Biosensing Vibrio cholerae with Genetically Engineered Escherichia

Aug 16, 2016 - Abstract | Full Text HTML | PDF w/ Links | Hi-Res PDF · A Cell–Cell Communication-Based Screening System for Novel Microbes with Targ...
3 downloads 10 Views 1MB Size
Subscriber access provided by Northern Illinois University

Article

Biosensing Vibrio cholerae with genetically engineered Escherichia coli. Maciej Bartosz Holowko, Huijuan Wang, PremKumar Jayaraman, and Chueh Loo Poh ACS Synth. Biol., Just Accepted Manuscript • DOI: 10.1021/acssynbio.6b00079 • Publication Date (Web): 16 Aug 2016 Downloaded from http://pubs.acs.org on August 18, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Synthetic Biology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 12

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Synthetic Biology

Biosensing Vibrio cholerae with genetically engineered Escherichia coli Maciej B. Holowko1, Huijuan Wang1, Premkumar Jayaraman1 and Chueh Loo Poh1* 1

School of Chemical and Biomedical Engineering, Nanyang Technological University, Singapore

KEYWORDS: Vibrio cholerae, Biosensor, CRISPRi, Cholera, Synthetic Biology

ABSTRACT: Cholera is a potentially mortal, infectious disease caused by Vibrio cholerae bacterium. Current treatment methods of cholera still have limitations. Beneficial microbes that could sense and kill the V. cholerae could offer potential alternative to preventing and treating cholera. However, such V. cholerae targeting microbe is still not available. This microbe requires a sensing system to be able to detect the presence of V. cholera bacterium. To this end, we designed and created a synthetic genetic sensing system using nonpathogenic Escherichia coli as the host. To achieve the system, we have moved proteins used by V. cholerae for quorum sensing into E. coli. These sensor proteins have been further layered with a genetic inverter based on CRISPRi technology. Our design process was aided by computer models simulating in vivo behavior of the system. Our sensor shows high sensitivity to presence of V. cholerae supernatant with tight control of expression of output GFP protein.

ACS Paragon Plus Environment

1

ACS Synthetic Biology

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Cholera disease, although virtually eradicated in the developed world, is still taking its toll in the third world countries. According to the 2015 report on cholera by WHO, there were over 190 thousand reported cases of cholera with mortality of over 1% in 2014 1. The causative agent of the disease is Vibrio cholerae, a marine bacterium 2 . The cornerstone therapy of choice – oral rehydration therapy relies on materials that are hard to obtain in the affected areas and at the same time it does not prevent other people from getting infected. Direct preventative methods (e.g., vaccines) unfortunately have low efficacy 2. Current detection methods of V. cholerae presence are either based on classic culture methods which are slow or based on modern molecular methods (e.g., PCR and antibodies) which are expensive and require trained staff 3. It was also proposed that the problem of cholera could be addressed by means of synthetic biology - Focareta et al. have created a genetically modified Escherichia coli which was able to extract cholera toxin from its immediate surroundings and March et. al proposed an engineered E. coli which prevents V. cholerae from producing toxin

Figure 1 Design of our V. cholerae sensor. A) In this diagram double plasmid sensor circuit in low V. cholerae cell density state is shown. CqsS, LuxU and LuxO are constitutively expressed and phosphorylated. Phosphorylated LuxO activates Qrr4 promoter which expresses gRNA. This gRNA joins constitutively expressed dCas9 and they together repress GFP expression from a constitutive promoter. B) In this diagram the same circuit is shown but in high V. cholerae cell density state. CAI-1 dephosphorylates CqsS which ultimately dephosphorylates LuxO. Dephosphorylated LuxO does not activate the Qrr4 promoter preventing gRNA expression. With no gRNA present GFP can be expressed efficiently. A) and B) In blue are parts of the sensor module, in red are parts of the inverter module and finally in green is the receiver module. Text in circles denotes respective plasmid origin.

Page 2 of 12

through quorum quenching 4,5. However, these methods do not kill the V. cholerae. Recent studies demonstrated that it is possible to develop beneficial microbes to sense and kill infection causing Pseudomonas aeruginosa 6,7. However, such a microbe that would effectively sense and kill V. cholerae is still not available. The microbe requires a sensing system which is able to sense the presence of V. cholerae. To this end, in this paper we engineered a synthetic sensing system in E. coli based on CAI-1 quorum sensing of V. cholerae. Vibrio cholerae controls its infection cycle using a sophisticated, redundant and parallelized quorum sensing system 8. There are two well established pathways, both using a two component sensor proteins, namely CqsS and LuxPQ which sense CAI-1 and AI-2 molecules respectively 9–11 . AI-2 is a molecule that is used for interspecies communication whereas CAI-1 is used for intraspecies communication. Different variants of CAI-1 are used to communicate between cells of the different specific Vibrio genus and are generally very selective to the genus that produces given variant 12–14. Recently, two additional systems of quorum sensing in V. cholerae have been described – VarS/VarA two component sensor based system and Fis protein based system 15–17. Although the mechanisms of these two systems are not fully understood, they have been shown to detect changes in density of V. cholerae cells. All these systems are proved to be working in parallel, meaning that even when only one of them is left working in the V. cholerae cell, the cell is still able to react to changes of its density in the medium. The CAI-1 based system uses three proteins – CqsS, LuxU and LuxO. This quorum sensing pathway starts with CqsS, a two component transmembrane sensor protein with kinase/phosphatase activity, while LuxU and LuxO channel the quorum regulated signal further along the signal pathway. CqsS can change phosphorylation level of LuxU, a phosphorelay protein, which in turn phosphorylates or dephosphorylates LuxO protein which is a transcription factor. In low cell density state, when CAI-1 is present in low concentration CqsS acts as kinase and phosphorylates LuxU. LuxU then phosphorylates LuxO which becomes an activator of expression from four associated promoters. These promoters then express quorum regulated RNAs (Qrr) which relay the quorum dependent signal further down the pathway 18–20. In high cell density state, concentration of CAI-1 in the medium is high and CqsS acts as a phosphatase and dephosphorylates LuxU 21. Dephosphorylated LuxU removes phosphate group from the regulatory domain of LuxO, making it unable to activate the Qrr promoters.

ACS Paragon Plus Environment

2

Page 3 of 12

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Synthetic Biology

To achieve V. cholerae sensing in E. coli, we designed and constructed a genetic circuit comprising CqsS, LuxU and LuxO proteins as the sensor, a genetic inverter and green fluorescence protein under Qrr4 promoter as reporter. pQrr4 is one of the Qrr promoters from V. cholerae. The system is depicted in Figure 1. The need for an inverter comes from the fact that the pQrr4 is downregulated by the sensor when CAI-1 is detected. Hence, we need to invert this signal for our circuit to be an effective biosensor. To this purpose, we decided to use CRISPR based genetic inverter. CRISPR is the new emerging technology which enables efficient and easy modification and binding of specifically targeted DNA 22,23. By creating two point mutations in the active centers of the Cas9 protein (the centerpiece of the CRISPR technology), it is rendered unable to cut target DNA, but at the same time, ability to bind the DNA is retained. This type of Cas9 is called dead Cas9 (dCas9) and the corresponding technology is called CRISPRi (CRISPR interference). It has been used for efficient repression of promoters and other genetic elements 24–26 . The biggest advantage of CRISPRi technology is the ability of repressing virtually any promoter to the point of near non-expression levels 26. CRISPRi has already found use in various genetic circuits 25,27,28. The three CAI-1 sensor proteins (CqsS, LuxU and LuxO) retained their ability to control gene expression by pQrr4, one of the Qrr promoters from V. cholerae which we have moved into E. coli. The pQrr4 promoter was able to turn on and off our CRISPRi based inverter which in turn controlled GFP expression. The CRISPRi system can also act as an amplifier because the output promoter of the inverter which drives the GFP expression can be much stronger than the input promoter to the inverter. As a result, we created an E. coli strain that normally represses GFP production, derepressing it only upon V. cholerae detection. Supernatant of V. cholerae has been reported to contain CAI-1 in concentrations high enough to induce density dependent change 11. We show that our final designed E. coli system is able to sense the presence of V. cholerae based on its supernatant, achieving up to a 2.5fold difference in reporter fluorescence intensity when cultures grown in E. coli and V. cholerae supernatant are compared. RESULTS In our work, we opted for systematic, rational design of our circuit. The system is divided into three modules – sensor, inverter and reporter (Figure 1, color coded). We have built and characterized a set of testing circuits for each of the modules. As the system is complex and involves a number of variables, we have developed a kinetic model of the system to guide us with the experiments and to optimize the final system. pQrr4 can be made selective towards phosphorylated LuxO by truncation To study the sensor’s (CqsS, LuxU and LuxO) performance in E. coli, we needed a suitable reporter protein coupled with LuxO dependent promoter. Being a wellestablished and easy to use reporter protein, green fluorescence protein GFPmut3b was chosen 29. LuxO, a tran-

Figure 2 Characterization of full and truncated Qrr4 promoter. A) Genetic circuit used in the study. LuxO or LuxOt is produced with arabinose present and activates full or truncated pQrr4 to produce GFP. B) Sequence of the Qrr4 promoter. LuxO binding sites are highlighted in red. Sigma factor binding sites (24 and -12) are underlined. Part of the sequence that constitutes truncated Qrr4 promoter is bolded and showed in blue. C) Comparison of GFP/OD600 readings after 4 hours of microplate cultures for all constructs. Truncated pQrr4 is much weaker, but is not activated by dephosphorylated LuxO. D) and E) Transition curves showing change in GFP/OD600 level in both full (D) and truncated (E) pQrr4 in different arabinose concentrations. The levels shown are after 4 hours of culture. Data is fitted with Hill type of curve, r square for all fits > 0.95.

scription factor protein, activates four different promoters in V. cholerae, one for each Qrr molecule. This activation is done by phosphorylated LuxO binding to σ54 factor of the RNA polymerase. Interestingly, the phosphorylated (active) state of LuxO can be mimicked by removing its regulatory domain18. In our study, this truncated LuxO (LuxOt) was used as activated state of LuxO. Among all the Qrr promoters, pQrr4 (sequence as shown in Figure 2B) which is responsible for expressing sRNA Qrr4 is the best studied one and it shows quite high expression relative to other Qrr promoters 30. To elucidate the performance of pQrr4 in E. coli, we characterized pQrr4 by expressing LuxO (inactivated state) or LuxOt (activated state) under pBAD arabinose promoter from plasmid which also harbored GFP under control of pQrr4 (Figure 2A). As our testing gene circuit does not include CqsS and LuxU, LuxO would stay unphosphorylated in the cell. pQrr4 is expected to only turn on when phosphorylated LuxO is present and remain off otherwise. However, our results show that the pQrr4 is not selective towards phosphorylated LuxO as the promoter is activated when LuxO is expressed (Figure 2C and D). Further, when LuxOt is expressed in the same setup the GFP expression is 2/3 times higher than in the case of LuxO (Figure 2C, second and fourth bar from the left). This implies that although pQrr4 is activated stronger by phosphorylated LuxO, it retains some of the activity with unphosphorylated LuxO present alone. To overcome this problem, we decided to shorten the pQrr4 so that it only

ACS Paragon Plus Environment

3

ACS Synthetic Biology

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

consists of the required minimum – binding sites for σ54 and phosphorylated LuxO. It was previously shown that for LuxO type activators the promoter is still activated in presence of the activator after removing the activator binding sites 31. We tested the truncated pQrr4 (tpQrr4 Figure 2B) with both LuxO (inactivated state) and LuxOt (activated state) using the aforementioned circuit (Figure 2C and E). Results show that although the tpQrr4 is much weaker compared to the full original pQrr4, it is much more selective towards LuxOt which mimics phosphorylated LuxO. Consequently, tpQrr4 was chosen as our promoter for further study. Design and Characterization of the CRISPRi based inverter system As previously mentioned, signal coming from the CAI-1 sensor is negative - the pQrr4 is down-regulated by the sensor upon CAI-1 detection. That means that by default the GFP under pQrr4 promoter will be expressed in low cell density state and its expression will stop upon high CAI-1 concentration detection. As it is generally more reliable to observe rise in GFP fluorescence rather than its reduction, it is required for this signal to be inverted. Moreover, since tpQrr4 has been shown to be a fairly weak promoter we needed to amplify the GFP expression as well. Consequently, we need an inverter that could invert and amplify the signal from the CqsS-LuxU-LuxO sensing cascade. There are many possibilities when it comes to choosing a genetic inverter, which in its most

Page 4 of 12

basic form comprises a repressor – promoter pair (eg. Lambda repressor, TetR like repressors, and CRISPRi) 32. Because CRISPRi gives us the flexibility to design gRNA to repress different promoters and has very high repressing efficiency, we decided to use a CRISPRi based inverter for our system. Furthermore, CRISPRi also gives us the ability to readily generate a library of inverters using promoters of very different strengths. This would enable us to match the required sensor output with a promoter with a suitable strength. To create the desired inverter, we first have to establish which of the two components of the CRISPRi system (i.e., dCas9 and gRNA) should be placed under the control of tpQrr4 promoter. It could be either the dCas9 or the gRNA, or even both. The other component would be expressed constitutively. The criterion for this choice was – by manipulating the cellular levels of which of the two components, a high level of repression of the target promoter at low induction gRNA/dCas9 level can be achieved. To elucidate the choice, we designed and constructed eight inverters using a set of four promoters from the widely used Anderson promoter library 33 (Figure 3B). The promoters we have chosen have two main qualities – they do not differ much between each other in terms of sequence but they have different promoter strength. For each of the promoters we designed two distinct gRNAs one targeting -10 region of the promoter and binding to the non-template strand and one targeting -35 region of

Figure 3 Characterization of CRISPRi system. A) Genetic circuit used in the study. IPTG induces gRNA production and ATc induces dCas9 production. When both are expressed they repress GFP expression from the target constitutive promoter. B) Table showing designations of promoters used in the study and their relative strengths. C) Heatmap of relative GFP expression levels for eight different inverters. The repression levels change considerably with increasing gRNA (IPTG) level, whereas they are insensitive to changes of dCas9 (ATc) level in most cases. First number in the plot title shows which promoter site is targeted (either -10 or -35) and second number shows the last three numbers from designation of the tested promoter (compare with B)). The repression levels shown are after 4 hours of culture. D) and E) plots showing repression curves for -10 (D) and -35 (E) targeting inverters with increasing gRNA (IPTG) level. dCas9 has been expressed with 0.5 nM ACS Paragon Environment ATc. Descriptions of curves as in C). F) Repression curve forPlus the -10 site targeting gRNA of JS23115 promoter in different gRNA (IPTG) concentrations. Data has been fitted with Hill type equation, r square > 0.99. dCas9 is been constitutively expressed under GabDP2 promoter.

4

Page 5 of 12

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Synthetic Biology

Figure 4 Simulation of sensor module aims to advise the optimal circuit design A) Simulated gRNA production to fulfil the needs of the inverter. In CRISPRi-based Js23115 inverter (Figure 3A), gRNA is transcribed by IPTG induced placI-o1 promoter. Figure 3F shows that the input dynamic range (90% repression) of JS23115 inverter is 0 - 25 µM IPTG. Simulation results suggest this input dynamic range corresponds to a gRNA production of 0-0.5 µM per OD600. B) Effects of LuxO (total and phosphorylated) on gRNA production. Proportion (% LuxOp) of phosphorylated LuxO (LuxOp) in total LuxO protein (LuxOTOT) and gRNA production is simulated in sensor circuit (CoE1 vector in Figure 1) which comprises a strong expression of CqsS, increasing expression of LuxU and IPTG-induced expression of LuxO. LuxOTOT is induced by 0, 3, 12, 100 µM IPTG respectively (from left to right). For all these four IPTG concentrations, phosphorylated LuxU (LuxUp) controls % LuxOp and gRNA production. gRNA threshold (0.5 µM) is suggested by Figure 4A. C) Effect of LuxU (total and phosphorylated) on gRNA production. Proportion (% LuxUp) of phosphorylated LuxU (LuxUp) in total LuxU protein (LuxUTOT) and proportion (% LuxU) of unphosphorylated LuxU (LuxU) in total LuxU protein is simulated in sensor circuit (CoE1 vector in Figure 1) which comprises an increasing expression of CqsS, strong expression of LuxU and 0 µM IPTG-induced expression of LuxO. Simulation result shows that both proportions are controlled by autophosphorylated CqsS (CqsSpp) and the turn-over of LuxUTOT into LuxUp could reach 99%. D) Effect of CAI-1 on gRNA production is simulated in sensor circuits (CoE1 vector in Figure 1) which comprises weak expression of CqsS and LuxU and 0 µM IPTG-induced expression of LuxO. Simulation result shows that CAI-1 controls the proportion (% CqsSpp) of CqsSpp in total CqsS (CqsSTOT) and gRNA production. Our model also predicts the response of GFP after simulation the inverter and reporter module.

the promoter and binding to the template strand. To study the characteristics of the inverters under different expression level of dCas9 and gRNA, we have placed dCas9 and gRNA under ATc (anhydrotetracycline) and IPTG (Isopropyl β-D-1-thiogalactopyranoside) inducible promoters respectively (Figure 3A).

promoter strength is high so as to amplify the output signal (GFP) but at the same time the promoter needs to be very strongly repressed by low gRNA levels which is expressed by tpQrr4. Both of these conditions have been met by inverter with JS23115 promoter with its respective gRNAs.

After examining all eight CRISPRi constructs, we generated eight heat map type plots, one for each construct (Figure 3C). In these heat maps we can see that very similar level of expression are achieved for all dCas9 levels at a given gRNA level, implying that the inverter is not sensitive to dCas9 level variations. On the other hand, for a given dCas9 level a range of repression levels are achieved for different gRNA levels. Based on this observation, we decided that the gRNA will be produced under tpQrr4, whereas dCas9 will be placed under a weak constitutive promoter. This has the added advantage of preventing dCas9 expression level going above toxicity threshold 28. Figure 3D and 3E shows the transition curves of promoters targeted at -10 and -35 site respectively. As can be seen in these graphs, a whole range of inverters with different final repression levels and repression profiles has been created. Since tpQrr4 is a weak promoter, it is necessary for the inverter to have a gRNA - promoter pair where the

To validate our hypothesis that the inverter will behave in a similar manner if the pLlacO-1 inducible promoter driving dCas9 expression is replaced by a weak constitutive promoter, we prepared a modified version of our construct in which GabDP2 constitutive promoter drives dCas9 expression (Figure 3F). As shown in Figure 3F, it is clear that upon adding IPTG in concentration of > 25µM, the expression of GFP is sharply repressed to nearly nondetectable levels. Further, compared to wild type MG1655 E. coli there is no change of growth in the investigated timeframe. This implies that the inverter system expressing dCas9 under GabDP2 and gRNA does not introduce significant metabolic burden to the host. Establishing the optimal sensor protein expression using in silico model With CRISPRi inverter and tpQrr4 characteristics established, we moved to design and characterize our sen-

ACS Paragon Plus Environment

5

ACS Synthetic Biology

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

sor module. This module comprises CqsS, LuxU, LuxO proteins and tpQrr4 promoter. In the final system, sensor module’s tpQrr4 will drive the expression of gRNA for the CRISPRi inverter; and the inverter will regulate the expression of the downstream reporter module. Our design goal for the sensor module is that the module needs to produce adequate gRNA to repress the reporter (GFP) production in the absence of V. cholerae, to respond to a change in density of V. cholerae (e.g. from low to high density) and reduce the production of gRNA so that GFP is produced in the presence of V. cholerae. However, designing such a genetic circuit is challenging because there are too many possible assembly plans, namely, various promoters, ribosome binding sites and all their combinations to express sensor proteins (CqsS, LuxU and LuxO). Among these assembly plans, only few of them could deliver gRNA levels that fall in the dynamic range of our inverter. To address this question, we developed an in silico kinetic model of the overall sensor system to identify key parameters so as to guide our experimental design in achieving the desired design. The model was developed using ordinary differential equations (ODEs); parameterized using literature, experimental data in this study and from our unpublished work; and validated using independent experimental data which were derived from a different set of genetic circuits (See Supporting Information Figure S5 and Figure S6B). We first determine the level of gRNA produced within the input dynamic range of the CRISPRi-based JS23115 inverter. Here, input dynamic range refers to the range of input signal (gRNA) over which JS23115 inverter could respond 34. As shown in Figure 3F, the reporter GFP expression reduces from 100% to 10% when IPTG concentration increases from 0 uM to 25 uM. As pLlac-O1 promoter drives the expression of gRNA, this 10% reporter expression level corresponds to a 90% repression caused by a gRNA level induced by 25 uM IPTG. Using our model, we simulated the dynamics of gRNA induced by IPTG. The result suggests that the dynamic range of repression corresponds to 0-0.5 µM of gRNA (Figure 4A). To finetune the sensor circuit design to achieve the expected gRNA level, we performed sensitivity analysis (SA) using the model to identify the key factor affecting the expression of gRNA. The SA results show that the promoter driving the expression of LuxO is a critical factor. Guided by this result, we constructed a genetic circuit in which LuxO is transcribed by an IPTG-inducible promoter (pLlac-O1) to study and determine the optimal expression level of LuxO by varying IPTG concentration (This construct was used in the study under “Full System Integration” section). We then run a number of simulations using the model to gain insights into what would the optimal amount of LuxO, LuxU and CqsS be. We first studied LuxO. Phosphorylated LuxO (LuxOp) activates the tpQrr4 promoter which drives the gRNA expression. Using our model, we found that the percentage of LuxOp with respect to total LuxO (LuxOTOT) correlates to the amount of gRNA presence (r-square > 0.96, Figure 4B). This correlation suggests that we can tune the gRNA level by adjusting the

Page 6 of 12

ratio between LuxOp and LuxOTOT. We also note that LuxOTOT induced by different amount of IPTG transform 60-80% of LuxOTOT into LuxOp. This amount of LuxOp produces enough gRNA (Figure 4B) because the saturated gRNA level is higher than the gRNA needed for full repression as shown in Figure 4A. Based on this simulation result, we hypothesize that the sensor can work properly with different amounts of LuxOTOT. As a result, we chose to express minimum amount of LuxOTOT for subsequent simulation to minimize the eventual metabolic burden to the cell caused by LuxOTOT production. Our model also suggests that phosphorylated LuxU (LuxUp) significantly affects LuxOp level. Figure 4B shows that only a small amount of LuxUp is needed to tune the LuxOp and gRNA level. LuxUp amount is determined by LuxUTOT amount and its phosphorylation / de-phosphorylation. Our model shows that autophosphorylated CqsS (CqsSpp) enables up to 99% of LuxUTOT to be LuxUp (Figure 4C). Together with the small amount of LuxUp needed (Figure 4B), we infer that weak expression of LuxU would keep the sensor working properly. CqsS is a membrane protein and is difficult to express in the genetically engineered E. coli. Since only a small amount of CqsSpp is required to transform 99% of LuxUTOT into LuxUp (Figure 4C), a weak expression of CqsS for the sensor circuits would be sufficient. Our model shows that with a low level of CqsSTOT, the sensor detects the presence of CAI-1 and prevents the formation of CqsSpp. As a result, gRNA production would be repressed (Figure 4D) which would then deactivate the inverter and allow the reporter to produce maximum amount of GFP upon the detection of CAI-1 (Figure 4D). We herein concluded that low level of LuxU and CqsS would be sufficient to adjust LuxOp level. Expressing low levels of CqsS also gives us an added advantage of avoiding the common problem with expressing membrane proteins in E. coli. Hence, the model proposes an optimal circuit design for the sensor module, which comprises weak constitutive expression of CqsS and LuxU, together with inducible expression of LuxO to retain our ability to change LuxO levels if needed. There were two reasons to do this: (i) one is that the LuxO, according to our model, is the key protein in the whole system and its level will have very high impact on the final system performance; (ii) the second reason being that by turning LuxO expression ON and OFF we can investigate the impact of the sensor has on the inverter. Full system Integration Based on the design suggested by our model we constructed a full system circuit which comprises the sensor layered with the inverter and reporter. We have prepared a double plasmid system showed in Figure 1. The final system was studied using E. coli and V. cholerae overnight supernatant. E. coli supernatant is CAI-1 free and V. cholerae supernatant is reported to contain 1.25 µM of CAI-1 when V. cholerae is cultured to high density 11. Also, this is the concentration found to shift V. cholerae cells from low to high density state 35.

ACS Paragon Plus Environment

6

Page 7 of 12

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Synthetic Biology

Our tests show that when our full system is grown in increasing concentrations of the supernatant (5, 30, 50 and 90 %) in a range of IPTG concentrations (from 0 to 500 µM), the highest achievable difference between final GFP/OD600 of cultures grown in E. coli and V. cholerae supernatant is 2.5 fold (90% supernatant, 0 µM IPTG) (see Figure 5A). Results show that increasing concentrations of supernatant increases the fold difference between the final GFP/OD600 of the cultures with V. cholerae supernatant and the cultures with E. coli supernatant. This is most probably the effect of increasing CAI-1 concentration in the medium and hence decreasing intracellular level of phosphorylated LuxO. On the other hand, when the intracellular level of LuxO is increased with addition of IPTG, the final level of GFP/OD600 decreases for both the cultures with E. coli and V. cholerae supernatant respectively. However, despite the fact that the decrease in the former case is lesser than in the latter, resulting decrease of fold difference with increasing IPTG concentration is very small (Figure 5B, compare Figure 5D and 5E). These results are also in line with predictions from our model, when it is used to predict fold difference (Figure 5C and 5F). This suggests that by increasing LuxO level one can decrease the basal (uninduced) GFP expression, as the higher level of phosphorylated LuxO produces

more gRNA. However, at the same time the system produces less GFP upon CAI-1 detection. This may be because the higher amount of LuxO proteins could no longer be efficiently dephosphorylated by the given amount of CqsS, as suggested by our model analysis. As a result, the final gRNA level in these two cases would be higher than when IPTG was absent. Further, when IPTG is absent, the baseline response of the sensor to E. coli supernatant looks almost like an induction curve (Figure 5D). This implies that the false positive rate of this sensor with no IPTG would be very high. On the other hand, although the absolute GFP/OD600 reading is lower with IPTG, the baseline response of the sensor to E. coli supernatant looks much more stable (Figure 5E). This suggests that the sensor with IPTG added would be more reliable. To confirm the response of the system to the V. cholerae supernatant, a variant of our circuit was constructed without the inverter module and with GFP being expressed directly by tpQrr4. As expected, GFP/OD600 reading of cultures of the said construct with V. cholerae supernatant added (50% of the medium) and without IPTG added are about 15% lower than those with E. coli supernatant (data not shown). This difference is likely due to the result of lower LuxOp level in the cell upon CAI-1 de-

