The Plastid Division Protein AtMinD1 Is a Ca2+-ATPase Stimulated by AtMinE1
2005; Elsevier BV; Volume: 280; Issue: 36 Linguagem: Inglês
10.1074/jbc.m505126200
ISSN1083-351X
AutoresCassie Aldridge, Simon Geir Møller,
Tópico(s)Enzyme Structure and Function
ResumoBacteria and plastids divide symmetrically through binary fission by accurately placing the division site at midpoint, a process initiated by FtsZ polymerization, which forms a Z-ring. In Escherichia coli precise Z-ring placement at midcell depends on controlled oscillatory behavior of MinD and MinE: In the presence of ATP MinD interacts with the FtsZ inhibitor MinC and migrates to the membrane where the MinD-MinC complex recruits MinE, followed by MinD-mediated ATP hydrolysis and membrane release. Although correct Z-ring placement during Arabidopsis plastid division depends on the precise localization of the bacterial homologs AtMinD1 and AtMinE1, the underlying mechanism of this process remains unknown. Here we have shown that AtMinD1 is a Ca2+-dependent ATPase and through mutation analysis demonstrated the physiological importance of this activity where loss of ATP hydrolysis results in protein mislocalization within plastids. The observed mislocalization is not due to disrupted AtMinD1 dimerization, however; the active site AtMinD1(K72A) mutant is unable to interact with the topological specificity factor AtMinE1. We have shown that AtMinE1, but not E. coli MinE, stimulates AtMinD1-mediated ATP hydrolysis, but in contrast to prokaryotes stimulation occurs in the absence of membrane lipids. Although AtMinD1 appears highly evolutionarily conserved, we found that important biochemical and cell biological properties have diverged. We propose that correct intraplastidic AtMinD1 localization is dependent on AtMinE1-stimulated, Ca2+-dependent AtMinD1 ATP hydrolysis, ultimately ensuring precise Z-ring placement and symmetric plastid division. Bacteria and plastids divide symmetrically through binary fission by accurately placing the division site at midpoint, a process initiated by FtsZ polymerization, which forms a Z-ring. In Escherichia coli precise Z-ring placement at midcell depends on controlled oscillatory behavior of MinD and MinE: In the presence of ATP MinD interacts with the FtsZ inhibitor MinC and migrates to the membrane where the MinD-MinC complex recruits MinE, followed by MinD-mediated ATP hydrolysis and membrane release. Although correct Z-ring placement during Arabidopsis plastid division depends on the precise localization of the bacterial homologs AtMinD1 and AtMinE1, the underlying mechanism of this process remains unknown. Here we have shown that AtMinD1 is a Ca2+-dependent ATPase and through mutation analysis demonstrated the physiological importance of this activity where loss of ATP hydrolysis results in protein mislocalization within plastids. The observed mislocalization is not due to disrupted AtMinD1 dimerization, however; the active site AtMinD1(K72A) mutant is unable to interact with the topological specificity factor AtMinE1. We have shown that AtMinE1, but not E. coli MinE, stimulates AtMinD1-mediated ATP hydrolysis, but in contrast to prokaryotes stimulation occurs in the absence of membrane lipids. Although AtMinD1 appears highly evolutionarily conserved, we found that important biochemical and cell biological properties have diverged. We propose that correct intraplastidic AtMinD1 localization is dependent on AtMinE1-stimulated, Ca2+-dependent AtMinD1 ATP hydrolysis, ultimately ensuring precise Z-ring placement and symmetric plastid division. Plastids are essential plant organelles vital for life on earth. They are not formed de novo but arise by binary fission from pre-existing plastids (1Possingham J.V. Saurer W. Planta. 1969; 86: 186-194Crossref PubMed Scopus (70) Google Scholar, 2Boasson R. Laetsch W.H. Price I. Am. J. Bot. 1972; 59: 217-233Crossref Google Scholar, 3Platt-Aloia K. Thomson W.W. New Phytol. 1977; 78: 599-605Crossref Scopus (35) Google Scholar); plastid division is therefore essential for the maintenance and accumulation of plastid populations within plant cells. Many proteins involved in plastid division are derived from bacterial components conserved from the cyanobacterial origins of higher plant chloroplasts (4Osteryoung K.W. Vierling E. Nature. 1995; 376: 473-474Crossref PubMed Scopus (246) Google Scholar, 5Colletti K.S. Tattersall E.A. Pyke K.A. Froelich J.E. Stokes K.D. Osteryoung K.W. Curr. Biol. 2000; 10: 507-516Abstract Full Text Full Text PDF PubMed Scopus (170) Google Scholar, 6Itoh R. Fujiwara M. Nagata N. Yoshida S. Plant Physiol. 2001; 27: 1644-1655Crossref Scopus (115) Google Scholar, 7Maple J. Chua N.H. Møller S.G. Plant J. 2002; 31: 269-277Crossref PubMed Scopus (80) Google Scholar), including FtsZ, an ancient tubulin-like protein that forms a Z-ring to which other components of the division machinery are recruited (8Bi E. Lutkenhaus J. Nature. 1991; 354: 161-164Crossref PubMed Scopus (1158) Google Scholar, 9Lutkenhaus J. Addinall S.G. Ann. Rev. Biochem. 1997; 66: 93-116Crossref PubMed Scopus (407) Google Scholar). The Z-ring is localized to the plastid midpoint (10Vitha S. McAndrew R.S. Osteryoung K.W. J. Cell Biol. 2001; 153: 111-119Crossref PubMed Scopus (235) Google Scholar, 11Fujiwara M. Yoshida S. Biochem. Biophys. Res. Commun. 2001; 287: 462-467Crossref PubMed Scopus (41) Google Scholar, 12McAndrew R.S. Froehlich J.E. Vitha S. Stokes K.D. Osteryoung K.W. Plant Physiol. 2001; 127: 1656-1666Crossref PubMed Scopus (133) Google Scholar), and correct Z-ring placement is mediated by the coordinated action of the prokaryotic-derived Min proteins. The Escherichia coli minB operon encodes MinC, MinD, and MinE, which together limit Z-ring placement to midcell (13de Boer P.A.J. Crossley R.E. Rothfield L.I. Cell. 1989; 56: 641-649Abstract Full Text PDF PubMed Scopus (598) Google Scholar, 14Bi E. Lutkenhaus J. J. Bacteriol. 1993; 175: 1118-1125Crossref PubMed Scopus (333) Google Scholar, 15Pichoff S. Lutkenhaus J. J. Bacteriol. 2001; 183: 6630-6635Crossref PubMed Scopus (93) Google Scholar): MinC is an antagonist of FtsZ polymerization (14Bi E. Lutkenhaus J. J. Bacteriol. 1993; 175: 1118-1125Crossref PubMed Scopus (333) Google Scholar, 15Pichoff S. Lutkenhaus J. J. Bacteriol. 2001; 183: 6630-6635Crossref PubMed Scopus (93) Google Scholar), and topological distribution of MinC is controlled by the ATPase MinD and the topological specificity factor MinE (16Hu Z. Lutkenhaus J. Mol. Microbiol. 1999; 34: 82-90Crossref PubMed Scopus (366) Google Scholar, 17Fu X. Shih Y.L. Zhang Y. Rothfield L.I. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 980-985Crossref PubMed Scopus (185) Google Scholar, 18Raskin D.M. de Boer P.A.J. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 4971-4976Crossref PubMed Scopus (589) Google Scholar). ATP-bound MinD recruits MinC to the membrane where the MinD-MinC complex forms a stable inhibition structure at the polar zone of the cell (19Szeto T.H. Rowland S.L. Rothfield L.I. King G.F. Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 15693-15698Crossref PubMed Scopus (191) Google Scholar, 20Hu Z. Lutkenhaus J. Mol. Microbiol. 2003; 47: 345-355Crossref PubMed Scopus (146) Google Scholar, 21Zhou H. Lutkenhaus J. J. Bacteriol. 2004; 186: 1546-1555Crossref PubMed Scopus (28) Google Scholar). Topological specificity is conferred on this complex through interaction of MinE with membrane-bound MinD whereby MinE stimulates MinD ATPase activity, causing MinD to disassociate from the membrane and oscillation to the opposite side of the cell (22Hu Z. Lutkenhaus J. Mol. Cell. 2001; 7: 1337-1343Abstract Full Text Full Text PDF PubMed Scopus (213) Google Scholar, 23Hale C.A. Meinhardt H. de Boer P.A.J. EMBO J. 2001; 20: 1563-1572Crossref PubMed Scopus (229) Google Scholar, 24Hu Z. Gogol E.P. Lutkenhaus J. Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 6761-6766Crossref PubMed Scopus (239) Google Scholar). The MinD dynamic behavior ensures the lowest MinD-MinC inhibitor complex concentration at midcell, resulting in FtsZ polymerization and appropriate placement of cell division (25Meinhardt H. de Boer P.A.J. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 14202-14207Crossref PubMed Scopus (295) Google Scholar). Higher plants contain MinD and MinE homologs (5Colletti K.S. Tattersall E.A. Pyke K.A. Froelich J.E. Stokes K.D. Osteryoung K.W. Curr. Biol. 2000; 10: 507-516Abstract Full Text Full Text PDF PubMed Scopus (170) Google Scholar, 6Itoh R. Fujiwara M. Nagata N. Yoshida S. Plant Physiol. 2001; 27: 1644-1655Crossref Scopus (115) Google Scholar, 7Maple J. Chua N.H. Møller S.G. Plant J. 2002; 31: 269-277Crossref PubMed Scopus (80) Google Scholar), and transgenic Arabidopsis plants with reduced AtMinD1 levels show Z-ring misplacement and asymmetric plastid division, whereas AtMinD1 overexpression leads to plastid division inhibition (5Colletti K.S. Tattersall E.A. Pyke K.A. Froelich J.E. Stokes K.D. Osteryoung K.W. Curr. Biol. 2000; 10: 507-516Abstract Full Text Full Text PDF PubMed Scopus (170) Google Scholar, 26Dinkins R. Reddy M.S. Leng M. Collins G.B. Planta. 2001; 214: 180-188Crossref PubMed Scopus (42) Google Scholar). These phenotypes are reminiscent of minicelling and filamenting E. coli deficient in or overexpressing MinD (13de Boer P.A.J. Crossley R.E. Rothfield L.I. Cell. 1989; 56: 641-649Abstract Full Text PDF PubMed Scopus (598) Google Scholar), suggesting functional conservation between E. coli MinD and AtMinD1. In agreement with this, the polar localization of E. coli MinD reflects the distinct intraplastidic localization pattern of AtMinD1, which localizes to a single spot or two spots at opposite poles of chloroplasts (7Maple J. Chua N.H. Møller S.G. Plant J. 2002; 31: 269-277Crossref PubMed Scopus (80) Google Scholar, 27Fujiwara M. Nakamura A. Itoh R. Shimada Y. Yoshida S. Møller S.G. J. Cell Sci. 