News & Views | Published:

Synthetic biology

The yin and yang of nature

Nature volume 457, pages 271272 (15 January 2009) | Download Citation

Oscillations in gene expression regulate various cellular processes and so must be robust and tunable. Interactions between both negative and positive feedback loops seem to ensure these features.

Periodic oscillations are the basis of timekeeping. For many millennia, the main timekeeper was the water clock, in which time is recorded by the regular dripping of water into or out of a basin. In the seventeenth century, the water clock was replaced by the pendulum clock, following Galileo's famous discovery that the period of a pendulum's swing is independent of the size of the swing. This clock offered a substantial improvement because pendulum oscillations are robust and the period can be altered by changing the length of the pendulum arm. As three papers1,2,3 now indicate, similar improvements are occurring in our ability to generate increasingly robust and tunable oscillations in biological systems. On page 309 of this issue, Tigges et al.1 demonstrate this feat in mammalian cells, and not long ago Stricker et al.2 reported it in bacteria. These two advances are nicely complemented by Tsai and colleagues' recent detailed theoretical study3, which elucidates the essential 'design principles' underlying oscillatory networks in nature.

The simplest way to generate oscillations is by negative feedback with a delay. We are probably all familiar with this phenomenon from our attempts to maintain the proper water temperature in the shower. Because of the delay inherent in the system, we often overshoot, leading to a sometimes comical oscillation between scalding and freezing temperatures. An early example of a synthetic biological oscillator was a network called the repressilator4, in which three genes sequentially repressed one another. The three repressive interactions led to net negative feedback, with a delay due to the multiple biochemical processes involved in gene expression. The repressilator did indeed oscillate, but the oscillations were not robust. Only half of the cells had observable oscillations, and those oscillations that did occur were variable. Moreover, the repressilator is not tunable; changing the rate constants of the various reactions generally abolishes the oscillation rather than changing its frequency3.

Nonetheless, both robustness and tunability are important features of an oscillatory system, whether it be a gene circuit or a clock. For example, despite slight variations in the individual components of a clock, the oscillation period must remain a precise number of seconds. Similarly, if the physical environment changes, it may be necessary to retune the system to compensate: moving a pendulum clock from the ground floor to the top floor of a tall building requires retuning the clock, to correct for the small change in the acceleration due to gravity.

In the recent set of papers1,2,3, a common theme is that supplementing the core negative feedback circuit with a positive feedback loop can make the oscillations both robust and tunable. Stricker et al.2 demonstrated this experimentally by implementing a simple transcriptional circuit in the bacterium Escherichia coli. Their oscillator was composed of two genes driven by the same hybrid promoter sequence: the gene encoding the LacI protein generated the core negative feedback loop to suppress transcription, whereas that encoding the AraC protein generated the positive feedback loop. The resulting oscillation period could be tuned between 13 and 100 minutes depending on the concentration of the molecules used to induce transcription and on the temperature of the system. This oscillator design was based on a previously published theoretical study5, although the authors2 found it necessary to explicitly model intermediate steps such as multimerization, translation and protein folding.

Tigges et al.1 construct a tunable oscillator in mammalian cells — using sense/antisense logic — by supplementing a central time-delayed negative feedback loop with a positive feedback loop. Here, negative feedback was provided by post-transcriptional repression of the gene encoding tTA by antisense RNA, whereas positive feedback was present because tTA enhanced its own transcription. The authors could tune the oscillation period by varying the number of copies of the genes encoding components of the oscillator. Whether varying the concentration of transcription inducers, while keeping the copy number of genes constant, would allow tuning of the oscillation period remains an intriguing question.

In their theoretical study, Tsai et al.3 computationally analysed many different network topologies that might lead to oscillations, and also concluded that a positive feedback loop is likely to be necessary for tunable oscillations. The authors pointed out that bioengineers are simply learning tricks discovered by evolution long ago. Many biological networks that drive oscillations of variable period (such as the cell cycle) have a positive feedback loop, in addition to the central negative feedback loop that is mainly responsible for generating the oscillation. Indeed, Tsai and colleagues found that such a positive feedback loop is present in many oscillatory networks that do not require tunability (such as the circadian rhythm that tracks the day–night cycle). A possible function of the extra feedback loop in these networks of fixed frequency could be to make the oscillations robust — that is, more resistant to changes in kinetic parameters — thus perhaps increasing the 'evolvability' of the oscillation.

Advances in generating biological oscillations are similar to those made in the seventeenth century that led to our widespread adoption of the pendulum clock. What advances lie in store for our ability to construct synthetic biological oscillations? In the latest experiments1,2,3, the phase of the oscillations was partly passed on to daughter cells, although individual cells gradually lost their synchronization. For any coordinated action, it is desirable for the population to oscillate in phase, thus requiring some mode of cell-to-cell communication to synchronize oscillations. More generally, a way to entrain or pause the oscillations is often necessary. For example, circadian oscillations must be entrained by daylight, and the cell-cycle oscillation must be paused under low-nutrient conditions. Further advances combining theoretical modelling with experimental synthetic biology will both increase our understanding of natural networks and allow us to use the cell as a platform for future developments in biological engineering.


  1. 1.

    , , & Nature 457, 309–312 (2009).

  2. 2.

    et al. Nature 456, 516–519 (2008).

  3. 3.

    et al. Science 321, 126–129 (2008).

  4. 4.

    & Nature 403, 335–338 (2000).

  5. 5.

    , , & Phys. Rev. Lett. 88, 148101 (2002).

Download references

Author information


  1. Jeff Gore and Alexander van Oudenaarden are in the Department of Physics, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, USA.

    • Jeff Gore
    •  & Alexander van Oudenaarden


  1. Search for Jeff Gore in:

  2. Search for Alexander van Oudenaarden in:

About this article

Publication history



Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Newsletter Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing