Integrating recombinase-based feedback and feedforward control for optimal resource decoupling

Rixin Zhang; Rong Zhang; Xiao-Jun Tian

doi:10.1093/nar/gkaf913

. 2025 Sep 17;53(17):gkaf913. doi: 10.1093/nar/gkaf913

Integrating recombinase-based feedback and feedforward control for optimal resource decoupling

Rixin Zhang ¹, Rong Zhang ², Xiao-Jun Tian ^3,^✉

PMCID: PMC12448909 PMID: 40966512

Abstract

Resource competition disrupts circuit modularity by introducing unintended coupling between otherwise independent gene modules, thereby compromising genetic circuit function. While various control strategies have been explored, their complexity or limited efficacy have hindered broader application. Here, we present the Re-NF-FF-Controller, a recombinase-based strategy that integrates negative feedback and feedforward regulation via promoter flipping to mitigate resource competition. Computational modeling and experimental validation demonstrate that Re-NF-FF-Controller effectively reduces resource coupling, ensuring robust gene expression and modularity. Moreover, its tunability allows for performance optimization through straightforward adjustments of recombinase enzyme levels. This strategy offers a versatile and easily implementable solution for designing reliable synthetic biological systems.

Graphical Abstract

Introduction

Synthetic biology offers a bottom-up approach for constructing genetic circuits to achieve tailored functions by leveraging previously characterized modules [1]. However, emergent or unintended behaviors often arise when multiple gene modules are assembled, making the engineering process lengthy and tedious [2–5]. A key challenge arises from the finite pool of cellular resources—competition for these resources can disrupt the intended function of individual modules, leading to unpredictable outcomes that deviate from the original design [6–9]. This resource limitation not only introduces uncertainty but also weakens the modularity principle that underpins synthetic biology.

To restore the modularity of genetic circuits, various methods have been proposed [2, 4, 10]. Using orthogonal resources can compartmentalize different gene modules into distinct groups, thereby reducing resource coupling by utilizing separate resource pools [11–13]. Resource reallocation is another strategy that allows for dynamic allocation of resources to different gene modules as needed, ensuring stable expression of the desired modules and enhancing the robustness of the gene circuit [14, 15]. Negative feedback and feedforward loops are also utilized for resource competition control [16–23]. For example, Darlington et al. developed a resource reallocator in which a negative feedback controller was incorporated [11]. By sensing changes in circuit demand, it adjusts the available ribosomes to maintain robustness. Shopera et al. introduced negative feedback into the operating system to decouple resource-coupled gene expression [24]. Additionally, Barajas et al. constructed a feedforward growth rate controller by co-expressing SpoTH with a gene of interest (GOI), where SpoTH played a role in replenishing ribosomes sequestered by GOI expression [25]. However, these strategies often offer limited improvements or rely on complex designs, reducing their versatility and broader applicability.

Site-specific recombinases modify DNA sequences by recognizing and acting on defined binding sites [26]. Depending on the orientation of the recombinase sites, recombination can occur in three ways: integration, excision, and inversion. While recombinase-mediated integration is commonly used for DNA assembly, its excision and inversion capabilities enable precise logical control in genetic circuits. For example, recombinase has been used to construct a rewriteable, recombinase-addressable data module, which reliably records digital information by fine-tuning the synthesis and degradation rates of recombinase proteins [27]. Recombinase-based negative feedback controllers (Re-NF-Controllers) have been developed to reduce gene expression variability [28]. Additionally, recombinase enzymes have been employed to create a feedback controller capable of near-perfect adaptation, with three distinct operational modes that respond to external signals, taking advantage of the switching properties of recombinase enzymes [29].

In this work, we harnessed the inversion mechanism of site-specific recombinases to dynamically regulate resource allocation across a range of resource-limited cellular conditions. We first implemented a Re-NF-Controller and demonstrated its effectiveness in mitigating resource coupling. Building on this, we incorporated a feedforward control layer using the same promoter flipping, which not only improved decoupling performance, particularly in high-copy plasmid systems, but also introduced greater tunability. Together, these recombinase-based strategies offer a robust and modular framework for reducing resource interference and restoring gene circuit modularity.

Materials and methods

Strains, media, and chemicals

Escherichia coli strains DH5α and K-12 MG1655ΔlacIΔaraCBAD were used for plasmid construction and gene circuit induction, respectively. DH5α colonies were inoculated into 5 ml of LB broth supplemented with 25 μg/ml chloramphenicol (LBC), 50 μg/ml kanamycin (LBK), or 100 μg/ml ampicillin (LBA), depending on the plasmid backbone. For plasmid extraction, cultures were grown in 15 ml tubes at 37°C with shaking at 250 rpm. For circuit induction and measurement, cells were induced by l-ara (l-(+)-arabinose, Sigma–Aldrich) in a 96-well plate. l-ara was dissolved in ddH₂O into a stock concentration of 25%, then diluted into appropriate working solutions for induction. The induction scheme is detailed further down in the “Materials and methods” section.

Plasmids construction

Gene circuits were constructed using either the pSB1A2, pSB1C3, or pSB3K3 plasmid backbones. The BioBrick parts used for circuit assembly are listed in Supplementary Table S1, and the sequences of specific parts are provided in Supplementary Table S2. The parts were digested as designed via restriction sites for EcoRI, XbaI, SpeI, and PstI (restriction enzymes from Thermo Fisher), then purified by polymerase chain reaction (PCR) cleanup (GenElute PCR Cleanup Kit from Sigma–Aldrich) or gel electrophoresis (GelElute Gel Extraction Kit from Sigma–Aldrich), depending on the length of the products. The digested parts were ligated using T4 DNA ligase (New England BioLabs) and then transformed into E. coli strain DH5α. Transformants were screened by colony PCR and cultured overnight. Plasmids were subsequently extracted using the GenElute Plasmids Miniprep Kit (Sigma–Aldrich) and verified by restriction digest (EcoRI and PstI). Circuit information and design details are provided in Supplementary Table S3, and the full sequence of circuit C53 is given as an example in Supplementary Table S4. The plasmid sequences are provided in the Supplementary data. All final constructs were confirmed by whole plasmid sequencing.

Circuit inductions

A single colony of E. coli strain K-12 MG1655ΔlacIΔaraCBAD was inoculated into 300 μl LB medium and supplemented with 25 μg/ml chloramphenicol, 50 μg/ml kanamycin, or 100 μg/ml ampicillin, depending on the plasmid backbone. Cultures were grown in 5 ml tubes at 37°C with shaking at 250 rpm for 5 h. Subsequently, 5 μl of the culture was transferred to each well of a 96-well plate containing 200 μl of LB medium and supplemented with the appropriate antibiotic (LBAnt), with three biological replicates per condition. Twenty-five percent l-ara was diluted into a gradient of 8.3%, 5%, 2.5%, 0.625%, 0.25%, 0.125%, and 0% to prepare the working solutions. Each of these was then diluted 1000-fold into the corresponding medium to induce the gene circuits. Plates were incubated overnight at 37°C with shaking at 250 rpm to allow gene expression to reach a steady state.

Following induction, optical density at 600 nm (OD) and fluorescence intensity were measured using a Synergy H1 Hybrid Reader (BioTek). Fluorescent proteins were detected using the following excitation/emission wavelengths: green fluorescent protein (GFP) (485/515 nm), red fluorescent protein (RFP) (546/607 nm), and cyan fluorescent protein (CFP) (438/485 nm). Wells containing 200 μl of LBAnt medium without cells were used as blank controls. To account for differences in cell density, fluorescence values were calculated by dividing the background-subtracted fluorescence by the corresponding OD. To facilitate comparison of resource coupling effects across different samples, GFP or CFP fluorescence values were normalized to their baseline expression levels in the absence of l-ara induction, reflecting relative expression trends. RFP fluorescence values were normalized to their respective maximum expression levels to enable comparison across different circuits. To evaluate cellular burden, cell growth was monitored dynamically by measuring OD. Instead of overnight culturing in a shaker, the assembled plate was incubated in a plate reader with orbital shaking at 807 CPM (circles per min) at 37°C. OD was recorded every 20 min.

Time-course analysis of promoter flipping

Plasmids carrying circuits C53 and C55 were transformed into E. coli K-12 MG1655ΔlacIΔaraCBAD. Colonies were inoculated into 2 ml of LBK medium and cultured overnight at 37°C with shaking at 250 rpm. Circuits C53 and C55 were induced with 5 × 10⁻³% l-arabinose at 0, 3, 6, 9, and 12 h. Finally, plasmids were extracted using the GenElute Plasmids Miniprep Kit (Sigma–Aldrich). Plasmids that underwent flipping into the attBP configuration contained the sequence from the Pbad promoter to the GFP gene, which was amplified by PCR using the primers: forward: 5′-ccgcttctagagacattgattatttgcacggcgtcacactttg-3′ and reverse: 5′-ggaacaggtagttttccagtagtgc-3′. Plasmids that remained in the attLR configuration contained the sequence from the Pbad promoter to the RFP gene, amplified using the same forward primer and a different reverse primer: 5′-tctggaattcgcggccgcttctagagttattaagcaccggtggagtg-3′. PCR with 16 cycles was performed to compare product concentrations at different time points. PCR products were loaded onto a 1% agarose gel, and electrophoresis was carried out in TAE buffer at 100 V for 1 h. The gel was imaged using Syngene PXi, and band intensities were analyzed using the Gel Analyzer tool in FIJI (NIH) [30]. The values were quantified relative to the baseline band intensity at 0 h.

Microscopy

To test the flipping efficiency of integrase, gene circuits C5 and C6 were induced with 6.25 × 10⁻⁴% and 5 × 10⁻³% l-ara, following a protocol similar to the one described earlier, except that after the 5-h incubation step, 2 μl of cell culture was added to a 15 ml culture tube containing 2 ml LBK. Then, the cells were induced with 6.25 × 10⁻⁴% and 5 × 10⁻³% l-ara overnight at 37°C in a shaker set to 250 rpm. Two microliters of cell culture was placed at the center of a 2.4 cm × 5 cm No. 1 coverslip and covered with a piece of agarose gel (1 cm × 1 cm × 0.2 cm) made with PBS to immobilize the cells. Phase, GFP, and RFP images were taken with a 100× oil objective on a Nikon Eclipse Ti inverted microscope (Nikon), equipped with an LED-based Lumencor SOLA SE and filter sets: 460–500 nm excitation/510–560 nm emission for GFP, and 532–587 nm excitation/608–683 nm emission for RFP.

Flow cytometry

All samples were analyzed using Stratedigm S1000EON Flow Cytometer with excitation/emission filters 480 nm/530 nm (FL1-A) for GFP detection and 480 nm/>670 nm (FL3-A) for RFP after overnight induction. Each sample was represented by three biological replicates, with 10,000 events recorded per replicate for analysis. The generated data was further analyzed using MATLAB (R2024b, MathWorks).

Coupling index

To quantify the coupling between GFP and RFP expression, we defined a coupling index (CI) as the average deviation of the GFP–RFP response curve from an ideal flat line representing complete resource decoupling. GFP levels were normalized to their baseline expression without RFP module induction, and RFP levels were normalized to their maximum expression. Experimental data were fitted using a piecewise cubic Hermite interpolating polynomial, and CI was computed as the average difference between the fitted curve and the ideal value of 1. For simulations, the density of data points allowed direct CI estimation without fitting. A CI of 0 indicates perfect decoupling, while negative or positive values reflect negative or positive coupling, respectively.

Parameter sensitivity analysis

To evaluate the sensitivity of recombinase-based resource controllers to parameter variations, we conducted both local and global sensitivity analyses. The local sensitivity analysis was performed to assess how variations in individual kinetic parameters affect controller performance by increasing or decreasing each parameter by 20% relative to its baseline value. The global sensitivity analysis was carried out to evaluate the impact of simultaneous variation in all parameters. A total of 1000 random parameter sets were generated using the Latin hypercube sampling method [31], with all parameters varying within ±20% of their nominal values. For the Re-NF controller, the degradation rate of excisionase (d_x) was set to its baseline value of 0.01. For the Re-NF-FF controller, d_x was set to 0.0216, which corresponds to the optimized value determined in Fig. 5D for achieving optimal performance.

Figure 5. — Additional feedforward control significantly enhances resource decoupling. (A) Schematic diagram of genetic circuits C53, C55, C57, and C59. Circuit C53 incorporates the Re-NF-FF-Controller, while C55 represents its corresponding open-loop system. Circuit C57 features a modified integrase start codon (ATG to GTG), resulting in reduced integrase expression. Circuit C59 includes a degradation tag (AAK) added to excisionase, enabling rapid degradation of excisionase. (B) Simulated steady-state GFP expression levels as a function of RFP expression (across varying inducer concentrations) and excisionase degradation rate (d_x). The boxed region highlights the effective resource decoupling achieved by the Re-NF-FF-Controller. (C) The CI between GFP and RFP as a function of the degradation rate of excisionase (d_x). (D) The optimal resource decoupling effect achieved by tuning the degradation rate of excisionase (black line) to correct the overcompensation (red line). The open-loop system (blue line) is shown as a reference. (E) Correlations between GFP and RFP expression levels in circuits C53, C55, and C59 across different l-ara concentrations: 0%, 1.25 × 10⁻⁴%, 2.5 × 10⁻⁴%, 6.25 × 10⁻⁴%, 2.5 × 10⁻³%, and 5 × 10⁻³%. (F) Correlations between GFP and RFP expression in high-copy circuits C53-H, C59-H, and C55-H. Induction conditions are as described in panel (E). Solid lines represent linear fits to experimental data; data points and error bars indicate mean ± SD (n = 3). GFP levels were normalized to their baseline expression without RFP induction, and RFP levels were normalized to their maximum expression. (G) CI calculated from experimental data, quantifying the resource decoupling performance of Re-NF and Re-NF-FF controllers under various conditions.

Mathematical models

Mathematical models based on ordinary differential equations (ODEs) were developed to describe and analyze system dynamics without a controller, as well as with the Re-NF-Controller or Re-NF-FF-Controller. Steady-state simulations were performed to assess GFP expression dependence on the RFP module across all scenarios. Further details are provided in the Supplementary data.

Results

Design of recombinase-based control strategy for minimizing resource coupling

In this study, we introduce a recombinase-based control strategy to minimize resource coupling between two seemingly unconnected modules in the cellular environment. For simplicity, two genes are designed in the synthetic gene circuit, GFP driven by a constitutive promoter and RFP controlled by an inducible promoter (Fig. 1A). In an ideal scenario with unlimited cellular resources, these two genes should be expressed independently. Specifically, GFP expression level should remain unchanged regardless of RFP expression level, resulting in a flat interdependence curve (red line, Fig. 1B). However, cellular resources are inherently limited, resulting in significant competition between the two genes and unintended coupling. Upon induction, the RFP expressing module sequesters a fraction of resources, leading to a reduction in GFP expression. This resource-driven interference is reflected as a negative correlation between the expression of the two genes (blue line, Fig. 1B).

Figure 1. — Design of a recombinase-based controller. (A) Schematic of the synthetic gene circuit, where GFP is driven by a constitutive promoter and RFP is controlled by an inducible promoter. The orientation of the inducible promoter is regulated by recombinase elements, integrase, and excisionase. Excisionase, co-expressed bicistronically with RFP, forms a complex with constitutively expressed integrase to flip the promoter by converting attL and attR sites into attB and attP sites, respectively. Integrase alone reverses this process, restoring the original promoter orientation. Re-NF-Controller and Re-NF-FF-Controller are distinguished by whether the flipped promoter drives GFP expression, with only the latter inducing GFP upon promoter inversion. (B) The extent of resource coupling is assessed by examining the relative expression of RFP and GFP. A negative correlation (blue line) indicates resource competition-driven coupling, whereas a flat relationship (red line) represents an ideal case where the controller effectively eliminates resource competition.

To address this issue, we redesigned the RFP expressing promoter by flanking it with recombinase attachment sites, specifically attL (attachment Left) and attR (attachment Right). In addition, we introduced both integrase and excisionase genes. While the former is constitutively expressed, the latter is co-expressed with RFP in a bicistronic manner, making its expression proportional to that of RFP. When the inducer concentration is high, the RFP module is strongly expressed and imposes a significant burden on cellular resources, and the expression level of excisionase also increases. The excisionase then forms a complex with integrase, reversing the promoter’s direction through cleavage, reversion, and re-ligation of the attachment sites, converting attL and attR into attB (attachment Bacteria) and attP (attachment Phage) [31, 32], respectively, as shown in Fig. 1A. As a result, the fraction of promoters oriented to drive RFP expression is reduced, automatically limiting RFP and excisionase’s expression as resource competition intensifies. This process reverses as excisionase level declines and the excisionase–integrase complex diminishes, allowing free integrase to catalyze the conversion of the attB and attP (BP) sites back to the attL and attR (LR) sites [26]. This bidirectional design enables a negative feedback mechanism, termed the Re-NF-Controller, which dynamically and reversibly adjusts gene expression in response to changing cellular resource levels.

Additionally, placing GFP downstream of the flipped promoter (flanked by attB/attP) enables its transcription to be enhanced by a dual-promoter effect [32, 33], compensating for the reduced expression caused by resource competition. This approach naturally establishes a feedforward mechanism using the flipped promoter without introducing additional genes that could further burden the system. By seamlessly integrating both feedback and feedforward regulation through a single promoter flip (referred to as Re-NF-FF-Controller thereafter), we anticipate that this new control strategy will effectively mitigate resource competition while offering a high level of tunability.

Evaluation of promoter flipping efficiency by integrase and excisionase

Bacteriophage Bxb1 is a temperate phage of Mycobacterium smegmatis and encodes the best-characterized serine integrase–excisionase system. The Bxb1 integrase is a member of the serine recombinase family and mediates recombination through short binding sites (<50 bp), without the need for complex higher order macromolecular structures. The Bxb1 excisionase has been shown to control recombination directionality with high efficiency [27, 34, 35]. We first constructed two circuits, C5 and C6, to assess Bxb1 integrase activity in our system (Fig. 2A). Circuit C6 is designed to switch from GFP to RFP expression upon induction through the promoter flipping mediated by integrase, which is absent in the reference circuit C5. GFP is placed downstream of the Pbad promoter, flanked by attB and attP sites, which serve as recognition sites for integrase-mediated promoter flipping. In the presence of the inducer l-ara, the transcription factor AraC activates the Pbad promoter, driving GFP expression. However, in circuit C6, integrase is also expressed, leading to promoter inversion and subsequent RFP expression. Induction experiments showed GFP expression in C5, whereas C6 exhibited a marked increase in RFP expression, confirming the integrase-mediated promoter reversion, as demonstrated by the fluorescence microscopy (Fig. 2B). Flow cytometry results further validated these observations, showing increased GFP expression in circuit C5 and increased RFP expression in circuit C6 when the circuits were induced with a higher concentration of inducer (Fig. 2C). These findings confirm that integrase can reliably flip the promoter flanked by BP sites to LR sites through site-specific recombination. The heterogeneous GFP expression observed in Fig. 2B is likely due to the use of degradation-tagged fluorescent proteins, as previously reported [36], where protein degradation exceeded synthesis after prolonged culture as arabinose became depleted. Here, degradation tag was used to ensure accurate detection of promoter flipping while minimizing potential artifacts from basal (leaky) expression.

Figure 2. — Evaluation of the flipping efficiency by the recombinase enzymes. (A) Schematic of genetic circuits C5 and C6 for testing integrase activity. Circuit C6 switches from GFP to RFP expression upon induction via promoter flipping by integrase, which is absent in the reference circuit C5. (B) Fluorescence microscopy images showing flipping efficiency. Representative results from three replicates after overnight culture with 5 × 10⁻³% l-ara concentration are shown. (C) Flow cytometry data show cell state transitions in C5 and C6 with increasing concentration of l-ara: 0%, 6.25 × 10⁻⁴%, and 5 × 10⁻³%. Representative results from three replicates are shown. (D) Schematic of genetic circuits C21, C24, and C25 for testing recombinase activity, with integrase constitutively expressed. Circuit C21 is the reference without recombinase. Circuit C25 uses a weaker start codon for integrase compared to circuit C24. (E) GFP expression levels in circuits C21, C24, and C25 under varying l-ara concentrations (0%, 6.25 × 10⁻⁴%, and 5 × 10⁻³%). Data are shown as means ± SD (n = 3).

The conversion of LR sites back to BP sites is facilitated by excisionase, which forms a complex with integrase to specifically recognize and mediate site-specific recombination. To examine the efficiency of the excisionase-directed recombination in our system, the Pbad promoter was designed to be flanked by LR sites, driving RFP expression (Fig. 2D). In reference circuit C21, excisionase was co-expressed with RFP in a bicistronic arrangement upon l-ara induction. However, without integrase, the directionality of the Pbad promoter remains unchanged. In circuit C24, constitutive integrase expression allows excisionase to form a complex with it, facilitating the conversion of LR sites back to BP sites. This recombination event reorients the Pbad promoter, leading to GFP expression instead of RFP. As shown in Fig. 2E, upon induction, GFP levels increased significantly in circuit C24 compared with circuit C21. To further assess the role of integrase, we constructed circuit C25, which contains a weaker start codon for integrase than that in circuit C24 (Fig. 2E). The results show that circuit C25 exhibited a more moderate increase in GFP expression upon l-ara induction compared to circuit C24, emphasizing the necessity of fine-tuning recombinase levels to achieve optimal regulatory control.

Recombinase-based negative feedback reduces resource coupling

To assess the resource decoupling capability of the Re-NF-Controller, we designed and constructed two circuits: C44 and C52 (Fig. 3A). In both circuits, GFP is expressed under a constitutive promoter, while RFP is regulated by the Pbad promoter. In circuit C44, negative feedback is implemented by co-expressing excisionase with the RFP module, along with a constitutively expressed integrase. Together, these enzymes can flip the orientation of the Pbad promoter, modulating RFP expression. In contrast, circuit C52 serves as a reference control, where a mutation in the dinucleotide sequence of the attR site prevents recombinase-mediated flipping of the Pbad promoter, thereby eliminating the feedback mechanism. The central dinucleotide of attL and attR functions as a proofreading element during recombination and is the key determinant of substrate identity [37]. Based on this, the specific dinucleotide sequence TG/AC in the attR sites was mutated to AC/TG. While recombinase can still recognize and excise both dinucleotides at the LR sites, the excised dinucleotide at the attL site (CA/GT) can only be re-ligated with its complement pair TG/AC, but not AC/TG, at the attR site to form the BP sites. In other words, the mutation disrupts the promoter reversion process in circuit C52 by creating a mismatch in the central dinucleotide sequences at the LR sites. As a result, the DNA fragments are unable to recombine into the BP configuration and instead re-ligate back to their original arrangement [37].

Figure 3. — Recombinase-based negative feedback reduces resource competition. (A) Schematic representation of the open-loop circuit C52 and the genetic circuit C44 incorporating the Re-NF-Controller. C52 is derived from C44 by mutating the dinucleotide sequence at the attR site, preventing promoter flipping while maintaining the overall circuit topology. Spacer sequence (pink part) was added between attL and promoter Pbad to avoid interference. (B) Simulated steady-state GFP expression levels as a function of RFP expression (across varying inducer concentrations) and negative feedback strength (v_off). (C) Simulated CI between GFP and RFP decreases as the negative feedback strength increases in the Re-NF controller. (D) Comparison of the GFP–RFP correlation between the open-loop system (v_off= 0) and the system with Re-NF-Controller (v_off= 0.75). (E) Correlations between GFP and RFP expression levels in circuits C52 and C44 across different l-ara concentrations: 0%, 1.25 × 10⁻⁴%, 2.5 × 10⁻⁴%, 6.25 × 10⁻⁴%, 2.5 × 10⁻³%, and 5 × 10⁻³%. Solid lines represent linear fits to experimental data, while data points and error bars indicate mean ± SD (n = 3). GFP levels were normalized to their baseline expression without RFP induction, and RFP levels were normalized to their maximum expression.

To investigate how negative feedback mitigates resource-mediated coupling, we developed a mathematical model (see the “Materials and methods” section and Supplementary data for details). We conducted simulations to examine how GFP levels depend on both the strength of negative feedback in the Re-NF-Controller (v_off) and the inducer dose for the RFP module (l-ara). As shown in Fig. 3B, in the absence of feedback (v_off = 0), corresponding to the open-loop system, GFP levels decrease sharply with increasing l-ara, indicating significant resource competition. However, as the negative feedback strength (v_off) increases, the dependence of GFP on l-ara weakens, reflecting reduced resource competition. To quantify this effect, we defined a CI (coupling index) as the mean change in GFP levels relative to 1. A CI of 0 indicates perfect decoupling, while a negative value suggests negative coupling. Figure 3C shows that the CI, calculated from the simulation data, is negative and increases with feedback strength until it reaches a plateau. This suggests that negative feedback can effectively reduce the dependence of GFP expression on RFP expression within a certain limit. Figure 3D further illustrates the relationship between GFP and RFP levels at the optimal decoupling condition. To experimentally validate this, we measured GFP and RFP expression after overnight induction of two circuits with varying l-ara concentrations. As shown in Fig. 3E, the GFP expression in C44 exhibited a weaker dependence on RFP levels compared to circuit C52, consistent with our modeling results. To evaluate the performance of the Re-NF-Controller under heightened resource competition, we constructed circuits C44 and C52 on a high-copy plasmid, designated C44-H and C52-H, respectively. The data show that its resource decoupling effect is diminished, revealing a limitation of this controller (Supplementary Fig. S1). To further validate this, we also constructed another reference circuit, C42, lacking recombinase genes, with RFP expressed under the Pbad promoter and GFP driven by a constitutive promoter (Supplementary Fig. S2A). While the GFP level as a function of RFP exhibits a flatter slope in circuit C44 compared to circuit C42 (Supplementary Fig. S2B), the decoupling efficiency of circuit C44 compared to C42 is not as pronounced as that observed when comparing C44 to circuit C52 (Fig. 3E). This is likely due to the resource consumption by controller genes, such as integrase and excisionase, in circuits C44 and C52. To experimentally validate this observation, we evaluated the cellular burden imposed by the controller genes by measuring the growth rates of strains harboring circuits C42b and C52, which share the same gene arrangement (Supplementary Fig. S3A). As shown in Supplementary Fig. S3B, reduced growth rate in C52 suggests that integrase and excisionase impose an unavoidable burden. This is further supported by simulation results showing that resource coupling is reduced when the resource consumption of integrase and excisionase is set to be negligible (Supplementary Fig. S4). This highlights a key design trade-off where controller genes improve system robustness but inevitably impose a resource burden [16, 19, 24]. The unnormalized expression levels of RFP in circuits C42 and C44 are presented in Supplementary Fig. S5. As expected, RFP expression is lower in the recombinase-mediated circuit C44 compared to the reference circuit C42, due to the negative feedback mechanism. For the comparison between circuits C52 and C44, the corresponding unnormalized RFP levels are shown in Supplementary Fig. S6. No significant difference in RFP expression was observed between these two circuits. Overall, theoretical and experimental analyses show that the Re-NF-Controller reduces resource coupling within certain limits.

Terminator malfunction causes unanticipated positive correlation in gene expressions

We also designed and constructed an alternative Re-NF-Controller topology by arranging gene placement to evaluate its function. In the new circuit, C43, the GFP module is positioned downstream of the attR site, separated by a terminator B0015 (Fig. 4A). Surprisingly, this change in gene placement results in a completely different gene expression pattern. As shown in Fig. 4B, GFP expression in circuit C43 now exhibits a positive correlation with RFP level, in contrast to the negative correlation observed previously, suggesting a change in the dynamics of resource allocation toward the GFP module.

Figure 4. — Malfunction of the terminator leads to overcorrection in resource decoupling. (A) Schematic of genetic circuit C43, an alternative Re-NF-Controller design where the GFP module is placed downstream of the attR site. Circuit C42 serves as a reference without recombinase elements. Spacer sequence (pink part) was added between attL and promoter Pbad to avoid interference. (B) Correlations between GFP and RFP expression levels in circuits C42 and C43 across different l-ara concentrations: 0%, 1.25 × 10⁻⁴%, 2.5 × 10⁻⁴%, 6.25 × 10⁻⁴%, 2.5 × 10⁻³%, and 5 × 10⁻³%. Solid lines represent linear fits to experimental data, while data points and error bars indicate mean ± s.d. (n = 3). (C) Schematic diagram of genetic circuits C45 and C46, incorporating constitutively expressed CFP as an independent indicator to assess resource coupling in the system. The correlation between GFP (D) or CFP (E) expression level and RFP expression level in circuits C45 and C46. The experimental conditions were the same as those described in panel (B). (F) Schematic diagram of genetic circuits C73 and C74, incorporating double terminators between the attR site and the GFP module. (G) The correlations between GFP and RFP expression levels in C73 and C74. The experimental conditions were the same as those described in panel (B). GFP or CFP levels were normalized to their baseline expression without RFP induction, and RFP levels were normalized to their maximum expression. The malfunctional B0015 terminators located downstream of the attR site are labeled in red to indicate the incomplete termination of its downstream GFP modules. Functional, modified dual terminators (B0015–B1002) are labeled with a double-bar symbol. Terminators shown in black are functional.

To validate the issue of placement dependency, we introduced an additional fluorescence reporter gene, CFP, into circuits C42 and C43, positioned far from the attR sites, resulting in circuits C45 and C46 (Fig. 4C). Experimental data show that GFP levels still increase abnormally with RFP expression in circuit C46 (Fig. 4D). In contrast, CFP levels display the expected negative correlation with RFP expression (Fig. 4E). These distinct expression patterns between CFP and GFP may strongly stem from the upstream spacer, which was added in front of the Pbad promoter in circuits C42 and C43, or from the loss of terminator function downstream of the attR site.

To assess whether the effect is caused by the spacers, we constructed circuits C34 and C35 with the spacers removed (Supplementary Fig. S7A). A similar induction pattern was observed (Supplementary Fig. S7B), except for differences in the relative expression of RFP and GFP. Upon spacer removal, the differential RFP expression between the unregulated and recombinase-regulated circuits is enhanced (Supplementary Fig. S7C) and GFP expression is reduced (Supplementary Fig. S7D). These data suggest that, while the spacer helps mitigate resource competition, it is not responsible for the abnormal relationship between GFP and RFP. We hypothesized that the abnormal expression pattern of GFP is associated with promoter flipping mediated by the integrase–excisionase complex. Therefore, we tested whether adding a fast degradation tag to excisionase would affect the circuit’s behavior (Supplementary Fig. S8A). We did not observe the increasing expression pattern of GFP after modification of excisionase; correspondingly, the resource coupling became unaffected (Supplementary Fig. S8B and C), indicating that the flipped promoter provided an additional transcriptional signal to the downstream GFP module, which was supposed to be terminated by our design. To validate our hypothesis, we further tested the functionality of terminator B0015 by constructing two simple circuits (C65 and C66) with GFP placed downstream of the flipped Pbad promoter, either with or without the terminator, respectively (Supplementary Fig. S9A). As shown in Supplementary Fig. S9B, no difference was observed in the induced GFP expression pattern as a function of l-ara dose between two circuits, suggesting that terminator B0015 had lost its function after the attR site.

To resolve this issue, we inserted an additional short terminator, B1002, downstream of the existing B0015 terminator, generating circuit C70 (Supplementary Fig. S9A). This dual-terminator design was observed to significantly reduce, though not completely eliminate, downstream GFP expression from the flipped Pbad promoter, as shown in Supplementary Fig. S9B. We then redesigned circuits C42 and C43 by incorporating the dual terminator, resulting in circuits C73 and C74 (Fig. 4F). In these redesigned circuits, we observed the expected negative correlation between GFP and RFP expression, with a reduced slope in C73 (Fig. 4G), as opposed to the positive correlation seen previously. Supplementary Fig. S10 shows the CI, illustrating the enhanced resource decoupling of circuit C74 compared to the reference circuit. The unnormalized RFP expression levels (circuits C45, C46, C73, and C74) are shown in Supplementary Fig. S11.

Additional feedforward control enhances mitigation of resource competition

The above observation was unexpected but can be leveraged for further minimizing resource coupling. By removing the B0015 terminator in C43, we constructed circuit C53 (Fig. 5A). In this new design, the promoter reversion induced by the integrase–excisionase complex not only limits RFP expression through a negative feedback mechanism but also compensates GFP expression via a feedforward control mechanism. We also constructed a reference circuit C55, which lacks the promoter-inverting function due to a mutated dinucleotide in the attR site. By combining the negative feedback and feedforward mechanisms, we hypothesize that this new design, “Re-NF-FF-Controller,” can enhance the efficiency of resource competition mitigation. By fine-tuning the recombinase elements, we can correct the overcompensation for the GFP module, ensuring that its expression is largely independent of RFP expression, exhibiting neither positive nor negative correlation.

We developed a mathematical model to demonstrate the tunability of the Re-NF-FF-Controller in regulating resource competition. GFP expression, mapped against RFP levels (under varying inducer concentration) and excisionase degradation rate (d_x), shows that optimal resource decoupling occurs within a tunable range of d_x (red boxed area, Fig. 5B). If the degradation rate of excisionase is too slow, the feedforward mechanism becomes excessively strong, leading to overcompensation and a positive CI (Fig. 5C). Conversely, if the degradation rate is too fast, the negative feedback loop becomes too weak to effectively reduce resource coupling, resulting in a system that still exhibits a negative CI (Fig. 5C). By optimizing excisionase level, the Re-NF-FF-Controller can maintain relatively stable GFP expression, minimizing its dependence on RFP expression (Fig. 5D). A similar strategy can be applied by fine-tuning integrase expression levels. As shown in Supplementary Fig. S12A–C, optimal integrase expression rate minimizes GFP perturbation and the CI. However, excessively high or low integrase expression leads to strong positive or negative correlations, respectively. These simulation results underscore the importance of maintaining balanced recombinase expression for the Re-NF-FF-Controller to effectively regulate resource competition.

Analysis of gene expression patterns in circuit C53 and the reference circuit C55 revealed distinct correlation trends: a strong positive correlation in C53, caused by overcompensation, and a strong negative correlation in C55, resulting from resource competition (Fig. 5E). Inspired by the tunability of the Re-NF-FF system demonstrated in the simulations, we modified circuit C53 into circuit C59 by introducing a degradation tag to excisionase (Fig. 5A). As expected, circuit C59 exhibited a more balanced expression profile, showing a relatively flattened GFP–RFP response curve compared to both C53 and C55 (Fig. 5E). To validate promoter flipping and its timing, we performed PCR amplification following induction of circuits C53 and C55 (see the “Materials and methods” section). Amplification of the attLR (nonflipped) and attBP (flipped) configurations was performed for samples from each time point. As shown in Supplementary Fig. S13A, an increasing proportion of flipped plasmids was observed in circuit C53 with prolonged induction time, while no flipping was detected in the reference circuit C55. A substantial increase was observed at 9 h post-induction compared to the 0–6 h time points (Supplementary Fig. S13B), indicating that the large-scale promoter flipping likely occurs between 6 and 9 h after induction. Circuits with lower recombination efficiency, such as C59, require a longer time to flip the promoter.

To evaluate the performance of the Re-NF-FF-Controller under heightened resource competition, we constructed high-copy versions of circuits C53 and C55, termed C53-H and C55-H, respectively. As shown in Fig. 5F, the reference circuit C55-H exhibited a very strong negative correlation between GFP and RFP expression, indicating increased resource competition. In contrast, C53-H showed minimal dependence of GFP on RFP levels, demonstrating effective decoupling. Adding a degradation tag to excisionase is unlikely to improve performance, as faster excisionase turnover reduces promoter flipping efficiency. To test this, we constructed C59-H, a high-copy version of C59, and observed reduced resource decoupling compared to C53-H, consistent with this expectation (Fig. 5F). To understand how excisionase degradation should be tuned under varying resource demands, we performed simulations to identify the optimal degradation rate across different RFP loads. Supplementary Fig. S14 shows that higher resource demand requires a slower excisionase degradation rate to maintain controller efficiency.

To test whether modulating integrase level could yield a similar tuning effect, we constructed circuit C57 by weakening integrase expression through a start codon mutation from ATG to GTG (Fig. 5A). As a result, GFP expression in C57 showed the weakest dependence on RFP compared to C53 and C55, indicating enhanced resource decoupling (Supplementary Fig. S12D). The unnormalized RFP expression levels are shown in Supplementary Fig. S15. Notably, higher RFP expression was observed in circuits C59 and C57, due to the reduced formation of the integrase–excisionase complex, which lowers the probability of promoter flipping. Together, experimental and simulation results highlight the importance of tuning recombinase levels to optimize the Re-NF-FF controller under resource-limited conditions.

