Construction of an Oscillator Gene Circuit by Negative and Positive Feedbacks
Construction of an Oscillator Gene Circuit by Negative and Positive Feedbacks
Journal of Microbiology and Biotechnology. 2016. Feb, 26(1): 139-144
Copyright © 2016, The Korean Society For Microbiology And Biotechnology
  • Received : July 06, 2015
  • Accepted : September 16, 2015
  • Published : February 28, 2016
Export by style
Cited by
About the Authors
Shihui Shen
Yushu Ma
Yuhong Ren
Dongzhi Wei

Synthetic oscillators are gene circuits in which the protein expression will change over time. The delay of transcription, translation, and protein folding is used to form this kind of behavior. Here, we tried to design a synthetic oscillator by a negative feedback combined with a positive feedback. With the mutant promoter PLacC repressed by LacIq and PLux activated by AHL-bound LuxR, two gene circuits, Os-LAA and Os-ASV, were constructed and introduced into LacI-deleted E. coli DH5α cells. When glucose was used as the carbon source, a low level of fluorescence was detected in the culture, and the bacteria with Os-ASV showed no oscillation, whereas a small portion of those carrying Os-LAA demonstrated oscillation behavior with a period of about 68.3 ± 20 min. When glycerol was used as the carbon source, bacteria with Os-ASV demonstrated high fluorescence value and oscillation behavior with the period of about 121 ± 21 min.
Synthetic biology is a new biological field where natural or artificial biodevices are used to construct new biological systems or to imitate natural systems for study or application [2] . In the past decades, different gene circuits such as genetic toggle switches [12] , genetic counters [8] , and various oscillators [5 , 7 , 9 , 11 , 18 , 20 , 22] have been designed and constructed. These devices have shown potential applications in many fields, including biofuel production, disease diagnosis, and gene therapy [16] . Among them, the synthetic oscillators are simplified models of a biological rhythm network, so that the mechanism and factors influencing on biological rhythm can be studied indirectly with the help of such gene circuits [7] . Moreover, oscillators can also be used as controllers of more complicated genetic circuits to adjust the expression of different genes [21] .
The simplest oscillator is a negative feedback composed of a self-inhibited gene, where the oscillation is generated by the delay of transcription, translation, and folding of the gene, and its corresponding mRNA and protein, respectively, [22] . The Goodwin oscillator is a typical oscillator model in which a repressor is used to repress its own gene to generate oscillation [13] . Such an oscillator was successfully constructed with PLac and its repressor protein LacI [18] . This kind of oscillators is not robust and the condition for oscillation is rigorous. For example, if PLac and LacI were replaced with PTet and TetR, respectively, no oscillation could be observed [3] . This might be associated with the different times needed for transcription repression. As the activated form of TetR is a dimer, whereas the activated form of LacI is a tetramer, LacI needs more time to repress the PLac promoter [4] .
Alternatively, multiple genes can be used for oscillator construction as well. The repressors LacI, TetR, and cI, together with the promoters regulated by them, have been used to construct a so-called repressilator, a gene circuit that generates oscillation by repressing the expression of one another [7] . The main principle for a multigene oscillator is that the finally expressed protein should repress the transcription of the first gene in order to form a negative feedback loop, so that an odd number of negative feedbacks should be used for such a purpose [19] . Furthermore, positive feedbacks can also be found in oscillator construction. Theoretically, oscillators constructed by negative feedbacks combined with positive feedbacks are more robust, and are more insensitive to parameter changes [21] .
Hybrid promoters had been used in most oscillators constructed by negative feedbacks combined with positive feedbacks. For example, by using a hybrid promoter activated by AraC and repressed by LacI, an oscillator with the oscillating period of 40 min was constructed [22] . In this study, we tried to use promoters that only response to one regulator, activator or repressor, to construct and analyze a genetic oscillator.
Materials and Methods
- Strains, Plasmids, and Chemicals
Escherichia coli DH5α competent cells were bought from TransGen Biotech. Plasmid pUC19 was bought from Takara. T4 ligase and Fastdigest restriction enzymes were the products of Thermo Scientific. PrimeStar Max and Hi-FI DNA polymerases were bought from Takara. Low melting agarose, casamino acid, and thiamine were bought from Solarbio Life Sciences. Acylhomoserine lactone (AHL) was brought from Sigma-Aldrich. LArabinose was bought from Shanghai Qianchen Biological Technology Co., Ltd. Other reagents were analytically pure.
- ΔPLacPlasmid Construction
The original PLac promoter was deleted from pUC19 by inverse PCR to construct the pUC19- plasmid. The following primers with Afl II restriction sites (underlined) were used: sense: 5’- CGGAA CTTAAG AAAGAACATGTGAGCAAAAGGCCAG-3’; antisense: 5’- CGGAA CTTAAGAAG CTTGCATGCCTGCAGGTC -3’.
After restriction, ligation, and transformation, the colonies were selected and the pUC19 plasmid was extracted for further construction.
- Oscillator Plasmid Construction
The design of the gene circuit and the structure of the oscillator plasmid are shown in Figs. 1 A and 1 B, respectively. For this purpose, the PLacC promoter was amplified from pUC19 with a point mutation (T14C) incorporated in its operator, by PCR. The repressor gene lacIq was obtained from E. coli BL21 (DE3). The PLux promoter and luxR were synthesized according to the sequences reported (GenBank: Y00509.1 and M19039.1, respectively). The reporter gene egfp was preserved in our laboratory. The RBS (underlined) and its downstream sequence AAGGAG ATATACAT were obtained from pET28a. Terminator rrnBT1T2 from pBV220 was used to separate promoters and the genes they controlled. Different ssrA -mediated tags were added to the C-termini of EGFP and LacIq to accelerate their degradation. Among them, EGFP was fused with AAV tag (5’-GCAGCAAACGACGAAAACTACGCTGCAGCAGTT-3’) with a medium half-life period, while LacIq was fused with two different tags. When ASV tag (5’- GCAGCAAACGACGAAAACTACGCTGCATCAGTT-3’) with a long half-life period was added, the circuit was designated as Os-ASV, and LAA tag (5’-GCAGCAAACGACGAAAACTACGCTTTAGCAGCT-3’) with a short half-life period was added in the circuit designated as Os-LAA [1] .
PPT Slide
Lager Image
Oscillator gene circuit design (A) and construction (B).

() Expression of EGFP and LuxR under the control of leads to the increase of fluorescence signal and accumulation of LuxR. When the concentration of LuxR reaches a threshold, it can turn on the promoter together with AHL added separately in the system. Accumulation of LacIq, in turn, represses the activity, resulting in the decrease of fluorescence signal and decay of LuxR. With the decline of LuxR concentration, returns to its nonactive state and the synthesis of LacIq is inhibited. Consequently, the system returns to its initial state with the degradation of LacIq. Repetition of the cycle will form an oscillation behavior of fluorescence. () controls the transcription of and , while controls the transcription of . ssrA represents the -mediated tags, which can reduce the half-life period of target proteins. Terminator was used to separate the two promoters and the genes they controlled.

All the fragments were digested separately and ligated consequently to obtain the full-length sequences. After validation by DNA sequencing, the pUC19- plasmid carrying the gene circuit Os-ASV or Os-LAA was transformed into the E. coli DH5α Δ lacI strain for further analysis.
- lacIKnockout Strain Construction
In order to avoid the possible interference caused by LacI expressed by the host strain, the Red recombination system from λ phage was used to construct an E. coli DH5α Δ lacI strain. A DNA fragment containing the kanamycin resistance gene and LacZ’was amplified by PCR with the following primers: sense: 5’- GTATGGCATGATAGCGCCCGGAAGAGAGTCAATTCAGGGTGGTGAATGTG GCGCAACGCAATTAATGTGAGTTAG-3’; antisense: 5’- TGAGCTAACTCACATTAATTGCGTTGCGCTCACTGCCCGCTTTCCAGTCG TTAGAAA AACTCATCGAGCATCAAATG-3’. The resultant sequence was flanked by a 50 bp homologous arm (underlined sequence in the primers) of lacI gene at either side and used as a knockout linear fragment.
The recombination procedure was carried out in E. coli DH5α with plasmid pKD46 as previously described [18] . The blue colonies that grew in media with kanamycin, X-Gal, and IPTG were screened by PCR using the following primers complementary to the cross-sites of chromosome and the inserted fragment: sense: 5’-GGTGAATGTGGCGCAACGC-3’; antisense: 5’-GCTTTCCAGTCGTTAGAAAAACTC-3’.
- Observation and Data Processing
The bacteria containing the oscillator plasmids were grown in M9 culture (0.1% casamino acid, 200 μM thiamine, 0.4% glucose or 0.5% glycerol, 10 μM IPTG, 100 μg/ml ampicillin) at 37℃ until the exponential phase. The cells were washed twice with fresh medium without IPTG, and then resuspended and concentrated at room temperature in 100 μl of fresh medium containing 30 nM AHL and 0.5% low-melt agarose, which was previously maintained at 55℃. The concentrated culture was dropped on single-well depressed slides with 2% agarose pad with 30 nM AHL and solidified. The slides were then covered by cover glasses and sealed to prepare the samples for observation.
A Nikon Ti series inverted fluorescence microscope was used for observation. The samples were incubated at 37℃. by a heater. DIC and green fluorescence (FITC) images were taken every 5 min for 5-6 h.
For data processing, the fluorescence intensities of the corresponding cells in FITC images were measured by Image Pro Plus. The raw data were smoothed using adjacent-averaging in Origin.
- Gene Circuit Design
The gene circuit designed is shown in Fig. 1 A. The transcription of egfp-ssrA and luxR is under the control of promoter PLacC , whereas the transcription of lacIq-ssrA is under the control of promoter PLux . Terminator rrnBT1T2 was used to separate the two promoters and the genes they controlled. In order to reduce binding between PLac and LacIq, PLacC , a mutant promoter with a point mutation (T14C) in the operator, was used [15] . The lacI gene on the bacterial chromosome was knocked out by Red recombination to inhibit its expression. To avoid circuit distraction by protein accumulation, ssrA -mediated tags that can reduce the half-life period of target proteins [1 , 17] were added to C-termini of all proteins in the circuit, except LuxR, which is unstable in E. coli . AHL was added additionally at the beginning of observation to activate LuxR and subsequently initiate the transcription from promoter PLux [6 , 10] .
By our design, the expression of LuxR and EGFP under the control of PLacC leads to the accumulation of the proteins and detection of fluorescence signal. When the concentration of LuxR reaches a threshold, it could turn on the PLux promoter in the presence of AHL and initiate the expression of LacIq-ssrA, a repressor of PLacC . Repression of PLacC is followed by the degradation of EGFP-ssrA and LuxR, and the decline of the fluorescence signal. With the decrease of LuxR amount in cells, the PLux promoter is turned off and the system returns to its initial state with the degradation of LacIq. Repetition of the cycle will form an oscillation behavior of the fluorescence signal.
- Validation oflacIStrains
The Red recombination system from phage λ was used for lacI knockout in E. coli DH5α. Blue colonies were selected and further analyzed by PCR, with primers complementary to the sequences formed only after recombination between the linear fragment and the chromosome. Among 20 colonies analyzed, only strain No. 10 and 17 (designated as zkd10 and zkd17, respectively) gave positive results (data not shown). For additional functional characterization, the plasmid with egfp controlled by PLacC was transformed into zkd10, zkd17, and original DH5α strains, respectively. Analysis of the fluorescence signal in the presence and absence of IPTG showed that the fluorescence in zkd10 and zkd17 was much higher than that in wild-type DH5α, and IPTG had no notable influence on EGFP expression ( Fig. 2 ). Based on the results, zkd17 was used for further experiment.
PPT Slide
Lager Image
Fluorescence in ΔlacI and wild-type E. coli DH5α strains.

Wild-type DH5α (DH5α), and knockout strains zdk10 and zdk17 were transformed with plasmid pUC19- harboring under the control of . The fluorescence in bacteria was analyzed in the absence (IPTG-) or presence (IPTG+) of 800 μM IPTG.

- Fluorescence Observation with Glucose as Carbon Source
Although PLacC was a weak promoter when glucose was used in M9 medium as the carbon source, the growth profile of bacteria was much better than that when glycerol was added (data not shown). Recording the fluorescence signal under the conditions showed that the fluorescence of bacteria zkd17 with Os-ASV declined during the observation, and no obvious rise was noticed ( Fig. 3 A). Since there was no remarkable fluorescence decline (less than 21%) in the observation without AHL, this behavior was not caused by fluorescence decay but by the expression of LacIq-ASV.
PPT Slide
Lager Image
Fluorescence behavior in bacteria with Os-ASV (A and D) or Os-LAA (B and C) when glucose (A, B, and C) or glycerol (D) was used as carbon source.

Bacteria were grown in M9 medium with 0.4% glucose or 0.5% glycerol and 30 nM AHL at 37oC during detection. Images were captured every 5 min and analyzed by Image Pro plus. The different lines represent fluorescence in different bacteria in the same experiment. No oscillation of fluorescence signals was detected in bacteria with Os-ASV () and in most bacteria with Os-LAA () when glucose was used as carbon source. Fluorescence oscillation with a period of 68.3 ±20 min was detected in about 10% bacteria with the Os-LAA gene circuit when glucose was used as carbon source (). When glycerol was used as carbon source, oscillation in fluorescence with a period of 121 ± 21 min in bacteria with Os-ASV was recorded ().

When the strain zkd17 with Os-LAA was grown and observed under the same conditions, the fluorescence of most bacteria, unlike that of bacteria with Os-ASV, increased at first and then started to decline. Similar to that in strain zkd17 with Os-ASV, no obvious rise was noticed after that in most bacteria ( Fig. 3 B). Detailed analysis of hundreds of bacteria showed that the fluorescence fluctuation from about 10% bacteria with Os-LAA exhibited a typical oscillating behavior during decline, with a period of 68.3 ± 20 min ( Fig. 3 C).
- Fluorescence Observation with Glycerol as Carbon Source
As PLacC is a strong promoter in medium with glycerol as the carbon source, a stronger fluorescence was expected. The fluorescence behavior of zkd17 with the Os-ASV gene circuit in M9 with glycerol was recorded in 6.5h ( Fig. 3 D). The results showed that the fluorescence of most bacteria increased with fluctuation. The oscillation period was 121 ± 21 min.
When glucose was added as carbon source, the fluorescence in bacteria with Os-ASV declined continuously ( Fig. 3 A), whereas that in bacteria with Os-LAA rose at first and then declined ( Fig. 3 B). We supposed that the different pattern might be associated with the different degradation rate of LacIq mediated by ssrA tags fused to the C-termini of the protein. As the half-life period of LacIq-ASV is longer than that of LacIq-LAA, the repressor would accumulate faster in bacteria with Os-ASV than that in bacteria with Os-LAA. Consequently, compared with bacteria with Os-LAA, less time would be needed for LacIq-ASV to reach the concentration threshold for PLacC repression in bacteria with Os-ASV, leading to the fluorescence decline at the beginning of observation. In contrast, as LacIq-LAA degraded more rapidly, the concentration of the protein at the beginning of observation was insufficient to repress the PLacC promoter in bacteria with the Os-LAA circuit, and the fluorescence continued to increase over time until LacIq-LAA reached the threshold concentration.
PLacC was a weak promoter with a relatively low transcription rate when glucose was used as the carbon source. In the presence of the repressor LacIq-ssrA, the expression rate of EGFP might be lower than its degradation rate, resulting in the continuous fluorescence decline as observed in most cases. Only about 10% of bacteria with Os-LAA showed oscillation during the signal decline ( Fig. 3 C). We supposed that in a small fraction of bacteria with rapidly degrading LacIq-LAA, a temporary accumulation of EGFP could occur, which was responsible for the oscillation observed.
When glycerol was used instead of glucose, the fluorescence in most bacteria containing Os-ASV increased with oscillating behavior during observation ( Fig. 3 D). As there are two different states of oscillation, unstable (oscillating) and stable [7] , once such parameters as protein and mRNA concentrations fell into the stable state range, no oscillation occurs. Based on the results observed, we supposed that the unstable range of bacteria with Os-LAA was wider than that of bacteria with Os-ASV in the presence of glucose, whereas the unstable range of bacteria containing Os-ASV with glycerol was much larger than that of bacteria with Os-LAA when glucose was used. Moreover, for the gene circuit in this study, it is more likely to oscillate when PLacC , the promoter responding to a repressor, is a strong promoter, or LacIq, the repressor, degrades rapidly. Furthermore, the differences between the gene circuits involving promoter transcription rate, protein degradation rates, etc . have influence not only on the possibility to oscillate but also on the oscillating period. According to the results, the period in bacteria containing Os-LAA with glucose was much shorter than that of bacteria containing Os-ASV with glycerol.
It has also been noticed that the fluorescence behaviors of different bacteria in the same culture were not fully synchronous, which was especially notable when glucose was used as the carbon source. We supposed that “noises” such as the discrepancy of the microenvironment around the bacteria and the diversity in protein expression were responsible for the phenomenon.
In this study, promoters responding to only one regulator were used for oscillating gene circuit construction, and several parameters influencing the oscillation and its period were analyzed. In further studies, a mathematical model simulating the gene circuit designed should be built to systemically analyze the influence of each element on the final oscillation of the gene circuit.
This research was financially supported by the Fundamental Research Funds for the Central Universities of China and the Open Funding Project of the State Key Laboratory of Bioreactor Engineering.
Andersen JB , Sternberg C , Poulsen LK , Bjorn SP , Givskov M , Molin S 1998 New unstable variants of green fluorescent protein for studies of transient gene expression in bacteria. Appl. Environ. Microbiol. 64 2240 - 2246
Andrianantoandro E , Basu S , Karig DK , Weiss R 2006 Synthetic biology: new engineering rules for an emerging discipline. Mol. Syst. Biol. 2 1 - 14    DOI : 10.1038/msb4100073
Becskei A , Serrano L 2000 Engineering stability in gene networks by autoregulation. Nature 405 590 - 593    DOI : 10.1038/35014651
Chen L , Wang R , Li C , Aihara K 2010 Modeling Biomolecular Networks in Cells: Structure and Dynamics 1st Ed. Springer 217 - 232
Danino T , Mondragón-Palomino O , Tsimring L , Hasty J. 2010 A synchronized quorum of genetic clocks. Nature 463 326 - 330    DOI : 10.1038/nature08753
Egland KA , Greenberg EP 1999 Quorum sensing inVibrio fischeri: elements of the luxI promoter. Mol. Microbiol. 31 1197 - 1204    DOI : 10.1046/j.1365-2958.1999.01261.x
Elowitz MB , Leibler S 2000 A synthetic oscillatory network of transcriptional regulators. Nature 403 335 - 338    DOI : 10.1038/35002125
Friedland AE , Lu TK , Wang X , Shi D , Church G , Collins JJ 2009 Synthetic gene networks that count. Science 324 1199 - 1202    DOI : 10.1126/science.1172005
Fung E , Wong WW , Suen JK , Bulter T , Lee SG , Liao JC 2005 A synthetic gene-metabolic oscillator. Nature 435 112 - 118    DOI : 10.1038/nature03508
Fuqua WC , Winans SC , Greenberg EP 1994 Quorum sensing in bacteria: the LuxR-LuxI family of cell densityresponsive transcriptional regulators. J. Bacteriol. 176 269 - 275
Garcia-Ojalvo J , Elowitz MB , Strogatz SH 2004 Modeling a synthetic multicellular clock: repressilators coupled by quorum sensing. Proc. Natl. Acad. Sci. USA 101 10955 - 10960    DOI : 10.1073/pnas.0307095101
Gardner TS , Cantor CR , Collins JJ 2000 Construction of a genetic toggle switch in Escherichia coli. Nature 403 339 - 342    DOI : 10.1038/35002131
Goodwin BC 1965 Oscillatory behavior in enzymatic control processes. Adv. Enzyme Regul. 3 425 - 428    DOI : 10.1016/0065-2571(65)90067-1
Hu FX , Ding R , Cui ZH , Yu JL , Liu LF , Ao YH , Zhang LJ 2013 Approaches and strategies of gene scarless knockout in theEscherichia coligenome. Lett. Biotechnol. 4 552 - 557
Jobe A , Sadler JE , Bourgeois S. 1974 lacRepressor-operator interaction: IX. The binding oflacrepressor to operators containing Oc mutations. J. Mol. Biol. 85 231 - 248    DOI : 10.1016/0022-2836(74)90362-3
Kaznessis YN 2007 Models for synthetic biology. BMC Syst. Biol. 1 1 - 47    DOI : 10.1186/1752-0509-1-47
Keiler KC , Waller PRH , Sauer RT 1996 Role of a peptide tagging system in degradation of proteins synthesized from damaged messenger RNA. Science 271 990 - 993    DOI : 10.1126/science.271.5251.990
Lutz R , Bujard H 1997 Independent and t ight regulation of transcriptional units inEscherichia coli viathe LacR/O, the TetR/O and AraC/I1-I2 regulatory elements. Nucleic Acids Res. 25 1203 - 1220    DOI : 10.1093/nar/25.6.1203
O’Briena EL , Italliea EV , Bennett MR 2012 Modeling synthetic gene oscillators. Math. Biosci. 236 1 - 15    DOI : 10.1016/j.mbs.2012.01.001
Sayuta DJ , Sun L 2010 Slow activator degradation reduces the robustness of a coupled feedback loop oscillator. Mol. Biosyst. 6 1469 - 1474    DOI : 10.1039/c003480k
Song K , Huang XT 2012 Introduction to Synthetic Biology 2nd Ed. Science Press Beijing, China. 65 - 83
Stricker J , Cookson S , Bennett MR , Mather WH , Tsimring LS , Hasty J 2008 A fast, robust and tunable synthetic gene oscillator. Nature 456 516 - 519    DOI : 10.1038/nature07389