This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Szeto, J.
Right arrow Articles by Dillon, J.-A. R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Szeto, J.
Right arrow Articles by Dillon, J.-A. R.

 Previous Article  |  Next Article 

Journal of Bacteriology, November 2004, p. 7175-7185, Vol. 186, No. 21
0021-9193/04/$08.00+0     DOI: 10.1128/JB.186.21.7175-7185.2004
Copyright © 2004, American Society for Microbiology. All Rights Reserved.

The N Terminus of MinD Contains Determinants Which Affect Its Dynamic Localization and Enzymatic Activity

Jason Szeto,1 Sudeep Acharya,1 Nelson F. Eng,1 and Jo-Anne R. Dillon1,2*

Department of Biochemistry, Microbiology, and Immunology,1 Centre for Research in Biopharmaceuticals University of Ottawa, Ottawa, Ontario, Canada2

Received 19 February 2004/ Accepted 30 July 2004


arrow
ABSTRACT
 
MinD is involved in regulating the proper placement of the cytokinetic machinery in some bacteria, including Neisseria gonorrhoeae and Escherichia coli. Stimulation of the ATPase activity of MinD by MinE has been proposed to induce dynamic, pole-to-pole oscillations of MinD in E. coli. Here, we investigated the effects of deleting or mutating conserved residues within the N terminus of N. gonorrhoeae MinD (MinDNg) on protein dynamism, localization, and interactions with MinDNg and with MinENg. Deletions or mutations were generated in the first five residues of MinDNg, and mutant proteins were evaluated by several functional assays. Truncation or mutation of N-terminal residues disrupted MinDNg interactions with itself and with MinE. Although the majority of green fluorescent protein (GFP)-MinDNg mutants could still oscillate from pole to pole in E. coli, the GFP-MinDNg oscillation cycles were significantly faster and were accompanied by increased cytoplasmic localization. Interestingly, in vitro ATPase assays indicated that MinDNg proteins lacking the first three residues or with an I5E substitution possessed higher MinENg-independent ATPase activities than the wild-type protein. These results indicate that determinants found within the extreme N terminus of MinDNg are implicated in regulating the enzymatic activity and dynamic localization of the protein.


arrow
INTRODUCTION
 
Bacterial cell division is a complex process involving the recruitment of different proteins to the division site (6). Septation is initiated by FtsZ, a protein that polymerizes into a constrictive ring-like structure (Z-ring) at midcell (1). Proper localization of the Z-ring is determined by nucleoid occlusion (42) and the Min protein system (3). In gram-negative organisms, such as Escherichia coli and Neisseria gonorrhoeae, MinC, MinD, and MinE regulate proper cell division patterns (3, 26, 38). Min system defects in the rod E. coli permit cytokinetic events in cell pole regions, resulting in small, anucleate minicells and rods of various lengths (4). Disruption of min genes in the coccus N. gonorrhoeae results in abnormal cell division, which is characterized by cells with different shapes and sizes, and reduced cell viability (26, 38).

MinD self-associates and interacts with MinC and MinE (15, 16, 18, 38). MinC inhibits FtsZ polymerization (10) and is recruited to the inner cell membrane by the ATP-binding protein MinD (11, 15, 18, 30). Green fluorescent protein (GFP)-E. coli MinD (MinDEc) fusions have been shown to undergo MinE-dependent pole-to-pole oscillations in E. coli (29, 31). Without MinE, GFP-MinDEc localizes along the entire inner cell membrane and does not oscillate (13). MinD oscillation presumably shuttles MinC to the inner membrane along one half of the cell to the other half, allowing Z-ring formation to occur exclusively at midcell (11, 29, 30). While it is involved in inhibiting the Z-ring, MinC plays no part in the oscillation of MinD (29, 31).

MinD dimerization is ATP dependent (15), and in vitro studies have demonstrated that MinD can bind to phospholipid vesicles in the presence of ATP (13, 18, 21). The extreme C terminus of MinD proteins consists of a conserved 8- to 12-residue sequence predicted to form an amphipathic helix that allows MinD to associate with the membrane (14, 39, 40). How ATP binding permits such an association is unclear at this time(14, 39, 40). MinE is able to stimulate MinD ATPase activity severalfold in the presence of phospholipids (12), resulting in the dissociation of MinD from lipid vesicles (13, 18, 37). It has been proposed that, in vivo, this ATPase stimulation allows MinD to dissociate from the membrane at one cell pole and begin migration towards the opposite cell half (13). Recently, fluorescent protein fusions to MinDEc, as well as MinEEc, were observed to localize as polymeric coiled arrays that extended the length of E. coli (35). This suggested that a portion of Min proteins may act as a basic scaffold for the incorporation of additional Min subunits into either cell half (35).

While the C terminus of MinD is involved in membrane association (14, 39, 40, 43), little is known about the role of its N terminus. The most characterized N-terminal region of MinD is its highly conserved Walker A ATP-binding motif (amino acids [aa] 10 to 17 in MinDEc and N. gonorrhoeae MinD [MinDNg]) (5, 38). In particular, mutation of the K16 residue disrupted MinD ATPase activity (13), protein dimerization (27, 43), and MinD association with membranes in vivo and in vitro (13, 27).

The ubiquity of MinD function, specifically among gram-negative organisms, has been demonstrated previously by observing that MinDNg behaves like MinDEc. Fusions of GFP to either protein exhibit MinE-dependent oscillations in E. coli (27, 29, 31). Overexpression of either MinDNg or MinDEc induces cell division arrest in E. coli (4, 38). Yeast two-hybrid assays revealed that both proteins can homodimerize (38). Furthermore, overexpression of MinDNg can partially complement an E. coli MinD mutant (38). Hence, there is definite functional similarity between the N. gonorrhoeae and E. coli MinD proteins (27, 38).

It has also been demonstrated previously that a 4-aa N-terminal truncation in MinDNg abrogated membrane localization and oscillation of GFP-MinDNg expressed in E. coli (27). The 4-aa deletion also disrupted MinDNg self-interaction and the interaction of this protein with MinENg (27). This phenotype was similar to that resulting from a K16Q mutation in the Walker A ATP-binding motif (27). Since the 4-aa truncated and K16Q mutants were generated by altering close, albeit separate, regions of MinDNg, their common phenotypes suggested that the N terminus of MinD may comprise a defined functional region. Hence, the hypothesis of the present study was that mutations of conserved N-terminal residues should produce phenotypes similar to those produced by the 4-aa N-terminal truncation and K16Q mutation.