To compare the overall resource-decoupling performance of the Re-NF and Re-NF-FF controllers across varying conditions, we calculated the CI based on experimental measurements. As shown in Fig. 5G, while the Re-NF controller exhibits strong decoupling, its performance declines under high resource competition with increased plasmid copy number. In contrast, the Re-NF-FF controller suffers from overcompensation but can be significantly improved through fine-tuning of integrase or excisionase expression. Moreover, in the high-copy context, the Re-NF-FF controller achieves effective decoupling as indicated by the low CI value. These results clearly demonstrate the limitations of the Re-NF controller and underscore the tunability and robustness of the Re-NF-FF controller in mitigating resource competition effects. We further validated these findings through theoretical simulations assessing the performance of both controllers under varying levels of resource competition. Re-NF-FF consistently outperforms the Re-NF controller across a range of resource competition conditions (Supplementary Fig. S16), highlighting the superior robustness of the Re-NF-FF controller.

To assess the influence of individual parameters on controller performance, we conducted a local parameter sensitivity analysis. The CI of both controllers was evaluated by varying each parameter by ±20% from its baseline value. This perturbation induced moderate fluctuations in performance across both systems (Supplementary Fig. S17). Notably, the Hill function parameters K_I1, K_I2, and K_X, which define the integrase and excisionase concentrations required to achieve half-maximal flipping rates, exhibited distinct effects. Variations in K_I1 showed negligible impact on the regulatory efficiency of either controller. In contrast, a 20% increase in K_I2 or K_X led to a slightly elevated CI compared to a 20% decrease, due to the requirement for higher concentrations of integrase and excisionase to form the integrase–excisionase complex. Nonetheless, both Re-NF and Re-NF-FF controllers maintained effective decoupling functionality under parameter perturbations, reflecting the robustness of recombinase-based control strategies. To further validate this conclusion, we performed a global parameter sensitivity analysis by generating 1000 random parameter sets, each sampled with a ± 20% variation relative to their nominal values. The resulting distribution of coupling indices across all scenarios is shown in Supplementary Fig. S18. The Re-NF-FF controller outperformed the Re-NF controller, exhibiting only modest overlap with the unregulated system. These findings demonstrate that the recombinase-based controllers, particularly Re-NF-FF, maintain robust performance in the face of parameter variability. Taken together, the Re-NF-FF-Controller integrates both negative feedback and feedforward mechanisms to achieve tunable resource distribution, effectively reducing resource competition in synthetic gene circuits.

Discussion

Resource competition adds another layer of complexity to engineering robust synthetic circuits. In this study, we observed substantial resource competition in open-loop systems, consistent with previous findings [6, 8]. To address this, we introduced a novel control strategy that integrates recombinase elements to restore the modularity of gene modules through a combination of negative feedback and feedforward regulation. Our results strongly suggest that combining feedforward with negative feedback effectively mitigates resource competition in a tunable manner.

Negative feedback is a fundamental regulatory motif in biological networks, known for its role in maintaining homeostasis and minimizing variability [28, 38–40]. Although this strategy can mitigate resource competition and reduce coupling, its capacity is inherently limited by controller saturation, where the regulatory elements (e.g. repressors or activators) reach a maximum effect and can no longer proportionally respond to increasing perturbations, thus failing to achieve full decoupling. This limitation is evident in several studies that employ negative feedback to mitigate resource coupling and is also reflected in our recombinase-based negative feedback systems under heightened resource competition condition, underscoring the need for a more efficient regulatory topology. Although not sufficient to achieve complete resource decoupling, the negative feedback mechanism can accelerate the system’s response to perturbations, reduce variance, and is particularly effective in handling large deviations. Owing to these characteristics, Frei et al. [16] integrated negative feedback with integral feedback to realize robust perfect adaptation in the presence of perturbations. However, constructing such a proportional-integral feedback circuit is challenging, as it requires an annihilation pair, two components that must be stable and capable of mutually nullifying each other’s function. Any leaky expression of these controller species can significantly impair the performance of the controller. Feedforward loops have been employed to achieve near-perfect adaptation in various contexts, such as cellular resource regulation, noise suppression, and growth burden control. Nonetheless, they are typically designed for single-input control, which limits their scalability for broader applications in the future. The combination of multiple regulatory mechanisms to achieve desired outcomes is a common strategy in synthetic biology. For example, negative feedback is coupled with positive feedback to achieve robust bistability, oscillation, and adaptation [41, 42]. By coupling a negative feedback loop with incoherent feedforward control, Yang et al. designed a circuit called “Equalizer” to buffer plasmid copy number variation [43].

However, combining these two regulatory mechanisms to control resource coupling in synthetic gene circuits presents challenges. First, more complex architecture inevitably introduces intricate interactions in circuit design and engineering. For instance, theoretical studies suggest that single-TF (transcription factor) negative feedback tends to offer greater robustness than multi-TF control strategies in metabolic regulation [44]. This implies that a simple design can outperform more complex alternatives without compromising circuit function. In our case, we implemented two regulatory mechanisms using a single promoter. This minimalistic approach not only maintains functional efficiency but also minimizes the risk of unpredictable interactions often associated with circuit complexity. Additionally, the use of more gene parts to construct complex structures to implement multiple control mechanisms consumes considerable cellular resources, imposing a significant burden on the host cells. Our design only requires the reversion of a single promoter by recombinase to achieve both regulatory mechanisms, avoiding the need for additional genes/promoters that would demand extra cellular resources. This minimizes the metabolic burden on the host cell. Furthermore, the simplicity of the system does not compromise its flexibility, as the desired output can be fine-tuned by adjusting the levels of integrase and excisionase. Integrase and excisionase form a complex in a stoichiometric manner. Therefore, in theory, modulating the expression of either component can achieve a tunable effect. In the current study, we implemented two strategies to adjust this balance: (i) a degradation tag was added to excisionase to reduce its intracellular concentration and (ii) the start codon strength of integrase was modified to alter its translation efficiency. In future studies, the tunability of this recombinase-based controller could be further explored using similar approaches, such as adjusting the degradation rate of integrase or modifying the translational activity of excisionase, thereby highlighting the flexibility of this regulatory system.

Like all control systems that incorporate negative feedback, the expression level of the regulated gene will be lower than the open-loop system [11, 16, 24, 28]. This challenge can be addressed by increasing the translational strength of regulated genes. Specifically, we observed a similar level of RFP expression in circuits C44 and C52 (Supplementary Fig. S6). Although RFP levels are comparable, GFP expression is substantially less reduced in C44 (Fig. 3E), suggesting that the recombined negative-feedback configuration effectively limits the burden imposed by the RFP module. In contrast, the greater reduction in GFP expression observed in C52 indicates that its RFP module imposes a higher translational load, even though this is not fully reflected in the measured RFP output. In the control circuits (C52), the recombinase can still bind the mutated recognition site and attempt ligation, which may lead to transcriptional interference or the production of nonfunctional or truncated transcripts. These incomplete products can consume transcriptional and translational resources without contributing to detectable RFP levels. Thus, the apparent similarity in RFP output masks a difference in underlying resource usage. Future studies incorporating alternative controls, such as catalytically inactive recombinases or recognition sites that fail to recruit recombinase, could help disentangle the individual contributions of recombinase binding, catalysis, and recombined configuration to circuit burden and resource redistribution.

Since recombinase expression imposes a cellular burden, excessively increasing recombinase levels to cope with heightened resource competition risks disrupting the trade-off between regulatory efficiency and controller burden. Exogenous expression or chromosomal integration offers a strategy to reduce the metabolic load on the host cell. Moreover, the availability of multiple orthogonal recombinases and their corresponding recombination sites makes the Re-NF-FF controller scalable for use in complex gene circuits, where cellular resources are limited. Zhao et al. proposed a binary ripple counter using integrase and its directionality factor to toggle states for event recording, which could be extended for controlling resource competition among more genes using multiple orthogonal recombination [45]. More than 25 putative large serine recombinases have been identified [46–48]. For each recombinase, up to six orthogonal recombination site pairs can be engineered by introducing distinct nonsymmetric 2-bp central overlap sequences. As a result, each attachment site recombines exclusively with its corresponding partner [49]. Utilizing orthogonal recombination site pairs, rather than multiple orthogonal recombinases, minimizes cellular resource consumption. This approach relies solely on DNA rearrangement in response to a single recombinase, thereby avoiding the metabolic burden associated with the expression of multiple recombinase proteins, which would otherwise consume various cellular resources. A comprehensive evaluation of these recombinases and their corresponding sites could enable the construction of scalable, resource-aware genetic circuits. Furthermore, recombinase elements have been successfully applied in species like mammals and yeast [50, 51], broadening the potential for recombinant-based regulation in diverse systems.

