Research ArticleAPPLIED SCIENCES AND ENGINEERING

E. coli “super-contaminates” narrow ducts fostered by broad run-time distribution

See allHide authors and affiliations

Science Advances  13 Mar 2020:
Vol. 6, no. 11, eaay0155
DOI: 10.1126/sciadv.aay0155
  • Fig. 1 Visualizing upstream super-contamination.

    (A) Sketch of the microfluidic device. The black arrow indicates the direction of motion of the setup during a scan, while the lens stays in place. (B) Image of swimmers, represented by green spots, near the exit of the bacterial reservoir at the beginning of one contamination experiment. (C) Reconstruction of the channel using images from a scan performed between 2 and 2.5 min from the beginning of the experiment: The pioneer swimmer has reached a distance of 3 mm from the bacterial reservoir. (D) An analogous reconstruction, taken from a scan performed from 12 to 13.5 min: The pioneer swimmer has reached a distance of 13 mm. The flow velocity at the center of the channel was 80 μm/s. As a size reference, the width of the channel is w = 40 μm.

  • Fig. 2 Bacterial paths represented by the superposition of photograms on fixed positions.

    Individual images were taken at intervals of 1/30 s from a 13-s-long video for (A) and (B), and 9-s-long video for (C). (A) Bacterial trajectories near the entrance of the channel from the bacterial reservoir, 15 s after starting the contamination experiment. (B) Zoom at the entrance of the channel: Red arrows indicate the path of one bacterium that first moves upstream along the left wall, then detaches from it, and then reattaches to the right wall, continuing its upstream motion. (C) A zoom at the end of the channel: Red arrows indicate the upstream trajectory of a bacterium that has reached the opposite extreme of the channel after swimming for a distance bigger than 15 mm in 15 min; the orange arrows indicate the paths of “inactive” beads moving downstream (the borders of the channel have been overdrawn in white for clarity). The flow velocity at the center of the channel was of 5 μm/s. As a reference, the width of the channel is w = 40 μm.

  • Fig. 3 Quantification of the upstream pioneer contamination.

    (A) Scanning position (crosses) for two experiments with different flow velocities. The scanning stops at the farthest bacteria (pioneers). The straight lines show the advance of the contamination pioneers, with a slope that gives the contamination velocity Vcont. (B) Positions of the pioneers as a function of Vbrt. The dotted line is the curve y = x. (C) Vcont/Vbr as a function of Vf for different experiments.

  • Fig. 4 Concentration profiles and boundary conditions.

    (A) and (B) show the concentration profiles along the channel at different times, for two different flows: Vf = 23 μm/s and Vf = 80 μm/s, respectively. The average bacteria velocity in the reservoir is Vbr = 22 μm/s in both cases. The thicker continuous lines are simulations based on the broad statistical distribution of run times (α = 0.5 and Vb = 22.0 μm/s) for (A) and (α = 0.5 and Vb = 19.5 μm/s) for (B). (C) Number of bacteria inside the channel for the set of scans of (B), using the same color code. The three snapshots show the increase in the concentration in the outlet (stock of bacteria) as time goes, therefore increasing the flux of bacteria into the channel.

  • Fig. 5 Modeling super-contamination.

    (A) Experimental spatial-temporal plot illustrating the two characteristic lengths associated with the upstream bacterial motion. (B) Sketch showing the mechanism of upstream contamination in 3D (top) and in the 1D version, supporting a biased random walk model (bottom). (C and D) Contamination profiles at different times generated by the biased random walk model based on exponential and broad distributions of run times, respectively. The flux of bacteria from the reservoir was kept constant in the simulations. The dimensionless time and length scales make the profiles good for comparison.

  • Fig. 6 Channel contamination in absence of a flow.

    Normalized spatial concentration P(x, t) ≡ c(x, t)/ ∫ c(y, t)dy of the bacteria in the channel, for time spanning between 400 and 3000, where c(x, t) are the concentration profiles. The black line is the result of a numerical random walk model using the switch time distribution ψPL(t). Inset: Contamination lengths, Λ(t) = ∫ x P(x, t)dx, for the experiment and the numerical model at t = 3000 s.

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/6/11/eaay0155/DC1

    Fig. S1. Optimization factor F as a function of the parameters α and Vb of the model.

    Fig. S2. Superposition of contamination experiments (thin lines) and simulations (thick lines) using the exponential statistics for bacterial detachments from the surfaces.

    Table S1. Summary of physical magnitudes.

    Movie S1. A sequence of three scans starting at times 0, 327, and 719 s for a contamination experiment with a maximum flow velocity of 80 μm/s.

    Movie S2. Bacterial upstream-downstream dynamics for a contamination experiment with a maximum flow velocity of 80 μm/s.

  • Supplementary Materials

    The PDF file includes:

    • Fig. S1. Optimization factor F as a function of the parameters α and Vb of the model.
    • Fig. S2. Superposition of contamination experiments (thin lines) and simulations (thick lines) using the exponential statistics for bacterial detachments from the surfaces.
    • Table S1. Summary of physical magnitudes.
    • Legends for movies S1 and S2

    Download PDF

    Other Supplementary Material for this manuscript includes the following:

    • Movie S1 (.avi format). A sequence of three scans starting at times 0, 327, and 719 s for a contamination experiment with a maximum flow velocity of 80 μm/s.
    • Movie S2 (.avi format). Bacterial upstream-downstream dynamics for a contamination experiment with a maximum flow velocity of 80 μm/s.

    Files in this Data Supplement:

Stay Connected to Science Advances

Navigate This Article