2004; 117: 2399-2410Crossref PubMed Scopus (71) Google Scholar). AtMinD1 encodes a protein of 326 amino acids and, based on amino acid similarity, belongs to the ParA ATPase protein family containing a Walker A motif involved in the binding and hydrolysis of ATP. Like many ParA proteins, AtMinD1 can dimerize (27Fujiwara M. Nakamura A. Itoh R. Shimada Y. Yoshida S. Møller S.G. J. Cell Sci. 2004; 117: 2399-2410Crossref PubMed Scopus (71) Google Scholar), and our studies on the accumulation and replication of chloroplasts 11 (arc11) mutant have demonstrated that the asymmetric plastid division observed in arc11 is due to a A296G missense mutation in AtMinD1 that leads to loss of dimerization and inappropriate intraplastidic localization (27Fujiwara M. Nakamura A. Itoh R. Shimada Y. Yoshida S. Møller S.G. J. Cell Sci. 2004; 117: 2399-2410Crossref PubMed Scopus (71) Google Scholar). This suggests that AtMinD1 dimerization and correct intraplastidic localization is in part important for correct Z-ring placement during Arabidopsis plastid division. Here we have expanded on our studies and demonstrated that AtMinD1 in Arabidopsis is an ATPase but, in contrast to its bacterial counterpart, is activated by Ca2+ rather than Mg2+. We further showed that the AtMinD1-mediated ATP hydrolysis is stimulated by AtMinE1 in the absence of envelope membrane lipids. Mutation analysis reveals that an active site AtMinD1(K72A) mutant leads to loss of ATPase activity and mislocalization within plastids and that, although AtMinD1(K72A) is able to dimerize, the interaction with AtMinE1 is abolished. Our findings reveal that appropriate AtMinD1 intraplastidic localization not only depends on AtMinD1 dimerization but on AtMinE1-stimulated ATP hydrolysis, which in turn governs correct Z-ring placement and symmetric plastid division. Site-directed Mutagenesis of AtMinD1—A full-length 981-bp AtMinD1 cDNA was PCR amplified using primers MIND/5 (5′-ATCATATGGCGGTCTGAGATTGTTC-3′; NdeI is underlined) and MIND/7 (5′-ATGGATCCTTAGCCGCCAAAGAAAGAGAAGAAGCC-3′; BamH1 is underlined) and cloned into pCRScript (Stratagene) to generate pCRScript-AtMinD1. Two oligonucleotide primers, MIND/20 (5′-TTGCGGTGGTTGTCGTCGCTCCAACACCGCCTTTTC-3′) and MIND/21 (5′-CGGAAAAGGCGGTGTTGGAGCGACGACAACCACCGC-3′) were designed spanning the AtMinD1 Walker A motif containing single point mutations (underlined) changing the active site lysine (K) at position 72 to alanine (A). PCR amplification, using pCRScript-AtMinD1 as a template and the primer pairs MIND/5-MIND/20 and MIND/21-MIND/7 generated two AtMinD1 overlapping fragments that were joined together by Splicing by Overlap Extension using the flanking primers MIND/5 and MIND/7 to generate AtMinD1(K72A). The 981-bp full-length AtMinD1(K72A) cDNA was ligated into pCRScript (Stratagene) to generate pCRScript-AtMinD1(K72A) and subjected to DNA sequencing to verify the incorporated K72A mutation. Protein Expression and Purification—Full-length wild-type AtMinD1 and AtMinD1(K72A) cDNAs were subcloned from pCRScript-AtMinD1 and pCRScript-AtMinD1(K72A) into pET14b (Novagen) to generate pET14b-AtMinD1 and pET14b-AtMinD1(K72A) and transformed into E. coli strain BL21 (DE3). 50-ml cultures were grown at 37 °C to a density of A600 = 0.6. Protein expression was induced with 1.5 mm isopropyl β-d-thiogalactopyranoside for 2 h at 37 °C. Both AtMinD1 and AtMinD1(K72A) were insoluble and purified using TALON metal affinity resin (BD Biosciences) under denaturing conditions following the user manual. The purity of the proteins was verified by SDS-PAGE and refolded by dialysis against sodium phosphate buffers (50 mm sodium phosphate, 50 mm NaCl, 0.1 m EDTA, 1.5 mm dithiothreitol, 10% glycerol, pH 7.2) containing 8-0 m urea. AtMinE1 was expressed as a translational fusion to the C terminus of glutathione S-transferase (GST) 1The abbreviations used are: GST, glutathione S-transferase; S.D., synthetic dropout medium; AD, activation domain; BD, binding domain; YFP, yellow fluorescence protein. from pGEX-AtMinE1 (7Maple J. Chua N.H. Møller S.G. Plant J. 2002; 31: 269-277Crossref PubMed Scopus (80) Google Scholar) in E. coli BL21 (DE3). Protein expression was performed as described above and the soluble AtMinE1-GST fraction purified using glutathione resin (BD Biosciences) following the user manual. The purity was verified by SDS-PAGE. As a control, GST was purified from pGEX-6P (Amersham Biosciences) in E. coli BL21 (DE3). The 267-bp full-length E. coli minE gene was PCR amplified using primers EcE/3 (5′-ATCATATGGCATTACTCGATTTCTT-3′; NdeI is underlined) and EcE/4 (5′-ATGGATCCTTATTTCAGCTCTTCTGCTTCC-3′; BamH1 is underlined), ligated into pET14b to generate pET14b-EcMinE, and transformed into BL21 (DE3). Protein expression was performed as described above. EcMinE was soluble and was purified under native conditions using TALON metal affinity resin (BD Biosciences) following the user manual and the purity verified by SDS-PAGE. ATPase