This web page was produced as an assignment for an undergraduate course at Davidson College.
The goal of this study was to improve synthetic genetic logic in hopes to provide a better understanding of how environmental and developmental signals control biological processes. Understanding cellular functions that operate similar to electrical circuits can provide numerous applications such as, pharmaceutical and fuel production. In the past, researchers have used genetically encoded logic to modify cellular functions, study living systems, and influence cellular development. However, most approaches use two-terminal device architecture with gate-gate layering, similar to electrical circuits that prove difficult for mass scaling and multiple input signals. Therefore, the authors attempted to engineer a device architecture in which the same regulatory molecules can be reused to implement all logic gates, where input and output signals are distinguishable, and all gates can be activated via a common input signal.
Using transcript logic a three terminal device was engineered where independent input signals govern transcriptor logic elements and regulate transcription. The three terminal device represents transcriptor logic, where a transcriptional element or engineered piece of DNA controls the flow of RNA polymerase along a piece of DNA. DNA represents an electrical wire and RNA polymerase’s rate of gene expression represents electrical current in an electrical circuit. The engineered device contains a transcriptional element, a segment of engineered DNA with three terminals: 1) input terminal 2) output terminal 3) control signal terminal. Interconnected transcriptional elements are logic gates that provide Boolean logic: AND, NAND, OR, NOR, XOR and all possible combinations there of. Selected unidirectional serine integrases control gate switching allowing output signal or transcription of reporter gene, GFP that can be measured using fluorescence.
In summary, the authors successfully designed a three terminal device that encodes single-layer digital logic and enables engineers to amplify transcription across different organisms. The three terminal device allows multiple control signals at a single spot on a wire like a multiple input transistor with different control signals at one spot. As a result, the new-engineered device enables mass scaling and a more comprehensive synthetic tool to study how environmental and developmental signals control biological processes.
Overall, I think the authors made a good attempt at representing their synthetic design through visual images. However, the authors failed to provide substantial background into their design. The concept of designing synthetic parts like an electrical circuit to mimic cellular processes is never explained. The authors assume the reader understands digital electronics. Ideally, I would have liked to seen a key explaining what each part of the device constructed represented. For example, the three terminal device engineered represents a transistor in an electrical circuit, etc. In addition, the use of engineering language made it difficult to follow logic of the device engineered. Without prior knowledge of digital electronic terms and synthetic biology the reader is lost.
Figure 1 explained the design of the terminal device in a logical manner. The use of color in Figure 1 was very beneficial for understanding the biological interactions taking place inside the device. Figure 1A-C clearly laid out the basis of the synthetic logic, but more detailed labels would have helped the reader connect the synthetic logic to biological parts. A key of what each biological part represents in an electrical circuit would provide the reader with a better visual representation of how the device works and relates to the overall goal of the study. Also, it is not obvious that the input, integrase, and output are terminals of the logic gate. Better labeling could help eliminate confusion and enhance the flow of the argument. Figure 1D did a great job of showing how the three terminal device with only one transcriptor for each gate can output all possible combinations of answers for a Boolean clause. Showing every input possible, the recombination and inversion of the terminator helps visualize biologically how a Boolean statement can be generated. Figure 1E is probably the least clear of the panels. The arrows from the integrases represent different combinations of control signals and are supposed to show that output is high only if control signals are different. However, there is no description of what the humps in the arrows represent or how the presence of the two integrases affects the output signal.
Figure 2 is a brilliant figure that at first glance is overwhelming. The truth table is vital to understanding what each Boolean gate or clause represents. Showing the output generated for each transcriptor element depending on the control signal was a clever way to relate biological input with mathematical results. In addition, it is the only place in the paper where the Boolean logic is explained. I suggest that instead of using ctrl in the figures and integrase in the text that the authors choose one and be consistent throughout the paper to avoid confusion over what represents the control signal. The predicted and actual population measurements provide clear visual representation of how the output expression is changing depending on changes in the input signal. However, it would have been nice to have incremental measurements on the x and y-axis showing how much the concentration of the input is changing. The single cell section clearly shows the digital threshold and the variation among level of expression. The bar graphs showing percent cells in addition to GFP intensity gives two perspectives on how the output signal is regulated and clearly displays the level of matching problem.
Figure 3 and Figure 4 were by far the easiest to grasp. The contour images provide clear explanation of how GFP expression changes individual cell expression and population expression compared to the control. The contour images make it slightly hard to see small differences between the control and output but make the general pattern clear. Figure 3C-D was a little difficult to differentiate high vs. low expression based on the bar, but standard deviation numbers helped summarize the amplification effect. Figure 4 does a great job at summarizing the amplification effect of control signals. However, the authors do not provide information on why they graphed colored boxes indicating the amplification of control signals to themselves.
In summary, the study is brilliantly designed and makes a great effort to conceptualize the experiments performed. The authors did a great job of showing what they did but failed to provide substantial background to the study. Clear labeling and a stronger introduction would enable the reader to better understand the function of the Boolean integrase logic gates designed.
Panel A
A visual representation of a three terminal transcriptor device used to study genetic logic gates and their regulation of transcription rates. The first terminal is the input terminal where RNA polymerase enters and if it can transcribe the engineered DNA flows out of the output terminal. Intergase (Int) that enter through the intergase terminal are independent control signals that regulate the transcriptional current or flow of RNA polymerase between the input and output gates represented by DNA. The gate input and output signals are specific transcription rates at positions on DNA marking the boundaries for logical elements in the transcriptor.
Panel B
Panel B shows a logical element that uses an asymmetric transcription terminator (T) that acts as a reversible check valve. The T disrupts the flow of RNA polymerase in one of two directions (right or left). On top the red T blocks the RNA polymerase flow to the right and on the bottom the gray T allows flow to the right. The opposing recombination sites (black/white triangles) flank the terminator and direct the flipping orientation of a T by an intergerase. If a control intergerase is present the red terminator is inverted and RNA polymerase’s activity is restored to the right.
Panel C
The logical element in panel C has aligned recombination sites that integrases can bind to and excise the asymmetric transcription factor (Red T) allowing RNA polymerase to flow to the right. Inversion and excision of the asymmetric transcription terminator occurs at recombination sites flanking the terminator and allows regulation of all gates without having to construct a connected series of simpler gates.
Panel D
Diagram of a transcriptor logical element XOR (only if input signals are different will transcription occur) with one asymmetric transcription terminator (T) flanked by two pairs of opposing recombination sites (orange/blue and black/white). Each of the opposing recombination sites is recognized by different integrases. As a result, depending on the integrase present or absent four distinct states can be engineered to control the orientation of the terminator and its function of the transcription.
Top:
If neither integrase A, which recognizes orange/blue opposing recombination sites or intergrase B, which recognizes black/white-opposing recombination sites are present the terminator blocks transcription.
Middle Left:
If integrase A is present the orange/blue recombination sites are recognized and inversion of the terminator occurs allowing transcription through the transcriptor.
Middle Right:
If integrase B is present the black/white recombination sites are recognized and inversion of the terminator occurs allowing transcription through the transcriptor.
Bottom:
If both integrases are expressed the terminator is inverted by one recombination site and then restored to its original orientation through inversion at the other recombination site, again blocking transcription.
Panel E
Panel E shows the logical element from D within a three terminal device. The XOR gate is controlled by two independent integrases. The output is high only if the control signals are different or from different integrases.
Figure 2 shows six transcriptor logic elements that encode for six different Boolean functions or gates within the three terminal device.
Truth Table:
The truth table shows, which control integrases (ctrl1 and ctrl2) must be present or absent to receive an output.
Architecture:
The architecture represents the design of the permanent gates that write or implement logical expressions.
Predicted:
The authors picked unidirectional serine integrases from bacteriophage TP901 (ctrl1) and Bxb1 (ctrl2) to control gate switching. TP901 and BXb1 were expressed in a recombination control plasmid under the control of arabinose (ara) and anhydrotetracycline (aTc) induction. Using the Hill function the expected behavior of multi-input gates after interaction with a single integrase were plotted. The graph shows the predicted fraction of cells that will be on or express GFP depending on the concentration of ctrl1 or ctrl2 controlled by aTc or ara. The x-axis shows the concentration of ctrl1 input increasing as the wedge becomes thicker and the y-axis shows the concentration of ctrl2 input increasing as the wedge become thicker. The heat scale on the right shows the fraction percent of cells turned on (red is high and blue is low).
Population Measurements:
A plasmid encoding for each of the six of the logical elements was designed to place one of the logical elements between a strong standard promoter (input signal source) and green fluorescent protein (GFP). The expression of GFP was measured as an output signal reporter. Overall the figure shows the gate output assay results from OD measurements using a plate reader. The x-axis shows the concentration of ctrl1 input increasing as the wedge becomes thicker and the y-axis shows the concentration of ctrl2 input increasing as the wedge become thicker. The heat scale on the right shows the level of GFP fluorescence/OD in a.u. (red is high and blue is low).
Single Cell:
Distribution of GFP was measured among single cells depending on the combination of inducer concentrations (200ng/ml for aTc and 1% w/v for ara) presented in the truth table. The x-axis shows the measured GFP output in arbitrary units. The y-axis shows combinations of input signals from the truth table. The red line indicates the common output threshold between high and low output signal. In summary, single cell GFP expression level was measured to define a threshold between low and high outputs depending on the exposure to low or high control signals.
% Cells On:
Quantifies the percent of cell turned on above common output threshold depending of input strength.
Gate AND
Truth Table:
The output of GFP expression is high both ctrl1 and ctrl2 are present.
Architecture:
The figure shows a transcriptor terminal device with two transcription terminators flanked by two different opposing recombination sites. If both integrases are present and recognize the two independent opposing recombination sites the transcriptional terminators are inverted and RNA polymerase is able to flow producing an output.
Predicted:
The graph predicts that the higher the concentration of ctrl1 and ctrl2 the greater the fraction of cells that are turned ON.
Population Measurements:
Similar to the predicted fraction of cells ON, GFP expression increases as the concentration of ctrl2 and ctrl1 increases.
Single Cell:
Only when both inputs are present (ctrl1 and ctrl1) is GFP’s intensity above the common threshold. This indicates that for a strong output from the device both integrases must be present to allow transcription.
% Cells On:
When ctrl1 is present a small percentage of cell are turned ON, but when both ctrl1 and ctrl2 are present there is a high percentage of cell turned ON indicting the transcriptor element works as a Boolean AND clause.
Gate NAND
Truth Table:
The output of GFP is low only when both input signals are present.
Architecture:
The figure shows a transcriptor terminal device with an inverted transcription terminator flanked by an opposing recombination site and a constitutive promoter flanked by a different pair of opposing recombination sites. If only one of the integrases (ctrl1 or ctrl2) is present and recognizes one of the two independent opposing recombination sites the transcriptional terminator and promoter are inverted and RNA polymerase is able to flow producing an output.
Predicted:
The lower the concentration of ctrl1 and ctrl2 the greater the fraction of cells that are turned on.
Population Measurements:
Similar to the predicted fraction of cells ON, the GFP expression decrease as the concentration of the ctrl2 and ctrl1 increases. There is some variation from the predicted at low concentrations of both integrases.
Single Cell:
Only when either ctrl or ctrl2 are present at different times is GFP’s intensity above the common threshold. This indicates that for a strong output from the device only one integrase must be present to allow transcription.
% Cells On:
When ctrl1 and ctrl2 are present a small percentage of cell are turned ON, but when only ctrl1 or ctrl2 are present there is a high percentage of cell turned ON indicting the transcriptor element works as a Boolean NAND clause.
Gate OR
Truth Table:
Output of GFP is high if either or both ctrl1 or ctrl2 are present.
Architecture:
The diagram shows a transcriptor terminal device with a transcription terminator flanked by two independent aligned recombination sites. If ctrl1 or ctrl2 are present and recognize one of the two independent opposing recombination sites the transcriptional terminator and promoter is excised and RNA polymerase is able to flow producing an output.
Predicted:
At low concentrations there is a low percentage of cells turned ON, but as the concentration of both ctrl1 and ctrl2 increases to about 50% the fraction of cells ON becomes high.
Population Measurements:
Compared to the control there is a lower percentage of cells expressing GFP as the concentration of ctr1 and ctr2 increase, but the overall the trend is the same.
Single Cell:
Only when either ctrl or ctrl2 is present the GFP intensity is above the common threshold. This indicates that for a strong output from the device only one integrase must be present to allow transcription.
% Cells On:
When ctrl1 or ctrl2 are present a large percentage of cell are turned ON indicting the transcriptor element works as a Boolean OR clause.
Gate NOR
Truth Table:
Output of GFP is high is the absence of both ctrl1 or ctrl2.
Architecture:
The figure shows a transcriptor terminal device with two inverted transcription terminators flanked by opposing recombination sites. If one or both of the integrases (ctrl1 or ctrl2) are present and recognize one of the two independent opposing recombination sites the transcriptional terminators are inverted and RNA polymerase is unable to flow.
Predicted:
The lower the concentration of ctrl1 and ctrl2 the greater the fraction of cells that are turned ON. When the concentration of ctrl1 or ctrl2 is greater than 50% zero cells are turned ON.
Population Measurements:
Similar to the predicted fraction of cells ON the GFP expression decrease as the concentration of the ctrl2 and ctrl1 increases.
Single Cell:
Only when ctrl1 or ctrl2 are absent the GFP intensity is above the common threshold. This indicates that for a strong output from the device neither integerase must be present.
% Cells On:
When ctrl2 is present a small percentage of cell are turned ON, but when neither ctrl1 or ctrl2 are present there is a high percentage of cell turned ON indicting the transcriptor element works as an Boolean NOR clause.
Gate XOR
Truth Table:
Output of GFP expression is high if one or the other but not both ctrl1 and ctrl2 are present.
Architecture:
The figure shows a transcriptor terminal device with a transcriptional terminator flanked by two opposing recombination sites. If only one of the integrases (ctrl1 or ctrl2) is present and recognizes one of the two independent opposing recombination sites the transcriptional terminator is inverted and RNA polymerase is able to flow producing an output.
Predicted:
At a low concentration of ctrl1 and a high concentration of ctrl2 there is large fraction of cells turned ON and at a high concentration of ctrl1 and low concentration of ctrl2 there is high fraction of cells turned ON.
Population Measurements:
Similar to the predicted fraction of cells ON there is a high GFP expression at a low concentration of ctrl1 and a high concentration of ctrl2 and at a high concentration of ctrl1 and low concentration of ctrl2.
Single Cell:
Only when exclusively ctrl1 or ctrl2 is present the GFP intensity is above the common threshold.
% Cells On:
When ctrl1 and ctrl2 are present a small percentage of cell are turned ON, but when only ctrl1 or only ctrl2 is present there is a high percentage of cell turned ON indicting the transcriptor element works as a Boolean XOR clause.
Gate XNOR
Truth Table:
Output of GFP expression is high only when the two input signals are the same and is false if they are different.
Architecture:
The figure shows a transcriptor terminal device with an inverted transcriptional terminator flanked by two opposing recombination sites. If both or neither of the integrases (ctrl1 or ctrl2) are present and recognize one of the two independent opposing recombination sites the transcriptional terminator is inverted and RNA polymerase is able to flow producing an output.
Predicted:
At low and high concentrations of ctrl1 and ctrl2 there is a large fraction of cells ON.
Population Measurements:
Similar to the predicted fraction of cells ON at low and high concentrations of ctrl1 and ctrl2 there is high GFP expression.
Single Cell:
Only when both ctrl1 and ctrl2 are present or absent the GFP intensity is above the common threshold.
% Cells On:
When both ctrl1 and ctrl2 are present or absent a large percentage of cells are turned ON. When either crtl1 or ctrl2 are present a small percentage of cells are turned ON.
The purpose of this figure is to define a digitalization error rate or the extent to which the gate outputs are more digital than the gate control signals across small changes in the control signals. The authors are asking how the output changes in comparison to changes in input signal and if the output changes are greater than the toggle in the control.
