Rotational dynamics of bound pairs of bacteria-induced membrane tubes

Abstract

We present experiments demonstrating tube formation in giant unilamellar vesicles that are suspended in a bath of swimming E. coli bacteria. We chose the lipids such that the bacteria form no adhering interactions with the membrane. The tubes are generated by the pushing force exerted by the bacteria on the membrane of the vesicles. Once a tube is generated, the bacterium is confined within it, resulting in long-lived tubes that protrude into the vesicle. We show that such tubes interact to form stable bound pairs that orbit each other. We speculate that the tubes are maintained by the persistent pushing force generated by the bacteria, and the rotating pairs are stabilized by a combination of curvature-mediated interaction and vorticity generated in the membrane by the rotation of the flagella.

---

Introduction

Several recent studies have shown that active particles exhibit diverse interactions with soft boundaries. For example, active colloidal particles show persistent orbital motion around giant unilamellar vesicles (GUVs).^1,2 When active particles are confined within a vesicle, they also induce significant shape deformations, such as tethers and dendritic structures.^3,4 Active particles interacting with a GUV can transfer both forces and torques, leading to GUV translation and rotation.¹ When encapsulated within a GUV, motile bacteria can deform the membrane by extruding tubes that, by coupling with their flagella, propel the entire vesicle.^4,5 Generally, the pressure exerted by active particles on curved walls is inhomogeneous and can lead to driven shape instabilities on flexible walls. However, at least in two-dimensions, an equation of state can be recovered for the average normal force.⁶ It is also known that asymmetric boundaries can rectify particle motion, generating currents and shear stresses, while pressure variations can destabilize flexible walls, leading to phenomena such as filament bending and self-propulsion.^5,6

Tubular structures emerge as a unifying theme in the strategies employed by diverse intracellular bacterial pathogens to manipulate the host cell processes for their benefit.^7–10Listeria monocytogenes exploits the host’s exocytic machinery to generate membrane-bound protrusions that resemble tubular extensions, facilitating its spread between epithelial cells.^7,8 Similarly, Salmonella enterica remodels the host’s membraneous organs to generate interconnected tubular membranes, including double-membrane Salmonella-induced filaments, which function in nutrient acquisition, immune evasion, and maintenance of the Salmonella-containing vacuoles.¹¹ As other examples, obligate intracellular bacteria like Chlamydia and Rickettsia leverage the host’s actin cytoskeleton to generate actin-rich protrusions or tunnels that promote cell invasion and intercellular dissemination.¹⁰ These studies collectively illustrate how pathogens co-opt hosts’ membraneous structures or form tubular networks as essential tools for intracellular survival, replication, and spread, highlighting the central role of tube-like structures in microbial pathogenesis. Thus, understanding the interaction between motile entities and soft boundaries is important both from a physical and biological point of view.

Our focus here is on membrane deformations induced by an active agent and the resulting interaction between such deformations. We have performed experiments where GUVs are suspended in a dilute bath of E. coli that are genetically modified to show persistent swimming. Lipids are chosen to deliberately avoid any adhering interaction between the bacterium and the membrane. We show that single E. coli can generate membrane tubes in floppy GUVs and become engulfed in them to form long-lived tubes. These tubes interact when they are near each other to form stable orbiting pairs. We analyse the structure of these tubes and the dynamics of orbiting pairs. We also provide some speculative arguments for the stability of bound pairs.

Materials and methods

Bacteria culture

We use a genetically modified laboratory strain of E. coli bacteria known as RP5232 (donated by Judith Armitage, University of Oxford).¹² This strain has a deletion of a cheY gene (ΔcheY) that makes the flagella rotate only counter-clock-wise, making them persistent swimmers.¹³ The bacteria are cultured using Tryptone Broth (TB) (cat. no. M463-500G, Himedia). The preparation of the bacteria starts with suspending 10 μl of bacterial stock in 10 ml of TB. This suspension is maintained in a shaker-incubator at 37 °C at 80 rpm. The culture is allowed to grow until the optical density, as measured via the OD600 method, reaches about 0.6. The optical density is measured using a UV-Visible spectrometer (NanoDrop 2000c, Thermo Scientific). The bacteria are then pelleted in a centrifuge and resuspended in a 56 mM glucose solution to be used for further experimentation.

