APPLICATION OF TEMPORAL LOGIC GATE FOR DIFFUSION-ACTIVATED SPATIAL PATTERNING
4.3 Spatial patterning with temporal logic gate
Chapter 3 introduced an integrase-based temporal logic gate (Figure 3.1) that used two integrases (Bxb1 and TP901-1) to implement logic with four unique genetic states (So, Sa, Sb, Sab). The final population fractions were characterized by man- ual addition of inducers into liquid culture. However, we hypothesized that slow diffusion on a 2D surface might also be sufficient for triggering event detection in temporal logic gate cells. Here, we show that diffusion-based delays are sufficient to create the timing differences necessary to trigger the temporal logic gate in a reliable way. Spatial patterns are characterized by testing plasmid-based and chromosomally integrated temporal logic gate strains on different types of media.
Figure 4.1A show the experimental set-up for 2D diffusion experiments. Standard petri dishes with 1% agar + M9CA minimal media were made and a biopsy punch (2mm diameter) was used to create point sources for liquid inducers, which were added at high concentration (arabinose 2%/vol, aTc 20 µg/ml). We expect these inducers (both small molecules) diffuse at a rate proportional to the density of the agar and the concentration of inducer, and to have some radius of induction.
Since the inducers do not diffuse instantaneously, cells spread uniformly over the surface would encounter different sequences of events (Figure 4.1B). In the non- overlap regions, cells closer to the inducer a source will see a only and switch to state Sa (RFP fluorescence), while cells closer to the b source switch to state Sb (no fluorescence). Within the overlapping region between the two inducers, there will be two populations: Cell closer to sourceawill encountera thenband switch to Sab (GFP fluorescence), while cells closer to source b will encounter b then a and switch to state Sb (no fluorescence). We tested both a plasmid-based version of the temporal logic gate (multiple copies of DNA targets, much higher fluorescence), and the chromosomally integrated version used in the previous chapter (single chromosomal copy of DNA target, lower fluorescence output). Additional information about plasmids and strains constructed are described in Materials and Methods (Table 4.1).
arabinose
Inducer a aTc
Inducer b arabinose
Inducer a aTc
Inducer b
Ea,Sa
a only Eb,Sb b only
Eab,Sab
a then b Eba,Sb b then a
A B
C D
Figure 4.1: 2D diffusion experiment with temporal logic gate. A) Diagram of inducer diffusion. Point sources of arabinose (inducer a) and aTc (inducer b) will induce cells within a certain radius. Only a small region in the center will be exposed to both inducers. B) Diagram of expected cell states based on overlapping inducer diffusion. C) Spatial patterning from a uniform lawn of plasmid-based temporal logic gate strain. RFP expression occurs where the cells encounteraonlyand switch to state Sa, while GFP expression occurs only where cell encounter a thenb and switch toSab. D) Diffusion pattern with chromosomally integrated temporal logic gate strain.
Diffusion experiments with a uniform lawn of the plasmid-based temporal logic gate created a bright red circle around the inducer a point source intersecting a bright green crescent centered around the inducerbpoint source (Figure 4.1C). The RFP expression corresponded to where cells encountered a only and switched to stateSa, while the GFP expression corresponded to where cells encounteredathen
of Figure 4.1C,D are shown in Figure 4.3BE.
D E F
A B
D E F
C
Figure 4.2: Microscopy of spatially differentiated colonies. Cell cultures of plasmid- based temporal logic gate were mixed with liquid media and spotted around inducer point sources. A) Plate imaged in ambient light shows overnight cell growth is independent of location on the plate. B) Blue light imaging shows RFP/GFP segmentation only between the two point sources. C) Cells in the overlap region (middle circle) were extracted and imaged at 10x magnification (Olympus IX81).
Images were taken of cells in the red, middle, and green quadrants (labeled D,E,F).
D) Within the overlap region, cells closer to the arabinose point source encounter induceraonly(Sa) and so express RFP. The densely grown cells show curly growth patterns. E) Cells that are equidistant from the two point sources are a mix of Sa
(RFP) and Sab(GFP). F) Cells in the overlap region that are closer to the aTc point source can senseathenb, and switch to stateSab(GFP).
Diffusion-differentiated cells were imaged at 10x magnification on a fluorescence microscopy (Figure 4.2). Cultures of plasmid-target strain were spotted onto an agar plate with inducer point sources and incubated overnight (Figure 4.2A). RFP fluorescence (Sastate,aonly) was observed for regions encircling the arabinose point source, GFP fluorescence (Sab, athenb) was only observed in the region between the the two point sources, and no fluorescence was observed for cells encircling the
aTc source or that were too far away from either source (Figure 4.2B). Fluorescence patterns were confirmed by further imaging cells in the overlapping region at 10x magnification (Figure 4.2C). Cells in the region between the two point sources show segmentation based on distance from each point source (Figure 4.2C) – cells closer to the arabinose source are predominantly red (Sa)(Figure 4.2D), those in the middle are a mixture of red and green (Figure 4.2E), and those closer to the aTc source are green (Sab)(Figure 4.2F). A sharp border between RFP and GFP populations was not observed; instead, there was a gradual transition where the population became predominantly red or green.