Article Figures & Data
Figures
- FIG 1
Unique haloed plaque morphology from which the coisolated novel B. bacteriovorus angelus and bacteriophage halo were identified by electron microscopy. (A) Haloed plaques containing both B. bacteriovorus angelus and bacteriophage halo on lawns of E. coli in YPSC double-layer agar plates. Scale bar, 1 cm. (B) Electron microscopy of B. bacteriovorus angelus, stained with 0.5% URA (pH 4.0). Scale bar, 500 nm. (C) Electron microscopy of a 0.22-μm filtrate of a predatory culture, showing the presence of phage particles with curved tails resembling bacteriophage RTP. Phage were stained with 0.5% URA pH 4.0. Scale bar, 50 nm.
- FIG 2
Kinetics of predation. Measured over 48 h on late-log-phase E. coli S17-1 by bacteriophage halo alone (green), B. bacteriovorus HD100 alone (red), and both bacteriophage halo and B. bacteriovorus HD100 combined (purple) versus E. coli plus buffer control (blue). (A) E. coli measured as the OD600 (B. bacteriovorus organisms are too small to register at OD600). (B) E. coli viable counts. (C) B. bacteriovorus HD100 enumeration by plaque counts. (D) Bacteriophage halo enumeration by plaque counts.
- FIG 3
Base model with one prey type. (A) Diagram of the model variables (populations and chemicals) in circles and their positive or negative interactions. The arrow colors match the colors of the terms in the equations in panel B, and the roman numerals refer to the list of processes in the text. (B) Set of differential equations defining the base model.
- FIG 4
Final model and model variants. (A) Diagram of the final model variables (populations and chemicals) and their positive or negative interactions. The arrow colors match the colors of the terms in the equations in panel B. (B) Set of differential equations defining the final model. (C) Top-level model variants with different prey phenotypes (models N1, N2, N3, and N4). (D) Midlevel model variants. (Di) Methods of development of plastic resistance to B. bacteriovorus; (Dii) methods of development of phage resistance. (E) Low-level model variants. The predation rate either saturates at high prey densities or does not (it can differ between B. bacteriovorus and phage).
- FIG 5
Hierarchical model selection process. This infers which model variants from Fig. 4 are best supported by the data (frequency of a variant winning out of 1,000). (A) Competition of models with different number of prey phenotypes. N1, one prey type sensitive to both predators (NS); N2, two prey types, NS and phage-resistant prey (NR); N3, three prey types, NS and NR and prey with plastic phenotypic resistance to B. bacteriovorus (NP); N4, four prey types, NS, NR, NP, and prey with dual resistance (ND). (B) Competition of models with different ways of converting between sensitive prey (NS) and plastic-resistant prey (NP) but the same saturating B. bacteriovorus attack rate (Pii) and nonsaturating phage attack rate (Vi). N3-IG-Pii-Vi, NS intrinsically (spontaneously) converts to NP and back conversion is coupled to growth. N3-S-Pii-Vi, NS conversion to NP is triggered by a signal and back conversion is spontaneous. N3-I-Pii-Vi, spontaneous conversion both ways. (C) The combined variant from panel B is in the middle and its “parent” variants are on either side. N3-SG-Pii-Vi, NS conversion to NP is triggered by a signal, and back conversion is coupled to growth. (D) Model variants, derived from the combined model in panel C but differing in the way the signal is produced. N3-SBG-Pii-Vi, signal derives from interaction of prey and B. bacteriovorus only. N3-SG-Pii-Vi, signal derives from interaction of prey with both predators. N3-SVG-Pii-Vi, signal derives from prey interaction with virus (phage) only. (E) Different ways of generating phage resistance. Phage-resistant prey were already present initially or prey developed resistance de novo or both. (F) Model variants, based on N3-SBG from panel D but differing in attack rate saturation. Pii, B. bacteriovorus attack rate saturates at high prey density (whereas Pi does not saturate), likewise with Vii and Vi for the virus (phage). (G) Mortality of B. bacteriovorus (phage assumed to be stable) was either set to that of Hespell et al. (39) or fitted by the ABC-SMC method. Less decisive competitions (B to D) were repeated 10 times (see Fig. S6).
- FIG 6
Comparison of experimental data (mean values) with fits of the best model variant (from Fig. 5). The model was fitted using either all experimental data (A to D) or all data without dual predation (E to H) and then used to predict the outcome of dual predation (shown in panel H). The parameter values for each case are given in Table S1 in the supplemental material. Experimental data are indicated by symbols; lines represent model simulations. Colors in panels A to H: blue, E. coli prey; red, B. bacteriovorus; green, bacteriophage halo; pink, medium (not experimentally measured). (I to L) Dynamics of the subpopulations of prey and predators predicted by the model that was fitted to all data, corresponding to panels A to D. Colors and lines in panels I to L (as they relate to models in Fig. 3 and 4): blue, E. coli prey (solid line, susceptible prey [NS]; dotted line, plastic-resistant prey [NP]; dashed line, bacteriophage-resistant prey [NR]); red, B. bacteriovorus (solid line, free B. bacteriovorus [P]; dashed line, bdelloplasts [B]); green, bacteriophage halo (solid line, free bacteriophage [V]; dashed line, bacteriophage-infected cells [I]); pink, medium.
Tables
Supplemental material
- Supplemental file 1 -
Supplemental text, Table S1, and Fig. S1 to S8
PDF, 1.7M
- Supplemental file 1 -
