Characterization of recombinase activity across cellular growth phases

Abstract

Recombinases are essential enzymes in synthetic biology and biomedical research, enabling site-specific DNA modifications for applications such as gene therapy, the generation of transgenic models, and the construction of genetic circuits. Recombination efficiency depends on several factors, including intracellular recombinase concentration and the growth phase of the host cells. Although recombination is typically studied during exponential growth, the effects of stationary-phase dynamics on recombinase activity remain poorly understood. In this study, we examine how bacterial growth phase influences recombinase-mediated DNA modifications, using the serine recombinase Bxb1 as a model. We engineered a genetic system in Escherichia coli to quantify intracellular Bxb1 levels and to measure recombination efficiency across different growth phases. Our results reveal a quasi-linear relationship between recombinase concentration and recombination efficiency during exponential growth, up to a saturation point. Notably, recombination continues in the stationary phase following recombinase induction in exponential phase, despite the decline in plasmid gene expression. Cells that undergo recombination during the stationary phase show significantly higher recombination efficiencies upon re-entering exponential growth than those maintained in exponential phase throughout. These findings highlight the importance of induction timing in optimizing recombinase-based genetic modifications. Specifically, inducing recombinase expression just before the onset of stationary phase can enhance recombination efficiency while minimizing the need for high expression levels. Moreover, the observed quasi-linear relationship during exponential growth provides a framework for tuning gene expression with precision. Overall, this work offers new insights into leveraging bacterial growth dynamics to improve the design and control of synthetic genetic systems.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-17024-y.

Introduction

Recombinases are enzymes that mediate the site-specific rearrangement of DNA by recognizing and binding to defined recombination sequences. These sequences serve as anchoring points at which recombinases precisely excise, insert, or invert DNA segments, ensuring that genetic modifications are both accurate and controllable^1,2.

In both biomedical research and synthetic biology, recombinases are powerful tools because of their ability to induce targeted genetic changes. Their precision makes them ideal for diverse applications, ranging from gene therapy to the development of advanced biological systems. In biomedical contexts, recombinases are key to innovative gene therapy approaches, enabling the insertion, deletion, or correction of specific genes within a genome. This capacity holds strong therapeutic potential for treating genetic disorders through direct modification of pathogenic mutations^3,4. Additionally, recombinases are instrumental in generating transgenic animal models^5–7 which serve as essential platforms for studying human diseases and evaluating novel therapies in vivo^8–10.

In synthetic biology, recombinases are widely used for designing complex genetic circuits that regulate gene expression in response to environmental or cellular signals^11–13. Moreover, recombinase-based biosensors enable sensitive and specific detection of environmental contaminants or disease biomarkers, opening new avenues for both environmental monitoring and clinical diagnostics^14–16. Recombinases are also central to emerging biocomputing systems, advancing the development of programmable cellular devices capable of autonomous regulation in response to physiological cues^17–20.

Understanding recombinase dynamics is thus critical for the optimal development of cellular devices and other biotechnological or therapeutic applications where recombination efficiency is a limiting factor.

In the context of synthetic biology, serine recombinases^21,22 have emerged as particularly advantageous because of their high site specificity, recombination efficiency, low toxicity, and independence from accessory co-factors. These features make them highly versatile, suitable for use across various cell types, and ideal for the streamlined engineering of genetic circuits²².

The efficiency of recombinase-mediated DNA modification depends on several key parameters: (i) the intracellular abundance of the recombinase, (ii) the number and accessibility of recombination sites, and (iii) the physiological state of the culture, specifically whether the cells are in exponential or stationary growth phase. Depending on these parameters, recombination may be incomplete, leaving a fraction of the target sites unmodified. This outcome may be beneficial or detrimental depending on the application. For instance, in cellular computing applications where a full population-level response is required, partial recombination can lead to heterogeneous behavior, compromising system performance. Conversely, partial recombination can be exploited to fine-tune gene expression levels, particularly when using multicopy plasmids in situations where a graded response may be desirable.

This study investigates two major factors influencing recombination efficiency: the intracellular expression level of recombinases and the growth phase of the bacterial culture. We report a strong positive correlation between recombinase abundance and recombination efficiency, modulated by the physiological state of the culture. Specifically, inducing recombinase expression immediately before entry into stationary phase, followed by incubation in stationary phase and subsequent return to exponential growth, results in recombination efficiency that is significantly higher than that following recombinase induction during exponential growth alone.

These findings suggest that the stationary phase provides a more favorable environment for recombination. This may be due to reduced replication, as well as decreased dilution of intracellular recombinase through cell division, thereby promoting the completion of recombination events.

As a model system, we analyzed the serine recombinase Bxb1^23,24, which is widely used in synthetic biology. Bxb1 recognizes attP and attB recombination sites and mediates two types of genetic modifications depending on their orientation: excision of the DNA segment when attP and attB are in the same direction, or inversion when they are in opposite directions²⁴. This study focuses on excision events and systematically evaluates how recombination efficiency is affected by Bxb1 expression levels and the cellular growth phase.

Results

To investigate the dynamics of recombination as a function of intracellular recombinase abundance and cellular growth phase, it was essential to implement a genetic system capable of quantitatively monitoring both recombinase levels and recombination efficiency in vivo.

To track intracellular levels of the Bxb1 recombinase in real time, we constructed a fusion protein comprising Bxb1 and a red fluorescent protein (RFP), connected via a flexible peptide linker (see Table S1 in Supplementary Materials). This design enables the fluorescence intensity of RFP to serve as a direct proxy for intracellular Bxb1 abundance.

The Bxb1-RFP fusion protein was expressed under the control of the arabinose-inducible P_BAD promoter²⁵. This expression cassette (GC1) was cloned into the low-copy-number plasmid pSB3K3²⁶ (see Fig. 1). The combination of the tightly regulated P_BAD promoter, characterized by minimal basal activity, and a low-copy backbone effectively minimized unintended background expression of Bxb1, thereby reducing the risk of spurious recombination in the absence of arabinose.

Fig. 1.

Schematic representation of the cell types used in this study. Cell type C1 carries two genetic constructs, GC1 and GC2. GC1 is located on the low-copy-number plasmid pSB3K3 and encodes the Bxb1-RFP fusion protein under the control of the arabinose-inducible promoter PBAD. GC2 is located on the high-copy-number plasmid pSB1AC3 and encodes GFP under the control of the PTet promoter. A terminator sequence (T), flanked by the recombination sites attB and attP, is inserted between the promoter and the ribosome binding site (RBS), thereby blocking gene expression. Cell type C2 is similar to C1 but contains GC1b instead of GC1. In GC1b, the Bxb1-RFP fusion protein is replaced by the wild-type Bxb1 recombinase. Cell type C3 carries GC1b and a modified version of GC2, termed GC2b, also on pSB1AC3. GC2b is similar to GC2 but includes an additional gene encoding RFP under the control of the PTet promoter, ensuring constitutive expression. Cell type C4 contains GC1 and a third construct, GC2c, also carried on pSB1AC3. GC2c expresses GFP under the control of the PTet promoter, but features an attL site positioned between the promoter and the RBS, mimicking a fully recombined state that allows GFP expression independently of recombinase activity.

To quantify recombination efficiency, we designed a second genetic construct (GC2) incorporating the green fluorescent protein (GFP) gene under the control of the constitutive P_Tet promoter, which remains active in the absence of the TetR repressor²⁷. A synthetic transcriptional terminator was inserted between the PTet promoter and the ribosome-binding site (RBS), flanked by Bxb1 recognition sites (attP and attB) in direct orientation. This terminator sequence effectively blocked GFP transcription prior to recombination. GC2 was cloned into the high-copy-number plasmid pSB1AC3, allowing robust signal amplification. Upon induction with arabinose, the expressed Bxb1-RFP mediates site-specific excision of the terminator sequence, thereby restoring transcriptional continuity and enabling GFP expression. As a result, GFP fluorescence can provide a quantitative measure of recombination efficiency.

Validation of the activity of the Bxb1-RFP fusion recombinase

The functionality of the Bxb1-RFP fusion protein was initially validated by comparing its recombination activity with that of the native Bxb1 recombinase. To this end, two distinct cell types were engineered. Figure 1 provides a schematic representation of their genetic architecture. The first, designated C1, harbored the genetic constructs GC1 and GC2. The second, designated C2, contained GC2 and a variant of GC1 (termed GC1b in Fig. 1), in which the P_BAD promoter drove expression of native Bxb1 instead of the Bxb1-RFP fusion protein.

To compare recombination efficiency, expression of the recombinases was induced by adding 10⁻⁴ M arabinose to the culture medium of both cell types. Induction was maintained for 24 h at 37 °C with shaking, after which GFP fluorescence was measured (see Methods for details). Figure S1 shows GFP expression levels in both cell types, with and without arabinose induction. In the absence of arabinose, both C1 and C2 exhibited similarly low GFP levels. Upon induction, GFP expression increased significantly in both cases, consistent with recombination mediated by the expressed recombinases. Notably, no statistically significant differences were observed between the two recombinases in either the absence (p = 0.259) or presence (p = 0.224) of arabinose, as determined by a t-test. These results indicate that the Bxb1-RFP fusion retains recombinase activity comparable to that of the native Bxb1 enzyme.

Effect of recombination on plasmid population stability

Following the confirmation that Bxb1-RFP retains the enzymatic activity of the native recombinase, we verified whether changes in GFP expression levels were solely due to excision of the terminator sequence located between the P_Tet promoter and the RBS. This verification was essential to rule out the possibility that recombinase-mediated modifications to plasmid structure could influence plasmid copy number, thereby affecting gene expression.

To address this, the genetic construct GC2 was modified to generate a new variant, GC2b (Fig. 1), in which a constitutively expressed RFP gene was inserted downstream of the P_Tet promoter. This configuration enables the monitoring of changes in plasmid abundance by measuring RFP fluorescence. A new cell type, designated C3 (Fig. 1), was then engineered that contained constructs GC1b and GC2b. C3 cells were cultured both in the absence and presence of 10⁻⁴ M arabinose, and RFP fluorescence was quantified. As shown in Figure S2a, no statistically significant difference in red fluorescence was observed between conditions (t-test yielded a p-value of 0.086), indicating that arabinose-induced expression of Bxb1 does not substantially affect RFP levels. By contrast, Figure S2b shows that GFP fluorescence was significantly elevated in the presence of arabinose (p = 0.009), consistent with excision of the terminator sequence and subsequent activation of GFP expression.

Together, these results demonstrate that although arabinose induction of Bxb1 leads to recombination and increased GFP expression, it does not significantly alter RFP levels. This suggests that plasmid population abundance remains stable regardless of recombination events and that structural modifications to the plasmid do not impact overall plasmid copy number.

Quantification of the relationship between recombinase abundance and recombination efficiency during exponential growth

To assess recombination efficiency, it is necessary to establish a reference of maximal recombination efficiency. For this purpose, a third genetic construct, GC2c, was generated using the high-copy-number plasmid pSB1AC3. This construct contains the GFP gene under the control of the P_Tet promoter, but with an attL sequence inserted between the promoter and the RBS. The attL site typically arises from Bxb1 recombinase-mediated recombination between attP and attB sites. Thus, GC2c simulates a fully recombined state, in which GFP expression is independent of recombinase activity and represents the maximum achievable fluorescence level.

A new cell type, C4 (Fig. 1), was constructed by co-transforming cells with GC2c and GC1. Although GFP expression in GC2c does not require recombinase activity, inclusion of GC1 accounts for any potential effects of Bxb1 expression on GFP levels, such as metabolic burden. Recombination efficiency can therefore be evaluated during the exponential phase by comparing GFP expression in C1 cells with that in C4 cells.

To standardize promoter induction while modulating recombinase levels, all experiments were conducted with a constant arabinose concentration of 10⁻⁴ M to ensure maximal P_BAD promoter activation. Recombinase abundance was varied by altering the duration of arabinose exposure prior to transferring cultures to arabinose-free medium (see Methods for details).

Figure S3 shows RFP levels, which serve as a proxy for Bxb1 expression, as a function of induction time (left axis). Notably, recombinase levels plateaued after 12 h of induction. This steady-state level is defined as the maximum recombinase abundance, R_max

Based on these data, the relative recombinase abundance µ as a function of induction time was calculated using the following equation (see Methods for details):

1

where R(t) represents the RFP fluorescence measured after an induction period t with 10⁻⁴ M arabinose, and R_max corresponds to the maximum fluorescence level observed after 12 h of induction. The right axis in Figure S3 displays the calculated values of µ at various induction times.

2

Finally, the recombination efficiency, denoted λ(µ), can be computed using the following expression:

where F(µ) represents the GFP fluorescence measured in cell type C1 at a given relative recombinase abundance µ, while F_max(µ) corresponds to the GFP fluorescence measured in cell type C4 under the same µ. This relationship enables quantification of recombination efficiency by comparing GFP expression across varying levels of recombinase accumulation.

To experimentally determine the relationship between µ and λ during the exponential growth phase, a stationary phase culture of cell type C1 grown in LB medium without arabinose was prepared and diluted 1:1000 in LB medium containing 10⁻⁴ M arabinose (see Methods for details). At hourly intervals, 1 mL culture samples were harvested and centrifuged to separate the cell pellet. Each pellet was then resuspended in 1 mL of fresh LB medium lacking arabinose. Finally, 1 µL of this resuspension was used to inoculate 5 mL of arabinose-free LB medium, and the cultures were incubated at 37 °C with shaking for 12 h.

This procedure generated a series of cultures in which the initial inocula had been exposed to arabinose for different durations, resulting in varying levels of intracellular recombinase accumulation. The final dilution step ensured that the subsequent cell growth occurred predominantly during the exponential phase, thereby allowing recombination events to occur predominantly during this phase.

The same procedure was applied to cell type C4 to determine F_max(µ) under each corresponding µ. Once cultures reached stationary phase, both RFP and GFP fluorescence were measured. Using Eqs. (1) and (2), values of µ and λ were calculated for each culture. Figure S4 shows GFP fluorescence as a function of recombinase induction time, along with the derived recombination efficiencies λ. Different induction times produced distinct levels of intracellular Bxb1, as assessed by measuring the fluorescence of the Bxb1-RFP fusion protein. Figure 2a shows the relationship between relative recombinase abundance µ and recombination efficiency λ. Three distinct regimes can be observed in Fig. 2a. In Region I (µ < 8%), recombination efficiency is practically zero. This is likely due to the dilution of intracellular recombinase through cell division and the effects of protein degradation, which together reduce recombinase levels below the threshold required for activity. In Region II (µ between 8% and 75%), a near-linear relationship between µ and λ is observed. The orange dashed line represents a linear fit of the data within this region (r² = 0.993), indicating that an increases in recombinase abundance directly translate to proportional increases in recombination efficiency. Finally, in Region III (µ > 75%), recombination efficiency saturates at approximately 90%, suggesting that most available recombination sites have been modified and additional recombinase no longer increases efficiency. Importantly, GFP expression levels in C4 remained stable across all µ values (see Figure S4), confirming that increased recombinase expression did not impose a measurable metabolic burden on GFP expression in the control construct.

Fig. 2.

(a) Relationship between relative recombinase abundance µ and recombination efficiency λ. Three distinct regions are observed. In Region I, no significant GFP expression is detected, indicating that intracellular recombinase levels are insufficient to trigger recombination-mediated gene expression. In Region II, a quasi-linear relationship is observed between µ and λ (r² = 0.993). In Region III, a plateau is reached, corresponding to the maximum achievable recombination efficiency. Data points represent the mean of three independent experiments; error bars indicate standard deviation. (b) Electrophoresis assay showing the presence of two plasmid populations: recombined and non-recombined. The right panel illustrates the genetic architecture of the GC2 construct before and after Bxb1 induction. The GFP expression cassette is flanked by EcoRI and PstI restriction sites. In the presence of these enzymes, the plasmid is digested, generating a DNA fragment that appears as a band on the agarose gel. The position of the band reflects the size of the digested DNA fragment. The figure shows the emergence of two distinct bands: an upper band corresponding to the non-recombined genetic configuration, and a lower band corresponding to the recombined architecture. The appearance of the recombined band increases with the duration of Bxb1 induction, while the intensity of the upper band diminishes, indicating a growing proportion of recombined plasmids.

To verify that the dependence of GFP expression on recombinase levels arises from recombinaseinduced DNA changes, plasmid DNA was extracted from cultures subjected to various Bxb1 induction times and analyzed by agarosegel electrophoresis to assess resulting DNA band patterns (see the Materials section for detailed procedures). Figure 2b shows these patterns. The GC2 construct, serving as a reporter for Bxb1-mediated excision events, is flanked by EcoRI and PstI restriction sites (Fig. 2b, right panel). Digestion of the DNA with these two restriction enzymes yields band patterns that are dependent on the duration of Bxb1 induction.

As depicted in the figure, two distinct bands correspond to the GC2 construct: the upper band represents the unrecombined construct, while the lower band corresponds to the recombined construct, which is smaller due to the excision of the sequence located between the attP and attB sites. In the absence of arabinose (Bxb1 induction time t = 0 h), only the upper band is observed, indicating that the unrecombined construct predominates in the plasmid population. However, as induction times increase, the intensity of the lower band, representing the recombined construct, becomes more pronounced at the expense of the upper band. This suggests the coexistence of two plasmid populations: those in which recombination has not yet occurred and those in which it has. The proportion of the latter increases with longer recombinase induction times. Consequently, the abundance of these populations can be modulated based on Bxb1 levels, thereby influencing GFP expression levels.

To verify that the appearance of these two bands corresponded to distinct plasmid populations, recombined (lower band) and non-recombined (upper band), DNA was extracted from each band and the region of interest (between the promoter and the RBS) was sequenced (see Materials and Methods for detailed procedures). Table S2 lists the primers used for sequencing the extracted DNA, together with the obtained sequences. As shown, sequences corresponding to the attP and attB sites flanking the terminator sequence were identified in DNA from the upper bands. In contrast, neither these sites nor the terminator sequence were present in DNA from the lower bands; instead, the attL sequence was detected, consistent with the expected recombination event.

To assess whether increasing Bxb1 induction times correlated with a higher number of recombined genetic sequences, we quantified the abundance of attL sequences resulting from recombination events. Quantification was performed using real-time fluorescent Loop-mediated Isothermal Amplification (qLAMP)²⁸. Table S3 shows the set of primers specifically designed for qLAMP (see Materials and Methods for detailed procedures). These primers were designed to selectively hybridize with recombined sequences, but not with non-recombined ones. Specifically, the BIP_P20 primer can only hybridize to recombined sequences containing the attL site, but not to non-recombined sequences. The lack of BIP_P20 hybridization prevents LAMP amplification, ensuring that amplification occurs exclusively in recombined sequences.

Figure S5a shows the qLAMP fluorescence curves over the amplification time for cultures exposed to different Bxb1 induction times (0 h, 2 h, 4 h, 6 h, 8 h, 10 h, and 12 h). From these fluorescence curves, the so-called Threshold time (Tt) can be defined, i.e., the time required to reach a defined fluorescence level. Typically, Tt can be defined as the mean basal signal plus 10 times the standard deviation of this baseline²⁹. The dashed line in Figure S5a represents this fluorescence threshold. The use of Tt enables quantification of the abundance of the target genetic sequence: the higher the abundance, the lower the Tt value. Figure S5b depicts the relationship between Tt values at different Bxb1 induction times and GFP levels expressed as a consequence of excision of the terminator region in the GC2 genetic construct. A clear inverse correlation is observed: lower Tt values correspond to higher GFP levels. These results demonstrate that the variations in GFP levels as a function of Bxb1 induction time are a direct consequence of Bxb1-mediated recombination events, indicating that fluorescence measurements provide a reliable indirect readout of recombination efficiency.

These findings demonstrate that gene expression can be modulated in a quasi-linear fashion by tuning recombinase accumulation via temporally controlled induction within a given range. Given the irreversible nature of recombinase-mediated genetic modifications, this strategy provides a precise method for adjusting constitutive gene expression levels. Notably, it offers finer control than traditional approaches based on combining promoters and RBS of varying strengths.

Effect of exponential phase duration on Bxb1-mediated gene expression

Previous results demonstrated that intracellular accumulation of Bxb1 reliably determines recombination efficiency and, consequently, the level of associated gene expression. However, the duration of the exponential phase is also critical for gene expression and may therefore influence recombination efficiency.

To assess this effect, recombinase expression was induced for 5 h in a culture of cell type C1 before it transitioned into stationary phase, during which plasmid-driven gene expression is known to decline significantly. Specifically, a saturated culture of C1 cells grown without arabinose was prepared and divided into two equal volumes. The first volume was centrifuged, and the resulting supernatant was filtered through a 0.22-µm membrane to remove residual cells, yielding a nutrient-depleted, cell-free, arabinose-free supernatant.

The second volume was diluted with fresh LB medium to an optical density (OD) of 0.45, and 10⁻⁴ M arabinose was added to induce recombinase expression. After 5 h of growth at 37 °C with shaking, the culture reached stationary phase. At that point, the cells were harvested by centrifugation and resuspended in the arabinose-free supernatant from the first volume. This procedure generated a stationary-phase culture of C1 cells that had experienced 5 h of recombinase induction during exponential growth and was subsequently maintained in a nutrient-depleted, arabinose-free environment.

To further investigate the impact of exponential phase duration, this stationary-phase culture was diluted into fresh LB medium at varying concentrations, generating a series of cultures with identical intracellular levels of Bxb1-RFP but different initial cell densities. These cultures were incubated overnight at 37 °C with shaking. Figure 3 shows recombination efficiency λ as a function of the initial OD of the cultures. The results show that although all cultures initially contained equivalent amounts of recombinase and no additional Bxb1 production occurred (due to the absence of arabinose), the observed λ values strongly depended on the initial cell density. Considering the relationship between the initial culture concentration and duration of exponential phase, shown in Figure S6, these differences indicated that cultures with higher initial concentrations have a shorter exponential phase, which results in lower GFP expression due to the culture entering stationary phase. As the initial concentration of the culture decreased and thus the duration of exponential phase increased, λ values rose.

Fig. 3.

Effect of initial optical density (OD) on recombination efficiency λ. Although all cultures initially contained equivalent intracellular recombinase levels, and no further recombinase was produced during the exponential phase due to the absence of arabinose, λ values varied significantly as a function of initial culture density. Data represent the mean of three independent experiments; error bars indicate standard deviation.

However, these findings do not clarify whether the dependence of GFP expression on exponential phase duration arises from differences in recombination efficiency itself, i.e., the number of recombined sites, or from reduced expression of the reporter gene during stationary phase. It remains possible that recombination events continue to occur during stationary phase, but that GFP expression is suppressed in this phase.

Analysis of recombinase activity during the stationary phase

To determine whether Bxb1 recombinase retains its activity during the stationary phase, even though recombination at this stage does not immediately lead to GFP expression, an experiment was performed using the protocol previously described. A saturated culture of cell type C1 grown without arabinose was diluted to an OD of 0.45, and 10⁻⁴ M arabinose was added to induce recombinase expression. The culture was incubated at 37 °C for 5 h, until it re-entered stationary phase. At this point, cells were pelleted, the supernatant was discarded, and the cells were resuspended in the arabinose-free, nutrient-depleted supernatant obtained from the original saturated culture. This ensured that cells remained in stationary phase, without external arabinose or fresh nutrients. At this stage, the cells contained an intracellular pool of recombinase that had accumulated during the 5-h induction. The culture was then incubated at 37 °C with shaking, and at defined time points, 1 µL aliquots were sampled and diluted into 5 mL of fresh, arabinose-free LB medium. These cultures, which had remained in stationary phase for varying durations, were incubated overnight at 37 °C with shaking before measuring GFP fluorescence.

If recombinase remains active during the stationary phase, the number of recombined sites should increase over time, even though GFP levels remain unchanged during this phase. Upon transfer to fresh medium and re-entry into exponential growth, the accumulated recombination events would result in corresponding increases in GFP expression. Conversely, if recombinase activity ceases during stationary phase, GFP levels should remain unchanged regardless of the duration of this phase.

Figure 4 shows recombination efficiencies λ across the different time points. Blue bars represent λ values measured in the stationary-phase cultures prior to dilution, while orange bars show λ after overnight growth in fresh medium. While λ values in stationary-phase cultures remained low (~ 18%) and changed only slightly with time, cultures re-exposed to fresh medium exhibited a marked increase in λ, reaching up to 100% after 15 h in stationary phase. These results indicate that recombination events accumulated progressively during the stationary phase, confirming that recombinase activity is maintained despite the absence of GFP expression.

Fig. 4.

Recombination efficiency λ as a function of stationary phase duration. Blue bars represent λ values measured in C1 cultures following Bxb1 induction and incubation for varying durations in stationary phase. Orange bars represent corresponding cultures that were returned to exponential phase by dilution in fresh medium after the same stationary-phase durations. The increase in λ upon re-entry into exponential phase indicates that Bxb1 remains active during stationary phase, even though GFP expression is not observed under these conditions. Data represent the mean of three independent experiments; error bars indicate standard deviation.

Interestingly, when comparing cultures with identical induction periods, recombination efficiency was significantly higher in cultures that transitioned to stationary phase than in those that remained in exponential growth. For example, after 4 h of induction, λ reached 30.6% in exponentially growing cultures (Figure S4), but increased to 73.7% in cultures that entered stationary phase and were later returned to exponential growth (Fig. 4).

This effect likely results from the dilution of intracellular recombinase due to cell division in exponentially growing cultures. By contrast, when cultures enter stationary phase following induction, recombinase levels decline only due to proteolytic degradation, which occurs more slowly. As a result, the intracellular recombinase pool is better preserved in stationary phase, enabling higher recombination efficiency. These findings suggest that recombination occurring during the stationary phase may be more efficient than during exponential growth.

To determine the minimum induction time required before entering stationary phase to achieve maximal recombination, an additional experiment was conducted. Starting from a saturated culture without arabinose, different dilutions were prepared and supplemented with 10⁻⁴ M arabinose to induce recombinase expression. The duration of the exponential phase prior to stationary phase onset was determined by the initial cell density: lower dilutions corresponded to shorter induction periods. Cultures were prepared with 1, 2, 3, and 4 h of induction. After each induction period, arabinose was removed and cells were transferred to the arabinose-free supernatant from the original culture to maintain stationary-phase conditions. After 20 h in stationary phase, 1 µL of each culture was diluted into 5 mL of fresh LB medium and incubated overnight at 37 °C to allow GFP expression.

The results presented in Fig. 5a indicate that a 1-h induction period was insufficient for efficient recombination. However, extending the induction to 2 and 3 h yielded λ values of 86% and 100%, respectively. Figure 5b illustrates the impact of varying Bxb1 induction durations on the plasmid population. The presence of two distinct bands in the agarose gel electrophoresis corresponds to two subpopulations: the upper band represents unrecombined plasmids, while the lower band corresponds to recombined plasmids, as detailed in the right panel of the figure. As the Bxb1 induction time increases, there is a noticeable shift favoring the recombined plasmid population, as evidenced by the intensifying lower band and diminishing upper band.

Fig. 5.

(a) Impact of Bxb1 induction duration prior to stationary phase on recombination efficiency λ. Recombination efficiency λ (orange bars) increases with longer Bxb1 induction times before the culture transitions into stationary phase, reflecting greater accumulation of intracellular recombinase before cell division ceases. Data represent the mean of three independent experiments; error bars indicate standard deviation. (b) Electrophoresis assay showing the presence of two plasmid populations: recombined and non-recombined. The right panel illustrates the genetic architecture of the GC2 construct before and after Bxb1 induction. The GFP expression cassette is flanked by EcoRI and PstI restriction sites. In the presence of these enzymes, the plasmid is digested, generating a DNA fragment that appears as a band on the agarose gel. The position of the band reflects the size of the digested DNA fragment. The figure shows the emergence of two distinct bands: an upper band corresponding to the non-recombined genetic configuration, and a lower band corresponding to the recombined architecture. The appearance of the recombined band increases with the duration of Bxb1 induction, while the intensity of the upper band diminishes, indicating a growing proportion of recombined plasmids. It is worth noting that the time scale for the appearance of the recombined band is shorter than that shown in Fig. 2b.

Notably, when comparing Fig. 5b to Fig. 2b, it is evident that recombination events occurring during the stationary phase result in a more pronounced accumulation of recombined plasmids than those during the exponential phase, even under identical induction durations. This observation suggests that relatively brief induction periods can suffice to accumulate recombinase levels that support maximal recombination, provided the recombination transpires during the stationary phase, where Bxb1 levels remain undiluted due to the cessation of cell division.

Discussion

Recombinases capable of mediating DNA modifications in response to external stimuli hold considerable promise in synthetic biology for the construction of complex genetic circuits with broad applications. These enzymes can be reprogrammed to respond to specific inputs, enabling precise modulation of cellular behavior. In many such applications, the recombination efficiency, defined as the fraction of modified recombination sites relative to the total available, must be tightly controlled. In some contexts, partial recombination (i.e., coexistence of recombined and non-recombined plasmids) can disrupt circuit behavior, while in others, controlled partial recombination may be desirable. Therefore, understanding the parameters that influence recombination efficiency is critical for designing predictable and controllable biological systems.

An important finding of this study is that recombination efficiency is influenced not only by the expression level of the recombinase but also by external factors, particularly the growth phase of the culture. The results demonstrate that recombination efficiency can be finely tuned by modulating the duration of exposure to the inducing signal. However, the phase during which the inducer is applied plays a crucial role in determining the final outcome of gene expression.

When recombinase expression is induced during exponential growth, gene expression correlates with recombination efficiency. Specifically, for Bxb1, a quasi-linear relationship was observed between the duration of induction and GFP output over a defined range. This allows precise temporal control over recombination outcomes by adjusting the length of exposure to the inducer (e.g., arabinose), thereby defining the fraction of plasmids undergoing recombination. Importantly, transient induction of Bxb1 during exponential phase proved more effective in driving gene expression than continuous induction throughout the same phase.

However, when the culture enters stationary phase before recombination is completed, gene expression becomes uncoupled from recombination efficiency due to the repression of plasmid-borne gene expression in this phase. Thus, both the level and the timing of recombinase induction are critical to ensure that gene expression accurately reflects the recombination state.

Notably, while recombinase activity persists in stationary phase, its effects are not immediately observable at the level of reporter gene expression. Yet, when these stationary-phase cultures are returned to exponential growth, recombination events that occurred during stationary phase result in increased gene expression. Because cell division is limited and plasmid replication is reduced during stationary phase, recombinase levels are not diluted, and recombination efficiency can reach 100% with lower enzyme concentrations than would be required during the exponential phase.

This observation supports a strategy to maximize recombination efficiency by inducing recombinase expression shortly before the onset of stationary phase. In our system, Bxb1 induction 2 and 3 h prior to stationary phase yielded recombination efficiencies of 86% and 100%, respectively, whereas induction after entry into exponential phase resulted in significantly lower efficiencies (3.4% and 12%, respectively). Implementing a two-phase strategy, comprising recombinase induction during exponential growth followed by maintenance in stationary phase, promotes efficient recombination. Reintroducing the culture into exponential growth then enables the selective expansion of cells harboring the desired modifications. This approach may be especially valuable in synthetic biology applications requiring robust and precise genetic editing, such as biosensor design, circuit construction, or therapeutic protein production.

While these results provide valuable insights into recombinase-based gene regulation, several limitations should be acknowledged. First, the experiments were performed in E. coli using Bxb1 as a model recombinase. Although both are commonly employed in synthetic biology, broader validation is needed across additional organisms and recombinases to assess the generality of these findings. Second, the study focused exclusively on one type of recombination event: DNA excision. Expanding this analysis to other recombination mechanisms (e.g., inversion, integration) would strengthen the applicability of the conclusions. Third, all genetic constructs in this study were plasmid-based. While plasmids are widely used for the development and testing of synthetic circuits because of their simplicity and high copy number, genome-integrated systems are preferred in applications requiring long-term genetic stability. Thus, further studies are needed to assess how recombinase induction and growth phase affect recombination efficiency and gene expression in chromosomally-integrated circuits.

Methods

Strains, media, and growth conditions

Escherichia coli TOP10 (Invitrogen, USA) was used for all cloning and expression experiments. Cultures were grown in LB medium at 37 °C with shaking at 200 rpm, and appropriate antibiotics were added for selection (35 µg/mL chloramphenicol or kanamycin; Sigma, USA). Bacterial strains were stored in LB medium supplemented with 20% (v/v) glycerol at − 80 °C.

All cultures were initiated from single colonies obtained from streaked glycerol stocks. These colonies were inoculated into fresh LB medium containing the appropriate antibiotic and incubated overnight at 37 °C with shaking at 200 rpm. The induction medium consisted of LB supplemented with the corresponding antibiotic and L-arabinose (Sigma).

All results represent the means of three independent experiments. Error bars indicate standard deviations.

Construction of genetic circuits

Genetic circuits were assembled using the BioBrick assembly method (Ginkgo Bioworks, USA). Constructs were cloned into one of two backbone plasmids: pSB1AC3 (high-copy number, ampicillin and chloramphenicol resistance) or pSB3K3 (low-copy number, kanamycin resistance). All transformations were performed using a chemical transformation protocol, and construct sequences were verified by Sanger sequencing.

Recombinase induction in cell cultures

Cultures were initiated from single colonies of the cell types obtained from streaked glycerol stocks and inoculated into LB medium containing 50 µg/mL kanamycin and chloramphenicol. To induce recombinase expression, 10⁻⁴ M L-arabinose was added to the medium. Cultures were incubated overnight at 37 °C with shaking at 200 rpm.

Fluorescence quantification in liquid cultures

GFP fluorescence was measured at an excitation wavelength of 485 ± 20 nm and an emission wavelength of 528 ± 20 nm using a Synergy MX microplate reader (BioTek Instruments, USA). Absorbance was measured at 600 nm. RFP fluorescence was measured at an excitation wavelength of 558 ± 20 nm and an emission wavelength of 583 ± 20 nm.

To minimize background interference from the LB medium, cultures were centrifuged at 14,000 rpm for 10 min prior to measurement. The supernatant was discarded, and the cell pellet was washed three times with phosphate-buffered saline (PBS) and finally resuspended in PBS. This procedure eliminates background fluorescence associated with the LB medium, thereby ensuring more accurate and reliable quantification.

In all cases, sample absorbance OD(S) and fluorescence f(S) were corrected using background signals from blank controls, OD(B) and f(B), respectively. The corrected fluorescence for reporter protein Θ was calculated using the following equation:

3

Quantification of relative recombinase abundance µ

The relative abundance of the recombinase, denoted as µ, was quantified using cell types C1 and C4, based on the following equation:

4

in which R(t) was calculated using the equation:

5

where RFP(t) represents the red fluorescence measured in cell type C1 following induction with 10⁻⁴ M arabinose for a duration t, RFP₀ denotes the background fluorescence, OD(t) is the optical density of the culture at 600 nm, and OD₀ corresponds to the background optical density measured at the same wavelength, and was calculated using the equation:

6

where RFP_max is the red fluorescence measured in cell type C4 after 12 h of induction with 10⁻⁴ M arabinose, corresponding to the maximum accumulation of recombinase, and OD_max is the optical density at 600 nm of this culture.

Quantification of recombination efficiency λ

The recombination efficiency, denoted λ, mediated by Bxb1-RFP was calculated by measuring GFP fluorescence levels using the following equations:

7

in which was calculated using the equation:

8

where GFP(µ) represents the green fluorescence measured in cell type C1 at a given recombinase abundance µ, GFP₀ is the background fluorescence, OD(µ) is the optical density of the culture at 600 nm, and OD₀ is the background optical density measured at the same wavelength, and F_max was calculated using the equation:

9

where GFP_max represents the green fluorescence measured in cell type C4, and OD_max is the optical density at 600 nm of this culture.

Induction of recombinases for varying time periods

Single colonies of cell types C1 and C4, obtained from streaked glycerol stocks, were inoculated into 5 mL of fresh LB medium supplemented with kanamycin and chloramphenicol (50 µg/mL each) and incubated overnight at 37 °C with shaking at 200 rpm.

The following day, saturated cultures of C1 and C4 were diluted by adding 40 µL of culture into 3,960 µL of fresh LB medium containing the same antibiotics. Arabinose was added to a final concentration of 10⁻⁴ M, and cultures were incubated at 37 °C with shaking at 200 rpm.

Every 60 min, a 1-mL aliquot was collected from each culture and centrifuged to separate the cell pellet from the supernatant. The pellet was resuspended in 1 mL of fresh LB medium without arabinose. Then, 1 µL of this resuspension was used to inoculate 5 mL of LB medium containing kanamycin and chloramphenicol, but lacking arabinose. These cultures were incubated overnight at 37 °C with shaking at 200 rpm.

Effect of exponential phase duration on Bxb1-mediated gene expression

Single colonies of cell type C1, obtained from streaked glycerol stocks, were inoculated into 40 mL of LB medium supplemented with kanamycin and chloramphenicol and incubated overnight at 37 °C with shaking at 200 rpm. The overnight culture was then divided equally into two 20-mL aliquots. Each aliquot was centrifuged at 14,000 rpm for 10 min, and the supernatant was collected and filtered through a 0.22-µm Millipore^® filter to ensure complete removal of cells.

One of the aliquots was diluted with fresh LB medium to reach an OD of 0.45 and supplemented with 10⁻⁴ M arabinose to induce Bxb1 expression. The culture was maintained at 37 °C with shaking at 200 rpm. OD at 600 nm was measured every 30 min until the culture re-entered stationary phase. At that point, the cells were pelleted by centrifugation (14,000 rpm, 10 min) and resuspended in the arabinose-free, cell-free supernatant obtained from the other aliquot.

Finally, a series of new cultures was prepared at different dilutions. These cultures were incubated overnight at 37 °C with shaking at 200 rpm.

Agarose gel electrophoresis assays

To analyze the structural changes in DNA induced by Bxb1, agarose gel electrophoresis assays were conducted. Specifically, various cultures were prepared as previously described, and plasmid DNA was extracted using the E.Z.N.A.^® Plasmid DNA Mini Kit I (Omega BioTek, USA). Subsequently, the extracted DNA underwent digestion with the restriction enzymes EcoRI and PstI (New England Biolabs) at 37 °C for 1 h, followed by enzyme inactivation at 80 °C for 20 min. A 0.8% agarose gel was then prepared in Tris-Borate-EDTA (TBE) buffer (w/v), and 1 µL of SYBR Safe Gel Stain (Invitrogen, USA) was added per milliliter of gel solution. Finally, the digested DNA samples from the different cultures were loaded onto the agarose gel, and electrophoresis was performed at 80 volts for 45 min.

DNA extraction, sequencing, and real-time fluorescence LAMP quantification

From agarose gel electrophoresis, DNA sequences can be separated according to their size, which enables discrimination between non-recombined and recombined plasmid populations, the latter being shorter due to the excision of the terminator region. Once resolved on the agarose gel, DNA corresponding to both types of sequences was carefully excised and purified using the QIAquick Gel Extraction Kit (QIAGEN, Germany), following the manufacturer’s instructions.

For sequencing of the different DNA samples, two primers were designed (Table S2) using the PrimerQuest™ Tool (Integrated DNA Technologies, IDT, USA). Primer synthesis was also carried out by IDT. The purified DNA samples together with the primers were then submitted to Eurofins Genomics (Germany) for Sanger sequencing.

For the quantification of recombination events by real-time fluorescence Loop-mediated Isothermal Amplification (qLAMP), plasmid DNA was extracted from cell cultures subjected to different Bxb1 induction times. Plasmid DNA was isolated using the E.Z.N.A.^® Plasmid DNA Mini Kit I (Omega Bio-Tek, USA) according to the supplier’s protocol. Subsequently, amplification and quantification were performed using the WarmStart^® Fluorescent LAMP/RT-LAMP Kit (New England BioLabs, USA) in a real-time fluorescence detection system.

Fluorescence emitted by the LAMP Fluorescent Dye, an intercalating dye that enables monitoring of LAMP reactions in real time, was measured at an excitation wavelength of λₑₓ = 497 nm and an emission wavelength of λₑₘ = 520 nm.

Primers for qLAMP (Table S2) were designed with the NEB LAMP Primer Design Tool, ensuring specific hybridization to recombined sequences while avoiding non-recombined templates.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

M.G.-C. and J.M. designed the circuits. M.G.-C. and J.M. performed the experiments. J.M. and M.G.-C. wrote the paper.

Funding

This study and the open access charge were funded by grants from the Spanish Ministry of Economy and Competitiveness [PID2023-151156NB-I00/MICIUN/AEI/10.13039/501100011033/FEDER, UE]. This work was supported by “Unidad de Excelencia María de Maeztu” and funded by the AEI (CEX2018-000792-M).

Data availability

Data are provided within the manuscript and supplementary information files.

Declarations

Competing interests

The authors declare no competing interests.

Supplementary Information

The Supplementary Material for this article can be found online at: