Dynamic Fingering in Adhered Lipid Membranes - Langmuir (ACS

Jan 24, 2018 - Department of Cell Systems & Anatomy, University of Texas Health Science Center San Antonio, San Antonio, Texas 78229, United States. â...
1 downloads 7 Views 3MB Size
Subscriber access provided by UNIV OF DURHAM

Article

Dynamic Fingering in Adhered Lipid Membranes Orrin Shindell, Natalie Mica, Kwan Kelvin Cheng, Exing Wang, and Vernita Gordon Langmuir, Just Accepted Manuscript • DOI: 10.1021/acs.langmuir.7b03708 • Publication Date (Web): 24 Jan 2018 Downloaded from http://pubs.acs.org on February 4, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Langmuir is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Dynamic Fingering in Adhered Lipid Membranes Orrin Shindell,∗,†,‡ Natalie Mica,†,§ Kwan H. Cheng,‡ Exing Wang,¶ and Vernita D. Gordon∗,† †Center for Nonlinear Dynamics and Department of Physics, University of Texas at Austin ‡Department of Physics and Astronomy, Trinity University, San Antonio, Texas ¶Department of Cell Systems & Anatomy, University of Texas Health Sciences Center San Antonio, San Antonio, Texas §School of Physics and Astronomy, University of St. Andrews, Saint Andrews, Scotland, United Kingdom E-mail: [email protected]; [email protected]

Abstract Artificial lipid membranes incorporating proteins have frequently been used as models for the dynamic organization of biological structures in living cells as well as in the development of biology-inspired technologies. We report here on the experimental demonstration and characterization of a pattern-forming process that occurs in a lipid bilayer membrane adhered via biotin-avidin binding to a second lipid membrane that is supported by a solid substrate. Adhesion regions are roughly circular with about 25 µm diameter. Using confocal fluorescence microscopy, we record time-series of dynamic fingering patterns that grow in the upper lipid membrane and inter-membrane biotin-avidin bonds. The fingers are µm-scale elongated pores that grow from the edge of an already-stabilized hole. Finger growth is saltatory on the scale of tens of seconds. We find that as the fingers grow and the density of adhesion proteins increases, the rate of finger growth decreases exponentially and the width of newly-formed fingers

1

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

decreases linearly. We show that these findings are consistent with a thermodynamic description of dynamic pore formation and stabilization.

Introduction Dynamic interactions between proteins and lipid membranes are ubiquitous in living systems 1,2 and are significant in biology-inspired technologies. 3,4 Living cells organize the components of their plasma membranes into transient and persistent functional structures, frequently associated with sites of cellular adhesion. 4–8 For example, the T-cell, upon adhesion, forms an immunological synapse which is dynamic on the scale of several minutes and then static for tens of minutes. 9 Understanding the physics leading to these dynamics and other membrane processes, such as dynamics associated with pinning of the plasma membrane to the cytoskeleton, has been the goal of many theoretical and experimental studies. 10–23 Lipid membranes are used in bio-inspired technologies for the encapsulation and controlled release of drugs and can be functionalized, like their living counterparts, to recognize and adhere to specific targets. 5 However, almost all applications of lipid membranes use them in their fluid phase, which is fragile under tension and easily lyses if a pore forms in the membrane. 24 Other work has shown that coating lipid membranes with a high density of proteins can change their physical state from fluid to solid-like. 25,26 Such a state is spatially and temporally static but does not alter the membrane’s fragility under tension. 26 A similar effect is observed in the adhesion region of vesicles that are adhered via biotin-avidin binding to solid supported membranes, where solid-like complexes of proteins can coexist with fluid proteins. 23,27 Here we demonstrate that protein-mediated adhesion of one lipid bilayer to a solidsupported lipid bilayer (SLB) via biotin-avidin binding can support the growth of pores that, in turn, cause the adhesion-mediating proteins to transition from a fluid state to a solid-like state that can sustain long-lived pores. When we adhere bilayers to SLBs, finger-

2

ACS Paragon Plus Environment

Page 2 of 27

Page 3 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

shaped pores grow in a saltatory fashion, invading the protein-bound membrane and, in so doing, compressing the binding proteins. After a time, the proteins are compressed into a solid state and the process arrests. What remains is a network of finger-shaped holes in the adhered membrane; this network remains stable for a day or longer.

Materials and Methods Materials The lipids used in this study were purchased from Avanti Polar Lipids: 1,2-dioleoyl-snglycero-3-phosphocholine (DOPC); Cholesterol; 1,2-dioleoyl-sn-glycero-3-phosphoethanolamineN-(cap biotinyl) (sodium salt) (DOPE-biotin).

R Whatman Polycarbonate Membranes,

19mm in diameter with 0.4µm and 0.03µm pore size, were also purchased from Avanti Polar Lipids and used to make small unilamellar vesicles for forming SLBs. From ThermoFisher, we R bought binding protein and two membrane dyes: avidin in the form of NeutrAvidin Biotin-

binding Protein (neutravidin); the amphiphilic membrane dye 1,1’-Dioctadecyl-3,3,3’,3’Tetramethylindocarbocyanine Perchlorate (DiI); and the lipophilic membrane dye 6-Dodecanoyl2-Dimethylaminonaphthalene (Laurdan). From Sigma Aldrich, we bought Atto 488 NHS ester (Atto-488), some reagents used for labeling neutravidin with Atto-488 - potassium chloride, sodium bicarbonate, and anhydrous dimethyl sulfoxide (DMSO) - and gaskets used to construct imaging chambers, Grace Bio-Labs Press-To-Seal silicone isolators (8-9 mm in diameter, 2.0 mm in depth). From Fisher Scientific, we bought other reagents - chloroform and sucrose, Dulbecco’s Phosphate Buffered Saline 1x without calcium chloride and without magnesium chloride (PBS), and Centri-Spin 20 Princeton columns used to separate aggregates during the fluorescent labeling process. Ultrapure 18.2 MΩ·cm water (DI Water) was obtained using a MilliQ Millipore system. SLBs were formed on Menzel-Glaser 24x32 mm #1 cover glasses.

3

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Instruments Confocal fluorescence microscopy and confocal reflection microscopy were performed using an Olympus IX81 inverted microscope with an Olympus UPlanFlN 100x oil immersion objective. Laurdan two-photon imaging was performed using an Ultima multiphoton microscopy system (Prairie Technologies, Inc.) with a Nikon Eclipse FN-1 upright microscope. We used a 780nm laser to excite the Laurdan via two-photon absorption. Two-photon images were taken with a Nikon 25X/NA 1.1 water immersion objective; 445/45 and 494/41 nm band pass emission filters and a 470 nm dichroic mirror (Semrock) were used. The stoichometry of Atto-488 bound to biotin was measured using a NanoDrop 2000, from Thermo Fisher. All experiments were conducted at room temperature, about 23◦ C.

Fluorescent labeling of neutravidin The neutravidin was fluorescently-labeled with Atto-488 using the Lys light method. 28 Briefly, Atto-488 was dissolved in DMSO at 10 mM and added to 140 mM neutravidin dissolved in labeling buffer. Labeling buffer was 150 mM sodium bicarbonate and 20 mM potassium chloride. The solution of Atto-488 and neutravidin was allowed to react at room temperature for 30 min. Aggregates were removed using a separation column. We ensured that the stoichometry of Atto-488 to neutravidin was 1:1 Atto-488:neutravidin by measureing the absorbance at 280 nm using the NanoDrop 2000. This stoichimetry was desirable because we found that commercially available fluorescently-labeled neutravidin, which has several fluorophores conjugated to each neutravidin molecule, slowed the dynamics of GUV adhesion and prevented fingering from occurring.

Preparation of membranes Giant unilamellar vesicles (GUVs) and supported lipid bilayers (SLBs) were prepared as previously described. 29 GUVs composed of 86.5% DOPC, 10% Cholesterol, 3% DOPE-biotin,

4

ACS Paragon Plus Environment

Page 4 of 27

Page 5 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

and 0.5% DiI were formed via electroformation 30 in 280 mM sucrose solution, which matched the osmolarity of PBS. SLBs were formed via vesicle fusion on cover glasses cleaned using piranha etching. 21 The GUVs were placed into a sample chamber constructed using a gasket sealed to a cover glass. On the cover glass there was an SLB composed of 97% DOPC and 2% DOPE-biotin and the sample chamber was filled with PBS. The SLB was functionalized by binding fluorescently-labeled neutravidin to the DOPE-biotin in the SLB. The molecular details of this experiment are discussed at length in Fenz et al. 31

Membrane adhesion and poration GUVs were used to form flat patches of bilayer membrane adhered to an SLB via biotinavidin binding; these flat patches became platforms for the formation of finger patterns. When GUVs where added to a sample chamber, they settled to the bottom of the chamber and adhered to the SLB, as shown in Figure 1 (left). When the GUVs adhered they were tensed, and sometimes finger-shaped pores would grow in the adhered region of an intact GUV. However, when fingers eventually penetrated the rim of the adhesion zone, the fingering pattern would collapse. Therefore, to facilitate studying fingering in adhered membranes without the confounding effect of the free membrane, we deliberately induced GUVs to rupture, thereby eliminating any remaining free portion of the membrane while leaving the flat portion of the membrane that was adhered to the SLB. We did this by waiting for the GUVs to fully adhere and then adding DI water on the top of a 0.4 µm polycarbonate membrane placed on the top of the gasket. DI water thus seeped into the sample chamber. This increased the osmotic pressure across intact GUV membranes, resulting in increased tension. The increased tension caused the GUVs to lyse, leaving behind the membrane that was adhered to the SLB by the biotin-avidin bonds, as shown in Figure 1 (right) and Figure 2C. When vesicles rupture, loosely represented by the center cartoon on the right side of Figure 1, some fraction of the membrane is pulled onto the surface of the SLB by biotin5

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 1: Cartoon of experimental setup. Samples are formed in gaskets placed on a microscope cover glass. On the left side of the cartoon, an initially flaccid GUV is adhered to an SLB. The GUV contains 280 mM sucrose solution and the surrounding solution is 280 mM PBS. On the right side, DI water is slowly added to the surrounding solution, which reduces the osmolarity of the external solution and causes water to flow into the GUV. First the GUV tenses and then the GUV ruptures. After the GUV ruptures, the remaining portion of the GUV is attached to the SLB. The molecular details of this system are detailed nicely by other authors. 31