GUV preparation

We prepare giant unilamellar vesicles (GUVs) using 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) lipids and 0.2 mole-percent 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindotri-carbocyanine iodide DiIC₁₈ dye, following a standard electroformation protocol.¹⁴ A signal of 1 V_rms at 25 Hz is applied for 3 hours for the preparation of GUVs, at the end of which a signal of 0.3 V_rms at 3 Hz is applied to detach the vesicles from the ITO plates. The GUVs are prepared in a 50 mM of sucrose solution to make them heavier than the surrounding medium. The GUVs then settle down on the bottom surface of the sample chamber, making observation using an inverted microscope easier. After preparation, the GUVs are added to the bacterial suspension and observed using video microscopy. The observation chamber consists of a bottom coverglass and a top quartz plate with the two plates separated by an aluminum spacer with air vents (grooves in the plate) to support the aerobic respiration mode of E. coli. To prevent vesicles from adhering, the coverglass is coated with Bovine Serum Albumin.

Video microscopy

Time-lapse video microscopy of GUVs in the presence of suspended bacteria is performed using an inverted microscope (Olympus IX70) equipped with a CCD camera (CoolSnap-EZ, Photometrics). Recordings are made at 7 frames-per-second either in fluorescence or phase-contrast mode using a 60× objective.

Observations

Giant unilamellar vesicles (GUVs) were maintained in an external bacterial bath consisting of genetically modified E. coli (RP5232 ¹²), which are persistent swimmers exhibiting interesting shape dynamics. We use a neutral lipid (POPC) to investigate the vesicle deformations arising out of forces and torques exerted by the bacteria. The swimming forces exerted by the bacteria on the membrane surface result in the formation of membrane tubes or invaginations that protrude into the vesicle. An example of tubes formed by such interaction, imaged using fluorescence microscopy, is shown in Fig. 1a (also see Videos 1 and 2 in the SI), and an enlarged view of two such tubes is shown in Fig. 1b. The length of the tube can vary from tube to tube. The base of the tubes can drift along the envelope of the vesicle, but the tubes always remain nearly normal to the envelop (pointed towards the GUV center). Phase contrast images show that each of these tubes engulfs at least one bacterium, as shown in Fig. 1c. A schematic of a vesicle with bacteria-containing tubes is shown in Fig. 1d. Usually, the number of bacteria-containing tubes is higher for floppy vesicles as compared to relatively less floppy ones. This is particularly clear in cases where a given vesicle transforms from a floppy state to a tense state via extrusion of membrane, as can be seen in Video 3 in the SI. Fig. 2 shows the distributions of lengths and diameters of tubes obtained using multiple GUV-bacteria preparations. It can be seen that the typical tube length is about 4.5 μm and the diameter peaks at around 1.4 μm.

Fig. 1 (a) An example of tube formation in a membrane vesicle suspended in a bath of E. coli. (b) An enlarged view of a couple of tubes. (c) Comparison of fluorescence (above) and phase-contrast (below) images show the presence of a bacterium inside the tube. Note that the two images were taken one after the other and the tube position has shifted during this period. (d) A schematic showing the deformation of the vesicle by bacteria (not to scale). Scale bars are 10 μm.

Fig. 2 Probability distributions of the tube diameter and length. A total number of 67 tubes were analyzed to construct these plots.

When there are extraneous objects stuck to the exterior of the vesicle, close to the base of a tube, such objects are seen to move around the tube axis. This suggests that there is circular hydrodynamic flow either within the lipid bilayer or in the external fluid, presumably generated by the rotation of the bacterial flagella. One such example is shown in Fig. 3 and in Video 4 in the SI.

Fig. 3 Time series of a vesicle bud moving around the base of a tube generated by a bacterium. This set of figures shows a complete cycle of the bud around the base.

In some cases, multiple bacteria enter a tube and the tube can get longer or even acquire branched geometries. In such cases, the tube appears slightly pinched between the two bacteria, as shown in Fig. 4. Pinching is also seen towards the base of long tubes containing a single bacterium, as in Fig. 1b.

