Synthetic Biology Platform for Sensing and Integrating Endogenous Transcriptional Inputs in Mammalian Cells

Abstract

One of the goals of synthetic biology is to develop programmable artificial gene networks that can transduce multiple endogenous molecular cues to precisely control cell behavior. Realizing this vision requires interfacing natural molecular inputs with synthetic components that generate functional molecular outputs. Interfacing synthetic circuits with endogenous mammalian transcription factors has been particularly difficult. Here, we describe a systematic approach that enables integration and transduction of multiple mammalian transcription factor inputs by a synthetic network. The approach is facilitated by a proportional amplifier sensor based on synergistic positive autoregulation. The circuits efficiently transduce endogenous transcription factor levels into RNAi, transcriptional transactivation, and site-specific recombination. They also enable AND logic between pairs of arbitrary transcription factors. The results establish a framework for developing synthetic gene networks that interface with cellular processes through transcriptional regulators.

Graphical abstract

Figure 1

Characterization of Open-Loop Transcriptional Sensors (A) Schematics of a sensor screening assay in HEK Tet-On cells. TATA indicates a core minimal promoter. DNA constructs and protein products are shown. Pointed arrow indicates transactivation. (B) Schematics of the HNF1A/B, HNF4A, and SOX9/10 sensors. The sequences of 1× REs and a spacer (sp) are shown. Only top strands are shown for HNF1A/B and HNF4A. (C) TF sensor population-averaged responses in HEK Tet-On cells in the presence and absence of Dox, as indicated, and in the HuH-7 cell line. Each bar represents mean ± SD of biological triplicates. (D) Mutual crosstalk between the TFs (y axis) and sensors (x axis). The color code indicates averaged AmCyan intensity. rtTA, reverse tetracycline-responsive transcriptional activator; CMV, cytomegalovirus promoter; TetOn, HEK293 Tet-On cells.

Figure 2

Initial Characterization of TF Sensors with Positive Feedback (A) Schematics of a feedback-amplified TF sensor. (B) Induction levels measured in HEK Tet-On cells and in HuH-7 cells compared to their open-loop counterparts (values reproduced from Figure 1). TF and cell line names are indicated. Open-loop and amplified sensor outputs are directly comparable, but the amplified output signal is underestimated by a factor of five due to the 2A sequence (Figure S3B). FB, feedback; TetOn, HEK293 Tet-On cells. Each bar represents mean ± SD of biological triplicates. (C) Crosstalk between the feedback-amplified sensors and different input TFs.

Figure 3

Characterization of the Composite Promoters (A) Experimental layout. Both the ectopic mammalian TF and the amplifier activator are induced with Dox. Different combinations of transcriptional inputs are achieved by withholding either the TF- or the amplifier activator-expressing cassette. (B) Schematics of different transcriptionally active complexes. DBD, DNA binding domain; AD, activation domain. (C) Expression levels reached with either or both transcriptional inputs provided to the promoter, as indicated. Different REs are compared. (D) Effect on synergy of swapping the DBD of the amplifier activator from PIT to ET for three REs of HNF4A. (E) Effect on synergy of replacing the TF input and its RE without changing the amplifier activator ET-VP16. (F) Effect on synergy of swapping the minimal promoter from the minimal TATA box to CMV_MIN. (G) Comparison between feedback-amplified sensors harboring a minimal TATA box (top) and those harboring CMV_MIN (bottom) with and without the TF input. 1×, 2×, and 3× HNF1A/B REs are compared. In all panels, each bar represents mean ± SD of biological triplicates.