6

ACS Paragon Plus Environment

Page 6 of 27

Page 7 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

avidin binding as seen in Figures 2C and 2F, and some fraction the membrane is lost to the surrounding fluid, as has been observed in rupture experiments in free-floating GUVs. 32 We regard the portion of the GUV remnant inside the ring of immobile biotin-avidin bonds, evident in 2C and represented by the right cartoon on the right side of Figure 1, as the experimental platform for our study of dynamic fingering. In Figure 2A, the areas under the bulging membrane (the bulges are shown in Figure 2D) are darker in the Atto-488 neutravidin channel than the area outside the adhered GUV, indicating they are depleted of both mobile and surface-aggregated neutravidin. We offer two speculations about why this may be: (1) The biotinylated GUV may have made contact with the supported lipid bilayer under the bulges during the initial adhesion process and depleted both fluid and aggregated neutravidin in the dark regions in Figure 2A. (2) The supported lipid bilayer under the bulges in the adhered GUV may have been depleted of neutravidin as a result of being surrounded by the biotin-contain GUV membrane, which acts as a sink for the neutravidin originally contained by the small amount of supportedbilayer membrane under the bulges. The area of supported bilayer outside the region of GUV adhesion is orders of magnitude larger than the area under the small bulges and therefore, while this area also loses some neutravidin to binding to the adhered GUV membrane, the percentage loss and the effective depletion is less, and therefore the external region appears brighter in the Atto-488 Neutravidin channel. The non-circular shape of the boundaries of the bulges seen in Figure 2A likely result from pinning effects, which we expect also to shape the fingers that are studied as the focus of this work. A brief discussion of the role of pinning in the structure of the patterns formed in this system, which are discussed in the Results and Discussion, is found in the Conclusion. However, the bulges themselves likely do not play a significant role in the pattern-forming (fingering) process. In Figure 2E, before the membrane has ruptured, the bulges are gone and the neutravidin concentration shown in Figure 2B appears uniform inside the adhering region. It appears, therefore, that the system in the final ruptured state (Figures 2C and

7

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Flaccid

A

Page 8 of 27

Increasing Tension

Ruptured

B

C

E

F

Atto-488 Neutravidin

D

DiI Membrane

Scale Bars = 10μm

Figure 2: (A-C) Confocal fluorescence micrographs of fluorescently-labeled neutravidin adhering a GUV to an SLB. Each image is a single confocal slice. In (A), an initially-flaccid GUV is adhered to the surface. The black spots in the adhesion zone are portions of the GUV that are buckled out-of-plane and therefore not imaged by the confocal imaging plane (see (D) and Supplementary Figure 1). DI water adjusts the osmolarity of the surrounding solution and water slowly seeps into the GUV and in (B), after 500s, the GUV has tensed and the adhesion zone has contracted. Finally, after 800s, in (C), the GUV ruptures, leaving behind the adhered remnant. (D-F) Confocal micrographs of DiI incorporated into the membrane. In (D) a median intensity z-projection of seven confocal slices (0.34 µm between slices) shows a bright rim of membrane immediately surrounding the black regions in (A). The appearance of the bright rim results from the fluorescent intensity contrast between the nearly-vertical edges of out-of-plane membrane bulges and the flat portion of the proteinadhered membrane; Supplementary Figure 1 shows the same confocal micrograph, with a vertical slice through two membrane bulges. In (E) and (F) single confocal slices show the configuration of the membrane corresponding to images (B) and (C) respectively.

8

ACS Paragon Plus Environment

Page 9 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

2F) has no memory of the out-of-plane buckling present in the initial state (Figures 2A and 2C).

Results and Discussion Finger-shaped heterogeneities grow in adhered regions After adhesion, we found that dark finger-shaped structures grew along the surface of the adhering membrane in a saltatory fashion, Figure 3 and Supplementary Movie 1. Fingers excluded both the membrane dye DiI and the Atto-488-labeled neutravidin proteins, Figure 4. As the fingers grew, they displaced proteins laterally into the remaining adhered membrane. This caused the protein-bound portion of the adhered region to become denser in biotin-avidin bonds, Figure 4A-D. While fingers were growing, local protein density and DiI intensity varied within the membrane. When fingering arrested, the protein density was the same in every intact membrane region whereas the DiI intensity was different between isolated regions. From this we conclude that the final density of the proteins, rather than the final lipid composition of the membrane, is responsible for the extinction of the fingering process. To characterize the mobility of the lipids in the fingering membrane and the biotinavidin bonds adhering the membrane to the SLB, we performed fluorecence recovery after photobleaching (FRAP) measurements. This was done by bleaching an ≈ 3µm diameter region of the sample, measuring the fractional fluorescence recovery, and fitting the data to a theoretical recovery curve, 33 Figure 5A). Measurements of the diffusion constants of DiI in the membrane in a before-fingering adhesion state, as in Figure 2D, and in a final finger pattern, as in Figure 4H, had the same value of 0.06 ± 0.01 µm2 s−1 , top line of Figure 5A. The diffusion constant of the Atto-488-labeled neutravidin binding the membranes together in a before-fingering adhesion state as in Figure 2A was 0.005 ± 0.010 µm2 s−1 , bottom line of Figure 5A. The bound neutravidin in the final finger state did not recover; Figure 5B shows a 9

ACS Paragon Plus Environment

Langmuir

A

B

18 16 14

Finger Area ( µm2 )

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 27

12 10 8 6 4 2 0

0

100

200

Time (s)

300

400

500

Figure 3: Saltatory growth of a fingering pattern in an adhered membrane. (A) Finger area versus time plot of a characteristic fingering region shows periods of slow growth followed by abrupt jumps in area. The insets are the finger shapes that correspond to the periods of slow growth. (B) The boundaries of the inset shapes from (A) are superimposed over a confocal micrograph of an evolved finger pattern. The fluorescent signal is from the amphiphilic membrane dye DiI incorporated in the adhered GUV, which is clearly absent in the finger pattern. bleached region of Atto-488-labeled neutravidin 5 minutes after bleaching that didn’t recover. From this we conclude that the biotin-avidin bonds are stationary (i.e. “frozen”) in the final finger pattern.