Deletions or substitutions were made in conserved residues within the N terminus of MinDNg. As expected, most of these mutations affected the MinDNg interaction with itself and with MinENg. However, in contrast to a previous study with the 4-aa N-terminal truncation and K16Q mutants of MinDNg (27), GFP fusions to most MinDNg variants displayed significantly faster oscillation than the wild-type protein, and this was coupled with increased propensities to remain in the cytoplasm. Furthermore, MinDNg proteins lacking the first three residues or having an I5E mutation had increased MinE-independent ATPase activity which was not further stimulated by MinENg, compared to wild-type MinDNg. Hence, mutation of the extreme N terminus of MinDNg affects determinants involved in regulating protein dynamism and ATPase activity, in contrast to the classic K16Q Walker A motif mutation, which completely abrogates MinD movement, membrane localization, and ATPase activity. These studies showed that the extreme N terminus of MinD is functionally distinct from the K16 Walker A residue, despite their proximity in the sequence.


arrow
MATERIALS AND METHODS
 
Strains and growth conditions. The strains used in this study are shown in Table 1. E. coli DH5{alpha} was used as a host strain for DNA cloning. E. coli PB114 ({Delta}minCDE) (3) was used in GFP-MinDNg localization studies. E. coli C41(DE3) was used as an expression strain for the purification of His-tagged MinDNg and MinENg proteins. Most E. coli strains were grown at 37°C in Luria-Bertani medium (Difco); the only exception was E. coli PB114, which was grown at 30°C and induced with 40 µM isopropyl-ß-D-thiogalactopyranoside (IPTG) as described previously (27). N. gonorrhoeae CH811 was cultured as described previously (17, 23, 38). Gonococcal cells suspended in deionized distilled water (0.5 McFarland equivalent turbidity standard; Remel) served as a source of template DNA for PCR. When required, media were supplemented with ampicillin at a concentration of 100 µg/ml or kanamycin at a concentration of 50 µg/ml. Saccharomyces cerevisiae SFY526 was used in yeast two-hybrid assays and was grown at 30°C on yeast extract-peptone-adenine-dextrose medium or on the appropriate synthetic dropout media as described by the manufacturer (Clontech).


View this table:
[in this window]
[in a new window]
  		TABLE 1. Strains and plasmids used

Construction of GFP fusions to MinDNg mutants. DNA sequences of all primers used in this study are available upon request. For localization studies, GFP fusions to wild-type and mutant MinDNg were constructed. Plasmids pSR15 (gfp-minDNg minENg) and pJDE1 (gfp-minDNg-K16Q minENg) (Table 1) were constructed previously (27) and were used as positive and negative controls, respectively, for GFP-MinDNg localization. Primers were designed to place the natural GTG start codon of minDNg directly upstream of the third and fourth codons, respectively. This generated minDNg genes with the first two or three codons (including the start codon [i.e., codon 1]) deleted and replaced with a single start codon. Since it was necessary to reintroduce the start codon to preserve protein translation, MinDNg-2aaNT and MinDNg-3aaNT essentially contained in-frame deletions of one and two residues, respectively. In addition, primers that incorporated K3E, K3I, I4Q, I5E, or I5A point mutations in MinDNg were also used. Each minDNg primer was used with an accompanying downstream primer that annealed to minENg to amplify mutant minDNg and minENg from the gonococcal chromosomal DNA template. All DNA amplicons were subsequently cloned into pDSW209 (Table 1) as previously described (27) to generate plasmids encoding specific gfp-minDNg variants and minENg (Table 1). These plasmids included pSIA16 (GFP-MinDNg-2aaNT), pSIA17 (GFP-MinDNg-3aaNT), pSIA18 (GFP-MinDNg-K3E), pJS30 (GFP-MinDNg-K3I), pJS18 (MinDNg-I4Q), pJS29 (GFP-MinDNg-I5E), and pJS19 (GFP-MinDNg-I5A) (Table 1). Plasmid pJS10, encoding only GFP-MinDNg (without MinENg), was also constructed (Table 1).

Microscopy. For GFP fusion localization studies, IPTG-induced E. coli PB114 cells were immobilized on coverslips as previously described (27). To maintain consistency, the oscillation cycles of each GFP-MinDNg fusion were measured in 30 cells with nearly equal lengths (~2.0 to 2.5 µm). In each cell observed, at least two complete oscillation cycles were used to verify the oscillation periods. Statistical analyses with unpaired Student's t tests were performed to determine whether differences in average oscillation times between GFP-MinDNg fusions were significant (P < 0.001). Fluorescence microscopy was performed with an Olympus BX61 microscope equipped with a Photometrics CoolSnap ES camera and the Image Pro software (version 5.0). Time-lapse images were taken every 5 s for all GFP fusion localization experiments. When required, raw images were enhanced by using standard options available in the Image Pro software (version 5.0). Contrast enhancement was used to increase contrast and to decrease gamma, and Gauss filtering using a 3 x 3 kernel size with four passes at strength 10 was used to help visualize intracellular substructures.

Construction of yeast two-hybrid vectors and yeast two-hybrid assays. Yeast two-hybrid vectors containing wild-type minDNg or minENg (pGADminD, pGBT9minD, and pGBT9minE) were constructed previously (Table 1) (27, 38). Similarly, mutant and truncated minDNg genes were also cloned into the yeast two-hybrid vectors pGAD424 and pGBT9 (Table 1). The vectors encoding fusions to the GAL4 DNA-binding domain included pSIA1 (MinDNg-2aaNT), pSIA2 (MinDNg-3aaNT), pSIA15 (MinDNg-K3E), pJS26 (MinDNg-K3I), pJS22 (MinDNg-I4Q), pJS24 (MinDNg-I5E), and pJS21 (MinDNg-I5A) (Table 1). The vectors encoding fusions to the GAL4 activation domain included pSIA3 (MinDNg-2aaNT), pSIA4 (MinDNg-3aaNT), pSIA14 (MinDNg-K3E), pJS25 (MinDNg-K3I), pSIA28 (MinDNg-I4Q), pJS11 (MinDNg-I5E), and pJS20 (MinDNg-I5A) (Table 1). Yeast two-hybrid experiments and ß-galactosidase assays were conducted as previously described (38).

Western blotting. GFP-MinDNg and MinENg protein levels in E. coli were monitored by Western blotting as described previously (27).

Protein purification. To purify MinDNg, wild-type minDNg was PCR amplified from N. gonorrhoeae CH811 chromosomal DNA and cloned in frame with a C-terminal six-His tag in pET30a (Novagen) to produce pSC9 (Table 1). We used a short six-His tag fused to the C terminus of MinDNg since we found that the protein yields were higher than those obtained with N-terminal tags. Plasmids encoding C-terminal His-tagged MinDNg-I5E (pJS33), MinDNg-3aaNT (pJS34), and MinDNg-K16Q (pSC10) were also cloned similarly into pET30a for protein purification (Table 1). Plasmid pEC1, encoding C-terminal His-tagged MinENg, was constructed previously (27) (Table 1).

Each plasmid was transformed into E. coli C41(DE3) for protein overexpression (Table 1). Purification of His-tagged MinDNg proteins was carried out as described previously (38); however, 50-ml cultures and 1 ml of His-Bind resin (Novagen) were used for each protein. The following solutions, described by the manufacturer (Novagen), were used for column washing: 10 ml of binding buffer (5 mM imidazole), 5 ml of wash solution (60 mM imidazole), and 1 ml of elution buffer (1 M imidazole). The eluted protein was dialyzed overnight at 4°C against buffer A (20 mM Tris [pH 7.5], 2 mM EDTA, 200 mM NaCl, 10% glycerol; final pH adjusted to 7.4).

For MinENg purification, 0.4 mM IPTG was used to induce a 350-ml log-phase culture of E. coli C41(DE3)(pEC1) transformants for 2 to 3 h at 250 rpm and 37°C. The soluble cell extract fraction was applied to 3 ml of His-Bind resin, the column was washed with buffers containing increasing concentrations of imidazole (5 to 100 mM; Novagen), and MinENg-His was eluted with 250 mM imidazole buffer. Purified protein was dialyzed in buffer B, which consisted of 50 mM Tris, 20 mM NaCl, and 1 mM EDTA (pH 7.4). MinENg-His was concentrated by using Biomax-5 centrifugal filter columns with a 5,000-molecular-weight cutoff (Millipore).

MinDNg ATPase stimulation assays. A rapid protocol for the formation of vesicles composed of anionic 1,2-dioleoyl-sn-glycero-3-[phospho-rac-(1-glycerol)] (phosphatidylglycerol [PG]) (Avanti Polar Lipids, Inc.) was developed. PG (10 mg/ml in chloroform) was dried under a stream of filtered air and resuspended in reaction buffer (25 mM Tris-Cl, 50 mM KCl; pH 7.5) to obtain a stock 5-mg/ml PG vesicle solution. PG vesicles were visualized by phase-contrast microscopy to verify their presence and stored at –20°C. As determined by phase-contrast microscopy, PG vesicles prepared in this manner appeared to be stable for at least 4 months.

Prior to each experiment, the concentrations of each MinNg protein were determined by using the Bio-Rad protein assay and were adjusted as required. In a typical 100-µl reaction mixture, the following reagents were added (final concentrations are indicated in parentheses): reaction buffer (the volume added depended on the other reagents included), MinDNg protein (0.012 mg/ml), ATP (1 mM), PG vesicles (0.5 mg/ml), and MgCl2 (1 mM). The reaction mixtures were incubated at room temperature for 5 min, and, if required, MinENg (0.012 mg/ml) was added, which was followed by an additional 5 min of incubation. Storage buffers for MinDNg or MinENg were used to adjust the reaction volumes to 100 µl when protein was not used. At specified times, 30 µl of each reaction mixture was removed and centrifuged at 10,000 x g for 1 min, and 25 µl of supernatant was removed and added to 50 µl of a malachite green solution in a microtiter dish, similar to the procedure described previously (8). Color development was allowed to proceed for 20 min at room temperature, and the absorbance of the reaction at 620 nm was determined with a TECAN Spectra Shell microplate reader (TECAN U.S. Inc.) and compared to the absorbance of a blank containing all the reagents in a standard reaction mixture, except that storage buffers were used in place of each protein. The amount of inorganic phosphate released was determined by comparing the absorbance values with the values obtained with inorganic phosphate standards prepared from dilutions of KH2PO4 in the blank buffer.


arrow
RESULTS
 
N-terminal truncations and point mutations in MinDNg affect dynamic localization of GFP-MinDNg fusions. Sequence alignment of MinD proteins revealed that the extreme N terminus is conserved (Fig. 1), suggesting that it is important for protein function. Gram-negative organisms have the consensus sequence M(A/G)(K/R)(I/V)(I) for the first five residues. Of particular note is that this N-terminal region is identical in MinD proteins from N. gonorrhoeae and E. coli, except that MinDEc contains an arginine residue instead of a lysine residue at position 3. Although our laboratory is focused on the study of Min proteins from the coccus N. gonorrhoeae, it has been demonstrated that MinDNg is functionally similar to E. coli MinD (27, 38). Hence, many functional studies with MinDNg have been facilitated by using the rod E. coli (26, 27, 28, 38), particularly since the gonococcus is much more fastidious and less amenable to genetic manipulation than E. coli. MinDNg and MinDEc exhibit high levels of identity (73%) and similarity (85%); therefore, structure-function analyses of one protein can serve as a paradigm for the other.



View larger version (78K):
[in this window]
[in a new window]
  		FIG. 1. Sequence alignment of the N termini of MinD proteins. Abbreviations: Mj, Methanococcus jannaschii; Af, Archaeoglobus fulgidus; Pf, Pyrococcus furiosus; Ph, Pyrococcus horikoshii; Ng, Neisseria gonorrhoeae; Nm, Neisseria meningitidis; Ec, Escherichia coli; St, Salmonella enterica serovar Typhimurium;Yp, Yersinia pestis; Vc, Vibrio cholerae; Pa, Pseudomonas aeruginosa; Bm, Brucella melitensis; Bsui, Bacillus suis; At, Agrobacterium tumefaciens; Hp, Helicobacter pylori; Aa, Aquifex aeolicus; Tm, Thermotoga maritima; Bs, Bacillus subtilis; Lm, Listeria monocytogenes; Cp, Clostridium perfringens; Sy, Synechocystis sp.; Gt, Guillardia theta; Dr, Deinococcus radiodurans; Ct, Chlamydia trachomatis. Residues of the Walker A ATP-binding motif are aligned below the solid bar.

It has been shown previously that deletion of the first four N-terminal residues of MinDNg localized a GFP-MinDNg fusion to the cytosol and made it unable to oscillate in E. coli (27). To further characterize how the N terminus of MinDNg is involved in protein function, N-terminal deletions in minDNg were constructed. In order to preserve protein translation, a single start codon was used to replace the deleted codons. The resulting proteins were designated MinDNg-2aaNT and MinDNg-3aaNT, where 2aaNT and 3aaNT indicate that two and three amino acids (or codons), respectively, were removed from the N terminus and replaced with a start codon. Amino acid substitutions were also made to alter the charge and/or side chain bulkiness of selected N-terminal residues. These mutations included K3E, K3I, K3A, I4Q, I5A, and I5E. Each MinDNg variant was fused to GFP to determine its dynamic behavior in E. coli. MinENg was encoded on the same plasmid as each GFP-MinDNg fusion to induce in vivo movement. Localization studies were performed with E. coli PB114 ({Delta}minCDE) (Table 1); hence, interference from endogenous Min proteins was avoided.

Positive control cells expressing wild-type GFP-MinDNg displayed dynamic intracellular fusion protein movement that clearly alternated along the inner cell periphery from one cell pole to the other (Fig. 2A), and the average oscillation period was 33.2 ± 5.8 s (Table 2). GFP-MinDNg-2aaNT also exhibited an oscillation pattern similar to that of the positive control (Fig. 2C; Table 2).



View larger version (91K):
[in this window]
[in a new window]
  		FIG. 2. Localization of wild-type and N-terminal deletion derivatives of MinDNg in E. coli PB114. Oscillation cycles (from one pole to the other and back) of GFP-MinDNg fusions were measured by using E. coli rods that were 2.0 to 2.5 µm long. (A) Distinct pole-to-pole movement of wild-type GFP-MinDNg in E. coli. In the cell shown, one cycle of fusion protein movement required 30 s. Note the U-shaped fluorescent signal of the fusion protein alternately lining each cell polar region. The left panel shows a differential interference contrast image, and the remaining panels show corresponding fluorescence images. Bar = 5 µm (the magnifications for all other images are similar). (B) GFP-MinDNg localizes in longer E. coli cells as regularly spaced bands. Each band contains the fusion protein arranged within adjacent polymeric arrays (arrows) that are suggestive of a coil-like structure. (C) GFP-MinDNg-2aaNT exhibits pole-to-pole oscillation (arrows). The time required for the fusion protein to complete one oscillatory cycle in this cell was 30 s. (D) GFP-MinDNg-3aaNT can also exhibit pole-to-pole oscillation (arrows), and one cycle required only 15 s in the cell shown. However, much of the fusion protein signal is distributed throughout the cytosol, making visualization of GFP-MinDNg-3aaNT oscillation more difficult. (E) Raw image of E. coli PB114 expressing GFP-MinDNg-3aaNT. Note the nearly uniform cytosolic localization of the protein. (E') Image enhancement of panel E, revealing the presence of GFP-MinDNg-3aaNT localizing within polymeric bands. (F) In the absence of MinENg, GFP-MinDNg localizes along the entire inner cell periphery with no evidence of oscillation or polymeric arrays. (G) Western blot obtained by using anti-MinDNg antisera to detect GFP-MinDNg proteins in E. coli PB114. Lane 1, cell extract from untransformed E. coli PB114; lane 2, wild-type GFP-MinDNg (pSR15); lane 3, GFP-MinDNg-2aaNT (pSIA16); lane 4, GFP-MinDNg-3aaNT (pSIA17). (H) Western blot obtained by using anti-MinENg antisera to detect MinENg in E. coli PB114 transformed with no plasmid (lane 1), pSR15 (lane 2), pSIA16 (lane 3), or pSIA17 (lane 4).


View this table:
[in this window]
[in a new window]
  		TABLE 2. Dynamic localization characteristics of GFP fusions to MinDNg N-terminal mutants

At first glance, GFP-MinDNg-3aaNT did not appear to move in E. coli PB114 cells, and much of the fluorescent signal was distributed throughout the cytoplasm (Fig. 2D). However, closer inspection revealed faint oscillatory signals shifting from one end of the cell to the other (Fig. 2D), amidst the increased cytosolic fluorescence relative to that of wild-type GFP-MinDNg and GFP-MinDNg-2aaNT. Interestingly, the average oscillation time of GFP-MinDNg-3aaNT (19.8 ± 3.8 s) was significantly less than that of GFP fusions to both the wild-type protein and MinDNg-2aaNT (P < 0.001) (Table 2).

E. coli GFP-MinD has been shown to localize in a MinE-dependent manner within a membrane-associated coiled array (35). Distinct fluorescent polymeric bands that were suggestive of GFP-MinDNg localizing in coiled segments were visualized in cells expressing GFP-MinDNg (Fig. 2B). In general, GFP-MinDNg-decorated coiled arrays appeared to be more easily visualized in elongated E. coli cells (length, >5 µm). Similar to wild-type GFP-MinDNg, GFP-MinDNg-2aaNT was also localized within bands that suggested the presence of a polymeric array (Table 2; data not shown). As expected, in the absence of MinENg, wild-type GFP-MinDNg was localized along the entire inner cell periphery, and there was no evidence of oscillation or distribution within polymeric arrays (Fig. 2F).

Despite the increased cytosolic distribution of GFP-MinDNg-3aaNT (Fig. 2E), image contrast enhancement revealed that a small proportion of fusion protein could also localize in polymeric structures suggestive of a helical array (Fig. 2E'; Table 2). Hence, despite a tendency to localize mainly in the cytosol, GFP-MinDNg-3aaNT retained the ability to localize within an intracellular array, likely permitting the faint oscillation patterns observed. As expected, the negative control GFP-MinDNg-K16Q was distributed throughout the cytoplasm (27), and there was no evidence of localization as polymeric arrays (data not shown).

Western blotting with anti-MinDNg antisera verified the overexpression of wild-type GFP-MinDNg (Fig. 2G, lane 2), GFP-MinDNg-2aaNT (lane 3), and GFP-MinDNg-3aaNT (lane 4) in E. coli transformants relative to the expression in untransformed cells (lane 1). Densitometric analysis indicated that the GFP-MinDNg-2aaNT and GFP-MinDNg-3aaNT levels were ~92% and ~86% of the wild-type GFP-MinDNg level, respectively. Western blotting detected similar levels of MinENg in all E. coli PB114 transformants used in these studies (Fig. 2H, lanes 2, 3, and 4) and did not detect any MinENg in untransformed cell extract (lane 1).

GFP fusions to MinDNg bearing K3E, K3I, I4Q, I5A, and I5E substitutions were also examined in E. coli PB114. Interestingly, as point mutations progressed from the third to the fourth residue of MinDNg, there was a corresponding increase in the tendency of each mutant (K3E, K3I, and I4Q mutants) to oscillate significantly faster than the wild type (Fig. 3A; Table 2). In addition, there were differences between the mutant fusions as well, and the average oscillation cycle of GFP-MinDNg-I4Q was significantly faster than that of GFP-MinDNg-K3E (P < 0.001) but not faster than that of GFP-MinDNg-K3I (Table 2). Hence, the average oscillation cycles of GFP-MinDNg fusions can be represented as follows: wild-type MinDNg > MinDNg-K3E = MinDNg-K3I > MinDNg-I4Q (Table 2). Similar to the findings for GFP-MinDNg-3aaNT, the faster oscillation of the MinDNg point mutants was accompanied by increased tendencies to remain in the cytosol (particularly for GFP-MinDNg-K3I and GFP-MinDNg-I4Q) (Fig. 3A and B; Table 2) and, hence, lowered tendencies to remain associated along the membrane in vivo. Image contrast enhancement showed that GFP fusions to MinDNg containing K3E, K3I, and I4Q could still localize as distinct bands, which was suggestive of dynamic movement along a polymeric array (Fig. 3B' and Table 2).



View larger version (49K):
[in this window]
[in a new window]
  		FIG. 3. Localization of GFP-MinDNg-I4Q in E. coli PB114. (A) Raw images of E. coli PB114 expressing GFP-MinDNg-I4Q. Bar = 5 µm (the magnifications for all other images are similar). Note the increased cytosolic localization of GFP-MinDNg-I4Q relative to wild-type GFP-MinDNg (Fig. 2A). GFP-MinDNg-I4Q required only 15 s to complete one cycle of oscillation. (B) Localization of GFP-MinDNg-I4Q in a longer E. coli cell. Note the increased cyosolic distribution of fluorescent signal relative to that of wild-type GFP-MinDNg (Fig. 2B). (B') Image enhancement of panel B, showing that GFP-MinDNg-I4Q can still localize within bands, which is suggestive of a polymeric array (arrows). (C) GFP-MinDNg-I5A is mostly localized in the cytosol. (C') Contrast enhancement of the image in panel C, showing that GFP-MinDNg-I5A retains the ability to localize as bands (arrows). (D) GFP-MinDNg-I5E is uniformly distributed throughout the cytoplasm. (E) Western blot obtained by using anti-MinDNg antisera to detect GFP-MinDNg in E. coli PB114 transformed with wild-type GFP-MinDNg (pSR15) (lane 1), GFP-MinDNg-K3E (pSIA18) (lane 2), MinDNg-K3I (pJS30) (lane 3), GFP-MinDNg-I4Q (pJS18) (lane 4), GFP-MinDNg-I5E (pJS29) (lane 5), GFP-MinDNg-I5A (pJS19) (lane 6), or no plasmid (lane 7). (F) Western blotting to detect MinENg in E. coli transformed with no plasmid (lane 1), pSR15 (lane 2), pSIA18 (lane 3), pJS30 (lane 4), pJS18 (lane 5), pJS29 (lane 6), or pJS19 (lane 7).

GFP-MinDNg-I5A tended to remain almost entirely in the cytoplasm (Fig. 3C). While evidence of faint fusion protein movement in bands could be discerned in longer filamentous E. coli PB114 cells (Table 2), it was not possible to clearly measure pole-to-pole movement in shorter E. coli rods due to the increased cytosolic fluorescence signal (data not shown). However, evidence of GFP-MinDNg-I5A localizing as bands (Fig. 3C') suggested that it localized as a polymeric structure(s). Finally, GFP-MinDNg-I5E appeared to be wholly distributed in the cytosol, presenting uniform fluorescence throughout the cell, and there was no evidence of dynamic movement or polymeric array formation (Fig. 3D; Table 2), similar to the negative control GFP-MinDNg-K16Q (data not shown). Western blotting confirmed that similar levels of GFP-MinDNg fusions (Fig. 3E) and MinENg (Fig. 3F) were found in each E. coli PB114 transformant used in these studies.

Hence, as mutations progressed from the third to the fifth residue of MinDNg, there was an accompanying decrease in the time required for GFP-MinDNg to oscillate from pole to pole and an apparent decrease in the ability of the protein to associate with the E. coli inner cell periphery and/or membrane-associated polymer. This was accompanied by increased tendencies of GFP-MinDNg mutants to remain in the cytosol.

Amino acid deletions or mutations in the extreme N terminus of MinDNg affect dimerization and interaction with MinENg. The yeast two-hybrid system was used to determine whether N-terminal deletions or point mutations affected interactions of MinDNg with itself and with MinENg. Similar to wild-type MinDNg (38), MinDNg-2aaNT could self-associate (Table 3). Furthermore, MinDNg-2aaNT retained a strong interaction with MinENg relative to the wild-type control (Table 3). In contrast, MinDNg-3aaNT did not interact with itself and had a significantly diminished interaction with MinENg (Table 3). The MinDNg point mutants were also assayed for Min protein interactions. Each of the mutants except MinDNg-I5E retained the ability to self-associate (Table 4); however, the strength of the mutant MinDNg self-interaction, as assessed by ß-galactosidase activities, was decreased relative to that of wild-type MinDNg (Table 4). Most of the mutants also had noticeably less interaction with MinENg, as determined by yeast ß-galactosidase assays; the exceptions were MinDNg-K3E and MinDNg-I4Q (Table 4).


View this table:
[in this window]
[in a new window]
  		TABLE 3. Yeast two-hybrid assays to determine self-association and MinE binding of N-terminal deletion derivatives of MinDNg


View this table:
[in this window]
[in a new window]
  		TABLE 4. Yeast two-hybrid assays to determine self-association and MinE binding of N-terminal point mutants of MinDNg

Since each of the N-terminal mutants except MinDNg-I5E interacted at least with MinENg or with itself, the data suggest that these MinDNg mutants were likely folded properly, without gross structural perturbations. Overall, our results show that residues in the extreme N terminus of MinDNg are implicated in MinDNg dimerization and binding of MinENg.

N-terminal mutation derivatives of MinDNg exhibit increased basal ATPase activity that is not affected by MinENg. The increasingly rapid oscillation of the MinDNg N-terminal mutants was surprising since decreasing and/or disrupting the MinD-MinE interaction should interfere with MinD pole-to-pole motion, as demonstrated with E. coli MinE mutants (20). Hence, to elucidate a possible biochemical explanation for these observations, ATPase stimulation assays were conducted with purified MinDNg-3aaNT, a mutant with increased cytosolic distribution and faster oscillation, and MinDNg-I5E, a mutant which failed to oscillate and which was entirely cytosolic. MinDNg-K16Q served as a negative control in these experiments, since it has disrupted ATPase activity, even in the presence of MinENg and phospholipids (J. Szeto, N. F. Eng, S. Acharya, M. D. Rigden, J. R. Dillon, Res. Microbiol., in press).

The short C-terminal His tag added to MinDNg did not affect the MinDNg function, since His-tagged wild-type MinDNg could induce cell division arrest when it was overexpressed in wild-type E. coli, while His-tagged MinDNg-K16Q could not, as expected (data not shown). His-tagged MinENg was also active and caused an obvious minicell phenotype when it was overexpressed in E. coli (data not shown), similar to overexpression of untagged E. coli MinE (4).

E. coli MinD was previously shown to preferentially bind anionic phospholipids (21, 40); therefore, PG vesicles were made and used for the ATPase assays. Purified wild-type MinDNg, MinDNg-3aaNT, MinDNg-I5E, MinDNg-K16Q, and MinENg were incubated in the presence of ATP and PG vesicles to measure the release of inorganic phosphate (Pi) from ATP. In the absence of MinENg, wild-type MinDNg displayed a low level of enzymatic activity (Fig. 4). Addition of MinENg increased the amount of Pi released from wild-type MinDNg to levels that were up to three times the basal level after 90 min (Fig. 4). In contrast, the MinDNg-K16Q negative control exhibited little ATPase activity irrespective of the presence of MinENg (Fig. 4).



View larger version (24K):
[in this window]
[in a new window]
  		FIG. 4. MinDNg ATPase stimulation assays. Equal amounts of purified MinDNg, MinDNg-3aaNT, MinDNg-I5E, and MinDNg-K16Q were incubated with PG vesicles and 1 mM ATP. The ATPase activities of each mixture were tested in the presence and absence of MinENg over a 90-min period. Inorganic phosphate released due to ATP hydrolysis was monitored by using a malachite green-based method. Buffer A was MinDNg storage buffer, and buffer B was MinENg storage buffer. WT, wild type.

In the absence of MinENg, both MinDNg-3aaNT and MinDNg-I5E exhibited markedly greater ATPase activities than the basal wild-type MinDNg protein (Fig. 4). After 45 min, wild-type MinDNg had not released any measurable amount of Pi in the absence of MinENg, and both MinDNg-3aaNT and MinDNg-I5E had released significantly more Pi (>350 pmol), despite the lack of MinENg (Fig. 4). By the end of the assay, both mutants had released approximately twice as much Pi as the wild type (Fig. 4). The addition of MinENg did not result in any significant increase in the ATPase activities of MinDNg-3aaNT and MinDNg-I5E (Fig. 4).

Therefore, the 3-aa N-terminal truncation and I5E mutations in MinDNg that decreased dynamic oscillatory periods and/or promoted cytosolic localization in vivo also conferred higher MinENg-independent ATPase activity to these proteins compared to wild-type MinDNg activity, which was not further increased by the addition of MinENg. These studies indicate there are determinants at the extreme N terminus of MinDNg involved in regulating the localization and intrinsic enzymatic activity of the protein.


arrow
DISCUSSION
 
Here we describe for the first time MinD proteins that have increased oscillation rates despite having diminished interactions with MinE and the inability of MinE to further stimulate the ATPase activity of MinD. Results from a previous study with 4-aa N-terminal truncation and K16Q mutants suggested that mutations and truncations in the N terminus of MinDNg would result in similar protein phenotypes, as exemplified by disrupted MinD self-association and interaction with MinE, abrogated pole-to-pole oscillation, and lost membrane localization (27). In the present study, we found that mutations in the extreme N terminus of MinDNg do not produce phenotypes that resemble the phenotype of the K16Q Walker A motif mutant. Most of our N-terminal MinDNg mutants still exhibited dynamic movement, interacted with themselves and/or MinE, and, in the case of MinDNg-3aaNT and MinDNg-I5E, possessed increased MinE-independent ATPase activity compared to the wild-type protein. Thus, this study extended our previously limited knowledge of the extreme N terminus of MinD, and showed that this region is functionally distinct from the proximal Walker A ATP-binding motif residue K16.

It has been proposed that under normal circumstances, the MinE-induced ATPase activity of MinD is the rate-limiting step for GFP-MinD oscillation (12). A few factors have been shown to affect the MinD oscillation frequency in vivo. These include mutations within MinE or changes in intracellular MinD/MinE ratios (12, 29). Inadequate stimulation of E. coli MinD ATPase activity by MinEEc mutants was shown to slow the oscillation of GFP-MinDEc (12). Furthermore, mutations in MinEEc that disrupted its binding to MinDEc resulted in a significantly lower rate of GFP-MinDEc movement from cell pole regions (20). Strikingly, in the present study, we found several mutations in MinDNg that also diminished its interaction with MinENg but resulted in significantly faster oscillation than wild-type MinDNg. Although the oscillation period of wild-type GFP-MinDNg in this study was significantly shorter than that in a previous study (27), the difference was most likely due to the shorter time interval and automation used to record fusion protein oscillations in the present study (5 s, compared to the 15-s intervals used in the previous study [27]). In addition, the use of different E. coli ({Delta}minCDE) strains (strains PB114 and WM1032 in the present and previous studies, respectively) and the strict limitations on cell length (2.0 to 2.5 µm) may have also resulted in differences in oscillation periods compared to those in the previous study.

An increased MinD/MinE ratio has also been shown to extend oscillation cycles, while a decreased ratio has been proposed to have the opposite effect (29). In the present study, the intracellular levels of all GFP-MinDNg variants (except GFP-MinDNg-3aaNT) and of the coexpressed MinENg protein were similar; hence, the decreased oscillation periods of our MinDNg mutants were due not to gross changes in MinD/MinE ratios but rather to changes intrinsic to MinDNg.

We believe that the observed changes in the dynamism of N-terminal MinDNg mutants were not due to diminished interactions with MinENg for several reasons. First, almost all of our MinDNg mutants had faster oscillation than the control, regardless of whether they interacted strongly or weakly with MinENg. Second, it has been demonstrated that diminished or abrogated MinD interactions with mutant MinE proteins cause GFP-MinDEc (20) or GFP-MinDNg (Eng and Dillon, unpublished results) to oscillate more slowly, presumably due to inefficient stimulation of MinD ATPase. However, in the present study, it was clear that MinDNg mutants having diminished interactions with MinENg had significantly faster oscillation than wild-type MinDNg. Hence, this finding supports a MinENg-independent mechanism responsible for the observed localization phenotypes. Third, the increased ATPase activity of two of the mutants, MinDNg-3aaNT and MinDNg-I5E, was not affected by the presence MinENg. Hence, the phenotypes of the mutant MinDNg proteins were likely MinENg independent. Finally, while N-terminal mutations in MinDNg did affect the binding of MinENg, the residues are probably not directly involved in this process, as yeast two-hybrid assays and protein modeling by other workers have indicated that there is a more direct role for the distal {alpha}-7 helix of MinD (approximately aa 146 to 162 in MinDEc) in its interaction with MinE (G. F. King, personal communication). Although we also have evidence that mutations within the N terminus of MinDNg affect the interaction with MinC (data not shown), this could not possibly have contributed to the faster oscillation and increased cytosolic localization of the GFP-MinDNg variants since all localization studies were conducted in an E. coli system devoid of MinC. In addition, it has been established that GFP-MinD membrane association and oscillation are independent of MinC (29).

What might account for the marked changes in intracellular MinDNg dynamism? The MinE-independent ATPase activities of MinDNg-3aaNT and MinDNg-I5E offer a possible explanation. We propose that the increased basal ATPase activity of these two mutants can account for their faster oscillation cycles and/or increased cytosolic distribution relative to wild-type MinDNg. Normally, the MinE-stimulated ATPase activity of MinD is required for dissociation of the latter protein from membranes (13). As a result of increased intrinsic ATPase activity at the membrane, it is possible that GFP-MinDNg-3aaNT and GFP-MinDNg-I5E were more prone to dissociate from their membrane targets regardless of MinENg. It is also possible that the other MinDNg N-terminal mutants constructed in the present study possess similar MinE-independent ATPase activity. As a result, GFP fusions to N-terminal MinDNg mutants would be released sooner into the cytosol than wild-type GFP-MinDNg, allowing them to migrate towards the opposite cell pole and produce faster oscillation cycles, as observed. This would account for the positive correlation between the amount of each GFP-MinDNg variant found in the cytoplasm and its rate of protein oscillation. In contrast, wild-type MinDNg would require sufficient MinENg recruitment and stimulation prior to commencing a cycle of oscillation in vivo. This delay would result in more stable MinDNg membrane association and longer oscillation cycles than those of the N-terminal mutants.

Interestingly, both GFP-MinDNg-I5E and the negative control GFP-MinDNg-K16Q were found entirely in the cytosol. However, since MinDNg-I5E had increased ATPase activity, in contrast to the low activity of MinDNg-K16Q, the data suggest that their common localization pattern was due to different factors. While a K16Q mutation in MinD does not affect ATP binding and has been proposed to render the protein unresponsive to this nucleotide (13), it is conceivable that the increased intrinsic ATPase activity of MinDNg-I5E prevents sufficient stable MinD-ATP to localize effectively along the membrane.

MinD oscillation and ATPase activity are intimately linked and require that MinD self-associate (12, 15, 19); hence, the faster oscillation cycles of most of our MinDNg mutants and the increased ATPase activities of MinDNg-I5E and MinDNg-3aaNT clearly show that there is still some level of functional self-association. Results from our in vitro ATPase assays also provide an explanation for the decreased self-affinity of our mutants, as detected by the yeast two-hybrid system. While ATP binding induces MinD dimerization, ATP hydrolysis should lead to the dissociation of MinD dimers (13, 15). Hence, while the molecules are still able to interact in the yeast reporter strain, the increased MinE-independent ATPase activity of the MinDNg-3aaNT and MinDNg-I5E mutants (and perhaps the other mutants) may have destabilized the self-association of these proteins, resulting in the diminished strengths of interaction observed. Furthermore, based on the proposed arrangement of a MinD dimer (19, 37a), generated from a three-dimensional alignment with its closest structural neighbor, dimeric NifH (2, 9, 32, 33), the first five residues of MinD are situated opposite the putative dimerization interface; therefore, these residues may not be directly involved in self-association.

It is unlikely that changes in the dynamic localization of our MinDNg mutants were due to deficiencies in ATP binding. First, many of the N-terminal MinDNg mutants displayed intracellular oscillation, which requires ATP (12). Second, two of the mutants, MinDNg-3aaNT and MinDNg-I5E, had increased ATPase activities, which would first require ATP binding (12). Third, archaeal MinD crystal structures show that the extreme N terminus of MinD does not contain residues that directly participate in nucleotide binding (9, 32). Finally, the MinC-MinD interaction appears to require ATP (15, 18), and we found that almost all of our MinDNg N-terminal mutants could still interact with MinC, as determined by yeast two-hybrid assays (data not shown).

What role might the N terminus of MinDNg play in the functionality of this protein? It is possible that the N terminus of MinDNg may be involved in controlling intrinsic ATPase activity. The crystal structure of MinD from Archaeoglobus fulgidus (2) revealed that residues at the extreme N terminus (Fig. 5) (aa 1 to 5) form one end of a buried ß-strand that connects directly to the P loop (Walker A ATP-binding motif) found at the opposite side of the protein. Hence, it is possible that some of the effects of ATP binding are transmitted from the ATP-binding face (Fig. 5), through the N-terminal residues, to other regions of the protein for ATP hydrolysis. Mutation of this N-terminal region may therefore render MinD more sensitive to the effects of ATP binding, perhaps in a more enzymatically active state, independent of MinE. This would be in contrast to the desensitization to ATP binding proposed for a K16Q mutation (13). Since the extreme N terminus of MinD is highly conserved, its function is likely shared by the N termini of MinD proteins from various organisms, particularly gram-negative bacteria. Other workers have also documented MinD mutants with increased basal ATPase activity, specifically mutants bearing mutations in the switch I region (aa 38 to 46 in MinDEc); however, it is still not clear how such mutations affect intracellular MinD movement, since dynamic localization studies were not carried out with these mutants (43). Furthermore, the switch I region mutants remained sensitive to MinE stimulation (43), in contrast to the mutant proteins in the present study. Hence, our MinD variants represent a new class of mutants.



View larger version (149K):
[in this window]
[in a new window]
  		FIG. 5. Structure of A. fulgidus MinD (PDB accession no. 1HYQ) (2), highlighting the N-terminal residues. Yellow indicates the extreme N-terminal amino acids at positions 1 to 5. The red residues (aa 6 to 9), in conjunction with the yellow residues (aa 1 to 5), form a ß-strand that connects to the P-loop (Walker A ATP-binding motif [green]). The general position of the ATP-binding face of MinD is indicated. The ribbon diagram was generated with the RasMol molecular graphics visualization tool (version 2.7.2.1).

There are several examples of ATPases in which the N terminus is involved in regulating enzymatic activity. Progressive removal of the first five residues of the chloroplast ATP synthase epsilon subunit resulted in significant increases in the ATPase activity of the holoenzyme (34). Deletion of the N-terminal A domain of XylR was also shown to fully activate the transcriptional regulator (7), which displays strongly stimulated ATPase activity in the presence of DNA activation sequences (24). In addition, the N-terminal domain of the DmpR transcriptional activator has been shown to act as a repressor of its ATPase activity (36).

Interestingly, our studies also showed that the ability to localize within polymeric arrays is not found exclusively in E. coli MinD, since gonococcal MinD was observed within such structures. We noted that each GFP-MinDNg mutant that displayed oscillation, no matter how faint, could still localize within such bands, suggesting that a basic array may be required to direct any degree of intracellular oscillation. The ability of MinDNg variants to localize within these bands corresponded with their ability to interact with MinENg, even if this interaction was severely diminished (e.g., MinDNg-I5A). Hence, these studies indicated that a minimum interaction with MinE must be maintained in order to form MinD-containing arrays that direct protein oscillation in vivo. Since MinDNg can oscillate distinctly from pole to pole in E. coli (27; this study), it is possible that the observed GFP-MinDNg-decorated polymers are actually coils, as observed with GFP-MinDEc (35); however, three-dimensional reconstruction of stacked images should be done to confirm this.

This study demonstrated that mutations or truncations within the first five residues of the conserved N terminus of MinDNg affect the movement and localization characteristics of the protein. Significantly, we demonstrated that alterations to the extreme N terminus of MinD produce phenotypes that differ fundamentally from the phenotype of the classic K16Q Walker A ATP-binding motif mutation. This study further highlighted the complexity of MinD function, showing that there are two regions that contribute to distinctly different phenotypes, despite their sequence proximity. For the first time, we found GFP-MinD variants that exhibit faster oscillation cycles than the wild-type control due to mutations within MinD itself and not due to mutated MinE or altered intracellular MinD/MinE ratios. Here we provide evidence that the altered dynamism of these MinDNg mutants stems from increased intrinsic basal ATPase activities that are independent of MinE. Further work, including in vitro membrane binding assays with vesicles that closely approximate bacterial cell membranes, as well as combined structural and enzymatic studies, should be useful in further elucidating the function of the MinD N terminus. We noted that residues 6 to 9 of MinD are even more conserved than the first five residues, and it would be interesting to study their role(s) as well. Since correct cell division site selection is such an important event in the bacterial cell cycle, there should conceivably be several mechanisms and/or factors that regulate dynamic Min protein movement, including determinants within MinD itself that ensure its proper ATPase activity.


arrow
ACKNOWLEDGMENTS
 
This project was funded by a Canadian Institutes of Health Research (CIHR) operating grant to J. R. Dillon (grant 200209MOP-106682-MI-CECA-22130). J. Szeto was funded by a CIHR Doctoral Research award, and N. F. Eng was funded by a Natural Sciences and Engineering Research Council Doctoral Canada Graduate Scholarship.

We thank Sheila Costford for constructing plasmids pSC9 and pSC10 and for initiating protein purification protocols.


arrow
FOOTNOTES
 
* Corresponding author. Present address: College of Arts and Science, c/o Arts Room 226, 9 Campus Drive, University of Saskatchewan, Saskatoon SK, S7N 5A5 Canada. Phone: (306) 966-4232. Fax: (306) 966-8839. E-mail: j.dillon{at}usask.ca. Back


arrow
REFERENCES
 
    1
  1. Bi, E., and J. Lutkenhaus. 1991. FtsZ ring structure associated with division in Escherichia coli. Nature 354:161-164.
    2. 2
  2. Cordell, S. C., and J. Löwe. 2001. Crystal structure of the bacterial cell division regulator MinD. FEBS Lett. 492:160-165.[CrossRef][Medline]
    3. 3
  3. de Boer, P., R. E. Crossley, and L. I. Rothfield. 1988. Isolation and properties of minB, a complex genetic locus involved in correct placement of the division site in Escherichia coli. J. Bacteriol. 170:2106-2112.
    4. 4
  4. de Boer, P., 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 Escherichia coli. Cell 56:641-649.
    5. 5
  5. de Boer, P., 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.[Medline]
    6. 6
  6. Errington, J., R. A. Daniel, and D. J. Scheffers. 2003. Cytokinesis in bacteria. Microbiol. Mol. Biol. Rev. 67:52-65.[Abstract/Free Full Text]
    7. 7
  7. Fernandez, S., V. de Lorenzo,and J. Perez-Martin. 1995. Activation of the transcriptional regulator XylR of Pseudomonas putida by release of repression between functional domains. Mol. Microbiol. 16:205-213.[CrossRef][Medline]
    8. 8
  8. Harder, K. W., P. Owen, L. K. Wong, R. Aebersold, I. Clark-Lewis, and F. R. Jirik. 1994. Characterization and kinetic analysis of the intracellular domain of human protein tyrosine phosphatase beta (HPTP beta) using synthetic phosphopeptides. Biochem. J. 298:395-401.
    9. 9
  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.[CrossRef][Medline]
    10. 10
  10. 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.[Abstract/Free Full Text]
    11. 11
  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.[CrossRef][Medline]
    12. 12
  12. 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.[CrossRef][Medline]
    13. 13
  13. Hu, Z., E. P. Gogol, and J. Lutkenhaus. 2002. Dynamic assembly of MinD on phospholipid vesicles regulated by ATP and MinE. Proc. Natl. Acad. Sci. USA 99:6761-6766.[Abstract/Free Full Text]
    14. 14
  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.[CrossRef][Medline]
    15. 15
  15. 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.[Abstract/Free Full Text]
    16. 16
  16. Huang, J., C. Cao, and J. Lutkenhaus. 1996. Interaction between FtsZ and inhibitors of cell division. J. Bacteriol. 178:5080-5085.[Abstract/Free Full Text]
    17. 17
  17. Kellogg, D. S., Jr., W. L. Peacock, Jr., W. E. Deacon, L. Brown, and D. I. Pirkle. 1963. Neisseria gonorrhoeae. I. Virulence genetically linked to clonal variation. J. Bacteriol. 85:1274-1279.[Abstract/Free Full Text]
    18. 18
  18. 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.[Abstract/Free Full Text]
    19. 19
  19. 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.[CrossRef][Medline]
    20. 20
  20. Ma, L.-Y, G. King, and L. Rothfield. 2003. Mapping the MinE site involved in interaction with the MinD division site selection protein of Escherichia coli. J. Bacteriol. 185:4948-4955.
    21. 21
  21. 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.
    22. 22
  22. Miroux, B., and J. E. Walker. 1996. Over-production of proteins in Escherichia coli: mutant hosts that allow synthesis of some membrane proteins and globular proteins at high levels. J. Mol. Biol. 260:289-298.[CrossRef][Medline]
    23. 23
  23. Pagotto, F. J., H. Salimnia, P. A. Totten, and J. R. Dillon. 2000. Stable shuttle vectors for Neisseria gonorrhoeae, Haemophilus spp. and other bacteria based on a single origin of replication. Gene 244:13-19.[CrossRef][Medline]
    24. 24
  24. Perez-Martin, J., and V. de Lorenzo. 1996. In vitro activities of an N-terminal truncated form of XylR, a sigma 54-dependent transcriptional activator of Pseudomonas putida. J. Mol. Biol. 258:575-587.
    25. 25
  25. Picard, F. J., and J. R. Dillon. 1989. Biochemical and genetic studies with arginine and proline auxotypes of Neisseria gonorrhoeae. Can. J. Microbiol. 35:1069-1075.
    26. 26
  26. Ramirez-Arcos, S., J. Szeto, T. J. Beveridge, C. Victor, F. Francis, and J. R. Dillon. 2001. Deletion of the cell division inhibitor MinC results in lysis of Neisseria gonorrhoeae. Microbiology 147:225-237.
    27. 27
  27. Ramirez-Arcos, S., J. Szeto, J. R. 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. 28
  28. Ramirez-Arcos, S., V. Greco, H. Douglas, D. Tessier, D. Fan, J. Szeto, J. Wang., and J. R. Dillon. 2004. Conserved glycines in the C terminus of MinC proteins are implicated in their functionality as cell division inhibitors. J. Bacteriol. 186:2841-2855.[Abstract/Free Full Text]
    29. 29
  29. Raskin, D. M., and P. A. J. de Boer. 1999. Rapid pole-to-pole oscillation of a protein required for directing division to the middle of Escherichia coli. Proc. Natl. Acad. Sci. USA 96:4971-4976.
    30. 30
  30. Raskin, D. M., and P. A. J. de Boer. 1999. MinDE-dependent pole-to-pole oscillation of division inhibitor MinC in Escherichia coli. J. Bacteriol. 181:6419-6424.
    31. 31
  31. Rowland, S. L., X. Fu, M. A. Sayed, Y. Zhang, W. R. Cook, and L. I. Rothfield. 2000. Membrane redistribution of the Escherichia coli MinD protein induced by MinE. J. Bacteriol. 182:613-619.[Abstract/Free Full Text]
    32. 32
  32. 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.
    33. 33
  33. Schindelin, H., C. Kisker, J. L. Schlessman, J. B. Howard, and D. C. Rees. 1997. Structure of ADP x AIF4(-)-stabilized nitrogenase complex and its implications for signal transduction. Nature 387:370-376.[CrossRef][Medline]
    34. 34
  34. Shi, X. B., J. M. Wei, and Y. K. Shen. 2001. Effects of sequential deletions of residues from the N- or C-terminus on the functions of epsilon subunit of the chloroplast ATP synthase. Biochemistry 40:10825-10831.[CrossRef][Medline]
    35. 35
  35. 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.[Abstract/Free Full Text]
    36. 36
  36. Shingler, V., and H. Pavel. 1995. Direct regulation of the ATPase activity of the transcriptional activator DmpR by aromatic compounds. Mol. Microbiol. 17:505-513.[CrossRef][Medline]
    37. 37
  37. Szeto, J., N. F. Eng, S. Acharya, M. D. Rigden, and J. R. Dillon. Res. Microbiol., in press.
    38. 37
  38. 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.[Abstract/Free Full Text]
    39. 38
  39. Szeto, J., S. Ramirez-Arcos, C. Raymond, L. D. Hicks, C. M. Kay, and J. R. Dillon. 2001. Gonococcal MinD affects cell division in Neisseria gonorrhoeae and Escherichia coli and exhibits a novel self-interaction. J. Bacteriol. 183:6253-6264.[Abstract/Free Full Text]
    40. 39
  40. 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.[Abstract/Free Full Text]
    41. 40
  41. 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.[Abstract/Free Full Text]
    42. 41
  42. Weiss, D. S., J. C. Chen, J. Ghigo, D. Boyd, and J. Beckwith. 1999. Localization of FtsI (PBP3) to the septal ring requires its membrane anchor, the Z ring, FtsA, FtsQ, and FtsL. J. Bacteriol. 181:508-520.[Abstract/Free Full Text]
    43. 42
  43. Woldringh, C. L., E. Mulder, J. A. Valkenburg, F. B. Wientjes, A. Zaritsky, and N. Nanninga. 1990. Role of the nucleoid in the toporegulation of division. Res. Microbiol. 141:39-49.[Medline]
    44. 43
  44. 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.[Abstract/Free Full Text]

Journal of Bacteriology, November 2004, p. 7175-7185, Vol. 186, No. 21
0021-9193/04/$08.00+0     DOI: 10.1128/JB.186.21.7175-7185.2004
Copyright © 2004, American Society for Microbiology. All Rights Reserved.



This article has been cited by other articles:

  • Rigden, M. D., Baier, C., Ramirez-Arcos, S., Liao, M., Wang, M., Dillon, J.-A. R. (2008). Identification of the Coiled-coil Domains of Enterococcus faecalis DivIVA that Mediate Oligomerization and their Importance for Biological Function. J Biochem 144: 63-76 [Abstract] [Full Text]  

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Szeto, J.
Right arrow Articles by Dillon, J.-A. R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Szeto, J.
Right arrow Articles by Dillon, J.-A. R.

Copyright © 2009 by the American Society for Microbiology.   For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»