Supplementary Material

gkaf913_Supplemental_Files

gkaf913_supplemental_files.zip^{(5.6MB, zip)}

Acknowledgements

Author contributions: Rixin Zhang (Data curation [equal], Formal analysis [equal], Investigation [equal], Methodology [equal], Validation [equal], Visualization [equal], Writing—original draft [equal], Writing—review & editing [equal]), Rong Zhang (Data curation [supporting], Formal analysis [supporting], Investigation [supporting], Methodology [equal], Supervision [supporting] Validation [supporting], Visualization [supporting], Writing—original draft [supporting], Writing—review & editing [supporting]), and Xiao-Jun Tian (Conceptualization [lead], Data curation [equal], Formal analysis [equal], Funding acquisition [equal], Investigation [equal], Methodology [equal], Project administration [lead], Resources [lead], Software [equal], Supervision [lead] Validation [equal], Visualization [equal], Writing—original draft [equal], Writing—review & editing [equal]).

Contributor Information

Rixin Zhang, School of Biological and Health Systems Engineering, Arizona State University, Tempe, AZ 85281, United States.

Rong Zhang, School of Biological and Health Systems Engineering, Arizona State University, Tempe, AZ 85281, United States.

Xiao-Jun Tian, School of Biological and Health Systems Engineering, Arizona State University, Tempe, AZ 85281, United States.

Supplementary data

Supplementary data is available at NAR online.

Conflict of interest

None declared.

Funding

This work was supported by grants from the US National Institutes of Health (R35GM142896 to X.-J.T.). Funding to pay the Open Access publication charges for this article was provided by National Institutes of Health.

Data availability

All data produced or analyzed for this study are included in the article and its Supplementary information files. The simulation code is shared at GitHub (https://github.com/TianLab-ASU/Re-NF-FF-Controller) and Zenodo at https://doi.org/10.5281/zenodo.16955061. Plasmids are available through Addgene (ID 245095-245101).

References

1. Purnick PEM, Weiss R The second wave of synthetic biology: from modules to systems. Nat Rev Mol Cell Biol. 2009; 10:410–22. 10.1038/nrm2698. [DOI] [PubMed] [Google Scholar]
2. Stone A, Youssef A, Rijal S et al. Context-dependent redesign of robust synthetic gene circuits. Trends Biotechnol. 2024; 42:895–909. 10.1016/j.tibtech.2024.01.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
3. Kumar S, Hasty J Stability, robustness, and containment: preparing synthetic biology for real-world deployment. Curr Opin Biotechnol. 2023; 79:102880. 10.1016/j.copbio.2022.102880. [DOI] [PMC free article] [PubMed] [Google Scholar]
4. Shakiba N, Jones RD, Weiss R et al. Context-aware synthetic biology by controller design: engineering the mammalian cell. Cell Syst. 2021; 12:561–92. 10.1016/j.cels.2021.05.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
5. Şimşek E, Yao Y, Lee D et al. Toward predictive engineering of gene circuits. Trends Biotechnol. 2023; 41:760–8. [DOI] [PubMed] [Google Scholar]
6. Qian Y, Huang H-H, Jiménez JI et al. Resource competition shapes the response of genetic circuits. ACS Synth Biol. 2017; 6:1263–72. 10.1021/acssynbio.6b00361. [DOI] [PubMed] [Google Scholar]
7. Zhang R, Goetz H, Melendez-Alvarez J et al. Winner-takes-all resource competition redirects cascading cell fate transitions. Nat Commun. 2021; 12:853. 10.1038/s41467-021-21125-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
8. Gyorgy A, Jiménez JI, Yazbek J et al. Isocost lines describe the cellular economy of genetic circuits. Biophys J. 2015; 109:639–46. 10.1016/j.bpj.2015.06.034. [DOI] [PMC free article] [PubMed] [Google Scholar]
9. Carbonell-Ballestero M, Garcia-Ramallo E, Montañez R et al. Dealing with the genetic load in bacterial synthetic biology circuits: convergences with the Ohm’s law. Nucleic Acids Res. 2016; 44:496–507. 10.1093/nar/gkv1280. [DOI] [PMC free article] [PubMed] [Google Scholar]
10. McBride CD, Grunberg TW, Del Vecchio D Design of genetic circuits that are robust to resource competition. Curr Opin Syst Biol. 2021; 28:100357. 10.1016/j.coisb.2021.100357. [DOI] [Google Scholar]
11. Darlington APS, Kim J, Jiménez JI et al. Dynamic allocation of orthogonal ribosomes facilitates uncoupling of co-expressed genes. Nat Commun. 2018; 9:695. 10.1038/s41467-018-02898-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
12. Kushwaha M, Salis HM A portable expression resource for engineering cross-species genetic circuits and pathways. Nat Commun. 2015; 6:7832. 10.1038/ncomms8832. [DOI] [PMC free article] [PubMed] [Google Scholar]
13. An W, Chin JW Synthesis of orthogonal transcription-translation networks. Proc Natl Acad Sci USA. 2009; 106:8477–82. 10.1073/pnas.0900267106. [DOI] [PMC free article] [PubMed] [Google Scholar]
14. Segall-Shapiro TH, Meyer AJ, Ellington AD et al. A ‘resource allocator’ for transcription based on a highly fragmented T7 RNA polymerase. Mol Syst Biol. 2014; 10:742. 10.15252/msb.20145299. [DOI] [PMC free article] [PubMed] [Google Scholar]
15. Venturelli OS, Tei M, Bauer S et al. Programming mRNA decay to modulate synthetic circuit resource allocation. Nat Commun. 2017; 8:15128. 10.1038/ncomms15128. [DOI] [PMC free article] [PubMed] [Google Scholar]
16. Frei T, Chang C-H, Filo M et al. A genetic mammalian proportional–integral feedback control circuit for robust and precise gene regulation. Proc Natl Acad Sci USA. 2022; 119:e2122132119. 10.1073/pnas.2122132119. [DOI] [PMC free article] [PubMed] [Google Scholar]
17. Ceroni F, Boo A, Furini S et al. Burden-driven feedback control of gene expression. Nat Methods. 2018; 15:387–93. 10.1038/nmeth.4635. [DOI] [PubMed] [Google Scholar]
18. Stone A, Ryan J, Tang X et al. Negatively competitive incoherent feedforward loops mitigate winner-take-all resource competition. ACS Synth Biol. 2022; 11:3986–95. 10.1021/acssynbio.2c00318. [DOI] [PMC free article] [PubMed] [Google Scholar]
19. Jones RD, Qian Y, Siciliano V et al. An endoribonuclease-based feedforward controller for decoupling resource-limited genetic modules in mammalian cells. Nat Commun. 2020; 11:5690. 10.1038/s41467-020-19126-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
20. Jones RD, Qian Y, Ilia K et al. Robust and tunable signal processing in mammalian cells via engineered covalent modification cycles. Nat Commun. 2022; 13:1720. 10.1038/s41467-022-29338-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
21. Huang H-H, Qian Y, Del Vecchio D A quasi-integral controller for adaptation of genetic modules to variable ribosome demand. Nat Commun. 2018; 9:5415. 10.1038/s41467-018-07899-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
22. Chakravarty S, Guttal R, Zhang R et al. Mitigating winner-take-all resource competition through antithetic control mechanism. ACS Synth Biol. 2024; 13:4050–60. 10.1021/acssynbio.4c00476. [DOI] [PMC free article] [PubMed] [Google Scholar]
23. Stone A, Zhang R, Tian X-J Coupling shared and tunable negative competition against winner-take-all resource competition via CRISPRi moieties. 2021 American Control Conference (ACC). 2021; 10.23919/ACC50511.2021.9483381. [DOI] [Google Scholar]
24. Shopera T, He L, Oyetunde T et al. Decoupling resource-coupled gene expression in living cells. ACS Synth Biol. 2017; 6:1596–604. 10.1021/acssynbio.7b00119. [DOI] [PubMed] [Google Scholar]
25. Barajas C, Huang H-H, Gibson J et al. Feedforward growth rate control mitigates gene activation burden. Nat Commun. 2022; 13:7054. 10.1038/s41467-022-34647-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
26. Grindley NDF, Whiteson KL, Rice PA Mechanisms of site-specific recombination. Annu Rev Biochem. 2006; 75:567–605. 10.1146/annurev.biochem.73.011303.073908. [DOI] [PubMed] [Google Scholar]
27. Bonnet J, Subsoontorn P, Endy D Rewritable digital data storage in live cells via engineered control of recombination directionality. Proc Natl Acad Sci USA. 2012; 109:8884–9. 10.1073/pnas.1202344109. [DOI] [PMC free article] [PubMed] [Google Scholar]
28. Folliard T, Steel H, Prescott TP et al. A synthetic recombinase-based feedback loop results in robust expression. ACS Synth Biol. 2017; 6:1663–71. 10.1021/acssynbio.7b00131. [DOI] [PubMed] [Google Scholar]
29. Steel H, Papachristodoulou A Low-burden biological feedback controllers for near-perfect adaptation. ACS Synth Biol. 2019; 8:2212–9. 10.1021/acssynbio.9b00125. [DOI] [PubMed] [Google Scholar]
30. Schneider CA, Rasband WS, Eliceiri KW NIH image to ImageJ: 25 years of image analysis. Nat Methods. 2012; 9:671–5. 10.1038/nmeth.2089. [DOI] [PMC free article] [PubMed] [Google Scholar]
31. Helton JC, Davis FJ Latin hypercube sampling and the propagation of uncertainty in analyses of complex systems. Reliab Eng Syst Saf. 2003; 81:23–69. 10.1016/S0951-8320(03)00058-9. [DOI] [Google Scholar]
32. Sanford EM, Emert BL, Coté A et al. Gene regulation gravitates toward either addition or multiplication when combining the effects of two signals. eLife. 2020; 9:e59388. 10.7554/eLife.59388. [DOI] [PMC free article] [PubMed] [Google Scholar]
33. Liu L, Ning N, Xu S et al. Double promoter and tandem gene strategy for efficiently expressing recombinant FGF21. Microb Cell Fact. 2024; 23:171. 10.1186/s12934-024-02447-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
34. Ghosh P, Wasil LR, Hatfull GF Control of phage Bxb1 excision by a novel recombination directionality factor. PLoS Biol. 2006; 4:e186. 10.1371/journal.pbio.0040186. [DOI] [PMC free article] [PubMed] [Google Scholar]
35. Kim AI, Ghosh P, Aaron MA et al. Mycobacteriophage Bxb1 integrates into the Mycobacterium smegmatis groEL1 gene. Mol Microbiol. 2003; 50:463–73. 10.1046/j.1365-2958.2003.03723.x. [DOI] [PubMed] [Google Scholar]
36. Morgan-Kiss RM, Wadler C, Cronan JE Long-term and homogeneous regulation of the Escherichia coli araBAD promoter by use of a lactose transporter of relaxed specificity. Proc Natl Acad Sci USA. 2002; 99:7373–7. 10.1073/pnas.122227599. [DOI] [PMC free article] [PubMed] [Google Scholar]
37. Ghosh P, Bibb LA, Hatfull GF Two-step site selection for serine-integrase-mediated excision: DNA-directed integrase conformation and central dinucleotide proofreading. Proc Natl Acad Sci USA. 2008; 105:3238–43. 10.1073/pnas.0711649105. [DOI] [PMC free article] [PubMed] [Google Scholar]
38. Nevozhay D, Adams RM, Murphy KF et al. Negative autoregulation linearizes the dose–response and suppresses the heterogeneity of gene expression. Proc Natl Acad Sci USA. 2009; 106:5123–8. 10.1073/pnas.0809901106. [DOI] [PMC free article] [PubMed] [Google Scholar]
39. Dublanche Y, Michalodimitrakis K, Kümmerer N et al. Noise in transcription negative feedback loops: simulation and experimental analysis. Mol Syst Biol. 2006; 2:41. 10.1038/msb4100081. [DOI] [PMC free article] [PubMed] [Google Scholar]
40. Montefusco F, Akman OE, Soyer OS et al. Ultrasensitive negative feedback control: a natural approach for the design of synthetic controllers. PLoS One. 2016; 11:e0161605. 10.1371/journal.pone.0161605. [DOI] [PMC free article] [PubMed] [Google Scholar]
41. Sun Z, Wei W, Zhang M et al. Synthetic robust perfect adaptation achieved by negative feedback coupling with linear weak positive feedback. Nucleic Acids Res. 2022; 50:2377–86. 10.1093/nar/gkac066. [DOI] [PMC free article] [PubMed] [Google Scholar]
42. Tian X-J, Zhang X-P, Liu F et al. Interlinking positive and negative feedback loops creates a tunable motif in gene regulatory networks. Phys Rev E. 2009; 80:11926. 10.1103/PhysRevE.80.011926. [DOI] [PubMed] [Google Scholar]
43. Yang J, Lee J, Land MA et al. A synthetic circuit for buffering gene dosage variation between individual mammalian cells. Nat Commun. 2021; 12:4132. 10.1038/s41467-021-23889-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
44. Oyarzún DA, Stan G-B Design tradeoffs in a synthetic gene control circuit for metabolic networks. 2012 American Control Conference (ACC). 2012; 2743–8. [Google Scholar]
45. Zhao J, Pokhilko A, Ebenhöh O et al. A single-input binary counting module based on serine integrase site-specific recombination. Nucleic Acids Res. 2019; 47:4896–909. 10.1093/nar/gkz245. [DOI] [PMC free article] [PubMed] [Google Scholar]
46. Brown WRA, Lee NCO, Xu Z et al. Serine recombinases as tools for genome engineering. Methods. 2011; 53:372–9. [DOI] [PubMed] [Google Scholar]
47. Yang L, Nielsen AAK, Fernandez-Rodriguez J et al. Permanent genetic memory with >1-byte capacity. Nat Methods. 2014; 11:1261–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
48. Roquet N, Soleimany AP, Ferris AC et al. Synthetic recombinase-based state machines in living cells. Science. 2016; 353:aad8559. 10.1126/science.aad8559. [DOI] [PubMed] [Google Scholar]
49. Colloms SD, Merrick CA, Olorunniji FJ et al. Rapid metabolic pathway assembly and modification using serine integrase site-specific recombination. Nucleic Acids Res. 2014; 42:e23. 10.1093/nar/gkt1101. [DOI] [PMC free article] [PubMed] [Google Scholar]
50. Roelle SM, Kamath ND, Matreyek KA Mammalian genomic manipulation with orthogonal Bxb1 DNA recombinase sites for the functional characterization of protein variants. ACS Synth Biol. 2023; 12:3352–65. 10.1021/acssynbio.3c00355. [DOI] [PMC free article] [PubMed] [Google Scholar]
51. Cautereels C, Smets J, Bircham P et al. Combinatorial optimization of gene expression through recombinase-mediated promoter and terminator shuffling in yeast. Nat Commun. 2024; 15:1112. 10.1038/s41467-024-44997-7. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