Diagnosing fingers to be holes In previous work we used the same approach (biotin-avidin binding) to adhere GUVs to SLBs; unlike the present case, wherein DOPC-cholesterol binary membranes are far from any demixing phase boundary, our earlier work used ternary membranes deliberately tuned to be close to a demixing transition. 29,34 In that work, we observed that adhesion induced the formation of heterogeneities that excluded both proteins and membrane dye; we thought these were lipid-ordered-phase domains (See Reference Note 41). The lipid composition of the GUV membranes we use in the present work should be miscible at all temperatures 34 hence, we do not expect the formation of an ordered lipid phase upon adhesion. Therefore, to diagnose the nature of these fingers, we used two separate approaches: confocal fluorescence microscopy combined with confocal reflection microscopy, 35 and two10

ACS Paragon Plus Environment

Page 11 of 27

Time = 0s

Time = 100s

A

Time = 200s

B

C

Time = 445s 1.0

D

0.8

0.6

Atto-488 Neutravidin

0.4

0.2

0.0

E

F

G

1.0

H

0.8

0.6

DiI Membrane

0.4

0.2

0.0

Scale Bars = 10μm

Figure 4: (A-D) Confocal fluorescence micrographs; false-colored fluorescent intensity of fluorescently-labeled neutravidin. (E-H) Confocal fluorescence micrographs; false-colored fluorescent intensity of the membrane dye DiI. (A,E) An initial finger-shaped tear in the membrane. Intermediate fingering states at 100s later, (B,F), and at 200s later, (C,G). Growing fingers partition some regions off from the larger adhered-membrane area until the density of neutravidin in segregated regions increases to a limiting value. (D) Fingering growth arrests at 445 s and all regions have reached the same final density, shown as dark red, of biotin-avidin bonds. (H) Finger growth has arrested and the DiI signal is different in different regions, which indicates variation in the DiI composition. The microscopy movie from which (A-D) were taken is Supplementary Movie 2 and from which (E-H) were taken is Supplementary Movie 3.

A Fractional Recovery

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

B

1

DiI in Membrane 0.5

Atto-488-Labeled Neutravidin

0 0

2

4

6

8

10

12

Time/Area ( s μm-2 )

Figure 5: (A) Fractional recovery curves of DiI in adhered membranes before fingering (pluses, top line) and after fingering (circles, top line) and of Atto-488-labeled neutravidin before fingering (crosses, bottom line). The data are fitted to a theoretical recovery function. 33 (B) Confocal fluorescence micrograph of fluorescently-labeled neutravidin; a bleached region in a final static finger pattern fails to recover after 5 minutes. 11

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 27

photon fluorescence microscopy. 36 Both types of experiment indicated that the fingers are holes in the GUV membrane, as follows.

Reflection evidence that fingers are holes

Binders

A

Membrane

B

Reflection

C

Scale Bars = 5μm Figure 6: Confocal fluorescence and reflection micrographs of an adhered membrane region in which one finger-like heterogeneity has formed. (A) Fluorescently-labeled neutravidin is excluded from the heterogeneity. (B) The membrane dye DiI is excluded from the heterogeneity. (C) The heterogeneity does not reflect the confocal laser. For confocal reflection microscopy in conjunction with confocal fluorescence microscopy, the reflection signal obtained from the fingers was indistinguishable from the reflection signal obtained from the SLB outside the adhered region, Figure 6. This is consistent with fingers being regions that do not contain any membrane from the GUV. Had the fingers been an ordered phase rather than a hole, they should have reflected the light from the incident laser and thus had a signal that was distinct from the SLB outside the adhered region of the GUV. From this we infer the fingers are likely holes rather than an ordered lipid phase.

Fluorescence evidence that fingers are holes As a second test, we used two-photon microscopy of GUVs composed of 33.5% DOPC, 33.5% DPPC, 30% Cholesterol, and 3% DOPE-biotin. Membranes of this composition exhibit coexisting liquid-ordered and liquid-disordered phases at room temperature. 34 These adhered

12

ACS Paragon Plus Environment

Page 13 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

GUVs also formed fingers in the adhered region. Thus, we could compare phase separation in the free portion of the membrane with fingers formed in the adhered portion of the membrane. We treated these adhered vesicles with 2 µM Laurdan. Laurdan has an environmentallysensitive emission spectrum, with a major peak near 440 nm when the dye is in a gel phase and a shifted peak near 490 nm when the dye is in a fluid-disordered phase. 36 When Laurdan is in a fluid-ordered phase the peak is still shifted, but less so than for the fluid-ordered phase; the exact amount of shift is determined by the composition of the fluid-disordered phase.

445 nm

494 nm

GP

A

B

C

D

E

F

Scale Bars = 5μm Figure 7: (A,B,D,E) Two-photon fluorescence micrographs. (C,F) Fluorescence intensity values have been used to calculate the spatially-resolved generalized polarization (GP), where red indicates GP > 0 and blue indicates GP < 0. (A, B) The Laurdan signal from the non-adhered portion of the GUV yields (C) a generalized polarization that indicates the coexistence of ordered and disordered lipid phases. (D, E) The Laurdan signal from the adhered portion of the membrane, containing fingers, yields (F) a generalized polarization that is constant regardless of whether the location is a finger. This indicates the fingers are holes and not ordered-phase domains.

