Synthetic Membrane Shaper for Controlled Liposome Deformation

Shape defines the structure and function of cellular membranes. In cell division, the cell membrane deforms into a “dumbbell” shape, while organelles such as the autophagosome exhibit “stomatocyte” shapes. Bottom-up in vitro reconstitution of protein machineries that stabilize or resolve the membrane necks in such deformed liposome structures is of considerable interest to characterize their function. Here we develop a DNA-nanotechnology-based approach that we call the synthetic membrane shaper (SMS), where cholesterol-linked DNA structures attach to the liposome membrane to reproducibly generate high yields of stomatocytes and dumbbells. In silico simulations confirm the shape-stabilizing role of the SMS. We show that the SMS is fully compatible with protein reconstitution by assembling bacterial divisome proteins (DynaminA, FtsZ:ZipA) at the catenoidal neck of these membrane structures. The SMS approach provides a general tool for studying protein binding to complex membrane geometries that will greatly benefit synthetic cell research.


77
(ii) (iii) group with respect to the head group of chol moiety showed a significant decrease of the first peak for the 103 oligo-chol system ( Figure S3a), whereas such a decrease was not observed for the terminal beads ( Figure   104 S3b) -which indicated that the came from an expansion in the head region.

106
Based on these results we can make a rough estimate of the curvature induced by a single oligo-chol Δ

107
compared to the one induced by chol, see Figure S4. In this system, for Δ = -~2ℎ / ≫ ℎ 108 (or small curvature), since and . Therefore, the curvature of = ( + ℎ) 2 sin 2 ( 2 ) = 2 sin 2 ( 2 ) 109 the midplane will be As we do not know the exact value of , and therefore, we obtain with respect to a reference point which Δ 112 is the curvature of the chol molecule. The RDF of the tail bead suggest that for cholesterol is equal to 113 for chol-oligo. Therefore To systemically investigate if chol-oligo can lead to curvature generation, we introduce three membrane 117 curvature identifiers; positive curvature index (PCI), negative curvature index (NCI), and total curvature 118 (TC: in nm -1 ), which is the sum of the two. The PCI measures an average of the number of instances that a 119 molecule resides on the part of the membrane with positive curvature. PCI=1 indicates that the molecule 120 has no preference for positive or negative curvature while a value larger than one indicates that the molecule 121 prefers positive curvature. NCI is the same measure but for negative curvature. TC is the average value of 122 the membrane curvature where the molecule is located.
Where is the number of molecules located on membrane regions with positive/negative ( / ) 129 mean curvature, is the area of the membrane that has positive/negative mean curvature, 130 is the total membrane area, and is the total number of the molecules. Figure S5 shows the time evolution 131 of these quantities. It is clear that chol-oligo is associated with positive curvature.

133
To further explore this association, we performed simulations of POPC membranes containing different  (Table S1).

140
Overall, we envision two important roles for the oligo moiety of the construct for the generation of 141 membrane curvature; i) making a large head group for an amphiphilic chol-oligo construct, leading to a 142 wedging effect, ii) being a highly charged and polar group and preventing the construct from flipflopping.

143
These lead the chol-oligo bound monolayer to prefer a bent shape.  Table 1 "10 chol-oligo/flat" row. 164 Buckling amplifies the molecules curvature behaviors as the membrane can exhibits large curvature. All 165 three measures of curvature indicate that chol-oligo are associated with positive mean curvature.  Table 1 "10 170 chol" row. All three measures of curvature indicate that chol induces negative curvature and therefore the 171 positive curvature of chol-oligo does not come from the chol moiety of the construct (see Figure SI6). .
(2) = 212 213 Combining (1) and (2)  We model bulk diffusion across the neck of a dumbbell as follows: 277 a) Two vesicles are connected by one channel of length L p and diameter . 278 b) The first vesicle (V 1 ) starts with a much lower amount of fluorescent molecules, due to photobleaching 279 step, compared to the second vesicle (V 2 ), namely , where and are the 1, ≪ 2, 1, 2, 280 initial concentrations of fluorescent molecules present in V 1 and V 2 , respectively. 281 c) The total number of fluorescent molecules present in the dumbbell system (after photobleaching) 282 is constant over time.

284
We define the flux which represents the net flux of molecules going from V 2 to V 1 . According to the first 285 Fick's law: 286 287 , where is the concentration gradient, and is the diffusion constant of the molecule. By definition of 290 flux, namely number of molecules per unit of surface per unit of time , we have that 291

301
, (11) = - where is the influx rate, i.e., the number of molecules with diffusion constant crossing a pore of length 304 and area ( pore radius) per unit of time . We can write the molecule concentration in = 2 =

305
V 1 as , where is the volume of V1. We also note that the total number of molecules 1 ( ) = 1 ( ) 1 1 306 in V 1 and V 2 is constant over time. We end up with the following system of equations: 307
We can substitute the expression for in (12) and rewrite as a function of only, finding 2 ( ) .
To minimize the number of fit parameters, we express (15)

348
(16) to the normalized intensity data . From the fit, a pore diameter of 8.6 nm was estimated.