Fig. 4 Example images of nearly straight tubes containing more than one bacterium (left two images) and ones with branches (right two images). Note that the tubes show narrow necks in the region between bacteria in such cases. The scale bars are 5 μm.

Remarkable dynamics are observed when two membrane protrusions come near each other. In such cases, the tubes form a bound pair and continuously orbit each other (see Fig. 5). Orbiting tubes appear to be tilted away from the normal to the vesicle envelope, in the direction of their motion. The tubes have been seen to remain bound for observation times of up to a minute or more. The distribution of orbital periods for such pairs is shown in Fig. 6a and that for the tilt angles is shown in Fig. 6b. Orbiting pairs are seen to remain bound while the central axis drifts around the envelope of the vesicle (see Video 5 in SI). The handedness of the orbit can be determined by observing pairs with their axis of orbit oriented normal to the plane of focus. This is particularly convenient when observing tubes near the bottom surface of the vesicle. Both senses of rotation could be observed for bound pairs.

Fig. 5 (a) Image sequence showing two tubes that orbit each other as seen nearly side on. The sequence shows almost half an orbit. Scale bars are 5 μm. (b) Image sequence showing a pair of orbiting tubes as seen almost along the axis of the orbit for almost half an orbit. Scale bars are 5 μm. (c) A schematic of the orbiting tubes (not to scale). The red arrow indicates the axis about which the tubes orbit.

Fig. 6 (a) Probability distribution of orbital frequencies of bound pairs of membrane tubes. A total of more than 50 pairs were measured. (b) Probability distribution of orientation of tubes within bound pairs, measured as the obtuse angle made by each tube with respect to the equatorial tangent to the GUV.

Discussion

The experiments mentioned in the previous sections can be summarized as follows. Bacteria exert a force on the GUV membrane that creates invaginations that trap them. This results in long-lived membrane tubes that protrude into the vesicle, and the tubes “diffuse” around the surface of the vesicle. When two tubes come close to each other, they form bound pairs that orbit each other.

Our experiments provide direct evidence for the absence of stable adhering interactions between the POPC membrane and the bacteria. The only persistent association we observed occurs when bacteria become trapped within the tubes they themselves generate. It is also known from earlier studies that E. coli adheres poorly to pure POPC membrane.¹⁵

E. coli swim by spinning their flagella using rotary motors that are located at the base of each flagellum.^16,17 When the motors spin in the counter-clockwise direction, as viewed from outside (from the tip from the flagellum towards the cell body), all the flagella form a coherent helical bundle, which spins to propel the cell body. The cell body itself will rotate with the opposite sense. Even for a given number of flagella, the drag force may vary, as the cell body has a broad size distribution.¹⁸ When the motors spin in the clockwise direction, the flagella become incoherent and splay out, and the bacterium performs a tumbling motion. The genetically modified strain we use (RP5232) predominantly consists of swimmers, which rarely tumble. The force generated by a swimming E. coli is about a pico-Newton and several parameters, like force, torque, rotational frequency, and efficiency, have been estimated.^17–19

It is known that a point force applied to a lipid bilayer membrane can result in the formation of a cylindrical membrane tube.²⁰ With a membrane bending modulus κ and in-plane tension σ, the force needed to maintain a tube can be estimated as , and its equilibrium tube diameter as .²⁰ In our experiments, we observe that only floppy vesicles (low σ) form bacteria-induced tubes and tubes disappear when vesicles become tense. Therefore, in low-tension vesicles, tube formation can be induced easily by the swimming force of bacteria, which is about 1 pN, acting against the membrane. The length of the tube depends on the available excess area, and may vary depending on how many tubes are formed.

The lengths of the tubes we observe fall within the range of 4–8 μm, which is significantly longer than the length of the cell body of a bacterium (length ranges from 2–4 μm and diameter is about 1 μm).¹⁸ This suggests that both the cell body and the flagella are within the tubes we observe, at least in most of the cases. The membrane invaginations show varying morphologies. When the tubes are on the shorter side of the distribution (a few microns), the tube broadens towards the base and has a wide neck (Fig. 1b). However, longer tubes show a narrowing of their necks (Fig. 1b). This narrowing is also observed near the midsections of very long tubes that carry two bacteria (Fig. 4). This indicates that the diameter of the thicker part of the tube is restricted by the steric hindrance caused by the bacterium.

One of the most interesting observations is that a pair of tubes can come together and form a bound pair that orbit each other. This implies a short-range repulsion and a long-range attraction between tubes. The observation of lipid/fluid flow near the base of the tube (Fig. 3) suggests that there is a vorticity at the base of the tube, possibly generated by the rotation of the bacterial flagella. This may lead to an interaction between the two tubes that results in the orbital motion. An additional curvature-mediated interaction can also play a role. Although elastic forces are known to cause the repulsion of inclusions in a membrane, experiments and simulations have shown that in a vesicle, with finite curvature, the interaction between inclusions can be repulsive or attractive depending on whether the inclusions are adsorbed inside or outside the vesicle.^21,22 Simulation and experimental studies have also been performed on the adsorption of passive spherocylinders on vesicles and the resulting shape changes.^23,24 However, bacteria are not passive objects and, in the case reported here, they do not adhere to the membrane. The tube is most certainly the result of a pushing force exerted by a bacterium on the membrane. It would be interesting to study the effects of spherocylinders that swim along their symmetry axis, with or without associated torque generation. It is known that tethers being pulled from a membrane with external forces results in barrier-free attraction with a force proportional to the product of the forces and inversely proportional to the distance between the tethers.²⁰ It is possible that this force, compounded by the repulsion between vortical flows induced by the bacteria, is responsible for the bound-pair formation. A detailed study to measure the various forces responsible for the stable bound pairs is underway.

Our investigation highlights the importance of physical interactions – forces, torques, and hydrodynamics – in bacteria–membrane interactions. The observation of orbiting pairs of membrane tubes demonstrates the non-trivial effects of active forces and torques in membrane-mediated interactions between active chiral inclusions. Further experiments and theoretical studies are required to elucidate these mechanisms, which could be very relevant to understanding how pathogens invade host cells.

Conflicts of interest

There are no conflicts of interest to declare.

Data availability

Supplementary information (SI) is available: the raw videos for the data presented in the paper are included as Videos 1–5. See DOI: https://doi.org/10.1039/d5sm00701a.

Acknowledgements

The authors acknowledge Hareesh Kumar and Divyang Trivedi for their help in conducting the experiments. We thank V. A. Raghunathan and K. Vijay Kumar for discussions, and the former also for help with designing experiments.

References

1. V. Sharma, E. Azar, A. P. Schroder, C. M. Marques and A. Stocco, Active colloids orbiting giant vesicles, Soft Matter, 2021, 170(16), 4275–4281.
2. O. Vincent, A. Sreekumari, M. Gopalakrishnan, V. V. Vasisht and B. Ranjan Sarangi, Curvature-dependent dynamics of a bacterium confined in a giant unilamellar vesicle, Phys. Rev. E, 2025, 1120(2), 024408.
3. H. Rao Vutukuri, M. Hoore, C. Abaurrea-Velasco, L. van Buren, A. Dutto, T. Auth, D. A. Fedosov, G. Gompper and J. Vermant, Active particles induce large shape deformations in giant lipid vesicles, Nature, 2020, 5860(7827), 52–56.
4. S. C. Takatori and A. Sahu, Active contact forces drive nonequilibrium fluctuations in membrane vesicles, Phys. Rev. Lett., 2020, 1240(15), 158102.
5. L. Le Nagard, A. T. Brown, A. Dawson, V. A. Martinez, W. C. K. Poon and M. Staykova, Encapsulated bacteria deform lipid vesicles into flagellated swimmers, Proc. Natl. Acad. Sci. U. S. A., 2022, 1190(34), e2206096119.
6. N. Nikola, A. P. Solon, Y. Kafri, M. Kardar, J. Tailleur and R. Voituriez, Active particles with soft and curved walls: Equation of state, ratchets, and instabilities, Phys. Rev. Lett., 2016, 1170(9), 098001.
7. R. Jennifer, Robbins, Angela I Barth, Hélène Marquis, Eugenio L de Hostos, W James Nelson, and Julie A Theriot. Listeria monocytogenes exploits normal host cell processes to spread from cell to cell, J. Cell Biol., 1999, 1460(6), 1333–1350.
8. G. C. Dowd, R. Mortuza, M. Bhalla, H. Van Ngo, Y. Li, L. A. Rigano and K. Ireton, Listeria monocytogenes exploits host exocytosis to promote cell-to-cell spread, Proc. Natl. Acad. Sci. U. S. A., 2020, 1170(7), 3789–3796.
9. P. M. Colonne, C. G. Winchell and D. E. Voth, Hijacking host cell highways: manipulation of the host actin cytoskeleton by obligate intracellular bacterial pathogens, Front. Cell. Infect. Microbiol., 2016, 6, 107.
10. E. Martinez, F. Ayenoue Siadous and M. Bonazzi, Tiny architects: biogenesis of intracellular replicative niches by bacterial pathogens, FEMS Microbiol. Rev., 2018, 420(4), 425–447.
11. V. Liss and M. Hensel, Take the tube: remodelling of the endosomal system by intracellular s almonella enterica, Cell. Microbiol., 2015, 170(5), 639–647.
12. R. Barak and M. Eisenbach, Chemotactic-like response of escherichia coli cells lacking the known chemotaxis machinery but containing overexpressed chey, Mol. Microbiol., 1999, 310(4), 1125–1137.
13. J. S. Parkinson and S. E. Houts, Isolation and behavior of escherichia coli deletion mutants lacking chemotaxis functions, J. Bacteriol., 1982, 1510(1), 106–113.
14. Z. Boban, I. Mardešić, W. Karol Subczynski and M. Raguz, Giant unilamellar vesicle electroformation: What to use, what to avoid, and how to quantify the results, Membranes, 2021, 110(11), 860.
15. S. Dai, X. Tang, N. Zhang, H. Li, C. He, Y. Han and Y. Wang, Lipid giant vesicles engulf living bacteria triggered by minor enhancement in membrane fluidity, Nano Lett., 2023, 230(1), 371–379.
16. M. Siva Kumar and P. Philominathan, The physics of flagellar motion of e. coli during chemotaxis, Biophys. Rev., 2010, 20(1), 13–20.
17. Y. Sowa and R. M. Berry, Bacterial flagellar motor, Q. Rev. Biophys., 2008, 410(2), 103–132.
18. S. Chattopadhyay, R. Moldovan, C. Yeung and X. L. Wu, Swimming efficiency of bacterium escherichia coli, Proc. Natl. Acad. Sci. U. S. A., 2006, 1030(37), 13712–13717.
19. D. J. Armstrong, T. A. Nieminen, A. B. Stilgoe, A. V. Kashchuk, I. C. D. Lenton and H. Rubinsztein-Dunlop, Swimming force and behavior of optically trapped micro-organisms, Optica, 2020, 70(8), 989–994.
20. I. Derényi, F. Jülicher and J. Prost, Formation and interaction of membrane tubes, Phys. Rev. Lett., 2002, 880(23), 238101.
21. C. Van Der Wel, A. Vahid and A. Šarić, Timon Idema, Doris Heinrich, and Daniela J Kraft. Lipid membrane-mediated attraction between curvature inducing objects, Sci. Rep., 2016, 60(1), 32825.
22. A. H. Bahrami and T. R. Weikl, Curvature-mediated assembly of janus nanoparticles on membrane vesicles, Nano Lett., 2018, 18, 1259–1263.
23. A. Sharma, Y. Zhu, E. J. Spangler, J.-M. Y. Carrillo and M. Laradji, Membrane-mediated dimerization of spherocylindrical nanoparticles, Soft Matter, 2023, 190(8), 1499–1512.
24. S. Van der Ham, J. Agudo-Canalejo and H. Rao Vutukuri, Role of shape in particle-lipid membrane interactions: from surfing to full engulfment, ACS Nano, 2024, 180(15), 10407–10416.

---