FtsZ

MAb Subclass Amino acid EpitoperesidueseEioe

F168-4 IgG2b 79-97, 243-302 I, IIIF168-8 IgGl 144-158 IIF168-11 IgGI 303-332 IVF168-12 IgGI 337-383 VF168-13 IgGl 144-158 IIF168-18 IgGl 144-158 II

a The summarized results of epitope mapping (Fig. 2 and Table 2).b Designated epitope numbers of FtsZ.

MAbs belong to the IgGl subclass; only Mab F168-4 belongsto the IgG2b subclass.

Detection of FtsZ in a wild-type strain. Cytoplasm andmembranes were separated from wild-type strain LMC500 andwere used to test the specificity of the MAbs. In Western blots,all six MAbs reacted with a 40-kDa band in both fractions (Fig.4). In addition, MAbs F168-8, F168-13, and F168-18 reactedwith an unknown cytoplasmic 50-kDa protein (Fig. 4B, lanes 2,5, and 6). MAb F168-8, eluted from the 40-kDa band inWestern blot, reacted with the 40-kDa band and the 50-kDaband de novo (data not shown). MAb F168-8, eluted from the50-kDa band, reacted in the same way, indicating that the40-kDa protein and the 50-kDa protein share a commonepitope.The 50-kDa protein was detected exclusively in the cyto-

plasm fractions of different wild-type strains. This result indi-cates that the presence of FtsZ in the membranes cannot beexplained by trapping of cytoplasmic proteins in membranevesicles. Therefore, a certain amount of FtsZ molecules mustbe associated with the membrane fraction.

Analysis of VIP2 and derivation of the 50-kDa protein froma gene other thanftsZ. To investigate a possible relationshipbetween FtsZ and the 50-kDa protein, we performed a tem-perature-shift experiment with wild-type strain LMC500 andstrain VIP2, which harbors a kanamycin cassette in the chro-mosomal gene ftsZ (27). For complementation of this genedisruption, the wild-type gene was cloned on a plasmid with atemperature-sensitive replicon. After a temperature shift to42°C, the plasmid cannot replicate and will be diluted over thedaughter cells. Cells lacking FtsZ are formed and start to makefilaments (9, 27). Samples were taken from such an experi-ment, before and after the temperature shift, and analyzed inWestern blots (Fig. 5). Using MAb F168-8, three bands werevisible, i.e., a 30-kDa band, a 40-kDa band, and a 50-kDa band.The 30-kDa band was recognized by the sheep anti-mouseperoxidase conjugate (data not shown). The intensity of thisprotein band was similar in all lanes, indicating that each lanecontained about the same amount of proteins. Figure 5 showsLMC500 (lanes 1 to 6). The intensities of the 40-kDa band andthe 50-kDa band were about equal (cells grown in steady-statecultures at 30°C) (lane 1). After a temperature shift to 42°C,the 50-kDa band became fainter, whereas the 40-kDa bandremained unchanged (lane 2). The 50-kDa band becamegradually more intense in the 2 h after shifting back to 30°C(lanes 3, 4, 5, and 6).

Figure 5, lanes 7 to 12, show strain VIP2 analyzed in thesame way as LMC500 was. Lane 7 shows the presence of FtsZand the 50-kDa band of strain VIP2 grown in steady-statecultures at 30°C. After the shift to 42°C, both bands becamefainter (lane 8). The 50-kDa band restored gradually in the 2 hafter shifting back to 30°C, and the 40-kDa FtsZ band disap-peared (lanes 9 to 12). Apparently, the VIP2 strain lost the

VOL. 176, 1994

1890 VOSKUIL ET AL.

100 200 300

2

10

1

-2-3

ELIZIFA

Epitopes

I) H 1JV I

I G-box

I II11 I I ,,Io1o00 200

llE

I II

FIG. 3. Hydrophobicity plot of FtsZ by the method of Kyte and Doolittle (20). The grey area represents the hydrophobic domain containingthe GTP-binding sequence (G box). Below the plot is a schematic representation of FtsZ in which the positions of the methionine residues are

indicated as vertical lines. The epitopes (I to V) defined by the MAbs are indicated by the shaded boxes. The letters (A to E) refer to the CNBrfragments analyzed (see Table 2).

plasmid with its wild-type ftsZ gene and could not synthesizeFtsZ de novo in the 2 h after shifting back to 30°C. This resultdemonstrates that the gene which codes for the 50-kDa proteinis different and that its expression is independent offtsZ. Thiswas confirmed by the observation that the 50-kDa protein didnot become more intense after overexpressing the ftsZ gene

(data not shown).

DISCUSSION

Our data characterize five surface sites accessible to differ-ent MAbs. We believe that some of these epitopes are

B 67-

43-

n a n a a a

potential sites for interaction with other proteins (see below),for interaction with the cellular membrane, and perhaps forinteraction with other FtsZ molecules to form a polymer ringat the site of division (4).

For mapping the epitopes of FtsZ, the exonuclease IIImethod was complemented by chemical cleavage of the puri-fied protein with CNBr. Using the conditions for gel electro-phoresis of small proteins (30), we were able to separate thefive larger CNBr fragments (Table 2). The combination ofmethods enabled us to localize epitopes in areas as small as 15to 18 amino acids with less effort than would have beennecessary with one of the methods alone. Besides, the CNBranalysis revealed the existence of two distinct epitopes thatwere recognized by the same MAb (see below). The exonucle-ase III truncations would indicate a second epitope only if

94-

67

43

30 -- _ -_

;k

30- 30-1 2 3 4 5 6 1 2 3 4 5 6

FIG. 4. Detection of FtsZ in cytoplasm and membrane fractions.Western blot of E. coli LMC500. Purified membranes (A) andcytoplasm (B) were loaded on preparative slots (both panels show one

sample each). Proteins (50 ,ig per lane on average) were loaded on thegel. Small strips of the Western blots were exposed to the differentculture supernatants, each of them containing MAbs. Lane 1, F168-4;lane 2, F168-8; lane 3, F168-11; lane 4, F168-12; lane 5, F168-13; lane6, F168-18. To the left of the gels are the positions of molecular sizemarkers (in kilodaltons).

1 2 3 4 5 6 7 8 9 10 11 12

FIG. 5. Temperature-shift experiment with E. coli LMC500 andVIP2 in minimal salts medium containing 0.05% Casamino Acids.Steady-state cultures of LMC500 (lanes 1 to 6) and VIP2 (lanes 7 to12) were grown at 30°C (lanes 1 and 7), shifted to 42°C for 80 min(lanes 2 and 8), and then shifted back to 30°C. After the shift back,samples were withdrawn from the cultures at 30 (lanes 3 and 9), 60(lanes 4 and 10), 90 (lanes 5 and 11), and 120 (lanes 6 and 12) min.Temperature-shift experiments were performed by diluting the cul-tures 5- to 10-fold in prewarmed medium. To the left of the gel are thepositions of molecular size markers (in kilodaltons).

2

-1

-2-3

YuI.IIIIIII

III IV V

A 67 - iZ ;'4;.0.'DD:

:0-0-0,-:

::.;

43-':

0: ;;:: X, <.

A_M.:_ 's_

dk

,,,,'S. " g

-11. . -11-

J. BACTERIOL.

--fto

EPITOPE MAPPING OF FtsZ 1891

truncations were constructed from the amino and carboxytermini. However, constructions of amino-terminal truncationswere not available.

Epitopes I and III reacted with the same MAb. MAb F168-4reacted with two different epitopes in FtsZ. No identical aminoacid sequences were found. The existence of two epitopes thatbind to MAb F168-4 would indicate that MAb F168-4 containsa mixture of two different MAbs. Remarkably though, if thatwere the case, both MAbs would belong to the same Ig class(IgG2b), which is different from all otherMAbs (Table 3). Inour opinion, the method used to isolate these MAbs shouldexclude the possibility of a mixture, since hybridoma cloningwas performed in three steps: once with one cell per well andthen twice with 0.3 cell per well. Competition experiments withMAb F168-4 revealed that purified fusion protein 24D pre-vented labeling of both 24D and the entire fusion protein inWestern blots (data not shown). Fusion protein 24D containsonly epitope I (Fig. 2 and 3). Apparently, all antibodies formeda complex with fusion protein 24D, leaving no antibodiesbehind to react with epitope III. Besides, hybridoma cells wererecloned at 0.75 cell per well. After growth, all supernatantsfrom the wells that contained cells reacted with 24D inWestern blots. Thus, there were no supernatants containingepitope III-specific antibodies that did not react with protein24D. These observations again indicate that this hybridomacell line was derived from one cell only (i.e., a monoclonal cellline). Because F168-4 belongs to the IgG2b subclass, this MAbmight be a hybrid molecule (17). This could explain thereaction with two apparently different epitopes.

Epitope location and surface model of FtsZ. Our analysisdemonstrates that epitope I, which is recognized by F168-4,and epitope II, which is recognized by F168-8, F168-13 andF168-18, are located on each side of a hydrophobic domain(Fig. 3). Presumably, this domain is situated deeper in theprotein as a hydrophobic pocket. Thus, the hydrophilicepitopes I and II seem to be located on the edges of thispocket. The glycine-rich sequence from amino acids 103 to 111is likely to be involved in its interaction with GTP (11, 24, 28).It is located in the hydrophobic region (Fig. 3), suggesting thatthe hydrolase activity is located in or near this region. There-fore, epitopes I and II may play a role in this hydrolase activity.Other GTP-binding elements may be present, like the regionsfrom aspartic acid 158 to leucine 172 and from glycine 204 tomethionine 217 that were aligned with conserved regions intubulins (22). However, consensus sequences as found in otherG proteins (6) have not been clearly found in FtsZ. Apart fromthe region between epitope I and epitope II, other hydropho-bic parts may be involved in the formation of the pocket, suchas a glycine-rich area between isoleucine 16 and asparagine 24(43). Since FtsZ is a cytoplasmic protein, the hydrophilic parts,like the described epitopes, are likely to be situated on itssurface. This is consistent with the strong reactivity of theMAbs in ELISA under conditions that were considered to benative (Materials and Methods). The recognition of epitopes I,III, and V by rabbit antiserum MJV8 against FtsZ (27) and therecognition of epitopes I and III by rabbit antiserum KAS287(this study) indicate the strong antigenicity of these epitopes,confirming the positions of these epitopes on the surface of thenative protein. All these considerations have led us to proposea provisional model of FtsZ (Fig. 6).Our computer analysis, done with the algorithms of Eisen-

berg et al. (14), revealed that FtsZ is mainly a globular protein(PlotA/TMH program). The positions of the epitopes I and IIsuggest that they constitute the borderline between the hydro-phobic pocket and the surface of FtsZ. Epitope III is not wellcharacterized yet, but it is likely to be located in a hydrophilic

FIG. 6. Clay model of FtsZ. The black spots (Ito V) represent theepitopes defined by the MAbs. The putative hydrophobic pocket withthe G box inside, is seen on top of the model, with epitopes I and II oneach side of the opening.

domain within CNBr fragment C. Its position on the surface ofthe protein may be proximal to epitope I, since in the ftsZJOO(Rsa) strain (originally designated sfiB* [16]), two mutationswere responsible for higher resistance to SfiA than other Rsastrains (3, 16). The mutated amino acids were valine 81 (3),which is situated in epitope I, and cysteine 268 (3), which issituated in CNBr fragment C that contains epitope III (Fig. 4).If these mutated amino acids were folded together to form adomain that interacts with SfiA, that would explain the higherresistance. This notion then favors epitopes I and III to form astructural domain. Alternatively, the interaction between FtsZand SfiA might involve different domains, which would indi-cate that epitopes I and III are situated in different positionson the surface of FtsZ. Hence, the distance between epitopesI and III on the protein's surface remains to be elucidated. Inour model, epitopes IV and V may be located anywhere on theoutside of the protein. Consequently, the positions of theepitopes III, IV, and V on the model are arbitrary (Fig. 6). Ourprediction that these epitopes are located on the surface ofFtsZ implies that they are available for interactions with othercell division proteins.

Cross-reaction of MAbs with a 50-kDa protein. The 50-kDaprotein, which was recognized by MAbs F168-8, F168-13, andF168-18, is encoded by a gene other thanftsZ, since in VIP2 itspresence was independent offtsZ and its gene product (Fig. 5).This 50-kDa protein was never detected with various publishedpolyclonal antisera against FtsZ (3, 27, 28; this study), hencethere is no relation between FtsZ and the 50-kDa protein. Sofar, we have no indications that the cytoplasmic 50-kDa proteinhas any regulatory function with respect to the cell cycle, andit may be a coincidence that this protein shares an epitope withFtsZ. Since its gene expression seems to be repressed byelevated temperature, the 50-kDa protein may possess aphysiological function in E. coli.

Interactions between FtsZ and other proteins. Our cellfractionation analysis indicated that FtsZ was present in thecytoplasm and in a membrane-containing fraction (Fig. 4). Thisresult is in agreement with observations by immunogold label-

VOL. 176, 1994

1892 VOSKUIL ET AL.

ing on thin sections (4). Other investigators suggest a strongdependence of detection of FtsZ in the membranes on theduration of the ultrasonication used to disrupt the cells (27).Both the immunogold labeling studies (4) and our cell frac-tionation experiments demonstrated the association of FtsZwith the membrane. Therefore, we consider this associationnot to be related to the method of cell disruption.

It is likely that positioning of FtsZ at the membrane triggersdivision-specific peptidoglycan synthesis (25). Thus, one do-main might interact with a membrane component that isdirectly or indirectly involved in peptidoglycan synthesis. Theresistance of MAbs to ND buffer will contribute to the searchfor such membrane components. Other domains might beinvolved in the formation of the FtsZ ring structure or itsGTPase activity. Since cell division is only useful if thenucleoids have separated, FtsZ might (through other proteins)sense decatenation following termination of DNA replication.Therefore, it is likely that there are domains on FtsZ which areavailable for interactions with such proteins. Considering thefive MAb-defined epitopes on the surface of FtsZ, our ongoingresearch is directed to deduce the physiological significance ofthese epitopes.

ACKNOWLEDGMENTSWe are grateful to M. Vicente and J. Pla for providing MJV8

antiserum and strain VIP2. The expert technical assistance of S.Sondij, J. van Leeuwen, S. Kuijper, M. Haring, and A. Maarsse isgratefully acknowledged. We also thank J. Leutscher for preparing thephotographs and W. van Noppen for reading this manuscript critically.

REFERENCES1. Aldea, M., T. Garrido, J. Pla, and M. Vicente. 1990. Division genes

in Escherichia coli are expressed coordinately to cell septumrequirements by gearbox promoters. EMBO J. 9:3787-3794.

2. Begg, K. J., and W. D. Donachie. 1985. Cell shape and division inEscherichia coli: experiments with shape and division mutants. J.Bacteriol. 163:615-622.

3. Bi, E., and J. Lutkenhaus. 1990. Analysis of ftsZ mutations thatconfer resistance to the cell division inhibitor SulA (SfiA). J.Bacteriol. 172:5602-5609.

4. Bi, E., and J. Lutkenhaus. 1991. FtsZ ring structure associatedwith division in Escherichia coli. Nature (London) 354:161-164.

5. Bi, E., and J. Lutkenhaus. 1991. Genetics of bacterial cell division,p. 123-152. In S. Mohan, C. Dow, and J. A. Cole (ed.), Prokaryoticstructure and function: a new perspective. Society for GeneralMicrobiology Symposium 47. Cambridge University Press, Cam-bridge.

6. Bourne, H. R., D. A. Sanders, and F. McCormick. 1991. TheGTPase superfamily: conserved structure and molecular mecha-nism. Nature (London) 349:117-127.