Figure 4

Analysis of Synergy Data Synergy is calculated as a ratio of the expression in the presence of both TF and amplifier activator, divided by the sum of expression with TF only and amplifier activator only. To increase accuracy, TF-only data from Figure 1 is used for calculations (Supplemental Experimental Procedures). (A) Transactivating efficiency of PIT-RelA amplifier activator as a function of PIR distance from the TATA box. The intervening RE for the endogenous TF input coupled to the low-leakage promoter (color coded) and the sequence of a minimal promoter (minimal TATA box versus CMV_MIN, inset) are compared. (B) Fully induced expression levels from a composite promoter achieved by providing the ectopic TF input and PIT-RelA, as a function of PIR distance from the TATA box. (C) Synergy scores for different composite promoters with PIT-RelA as amplifier activator, as a function of PIR distance from the TATA box. (D) Schematics of the synergistic positive-feedback amplifier. (E and F) Simulated (E) and experimental (F) responses of open- and feedback-amplified sensors to varying TF input. Comparison to an amplified loop without synergy is shown. (G) Simulated open-loop and amplified response dependency on the binding affinity of the TF input. (H) Simulated relationship between open- and amplified-loop responses. (I) Experimental correlation between open- and closed-loop responses for HNF1A/B and SOX9/10 constructs in HEK Tet-On cells. (J) Simulated off and on responses of feedback-amplified loops as a function of synergy (left) and on-to-off ratio as a function of synergy (right). (K) Experimental on-to-off ratio of feedback-amplified sensors as a function of synergy (black dots). The straight line is a linear fit forced through zero. In (A), (B), and (F), error bars represent SD of biological triplicates.

Figure 5

Forward Sensor Engineering (A–C) For each TF, we show the response of the composite promoter in open-loop configuration (top) and the response of the corresponding feedback loops (middle) and then compare the distance dependency of PIT-RelA transactivation with the synergized expression (bottom). The cofactors used for induction are indicated. For NFATC1, induction means expression of the TF cDNA combined with ionomycin. (D) Synergy for the composite promoters tested in this panel as a function of PIR distance from the TATA box. (E) On-to-off ratio as a function of synergy for this dataset (red), overlaid on Figure 4K data (gray). In (A)–(D), error bars represent SD of biological triplicates.

Figure 6

AND Logic on the Composite Promoter (A) Experimental setup for independent input modulation and the logic circuit abstraction of this setup (left). Promoter response as a function of varying levels of the TF and the amplifier activator inputs (right). Abx, antibiotic; PI, pristinamycin; ET, erythromycin. (B) Schematics of an AND gate between two ectopic transcriptional inputs. (C) AND gate between SOX10 and HNF1A. (D) AND gate between HNF4A and SOX10. Left, the gate that used PIT-RelA. Right, the same gate employing PIT-VP16. (E) AND gate between HNF1A/B and SOX9/10 discriminates between cell lines using endogenous levels of TF inputs. The bar chart shows AmCyan output levels in different cell lines (each bar represents mean ± SD of biological triplicates), and the micrographs show the expression of transfection marker mCherry (red) and the AmCyan output (green). Scale bars, 100 μm.

Figure 7

Transduction of Transcriptional Activity into Various Downstream Processes (A) Multipronged actuation combining downstream RNAi with transactivation, including a low-leakage RNAi target with recombinase delay and a transactivation target, PIR-driven mCitrine. (B) The improvement in RNAi knockdown with the delayed switch, comparing between on and off mCherry readouts in the delayed and standard configurations (top) and flow cytometry plots (bottom). (C) The outputs generated by the multipronged sensor in three cell lines triggered by the endogenous HNF1A/B activation. The rows show representative images of AmCyan (cyan), Citrine (yellow), and mCherry (red). Transfection marker infrared fluorescent protein (iRFP; purple) is in a separate row. Scale bars, 100 μm. Note the low transfection efficiency in HeLa cells. (D) Schematics of TF-driven, feedback-amplified expression of Cre recombinase. (E) Response of two amplified Cre sensors to ectopic induction of their respective cognate TF (HNF1A and SOX10) (left) and corresponding flow cytometry plots (right). (F) Endogenous TF-driven, recombinase-triggered gene inversion in three cell lines. The figure shows the response of the sensor for HNF1A/B (top) and SOX9/10 (bottom). Representative images are shown. The Citrine signal (yellow) is represented over the corresponding image for the iRFP transfection marker (purple). Scale bars, 100 μm. HeLa cell images have higher background when similar settings are applied to all images in this panel. In (B), (C), (E), and (F), each bar represents mean ± SD of biological triplicates.