Confinement discerns swarmers from planktonic bacteria

  1. Weijie Chen
  2. Neha Mani
  3. Hamid Karani
  4. Hao Li
  5. Sridhar Mani
  6. Jay X Tang

  • Curated by eLife

    eLife logo

    Summary: In this paper, the authors proposed a new approach by mounting a PDMS microwells of specific sizes on agar surface to confine swarming and planktonic SM3 cells. They found swarming bacteria exhibit a "single-swirl" motion pattern and concentrated planktonic bacteria exhibit "multi-swirls" motion pattern in the diameter range of 31-90 μm. The phase diagram shows that in smaller wells concentrated planktonic SM3 forms a single vortex and in larger wells swarming SM3 also breaks into mesoscale vortices.

    In addition, they conducted systematic experiments to explore parameters defining the divergence of motion patterns in confinement including cell density, cell length, cell speed and surfactant. They concluded that the single swirl pattern depends on cohesive cell-cell interaction mediated by biochemical factors removable through matrix dilution.

    This paper gives a new method to discern swarmers from planktonic bacteria and carefully studies the factors that influence the formation of bacterial vortices under restriction.

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

Powered by flagella, many bacterial species exhibit collective motion on a solid surface commonly known as swarming. As a natural example of active matter, swarming is also an essential biological phenotype associated with virulence, chemotaxis, and host pathogenesis. Physical changes like cell elongation and hyper-flagellation have been shown to accompany the swarming phenotype. Less studied, however, are the contrasts of collective motion between the swarming cells and their counterpart planktonic cells of comparable cell density. Here, we show that confining bacterial movement in circular microwells allows distinguishing bacterial swarming from collective swimming. On a soft agar plate, a novel bacterial strain Enterobacter sp. SM3 in swarming and planktonic states exhibited different motion patterns when confined to circular microwells of a specific range of sizes. When the confinement diameter was between 40 μm and 90 μm, swarming SM3 formed a single-swirl motion pattern in the microwells whereas planktonic SM3 formed multiple swirls. Similar differential behavior is observed across several other species of gram-negative bacteria. We also observed ‘rafting behavior’ of swarming bacteria upon dilution. We hypothesize that the rafting behavior might account for the motion pattern difference. We were able to predict these experimental features via numerical simulations where swarming cells are modeled with stronger cell–cell alignment interaction. Our experimental design using PDMS microchip disk arrays enabled us to observe bacterial swarming on murine intestinal surface, suggesting a new method for characterizing bacterial swarming under complex environments, such as in polymicrobial niches, and for in vivo swarming exploration.

Article activity feed

  1. Version published to 10.7554/elife.64176 on eLife
  2. eLife

    Summary: In this paper, the authors proposed a new approach by mounting a PDMS microwells of specific sizes on agar surface to confine swarming and planktonic SM3 cells. They found swarming bacteria exhibit a "single-swirl" motion pattern and concentrated planktonic bacteria exhibit "multi-swirls" motion pattern in the diameter range of 31-90 μm. The phase diagram shows that in smaller wells concentrated planktonic SM3 forms a single vortex and in larger wells swarming SM3 also breaks into mesoscale vortices.

    In addition, they conducted systematic experiments to explore parameters defining the divergence of motion patterns in confinement including cell density, cell length, cell speed and surfactant. They concluded that the single swirl pattern depends on cohesive cell-cell interaction mediated by biochemical factors removable through matrix dilution.

    This paper gives a new method to discern swarmers from planktonic bacteria and carefully studies the factors that influence the formation of bacterial vortices under restriction.

  3. eLife

    Reviewer #2:

    The authors address the vortex formation of bacteria in circular confinements with a particular focus on the difference of swarming vs. swimming (planktonic) motility of individuals. In the field of active matter, this critical distinction has rarely been studied so far but it is oftentimes ignored in modeling studies. Chen et al. show that qualitatively different patterns emerge for swarming and swimming bacteria. I do therefore believe that the work could have substantial influence on future studies devoted to bacterial pattern formation.

    I have two main concerns detailed in the following.

    1. A central finding of the present study is that the number of vortices/swirls as a function of the well diameter differs for swarming vs. swimming bacteria. The authors argue and show experimentally (Fig. 2) that the behavior is identical for small and large diameters. For intermediate values, however, they report that a single swirl is observed for swarming bacteria whereas swimming bacteria show multiple swirls.

      The fact that the behavior is identical for large wells suggests that the bulk behavior is identical. This is also confirmed by Fig. 2E which shows that the spatial correlation function of the velocity is identical in large wells. To me, that suggests that the boundary conditions play a central role for understanding how the observed phenomenology emerges. [Indeed, it was shown in the past that the interaction of bacteria with boundaries crucially determines the formation of swirls in confinement (Lushi, Wieland & Goldstein PNAS 111 9733 (2014). The authors of this work assume reflecting boundary conditions, which -- to my knowledge -- contradicts the finding of Lushi et al.]. The authors, however, explain the difference of the observed patterns within their modeling study in a different way, namely by a different strength of the (anti-)alignment interactions. Changing the interaction at the level of individual cells will, however, change the bulk behavior too. Accordingly, the numerically observed bulk behavior (Fig. 5B ) is very different in both cases (at a qualitative level). It is difficult to judge the difference in detail because the correlation function was not calculated for the simulations.

      In short:

      The model (Fig. 5A) reproduces the experimental results partially (Fig. 2C), but the modeling analogue to Fig. 2E is missing. The line of arguments seems to me not to be entirely consistent.

    2. Inferring the interactions of active particles from observations of the emergent patterns is a highly non-trivial task. In view of this I am not entirely convinced by the arguments put forward by the authors that "more substantial cell-cell cohesive interaction[s]" are the reason why the swirling patterns formed by swarming/swimming bacteria differ. In this context, I want to raise the attention of the authors to Ref. [Peruani, Deutsch & Bär: Phys. Rev. E 74 030904(R) 2006]. In this work, a clustering transition of self-propelled rods was described. "Rafts", referred to as clusters by Peruani et al., are observed as the aspect ratio of rods is increased. Notably, a kinetic transition towards clustering can emerge even in the absence of any attractive interactions. In short, the observation that cells move in parallel (polar clusters) next to each other does not allow to conclude that cohesive interactions are present. The movies S3 and S4 provided by the authors show that the particle shape of swarming and swimming particles is clearly different. In particular, the elongated swarming bacteria show pronounced clusters (Movie S3) whereas the shorter planktonic cells (Movie S4) do not. The difference in aspect ratio does indeed suggest that swarming and swimming bacteria differ in their alignment interaction. However, this contradicts the observation that spatial correlations in large wells are indistinguishable (see comment 1 and Fig. 2E). Side remark: in the main text, the authors argue that changes of the aspect ratio are not the reason for an increased alignment interaction, however, in the discussion section cell morphology changes (e.g. cell elongation and hyper-flagellation) are mentioned as an indicator that swarming is a different phenotype from swimming.

    In summary, I believe that the connection of experimental observations and modeling are not entirely convincing.

  4. eLife

    Reviewer #1:

    In this paper, the authors proposed a new approach by mounting a PDMS microwells of specific sizes on agar surface to confine swarming and planktonic SM3 cell, they found swarming bacteria exhibit a "single-swirl" motion pattern and concentrated planktonic bacteria exhibit"multi-swirls" motion pattern in the diameter range of 31-90 μm. The phase diagram shows that in smaller wells concentrated planktonic SM3 forms a single vortex and in larger wells swarming SM3 also breaks into mesoscale vortices.

    After that, they conducted systematic experiments to explore parameters defining the divergence of motion patterns in confinement including cell density, cell length, cell speed and surfactant. They concluded that the single swirl pattern depends on cohesive cell-cell interaction mediated by biochemical factors removable through matrix dilution.

    This paper gives a new method to discern swarmers from Planktonic Bacteria and carefully studies the factors that influence the formation of bacterial vortices under restriction. However, major revisions are required to improve the quality of this paper.

    Major questions and comments:

    1. When the authors put the PDMS chip mounting on the edge of the swarming colony, the PDMS chip is completely attached to agar or suspended in a bacterial solution. The distance between PDMS chip and agar surface should be quantified. It is better to have a schematic diagram of the experimental device.

    2. Is the bacteria still expanding outward after a PDMS chip was mounted on agar surface? The effect of PDMS chips on the expansion of bacteria on the agar surface needs to be discussed.

    3. "Diluted swarming SM3 show unique dynamic clustering patterns". In the diluted bacteria experiment, the authors found that the diluted swarming bacteria can form bacterial rafts and the concentrated planktonic SM3 disperse uniformly and move randomly. Hence, when bacteria expand and gradually fill up new empty microwells, is there a process of transition from raft to single vortex state?

    4. In the experiment of altering the conditions of swarming SM3, the authors diluted the swarming cells in Lysogenic Broth (LB) by 20-fold, re-concentrated the cells by centrifugation and removed extra LB to recover the initial cell density. After these operations, they found the previous single swirl turned to multiple swirls and got a conclusion that matrix dilution can affect single swirl patterns. The authors think centrifugation may wash away some surrounding matrix or polymers on the surface of bacteria. Therefore, the steps of centrifugation need to be presented and the effect of centrifugation on the physiological behavior of bacteria should be discussed.

    5. This article covers the PDMS chip directly on the agar surface and finds that swarm and planktonic bacteria have different spatial correlation scales in the restricted microwells. The authors have done a lot of experiments to prove the difference between clusters and planktonic bacteria and explain the reason for the single vortex. However, the conclusion is not clear. Therefore, the authors should focus more on the analysis of this new experimental phenomenon, such as critical length and vortex phase diagram, rather than just describing the experiments they did.

    6. The authors mentioned the critical length for swarming SM3 is ~ 49 μm, whereas, for concentrated planktonic SM3, it is ~ 17 μm. Does this quoted data match what you get from their experimental method? I do not see any follow-up discussion and evidence.

    7. As shown in Figure 1 and Movie_S1_mp4, the direction of the single vortex motion of bacteria is clockwise. However, the article simply ignores that the single vortexes of bacteria all present the same direction, and there is no analysis and reasonable explanation on the vortex direction. As shown in Movie_S5_mp4 on the numerical simulations of circularly confined SM3, simulated bacteria vortex counterclockwise in completely opposite directions. The influence of the microwell boundary on the direction of the vortex should be clearly explained at the level of bacterial movement and preferentially with theoretical simulation.

    8. Swarming and concentrated planktonic Bacillus subtilis 3610 show the same motion pattern across different confinement sizes. However, the authors did not give definitive conclusions and evidence. As shown in Figure S1, bacillus subtilis 3610 show completely different cluster behavior. Therefore, the discussion of 3601WT may cause readers' confusion on the article. It may be better to put it in the supporting material.

    Minor questions and comments

    1. Figure 1C, 1D, 6A, 6B may be more convenient to have a scale bar.
  6. Version published to 10.1101/2020.08.30.274316v2 on bioRxiv
  7. Version published to 10.1101/2020.08.30.274316v1 on bioRxiv