gkaf913_Supplemental_Files

gkaf913_supplemental_files.zip^{(5.6MB, zip)}

Data Availability Statement

[B1] 1. Purnick PEM, Weiss R The second wave of synthetic biology: from modules to systems. Nat Rev Mol Cell Biol. 2009; 10:410–22. 10.1038/nrm2698. [DOI] [PubMed] [Google Scholar]

[B2] 2. Stone A, Youssef A, Rijal S et al. Context-dependent redesign of robust synthetic gene circuits. Trends Biotechnol. 2024; 42:895–909. 10.1016/j.tibtech.2024.01.003. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B3] 3. Kumar S, Hasty J Stability, robustness, and containment: preparing synthetic biology for real-world deployment. Curr Opin Biotechnol. 2023; 79:102880. 10.1016/j.copbio.2022.102880. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B4] 4. Shakiba N, Jones RD, Weiss R et al. Context-aware synthetic biology by controller design: engineering the mammalian cell. Cell Syst. 2021; 12:561–92. 10.1016/j.cels.2021.05.011. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B5] 5. Şimşek E, Yao Y, Lee D et al. Toward predictive engineering of gene circuits. Trends Biotechnol. 2023; 41:760–8. [DOI] [PubMed] [Google Scholar]

[B6] 6. Qian Y, Huang H-H, Jiménez JI et al. Resource competition shapes the response of genetic circuits. ACS Synth Biol. 2017; 6:1263–72. 10.1021/acssynbio.6b00361. [DOI] [PubMed] [Google Scholar]

[B7] 7. Zhang R, Goetz H, Melendez-Alvarez J et al. Winner-takes-all resource competition redirects cascading cell fate transitions. Nat Commun. 2021; 12:853. 10.1038/s41467-021-21125-3. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B8] 8. Gyorgy A, Jiménez JI, Yazbek J et al. Isocost lines describe the cellular economy of genetic circuits. Biophys J. 2015; 109:639–46. 10.1016/j.bpj.2015.06.034. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B9] 9. Carbonell-Ballestero M, Garcia-Ramallo E, Montañez R et al. Dealing with the genetic load in bacterial synthetic biology circuits: convergences with the Ohm’s law. Nucleic Acids Res. 2016; 44:496–507. 10.1093/nar/gkv1280. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B10] 10. McBride CD, Grunberg TW, Del Vecchio D Design of genetic circuits that are robust to resource competition. Curr Opin Syst Biol. 2021; 28:100357. 10.1016/j.coisb.2021.100357. [DOI] [Google Scholar]

[B11] 11. Darlington APS, Kim J, Jiménez JI et al. Dynamic allocation of orthogonal ribosomes facilitates uncoupling of co-expressed genes. Nat Commun. 2018; 9:695. 10.1038/s41467-018-02898-6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B12] 12. Kushwaha M, Salis HM A portable expression resource for engineering cross-species genetic circuits and pathways. Nat Commun. 2015; 6:7832. 10.1038/ncomms8832. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B13] 13. An W, Chin JW Synthesis of orthogonal transcription-translation networks. Proc Natl Acad Sci USA. 2009; 106:8477–82. 10.1073/pnas.0900267106. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B14] 14. Segall-Shapiro TH, Meyer AJ, Ellington AD et al. A ‘resource allocator’ for transcription based on a highly fragmented T7 RNA polymerase. Mol Syst Biol. 2014; 10:742. 10.15252/msb.20145299. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B15] 15. Venturelli OS, Tei M, Bauer S et al. Programming mRNA decay to modulate synthetic circuit resource allocation. Nat Commun. 2017; 8:15128. 10.1038/ncomms15128. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B16] 16. Frei T, Chang C-H, Filo M et al. A genetic mammalian proportional–integral feedback control circuit for robust and precise gene regulation. Proc Natl Acad Sci USA. 2022; 119:e2122132119. 10.1073/pnas.2122132119. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B17] 17. Ceroni F, Boo A, Furini S et al. Burden-driven feedback control of gene expression. Nat Methods. 2018; 15:387–93. 10.1038/nmeth.4635. [DOI] [PubMed] [Google Scholar]