13

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Laurdan emission spectrum was collected simultaneously through 445 nm and 495 nm channels. Using the intensity in the 445 nm channel, I445 , and the intensity in the 495 nm channel, I495 , we calculated the generalized polarization GP = (I445 − I495 ) / (I445 + I495 ). 36 If Laurdan is in an ordered phase, then GP > 0, but if Laurdan is in a disordered phase, then GP < 0. We found that in the phase-separated, free portion of the membrane, the GP confirmed the co-existence of ordered and disordered lipid phases, Figures 7A-C. In the adhered portion of the GUV, fingers were clearly visible, Figure 7D and 7E. However, the GP measure indicated that there was no ordered phase present in the adhered region, Figure 7F. From this, we conclude the fingers are holes rather than an ordered lipid phase. Thus, to our knowledge, our experiment is the first reported case of dynamic holes forming at the interface between protein-adhered lipid membranes.

Analysis and Modeling To probe the interrelated dynamics of finger growth and increasing biotin-avidin protein density, and how these link to long-lived, stable pores that do not lead to membrane lysis, we tracked distinct membrane regions backward in time. Our analysis is founded on three points. First, the terminal density of proteins in the adhered membrane was the same in all intact parts of the membrane, as in Figure 4D. Second, the density of proteins in the fingers was zero. Third, we assumed that the total number of binding proteins in the adhering region was conserved during the fingering process, which is justified because the binding energy of the biotin-neutravidin pair is at least 10 kT in adhered membranes (k is Boltzmann’s constant and T is the absolute temperature in Kelvin). 27 Thus, we could track the decrease in protein density of each region of the membrane backward in time from the final density. The final density was common throughout the adhered membrane. We identified 23 intact regions of Figure 4D that remained when fingering arrested. For each identified region, the density of proteins increased over time. This agrees with our finding that the size of the intact regions decreases with time, up to the point that local finger 14

ACS Paragon Plus Environment

Page 14 of 27

Page 15 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

growth arrests. Furthermore, at earlier times the intact regions contained some biotin-avidin bonds that ultimately formed the immobile boundaries between fingers. Those boundaries had a protein density equal to the final density of the stationary membrane regions rather than the lesser densities internal to the intact regions at intermediate timepoints. Taking all this into account, we reconstructed the density of each region, at each measured timepoint, normalized to the final density (i.e. the final density was assigned a value of unity). We model the dynamics of finger growth by treating the formation of fingers as intermittentlyformed, tension-induced semicircular pores that form at the interface between an existing finger and the protein-bound membrane. In pore-forming experiments with GUVs suspended in fluid, one of two typical outcomes occurs: either a temporarily stable pore forms and allows fluid to escape, thus reducing tension in the GUV and allowing the pore to close; 32,37,38 or else the pore grows indefinitely and the vesicle is annihilated. 39 When membranes are adhered the situation is more complicated. For example, G¨ozen et al., 40 found that multilammelar membranes, bound via generic surface forces to cover glass, could exhibit saltatory growth as a double bilayer membrane was unrolled onto the surface into a single bilayer membrane. 41 Therefore, we posit the following model, diagrammed in Fig. 8, to describe fingering in adhered membranes: Saltatory finger growth occurs at the interface between a hole in the membrane and a region of intact membrane, which is adhered by proteins with a density ρ1 and surrounded by a rim of immobile protein binders, Figure 8 (left). After a time ∆t1 , there is free energy fluctuation ∆F1 that results in the formation of a pore of width λ1 and a binding protein density ρ2 > ρ1 , Figure 8 (center). After another time ∆t2 , another free energy fluctuation ∆F2 occurs and pore forms resulting in a static finger of width λ2 , Figure 8 (right). The saltatory nature of the growing pattern is represented by the series of finite time intervals between each growing pore, i.e., ∆t1 , ∆t2 , · · · . As a proxy for the rate at which fingers grew, we measured the number of pores that formed in a given region with a given

15

ACS Paragon Plus Environment

Langmuir

A Δt1 Immobile Proteins

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ρ1

ΔF1

Page 16 of 27

Δt2

ΔF2

ρ2 Membrane

λ1

λ2

Hole

B

Scale Bar = 5 μm Figure 8: A model for dynamic fingering by saltatory poration. In (A), the left side imagines a membrane patch adhered by proteins with density ρ1 bordered on three sides by immobile proteins and on one side by a hole. After a time ∆t1 a free energy fluctuation ∆F1 causes a transition to the center frame where a semicircular pore of diameter λ1 has formed and the protein density has increased to ρ2 . After another time ∆t2 another free energy fluctuation ∆F2 causes a transition to the right frame where another semicircular pore of diameter λ2 has formed. In (B), a confocal micrograph shows a characteristic fingering pattern that results from the process depicted in (A); the fluorescence signal is from Atto-488-labeled neutravidin.

16

ACS Paragon Plus Environment

Page 17 of 27

boundary length in a given time interval. The result of this analysis was that the rate of finger growth decreased monotonically with the density of binding proteins in the region being invaded, Figure 9. We also found that the width of the fingers decreased monotonically with the local density of proteins in the region being invaded, Figure 10.