Assays—For all assays, the reaction mixture contained 100 mm Tris-Cl (pH 7.4), 50 mm NaCl, 0.1 mm EDTA, 1.5 mm dithiothreitol, 10% glycerol, and 5 mm CaCl2, except for the cation effects assays where 5 mm CaCl2 was replaced with either 5 mm MgCl2, KCl, or MnCl2. In experiments testing the effects of the AtMinD1(K72A) mutation, different cation effects, and pH dependence, 10 μm [γ-32P]ATP (specific activity 10 mCi/mmol) and 0.1 μm AtMinD1 or AtMinD1(K72A) were used, and reactions (20 μl) were incubated for 1 h at 35 °C and stopped with 1 μl of 1 m formic acid. In the time course assays for the double reciprocal plot 10-80 μm [γ-32P]ATP and 0.1 μm AtMinD1 were used, and reactions were incubated at 35 °C and stopped at the specified time. To analyze the effect of AtMinE1 and EcMinE on AtMinD1 ATP hydrolysis, 0.1 μm AtMinE1 (1 pmol), 0.1 μm EcMinE (1 pmol), and 0.1 μm AtMinD1 (1 pmol) were used. Reactions were incubated at 35 °C and stopped after 10 min. In all assays a no-enzyme control was used to assess the background. Samples were spotted onto PEI-cellulose (POLYGRAM CEL 300 PEI; Macherey-Nagel) TLC plates and developed using 0.5 m LiCl and 0.5 m formic acid. Radioactive nucleotides were visualized by autoradiography using x-ray film (Kodak). For quantification purposes plates were scanned using phosphorimaging (Cyclone Storage Phosphor System; Packard). AtMinD1/AtMinD1(K72A) Localization Analysis—Full-length AtMinD1 and AtMinD1(K72A) cDNAs were PCR amplified using primers MIND/1 (5′-TACTCGAGATGGCGTCTCTGAGATTGTTC-3′; XhoI underlined) and MIND/6 (5′-ATGGTACCGCCGCCAAAGAAAGAGAAGAAGCC-3′; KpnI underlined), removing the termination codon, and cloned into pWEN18 as N-terminal fusions to the YFP to generate pWEN18/AtMinD1 and pWEN18/AtMinD1(K72A). pWEN18/AtMinD1 and pWEN18/AtMinD1(K72A) were transiently expressed in tobacco leaves by particle bombardment (28Kost B. Spielhofer P. Chua N.H. Plant J. 1998; 16: 393-401Crossref PubMed Google Scholar) and visualized by fluorescence microscopy using a Nikon TE2000U inverted microscope, and image analysis was performed by using OPENLAB software (Improvision, Coventry, UK). The number of fluorescent spots in each chloroplast was recorded for each bombardment. Yeast Two-hybrid Analysis—Full-length AtMinD1(K72A) was PCR amplified using MIND/5 and MIND/7 and cloned into pGADT7 (GAL4 activation domain; AD) and pGBKT7 (GAL4 DNA-binding domain; BD) (Matchmaker two-hybrid system, version 3; Clontech). The resulting constructs, pGADT7/AtMinD1(K72A) and pGBKT7/AtMinD1(K72A) together with pGADT7/AtMinD1 and pGBKT7/AtMinD1 (27Fujiwara M. Nakamura A. Itoh R. Shimada Y. Yoshida S. Møller S.G. J. Cell Sci. 2004; 117: 2399-2410Crossref PubMed Scopus (71) Google Scholar), pGADT7/ARC11 and pGBKT7/ARC11 (27Fujiwara M. Nakamura A. Itoh R. Shimada Y. Yoshida S. Møller S.G. J. Cell Sci. 2004; 117: 2399-2410Crossref PubMed Scopus (71) Google Scholar), pGADT7/AtMinE1 and pGBKT7/AtMinE1 (29Maple J. Aldridge C. Møller S.G. Plant J. 2005; (in press)PubMed Google Scholar), and empty vector controls (pGADT7 and pGBKT7), were transformed into HF7c yeast cells in different combinations (Fig. 5). Double transformants were selected on minimal synthetic drop-out medium (S.D. medium) lacking tryptophan (T) (pGBKT7 vectors) and leucine (L) (pGADT7 vectors). Restoration of histidine (H) auxotrophy was used as a marker for protein-protein interactions. Quantitative data were obtained by growing 3-day-old double transformants in liquid S.D.-LT medium in a shaking incubator at 30 °C for 24 h. Cultures were grown to A600 of 1.0 before spotting 5 μl on to S.D.-LT and S.D.-LTH media plates. The plates were incubated for 4 days at 30 °C, and growth was assessed visually. For quantification each yeast spot was suspended in 1 ml of liquid S.D. medium and the A600 of the suspension recorded. Bimolecular Fluorescence Complementation—Full-length AtMinD1(K72A) cDNA was PCR amplified using primers MIND/1 and MIND/6 and cloned into pWEN-NY (29Maple J. Aldridge C. Møller S.G. Plant J. 2005; (in press)PubMed Google Scholar) as a fusion to the N-terminal fragment of YFP (containing amino acids 1-154 of YFP). The resulting construct pWEN-NY/AtMinD1(K72A) was co-bombarded into tobacco along with pWEN-CY/AtMinD1 (29Maple J. Aldridge C. Møller S.G. Plant J. 2005; (in press)PubMed Google Scholar) (containing amino acids 155-238 of YFP), and bimolecular fluorescence complementation was visualized by fluorescence microscopy. As a positive control tobacco was also co-bombarded with pWEN-NY/AtMinD1 and pWEN-CY/AtMinD1 (29Maple J. Aldridge C. Møller S.G. Plant J. 2005; (in press)PubMed Google Scholar). As a negative control pWEN-CY/AtMinD1, pWEN-NY/AtMinD1, and pWEN-NY/AtMinD1(K72A) were bombarded separately into tobacco cells. AtMinD1 Is a Ca2+-dependent ATPase—AtMinD1 contains a putative Walker A motif, and to test whether AtMinD1 is an ATPase a His6-AtMinD1 fusion protein was expressed in E. coli followed by denaturing Co2+ affinity chromatography purification and refolding by dialysis. The purity of refolded AtMinD1 was verified by SDS-PAGE (Fig. 1A) followed by incubation with radiolabeled [γ-32P]ATP at pH 7.4 in the presence of 5 mm CaCl2. TLC analysis and autoradiography showed clear AtMinD1-induced radiolabeled inorganic phosphate (Pi) release compared with a no-enzyme control, revealing that AtMinD1 is an ATPase (Fig. 1B). To ensure the measured ATP hydrolysis was due to AtMinD1 and not a contaminating E. coli ATPase, we generated an active site AtMinD1 mutant by substituting the conserved Walker A lysine for alanine, creating AtMinD1(K72A). AtMinD1(K72A) was expressed and purified as a His6 fusion protein as for wild-type AtMinD1 (Fig. 1A), incubated with [γ-32P]ATP in the presence of 5 mm CaCl2, and analyzed by TLC. Autoradiography revealed no significant Pi release above the no-enzyme control reaction (Fig. 1B), confirming that the K72A mutation inactivates AtMinD1 ATPase activity and that AtMinD1 ATP hydrolysis is mediated through the Walker A domain. To more fully characterize the catalytic activity of AtMinD1, we determined key kinetic parameters of AtMinD1-mediated ATP hydrolysis. The extent of ATP hydrolysis as a function of protein concentration was calculated by measuring Pi release in response to increasing AtMinD1 amounts (Fig. 2A). As seen from Fig. 2A, the extent of ATP hydrolysis was proportional to the amount of input AtMinD1, and in the linear range of enzyme dependence 7 pmol of ATP was hydrolyzed per pmol AtMinD1, which translates to a turnover number of ∼2 fmol s-1. Using input [γ-32P]ATP concentrations from 10-80 μm, we quantified Pi release as a function of time in separate time course assays (supplemental Fig. 1). From initial reaction rates a reciprocal plot revealed that AtMinD1 has a Km of 500 μm ATP and a Vmax of 0.2 pmol ATP s-1 (Fig. 2B). These values show that AtMinD1 is a weak ATPase. Divalent cations are known to influence the activity of ATPases (30Berger G. Girault G. J. Bioenerg. Biomembr. 2001; 33: 93-98Crossref PubMed Scopus (8) Google Scholar), and we investigated whether different cations affected AtMinD1 activity. Surprisingly, Mg2+ had no significant effect on AtMinD1 activity (Fig. 2C) in contrast to the Mg2+-dependent activity of E. coli MinD (31de Boer P.A. Crossley R.E. Hand A.R. Rothfield L.I. EMBO J. 1991; 10: 4371-4380Crossref PubMed Scopus (260) Google Scholar). Similarly, K+ and Mn2+ did not significantly stimulate AtMinD1-mediated ATP hydrolysis (Fig. 2C). However, addition of Ca2+ ions had a dramatic effect leading to an ∼5-fold increase in ATP hydrolysis compared with reactions containing Mg2+ (Fig. 2C). Further analysis showed maximum ATP hydrolysis between pH 7.5-8 (Fig. 2D). Combined, these data demonstrate that AtMinD1 is a Ca2+-dependent ATPase. AtMinE1 Stimulates AtMinD1 ATPase Activity—Because of the low basal ATPase activity of AtMinD1, we analyzed whether AtMinE1 can stimulate AtMinD1-mediated ATP hydrolysis. Equimolar amounts of purified AtMinE1-GST and AtMinD1 were incubated with CaCl2 and [γ-32P]ATP for 10 min. The inclusion of AtMinE1 had a marked effect on the amount of Pi release, showing an ∼3-fold increase in ATP hydrolysis compared with AtMinD1 alone and taking into account background Pi release in assays only containing AtMinE1 (Fig. 3), demonstrating that AtMinE1 can stimulate AtMinD1 activity. To ensure the increase in ATP hydrolysis was not due to either inherent AtMinE1 ATPase activity or a contaminating E. coli ATPase, we performed ATPase assays using only purified AtMinE1-GST. Although AtMinE1-GST assays did result in low background ATP hydrolysis (Fig. 3), probably due to contaminating E. coli protein(s), it is clear that the measured increase in ATP hydrolysis in reactions containing both AtMinD1 and AtMinE1 is because of AtMinD1 activity. To further verify this, we performed assays with AtMinD1(K72A) and AtMinE1 that resulted in similar ATP hydrolysis levels as observed for AtMinE1-GST alone (data not shown). As a control for AtMinE1-GST, we performed assays using purified GST that resulted in no Pi release (Fig. 3). To investigate the evolutionary conservation of AtMinE1-stimulated AtMinD1 activity, E. coli MinE (EcMinE) was purified and substituted for AtMinE1 in the ATPase assays. Although we have previously demonstrated that AtMinE1 can function as a topological specificity factor in E. coli (7Maple J. Chua N.H. Møller S.G. Plant J. 2002; 31: 269-277Crossref PubMed Scopus (80) Google Scholar), experiments revealed EcMinE was unable to stimulate AtMinD1 activity (Fig. 3). Loss of ATP Hydrolysis Results in Abnormal AtMinD1 Localization—AtMinD1 shows distinct intraplastidic localization patterns localizing into one or two discrete spots at polar regions of ellipsoidal chloroplasts (7Maple J. Chua N.H. Møller S.G. Plant J. 2002; 31: 269-277Crossref PubMed Scopus (80) Google Scholar, 27Fujiwara M. Nakamura A. Itoh R. Shimada Y. Yoshida S. Møller S.G. J. Cell Sci. 2004; 117: 2399-2410Crossref PubMed Scopus (71) Google Scholar). In contrast, the A296G mutation in AtMinD1/ARC11 results in mislocalization, forming large and distorted fluorescent aggregates and/or multiple speckles (27Fujiwara M. Nakamura A. Itoh R. Shimada Y. Yoshida S. Møller S.G. J. Cell Sci. 2004; 117: 2399-2410Crossref PubMed Scopus (71) Google Scholar). To test whether an active site mutation affects AtMinD1 localization, translational fusions of AtMinD1 and AtMinD1(K72A) to YFP were created and transiently expressed in tobacco leaves by particle bombardment. As expected, the majority of AtMinD1-YFP (75%) localized into one or two discrete spots (Fig. 4A). In contrast, AtMinD1(K72A)-YFP forms multiple speckles often up to six or more speckles within chloroplasts (Fig. 4). Unlike the mislocalization of AtMinD1(A296G), protein aggregation is rarely observed and AtMinD1(K72A)-YFP frequently forms more speckles per chloroplast than reported for AtMinD1(A296G)-GFP (27Fujiwara M. Nakamura A. Itoh R. Shimada Y. Yoshida S. Møller S.G. J. Cell Sci. 2004; 117: 2399-2410Crossref PubMed Scopus (71) Google Scholar), suggesting a different mechanism is responsible for the mislocalization. It is clear from these results that ATP binding/hydrolysis plays an important role in ensuring correct intraplastidic localization of AtMinD1. AtMinD1(K72A) Can Dimerize—Yeast two-hybrid and fluorescence resonance energy transfer analyses have demonstrated AtMinD1 is able to form homodimers (27Fujiwara M. Nakamura A. Itoh R. Shimada Y. Yoshida S. Møller S.G. J. Cell Sci. 2004; 117: 2399-2410Crossref PubMed Scopus (71) Google Scholar). In line with this, E. coli MinD undergoes self assembly on phospholipid vesicles forming filamentous polymeric structures (24Hu Z. Gogol E.P. Lutkenhaus J. Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 6761-6766Crossref PubMed Scopus (239) Google Scholar). Both ARC11/AtMinD1(A296G) and AtMinD1(K72A) exhibit mislocalization within plastids, and we have previously shown that ARC11/AtMinD1(A296G) mislocalization is due to loss of dimerization (27Fujiwara M. Nakamura A. Itoh R. Shimada Y. Yoshida S. Møller S.G. J. Cell Sci. 2004; 117: 2399-2410Crossref PubMed Scopus (71) Google Scholar). To investigate whether AtMinD1(K72A) mislocalization was due to loss of dimerization, we performed yeast two-hybrid studies using restoration of histidine auxotrophy as a marker for interaction. In agreement with our previous data (27Fujiwara M. Nakamura A. Itoh R. Shimada Y. Yoshida S. Møller S.G. J. Cell Sci. 2004; 117: 2399-2410Crossref PubMed Scopus (71) Google Scholar) we found His auxotrophy restoration in cells expressing AD-AtMinD1 and BD-AtMinD1 (Fig. 5A). Similarly, HF7c expressing AD-AtMinD1(K72A) and BD-AtMinD1 (Fig. 5A) or AD-AtMinD1 and BD-AtMinD1(K72A) (data not shown) showed restoration of His auxotrophy, revealing the K72A mutation does not affect AtMinD1 dimerization. As a negative control we expressed AD-ARC11/AtMinD1(A296G) and BD-AtMinD1 (Fig. 5A) and AD-AtMinD1 and BD-ARC11/AtMinD1(A296G) (data not shown), which showed no growth without His. BD-ARC11/AtMinD1(A296G) and AD-AtMinD1(K72A) (Fig. 5A) and AD-ARC11/AtMinD1(A296G) and BD-AtMinD1(K72A) (data not shown) were also co-expressed, showing no growth without His. To ensure the interactions detected were not because of autoactivation, each construct was co-expressed with the empty vector controls and showed no restoration of His auxotrophy. To verify that AtMinD1(K72A) can dimerize inside living chloroplasts, bimolecular fluorescence complementation assays were carried out. Separate, non-fluorescent N-terminal (NY) and C-terminal (CY) YFP protein domains can associate to form a functional fluorescent bimolecular complex when brought into proximity by interacting proteins. As previously observed (29Maple J. Aldridge C. Møller S.G. Plant J. 2005; (in press)PubMed Google Scholar), tobacco co-bombarded with pWEN-NY/AtMinD1 and pWEN-CY/AtMinD1 showed clear fluorescence (Fig. 5B). Tobacco co-bombarded with pWEN-NY/AtMinD1(K72A) and pWEN-CY/AtMinD1 also showed clear fluorescence (Fig. 5B), demonstrating that AtMinD1(K72A) is able to dimerize in planta and showing that AtMinD1(K72A) mislocalization is not because of loss of dimerization as observed for ARC11/AtMinD1 (A296G). Tobacco cells bombarded with single vectors (negative controls) showed no fluorescence, as expected. AtMinD1(K72A) Is Unable to Interact with AtMinE1—The MinE binding site in E. coli MinD is in close proximity to the ATP binding site on α-helix 7 (32Ma L. King G.F. Rothfield L. Mol. Microbiol. 2004; 54: 99-108Crossref PubMed Scopus (49) Google Scholar) and lysine 11 within the Walker A motif (equivalent to AtMinD1 lysine 67) interacts with residues within α-helix 7 competing with MinE (32Ma L. King G.F. Rothfield L. Mol. Microbiol. 2004; 54: 99-108Crossref PubMed Scopus (49) Google Scholar), suggesting the Walker A motif is involved in mediating MinD-MinE interaction. To test whether the AtMinD1 K72A mutation affects interaction with AtMinE1 we expressed AD-AtMinE1 and BD-AtMinE1 with BD-AtMinD1(K72A) and AD-AtMinD1(K72A), respectively, in HF7c. In contrast to yeast cells expressing AD-AtMinE1 and BD-AtMinD1 (Fig. 5A) or AD-AtMinD1 and BD-AtMinE1 (data not shown) showing growth on His-free media, cells containing AD-AtMinE1 and BD-AtMinD1(K72A) or AD-AtMinD1(K72A) and BD-AtMinE1 (Fig. 5A) showed no restoration of His auxotrophy, demonstrating that the K72A mutation in AtMinD1 abolishes the interaction with AtMinE1. AD-AtMinE1 and BD-ARC11/AtMinD1(A296G) were also co-expressed in HF7c and demonstrated the AtMinD1 A296G mutation has no effect on AtMinD1-AtMinE1 interaction (Fig. 5A). To ensure the interactions detected were not due to AtMinE1 autoactivation, empty BD and AD vector were expressed with AD-AtMinE1 and BD-AtMinE1, revealing no His auxotrophy restoration (Fig. 5A). The importance of AtMinD1 in plastid division has been shown by studies demonstrating that a disequilibrium of AtMinD1 levels in transgenic plants results in Z-ring misplacement and inappropriate plastid division (5Colletti K.S. Tattersall E.A. Pyke K.A. Froelich J.E. Stokes K.D. Osteryoung K.W. Curr. Biol. 2000; 10: 507-516Abstract Full Text Full Text PDF PubMed Scopus (170) Google Scholar, 26Dinkins R. Reddy M.S. Leng M. Collins G.B. Planta. 2001; 214: 180-188Crossref PubMed Scopus (42) Google Scholar). Although little is known about how AtMinD1 ensures correct Z-ring placement, we recently reported, through the cloning of the disrupted locus in arc11, that AtMinD1 dimerization is important for correct interplastidic localization and central Z-ring positioning during plastid division (27Fujiwara M. Nakamura A. Itoh R. Shimada Y. Yoshida S. Møller S.G. J. Cell Sci. 2004; 117: 2399-2410Crossref PubMed Scopus (71) Google Scholar). In this report we have revealed the importance of the biochemical activity of AtMinD1 and shown how the topological specificity factor AtMinE1 modulates both AtMinD1 activity and AtMinD1 localization dynamics. We showed that AtMinD1, in contrast to its bacterial counterpart, is a Ca2+-dependent ATPase and that the Walker A motif is important for both ATPase activity and correct intraplastidic localization. Although an active site AtMinD1 mutant can still dimerize, loss of ATPase activity abolishes its interaction with AtMinE1. Together with the fact that AtMinE1 can stimulate the ATPase activity of AtMinD1, our data suggest that AtMinE1 not only modulates ATP hydrolysis but also ensures correct AtMinD1 localization within plastids during division. AtMinD1 ATP hydrolysis is mediated through the Walker A domain, as a K72A mutation within this motif leads to a complete loss of ATPase activity (Fig. 1B). The high Km and low Vmax values show that AtMinD1 is a weak ATPase, like the MinD-related protein ParA (33Bouet J.Y. Funnell B.E. EMBO J. 1999; 18: 1415-1424Crossref PubMed Scopus (138) Google Scholar). The weak activity exhibited by AtMinD1 may be because of electrostatic properties of the ATP binding site. In the F1 ATPase (34Abrahams J.P. Leslie A.G. Lutter R. Walker J.E. Nature. 1994; 370: 621-628Crossref PubMed Scopus (2764) Google Scholar) and in hydrolases (35Coleman D.E. Berghuis A.M. Lee E. Linder M.E. Gilman A.G. Sprang S.R. Science. 1994; 265: 1405-1412Crossref PubMed Scopus (757) Google Scholar), basic amino acids near the ATP γ-phosphate are responsible for stabilization of the transition state negative charge (36Hayashi I. Oyama T. Morikawa K. EMBO J. 2001; 20: 1819-1828Crossref PubMed Scopus (108) Google Scholar); both E. coli MinD and AtMinD1 lack these basic amino acids, possibly explaining the weak ATP turnover. The low basal activity may, however, be an important feature of the AtMinD1 mode of action, as in E. coli MinD membrane dissociation only occurs after stimulation by MinE and this plays an essential role in the MinCDE oscillatory cycle (22Hu Z. Lutkenhaus J. Mol. Cell. 2001; 7: 1337-1343Abstract Full Text Full Text PDF PubMed Scopus (213) Google Scholar, 24Hu Z. Gogol E.P. Lutkenhaus J. Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 6761-6766Crossref PubMed Scopus (239) Google Scholar). Although AtMinD1 and AtMinE1 oscillatory behavior has not been reported, based on the evolutionary conservation of the division machinery it is probable that a similar mechanism occurs during plastid division. Although E. coli MinD and AtMinD1 show a high degree of similarity at the amino acid level, we have found significant differences in the functioning of the two proteins, First, a difference in cation dependence of ATPase activity between Ca2+-dependent AtMinD1 and Mg2+-dependent E. coli MinD implies an important functional difference, probably signifying evolutionary adaptation as many plant processes are regulated by calcium. Indeed, studies have suggested a regulatory role for plastidic Ca2+ fluxes (37Sai J. Johnson C.H. Plant Cell. 2002; 14: 1279-1291Crossref PubMed Scopus (132) Google Scholar), and our findings suggest plastidic Ca2+ levels regulate AtMinD1 activity during plastid division. Second, although we have demonstrated that AtMinE1 stimulates the activity of AtMinD1, this stimulation can occur independently of membrane binding (Fig. 3) in contrast to the phospholipid-dependent MinE stimulation of E. coli MinD (31de Boer P.A. Crossley R.E. Hand A.R. Rothfield L.I. EMBO J. 1991; 10: 4371-4380Crossref PubMed Scopus (260) Google Scholar), suggesting functional divergence in the mechanism of AtMinE1-stimulated AtMinD1 ATP hydrolysis during plastid division in Arabidopsis. In addition, purified MinE from E. coli is unable to stimulate AtMinD1 ATPase activity in vitro (Fig. 3), further indicating that at least in part the mechanism of AtMinE1-stimulated AtMinD1 activity is different from that in prokaryotes. In E. coli, wild-type MinD localizes to the cell periphery, whereas the active site MinD(K16Q) mutant is distributed throughout the cell (24Hu Z. Gogol E.P. Lutkenhaus J. Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 6761-6766Crossref PubMed Scopus (239) Google Scholar). MinD(K16Q) is, however, still able to bind ATP but cannot bind phospholipids (24Hu Z. Gogol E.P. Lutkenhaus J. Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 6761-6766Crossref PubMed Scopus (239) Google Scholar). In agreement with this we have shown that the active site AtMinD1(K72A) mutant exhibits aberrant localization patterns, distributed as speckles throughout plastids (Fig. 4). This mislocalization is not because of loss of AtMinD1 dimerization (Fig. 5) as observed for ARC11 (27Fujiwara M. Nakamura A. Itoh R. Shimada Y. Yoshida S. Møller S.G. J. Cell Sci. 2004; 117: 2399-2410Crossref PubMed Scopus (71) Google Scholar) but most probably because of lack of interaction with AtMinE1 (Fig. 5A). The loss of interaction between AtMinD1(K72A) and AtMinE1 may either be because lysine 72 is directly involved in the AtMinE1 interaction or because AtMinD1(K72A) is unable to adopt the correct conformation upon binding ATP necessary for AtMinE1 interaction. In E. coli, the α-helical region of the MinE anti-MinCD domain interacts with MinD α-helix 7 forming a coiled-coil structure (38Ma L. King G. Rothfield L. J. Bacteriol. 2003; 185: 4948-4955Crossref PubMed Scopus (49) Google Scholar, 32Ma L. King G.F. Rothfield L. Mol. Microbiol. 2004; 54: 99-108Crossref PubMed Scopus (49) Google Scholar), and lysine 11 within the Walker A region (P-loop) competes with MinE for residues within α-helix 7 (32Ma L. King G.F. Rothfield L. Mol. Microbiol. 2004; 54: 99-108Crossref PubMed Scopus (49) Google Scholar). MinE-mediated disruption of the non-covalent interaction between lysine 11 and α-helix 7 changes the lysine 11 side-chain orientation and the P-loop conformation. This transmits an activation signal to the neighboring catalytic domain or to the bound ATP, bringing about ATP hydrolysis (32Ma L. King G.F. Rothfield L. Mol. Microbiol. 2004; 54: 99-108Crossref PubMed Scopus (49) Google Scholar). This model suggests MinE stimulation is through conformational change in the Walker A motif rather than through direct interaction between MinE and Walker A residues. Based on this model it is unlikely that AtMinE1 interacts directly with the Walker A motif; we favor the theory that a Walker A mutation in AtMinD1 changes the overall conformation, ultimately disabling its interaction with AtMinE1. Based on our data we propose a working model for the AtMinD1 mode of action during plastid division in Arabidopsis (Fig. 6). We suggest AtMinD1 undergoes dimerization and binds ATP and that this AtMinD1 dimer complex exhibits low basal Ca2+-dependent ATPase activity. AtMinD1 then interacts with AtMinE1, stimulating ATP hydrolysis. How does this impinge on the function of AtMinD1 in ensuring the correct placement of the Z-ring? In line with the prokaryotic model, dimerized ATP-bound AtMinD1 may bind to the chloroplast envelope before AtMinE1 interaction, which stimulates ATP hydrolysis and membrane release followed by protein relocation. However, in contrast to E. coli, AtMinE1 can stimulate ATP hydrolysis in the absence of envelope lipids; therefore, AtMinE1 can enhance AtMinD1 activity prior to envelope binding. In addition, plants do not harbor MinC, suggesting that AtMinD1 and AtMinE1 modes of action differ from those in prokaryotes. Together with the fact that AtMinD1 is dependent on Ca2+ and not Mg2+, it is clear that AtMinD1 has evolved at the biochemical and cell biological level, presumably to adapt from being a part of cell division machinery in free living prokaryotes to becoming an integral component of the plastid division machinery in higher plants. We thank Jodi Maple and Xiang Ming Xu for helpful discussions. Download .pdf (.04 MB) Help with pdf files
Referência(s)