ARTICLE IN PRESS

Reconstituting geometrymodulated protein patterns in membrane compartments Katja Zieske1, Petra Schwille1 Department of Cellular and Molecular Biophysics, Max Planck Institute for Biochemistry, Martinsried, Germany 1

Corresponding author: E-mail: [email protected], [email protected]

CHAPTER OUTLINE Introduction ................................................................................................................ 2 1. Reagents ............................................................................................................... 4 2. Methods ................................................................................................................ 4 2.1 Min Protein Expression and Purification .................................................... 4 2.2 Microfabrication of Microcompartments .................................................... 5 2.3 Supported Lipid Membranes .................................................................... 6 2.4 Reconstitution of Min Protein Oscillations ................................................. 8 2.5 Analysis of Min Protein Oscillations in Dependence of Compartment Length 9 Discussion and Future Perspectives ........................................................................... 11 Acknowledgments ..................................................................................................... 14 References ............................................................................................................... 14

Abstract The MinCDE protein system from Escherichia coli has become one of the most striking paradigms of protein self-organization and biological pattern formation. The whole set of Min proteins is functionally active to position the divisome machinery by inhibiting Z ring assembly away from mid-cell. This is accomplished by an oscillation behavior between the cell poles, induced by the reaction between the two antagonistic proteins MinD and MinE, which has long caught the attention of quantitative biologists. Technical advances in fluorescence microscopy and molecular biology have allowed us in the past years to reconstitute this MinDE self-organization in cell-free environments on model membranes. We verified the compositional simplicity of protein systems principally required for biological pattern formation, and subjected the mechanism to quantitative biophysical analysis on a single-molecule level. On flat extended membranes, MinD and MinE self-organized Methods in Cell Biology, Volume 128, ISSN 0091-679X, http://dx.doi.org/10.1016/bs.mcb.2015.02.006 © 2015 Elsevier Inc. All rights reserved.

1

ARTICLE IN PRESS

2

Reconstituting geometry-modulated protein patterns

into parallel propagating waves. Moreover, employing microsystems technology to construct membrane-clad soft polymer compartments mimicking the shape of native E. coli cells has further enabled us to faithfully reproduce Min protein oscillations. We further investigated the response of this self-organizing molecular system to three-dimensional compartment geometry. We could show that Min protein patterns depend strongly on the size and shape of the compartment, and the oscillation axis can only be preserved within a certain length interval and narrow width of the compartment. This renders the Min system a perfectly adapted oscillator to the bacterial cell geometry.

INTRODUCTION Among the most essential features of the molecular systems that constitute cellular life is their ability to self-organize within the three-dimensional framework of the compartment that they are enclosed in. This includes the spontaneous generation of ordered structures, either as self-assembled macromolecular complexes, ordeven more bewildering at first glancedthe constitution of gradients and patterns of soluble molecules, which apparently escape the dictate of entropy, i.e., of spatial equilibration. The thermodynamic preconditions of such a spontaneous generation and homeostasis of ordered systems have most popularly been formulated by Schro¨dinger in his famous Dublin lecture series in 1944, “What is life”: Mainly, the continuous flow of energy through a chemical system, on the expense of external order (Schro¨dinger, 1944). The kinetic requirements for a reaction system to actually exhibit self-organization and pattern formation, particularly in the presence of molecular diffusion of its reactants, have in the early 1950s been first described by Alan Turing, in his seminal and visionary article on the chemical basis of morphogenesis (Turing, 1952). However, although it is evident that spatial unmixing of reactants has to occur at the beginning of every life form, and that these autonomously formed gradients are playing essential roles during the differentiation and development of higher organisms, few biochemical reaction systems have so far been directly observed displaying such fundamental self-organization when reconstituted in dramatically simplified cell-free environments. One that has recently caught a lot of attention is the set of Min proteins, MinCDE, found in various microorganisms and most prominently in Escherichia coli, where it faithfully positions the cell division machinery to mid-cell, ensuring symmetric division into equally sized daughter cells (de Boer, Crossley, & Rothfield, 1989). Remarkably, the mode of action of the Min proteins in E. coli involves their oscillations between the cell poles on a timescale of minutes, thereby directing the cell division plane to mid-cell by inhibiting the formation of the Z ring away from the geometric center (Lutkenhaus, 2007). The reaction mechanism, although still not fully understood, is based on an ATPdependent cycling of the ATPase MinD between the bacterial cytosol and the inner membrane, presumably based on a conformational transition which allows an

