Journal of Bacteriology, March 2004, p . 1546-1555, Vol . 186, No . 5

The Switch I and II Regions of MinD Are Required for Binding and Activating MinC

Huaijin Zhou and Joe Lutkenhaus^*

Department of Microbiology, Molecular Genetics, and Immunology, University of Kansas Medical Center, Kansas City, Kansas 66160

Received 1 October 2003/ Accepted 19 November 2003

 
MinD and MinC cooperate to form an efficient inhibitor of Z-ring formation that is spatially regulated by MinE . MinD activatesMinC by recruiting it to the membrane and targeting it to aseptal component . To better understand this activation, we haveisolated loss-of-function mutations in minD and carried out site-directed mutagenesis . Many of these mutations block MinC-MinD interaction; however, they also prevent MinD self-interactionand membrane binding, suggesting that they affect nucleotideinteraction or protein folding . Two mutations in the switchI region [MinD box] and one mutation in the switch II regionhad little affect on most MinD functions, such as MinD self-interaction,membrane binding, and MinE stimulation; however, they did eliminateMinD-MinC interaction . Two additional mutations in the switchII region did not affect MinC binding . Further study revealedthat one of these allowed the MinCD complex to target to theseptum but was still deficient in blocking division . These resultsindicate that the switch I and II regions of MinD are requiredfor interaction with MinC but not MinE and that the switch IIregion has a role in activating MinC.

 
In Escherichia coli, the min system prevents Z-ring assembly away from the midcell [1, 6] . This spatial regulation of Z-ringassembly requires all three products of the min locus, MinC,MinD, and MinE . MinC and MinD cooperate to form a division inhibitorthat is topologically regulated by MinE [6] . MinC is an antagonistof FtsZ assembly [15] that is recruited to the membrane by MinDand, together, induced to oscillate between the poles of thecell by MinE [11, 28] . Through this oscillation, the time-averaged concentration of MinC and MinD at the membrane is highest atthe poles and lowest at the midcell [24] . Although it appears that FtsZ has the capacity to assemble into a Z ring at many points on the cell membrane, the Min system helps to limit the position to the midcell [37].

Overexpression of MinC, but not MinD, in a min strain preventsZ-ring formation, establishing MinC as an inhibitor of division[1, 7] . In vitro, MinC antagonizes FtsZ assembly without affectingGTP hydrolysis, indicating that it does not block assembly butinstead promotes disassembly [15] . MinC can be divided intotwo domains, an N-terminal domain responsible for antagonismof FtsZ assembly and a C-terminal domain responsible for dimerizationand interaction with MinD [12, 34] . The crystal structure ofMinC confirmed that it is a dimer with two structurally independentdomains and that it dimerizes through its C-terminal domain[3].

Although MinC is the inhibitor and can antagonize FtsZ assemblyin vitro, MinD is required for it to be effective at its physiological concentration [7] . MinD recruits MinC to the membrane [11, 16,21, 28] and confers a high affinity on the complex for someseptal component [19] . This affinity was revealed by examiningthe localization of a green fluorescent protein [GFP]-taggedMinC mutant that is unable to antagonize FtsZ assembly . TheMinD-dependent localization of such GFP-tagged MinC mutantsto Z rings is readily observed, since the complex localizesto rings but is unable to cause their destruction . DicB, encodedby the defective Kim prophage, can also confer high affinityto MinC for a septal component [19].

MinD is at the center of the min system and has been shown to interact with itself, the membrane, MinC, and MinE . MinD self-interaction and interaction with MinC and MinE have been demonstrated with the yeast two-hybrid system [17, 32] . Mutations that affectthe nucleotide binding site of MinD reduce the interaction ofMinD with MinC and MinE as well as self-interaction and bindingto the membrane, suggesting that ATP is required for all MinDinteractions [9, 10, 14, 27].

In vitro studies have shown that MinD binds cooperatively to vesicles in an ATP-dependent manner and is released by MinE stimulation of the MinD ATPase [10, 21, 25] . MinE stimulationof the MinD ATPase requires the presence of phospholipid vesicles,indicating that only the membrane-bound form of MinD is susceptibleto MinE stimulation [13] . MinC is recruited to the MinD-vesiclecomplex and can be displaced by MinE, even in the absence ofATP hydrolysis [16, 21] . However, MinC cannot displace MinEbound to the MinD-vesicle complex [16] . Furthermore, once MinD is bound to the vesicle, it is able to undergo assembly into polymers that are able to tubulate the vesicles [10] . Assemblyof MinD into polymers on vesicles could help to recruit MinD to the site of assembly and restrict MinD diffusion until itis released by MinE . Recently, GFP-MinD was observed in spiral-like patterns in vivo, indicating that the oscillation of MinD involved the formation of MinD polymers on the membrane [30].

Binding of MinD to the membrane is ATP dependent and requiresa short, conserved C-terminal region that encodes an amphipathichelix [14, 35] . Hydrophobic residues within this helix mediatebinding by inserting into the bilayer [38] . Deletion of thishelix or mutations that substitute charged residues for hydrophobicresidues prevent MinD from binding to the membrane [35, 38].Importantly, such mutations do not prevent MinD binding to MinCin the yeast two-hybrid system but decrease the targeting ofMinC to the septal ring [14] . This result indicated that MinD'sinteraction with MinC may not require the membrane but thatmembrane-bound MinD is required for efficient targeting of MinCDcomplexes to the septum . Together, these results have led toa model in which the MinCD complex assembles on the membraneat the poles of the cell and acquires a high affinity for a septal component . In this model, the polar targeting and subsequent polymerization of MinD restrict the MinCD complex from diffusing in the membrane so that it does not come into contact with the cell center [14] . However, a Z ring attempting to form in the polar MinCD zone would come into contact with MinCD and be destabilized.

MinD-like proteins have been crystallized as monomers from several Archaea species with ADP or no nucleotide bound [4, 9, 29].However, MinD has been reported to be a dimer . In one study,dimerization was ATP dependent [14], whereas in the other, itwas not [32] . The structure of MinD is similar to that of NifH,which has been compared to small G proteins [4, 8, 23] . Theseproteins have regions designated switch I and II that undergonucleotide-dependent conformational changes that are responsiblefor interactions with their partners . In this study, we haveisolated MinD mutants that are unable to cooperate with MinCto block division to further explore the MinD-MinC interaction.Our results demonstrate the importance of the switch I and IIregions in the interaction between MinD and MinC.

 
Strains and plasmids. The E . coli K-12 strain JS964 [MC1061 malP::lacI^q min::kan] was used in this study . The plasmids used in this study are listed in Table 1 . To construct the plasmid pGB2CD2, the minCDgenes, including the promoter region and ribosome binding siteupstream of minC, were cloned into the low-copy-number plasmidpGB2 [2] . The primers used to amplify the minCD fragment were5'MTNCD2, 5'-TAGCATGAATTCACGACGGCAATGGGTTGATTG-3', and 3'MTNCD2,5'-TAGCATAAGCTTAACTTATCCTCCGAAGAAGCG-3', with the locationsof restriction sites indicated by underlining . To constructthe plasmid pHJZ109, a fragment containing minC^116-231 and minDwas amplified from pJPB210 [minCDE] by PCR with the primers5'SalIminC^116-231, 5'-TAGCATGTCGACGCGCAAAATACAACGCCGGTC-3',and 3'HindIII minD, 5'-TAGCATAAGCTTCTTATCTCTCGAACAAGCGTTTGA-3'. The fragment was cloned such that the C-terminal fragment of MinC was in frame downstream of gfp [mut2] in pSEB181 . pSEB181 was constructed by S . Pichoff and contains the EcoRI-SspI fragment from pJC106 [11] cloned into pAM238 [11] between the FspI andEcoRI sites . pHJZ109-D9 [minD-E126A] was constructed similarlyto pHJZ109, except that pSEB12-D9 [minD-E126A] [as describedbelow] was used as the template for PCR.

 
Isolation of minD mutations. A PCR-based, random mutagenesis approach [22] was used to introduce mutations into minD. The mutagenic PCR contained 10 mM Tris-HCl [pH 8.3], 50 mM KCl, 2 mM MgCl₂, 0.25 mM MnCl₂, 20 µMconcentrations of each deoxynucleoside triphosphate, 1 mM concentrationsof each primer, 50 ng of template DNA [pGB2CD2], and 1.25 Uof Taq DNA polymerase in a 100-µl reaction mixture . The initial denaturing step was carried out at 94°C for 5 min followed by 40 cycles of temperature cycling [94°C for 30s, 55°C for 30 s, and 72°C for 90 s] and a final incubationat 72°C for 10 min . Primers used in the reaction were 3'MTNCD2and 5'MluC, 5'-TATTAATCGCGTTTTTGTGACCGGCGT-3', which amplifieda fragment of 1.16 kb containing minC^116-231 minD . The poolof mutated fragments was digested with BstXI and HindIII, generating a 0.7-kb fragment containing codons 26 to 270 of minD . By doing this, we excluded the first 25 codons to avoid mutations inthe deviant Walker A motif already known to affect the MinD-MinC interaction [5, 9] . The pool of mutated fragments was used toreplace the wild-type fragment on pGB2CD2 . The resultant plasmidswere electroporated into JS964 [min], and transformants wereselected on plates containing spectinomycin . The transformantswere inoculated into liquid culture, and cell lysates were analyzedby sodium dodecyl sulfate-polyacrylamide gel electrophoresis[SDS-PAGE] followed by Western blotting to assess the levelof MinD . Forty-nine of the 111 clones examined expressed MinD at a level comparable to the wild-type MinD expressed from pGB2CD2. Of these 49 clones, 30 were further analyzed by sequencing and 20 of them were found to contain a single mutation . DNA fragments containing these minD mutations were subcloned into different vectors by PCR to generate various fusions, pZH106 [gfp-minD], pJC41[AD-minD], pJC42-1 [BD-minD], and pZH115 [pJF118EH-minD], with the same primers that were used for cloning the wild-type minD into these vectors . In addition, several minD mutations were made by site-directed mutagenesis . The switch II mutations minD-IE125,126AA, minD-I125E, and minD-E126A were introducedinto minD contained on pZH106 by using the QuikChange site-directedmutagenesis kit [Stratagene, La Jolla, Calif.] . The minD-I126Emutation was also introduced into minD on pSEB12 [minCDE] bysite-directed mutagenesis, giving rise to pSEB12-D9 [minD-I126E].

Yeast two-hybrid assay. Interaction between MinD and itself and MinC can be detectedin the yeast two-hybrid system [17, 19, 32] . The effect of various minD mutations on the interaction between MinD and MinC and the interaction of MinD with itself were assessed by using theyeast two-hybrid assay as described previously [17] . The correspondingyeast two-hybrid plasmids described above were transformed invarious combinations into the yeast reporter strain SFY526 asdescribed in the Clontech manual [BD Bioscience, Palo Alto, Calif.] . Double transformants were selected on media without tryptophan and leucine . Transformants were examined for ß-galactosidase activity both qualitatively by the colony lift assay and quantitatively as described in the Clontech manual.

Fluorescence microscopy. The effect of minD mutations on the membrane localization ofGFP-MinD fusions were examined as described previously [13].To test the ability of minD9 to target GFP-MinC^116-231 to theseptum, plasmid pHJZ109-D9 [minD-E126A] was transformed intoJS964 [min] by selecting for Spc^r . pHJZ109 was used as a control. Cultures of the transformants were grown at 37°C until theoptical density at 600 nm reached 0.05 . At this point, 40 µMof isopropyl-ß-D-thiogalactopyranoside [IPTG] wasadded, and 1 h later, samples were removed and fixed with 2%glutaraldehyde . Aliquots were placed on microscope slides, and cells were photographed with a Nikon fluorescence microscope with a MagnaFire charge-coupled device camera [Optronics] . Images were imported into Adobe PhotoShop for assembly.

