Protocells are objects that mimic one or several functions of biological cells and may be embodied as solid particles, lipid vesicles, or droplets. Our work is based on using decanol droplets in an aqueous solution of sodium decanoate in the presence of salt. A decanol droplet under such conditions bears many qualitative similarities with living cells, such as the ability to move chemotactically, divide and fuse, or change its shape. This article focuses on the description of a shape-changing process induced by the evaporation of water from the decanoate solution. Under these conditions, the droplets perform complex shape changes, whereby the originally round decanol droplets grow into branching patterns and mimic the growth of appendages in bacteria or axon growth of neuronal cells. We report two outcomes: (i) the morphological changes are reversible, and (ii) multiple protocells avoid contact between each other during the morphological transformation. The importance of these morphological changes in the context of artificial life are discussed.
A living cell is the basic structural, functional, and biological unit of all known living organisms. The ability to maintain homeostasis, to self-reproduce, and to evolve are three characteristics that may define a natural living cell . In contrast, artificial (or synthetic) cells are man-made systems that mimic only some of the properties, functionalities, and processes of natural cells. However, the successful preparation of fully functional synthetic cells having the same features as their natural counterparts—in particular the ability to self-produce and to maintain themselves (the so-called autopoietic systems)—remains elusive . In fact, the creation of fully functional artificial cell-like structures that could self-reproduce and evolve might even pose a number of safety and ethical questions .
A simpler and less problematic task is to construct protocells that are not alive, but exhibit some lifelike properties. Protocells can be defined as simplified systems that mimic one or several (but not all) of the morphological and functional characteristics of biological cells. Their structure and organization are usually relatively simple and can be based on a different chemistry than any known living system. Protocells are useful tools in artificial life research because they represent “cells as they could be,” while biology focuses on “cells as we know them.” Protocells can also serve as model systems for research into the origin of life, where they can represent hypothetical precursors of the first natural living cells .
The material bases of protocells can be rather diverse. Protocells can be embodied as solid particles [9, 10], coacervates , polymersomes , hydrogel microcapsules , lipid vesicles , or droplets [13, 21]. The principal research focus using protocells may be the membrane , growth and division [24, 35], motility , evolution , or the ability of protocells to communicate with their living counterparts [18, 19]. Although gel- and fluid-based protocells in principle enable shape morphing, this aspect has scarcely been studied in the literature. One notable exception is the growth of elongated protocells, where the shape was found to play an essential role in the fission dynamics . However, the majority of published research assumes a spheroidal shape of protocells, and there has not been much focus on the shape or the morphogenesis of protocells . Living cells exhibit a plethora of intriguing shapes, and the ability to change shape can be considered an important manifestation of life. Therefore, in the present article we will describe a relatively simple chemical system that exhibits lifelike shape changes. We will discuss the observed shape changes and their relevance to living cells.
Recently it has been reported that simple droplets can morph from a spherical shape to various polygonal flat shapes when cooled . Here we focus on shape-shifting, multi-arm droplets observable in a simple system consisting only of decanol, water, sodium decanoate, and NaCl on a glass substrate. The lifelike chemotactic movement of decanol droplets in salt gradients has been shown on short time scales (a few minutes) , mimicking chemotactic behavior of living cells . Recently, we observed an interesting shape change of the same decanol droplet system over long time scales (several hours) . The evaporation of water from the decanoate solution in this open system induces fantastic shape changes in the droplets: The originally spheroidal decanol droplets develop branching patterns and mimic the appendage growth of bacteria or axon growth (Figure 1). The present article focuses on the discussion of a phenomenological similarity between shape-changing living cells and droplet protocells. Specifically, our aim is to report two new lifelike phenomena, namely (i) reversible morphological changes and (ii) how the protocells avoid contact between each other.
Decanoic acid, 1-decanol, oil red O, sodium hydroxide, and sodium chloride were obtained from Sigma Aldrich. All chemicals were used without further purification; water was purified by a Millipore Milli-Q system. Sodium decanoate solution was prepared by dissolving decanoic acid at 10 mM in water using 5 M NaOH to adjust the pH of the resulting solution to typically 12–13. Sodium chloride solution was prepared at saturation with final concentration 6.5 M. The decanol droplets were colored by Oil red O (approximately 2 mg/ml) for better visibility.
2.2 Experimental Procedure
The majority of experiments described in the present article were performed as follows (see Figure 2). Aqueous 10 mM decanoate solution was spread over a round glass coverslip (with diameter 10 or 18 mm). A decanol droplet was placed with a micropipette on the decanoate layer, and then the salt solution was added to the decanoate solution. The volumes of decanoate, decanol, and salt solutions were varied as described below. The experiments were performed under laboratory conditions (temperature around 23°C and evaporation by natural convection).
2.3 Observation Techniques
The pattern formation of decanol droplets could be clearly seen with the naked eye; however, the experiments were monitored using an ImagingSource video camera (DFK 23U274) from the top view. For microscopic observation, an Olympus CK40 microscope with phase contrast was used.
In the present article we focus on the tentacular pattern formation of evolving decanol droplets, and specifically on two previously unreported lifelike features: (i) reversible morphological changes, viz., the retraction of previously formed tentacles, and (ii) avoidance of growing tentacles in the presence of multiple droplets.
Figure 3 shows an example of the arm growth observed on both microscopic and macroscopic scales during the first few minutes of the experiment with a single droplet. The whole time series of the experiment can be viewed in Supplementary S1 Video (see Appendix 2). The initial conditions were as follows: A glass coverslip with diameter 10 mm was covered by a thin layer of 77.2 μl of 10 mM aqueous decanoate solution; then a 0.6-μl decanol droplet and 0.5 μl of 6.5 M sodium chloride solution were added. Initially, the decanol droplet has a round shape. Small spikes then begin to appear on the droplet's periphery; they bulge and also retract. A few minutes after, several spikes dominate over the others and start to prolong and grow. The spherical droplet transforms into a star shape. Then some spikes extend and form long tentacles that occasionally branch. As the tentacles elongate further, structural gaps appear in the elongations. The tentacles then can disintegrate, and their fragments shrink, coalesce, and rearrange into small droplets. At the end of the experiment, that is, after the complete evaporation of water from the decanoate solution (less than 2 h for the initial decanoate solution volume of 77 μl), the glass slide substrate is covered by several residues of gelled decanol and an unstructured dry decanol-decanoate film.
Since the shape changes described above were driven by the process of water evaporation from the system, in the next step we perturbed the system back to its initial condition by adding the appropriate quantity of water. Figure 4 shows a microscopic detail of decanol droplet behavior during the period of evaporation (time t = 0 to 1 h) and then the return into spherical shape during the rehydration phase. This experiment confirms that the process of pattern formation is reversible over at least one cycle and that it is possible to control this process by the timing of evaporation and hydration.
The second set of experiments focused on mutual interactions of two or more droplets during the growth of tentacles. Two particular outcomes were possible: The arms attract each other and intertwine, or the arms from different droplets avoid mutual contact. Figure 5a shows the progress of a typical experiment with two droplets. Initially, the droplets come together to form a doublet. Later on they repel each other and start to form tips that elongate into tentacles. It was observed that the tentacles from one droplet were never attracted by the second droplet. The droplets always avoided contact. In multiple droplet experiments the symmetrical star-shaped beginning of pattern formation was not observed. This contrasts with single-droplet experiments, in which the growth of tentacles was usually more or less symmetrical and there was no preferential direction of tentacle growth (see Figure 5b).
It is of interest to explore the role of shape in the context of artificial life. While most researchers use spherical droplets or vesicles as model protocells, phylogenetic studies indicate that the early bacteria on planet Earth were rod-shaped and cocci appear to possess a “dead-end” shape that has arisen many times independently . This indicates that non-spherical droplets could serve well as protocells, and it is important to focus on their properties and the process of shape change .
Living cells such as bacteria exhibit an immense variety of shapes, and even eukaryotic cells within multicellular organisms show an enormous diversity in size, shape, and internal organization. Bergey's Manual of Determinative Bacteriology  focuses on the morphological diversity of bacterial shapes and sizes. A recent article by Kysela et al.  nicely sums up the terms that can be used for the description of bacterial shapes:
In addition to the familiar coccoid, rod-shaped, or spirillar types, there are also dendroid, coryneform, cylindrical, bulbiform, fusiform, and vibrioid types. There are uniseriate or multiseriate filaments of cells that are flexible or rigid, flat or round, unbound or bound in hyaline or slime sheaths. Single cells are described as star-shaped, disk-shaped, hourglass-shaped, lemon-shaped, pear-shaped, crescent-shaped, or flask-shaped. Rods can be pleomorphic, straight, curved, or bent, with blunt, pointed, rounded, or tapered ends. Some cells grow appendages such as prosthecae, stalks, or spikes. (Kysela et al.  Introduction)
There is a relationship between shape and movement. For example, highly motile bacteria usually assume rodlike morphology, while movement through highly viscous fluids such as mucus seems to favor spiral shape. Non-spherical shape also helps a bacterium to avoid predation. For example, the strategy to avoid phagocytosis and facilitate enhanced attachment to host cell surfaces used by a number of different bacterial and fungal pathogens is to increase their cell size and overall surface area . The locomotion and transformation abilities of protocells in the forms of molecular aggregates (e.g., oil droplets, liquid crystalline droplets and tubes, giant vesicles) are discussed in the recent review of Toyota et al. .
An increase in the surface area also improves the rate of diffusion and nutrient uptake. A perfectly spherical shape of the cell minimizes the ratio of surface area to volume. Each change in the shape (elongation, flattening, growth of protrusions, etc.) leads to an increase of this ratio. Some cells can adapt to changes in the environmental conditions by adjusting their shape, for example by appendage formation in nutrient-poor conditions . By shape deformation or tentacle elongation the cells can reach multiple places or hardly accessible areas.
As was already shown in Figure 1, the tentacular structures of decanol droplets (c) qualitatively resemble the appendages of freshwater bacteria Ancalomicrobium (a), so let us compare the shape changes of decanol droplets with the appendage formation of this prosthecate bacteria. Under high nutrient concentrations, these cells have a rodlike or spherical shape. However, in nutrient-poor conditions they form appendages (cell envelope extensions) called prosthecae. Prosthecae allow a greater surface area with which to take up nutrients (and release metabolites) . This example of how living cells can modify their morphology in response to environmental changes shows that shape is an important parameter that affects the survival of cells in different conditions. Moreover, Ancalomicrobium is a unicellular bacterium, and its cells usually occur singly or in pairs prior to division, rarely forming aggregates . As one can see from Figure 5a, decanol droplets with appendages also prefer to stay independent, and they do not form aggregates.
Another qualitative comparison of decanol droplets is with neuronal cells. The extensive shape changes observed in our droplet system, with bifurcation and branching patterns, are at least superficially similar to biological neuronal architectures. However, in this case it is (even) harder to establish a link with the biological domain, because the electrical conductivity of the surrounding bulk electrolyte is now much higher (3.77 mS/cm) than the conductivity of decanol (nonconducting), so this is the opposite situation to what one finds in neuronal networks . In order to simulate neuronal connections and some signal propagation, we would need the tentacles to conduct electricity better than the surrounding solution. In addition, when two or more droplets are added to the same system and begin to change shape, the growing tentacles avoid each other. This results in droplets that do not touch each other and do not form an extensive interconnected network although they are individually extended in space (see Figure 5a and Supplementary Video S2 ( Appendix 2)). With further development using chemoattractants and interfacial tension gradients, it may be possible to use such a system for the propagation of signal using “artificial neurons.”
Artificial cells in the form of organic droplets can serve as chemical or liquid robots . They can be used, for example, for the delivery of molecules or small objects. Modular self-reconfiguring robotic systems are able to deliberately change their own shape by rearranging the connectivity of their parts, in order to adapt to new circumstances, perform new tasks, or recover from damage. If we suppose that our decanol-decanoate-salt system is a liquid robotic system, we can also imagine a self-reconfiguring system that can react according to the changes in the environment such as water activity (drying-rehydration cycles). As an example, when the water evaporates from the system and the droplets extend to form multi-armed structures, the addition of water returns the droplet to its original spheroid shape (Figure 4).
We are investigating how a normally spheroidal droplet can morph into non-spherical shapes and under what conditions. In our previous work, we have found a parameter space that can reliably produce droplets that proceed from spheroidal to highly extended tentacular structures . Although this shape change requires an open system with evaporation to produce such dynamics, the system is highly reproducible. Here we have shown that the process is reversible and the tentacular structures can be returned to a round shape simply by adding water to the system. In addition we observe that multiple droplets in the same environment are able to morph into extended structures. But since the extended appendages do not touch, this results in the morphology of one droplet influencing the morphology of neighboring droplets. It would be interesting to explore the effects of droplet crowding on morphology to determine if simple oil droplets capable of morphological plasticity would extend and pack together like living cells in specialized tissues.
In a similar manner to that in which living cells exhibit a variety of shapes and morphological changes allowing them to live and survive under certain conditions, we suggest that also the study of morphological changes of nonliving protocells is a key step in understanding the underlying principles. We hope to use our and other droplet systems as simple artificial life models of living cells to investigate the role of shape and shape change in simple chemical systems, artificial cells, and protocells. It may be that shape provides necessary information or regulation of protocell dynamics and therefore could have a selectable function in future applications.
J.Č. was financially supported by the Czech Science Foundation (Grant No. 17-21696Y). M.M.H. was financially supported in part by the European Commission FP7 Future and Emerging Technologies Proactive (EVOBLISS 611640) and the European Commission Horizon 2020 Future and Emerging Technologies Open Grant Living Architecture (LIAR).
Appendix 1: Note
This work was presented at the Artificial Life Conference 2016 (the 15th International Conference on the Synthesis and Simulation of Living Systems) in Cancún in Mexico, July 4–8. Jitka Čejková received the Best Poster Award. See the “commercial” for this poster in Supplementary Video S3 ( Appendix 2).
Appendix 2: Supplementary Videos
A2.1 Supplementary Video S1: https://youtu.be/qFWmzWsSgsQ
Macroscopic (left) and microscopic (right) temporal sequence of shape change of a decanol droplet. Initial conditions: a glass coverslip with diameter 10 mm, 77.2 μl of 10 mM aqueous decanoate solution, 0.6 μl of decanol, 0.5 μl of 6.5 M NaCl. (Note: 500× faster than real time.)
A2.2 Supplementary Video S2: https://youtu.be/zI5CbBnHwTo
Shape changes of two decanol droplets. Initial conditions: a glass coverslip with diameter 24 mm, 500 μl of 10 mM aqueous decanoate solution, two decanol droplets with volume 2.5 μl each, 3.2 μl of 6.5 M NaCl added. (Note: 1000× faster than real time.)
A2.3 Supplementary Video S3: https://youtu.be/ZKccS1kBIy4
The “commercial” for a poster titled “Shape changing multi-armed droplets” by Jitka Čejková, Martin M. Hanczyc, and František Štěpánek at the ALife 2016 Conference (the 15th International Conference on the Synthesis and Simulation of Living Systems, Cancún, Mexico, July 4–8).