From cell differentiation to cell collectives: Bacillus subtilis uses division of labor to migrate

Abstract

The organization of cells, emerging from cell-cell interactions, can give rise to collective properties. These properties are adaptive when together cells can face environmental challenges that they separately cannot. One particular challenge that is important for microorganisms is migration. In this study, we show how flagellum-independent migration is driven by the division of labor of two cell types that appear during Bacillus subtilis sliding motility. Cell collectives organize themselves into bundles (called "van Gogh bundles") of tightly aligned cell chains that form filamentous loops at the colony edge. We show, by time-course microscopy, that these loops migrate by pushing themselves away from the colony. The formation of van Gogh bundles depends critically on the synergistic interaction of surfactin-producing and matrix-producing cells. We propose that surfactin-producing cells reduce the friction between cells and their substrate, thereby facilitating matrix-producing cells to form bundles. The folding properties of these bundles determine the rate of colony expansion. Our study illustrates how the simple organization of cells within a community can yield a strong ecological advantage. This is a key factor underlying the diverse origins of multicellularity.

Fig 1. Colony expansion in wild type and biofilm-related mutants.

Left: time course experiment of colony growth in WT and colony expansion in srfA, tasA, eps, sigF, and hag mutants, which are defective in producing surfactin, TasA, EPS, sporulation, and motility, respectively. Colonies are toothpick inoculated onto MSggN medium as described in the Materials and Methods. Right: WT colony after 70 h. The different regions of the colony are named: dendrites, petals, and rays.

Fig 2. Colony expansion in chimeric colonies of sliding-deficient mutants.

(A) Chimeric colonies of different pairwise combinations of sliding-deficient mutants (imaged 52 h after inoculation). Strains were mixed 1:1, and 2 μl of the mix was spotted in the center of the plate. (B) Colony expansion of eps tasA + srfA chimeras when inoculated in different ratios of eps tasA:srfA (imaged after 24 h): 9:1, 4:1, 1:1, 1:4, and 1:9. (C) Colony expansion of eps tasA–YFP (false-colored green) + WT-mKate2 (colored red; see S1 Table for specifications) chimeras for different initial ratios of eps tasA:WT (imaged after 32 h): 19:1, 9:1, and 1:1. Yellowish regions in the colony correspond to colony parts where both eps tasA and WT cells occur. In the rightmost fluorescence image, only the colony edge is shown, as the colony was too big for a single microscopy image (the center of this colony is towards the lower left corner). Images were taken with a stereomicroscope.

Fig 3. Temporal gene expression dynamics of srfA and tapA during colony expansion in wild-type cells.

Surfactin- and matrix-producing cells are monitored in a WT strain harboring promoter fusions (P_srfA-YFP and P_tapA-CFP) of srfA and tapA to genes encoding yellow and cyan fluorescent proteins, respectively. (A) The average expression level of tapA and srfA was measured by microscopy 2, 4, 6, 8, 10, 12, 24, 31 h after inoculation. The average expression level is equal to the average fluorescence intensity in labeled WT cells (n = 20–50 microscopy images per time step) minus that in non-labeled WT cells (n = 10–30 microscopy images per time step). Fluorescence intensity data were acquired from segmented microscopy images (n = 439; containing many thousands of cells). (B) Representative microscopy images from colonies dissected at 10 h (1) and 31 h (2) after inoculation. AU, arbitrary units. Cells expressing srfA are false-colored red, and cells expressing tapA are false-colored green.

Fig 4. Spatial expression pattern of srfA and tapA during colony expansion in wild-type cells.

Top: colony growth 10, 21, 34, 45, and 57 h after inoculation. Colony expansion is divided in two growth periods: (1) dendrite formation (<11–13 h) and (2) petal-shaped colony outgrows (>11–13 h). (A) Microscopy image of a cellular aggregate that appeared in the first growth period. Red and green fluorescent cells represent, respectively, surfactin- and matrix-producing cells in the double-labeled P_tapA-CFP P_srfA-YFP WT strain (CFP and YFP are artificially colored green and red, respectively). (B) Cellular aggregate at the transition from the first to the second growth period. All microscopy images were made at the edge of the colony with an inverted microscope. Cellular aggregates were examined in colonies inoculated by toothpick or pipet and from a passaging experiment (see Materials and Methods). The observed clumping was qualitatively the same for all inoculation conditions.

Fig 5. Van Gogh bundles and tapA expression at the colony edge 34 h after inoculation.

(A) Composite image of van Gogh bundles at the colony edge, consisting of multiple microscopy frames. Left: phase-contrast image. Right: green cells represent matrix-producing cells (i.e., tapA expression) in the double-labeled P_tapA-CFP P_srfA-YFP WT strain (CFP is artificially colored green). srfA expression is not shown in the composite image because of bleaching problems with multi-image acquisition. (B) Phase-contrast microscopy images of van Gogh bundles at high magnification.

Fig 6. Van Gogh bundle formation in mutant chimeras.

Microscopy images were taken from the colony edge 21 h after inoculation, which is close to the temporal transition from the first to the second growth period in mutant chimeras mixed at a 1:1 ratio. One strain per chimera is labeled with constitutive expression of mKate2 (false-colored red) as indicated below. The top, middle, and bottom rows of images show, respectively, the phase-contrast, fluorescence, and overlay microscopy images. Four mutant chimeras were examined (columns): (1) srfA (mKate2) + eps, (2) srfA (mKate2) + tasA, (3) eps (mKate2) + tasA, and (4) eps tasA (mKate2) + srfA. For the eps-mKate2 + tasA chimera, the two columns show images from regions without (left) and with (right) visible van Gogh bundles. The left image without van Gogh bundles, but with strong cell clumps, was acquired from a dendrite that had not made the transition to petal growth yet (such dendrites were not present for the other chimeras, because they developed more quickly; see S8 Fig).

Fig 7. Localization of TasA protein in van Gogh bundles.

(A) Representative microscopy images of the WT strain with a protein fusion of TasA to mCherry (TasA-mCherry) at the colony edge 20 h after inoculation (phase-contrast [left] and fluorescent [right] images). Red corresponds to localization of TasA protein. White arrowheads indicate illustrative points in the images that show TasA localization at the pole-to-pole interaction zone between cells. (B) Chimera of TasA-mCherry + tasA mutant at the colony edge 26 h after inoculation (i.e., in the second growth phase). The tasA mutant is labeled with a constitutively expressed YFP gene (false-colored green). Van Gogh bundles consist of both strains. Phase-contrast and fluorescent images are shown.

Fig 8. Van Gogh bundles and the emergence of filamentous loops at the colony edge.

Images were taken at the colony edge of a WT strain grown for 28 h. The upper three images are insets of each other from a low (left) to high (right) magnification. The white arrow indicates the distance the van Gogh bundles spread over the agar plate. The lower image shows the van Gogh bundles at a higher magnification.

Fig 9. Van Gogh bundles and colony edge for different mutants.

The upper row of images shows colonies 18 h after inoculation. The middle row of images shows the presence or absence of loops at the colony edge. The lower row of images shows the presence or absence of van Gogh bundles at the colony edge. In total, five strains were examined, WT and four sliding-deficient mutants: srfA, eps, eps tasA, and tasA. Scale bars are identical for all images in a row.

Fig 10. Model of filament growth and migration.

Upper left: microscopy image of WT cells on the edge of the colony taken 18 h after inoculation. Lower left: relative migration rate for three parameter settings: (A) default parameter setting (grey), (B) high bending rigidity between cells (blue), and (C) large cell size (dark blue). Histograms and error bars show, respectively, mean and standard deviation (n = 10) in the extent of migration along the y-axis of the two-dimensional space (see images on the right; filament growth is initiated on the bottom). Right: filamentous loops at the end of the simulation for three representative runs, one for each of the three different parameter settings. See Materials and Methods for detailed model description and exact parameter settings (S2 Table).

Fig 11. Schematic overview of cell differentiation and collective properties in B. subtilis colony expansion.

Red and green cells represent, respectively, surfactin- and matrix-producing cells. Dendrites predominantly consist of surfactin-producing cells interspersed with clumps of matrix-producing cells. The petals of the colony consist predominantly of matrix-producing cells that form van Gogh bundles. We propose that surfactin mediates the expansion of van Gogh bundles by reducing the friction between the van Gogh bundles and substrate and that van Gogh bundle expansion is driven by cell division. The elastic and folding properties—dependent on matrix-producing cells—of the van Gogh bundles allow for an efficient colony expansion and prevent the bundles from breaking under increased compression.