7. Casadaban, M. J. 1976. Transcription and fusion of the lac genesto selected promoters in Escherichia coli using bacteriophagelambda and Mu. J. Mol. Biol. 104:541-555.

8. Chung, C. T., and R. H. Miller. 1988. A rapid and convenientmethod for the preparation and storage of competent bacterialcells. Nucleic Acids Res. 16:3580.

9. Dai, K., and J. Lutkenhaus. 1991. ftsZ is an essential cell divisiongene in Escherichia coli. J. Bacteriol. 173:3500-3506.

10. Dai, K., and J. Lutkenhaus. 1992. The proper ratio of FtsZ to FtsAis required for cell division to occur in Eschenichia coli. J.Bacteriol. 174:6145-6151.

11. de Boer, P., R. Crossley, and L. Rothfield. 1992. The essentialbacterial cell-division protein FtsZ is a GTPase. Nature (London)359:254-256.

12. Dewar, S. J., K. J. Begg, and W. D. Donachie. 1992. Inhibition ofcell division initiation by an imbalance in the ratio of FtsA toFtsZ. J. Bacteriol. 174:6314-6316.

13. Dewar, S. J., V. Kagan-Zur, K. J. Begg, and W. D. Donachie. 1989.Transcriptional regulation of cell division genes in Escherichia coli.Mol. Microbiol. 3:1371-1377.

14. Eisenberg, D., E. Swarz, M. Komaromy, and R. Wall. 1984.Analysis of membrane and surface protein sequences with thehydrophobic moment plot. J. Mol. Biol. 179:125-142.

15. Garrido, T., M. Sinchez, P. Palacios, M. Aldea, and M. Vicente.1993. Transcription of ftsZ oscillates during the cell cycle ofEscherichia coli. EMBO J. 12:3957-3965.

16. Huisman, O., R. D'Ari, and S. Gottesman. 1984. Cell-divisioncontrol in Escherichia coli: specific induction of the SOS functionSfiA protein is sufficient to block septation. Proc. Natl. Acad. Sci.USA 81:4490-4494.

17. Kohler, G., H. Hengartner, and M. J. Shulman. 1978. Immuno-globulin production by lymphocyte hybridomas. Eur. J. Immunol.8:82-88.

18. Kolk, A. H. J., R. Evers, D. G. Groothuis, H. Gilis, and S. Kuuper.1989. Production and characterization of monoclonal antibodiesagainst specific serotypes of Mycobacterium avium and the Myco-bacterium avium-Mycobactenium intracellulare-Mycobacterium scro-fulaceum complex. Infect. Immun. 57:2514-2521.

19. Kolk, A. H. J., M. L. Ho, P. R. Klatser, T. A. Eggelte, S. Kuiper,S. de Jonge, and J. van Leeuwen. 1984. Production and character-ization of monoclonal antibodies to Mycobactenium tuberculosis,M. bovis (BCG) and M. leprae. Clin. Exp. Immunol. 58:511-521.

20. Kyte, J., and R. F. Doolittle. 1982. A simple method for displayingthe hydropathic character of a protein. J. Mol. Biol. 157:105-132.

21. Laemmli, U. K. 1970. Cleavage of structural proteins during theassembly of the head of bacteriophage T4. Nature (London)227:680-685.

22. Lutkenhaus, J. 1993. FtsZ ring in bacterial cytokinesis. Mol.Microbiol. 9:403-409.

23. Masters, M., T. Paterson, A. G. Popplewell, T. Owen-Hughes, J. H.Pringle, and K. J. Begg. 1989. The effect of DnaA protein levelsand the rate of initiation at oriC on transcription originating in theftsQ and ftsA genes: in vivo experiments. Mol. Gen. Genet.216:475-483.

24. Mukherjee, A., K. Dai, and J. Lutkenhaus. 1993. Escherichia colicell division protein FtsZ is a guanine nucleotide binding protein.Proc. Natl. Acad. Sci. USA 90:1053-1057.

