Preprint: bacterial patterns followed a relocated embryo, and a computer model distinguished cultures containing an embryo, a xenobot, or bacteria alone

Preprint: bacterial patterns followed a relocated embryo, and a computer model distinguished cultures containing an embryo, a xenobot, or bacteria alone

A preprint published on September 3 described an experiment in which motile Bacillus subtilis bacteria changed their spatial pattern in liquid near a frog embryo and a xenobot, a body assembled from embryonic cells. When the embryo was moved, the region of bacterial accumulation moved with it. A computer model used the bacterial pattern to distinguish among three conditions: an embryo, a xenobot, and a culture containing bacteria alone.

The researchers tested whether the spatial pattern of the bacteria contained information about a neighboring living system. They compared a frog embryo with a xenobot. Both consist of cells from the same frog species and share the same genome, but a xenobot is assembled into a separate body and develops differently.

In cultures containing bacteria alone, motile cells spontaneously formed a branching pattern. A nonmotile strain and fluorescent microparticles produced almost no such pattern. Motile bacteria formed a fluorescent halo around a xenobot. After one hour, its mean area reached 9.4 mm², while the values in the two control groups remained within hundredths of a square millimeter. This control linked the halo to active behavior within the bacterial population.

The halo grew around a living embryo. Around a heat-inactivated embryo, the initial accumulation later declined. When the embryo was moved, the bacterial accumulation disappeared from its previous location and appeared near the new one. The bacterial pattern therefore changed with the target's current position and state.

The researchers then tested the role of the ionic environment. When the potassium chloride concentration was increased from 1.5 to 200 millimoles per liter, the halo area at hour 11 increased from 0.67 to 23.0 mm². Early accumulation was slower in bacteria lacking YugO, a potassium ion channel protein. In this system, the ionic environment altered the bacteria's early response to a living target.

To ensure that the model relied on the bacterial pattern, the researchers masked the target's outline in the video frames. Using only the remaining bacterial distribution in held-out videos, the model distinguished the three conditions with an accuracy of 61.13%, compared with a baseline proportion of 33%.

“The physiological state of one collective is partially recorded in the form of another,” the authors write.

In this laboratory system, the spatial distribution of bacteria provided a measurable response to the state and position of a neighboring multicellular target.