ARTICLE IN PRESS

Introduction

amphipathic helix and thus, a membrane-binding motif, to be exhibited in the ATPbound form (de Boer, Crossley, Hand, & Rothfield, 1991; Hu & Lutkenhaus, 2003; Lackner, Raskin, & de Boer, 2003; Szeto, Rowland, Rothfield, & King, 2002; Zhou & Lutkenhaus, 2003). Upon enrichment of MinD on the membrane, its antagonist MinE is being recruited, which catalyzes the ATP hydrolysis of MinD, presumably rendering the amphipathic helix to be retracted and MinD to be released from the membrane (Hu & Lutkenhaus, 2003; Lackner et al., 2003). MinE acts processively on the remaining MinDs, inducing a directionality of release, which on large-scale results in a traveling wave of MinD, and consequently also MinE, surface concentration (Loose, Fischer-Friedrich, Herold, Kruse, & Schwille, 2011; Park et al., 2011). Released molecules diffuse in the fluid phase until they undergo the next cycle of ATP binding and membrane attachment, at a distance away from their point of release which is encoded by diffusion, as well as the kinetic rate constants of solution reactions. The core set of functional modules, such as MinD, MinE, and ATP, when reconstituted on an artificial bilayer mimicking the bacterial inner membrane, selforganizes into striking patterns (Loose, Fischer-Friedrich, Ries, Kruse, & Schwille, 2008), resembling those found in inorganic self-organizing systems such as the famous BelousoveZhabotinsky reaction, or for heterogeneous catalysis of CO oxidation on platinum substrates (Ertl, 1991). In fact, it appears that the membrane in the MinDE oscillation machinery can, with some confidence also, be viewed as a heterogeneous catalyst, inducing at least one of the reactants to exhibit their antagonistic activity. Obviously, the patterns formed depend strongly on reaction conditions, such as concentrations, temperature, diffusion coefficients in solution and on the surface, but also on the composition, and the geometry of the membrane as the major template for several reaction steps. Thus, when searching for the set of parameters to initiate a particular pattern, it appears to be one straightforward approach to structure and manipulate the membrane shape. In order to convert the traveling waves observed on extended flat membranes into true oscillations, the system needs to be confined in volume. Remarkably, as we have recently shown, this confinement does not have to involve a closed membrane fully surrounding the reaction volume. Instead, open “bathtub” like compartments which are membrane coated on their bottom and interfaced with air on top fulfill all the functional geometric requirements for the establishment of true pole-to-pole oscillations of MinCDE (Zieske & Schwille, 2013, 2014). They can easily be engineered in large numbers on a chip, to parallelize experiments for much better statistics than at all possible on live cells. Their shape can be modified basically ad libitum, such that any length or length-to-width ratio can be realized, and even the closing septum geometry can be mimicked by inserting a small neck region between two half-cells (Zieske & Schwille, 2014). We could demonstrate that the oscillations responded to a simulated one-dimensional growth of the compartment as observed in vitro: The oscillation period grew up to a certain length, until a splitting into two synchronous oscillations could be observed. On the other side, in round compartments, i.e., without a clear long axis of the compartments, a preferential direction of the

3

ARTICLE IN PRESS

4

Reconstituting geometry-modulated protein patterns

oscillations could hardly be established, leading to random patterns unable to form gradients and position downstream processes such as FtsZ assembly (Zieske & Schwille, 2013). In this chapter, we describe the pattern-forming Min protein assay, and particularly focus on our recently developed procedures to employ microsystems technology to engineer cell-shape mimicries coated with lipid membranes. We will conclude with a short outlook on the coreconstitution of more complex systems, such as a minimal E. coli divisome, being directed by the oscillatory patterns to the center of a cell-like compartment.

1. REAGENTS • • • • • • •

• • • •

E. coli BL21 with overexpression plasmids for MinD and MinE LB medium Supplements for media: Kanamycin, IPTG (Isopropyl-b-D-thiogalactopyranosid) For buffers: Tris, Hepes, NaCl, KCl, MgCl2, imidazole, mercaptoethanol, glycerol, protease inhibitor (Roche), ADP, ATP, EDTA Ni-NTA column for purification Alexa488 Fluor C5 maleimide (molecular probes, Carlsbad, CA) For microstructures: Si-wafer (Si-Mat, Kaufering, Germany), photoresist (ma-P 1275, micro resist technology GmbH, Germany), developer solution (ma-D 531, micro resist technology GmbH, Germany), chlorotrimethylsilane (SigmaeAldrich) PDMS (Polydimethylsiloxane) and cross-linker (Sylgard184, Dow Corning) E. coli polar extract in chloroform (Avanti Polar Lipids, Alabaster, AL) Membrane dye (FAST DiI, Life Technologies, Carlsbad, CA) UV-Glue (Norland Optical Adhesive 63, Norland Products, Cranbury, NJ)

2. METHODS Below we present the detailed methods and protocols that have been established previously in our lab to reconstitute Min protein oscillations and protein gradient formation, targeting FtsZ filaments to the middle of specially manufactured compartments. This cell-free reconstitution of essential components of the E. coli divisome has enabled us to study the impact of geometry on protein self-organization of the Min/FtsZ protein system.

2.1 MIN PROTEIN EXPRESSION AND PURIFICATION MinD and MinE were expressed from the plasmids pET28a-MinD and pET28aMinE which code for MinD and MinE, respectively, connected to an N-terminal hexahistidine tag by a short linker sequence and a thrombin cleavage site.

ARTICLE IN PRESS

2. Methods

Step 1: protein induction. E. coli BL21 cells with the corresponding overexpression plasmids for MinD or MinE were inoculated in 10 mL LB medium (supplemented with 50 mg/mL kanamycin) and grown at 37  C overnight. The next morning the overnight culture was diluted in 1 L LB medium (with 50 mg/mL kanamycin) and grown at 37  C to an OD600 of 0.7. Then, protein expression was induced by adding 1 mM IPTG. After 4 h of incubation at 37  C, bacteria were collected by centrifuging bacteria at 4500 g for 10 min at 4  C. The bacteria pellet could be stored at 80  C or protein purification proceeded. Step 2: cell lysis. The pellet was resuspended in 45 mL lysis buffer (50 mM Tris pH 8.0, 300 mM NaCl, 10 mM imidazole, 10 mM mercaptoethanol, protease inhibitor, and 0.2 mM ADP (ADP only for MinD purification)). Afterward, the bacteria were broken by an EmulsiFlex-C3 (Avestin) and the lysate was centrifuged at 25,000 g at 4  C. Step 3: protein purification. The supernatant was transferred to a Ni-NTA column, washed with wash buffer (50 mM Tris pH 8.0, 300 mM NaCl, 20 mM imidazole, 10 mM mercaptoethanol, 10% glycerol, protease inhibitor (Roche)) and proteins were eluted with elution buffer (50 mM Tris pH 8.0, 300 mM NaCl, 250 mM imidazole, 10 mM mercaptoethanol, 10%, glycerol, protease inhibitor). Finally the protein buffer was exchanged to storage buffer (50 mM HEPES pH 7.25, 150 mM KCl, 10% glycerol, 0.1 mM EDTA, 0.2 mM ADP (ADP only for MinD purification)). Step 4: labeling of MinE. MinE harbors a single cysteine residue within its sequence, which was labeled with Alexa488 Fluor C5 maleimide (Molecular Probes) according to the manufacturers’ manual. Step 5: storing the proteins. MinD, MinE, and MinE.AlexaFluor488 were aliquotted to prevent multiple subsequent thawing and freezing circles, frozen in liquid nitrogen and stored at 80  C.

2.2 MICROFABRICATION OF MICROCOMPARTMENTS Microstructures of photoresist on top of Si-wafers were produced in a clean room. Alternatively to engineering the microstructures in house (Steps 1e2), which is cheaper but requires specialized equipment, wafers with microstructures might be purchased from a company specialized in microfabrication processes. Step 1: Design of micropatterns. The two-dimensional layout of the microstructures was designed using the software AutoCAD and a chrome mask comprising the respective structures purchased (Compugraphics Jena GmbH). To investigate compartment length dependence of Min protein oscillations, the twodimensional mask patterns had a constant width of 10 mm and different lengths of 12, 15, 20, 25, 35, 45, 55, 65, 95, 145, and 245 mm. Multiple copies of the different structures were arranged in an array and covered a macroscopic area of about 4  4 mm. Step 2: Engineering microstructures. Microstructures were produced by photolithography. A Si-wafer was placed for 30 s in an HMDS vapor chamber to

5

ARTICLE IN PRESS

6

Reconstituting geometry-modulated protein patterns

improve adhesion of resist to the wafer. Then a 10 mm thick photoresist layer was spin-coated on top of the wafer and baked for 30 min at 90  C. Afterward, the photoresist layer was exposed to UV-light through the chromium mask featuring the required structures. Finally, the photoresist patterns were developed by immersing the wafer in developer solution for about 15 min until the photoresist micropattern was developed. Then the wafer was rinsed with water and dried by spinning for 30 s. The following steps were performed in standard laboratory rooms. Step 3: Pouring PDMS on microstructures. To prevent sticking of PDMS to the wafer after curing of PDMS, the Si-wafers with resist microstructures was placed for 30 min in a vapor chamber with chlorotrimethylsilane. Then 15 g PDMS was mixed with 1.5 g cross-linker and degassed under vacuum until all air bubbles had disappeared. The wafer was placed in a frame made out of aluminum foil and the PDMS mixture was poured on top of the wafer. Step 4: Adding glass coverslips as PDMS supports. Clean glass coverslips (22  22 mm) were then manually pressed into the liquid PDMS. Thereby, a thin PDMS film of about 30 mm resides between the microstructured wafer and the glass coverslip (Figure 1(A)). It is important that the thickness of glass coverslip plus PDMS is small enough to be within the working distance of the objective. Step 5: Curing PDMS microstructures. The PDMS was then cured at 80  C. The exact time for curing the PDMS is not critical for this protocol. A few hours are sufficient to cure the PDMS, however, one can cure the PDMS overnight, as well. Step 6: Separating microstructures from the wafer. After curing, the bulk PDMS layer above the glass coverslip was peeled off. Thereby, the thin PDMS film and the glass coverslip remain on the wafer. Then, the glass coverslip with the thin microstructured PDMS film attached was carefully separated from the wafer. A razor blade is useful at this step to carefully lift the glass slip without breaking the thin glass (Figure 1(A)). The microstructured PDMS/glass devices were wrapped in aluminum foil and stored at room temperature.

2.3 SUPPORTED LIPID MEMBRANES Step 1: Vesicle preparation. E. coli polar lipids (64 mL of a 25 mg/mL stock solution in chloroform) were pipetted in a glass vial and dried under a weak nitrogen stream. Placing the vial with dry lipids 30 min in vacuum removed residual solvents. Subsequently, the lipids were rehydrated with 400 mL Min-buffer (25 mM Tris-HCl pH 7.5, 150 mM KCl, 5 mM MgCl2) and incubated for 30 min at 37  C. Afterward, the vial was vortexed. This step results in multilamellar vesicles. Then, the vesicle solution was sonicated in an ultrasonic bath for about 20 min. Sonification results in unilamellar vesicles, with the solution becoming more transparent. Depending on the strength of the sonicator, the duration of this step needs to be adjusted. The vesicle solution was aliquotted in PCR tubes and stored at 20  C until further use.

ARTICLE IN PRESS

2. Methods

FIGURE 1 Experimental setup: (A) Si-wafer with cured PDMS. Part of the bulk PDMS had been removed, enabling the separation of glass coverslips with the attached microstructured PDMS layer from the wafer surface. (B) Supported lipid membranes adapt to the topography of the microstructures, and dynamic Min protein waves can be reconstituted on these surfaces. (C) Min protein waves on microstructured membrane-clad surfaces self-assemble into dynamic wave-like patterns. (D) Removal of the bulk buffer reservoir results in confinement of the Min reaction system to small compartments of defined geometry, and leads to the oscillatory behavior of the Min proteins. Zieske & Schwille, eLife 2014; doi:10.7554/eLife.03949, creative commons (Zieske & Schwille, 2014).

7

ARTICLE IN PRESS

8

Reconstituting geometry-modulated protein patterns

Step 2: Preparation of sample chambers. Before the microstructured PDMS was used as a membrane support, it was sonicated for 5 min in ethanol, washed with water, and air-dried. Then the PDMS surface was treated with air plasma to render the surface hydrophilic. To create a reaction chamber, the tip and the lid of a laboratory plastic tube (0.5 mL) was cut off, and the remaining plastic ring was glued with UV-Glue on top of the PDMS surface. The glue was cured under an UV-lamp for 10 min. Step 3: Generating supported lipid membranes. 10mL of the vesicle solution was thawed and diluted 7.5-fold with buffer (25 mM Tris-HCl pH 7.5, 150 mM KCl, 5 mM MgCl2). The diluted vesicle solution was added in the plastic chamber on top of the PDMS surface and 3 mL of CaCl2 (100 mM stock solution) was added. The vesicles on top of the PDMS surface were then incubated for 20e30 min at 37  C. During this time the vesicles ruptured on top of the membrane support and fused to form a lipid bilayer, which adapted to the topography of the microstructured PDMS. Subsequently, residual vesicles were washed away by repeating cycles of diluting the vesicle solution with buffer, and subsequently removing a fraction of the buffer. In total, the membranes were washed with 2 mL buffer (25 mM Tris-HCl pH 7.5, 150 mM KCl, 5 mM MgCl2). After washing the membrane, the buffer volume in the reaction chamber was adjusted to 200 mL. Notably, for the reconstitution of Min protein patterns, it is important that the supported lipid membrane is intact and the lipids are freely mobile within the membrane plane. If the membrane assay is applied for the first time or if troubleshooting is otherwise required, the membrane should be labeled for visualization by fluorescence microscopy. This can be accomplished by adding a membrane dye (e.g., 0.05% DiI) before the lipid solvent is dried. The quality of the membrane can be analyzed by imaging the membrane, and the mobility of lipids can be probed by FRAP (simpler) or by FCS measurements. The overall appearance of the membrane should be homogeneously fluorescent across the whole PDMS surface (without interrupted membrane patches or holes) and the microcompartments should be clearly discernible if the membrane adapts faithfully to the topography of the support. In FRAP experiments with the membrane dye, the fluorescence intensity of a bleached region should recover.

2.4 RECONSTITUTION OF MIN PROTEIN OSCILLATIONS Step 1: Reconstituting Min protein patterns. Min protein patterns were induced by adding 1 mM MinD, 1 mM MinE (supplemented with 5e10% MinE.Alexa488), and 2.5 mM ATP to the reaction chamber. The proteins were mixed with the buffer by pipetting the buffer within the reaction chamber several times up and down. The Min proteins started to form patterns within a few minutes and these patterns were imaged with an inverted confocal microscope (Figure 1(B) and (C)). Initially the patterns are irregular, but after about 30 min to 1 h the patterns are remodeled into regular parallel waves and spirals.

ARTICLE IN PRESS

2. Methods

Important to consider: 1. Note that there is always a substantial amount of protein in the buffer. For visualizing the protein patterns occurring on the membrane with high contrast, it is recommended to discriminate against the signal of the proteins in the bulk by using a confocal or a total internal reflection fluorescence microscope. 2. If protein patterns are not forming, the functionality of the membrane should be checked. In other words, the membrane should form a homogeneous layer without holes that adapts to the topography of the surface and the lipid molecules should be mobile within the membrane plane. Step 2: Confining Min proteins in microcompartments. The buffer was reduced by manually removing the bulk buffer reservoir by pipetting (Figure 1(D)). After reducing the buffer it is important to immediately close the reaction chamber with a lid or with parafilm to limit evaporation of the buffer in the microcompartments. It might require some practice to carefully remove the appropriate amount of buffer. If the protocol is tried for the first time, the buffer should be reduced stepwise while observing how the meniscus of the buffer approaches the membrane surface. Step 3: Observing Min protein oscillations. After confining the protein solution within cell-shaped compartments, the protein patterns oscillated from one pole to the other (Figure 2). Time-lapse movies of the dynamic patterns were acquired with an inverted confocal microscope (LSM 780, Zeiss). Note that although the notion “Min protein oscillation” is well established in the literature, the dynamic behavior of the individual Min proteins is not a directed movement from one pole to the other. The observed oscillations along a well-defined direction are movements of protein patterns and not of individual proteins. Instead, individual membrane-attached proteins perform random movements on the membrane before they detach again (Loose et al., 2011). The appearance of the Min protein patterns can be described by a reaction diffusion mechanism.

2.5 ANALYSIS OF MIN PROTEIN OSCILLATIONS IN DEPENDENCE OF COMPARTMENT LENGTH Time-lapse images of Min protein patterns in compartments with different geometries contain a wealth of information. In compartments of different length but a constant width of 10 mm (Figure 2), time-lapse images reveal that Min protein patterns in longer compartments have higher oscillation modes than in shorter compartments. More quantitative information can be obtained by analyzing the same images in more detail. Apart from the shortest compartments, in which the oscillation axis switches occasionally, the protein patterns in compartment with different lengths always moved along the long axis of the compartments. Thus, the fluorescence intensities are mainly modulated with time and along one dimension. Therefore it is advisable to perform kymograph analysis, which is a graphical representation of the

9

ARTICLE IN PRESS

10

Reconstituting geometry-modulated protein patterns

FIGURE 2 Min protein oscillations in compartments with different length. Hundreds of compartments with systematically changing geometry can be imaged with the cell-free reconstitution assay for Min protein oscillations. In short compartments, the Min protein patterns oscillate from pole-to-pole. In long compartments, higher order modes with multiple maxima are generated. 1 mM MinD, 1 mM MinE (supplemented with 10% MinE.Alexa488) Scale bar: 50 mm.

localization along a specified spatial axis over time. For the Min proteins in different compartments, the fluorescence signal along the length axis of the compartments is represented over the elapsed time. Figure 3 depicts five kymographs of Min protein oscillations in compartments with increasing length, whereby the bright signal represents the concentration maxima of the fluorescently labeled protein MinE during the oscillations. Pole-to-pole oscillations of MinE are represented by the first two kymographs in Figure 3. By quantitatively analyzing the kymographs, also the velocities of spatiotemporal Min protein patterns can be determined. A shallow slope represents fast moving Min patterns, whereas an increasing slope represents slowly moving patterns. The MinE kymographs in Figure 3 demonstrate that the speed of the MinE maximum tends to be higher at the beginning of a newly assembled MinE maximum and at the compartment ends than at the other locations along the compartment length. An intriguing advantage of the cell-free assay for the reconstitution of Min protein oscillations is the simultaneous acquisition of protein patterns under the same conditions on one sample. While specific properties of the Min protein oscillations, such as the oscillation period, depend on many cellular and external factors that render a length-dependent characterization of the oscillation period in different bacteria difficult, these measurements can be easily performed in vitro. Applying the

ARTICLE IN PRESS

Discussion and future perspectives

FIGURE 3 Kymographs of MinE in compartments with different lengths. In the two shortest compartments, the characteristic kymograph patterns of pole-to-pole oscillations are visible. Higher order oscillation modes result in fish scales pattern in the kymographs (last three kymographs).

cell-free system, we demonstrated that the oscillation period increased in longer compartments (Figure 4).

DISCUSSION AND FUTURE PERSPECTIVES In the previous section, we described in detail how, on basis of the oscillating Min protein machinery from E. coli, simple biochemical oscillators can be constructed from the bottom-up. The antagonistic pair of proteins MinD and MinE, together with ATP and a membrane surface as a catalytically active substrate, assemble into propagating parallel waves, which can under confinement in a membraneclad compartment be converted to oscillations. As described, the establishment of the oscillation axis, as well as the period of the oscillation, depends on the spatial scales, such that the oscillator can in principle be tuned by geometry. In live cells, the role of the Min oscillator is to establish a time-averaged gradient of MinC, being a “passenger” of the described MinDE oscillations by binding to MinD (Hu & Lutkenhaus, 1999; Raskin & de Boer, 1999). MinC, on the other hand, serves as an inhibitor of Z ring assembly and thus directs all downstream processes connected to the composition of the Z division ring. The dynamic inhibition of FtsZ filament attachment to the membrane by MinC has been confirmed and mechanistically investigated in a reconstitution assay on flat extended membranes, with a membrane-targeted FtsZ mutant, FtsZ-mts (Arumugam, Petrasek, & Schwille, 2014).

11

ARTICLE IN PRESS

12

Reconstituting geometry-modulated protein patterns

FIGURE 4 The oscillation period of the Min protein patterns increases in longer compartments. The oscillation period of protein patterns on a chip with more than 100 compartments were evaluated in dependence of compartment length. The average oscillation period and standard deviation are depicted for different compartment lengths. Due to higher order oscillation in long compartments the number of data points for pole-to-pole oscillations in compartments of 55 mm length was limited and individual data points are depicted in the graph. Zieske & Schwille, eLife 2014; doi:10.7554/eLife.03949, creative commons (Zieske & Schwille, 2014).

Based on the compartment assay described above, we could show, in a co-reconstitution of the full MinCDE positioning machinery and FtsZ-mts, that time-averaged bipolar MinC gradients can be achieved, confining the localization of FtsZ bundles to the middle of the compartment (Zieske & Schwille, 2014, Figure 5). When introduced to the cell-shaped containers featuring Min oscillations, it could be shown that FtsZ-mts protoring fragments were aligned perpendicularly to the long axis in the central region of the compartment, an important precondition for assembly of the divisome machinery. Thus, while the reconstitution of FtsZ into freestanding tubular vesicles resulted into membrane-tethered Z rings, which provides an intriguing system to study force generation of membrane interacting proteins (Osawa, Anderson, & Erickson, 2008), the cell-shaped compartments provide a complementary cell-free system to systematically study geometry-modulated processes and the localization of FtsZ. Recent work on the self-organization of FtsZ coreconstituted with its native membrane adaptor FtsA, yielding amazing dynamics of swirling FtsZ protorings on a flat membrane surface (Loose & Mitchison, 2014)

ARTICLE IN PRESS

Discussion and future perspectives

FIGURE 5 Coreconstitution of Min oscillations and FtsZ results in large-scale gradient formation that confines Z protoring filaments to the center of the reaction compartment, as proposed for live cells. MinE: red (dark gray in print versions), FtsZ-mts: blue (white in print versions), Scale bar: 10 mm.

spurred new hypotheses that these dynamics may be constitutive for any sort of contractile force exerted by the Z ring. It will thus be an obvious goal toward identifying a minimal divisome machinery, to introduce native membrane adaptors, such as FtsA but also ZipA, into the above described assay. Finally, the possibility of massive parallelization and of designing compartments of variable length are only two of the many advantages that are conferred to bottom-up synthetic biology by microsystems technology. More complex shapes of cell mimicry, e.g., the simulation of a small neck region between two separate compartments, simulating the possible intermediate states during septum formation in cell division, could be realized. Preliminary work along these lines shows that the oscillations respond dramatically to the size of the neck region, changing from synchronous to asynchronous oscillations in the two half-compartments after a certain degree of neck closure (Figure 6). This behavior is also dependent on the total length of the cells or cell halves. Closing the artificial septum in very short cells results in a loss and new orientation of the oscillation axis (Figure 6(A), right). Taken together, we described here the details of our reconstitution approach involving the mimicry of cell geometry by microfabricated soft polymer

13

ARTICLE IN PRESS

14

Reconstituting geometry-modulated protein patterns

FIGURE 6 Simulated closing of a septum connecting two cell-like compartments, mimicking the stages of cell division. (A) Below a certain width of the septum, the pole-to-pole oscillations switch into two symmetric “daughter” oscillations (left). In very short simulated cells closing of the septum results in a loss of the oscillation axis (right). Scale bar: 20 mm (B) Kymographs of the third (*) and fourth (**) compartment in (A). (C) Pole-to-pole oscillation versus symmetric oscillations of the complete MinCDE system. MinE: red (dark gray in print versions), MinC: yellow (white in print versions), Scale bar: 10 mm.

compartments, tailored toward the pattern and oscillation producing Min protein system. Our long-term goal is the reconstitution of the full molecular machinery required for the division of a cell-like compartment. While for the last step of actual compartment splitting, the biochemical system will have to be transferred to a deformable compartment (e.g., a lipid vesicle), we believe that the assay described above based on membrane-coated PDMS compartments will be of great use for any quantitative analysis of this or similar pattern-forming or polarity-inducing systems.

ACKNOWLEDGMENTS The authors thank Martin Loose for establishing the protocol for Min protein purification and Min protein reconstitution on flat membranes in the lab and Simon Kretschmer for coments on the manuscript. KZ is supported by “The International Max Planck Research School for Molecular and Cellular Life Sciences” (IMPRS-LS). KZ and PS are associated with CeNS (Center for nanoscience) of the Ludwig-Maximilians University Munich, and the collaborative research center SFB 1032.

REFERENCES Arumugam, S., Petrasek, Z., & Schwille, P. (2014). MinCDE exploits the dynamic nature of FtsZ filaments for its spatial regulation. Proceedings of the National Academy of Sciences of the United States of America, 111(13), E1192eE1200. de Boer, P. A., Crossley, R. E., Hand, A. R., & Rothfield, L. I. (1991). The MinD protein is a membrane ATPase required for the correct placement of the Escherichia coli division site. The EMBO Journal, 10(13), 4371e4380.

ARTICLE IN PRESS

References

de Boer, P. A., Crossley, R. E., & Rothfield, L. I. (1989). A division inhibitor and a topological specificity factor coded for by the minicell locus determine proper placement of the division septum in E. coli. Cell, 56(4), 641e649. Ertl, G. (1991). Oscillatory kinetics and spatio-temporal self-organization in reactions at solid surfaces. Science, 254(5039), 1750e1755. Hu, Z., & Lutkenhaus, J. (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. Molecular Microbiology, 34(1), 82e90. Hu, Z., & Lutkenhaus, J. (2003). A conserved sequence at the C-terminus of MinD is required for binding to the membrane and targeting MinC to the septum. Molecular Microbiology, 47(2), 345e355. Lackner, L. L., Raskin, D. M., & de Boer, P. A. (2003). ATP-dependent interactions between Escherichia coli Min proteins and the phospholipid membrane in vitro. Journal of Bacteriology, 185(3), 735e749. Loose, M., Fischer-Friedrich, E., Herold, C., Kruse, K., & Schwille, P. (2011). Min protein patterns emerge from rapid rebinding and membrane interaction of MinE. Nature Structural & Molecular Biology, 18(5), 577e583. Loose, M., Fischer-Friedrich, E., Ries, J., Kruse, K., & Schwille, P. (2008). Spatial regulators for bacterial cell division self-organize into surface waves in vitro. Science, 320(5877), 789e792. Loose, M., & Mitchison, T. J. (2014). The bacterial cell division proteins FtsA and FtsZ selforganize into dynamic cytoskeletal patterns. Nature Cell Biology, 16(1), 38e46. Lutkenhaus, J. (2007). Assembly dynamics of the bacterial MinCDE system and spatial regulation of the Z ring. Annual Review of Biochemistry, 76, 539e562. Osawa, M., Anderson, D. E., & Erickson, H. P. (2008). Reconstitution of contractile FtsZ rings in liposomes. Science, 320(5877), 792e794. Park, K. T., Wu, W., Battaile, K. P., Lovell, S., Holyoak, T., & Lutkenhaus, J. (2011). The Min oscillator uses MinD-dependent conformational changes in MinE to spatially regulate cytokinesis. Cell, 146(3), 396e407. Raskin, D. M., & de Boer, P. A. (1999). MinDE-dependent pole-to-pole oscillation of division inhibitor MinC in Escherichia coli. Journal of Bacteriology, 181(20), 6419e6424. Schro¨dinger, E. (1944). What is life?: The physical aspect of the living cell. Based on lectures delivered under the auspices of the Dublin Institute for Advanced Studies at Trinity College, Dublin. in February 1943. Szeto, T. H., Rowland, S. L., Rothfield, L. I., & King, G. F. (2002). Membrane localization of MinD is mediated by a C-terminal motif that is conserved across eubacteria, archaea, and chloroplasts. Proceedings of the National Academy of Sciences of the United States of America, 99(24), 15693e15698. Turing, A. M. (1952). The chemical basis of morphogenesis. Philosophical Transactions of the Royal Society B, 237(641), 37e72. Zhou, H., & Lutkenhaus, J. (2003). Membrane binding by MinD involves insertion of hydrophobic residues within the C-terminal amphipathic helix into the bilayer. Journal of Bacteriology, 185(15), 4326e4335. Zieske, K., & Schwille, P. (2013). Reconstitution of pole-to-pole oscillations of min proteins in microengineered polydimethylsiloxane compartments. Angewandte Chemie International Edition England, 52(1), 459e462. Zieske, K., & Schwille, P. (2014). Reconstitution of self-organizing protein gradients as spatial cues in cell-free systems. eLife, 3.

15

Reconstituting geometry-modulated protein patterns in membrane compartments.

The MinCDE protein system from Escherichia coli has become one of the most striking paradigms of protein self-organization and biological pattern form...
1MB Sizes 3 Downloads 7 Views