25. Nanninga, N. 1991. Cell division and peptidoglycan assembly inEscherichia coli. Mol. Microbiol. 5:791-795.

26. Nikodem, V., and J. R. Fresco. 1979. Protein fingerprinting bySDS-gel electrophoresis after partial fragmentation with CNBr.Anal. Biochem. 97:382-386.

27. Pla, J., M. Sanchez, M. Vicente, and M. Aldea. 1991. Preferentialcytoplasmic location of FtsZ, a protein essential for Escherichiacoli septation. Mol. Microbiol. 5:1681-1686.

28. RayChaudhuri, D., and J. T. Park 1992. Escherichia coli cell-division gene ftsZ encodes a novel GTP-binding protein. Nature(London) 359:251-254.

29. Sanger, F., S. Nicklen, and A. R. Coulson. 1977. DNA sequencingwith chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA74:5463-5467.

30. Schigger, H., and G. von Jagow. 1987. Tricine-sodium dodecylsulfate-polyacrylamide gel electrophoresis for the separation ofprotein in the range from 1 to 100 kDa. Anal. Biochem. 166:368-379.

31. Smith, R. W. P., M. Masters, and W. D. Donachie. 1993. Celldivision and transcription of fisZ. J. Bacteriol. 175:2788-2791.

32. Stanley, K. K., and J. P. Luzio. 1984. Construction of a new familyof high efficiency bacterial expression vectors: identification ofcDNA clones coding for human liver proteins. EMBO J. 3:1429-1434.

33. Taschner, P. E. M., P. G. Huls, E. Pas, and C. L. Woldringh. 1988.Division behavior and shape changes in isogenic ftsZ, ftsQ, ftsA,pbpB, and ftsE cell division mutants of Escherichia coli duringtemperature shift experiments. J. Bacteriol. 170:1533-1540.

34. Tetart, F., R. Albigot, A. Conter, E. Mulder, and J.-P. Bouche.1992. Involvement of FtsZ in coupling of nucleoid separation withseptation. Mol. Microbiol. 6:621-627.

35. Tetart, F., and J.-P. Bouche. 1992. Regulation of the expression ofthe cell-cycle gene ftsZ by DicF antisense RNA. Division does notrequire a fixed number of FtsZ molecules. Mol. Microbiol. 6:615-620.

36. Verbon, A., S. Kuiper, H. M. Jansen, P. Speelman, and A. H. J.

J. BAcrERIOL.

EPITOPE MAPPING OF FtsZ 1893

Kolk. 1990. Antigens in culture supernatant of Mycobacteriumtuberculosis: epitopes defined by monoclonal and human antibod-ies. J. Gen. Microbiol. 136:955-964.

37. Vicente, M., S. R. Kushner, T. Garrido, and M. Aldea. 1991. Therole of the 'gearbox' in the trancription of essential genes. Mol.Microbiol. 5:2085-2091.

38. Vidal-Ingigliardi, D., and 0. Raiboud. 1985. The mac promoters:functional hybrid promoters activated by the malT product andrepression by the lacI product. Nucleic Acids Res. 13:1164-1174.

39. Wang, X., P. A. J. de Boer, and L. I. Rothfield. 1991. A factor thatpositively regulates cell division by activating transcription of themajor cluster of essential cell division genes of Escherichia coli.

EMBO J. 10:3363-3372.40. Ward, J. E., Jr., and J. Lutkenhaus. 1985. Overproduction of FtsZ

induces minicell formation in Escherichia coli. Cell 42:941-949.41. Wientjes, F. B., and N. Nanninga. 1989. Rate and topography of

peptidoglycan synthesis during cell division in Escherichia coli: theconcept of a leading edge. J. Bacteriol. 171:3412-3419.

42. Woldringh, C. L., E. Mulder, P. G. Huls, and N. Vischer. 1991.Toporegulation of bacterial division according to the nucleoidocclusion model. Res. Microbiol. 142:309-320.

43. Yi, Q.-M., and J. Lutkenhaus. 1985. The nucleotide sequence ofthe essential cell-division gene ftsZ of Escherichia coli. Gene36:241-247.

VOL. 176, 1994