Collapse of active nematic order through a two-stage dynamic transition

Abstract

We present a novel two-stage transition of the ordered active nematic state of a system of bundled microtubules into a biphasic active fluid. Specifically, we show that upon light-induced solidification of the underlying medium, microtubule-kinesin motor mixtures first undergo a cascade of folding in on themselves, which is then followed by a progressive formation of isotropic fluid domains, transforming the system into a biphasic active suspension. Using an active lyotropic model, we show that the two-stage transition in the system is governed by screening effects that become important upon substrate changes, leading to higher activity modes to become important. Specifically, the combined effect of friction and quadrupolar activity leads to the hierarchical folding that follows the intrinsic bend instability. Our results demonstrate the dynamics of the collapse of orientational order in active nematics and present a new route for controlling active matter by modifying the activity through changing the surrounding environment.

Living materials, such as eukaryotic cells and bacteria, are distinct from their inanimate counterparts in their ability to continuously convert chemical energy to mechanical work at the level of their constituent elements [1]. This activity, in turn, orchestrates the formation of collective patterns of motion on scales much larger than the size of the individuals [2, 3, 4, 5], and guides important physiological processes, such as organ development [6, 7] and collective invasion [8, 9, 10]. Similar activity-induced collective patterns are increasingly emulated in bio-inspired synthetic materials, with the aim of rendering them capable of self-organization and self-healing [11, 12].

In the bulk, the mechanistic basis and fundamental instabilities of activity-induced collective pattern formation have been intensely studied both in theoretical models [13, 14, 15] and in experiments [16]. Importantly, however, living materials in most natural setups are constrained by confining boundaries and substrates, where the dynamics of the active systems are closely controlled through their interaction with the surrounding environment. Striking examples include subcellular active flows confined to the cell boundaries [17], twitching motility of surface-attached bacteria [18], and dynamics of cellular monolayers interacting with their underlying substrate [19].

Dynamics of active matter confined in channels of varying shapes and sizes have been the focus of intense research. While various modes of collective self-organization have been identified in theoretical models [20, 21, 22, 23] and experimental systems including subcellular filaments [24, 25, 26], bacteria [27, 28], and confined epithelial and mesenchymal cells [29, 30], the impact of active matter interaction with a tunable and dynamically changing environment is not well understood. Theoretical studies, focusing on the frictional and momentum-damping effects induced by substrates, have suggested possibilities of stabilization, control, and significant alterations in steady-state collective patterns of active materials [31, 32, 33, 34, 35]. Nevertheless, experimental realization of such effects and deciphering the complex impact of the substrate beyond frictional effects, has remained a significant challenge. Moreover, while the focus so far has been on steady-state patterns of motion, the study of transient effects of active matter interacting with substrates, and the possibility of transitioning between different steady-states are yet to be explored.

Here, using a mixture of microtubule-kinesin motor proteins as a model active system residing on a tunable hydrogel material, we reveal a novel two-stage dynamic transition that extinguishes the nematic order of the system and transforms it into a biphasic, yet still active, suspension of the bundled microtubules. In the first stage, cascades of self-similar patterns emerge upon active nematic folding in on itself, which is then followed by the growth of the number of isotropic domains in the second stage (Fig. 1). We combine experimental observations with a continuum model of lyotropic active nematic to reveal the mechanisms that govern the transition between these two dynamical steady states and uncover the previously overlooked impact of the substrate on tuning the dominant forms of active stresses.

Figure 1: Activation of light leads to a two-stage dynamic transition. (a) Schematic representation of the experimental cell. The flow cell is filled with fluorinated oil and an aqueous solution containing the active material, monomer, and initiator of the hydrogel. A flow cell of 50 $\mu$m thickness is prepared using a superhydrophilic slide and a hydrophobic cover slide. Low-power UV light leads to the formation of the hydrogel, confining the active material between it and the oil phase. (b) – (d) Director correlation length, $L_{C_{Q}}(r)$, order parameter, $S$, and root-mean-squared velocity, $v_{\text{rms}}$, respectively, as a function of time. The gray dashed line marks the time when the light is turned on. $L_{C_{Q}}(r)$ and $v_{\text{rms}}$ are scaled by pre-light averages. (e) Snapshots from experiments, showing patterns for different transition checkpoints. The color of each frame corresponds to the corresponding time point indicated in panel (b) (see also Movie S1).

Two-stage transition from active nematic to active biphasic fluid.

In order to explore the effect of contact with the substrate in a controllable manner, we designed an experimental system, where we employ the standard formulation of a kinesin/tubulin active gel to which we add the precursors of a poly(ethylene glycol) (PEG)-based hydrogel that can be photopolymerized with UV light with the concourse of a photoinitiator [36] (see Supplementary Material for more details). Experiments are performed in a flow cell of 50 µm thickness, made with a superhydrophilic and a hydrophobic glass slide (Fig. 1a, and Supplemental Material). The cell is filled with the active mixture and fluorinated oil to allow the formation of the active nematic at the water-oil interface. The interfaced material displays, at this stage, the well-established, seemingly chaotic flow patterns characteristic of an active nematic layer [11, 37]. We then subject the cell to a sustained (120 s) low-power-illumination. The hydrogel polymerizes, leaving a wetting film that is still capable of accommodating the microtubule-based fluid (more details of the experimental protocols can be found in Supplementary Material).

Figure 2: Continuum simulations reproduce the two-stage transition from active nematic fluid to active biphasic fluid. (a) Schematic depiction of the simulation setup. In the beginning, simulation is carried out with only dipolar activity. At $t_{0}$, in addition to the dipolar activity, friction and quadrupolar activity are turned on to mimic the effect of turning on light in the experiments. (b)–(c) Director correlation length (rescaled by the value before $t_{0}$), $L_{C_{Q}(r)}$, and the order parameter, $S$, respectively, as a function of time (in simulation units, s.u.). The gray dashed line denotes the onset of friction and quadrupolar activity ($t_{0}^{f}=t_{0}^{q}$). Plots are averaged over 5 realizations. (d) Line integral convolution (LIC) render of snapshots of the director at different stages of transition (see Movie S2). Black regions correspond to the isotropic phase. The color of each frame corresponds to the corresponding stage indicated in panel (b).

The observed changes in the orientational textures following this simple intervention are drastic (see Fig. 1e and Supplementary Movie S1). The material starts to fold in on itself, following what can be interpreted as a cascade of hierarchical bending instabilities. The result is a highly corrugated material, as exhibited by the practical collapse of its orientation two-point correlations (see Fig. 1b). In the final state, the microtubule suspension seems to rarefy while losing nematic order (see Fig. 1c). We term this state active biphasic fluid, as despite being largely parceled by isotropic non-fluorescent domains, the material does not cease to move at any time, but at much lower velocities (see Fig. 1d). A closer inspection of the temporal evolution of the system reveals that upon turning on the light the crossover from active nematic to active biphasic fluid goes through a two-stage process: a cascade of self-similar folding patterns, followed by a drop in the orientational order and increase in the number of isotropic domains that span the system (see Supplementary Fig. 1a). A hallmark of this two-stage transition is also evident in the evolution of the director correlation length with time (see Fig. 1b): as hierarchical folding emerges during the first stage of the transition, the correlation length and the magnitude of the orientational order sharply drop (Fig. 1c). This is followed by a subsequent increase and a peak in the correlation length, as the isotropic domains start to elongate, which is then followed by a second drop in the correlation length and the magnitude of the orientational order. This identifies the second stage of the transition leading to an active biphasic fluid state being established.

Continuum theory reproduces the two-stage transition.

To investigate the physical mechanism driving the experimentally observed transition, we used a continuum model of lyotropic active nematics [38] (see Supplementary Material for model description). Here, a non-conserved scalar parameter, $\phi\in[0,1]$, is introduced to distinguish the active nematic phase driven by dipolar active stresses [33, 39, 40], from a passive, isotropic phase. When the UV light is on, the gel underneath the substrate solidifies. This leads to a friction-induced screening effect that affects the dominant forms of active stresses exerted by active particles [41]. Therefore, higher-order contributions to active stresses become important [41, 42, 43]. To account for this, we introduce quadrupolar activity to the momentum equation, together with the isotropic friction. To replicate the experimental condition as closely as possible, these terms are switched on instantaneously at a given time after the system has reached the active turbulence state (Fig 2a). This corresponds to switching the UV light on in the experiment.

Figure 3: Physical mechanism behind the two-stage dynamic transition. (a) LIC render of a stripe of nematic phase surrounded by isotropic phase (black) with bend instability evolved under different conditions: (– –) no friction and no quadrupolar activity; (+ –) friction only; (– +) quadrupolar activity only; (+ +) both friction and quadrupolar activity. Dipolar activity is on at all times. See Movies S3–S6. (b) Correlation of director before the friction and quadrupolar activity are activated (black dashed line) and after (pink line). (c) Director correlation length (rescaled by pre-light average) and the order parameter as a function of time for high and low values of dipolar activity obtained via simulations. Plots are the average over 5 realizations. (d) Same as (c) but for experiments (single realization, see Movie S7).

The model qualitatively reproduces the dynamics in the director field (see Fig. 2d and Supplementary Movie S2), accurately capturing the two-stage transition. Ordered regions undergo folding, breaking into smaller perpendicular domains. Afterward, the domains elongate until the active biphasic regime with smaller length scales is established. Remarkably, not only do the qualitative features mirror experimental observations, but the transition follows a two-stage dynamics as both the two-point director correlation and the order collapse (Fig. 2b–c), and the number of isotropic domains subsequently increases, in agreement with experiments (see also Supplementary Fig. 1b). Additionally, in line with the experimental measurements, the crossover between the two stages is marked by the temporal, step-like (see Supplementary Fig. 2) increase and a peak in the correlation length before the transition to biphasic active fluid. Together, the experimental observations and continuum modeling reveal a novel two-stage process for the dynamic transition between active nematic and active biphasic fluids that is mediated by the interaction with the substrate.

Physical mechanism.

We next investigate the mechanism for the emergence of self-similar patterns during the two-stage transition. To understand the separate roles of friction and quadrupolar activity in inducing hierarchical folding, we simulate a minimal setup that consists in setting a stripe of low-activity nematic fluid surrounded by an isotropic phase (Supplementary Movies S3–S6). Here, we let the system evolve through the well-established bend instability (Fig. 3a, panel (i)) under dipolar activity only, and then turn on friction and/or quadrupolar activity.

Turning on only isotropic friction stabilizes the system and suppresses the growth of the bend instability (Fig. 3a, panel (ii)). On the other hand, when quadrupolar activity acts on the system (and the friction is turned off), the nematic phase folds in on itself, creating mushroom-like structures (Fig. 3a, panel (iii)). Finally, when both friction and quadrupolar activities are turned on (Fig. 3a, panel (iv)), the hierarchical buckling of the existing bend instability is observed. This secondary buckling occurs at a much smaller wavelength, compared to the primary bend instability, as shown by the measurements of the orientation correlation (see Fig. 3b). These results indicate that the observed patterns during the two-stage transition result from a balance between the dipolar and quadrupolar activities (destabilizing) and friction (stabilizing) that leads to cascades and eventual decreases in length scales. Hence, the ratios between these three parameters determine the possibility of a substrate-induced transition.

Controlling the transition by tuning activity.

Having established the mechanism of the two-stage collapse of the orientation order in active nematics, we use the model to explore the possibility of controlling the dynamics by varying the activity of the system. The model (Fig. 3c) predicts that for a less active system with lower dipolar activity, the two-stage transition of a system is delayed, the decay to smaller length scales and collapse of the order are slowed down. The increase in the number of isotropic domains is also slowed down (see Supplementary Fig. 2a). Similar behavior can be seen in the experiments (Fig. 3d) when the activity of the microtubules is reduced by lowering ATP concentration before shining light. The resulting experimental patterns (see Supplementary Movie S7) show similar stages of the transition, including folding and elongation. The temporal evolution of the correlation length, order parameter and number of domains qualitatively confirm the model predictions (Fig. 3d and Supplementary Fig. 1b). Together, the modeling and experimental results show how the interplay of dipolar activity with frictional damping and quadrupolar activity shapes the dynamics of the transition from active nematic to active biphasic fluid.

Discussion.

In this work, we present novel controlled experiments, using a reconstituted microtubule-kinesin motor mixture system that is in contact with a tunable hydrogel material, to demonstrate the dynamic pattern evolution of an active fluid upon perturbations in the surrounding environment. Specifically, we show that upon controlled, light-induced solidification of the underlying hydrogel, the active material undergoes a two-stage dynamic transition in which an initially ordered active nematic fluid evolves into an active biphasic fluid. The two stages are marked by distinct pattern formation: in the first stage, quickly after solidification of the underlying hydrogel, active nematic domains evolve into cascades of self-similar patterns by folding in on themselves. This folding cascade is then followed by a significant increase in the number of isotropic domains that are interleaved with active nematic filaments until the final stage is characterized by an active biphasic fluid.

Our continuum model of lyotropic active fluid reveals the mechanistic basis of the two-stage evolution of the system upon solidification of the substrate and shows that increased friction alone is insufficient to explain the observed transition. Substrate friction not only extracts momentum from the system but also induces screening effects, which alter the dominant form of active stress from dipolar to quadrupolar stress. The emergence of significant quadrupolar contribution, induced by hydrodynamic screening effect of the substrate, is in line with theoretical predictions in the context of single active particles [41] and dense collections of active particles, both in the presence of hydrodynamics [42] and in over-damped systems [43]. To our knowledge, the experiments presented in this work provide the first evidence of the important role of quadrupolar active stresses in dense active matter systems under friction. The continuum model further demonstrates how such quadrupolar activity combined with frictional damping can induce the folding of active domains onto themselves, leading to the emergence of cascades of self-similar folding patterns.

Our findings present one of the first examples of the dynamics of active fluid interaction with a reconfigurable surrounding that transitions between fluid-like and solid-like states. Understanding the dynamics of active materials, such as cells and bacteria, requires accounting for the dynamic changes in their surrounding environment. Examples like the extracellular matrix (ECM) and bacterial biofilms underscore the importance of this dynamic interplay. The ECM, with its ability to transition between fluid-like and solid-like states, profoundly influences cell behavior, tissue remodeling, and disease progression [44]. Similarly, bacterial biofilms, through their dynamic matrix properties, govern bacterial community formation, resistance to external stresses, and biofilm-associated infections [45]. Ignoring the dynamic nature of the environment could oversimplify the complexities of cell and bacterial behaviors, hindering our ability to comprehend processes like wound healing, cancer metastasis, and microbial infections. Hence, integrating the understanding of dynamic environmental changes is crucial for unraveling the intricate dynamics of active materials in biological systems.

I Author contributions

J.I.-M and F.S. designed the experiments and A.D. designed the model. I.V.-C. performed the experiments and extracted the director field and velocity field data. A.A implemented the model, performed the simulations, analyzed the simulations and experimental data, and prepared figures with help from I.V.-C. and M.C.P., J.I.-M., F.S and A.D. all contributed to the interpretation of results. A.D. prepared the first draft. A.D. and A.A. wrote the paper with input from all authors. This collaborative effort was led by J.I.-M, F.S and A.D.

II Acknowledgements

We thank Kristian Thijssen for the helpful discussions. A.A. acknowledges support from the EU’s Horizon Europe research and innovation program under the Marie Sklodowska-Curie grant agreement No. 101063870 (TopCellComm). A.D. acknowledges funding from the Novo Nordisk Foundation (grant no. NNF18SA0035142 and NERD grant no. NNF21OC0068687), Villum Fonden Grant no. 29476, and the European Union via the ERC-Starting Grant PhysCoMeT, grant no. 101041418. We thank M. Pons, A. LeRoux, and G. Iruela (Universitat de Barcelona) for their assistance in the expression of motor proteins. I.V.-C. acknowledges funding from Generalitat de Catalunya though a FI-2020 PhD. Fellowship. I.V.-C., J.I.-M., and F.S. acknowledge funding from MICIU/AEI/10.13039/501100011033 (Grant No. PID2022-137713NB-C21). I.V.-C. and J.I.-M. acknowledge funding from MICIU/AEI/10.13039/501100011033 (Grant No. PDC2022-133625-I00).