Table 1.
Strain and plasmid used in this study.
Fig 1.
Single-mode imaging system prototypes.
(A) In the first prototype, the shutter of the SLR camera is controlled with a computer and synchronized with a timer-controlled power-supply used for the blue LED transilluminator. Green fluorescence is selected via an amber filter. (B) In a more complex prototype variation, a Raspberry Pi computer is used to control a Pi camera, a rail-mounted optical filter change plate and blue Illumination module ON/OFF. Green and red fluorescence signals are selected via a band-pass filter (green) and a long pass filter (red). A smart-plug controlled white LED tape is used for bright-field epi-illumination.
Fig 2.
Overall design of LOTUS and its components.
(A) x-axis actuator, (B) y-axis actuators, (C) stepping motor box, (D) supporting arm, (E) diffuser box, (F) white illumination module (for transillumination), (G) epi-illumination illumination module scaffold, (H) white illumination module (epi-illumination), (I) blue illumination module, (J) green illumination module, (K) relay modules wired to each illumination modules, (L) stepping motors attached to Illumination module scaffold and (M) filter wheel mounted with two optical glass filters and (N) camera module. The bottom panels reveal more details about 3D-printed components.
Fig 3.
Overview of control and circuit components.
(A) Raspberry Pi 4 model B, (B) GPIO extension board, (C) 5V DC power source for suppling power to stepping motor control module and relay module, (D) 18V DC power source for suppling power to illumination module, (E) bread board, (F-I) relay modules for different LEDs (Wayintop), (J) 64 MP camera (Arducam UCTRONICS-B0399), (K) filter wheel controlled by stepping motors, (L) illumination module controlled by stepping motors, (M) white LED transilluminator and support arm, (N-R) stepping motors (Shanxingan 28BYJ48) and control modules (ULN2003) attached to (S) y-axis actuator and (T) x-axis actuator or optical elements (K-L). These schematics were generated using Fritzing software and MS PowerPoint.
Fig 4.
Benchmark tests for imaging time and sample position reproducibility (drift).
(A) Variation in the image timestamps over 288 imaging cycles (count). The time stamp difference between transillumination images of well 1 (the first step of the imaging cycle) and white epi-illumination of well 9 (the last step of the cycle) is plotted as deviation from the expected cycle duration time (1200 sec) in seconds. (B) Analysis of well position alignment. Light intensity was measured along the orange arrow for each image, and kymographs (temporal heatmap) were generated to show edge position (mm or pixel) (horizontal axis), over time (288 x 12 min cycle, over 4 days) on the vertical axis. The pixel size was calculated from the diameter of the Petri dish. Pixel intensity is shown as normalized intensity. (C) Kymographs for each of nine well positions. (D) Detection of dish edge as maximum brightness variations position. (E) Well position variability. The dish edge position deviation from its initial position (t = 0) is shown for each well (n = 288 cycles). Box plots show the median, interquartile range (IQR) and whiskers extend to the most distant data point within 1.5 X IQR.
Fig 5.
Multimode imaging of biofilms with LOTUS.
Bright-field (epi-illumination and transillumination) and fluorescence images (green and red fluorescence channels) of biofilms containing one or two reporters were captured at biofilm growth experiment endpoint. Samples are: E. coli MG1655 promoter strains U66 (promoterless), U66+p5mCherry, rsd, and rsd+p5mCherry.
Fig 6.
Evaluation of fluorescence intensity flatfield uniformity using alignment disks.
(A) Alignment disk, 3D-printed adapter and 36.7 mm dish, shown separately and assembled. (B) Fluorescence intensity was measured at each well position using both green and red alignment disks at each pixel position along the four directions shown by arrows. (C-J) The resulting intensity profiles are shown for the green (C-F) and red (G-J) disks. (K and L). Directional variation in CV values of pixel intensity over a 250-pixel window that excludes the central hole and disk edges for all well positions (n = 9). Box-and-whisker plots show the median, IQR, and whiskers extend to the most distant data point within 1.5 X IQR.
Fig 7.
Spatio-temporal fluctuations of fluorescent reporters in growing biofilms.
(A) Time-lapse images collected during the growth of E. coli biofilms expressing GFPmut2 and mCherry (Fig 5) were analyzed along the orange arrow and the pixel intensity was measured. (B) Kymographs showing pixel intensity along the spatial (horizontal axis) and time (vertical axis) dimensions. Color intensity is displayed as the relative intensity in either bright-field (transillumination), green (GFPmut2) or red (mCherry fluorescence). (C) Transmitted light or fluorescence intensity of E. coli biofilm growth. Shading represents the SD. (D) Integrated intensity values for transillumination or fluorescence signals are plotted for the two time points highlighted by arrowheads in (C) (2320 and 7200 min). The error bars show the mean ±SD (n = 9, technical replicates).
Fig 8.
Effect of sub-MIC ampicillin treatment on GFP expression level, for selected promotor strains in growing E. coli biofilms.
(A) Temporal changes in biomass proxy normalized fluorescence intensity for (U66 promoterless) and five selected promoter strains. Error bars indicate the standard deviation (n = 3, biological replicates). (B) Change in integrated fluorescence intensity over the whole biofilm area. The AUC was calculated from the first time point at which normalized fluorescence intensity was detected until experimental endpoint (n = 3). The bar shows the mean and error bars indicate SD (n = 3). (C) Heatmap of fold-change in normalized fluorescence intensity during biofilm development (ampicillin presence vs absence).