[B18] 18. Stone A, Ryan J, Tang X et al. Negatively competitive incoherent feedforward loops mitigate winner-take-all resource competition. ACS Synth Biol. 2022; 11:3986–95. 10.1021/acssynbio.2c00318. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B19] 19. Jones RD, Qian Y, Siciliano V et al. An endoribonuclease-based feedforward controller for decoupling resource-limited genetic modules in mammalian cells. Nat Commun. 2020; 11:5690. 10.1038/s41467-020-19126-9. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B20] 20. Jones RD, Qian Y, Ilia K et al. Robust and tunable signal processing in mammalian cells via engineered covalent modification cycles. Nat Commun. 2022; 13:1720. 10.1038/s41467-022-29338-w. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B21] 21. Huang H-H, Qian Y, Del Vecchio D A quasi-integral controller for adaptation of genetic modules to variable ribosome demand. Nat Commun. 2018; 9:5415. 10.1038/s41467-018-07899-z. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B22] 22. Chakravarty S, Guttal R, Zhang R et al. Mitigating winner-take-all resource competition through antithetic control mechanism. ACS Synth Biol. 2024; 13:4050–60. 10.1021/acssynbio.4c00476. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B23] 23. Stone A, Zhang R, Tian X-J Coupling shared and tunable negative competition against winner-take-all resource competition via CRISPRi moieties. 2021 American Control Conference (ACC). 2021; 10.23919/ACC50511.2021.9483381. [DOI] [Google Scholar]

[B24] 24. Shopera T, He L, Oyetunde T et al. Decoupling resource-coupled gene expression in living cells. ACS Synth Biol. 2017; 6:1596–604. 10.1021/acssynbio.7b00119. [DOI] [PubMed] [Google Scholar]

[B25] 25. Barajas C, Huang H-H, Gibson J et al. Feedforward growth rate control mitigates gene activation burden. Nat Commun. 2022; 13:7054. 10.1038/s41467-022-34647-1. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B26] 26. Grindley NDF, Whiteson KL, Rice PA Mechanisms of site-specific recombination. Annu Rev Biochem. 2006; 75:567–605. 10.1146/annurev.biochem.73.011303.073908. [DOI] [PubMed] [Google Scholar]

[B27] 27. Bonnet J, Subsoontorn P, Endy D Rewritable digital data storage in live cells via engineered control of recombination directionality. Proc Natl Acad Sci USA. 2012; 109:8884–9. 10.1073/pnas.1202344109. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B28] 28. Folliard T, Steel H, Prescott TP et al. A synthetic recombinase-based feedback loop results in robust expression. ACS Synth Biol. 2017; 6:1663–71. 10.1021/acssynbio.7b00131. [DOI] [PubMed] [Google Scholar]

[B29] 29. Steel H, Papachristodoulou A Low-burden biological feedback controllers for near-perfect adaptation. ACS Synth Biol. 2019; 8:2212–9. 10.1021/acssynbio.9b00125. [DOI] [PubMed] [Google Scholar]

[B30] 30. Schneider CA, Rasband WS, Eliceiri KW NIH image to ImageJ: 25 years of image analysis. Nat Methods. 2012; 9:671–5. 10.1038/nmeth.2089. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B31] 31. Helton JC, Davis FJ Latin hypercube sampling and the propagation of uncertainty in analyses of complex systems. Reliab Eng Syst Saf. 2003; 81:23–69. 10.1016/S0951-8320(03)00058-9. [DOI] [Google Scholar]

[B32] 32. Sanford EM, Emert BL, Coté A et al. Gene regulation gravitates toward either addition or multiplication when combining the effects of two signals. eLife. 2020; 9:e59388. 10.7554/eLife.59388. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B33] 33. Liu L, Ning N, Xu S et al. Double promoter and tandem gene strategy for efficiently expressing recombinant FGF21. Microb Cell Fact. 2024; 23:171. 10.1186/s12934-024-02447-5. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B34] 34. Ghosh P, Wasil LR, Hatfull GF Control of phage Bxb1 excision by a novel recombination directionality factor. PLoS Biol. 2006; 4:e186. 10.1371/journal.pbio.0040186. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B35] 35. Kim AI, Ghosh P, Aaron MA et al. Mycobacteriophage Bxb1 integrates into the Mycobacterium smegmatis groEL1 gene. Mol Microbiol. 2003; 50:463–73. 10.1046/j.1365-2958.2003.03723.x. [DOI] [PubMed] [Google Scholar]

[B36] 36. Morgan-Kiss RM, Wadler C, Cronan JE Long-term and homogeneous regulation of the Escherichia coli araBAD promoter by use of a lactose transporter of relaxed specificity. Proc Natl Acad Sci USA. 2002; 99:7373–7. 10.1073/pnas.122227599. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B37] 37. Ghosh P, Bibb LA, Hatfull GF Two-step site selection for serine-integrase-mediated excision: DNA-directed integrase conformation and central dinucleotide proofreading. Proc Natl Acad Sci USA. 2008; 105:3238–43. 10.1073/pnas.0711649105. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B38] 38. Nevozhay D, Adams RM, Murphy KF et al. Negative autoregulation linearizes the dose–response and suppresses the heterogeneity of gene expression. Proc Natl Acad Sci USA. 2009; 106:5123–8. 10.1073/pnas.0809901106. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B39] 39. Dublanche Y, Michalodimitrakis K, Kümmerer N et al. Noise in transcription negative feedback loops: simulation and experimental analysis. Mol Syst Biol. 2006; 2:41. 10.1038/msb4100081. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B40] 40. Montefusco F, Akman OE, Soyer OS et al. Ultrasensitive negative feedback control: a natural approach for the design of synthetic controllers. PLoS One. 2016; 11:e0161605. 10.1371/journal.pone.0161605. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B41] 41. Sun Z, Wei W, Zhang M et al. Synthetic robust perfect adaptation achieved by negative feedback coupling with linear weak positive feedback. Nucleic Acids Res. 2022; 50:2377–86. 10.1093/nar/gkac066. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B42] 42. Tian X-J, Zhang X-P, Liu F et al. Interlinking positive and negative feedback loops creates a tunable motif in gene regulatory networks. Phys Rev E. 2009; 80:11926. 10.1103/PhysRevE.80.011926. [DOI] [PubMed] [Google Scholar]

[B43] 43. Yang J, Lee J, Land MA et al. A synthetic circuit for buffering gene dosage variation between individual mammalian cells. Nat Commun. 2021; 12:4132. 10.1038/s41467-021-23889-0. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B44] 44. Oyarzún DA, Stan G-B Design tradeoffs in a synthetic gene control circuit for metabolic networks. 2012 American Control Conference (ACC). 2012; 2743–8. [Google Scholar]

[B45] 45. Zhao J, Pokhilko A, Ebenhöh O et al. A single-input binary counting module based on serine integrase site-specific recombination. Nucleic Acids Res. 2019; 47:4896–909. 10.1093/nar/gkz245. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B46] 46. Brown WRA, Lee NCO, Xu Z et al. Serine recombinases as tools for genome engineering. Methods. 2011; 53:372–9. [DOI] [PubMed] [Google Scholar]

[B47] 47. Yang L, Nielsen AAK, Fernandez-Rodriguez J et al. Permanent genetic memory with >1-byte capacity. Nat Methods. 2014; 11:1261–6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B48] 48. Roquet N, Soleimany AP, Ferris AC et al. Synthetic recombinase-based state machines in living cells. Science. 2016; 353:aad8559. 10.1126/science.aad8559. [DOI] [PubMed] [Google Scholar]

[B49] 49. Colloms SD, Merrick CA, Olorunniji FJ et al. Rapid metabolic pathway assembly and modification using serine integrase site-specific recombination. Nucleic Acids Res. 2014; 42:e23. 10.1093/nar/gkt1101. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B50] 50. Roelle SM, Kamath ND, Matreyek KA Mammalian genomic manipulation with orthogonal Bxb1 DNA recombinase sites for the functional characterization of protein variants. ACS Synth Biol. 2023; 12:3352–65. 10.1021/acssynbio.3c00355. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B51] 51. Cautereels C, Smets J, Bircham P et al. Combinatorial optimization of gene expression through recombinase-mediated promoter and terminator shuffling in yeast. Nat Commun. 2024; 15:1112. 10.1038/s41467-024-44997-7. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Integrating recombinase-based feedback and feedforward control for optimal resource decoupling

Rixin Zhang

Rong Zhang

Xiao-Jun Tian

Roles

Abstract

Graphical Abstract

Graphical Abstract.

Introduction

Materials and methods

Strains, media, and chemicals

Plasmids construction

Circuit inductions

Time-course analysis of promoter flipping

Microscopy

Flow cytometry

Coupling index

Parameter sensitivity analysis

Figure 5.

Mathematical models

Results

Design of recombinase-based control strategy for minimizing resource coupling

Figure 1.

Evaluation of promoter flipping efficiency by integrase and excisionase

Figure 2.

Recombinase-based negative feedback reduces resource coupling

Figure 3.

Terminator malfunction causes unanticipated positive correlation in gene expressions

Figure 4.

Additional feedforward control enhances mitigation of resource competition

Discussion

Supplementary Material

Acknowledgements

Contributor Information

Supplementary data

Conflict of interest

Funding

Data availability

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases