A signal filter that simplifies complex vibration is vital in information processing [1]. This summer, we aim to build one and explore its application in biocomputing and oral health care. Light inputs are achieved by our light-sensing protein and then further processed by a negative feedback circuit [2]. The criticality detector generates pulse output when the input’s vibration exceeds predefined threshold. For biocomputing, the detector is equipped with a binary system to store information in a “living register”. For oral health care, the detector is used to control the dosage of a powerful drug ^[3] in the dosage control system .

Details

Criticality detection system, which we intend to construct, combines biosensor with signal converter. The work pattern is that the system receives the specific input signal and then generates pulse output when the input’s vibration exceeds predefined threshold. Fig.1 The schematic diagram indicating the relationship between input and output.

The detection of threshold is achieved with the help of a red-light-sensing part and a specially designed negative feedback circuit. A fusion protein called Cph8 senses red light and activates three OmpC promoters when red light intensity in the environment drops below a threshold. An output (GFP) keeps being generated until it is inhibited by the negative feedback circuit we built (see Fig.2). A relatively long feedback circuit was designed to generate pulse of good intensity and modularity. Considering the system construction and work pattern, our criticality detection system can function as a pulse generator, a signal filter and so on. In our case, the module is proved to have potential application in biocomputer development and health care. Our system can be used as a signal filter in the biocomputer development. Combin-ing criticality detection system with pairs of integrase and excisionase, we set up a binary counting system composed of some two-state latch modules that switch between states "0" and "1". Working as a pulse generator in dental caries prevention, the system release a suita-ble amount of antimicrobial peptides by adjusting the length and intensity of the pulse and maintain the population of oral flora at a satisfactory level. Fig.2 The schematic diagram showing how the criticality detector works.

The design of our criticality detector came occasionally from a paper about edge detection (A Synthetic Genetic Edge Detection Program.Jeffrey J. Tabor etc. j.cell.2009.04.048 ). Their gene circuit detects edge by recognizing the diffusion near the boundary of an object. Essentially speaking, they use a limitation of diffusion and the spatial distribution it causes to mark the edge. Inspired by their work, we design the negative feedback circuit to create a delay to mark the threshold. Therefore, the criticality detector could be divided into two parts: a series of protein for light-sensing and a negative feedback for pulse generating. The light-sensing protein we use is Cph8, a chimeric protein which produces strong response to light. It consists of a light-sensing domain Cph1 and an EnvZ domain. Phococyanobilin is necessary for Cph1 to response to light, but it’s not naturally produced in E. coli. Another two genes, ho1 and pcyA, can produce two enzymes that convert the endogenous haem into phococyanobilin. Ho1 is one member of the heme oxygenases family. It functions in producing BV IXa from endogenous heme in E. coli. Ho1 catalyses stereospecific cleavage of heme and releases Fe2+ and carbon monoxide, which is the first step of phycocyanobilin synthesis. The second step is conducted by a phycocyanobilin: ferredoxin oxidoreductase (pcyA) which functions in reducing BV IXa. Functioning as a kinase, the EnvZ domain could lead to autophosphorylation of an endogenous regulator OmpR when light intensity remains below threshold. Then the phosphorylated OmpR activates the ompC promoter. But when light intensity is high, autophosphorylation is inhibited and therefore, the expression stops. Fig.1

A negative feedback circuit is constructed to shut down output while maintain it for a while, which generates a suitable pulse output. Three OmpC promoters are set to receive input. The first one controls the transcription of a CI protein with the RBS locked in a cr loop of an artificial riboregulator system. The second one controls the transcription of a taRNA which unlocks the cr loop and starts the expression of CI. The last one expresses GFP, before which is a CI binding site. When CI binds to it, GFP output would be shut down and hence a pulse is generated. Fig.2 Negative feedback circuit. First two OmpC promoters receive input signals and generate taRNA and output signal. The taRNA will change the structure of an artificial riboregulator and expose the RBS to make CI protein express. The CI protein will inhibit the expression of the output. The different work patterns of the negative circuit in states of reaching the threshold and after reaching the threshold for a while are shown in Fig.5 and Fig.6 below. Fig.3 The work pattern of the negative circuit when reaching the threshold. The OmpC promoters receive the signal. The output and taRNA are generated. Because taRNA used to unlock riboregulator and CI protein haven’t accumulated to the predefined level, the output is maintained for a while.

Fig.4 The work pattern of the negative circuit after reaching the threshold for a while. Remaining below the threshold for a while, taRNAs accumulate and unlock crloops to express CI proteins. When CI protein binds to its binding site, the output is blocked.

Inspired by the bidirectional DNA flipping system illustrated by Jerome Bonnet, etc in 2012[1], we are planning a binary counting system, using the mechanism of serine recombinases. The counting system is expected to record the number of certain stimulation by the direction states of the registers, which are under the control of serine recombinases. The stimulation should be normalized by the criticality detection system to ensure that the stimulation is properly dealt with.

Basic mechanisms

Recombinases used in this system are large serine recombinases that mediate integrative and excisive site-specific recombination of temperate phage genomes [2]. For a species of phage, the recombinases consist of a strictly directional integrase (Int), and a recombination directionality factor (RDF), or called excisionase (Xis), which can change the directionality of Int. The Int can specifically flip the DNA between specific sites (attB and attP), and the sites turn into attL and attR. If a promoter is placed between the pair of att sites, the register is formed, which registers by the directionality of the promoter. As flipping in either direction needs different sets of enzymes, the register may take control in the flipping procedure, by which we can make flip in both direction using the same induction (some standardization work certainly needed). Pairs of recombinases have been found in some phages, and the enzymes as well as att sites are different between each other. Based on this, if we stack up a number of recombinase systems like a binary system, in which the two directionality of each register represents 0 or 1 on one digit, and the overflow to the higher digit is performed by specific RNA polymerases which are highly specific for the exact corresponding promoter. T7 RNA polymerase, for example, can specifically start the transduction after T7 Promoter.

How does it count

We chose two pairs of recombinases, from phage Bxb1 and φC31, and intend to build a counting system with two digits. The two digits are linked by T7 RNA polymerase. As in the figure, the initial status is 00. The first stimulation can induce the expression of Bxb1 Int and Xis, which flips the register from L/R state to B/P state, just like the binary digit is switched from 0 to 1. The next stimulation will flip the register back, and express T7 RNA Polymerase, which induces the flip of the next register: attL/R of φC31. This step is like the binary 01 is added to 10. And the same mechanism goes on. If the registers are more, there should be much larger number that the system can count.

Future work

We have to work out the intensity and duration under which each flipping system works best. Hopefully the conditions for both directions of one flipping system are similar, or some adjustments should be taken, to standardize the input and to make good use of the criticality system, which helps standardize the input. Meanwhile, the RNA polymerases are usually rather stable, so the degradation of the RNAP should be accelerated.

[1] Bonnet J, Subsoontorn P, Endy D. Rewritable digital data storage in live cells via engineered control of recombination directionality. Proceedings of the National Academy of Sciences of the United States of America. 2012;109(23):8884-8889. doi:10.1073/pnas.1202344109. [2] Singh S, Rockenbach K, Dedrick RM, VanDemark A, Hatfull GF. Cross-talk between diverse serine integrases. Journal of molecular biology. 2014;426(2):318-331. doi:10.1016/j.jmb.2013.10.013.

Idea to put the criticality detector into practice

Our criticality detector can be use to control doses of some strong medicine precisely by the impulse signals produced by it. To prove this property creatively, we construct a antimicrobial peptide producing circuit relating to dental caries prevention and homeostasis maintenance. To be specific, when the concentration of Streptococcus mutans increase to a certain level, the criticality detector can activate the expression of targeted antimicrobial peptide to inhibit the growth of S.mutans, controlling its population in oral cavity under a healthy level.

Relationship with the criticality detector

There is a large microbial population living in our oral environment playing a major role in our fine body condition undoubtedly. However a big part of that population, the Streptococcus mutans ,would be the foundation stone of the biofilm causing tooth decay when its population size loses control. S.mutans is a kind of gram positive pathogen living in our oral cavity. It can colonize on the surface of tooth and produce acidic metabolites which can destroy tooth enamel. It can also create a special low pH micro-environment to help other oral pathogens to colonize on tooth. The growth of these pathogens as well as the metabolites they produce can gradually cause the formation of dental caries. Therefore the population of the S.mutans needs to stay in a heathy range,which means a strict control of the dose of the medicine targeting on the S.mutans. And our criticality detector can act as a dose controller perfectly by detect the pH in the micro-environment around the S.mutans population, for the pH implies the size of its population.

How to control the population of S.mutans

Antimicrobial peptides are recently regarded as a promising choice to kill microbes due to their strong microbicidal effects and their unlikeliness of triggering resistance in microorganisms. Since there are also various kinds of probiotics living in our mouth besides oral pathogens, and since antimicrobial peptides have effect on general kinds of microorganisms, it’s more wise to modify the antimicrobial peptide into a targeted one by adding a S.mutans targeted sequence on the N-terminal, so that only S.mutans can be killed. Thus we design two fused peptides, CAP-glyglygly-Bac8c and CAP-glyglygly-APP2, to targetedly inhibit the growth of S.mutans in oral cavity, where CAP is an optimized peptide that can specifically bind to the comD receptor on the surface of S.mutans, while Bac8c and APP2 are two kinds of peptides that assume strong bactericidal effect in acidic environment.

How to construct the antimicrobial peptide expression circuit

We first fuse CAP and Bac8c(or APP2) with linker “-glyglygly-” to construct targeted antimicrobial peptides. Then in order to prove their function, we express them in Pet28a with the induction of IPTG, and observe their bactericidal effect on E.coli and S.mutans to see if they have targeted effect. And finally we add criticality detector circuit on the upstream of antimicrobial peptide expression gene to control its expression, and use acid sensitive promoter AsR as the input signal sensor of criticality detector to complete the whole system.

Fig.1 Two chimeric protein we designed.

How does the whole system work

When the concentration of S.mutans in mouth increases, the pH of the micro-environment decreases as well, and reaching to certain level it will activate promoter AsR and induce the reaction downstream in different intensity. The signal intensity will be processed by our criticality detector and an final output signal containing the information whether the antimicrobial peptide gene should or should not be expressed will transfer to the peptide expression circuit to control the expression of antimicrobial peptide. Only when the concentration of S.mutans reaches certain level, can the peptide expression circuit be activated, so that the targeted antimicrobial peptide can inhibit the growth of S.mutans, reducing its population to a lower and healthy level.
Fig.2 The work pattern of the whole dosage control system.