Tubular Liposomes with Variable Permeability for Reconstitution of FtsZ Rings

Masaki Osawa; Harold P. Erickson    Department of Cell Biology, Duke University Medical Center, Durham, North Carolina, USA

1 INTRODUCTION

FtsZ is a bacterial tubulin homologue that forms a ring structure called the “Z ring” at the division plane in bacteria. The Z ring is anchored to the membrane and constricts to divide the bacteria. FtsZ recruits a dozen other essential division proteins, which are mostly involved in remodeling the peptidoglycan cell wall. We recently succeeded in reconstituting Z rings inside tubular liposomes, and found that they generated a constriction force on the liposome wall (Osawa et al., 2008). The assembly of Z rings and the generation of the constriction force were achieved with FtsZ alone, and did not require any other division protein. This was an important discovery itself for understanding the mechanism of bacterial cell division. Now the liposome system we developed should provide a simple in vitro system for studying molecular details of how FtsZ works.

To achieve these results we had to overcome two technical problems. The first problem was to tether FtsZ to the membrane. Pichoff and Lutkenhaus (2005) discovered that the carboxy terminus of FtsZ binds to FtsA, and FtsA has an amphipathic helix at its carboxy terminal that inserts into the membrane. We made an FtsZ that could tether itself to the membrane by fusing an amphipathic helix (membrane targeting sequence: mts) to the carboxy terminus of FtsZ. To visualize the protein, we inserted a yellow fluorescent protein (YFP) before the mts, giving FtsZ-YFP-mts. Here we provide detailed protocols for the purification of FtsZ-YFP-mts.

The second problem was how to get the protein inside liposomes. We have succeeded in getting FtsZ-YFP-mts inside spherical unilamellar liposomes using the emulsion method (Noireaux and Libchaber, 2004; Pautot et al., 2003). However, we have never found Z rings assembled in such spherical unilamellar liposomes.

Eventually, we discovered a procedure that produced tubular multilamellar liposomes, and incorporated FtsZ-YFP-mts inside, where it formed Z rings. Initially this was a fortunate accident, since the cylindrical geometry was not designed, and the FtsZ was initially on the outside. We have since refined the procedures for producing tubular multilamellar liposomes, and we now understand some aspects of the permeability or leakiness that lets FtsZ inside. We describe here our protocols for producing the tubular liposomes and the tests of permeability.

2 REAGENTS

The following reagents are used in our experiments:

• Column buffer: 50 mM Tris/HCl, pH 7.9, 50 mM KCl, 1 mM EDTA, 10% (v/v) glycerol

• HMKCG buffer: 50 mM HEPES/KOH, pH 7.7, 5 mM MgAc, 300 mM KAc, 50 mM KCl 10% (v/v) glycerol

• HMK50-350 buffer: 50 mM HEPES/KOH, pH 7.7, 5 mM MgAc, 50–350 mM KAc

• DOPG:1,2-dioleoyl-sn-glycero-3-[phospho-rac-(1-glycerol)] (Avanti)

• Egg PC: phosphatidylcholine (Avanti)

• HccA: 7-Hydroxycoumarin-3-carboxylic acid (Invitrogen)

• Teflon disc, 37 mm diameter

3 BACTERIAL EXPRESSION OF MEMBRANE TARGETING FTSZ

The FtsZ-YFP-mts is expressed from a pET-11b expression vector, with FtsZ366-YFP-mts or FtsZ366-mts genes inserted at NdeI/BamHI sites (366 indicates that the FtsZ was truncated there, removing the FtsA-binding C-terminal peptide). The YFP we use is the variety Venus (Nagai et al., 2002), which gave superior results in FtsZ fusions in E. coli (Osawa and Erickson, 2005). The mts used here is the amphipathic helix from E. coli MinD (Szeto et al., 2003). We have not yet tested the amphipathic helix from FtsA, which has three to five additional extra amino acids that extend the amphipathic helix (Pichoff and Lutkenhaus, 2005). The expression vector is transformed into E. coli strain C41 (Miroux and Walker, 1996), which gives better yields of soluble proteins than BL21.

After transforming, colonies are selected on an LB (Luria broth) agar plate containing 100 μg/ml ampicillin. A colony is picked and cultured overnight in 50 ml LB media with 100 μg/ml ampicillin at 37 °C.

Five milliliters of the overnight culture is diluted in 500 ml LB and cultured at 37 °C until the optical density at 600 nm reaches 0.8–1.0. Protein expression is induced by addition of 0.5 mM IPTG and at the same time the temperature of the shaker is set to 20 °C (our shaker takes 1–2 h to reach 20 °C).

The cells are cultured overnight and spun down at 3750 rpm for 45 min in a Beckman GPR rotor.

4 PURIFICATION OF FTSZ-MTS AND FTSZ-YFP-MTS

Since FtsZ-mts and FtsZ-YFP-mts are expressed as soluble proteins, we purify them using the same protocol as for wild-type FtsZ.

The packed cells are resuspended in a final volume of 20 ml column buffer, and 1 mM phenylmethanesulphonylfluoride (PMSF) and 0.1–0.2 mg/ml lysozyme are added. The mixture is then incubated on a rotator at 4 °C for 15 min. They are frozen at 80 °C overnight or longer.

Two cycles of freeze–thaw (fresh 1 mM PMSF is added after each thawing) method are performed. The resultant mixture is sonicated on ice until the viscosity is reduced. We usually sonicate it for three cycles of 20 s, with 1 min cooling intervals.

It is then centrifuged at 32,000 rpm for 20 min at 4 °C (Beckman 42.1 Ti rotor). The supernatant is collected and ammonium sulfate is added to 30% saturation (3.52 g dry ammonium sulfate to the 20 ml volume). This mixture is incubated for 20 min on ice and again centrifuged at 32,000 rpm for 20 min at 4 °C (Beckman 42.1 Ti rotor). The supernatant is discarded and the pellet is resuspended in 10 ml column buffer and passed through a 0.22 μm filter.

The protein is purified on an anion exchange column. A 1 × 10 cm Source Q column (Source 15Q, GE Healthcare) is used. The column is eluted with a 100 ml gradient from 50–500 mM KCl in column buffer.

4.1 For FtsZ-mts

FtsZ has very low UV absorbance, so the peak is located by running each fraction on SDS–PAGE.

The peak fractions are pooled and dialyzed into HMK350.

The protein concentration is determined by the BCA method (Pierce). FtsZ produces 75% as much color as BSA (Lu et al., 1998), so it is necessary to correct for this.

Aliquots are frozen and stored at − 80 °C.

4.2 For FtsZ-YFP-mts

After elution from the Source 15Q column, the peak fractions are pooled. There are typically two peaks: a large main peak and a following small peak, and both peaks have an indistinguishable activity. These peaks can be identified by yellow fluorescence and confirmed by SDS–PAGE.

They are concentrated using an Amicon Ultra-15 with centrifugation at 5000×g.

We have noted that incomplete boiling of FtsZ-YFP-MTS with SDS sample buffer generates two bands on the gel. The upper band (68 Kd) results from completely denatured protein and the lower band (60 Kd), which still has yellow fluorescence in the gel, is due to FtsZ-YFP-mts where the YFP is not denatured.

The concentration of FtsZ-YFP-mts can be determined from its absorption at 515 nm, using the extinction coefficient for YFP-Venus 92,200 M− 1 cm− 1.

Our preferred buffer for FtsZ-YFP-mts is now HMKCG because FtsZ-YFP-mts seems to be more stable, as described below; we now use HMKCG for dialysis, dilution, reaction, and storage buffer.

5 RENATURED PREPARATION OF FTSZ-YFP-MTS

In our previous study (Osawa et al., 2008), we used a renaturing technique to prepare FtsZ366-YFP-mts. We developed this protocol because an early preparation...