0.04

Rate of Finger Formation ( μm-1 s-1 )

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

0.03

0.02

0.01

0

0

0.1

0.2

0.3

0.4

0.5

0.6

1 - Protein Density (dimensionless) Figure 9: Finger formation rate versus protein density. 172 measurements of finger for√ mation rates were binned by density into 172 ≈ 13 groups each with 13 or more data points. The error bars in protein density are the half width of the bins and the error bars in finger formation are the standard error of the mean finger formation rate in each bin. The dashed line is an exponential fit, based on our model, given by Rate of Finger Growth = (0.0035µm−1 s−1 ) exp [3.8 (1 − Protein Density)]. The thermodynamic state of a semicircular pore of radius r ≥ 0 formed at a straight boundary is specified by a free energy function which we write as, 32   r2 π F(r) = − σ 1 − 2 r2 + (π − 2) γr 2 2r0

(1)

where σ is the surface tension jump across the boundary of the membrane, γ is the line 17

ACS Paragon Plus Environment

Langmuir

1.2 1.0

Finger Width ( μm )

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 27

0.8 0.6 0.4 0.2 0

0

0.1

0.2

0.3

0.4

0.5

0.6

1 - Protein Density (dimensionless) Figure 10: Finger width versus√protein density. 524 measurements of finger widths were binned by protein density into 524 ≈ 22 groups each with 23 or more data points. The error bars in density are the half width of the bins and errors in the finger widths are the standard errors of the mean finger width in each bin. The dashed line is a linear fit, based on our model, given by Finger Width = 0.27µm + 1.0µm (1 − Protein Density).

18

ACS Paragon Plus Environment

Page 19 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

tension along the pore boundary, and r0 is the radius of a stable pore in the absence of a restoring force due to γ. In this model, the formation of a pore decreases the surface tension. For another pore to form, the surface tension in the membrane must again increase. We suggest this occurs because the material in the bilayer, i.e., DOPC, cholesterol, and DiI, is lost into the surrounding solution. 32 As can be seen by comparing Figures 4E and 4H, only about half of the original adhering material remains when the fingering process reaches the final static state. The pore free energy, Equation 1, exhibits a maximum of F ∗ ≡ max F, and a minimum at r = λ/2, with λ identified as the width of an arrested finger. We can justify that the energy terms in Equation 1 are sufficient to account for the formation of pores by inserting some typical values for the lysis tension and line tension. If we estimate σ ≈ 10 mN m−1 , 42 γ ≈ 10 pN, 43 and let r0 ≈ 0.5 µm (F ∗ is insensitive to r0 ), then we have F ∗ ∼ kT where k is Boltzmann’s constant and T (≈ 296K) is the absolute temperature. To compare the model with the experiment, we assume F ∗ and λ can be expanded as Taylor functions about ρ = 1, which is justified near ρ = 1. Thus, λ ≈ λ0 + λ1 (1 − ρ) and a fit to the data yields λ0 = 0.27 ± 0.04 µm and λ1 = 1.0 ± 0.1 µm, Figure 10. For a pore to form, a free energy fluctuation at least as large as F ∗ ≈ F0∗ − F1∗ (1 − ρ) must occur at the interface between a hole and the protein-bound membrane. For a given boundary length L, time interval ∆t, and density ρ, the number of fingers N that form per unit length per unit time will follow an Arrhenius rate law. Thus, from our data we fit N ∗ = AeF1 (1−ρ) L∆t

(2)

The results of this fit are F1∗ = 3.8 ± 0.8 kT, and A = 0.0035 ± 0.0010 µm−1 s−1 , Figure 9. The finding that the activation energy F ∗ is on the order of the thermal energy kT when experimental data is fit with the model and when typical values for edge tension and lysis tension are used in the model, underscores that our model and underlying assumptions are

19

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

physically reasonable.

Conclusion We have reported on the experimental demonstration of a novel dynamic fingering process in adhered lipid membranes. Here we focused on fingering in membranes composed of a binary lipid mixture of 10% Cholesterol, 87.5% DOPC, and additives. However, we have also shown that similar phenomena occur in ternary lipid membranes, both those above their demixing transition temperature (42% cholesterol, 27.5% DOPC, 27.5% DPPC, and additives; see Panels K-O in Figure S3 of our previous publication) 29 and those below (30% Cholesterol, 33.5% DOPC, 33.5% DPPC, and additives; see Figure 7 and associated discussion in Section “Fluorescence evidence that fingers are holes”). Thus, we expect the dynamic fingering in lipid membranes adhered via biotin-avidin binding we study here or similar patterns to occur in many different lipid mixtures. The process studied in our system resembles fingering in other systems, e.g. viscous fingering in Hele-Shaw cells 44 and the avalanche ruptures in lipid membranes. 40 However, the fingering that occurs in our membranes differs from these examples in important ways: viscous fingering is a continuous growth process whereas ours is saltatory; avalanche ruptures in fluid membranes have saltatory growth but don’t compress laterally-mobile adhesion proteins. In our system the fingering process compresses laterally-mobile binding proteins from a fluid state into a solid-like state. The irregular shapes of the fracture patterns obtained in fluid membranes in G¨ozen et al. 40 were presumably due to generic pinning forces between the two bilayers; in our system, it is likely that localized regions of immobile proteins are responsible for pinning points along the irregular boundaries of the finger pattern, while the bulk of the binding proteins are still fluid. As the binding proteins in our system are compressed by the saltatory growth of pores, the dynamics and the geometry of the fingering pattern are altered: as the binding protein density increases, the finger width decreases linearly and the

20

ACS Paragon Plus Environment

Page 20 of 27

Page 21 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

rate of finger formation decreases exponentially. Both of these observations are consistent with a thermodynamic model of pores forming at the interface between previously formed fingers and the adhered membrane.

Acknowledgement This work was funded by start-up funds from The University of Texas at Austin (UT Austin) to VDG. OS was supported by a Downer-Focht summer fellowship from the Department of Physics at UT Austin. Two-photon images were generated in the Core Optical Imaging Facility which is supported by UTHSCSA and NIH-NCI P30 CA54174 (CTRC at UTHSCSA). The authors thank Wilton Snead and Professor Jeanne Stochowiak (Biomedical Engineering, UT Austin) for helpful conversations and their technical support helping us work with fluorescently-labeled proteins.

Supporting Information Available Supporting Information Additional images showing bulges and microscopy time-series showing fingering in both DiI and Atto-488 channels (EPS, AVI, and TIF) This material is available free of charge via the Internet at http://pubs.acs.org/.

Notes and References (1) Simons, K.; Ikonen, E. Functional rafts in cell membranes. nature 1997, 387, 569. (2) McMahon, H. T.; Boucrot, E. Molecular mechanism and physiological functions of clathrin-mediated endocytosis. Nature reviews Molecular cell biology 2011, 12, 517– 533. 21

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(3) Puri, A.; Loomis, K.; Smith, B.; Lee, J.-H.; Yavlovich, A.; Heldman, E.; Blumenthal, R. Lipid-based nanoparticles as pharmaceutical drug carriers: from concepts to clinic. Critical Reviews in Therapeutic Drug Carrier Systems 2009, 26 . (4) Liu, Q.; Boyd, B. J. Liposomes in biosensors. Analyst 2013, 138, 391–409. (5) Gordon, V. D.; O’Halloran, T.; Shindell, O. Membrane adhesion and the formation of heterogeneities: biology, biophysics, and biotechnology. Physical Chemistry Chemical Physics 2015, 17, 15522–15533. (6) Carquin, M.; DAuria, L.; Pollet, H.; Bongarzone, E. R.; Tyteca, D. Recent progress on lipid lateral heterogeneity in plasma membranes: From rafts to submicrometric domains. Progress in lipid research 2016, 62, 1–24. (7) Kroll, A. V.; Fang, R. H.; Zhang, L. Biointerfacing and Applications of Cell MembraneCoated Nanoparticles. Bioconjugate chemistry 2016, 28, 23–32. (8) Barile, L.; Vassalli, G. Exosomes: Therapy delivery tools and biomarkers of diseases. Pharmacology & therapeutics 2017, (9) Grakoui, A.; Bromley, S. K.; Sumen, C.; Davis, M. M.; Shaw, A. S.; Allen, P. M.; Dustin, M. L. The immunological synapse: a molecular machine controlling T cell activation. Science 1999, 285, 221–227. (10) Qi, S.; Groves, J. T.; Chakraborty, A. K. Synaptic pattern formation during cellular recognition. Proceedings of the National Academy of Sciences 2001, 98, 6548–6553. (11) Raychaudhuri, S.; Chakraborty, A. K.; Kardar, M. Effective membrane model of the immunological synapse. Physical review letters 2003, 91, 208101. (12) Lin, L. C.-L.; Brown, F. L. Dynamics of pinned membranes with application to protein diffusion on the surface of red blood cells. Biophysical journal 2004, 86, 764–780.

22

ACS Paragon Plus Environment

Page 22 of 27

Page 23 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

(13) Weikl, T. R.; Lipowsky, R. Pattern formation during T-cell adhesion. Biophysical journal 2004, 87, 3665–3678. (14) Ma˜ nes, S.; Viola, A. Lipid rafts in lymphocyte activation and migration. Molecular membrane biology 2006, 23, 59–69. (15) Sheetz, M. P.; Sable, J. E.; D¨obereiner, H.-G. Continuous membrane-cytoskeleton adhesion requires continuous accommodation to lipid and cytoskeleton dynamics. Annu. Rev. Biophys. Biomol. Struct. 2006, 35, 417–434. (16) Balasubramanian, N.; Scott, D. W.; Castle, J. D.; Casanova, J. E.; Schwartz, M. A. Arf6 and microtubules in adhesion-dependent trafficking of lipid rafts. Nature cell biology 2007, 9, 1381–1391. (17) Viola, A.; Gupta, N. Tether and trap: regulation of membrane-raft dynamics by actinbinding proteins. Nature Reviews Immunology 2007, 7, 889–896. (18) Lajoie, P.; Goetz, J. G.; Dennis, J. W.; Nabi, I. R. Lattices, rafts, and scaffolds: domain regulation of receptor signaling at the plasma membrane. The Journal of cell biology 2009, 185, 381–385. (19) Ehrig, J.; Petrov, E. P.; Schwille, P. Near-critical fluctuations and cytoskeleton-assisted phase separation lead to subdiffusion in cell membranes. Biophysical journal 2011, 100, 80–89. (20) Machta, B. B.; Papanikolaou, S.; Sethna, J. P.; Veatch, S. L. Minimal model of plasma membrane heterogeneity requires coupling cortical actin to criticality. Biophysical journal 2011, 100, 1668–1677. (21) Nair, P. M.; Salaita, K.; Petit, R. S.; Groves, J. T. Using patterned supported lipid membranes to investigate the role of receptor organization in intercellular signaling. Nature protocols 2011, 6, 523–539. 23

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(22) Yu, Y.; Fay, N. C.; Smoligovets, A. A.; Wu, H.-J.; Groves, J. T. Myosin IIA modulates T cell receptor transport and CasL phosphorylation during early immunological synapse formation. PloS one 2012, 7, e30704. (23) Schmidt, D.; Bihr, T.; Fenz, S.; Merkel, R.; Seifert, U.; Sengupta, K.; Smith, A.-S. Crowding of receptors induces ring-like adhesions in model membranes. Biochimica et Biophysica Acta (BBA)-Molecular Cell Research 2015, 1853, 2984–2991. (24) Needham, D.; Nunn, R. S. Elastic deformation and failure of lipid bilayer membranes containing cholesterol. Biophysical journal 1990, 58, 997–1009. (25) Ratanabanangkoon, P.; Gropper, M.; Merkel, R.; Sackmann, E.; Gast, A. P. Twodimensional streptavidin crystals on giant lipid bilayer vesicles. Langmuir 2002, 18, 4270–4276. (26) Ratanabanangkoon, P.; Gropper, M.; Merkel, R.; Sackmann, E.; Gast, A. P. Mechanics of streptavidin-coated giant lipid bilayer vesicles: A micropipet study. Langmuir 2003, 19, 1054–1062. (27) Fenz, S. F.; Smith, A.-S.; Merkel, R.; Sengupta, K. Inter-membrane adhesion mediated by mobile linkers: effect of receptor shortage. Soft Matter 2011, 7, 952–962. (28) Modesti, M. Fluorescent labeling of proteins. Single Molecule Analysis: Methods and Protocols 2011, 101–120. (29) Shindell, O.; Mica, N.; Ritzer, M.; Gordon, V. D. Specific adhesion of membranes simultaneously supports dual heterogeneities in lipids and proteins. Physical Chemistry Chemical Physics 2015, 17, 15598–15607. (30) Angelova, M. I.; Dimitrov, D. S. Liposome electroformation. Faraday discussions of the Chemical Society 1986, 81, 303–311.

24

ACS Paragon Plus Environment

Page 24 of 27

Page 25 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

(31) Fenz, S. F.; Merkel, R.; Sengupta, K. Diffusion and intermembrane distance: case study of avidin and E-cadherin mediated adhesion. Langmuir 2008, 25, 1074–1085. (32) Sandre, O.; Moreaux, L.; Brochard-Wyart, F. Dynamics of transient pores in stretched vesicles. Proceedings of the National Academy of Sciences 1999, 96, 10591–10596. (33) Soumpasis, D. Theoretical analysis of fluorescence photobleaching recovery experiments. Biophysical journal 1983, 41, 95–97. (34) Veatch, S. L.; Keller, S. L. Separation of liquid phases in giant vesicles of ternary mixtures of phospholipids and cholesterol. Biophysical journal 2003, 85, 3074–3083. (35) Paddock, S. Confocal reflection microscopy: the” other” confocal mode. Biotechniques 2002, 32, 274–278. (36) Parasassi, T.; De Stasio, G.; Ravagnan, G.; Rusch, R.; Gratton, E. Quantitation of lipid phases in phospholipid vesicles by the generalized polarization of Laurdan fluorescence. Biophysical journal 1991, 60, 179–189. (37) Zhelev, D. V.; Needham, D. Tension-stabilized pores in giant vesicles: determination of pore size and pore line tension. Biochimica et Biophysica Acta (BBA)-Biomembranes 1993, 1147, 89–104. (38) Chabanon, M.; Ho, J. C.; Liedberg, B.; Parikh, A. N.; Rangamani, P. Pulsatile lipid vesicles under osmotic stress. Biophysical Journal 2017, 112, 1682–1691. (39) Levadny, V.; Tsuboi, T.-a.; Belaya, M.; Yamazaki, M. Rate constant of tension-induced pore formation in lipid membranes. Langmuir 2013, 29, 3848–3852. (40) G¨ozen, I.; Dommersnes, P.; Czolkos, I.; Jesorka, A.; Lobovkina, T.; Orwar, O. Fractal avalanche ruptures in biological membranes. Nature materials 2010, 9, 908–912. (41) In G¨ozen et al., the authors discuss two modes of rupture growth. One is the intermittent (saltatory) mode already mentioned and the other is continuous floral-patterned growth. 25

ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

In our previous work 29 we suggested that floral patterns that appeared in the adhesion zone of GUVs bound to an SLB via biotin-avidin binding were ordered phase lipid domains that formed during adhesion-induced lipid phase separation. Based on the findings of our current study and those of G¨ozen et al., we suspect we may have been mistaken. The domains, which were dark to fluorescence microscopy, which we took to indicate ordered lipid phase domains, were likely floral-patterned pores. (42) Evans, E.; Heinrich, V.; Ludwig, F.; Rawicz, W. Dynamic tension spectroscopy and strength of biomembranes. Biophysical journal 2003, 85, 2342–2350. (43) Karal, M. A. S.; Yamazaki, M. Communication: Activation energy of tension-induced pore formation in lipid membranes. 2015. (44) Homsy, G. M. Viscous fingering in porous media. Annual review of fluid mechanics 1987, 19, 271–311.

26

ACS Paragon Plus Environment

Page 26 of 27

Page 27 of 27

Graphical TOC Entry Finger area

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Saltatory Growth

Time

27

ACS Paragon Plus Environment