Panel A
The panel shows the distribution of XOR gates outputs among singe cells (red) in comparison to control signal, ctrl1 (blue). Each red contour represents 5% of all cells. Thick contours surround 50% of the cell population. The y-axis shows the amount of GFP output measured in arbitrary units. The x-axis shows scattered inducer concentrations in log units. As the concentration of the control signal increases the amount of GFP expressed by single cells remains consistent with the control until the control signal switched the gate. When the control signal switched the gate at 2 ng/ml the GFP expression increased 5 fold and continued to increase with an increase in concentrations of the control one.
Panel B
The panel shows the distribution of XOR gates outputs among singe cells (red) in comparison to control signal, ctrl2 (blue). Each red contour represents 5% of all cells. The y-axis shows the amount of GFP output measured in arbitrary units. The x-axis shows the log of scattered inducer concentrations. As the concentration of the control signal increases the amount of GFP expressed by single cells remains consistent with the control until the control signal witched the gate. When the control signal switched the gate at 0.5 X 10-3 ng/ml the GFP expression increased 5 fold and continued to increase with an increase in concentrations of the control.
Panel C
This panel shows the gate switching and digitization errors across an intermediate control signal change (0.2-2 ng/ml aTc, left/right of each frame). The horizontal bars represent the gate-specific digitization thresholds used to quantify the fraction of false “high” cells given a low control signal, and vise versa. Numbers within frames are high-given-low and low-given-high error rates. Numbers above boxes are combined error rates with standard deviation from three independent experiments. AND, OR, XOR, and XNOR gates digitized ctrl1 signals to varying degrees. None of the gates reduced digitization and all gates realized digital outputs in response to low/high control signals. This indicates that there is a narrow range of amplification for the output but that a gate switch initiates changes in output.
Panel D
This panel shows the gate switching and digitization errors across an intermediate control signal change (1E-4 to 0.5E-2 arabinose, left/right of each frame). The horizontal bars represent the gate-specific digitization thresholds used to quantify the fraction of false “high” cells given a low control signal, and vise versa. Numbers within frames are high-given-low and low-given-high error rates. Numbers above boxes are combined error rates with standard deviation from three independent experiments. AND, OR, NOR, XOR, and XNOR gates digitized ctrl2 signals to varying degrees. None of the gates reduced digitization and all gates realize digital outputs in response to low/high control signals. This indicates that there is a narrow range of amplification for the output but that a gate switch initiates changes in output.
The purpose of this figure is to compare gate outputs to changes in gate control signals to determine if gates function as amplifiers of GFP expression. The authors are looking to see how the control signal is changing on the same physical basis as how the gate output is changing. The figure shows the population average response of amplifier gates. The population-average of GFP levels for each control signal and gate output were calculated. The changes in gate outputs were compared to changes in gate control signals needed to activate gate switching.
Panel A
This panel shows the changes in output GFP levels produced by gates (colored lines) directly compared to changes in ctrl1 signal required for gate switching (increasing straight dashed lines). The gray boxes represent the amplification of GFP by ctrl1 signal and the colored boxes show the gate-specific amplification of GFP compared to the ctrl1 signals. All gates increased the fold change of GFP compared the ctrl1.
Panel B
This panel shows the changes in output GFP levels produced by gates (colored lines) directly compared to changes in ctrl2 signal required for gate switching (increasing straight dashed lines). The gray boxes represent the amplification of GFP by ctrl2 signal and the colored boxes show the gate-specific amplification of GFP compared to the ctrl2 signals. All gates increased the fold change of GFP compared the ctrl2.
Panel C
This panel shows responses of inverting amplifier gates, as in A except that fold changes for control signals are inverted (decreasing straight dashed lines). All gates increased the fold change of GFP compared the ctrl1.
Panel D
This panel shows responses of inverting amplifier gates, as in B except that fold changes for control signals are inverted (decreasing straight dashed lines). All gates increased the fold change of GFP compared the ctrl2.
Genomics Page
Biology Home Page
Email Questions or Comments to cadoyle@davidson.edu
© Copyright 2013 Department of Biology, Davidson College, Davidson, NC 28035