Phenotypic analysis of minD mutants. To test the effect of minD mutations on the ability of MinDto activate MinC to block cell division, derivatives of pZH106[gfp-minD] were cotransformed with pZH110 [minC] into JS964[min] and selected on Luria-Bertani plates containing ampicillin, spectinomycin, and 0.2% glucose . Colonies were restreaked ontoplates with ampicillin and spectinomycin with arabinose addedat 0.0001% and 0.001% as indicated . The effect of minD mutationson cell morphology was determined by phase-contrast microscopyafter overnight incubation.

Protein purification. MinD-S121T, MinD-R44G, MinD-G42A, MinD-G42D, MinD-I141N, andMinD-E126A were overexpressed and purified with plasmids pJF118EH-minD-S121T,-minD-R44G, -minD-G42A, -minD-G42D, -minD-I141N, and -minD-E126A, respectively . The purification procedure was the same as thatfor the wild-type MinD protein . All of the mutants behaved thesame as the wild-type protein during the purification exceptMinD-G42D, which proved to be unstable . Protein concentrationswere determined by protein assay [Bio-Rad Laboratories, Hercules,Calif.], and the purity was judged by sodium dodecyl sulfate-polyacrylamidegel electrophoresis . MinE and MalE-MinC^116-231 were purified as described before [16].

MinD ATPase and membrane binding assay. The MinD ATPase assay was as described previously [13] . MinD,MinD mutants [9 µM], and phospholipid vesicles [400 µg/ml]were mixed with 1 mM [-³²P]ATP in an ATPase buffer [25 mM Tris· HCl [pH 7.5], 50 mM KCl, 5 mM MgCl₂] . MinE [9 µM]was added to reaction mixtures as indicated . Reaction mixtureswere incubated at 30°C, the amount of released P_i was determined,and the specific activities were calculated . In the sedimentationassay, MinD mutants [4 µM], phospholipid vesicles [400 µg/ml], and nucleotide [1 mM ADP or ATP] were mixed atroom temperature in 50 µl of ATPase buffer . MalE-MinC^116-231 and MinE were added at 4 µM as indicated in the figurelegends . The reaction mixtures were incubated at room temperaturefor 20 min . For reaction mixtures containing MinE, the sampleswere incubated at 30°C . Samples were centrifuged at 10,000x g at room temperature in a tabletop centrifuge for 2 min.The supernatants were carefully removed, and the pellets wereresuspended in 50 µl of SDS sample buffer . Twenty-microliteraliquots of the samples were electrophoresed on SDS-12.5% PAGEand stained with Coomassie brilliant blue.

 
Isolation of MinD mutants unable to activate MinC. A low-copy-number plasmid containing minCD cannot be introduced into a min strain in the absence of MinE . This offers a simpleapproach to isolate mutations that inactivate minC or minD.We used a plasmid containing the minC and minD genes cloned into the low-copy-number plasmid pGB2 [Spc^r] . This plasmid, pGB2CD2, could not transform JS964 [min] unless this strainalso contained pJPB216, which expresses minE under lac promotercontrol, and was grown in the presence of IPTG . To generatemutations, the minD gene was amplified from pJPB210 [minCDE]by a mutagenic PCR procedure as described in Materials and Methods.The approach we used excluded the first 25 codons to avoid mutationsin the deviant Walker A motif already known to be required forMinD function . The pool of PCR fragments was used to replaceminD in pGB2CD2, and transformants of JS964 [min] were selectedon plates containing spectinomycin . Transformants were screenedby immunoblot analysis for the level of expression of MinD totry to eliminate mutations that lead to instability . Of 111 transformants, 49 were found to express MinD at a level comparable to a control containing the wild-type minD [pGB2CD2] . DNA sequence analysis of 30 of these 49 revealed that 20 contained a single mutation . These 20 were chosen for further study . The locationsof the mutations are summarized in Fig . 1 . Despite screening for protein stability, many of these mutations contain proline substitutions which are likely to disrupt the protein structure. However, other mutations affect charged residues that probablydo not affect folding and are likely to have other defects.

 
Effect of minD mutations on self-interaction and interaction with MinC. One possible explanation for the inability of the MinD mutants to inhibit division and prevent colony formation is that they no longer interact with MinC . To test this possibility, we utilized the yeast two-hybrid test in which a strong interaction is observed between MinD and MinC . We found that MinD-I141N interacted strongly with MinC, although somewhat less than the wild type . The other19 mutants, however, showed no interaction with MinC [Table 2] . Possible explanations for the failure of these mutants to bind MinC include that they don't fold properly, they don't interact with ATP, which is required for MinD to bind MinC,or they have an alteration in the surface of MinD that interactswith MinC.

 
MinD self-interaction can also be assayed by the yeast two-hybrid system [19, 32] . This interaction may reflect the interactionto form dimers . We observed that the MinD self-interaction isweaker than the interaction observed between MinC and MinD inthis test system . Of the 20 mutants, only 4, MinD- R44G, MinD-G42A,MinD-G42D, and MinD-I141N, displayed self-interaction . In fact,two of these, MinD-R44G and MinD-G42A, displayed stronger interactionthan the wild type . If we assume that this test reflects dimerization,then these four mutants are able to dimerize, whereas the other16 are not [Table 2, three are listed under mutations affectingATP binding or hydrolysis motifs, and the other 13 are groupedseparately] . Since three of the four mutants that self-interactdo not interact with MinC, the corresponding mutations may defineresidues specifically required for MinC interaction . The fourthmutant, MinD-I141N, binds MinC and must be defective in anotherstep.

Effect of MinD mutations on MinD localization. MinD is a peripheral membrane protein that binds to the membranethrough a C-terminal amphipathic helix [14, 35] . In vitro ithas been shown that this binding requires ATP, which also promotesMinD dimerization [16] . To determine the effect of the mutationsisolated in this study on MinD localization, each of the mutationswas introduced into a gfp-minD fusion downstream of the arabinosepromoter . Each of the plasmids was introduced into JS964 [min] and examined by fluorescence microscopy.

The 16 mutants, such as MinD-S121T [Fig . 2], that failed to display self-interaction in the yeast two-hybrid test also failed to localize to the membrane . In contrast, three of the mutants which displayed self interaction, MinD-R44G, MinD-G42A, and MinD-G42D, were clearly on the membrane, as they produced ahalo-like appearance comparable to that seen with the wild-typeMinD [Fig. 2] . However, MinD-I141N, the other mutant that displayed self interaction and the only one that bound MinC, was present throughout the cytoplasm . The failure of MinD-I141N to bindto the membrane would explain its reduced ability to activateMinC . The behavior of MinD-I141N is similar to mutants reportedpreviously [14] that fail to bind to the membrane due eitherto deletion of the C-terminal helix or to substitution of chargedfor hydrophobic amino acid residues within this helix . Suchmutants are still able to bind MinC in the yeast two-hybridsystem but are unable to activate it as effectively as the wild-typeMinD due to a deficiency in recruiting it to the membrane andtargeting it to the septum [14] . Interestingly, the minD-I141Nmutation does not lie in the C-terminal helix but lies nearthe middle of the protein [Fig. 1].

 
Biochemical analysis of MinD mutants. To determine the effects of minD mutations on the biochemicalactivities of MinD, selected minD alleles were cloned into anexpression vector for purification of the mutant proteins . Thefour mutants that displayed self-interaction [MinD-R44G, MinD-G42A, MinD-G42D, and MinD-I141N] and one that did not [MinD-S121T]were chosen . MinD-G42D proved to be unstable during purificationand was not studied further . The other mutant proteins werestable and behaved similarly to the wild-type protein duringpurification.

Wild-type MinD displays a basal ATPase activity that is stimulated about 10-fold by MinE in the presence of phospholipid vesicles[13, 21, 31] . MinD-R44G and MinD-G42A expressed a basal ATPasethat was three to five times the level of the wild-type protein[Fig . 3] . Nonetheless, this basal level was stimulated by MinEsuch that the stimulated level was higher than the wild-typestimulated level . In contrast, the basal ATPase activity ofMinD-I141N was slightly lower than the wild-type level and wasnot stimulated by MinE . Also, the basal ATPase activity of MinD-S121Twas comparable to the wild-type activity but was only stimulatedtwofold by MinE [Fig . 3].

 
Of the four mutant proteins examined for ATPase activity, thetwo that showed a level of stimulation by MinE similar to thatof the wild type localized to the membrane in vivo . In contrast,the two that showed little or no stimulation failed to localizeto the membrane . This result is consistent with the previousfinding that ATPase stimulation by MinE requires that MinD bindsto the membrane [13, 14] . Further support for this interpretationwas provided by examining binding to phospholipid vesicles invitro . Both MinD-R44G and MinD-G42A showed ATP-dependent bindingto phospholipid vesicles, whereas MinD-I141N and MinD-S121Tdid not display significant binding [Fig . 4A] . Since MinD-R44Gand MinD-G42A bound to vesicles, we could also test whetherthey were released by MinE and whether they were able to recruitMinC . Both MinD-R44G and MinD-G42A displayed an approximately50% reduction in the amount of bound protein when MinE was added[Fig . 4B] . Importantly, both mutants failed to recruit MinC[Fig. 4C] to the vesicles . This latter result is consistent with their failure to interact with MinC in the yeast two-hybrid system and explains their failure to inhibit division in the presence of MinC.

 
Mutations in the switch II region. Proteins in the extended ParA family, NifH, ArsA, and MinD,have been compared to G proteins, and regions have been identifiedthat correspond to the switch I and II regions [4, 8] . Theseregions correspond to signal transduction pathways and communicate information about the state of the bound nucleotide to otherregions of the protein . Each region has an aspartate at oneend of the switch that contacts the bound nucleotide indirectlythrough H₂O molecules bound to an Mg²⁺ ion . Two of the mutationsthat we isolated in the first part of this work contain mutationsthat alter these aspartic acid residues . MinD-D38A at the baseof switch I and MinD-D120G at the base of switch II produceproteins that do not bind MinC or the membrane [Table 2] . Thesemutants also fail to display self-interaction, suggesting thatthey fail to dimerize . Presumably, such mutants do not bindATP or are unable to differentiate ATP from ADP . This is incontrast to mutations that alter residues in the switch I regionadjacent to D38A and produce proteins [MinD-R44G and MinD-G42A]that failed to bind MinC but retain other functions such asself-interaction, membrane binding, and MinE stimulation . We,however, obtained no mutations in the switch II region . We thereforeconstructed three mutations in the switch II region by site-directedmutagenesis [MinD- I125E, MinD-E126A, and MinD-IE125,126AA]to try to determine what role, if any, the switch II regionhas in MinD function.

The MinD mutants MinD-I125E, MinD-E126A, and MinD-IE125,126AAwere constructed in the GFP-MinD expression vector under arabinose promoter control . All three mutant proteins localized to the membrane, suggesting that the switch II region, like the switchI region, did not have a significant role in membrane binding[Fig. 5] . Next, the three mutants were examined for the ability to activate MinC to block cell division [Table 2] . Plasmidscontaining the corresponding mutations were transformed into JS964 [min] containing pZH110 [minC] and selected on platescontaining glucose, which causes maximum repression of the arabinosepromoter . Under these conditions, a plasmid containing the wild-typeGFP-MinD fusion only yields a few transformants and the cellsare extremely filamentous . Thus, the glucose plates probablymost closely represent the physiological situation, since minCand minD cannot be introduced together in a single copy intoa min strain . In contrast, the switch II mutants yielded 1,000 transformants with a minicell phenotype similar to that of a plasmid control containing GFP-MinD-K16Q, a mutant that appearsto be completely inactive [5, 10].

 
Restreaking the transformants on plates lacking glucose or containing various amounts of arabinose gives some indication of the degreeof MinD attenuation . For example, transformants with the MinD-K16Q mutation gave a minicell phenotype at all arabinose concentrations, which is consistent with MinD-K16Q being completely inactive.In contrast, transformants with MinD-I125E started to filamentat high arabinose concentrations, whereas transformants withMinD-E126A or MinD-IE125,126AA started to filament in the absenceof both glucose and arabinose . The degree of attenuation ofthese latter mutants is similar to that observed with MinD3, which is missing the three carboxyl amino acids and binds poorly to the membrane [14] . Thus, all three switch II mutants areattenuated, which suggests that they are either deficient in binding MinC, similar to switch I mutants, or deficient in a subsequent step such as targeting the MinCD complex to the septum.

The yeast two-hybrid test revealed that one of the three switchII mutants, MinD-I125E, failed to bind MinC while the othertwo still bound MinC [Table 2] . This result indicated that the switch II region has some role in interaction with MinC . However, the fact that two of the switch II mutants still bound MinCand the membrane indicated that the switch II region is alsorequired at a step after the binding of MinD to MinC and themembrane.

To further explore the ability of these switch II mutants to activate minC, we examined the ability of MinD-E126A to target the C terminus of MinC to the septal ring . Plasmids were constructed with minD [pHJZ109] or minD-E126A [pHJZ109-D9] downstream of a gfp-minC^116-231 fusion and transformed into JS964 [min] . Cellsfrom the colonies were examined by fluorescence microscopy to determine whether the GFP fusion was localized in bands indicative of localization to septal rings . Control cells containing MinD contained fluorescent bands as described previously [19], indicatingthat the GFP-MinC^116-231-MinD complexes were being targetedto septal rings [Fig . 6] . Cells containing MinD-E126A also containedfluorescent bands, indicating that it could still target GFP-MinC^116-231 to the septal ring.

 
Since MinD-E126A could still target GFP-MinC^116-231 to septal rings, we checked the Min phenotype of this mutation following introduction on a single-copy plasmid carrying the intact min operon . Introduction of a plasmid carrying the wild-type min operon conferred a wild-type morphology on JS964 [min] [Fig.7A] . In contrast, JS964 [min] carrying the plasmid pSEB12-D9[minC minD-E126A minE] displayed a typical Min phenotype withminicells and nucleated cells of heterogeneous length [Fig.7B] . This result confirms that the minD-E126A mutation confersa typical Min phenotype and is unable to spatially regulatecell division.

 
To support the in vivo results, MinD-E126A was purified andexamined for MinD functions . MinD-E126A expressed a basal ATPaseactivity that was stimulated by MinE similar to wild-type MinD[data not shown] . As expected from this result and the in vivolocalization, MinD-E126A bound to phospholipid vesicles in anATP-dependent fashion [Fig. 4C] . Furthermore, MinD-E126A wasable to recruit MinC to the vesicles, consistent with its abilityto interact with MinC in the yeast two-hybrid system [Fig . 4C;Table 2] . Thus, this mutant is able to bind to the membrane, recruit MinC, and undergo MinE stimulation . It is also ableto target the MinCD complex to the septal ring, leading to the conclusion that the MinC-MinD-E126A complex is unable to carryout destruction of the Z ring.

 
MinC and MinD associate to form an inhibitor of division that prevents Z rings from forming away from the midcell in the presence of MinE . The formation of an efficient MinCD inhibitor involves several steps, including [i] the interaction of MinD with ATP,[ii] the binding of MinD to itself, to MinC, and to the membrane,[iii] the acquisition by the MinCD inhibitor of high affinityfor some septal component following membrane binding, and [iv]destabilization of Z rings upon MinCD binding to a septal component.Also, MinD must interact with MinE for MinCD to be spatiallyregulated.

The purpose of this study was to investigate the requirementsfor MinD activation of MinC . We isolated loss-of-function mutationsin minD coupled with site-directed mutagenesis to gain information about these steps and to identify the regions of MinD that are required . Several loss-of-function mutations affected residues important for interaction with ATP . As a result, they failedto self-interact, bind to the membrane, or interact with MinC, indicating that ATP is required for all MinD interactions . Several mutations were isolated that affected residues that correspondto the switch I and II regions of MinD . Analysis of these mutationsrevealed that the switch I and II regions are primarily involvedin the binding and activation of MinC, as the other activitiesof MinD were retained . Another mutation [minD-I141N] affectedmembrane binding without significantly affecting MinC binding.

Mutations leading to loss of MinD interaction with itself, the membrane, and MinC. Most of the loss-of-function mutations we isolated are defectivein self-interaction, localization to the membrane, and bindingto MinC . Many of these mutations result in either proline substitutionsthat probably alter protein structure or substitutions thatalter residues located in motifs required for ATP binding and/or hydrolysis, including S121T, D38A, and D120G . The two aspartic acid residues are at the bases of the switch I and switch IIregions [Fig . 8] and are involved in hydrogen bonding to H₂O molecules that in turn are bound to the Mg²⁺ ion required for ATP binding and hydrolysis [9] . Loss of either of these asparticacid residues would likely disrupt ATP binding . Serine 121 isat the base of the pocket near the -phosphate of the bound ATP.It is likely that replacing it with the larger threonine residueoccludes the space that would be occupied by the -phosphate and would therefore interfere with ATP binding . These results suggest that ATP binding is required for dimerization, bindingto the membrane, and binding to MinC.

 
Other mutations that affect residues in the deviant Walker Amotif involved in ATP binding and hydrolysis have been analyzedand found to be deficient in interaction with MinC [5, 9] . Inthis study, we have analyzed one of these, MinD-K16Q, and found that it is also deficient in self-interaction in the yeast two-hybrid test . Previous results have shown that this mutant is also unable to bind to the membrane [10] . Thus, the behavior of this mutantis typical of MinD mutants that are defective in interactionwith ATP; they don't interact with themselves, MinC, or themembrane.

The effects of mutations in minD genes from other bacteria have also been reported . Mutations in the minD gene from Bacillus subtilis that alter the same conserved aspartic residues at the bases of the switch I and switch II regions investigatedhere have been studied previously [20] . The mutations were found to prevent the polar targeting of MinD [which is dependent upon DivIVA], but they did not appear to prevent MinD from bindingto the membrane . We are not sure of the basis for this apparentdifference, since MinD from B . subtilis also binds to the membranethrough a C-terminal amphipathic helix [35] . A K16Q substitution in the Neisseria gonorrhoeae minD largely eliminated the self-interactionin the yeast two-hybrid test [27], similar to what we observedhere.

MinD binding to the membrane. Analysis of MinD has revealed that a C-terminal motif, whichhas the potential to form an amphipathic helix, is requiredfor MinD to bind to the membrane [14, 33, 35] . More recently, evidence for this model was provided by demonstration that tryptophan residues, substituted for the hydrophobic residues within thishelix, are embedded in the bilayer [38] . This has led to a model in which ATP binding by MinD activates this helix so that it can interact with the membrane . One possibility is that ATPbinding promotes dimerization, resulting in a tighter bivalentassociation with the membrane [33, 38].

Although most of the loss-of-function mutations we isolatedfailed to bind to the membrane, we isolated one mutation, minD-I141N, which primarily affected membrane binding, since it displayed self-interaction and interaction with MinC in the yeast two-hybrid system . In vitro, this mutant displayed poor binding to vesiclesand expressed a basal ATPase that was poorly stimulated by MinE.These results are very similar to a mutant, MinD10, with a deletionof the C-terminal amphipathic helix [14] . Interestingly, however,the minD-I141N mutation does not affect a residue in the C-terminalamphipathic helix but one that maps near the middle of the primarysequence [Fig . 1] . Since this mutant appears to dimerize, itis likely that it is unable to regulate the C-terminal amphipathichelix in response to ATP binding.

Interaction of MinD with MinC and MinE. The mutations we isolated that mapped to the switch I and IIregions of MinD are primarily defective in the interaction withMinC . Three of these mutations [G42A, R44G, and I125E] leadto a loss of interaction between MinD and MinC in the yeasttwo-hybrid system; however, none of the mutations affected MinDself-interaction or localization to the membrane . Detailed studyof the effect of two of the switch I mutations on MinD functionin vitro revealed that the mutant proteins had an elevated basalATPase activity but were still stimulated by MinE . Consistentwith this, the proteins underwent ATP-dependent vesicle bindingand were released by MinE . The mutants, however, were unableto recruit MinC to the membrane . These results are consistent with these mutants being primarily deficient in the bindingof MinC . The switch I region has been designated the MinD box,as it is highly conserved among MinD proteins but not amongother members of the ParA family, of which MinD is a member[36].

Interestingly, MinE can displace MinC from a MinC-MinD-vesicle complex made with a nonhydrolyzable ATP analogue, suggestingthat it may compete with MinC for binding to MinD [16, 21].However, the binding of the two proteins to MinD must be distinct,since MinE stimulates the MinD ATPase, whereas MinC has no stimulatoryeffect and does not interfere with the MinE stimulation [13]. Furthermore, MinC cannot displace MinE from a MinE-MinD-vesicle complex . This distinction between the binding of MinC and MinEis further documented here, as the switch I mutations eliminateMinC binding without having an observable effect on the interaction between MinD and MinE . These results are consistent with theprevious suggestion that MinE induces a conformation changein MinD, resulting in release of MinC [16] . Furthermore, wehave also observed that the switch I mutants can assemble onvesicles to cause tubulation, indicating that the switch I regionis not required for MinD polymerization that occurs on vesicles[data not shown].

In addition to the two switch I mutations, one of the mutationsin the switch II region [I125E] prevented MinC binding . Thismutation affects a residue, based upon homology modeling withthe MinD-like protein from Pyrococcus furiosus, which lies onthe surface of the MinD molecule [Fig . 8] . Although similarin sequence, it should be noted that the P . furiosus proteinis unlikely to carry out the same function as MinD, since itlacks the C-terminal amphipathic helix and therefore is unlikelyto bind to the membrane . It has not been tested in vivo butwas shown to bind very poorly to phospholipid vesicles in vitro[31] . Although the residues in the switch I region are on thesurface in a monomer, they are not readily accessible to thesurface in the postulated dimer model [23] . Thus, these residues are unlikely to interact with MinC directly but may functionas a link to surface residues . This would be similar to whatis postulated for the switch I region of NifH [18] . On the other hand, residue I125 in the switch II region is on the surfacein both the monomer and dimer models and therefore might interactdirectly with MinC.

MinD activation of MinC. The MinCD complex acquires a high affinity for some septal componentupon binding to the membrane [19] . This affinity for a septalcomponent is easy to visualize experimentally provided thata mutant MinC is used that is unable to cause disassembly ofthe Z ring . In contrast, MinCD complexes formed between MinCand MinD mutants unable to bind to the membrane fail to acquirethis high affinity and to decorate the septum [14] . One of theswitch II mutants [E126A] was able to target the GFP-MinC^116-231 fusion; however, it was unable to inhibit division efficiently, since placing the minD-E126A mutation in the context of the min operon resulted in a Min phenotype . There appears to be two possible explanations for this mutant . Either MinD is required for an additional step beyond targeting MinC to the septum orthe MinD-E126A-MinC complex has a lower affinity for the septum,but this is overcome by the expression of the GFP fusion protein,which is visualized at a level higher than the physiologicallevel.

Although MinC or the N-terminal half of MinC is able to antagonize FtsZ assembly in vitro [15], this must represent a basal activityof this inhibitor . In vivo, MinC requires MinD to be an efficientinhibitor . This activation involves MinD bringing MinC to themembrane and targeting it to the septum, steps that require the C-terminal domain of MinC [19] . However, the behavior ofMinD-E126A raises the possibility that MinD may have an additionalrole . It may be required at a step subsequent to targeting,possibly activating the N-terminal domain of MinC followingbinding to the C-terminal domain.

Order of MinD interactions. The binding of MinD to itself, the membrane, and MinC are alldependent upon ATP; however, the order of these interactionsis not clear . Earlier experiments demonstrated that MinD recruitedMinC to vesicles [16, 21] but did not address whether MinD boundto MinC in the absence of vesicles . ATP-dependent but vesicle-independent dimerization of MinD and binding to MinC have been observedby size-exclusion chromatography; however, that analysis wasdone at relatively high protein concentrations [30 µM][16] . Interestingly, we have found that MinD mutants that donot bind to the membrane, such as MinD-I141N [this study] andMinD10 [14], are able to self-interact and bind MinC quite well in the yeast two-hybrid system . These results indicate that these MinD mutants dimerize and bind MinC quite well independentof the membrane.

MinD binding to vesicles and its ATPase both show cooperative behavior suggesting oligomerization [13, 21, 25] . MinD is amonomer in the presence of ADP but undergoes ATP-dependent dimerizationat high concentrations [30 µM] in the absence of phospholipidvesicles [14] . It is quite possible, however, that at lowerconcentrations the membrane promotes MinD dimerization . Evidencefor membrane-dependent dimerization has been obtained from analysisof MinD proteins tagged with different fluorophores . Fluorescenceresonance energy transfer between the differentially taggedMinD proteins was only observed after the addition of phospholipidvesicles to MinD at a concentration of 1 µM [25].

How can we reconcile the observations that MinD may requirea membrane for dimerization and MinC interaction, whereas MinD-I141N and MinD10 do not? One possible explanation may lie in the recentobservation that removal of the amphipathic helix from MinDfacilitates dimerization . It is possible that removing the helix,or possibly affecting its availability [as with MinD-I141N],bypasses the membrane requirement [C . Saez and J . Lutkenhaus,unpublished data] . This result suggests that the C-terminalamphipathic helix masks the dimerization domain in the wild-typemonomer and that the membrane promotes MinD dimerization uponthe addition of ATP, and therefore MinC binding, by sequesteringthe helix away from the dimer interface.

 
This work was supported by grant GM2974 from the National Institutes of Health.

We thank Gerry Lushington for modeling the E . coli MinD structure.

 
* Corresponding author . Mailing address: Department of Microbiology, Molecular Genetics, and Immunology, University of Kansas Medical Center, Kansas City, KS 66160 . Phone: [913] 588-7054 . Fax: [913] 588-7295 . E-mail: jlutkenh@kumc.edu.

1. Bi, E., and J . Lutkenhaus. 1993 . Cell division inhibitors SulA and MinCD prevent formation of the FtsZ ring . J . Bacteriol . 175:1118-1125.
2. Churchward, G., D . Belin, and Y . Nagamine. 1984 . A pSC101-derived plasmid which shows no sequence homology to other commonly used cloning vectors . Gene 31:165-171.
3. Cordell, S . C., R . E . Anderson, and J . Lowe. 2001 . Crystal structure of the bacterial cell division inhibitor MinC . EMBO J . 20:2454-2461 .
4. Cordell, S . C., and J . Lowe. 2001 . Crystal structure of the bacterial cell division regulator MinD . FEBS Lett . 492:160-165.
5. de Boer, P . A., R . E . Crossley, A . R . Hand, and L . I . Rothfield. 1991 . The MinD protein is a membrane ATPase required for the correct placement of the Escherichia coli division site . EMBO J . 10:4371-4380.
6. de Boer, P . A., R . E . Crossley, and L . I . Rothfield. 1989 . A division inhibitor and a topological specificity factor coded for by the minicell locus determine proper placement of the division septum in E . coli. Cell 56:641-649.
7. de Boer, P . A . J., R . E . Crossley, and L . I . Rothfield. 1992 . Roles of MinC and MinD in the site-specific septation block mediated by the MinCDE system of Escherichia coli. J . Bacteriol . 174:63-70.
8. Georgiadis, M . M., H . Komiya, P . Chakrabarti, D . Woo, J . J . Kornuc, and D . C . Rees. 1992 . Crystallographic structure of the nitrogenase iron protein from Azotobacter vinelandii . Science 257:1653-1659.
9. Hayashi, I., T . Oyama, and K . Morikawa. 2001 . Structural and functional studies of MinD ATPase: implications for the molecular recognition of the bacterial cell division apparatus . EMBO J . 20:1819-1828 .
10. Hu, Z., E . P . Gogol, and J . Lutkenhaus. 2002 . Dynamic assembly of MinD on phospholipid vesicles regulated by ATP and MinD . Proc . Natl . Acad . Sci . USA 99:6671-6676.
11. Hu, Z., and J . Lutkenhaus. 1999 . Topological regulation of cell division in Escherichia coli involves rapid pole to pole oscillation of the division inhibitor MinC under the control of MinD and MinE . Mol . Microbiol . 34:82-90.
12. Hu, Z., and J . Lutkenhaus. 2000 . Analysis of MinC reveals two independent domains involved in interaction with MinD and FtsZ . J . Bacteriol . 182:3965-3971 .
13. Hu, Z., and J . Lutkenhaus. 2001 . Topological regulation of cell division in E . coli . Spatiotemporal oscillation of MinD requires stimulation of its ATPase by MinE and phospholipid . Mol . Cell 7:1337-1343.
14. Hu, Z., and J . Lutkenhaus. 2003 . A conserved sequence at the C-terminus of MinD is required for binding to the membrane and targeting MinC to the septum . Mol . Microbiol . 47:345-355.
15. Hu, Z., A . Mukherjee, S . Pichoff, and J . Lutkenhaus. 1999 . The MinC component of the division site selection system in Escherichia coli interacts with FtsZ to prevent polymerization . Proc . Natl . Acad . Sci . USA 96:14819-14824 .
16. Hu, Z., C . Saez, and J . Lutkenhaus. 2003 . Recruitment of MinC, an inhibitor of Z-ring formation, to the membrane in Escherichia coli: role of MinD and MinE . J . Bacteriol . 185:196-203 .
17. Huang, J., C . Cao, and J . Lutkenhaus. 1996 . Interaction between FtsZ and inhibitors of cell division . J . Bacteriol . 178:5080-5085.
18. Jang, S . B., L . C . Seefeldt, and J . W . Peters. 2000 . Insights into nucleotide signal transduction in nitrogenase: structure of an iron protein with MgADP bound . Biochemistry 39:14745-14752.
19. Johnson, J . E., L . L . Lackner, and P . A . de Boer. 2002 . Targeting of ^DMinC/MinD and ^DMinC/DicB complexes to septal rings in Escherichia coli suggests a multistep mechanism for MinC-mediated destruction of nascent FtsZ rings . J . Bacteriol . 184:2951-2962 .
20. Karoui, M . E., and J . Errington. 2001 . Isolation and characterization of topological specificity mutants of minD in Bacillus subtilis . Mol . Microbiol . 42:1211-1221.
21. Lackner, L . L., D . M . Raskin, and P . A . de Boer. 2003 . ATP-dependent interactions between Escherichia coli Min proteins and the phospholipid membrane in vitro . J . Bacteriol . 185:735-749 .
22. Lin-Goerke, J . L., D . J . Robbins, and J . D . Burczak. 1997 . PCR-based random mutagenesis using manganese and reduced dNTP concentration . BioTechniques 23:409-412.
23. Lutkenhaus, J., and M . Sundaramoorthy. 2003 . MinD and role of the deviant Walker A motif, dimerization and membrane binding in oscillation . Mol . Microbiol . 48:295-303.
24. Meinhardt, H., and P . A . de Boer. 2001 . Pattern formation in Escherichia coli: a model for the pole-to-pole oscillations of Min proteins and the localization of the division site . Proc . Natl . Acad . Sci . USA 98:14202-14207 .
25. Mileykovskaya, E., I . Fishov, X . Fu, B . D . Corbin, W . Margolin, and W . Dowhan. 2003 . Effects of phospholipid composition on MinD-membrane interactions in vitro and in vivo . J . Biol . Chem . 278:22193-22198 .
26. Pichoff, S., B . Vollrath, C . Touriol, and J . P . Bouche. 1995 . Deletion analysis of gene minE which encodes the topological specificity factor of cell division in Escherichia coli. Mol . Microbiol . 18:321-329.
27. Ramirez-Arcos, S., J . Szeto, J . A . Dillon, and W . Margolin. 2002 . Conservation of dynamic localization among MinD and MinE orthologues: oscillation of Neisseria gonorrhoeae proteins in Escherichia coli . Mol . Microbiol . 46:493-504.
28. Raskin, D . M., and P . A . de Boer. 1999 . MinDE-dependent pole-to-pole oscillation of division inhibitor MinC in Escherichia coli. J . Bacteriol . 181:6419-6424 .
29. Sakai, N., M . Yao, H . Itou, N . Watanabe, F . Yumoto, M . Tanokura, and I . Tanaka. 2001 . The three-dimensional structure of septum site-determining protein MinD from Pyrococcus horikoshii OT3 in complex with Mg-ADP . Structure [Cambridge] 9:817-826.
30. Shih, Y . L., T . Le, and L . Rothfield. 2003 . Division site selection in Escherichia coli involves dynamic redistribution of Min proteins within coiled structures that extend between the two cell poles . Proc . Natl . Acad . Sci . UsA 100:7865-7870 .
31. Suefuji, K., R . Valluzzi, and D . RayChaudhuri. 2002 . Dynamic assembly of MinD into filament bundles modulated by ATP, phospholipids, and MinE . Proc . Natl . Acad . Sci . USA 99:16776-16781 .
32. Szeto, J., S . Ramirez-Arcos, C . Raymond, L . D . Hicks, C . M . Kay, and J . A . Dillon. 2001 . Gonococcal MinD affects cell division in Neisseria gonorrhoeae and Escherichia coli and exhibits a novel self-interaction . J . Bacteriol . 183:6253-6264 .
33. Szeto, T . H., S . L . Rowland, C . L . Habrukowich, and G . F . King. 2003 . The MinD membrane targeting sequence is a transplantable lipid-binding helix . J . Biol . Chem . 278:40050-40056 .
34. Szeto, T . H., S . L . Rowland, and G . F . King. 2001 . The dimerization function of MinC resides in a structurally autonomous C-terminal domain . J . Bacteriol . 183:6684-6687 .
35. Szeto, T . H., S . L . Rowland, L . I . Rothfield, and G . F . King. 2002 . Membrane localization of MinD is mediated by a C-terminal motif that is conserved across eubacteria, archaea, and chloroplasts . Proc . Natl . Acad . Sci . USA 99:15693-15698 .
36. Yamaichi, Y., and H . Niki. 2000 . Active segregation by the Bacillus subtilis partitioning system in Escherichia coli . Proc . Natl . Acad . Sci . USA 97:14656-14661 .
37. Yu, X . C., and W . Margolin. 1999 . FtsZ ring clusters in min and partition mutants: role of both the Min system and the nucleoid in regulating FtsZ ring localization . Mol . Microbiol . 32:315-326.
38. Zhou, H., and J . Lutkenhaus. 2003 . Membrane binding by MinD involves insertion of hydrophobic residues within the C-terminal amphipathic helix into the bilayer . J . Bacteriol . 185:4326-4335 .

Free Online Full-text Article

What Is Anthrax?, What Is Yeast?, What Is Dna?, What Is Prokaryote?, What Is Cell Biology?, e, Microbes, o, Microbiology, o, Microorganisms, e, Microbe, c, Microorganism, i, Haemophilus, c, Staphylococcus, c, Wastewater, n, Salmonella typhimurium, a, Microorganism, c, Bacteria, e, Clostridia, o, Salmonella typhimurium, o, Campylobacter, r, Schizosaccharomyces, a, Microorganism, o, Cell cultures, a, Schizosaccharomyces, s, Bacillus, a, Escherichia coli, a, Bioreactor, r, Yeasts, a, S. cerevisiae, e, Corynebacterium, c, Pathogenic bacterium, a, Haemophilus




 

© 2005 Transgalactic Ltd (manufacturer of Bioscreen C software) | Privacy Statement | P.O. Box 1393, 00101 Helsinki, Finland, phone: +358 9 85172920, fax: +358 9 8749481, e-mail: microbiology@bionewsonline.com
 

Last modified: May 25, 2005