Figure 5 Full System Integration A) Bar graph showing the inverted sensor behaviour following increasing IPTG (LuxO) and supernatant concentrations. The fold difference in GFP/OD600 after 6 hours between cultures with V. cholerae supernatant and E. Coli supernatant added increases with increasing supernatant percentage. The same fold difference does not change substantially with increasing IPTG (LuxO) level. B) Graph showing final level of GFP/OD600 of cultures with different IPTG levels and with 90% addition of supernatants. Although the fluorescence levels differ between different IPTG levels the fold difference is maintained for all the IPTG concentrations. C) Prediction of our model for cultures as in B). Our model predicts the shape of the curves; however, it indicates higher fold difference between the two supernatants. D) and E) Graphs showing characteristics of the tested system in time. Both plots show cultures with 90% supernatant content, D) is with no IPTG added and E) is with 500 µM IPTG. As far as final fold difference in GFP/OD600 is very similar for both, the final levels of GFP/OD600 differ substantially between these two cases. F) Result of in silico simulation of cultures with 90% V. cholerae and E. coli supernatants with no IPTG added. The model accurately predicts behaviour of the system (compare with 5D) after the difference between the two cultures becomes visible – it does not model the lag phase seen in the actual experiment.

ACS Paragon Plus Environment

7

ACS Synthetic Biology

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

tection. After addition of IPTG to the medium (500 µM), the difference is lowered to around 10% with slight increase in absolute expression (effect of overall higher LuxOp level). However, this expression is much lower than in the system with inverter as the inverter doubles as an amplifier for the weak tpQrr4. Direct final GFP/OD600 reading comparison of the two constructs (inverted and non-inverted, with no IPTG and supernatant added to the culture) shows that the amplification ratio of the inverter module is around 30 (see Supporting Information). Conclusion In this work we have developed a whole cell biosensor based on V. cholerae quorum sensing mechanism and CRISPRi technology for V. cholerae detection. As Cecchini et al. have pointed out in their recent review of V. cholerae detection methods, the gold standard for this detection is still the laborious culture method which is characterized by low sensitivity 36. The more recent molecular methods (mainly V. cholerae DNA and toxin detection) which show the advantage of being more sensitive require the use of advanced equipment, limiting their use in affected areas. And most of the methods that are easy to use, such as LPS (Lipopolysaccharide) rapid diagnostic tests show either low specificity or sensitivity or even both 37. Here, our sensor shows sensitivity to the presence of V. cholerae supernatant with tight control of expression of output GFP protein. Our sensor uses part of the V. cholerae quorum sensing system (i.e. CqsS) which has been shown by different authors to be very specific to CAI-1 9,21. Further, it has been shown in an earlier work that CAI-1 is a unique molecule to the V. cholerae species 14. In addition, our sensor design would be applicable for sensing other bacterial pathogen that has a unique sensing molecule like CAI-1. This could be achieved by replacing the quorum sensing proteins in the current design with other sensing proteins from an organism which produces similarly specific molecule. This work lays the foundation for future development of a highly sensitive, specific and easy-to-use whole cell biosensor for the detection of V. cholerae. In the current form our sensor requires only two specialized elements to be used – our modified cells (potentially in lyophilized form for easier handling and longer shelf life) and basic fluorescence reading device. In the future, the GFP could be replaced with some other, easier to detect protein, or a protein that would produce a distinctive change in the medium (e.g. color or smell). Finally, the sensor could be repurposed to produce anticholera infection molecules e.g. molecules targeting the cholera toxin or the V. cholerae cells themselves. This could turn our sensor into new V. cholerae treatment option. MATERIALS AND METHODS Strains, DNA sequences and plasmids All the transformations were made using TOP10 (Invitrogen) or Beta10 (New England Biolabs) E. coli strains with cells used after maximum two passages from the commercial tube supplied by vendor. All microplate experiments were done in MG1655 (K-12) E. coli except the

Page 8 of 12

combinatorial and pQrr4 study which was done using TOP10 E. coli strain. All plasmids besides the ones for sensor combinatorial study has been prepared using Gibson assembly method 38. Plasmids for sensor combinatorial study were synthesized and cloned by GenScript (Piscataway, NJ, USA). Primers and longer oligonucleotides were supplied by Integrated DNA Technologies (Coralville, Iowa, USA). Sequences of CqsS, LuxU, LuxO, pQrr4, GFPmut3b, promoters, RBS and terminators used in the study were taken from Genbank or iGEM parts registry 33,39,40 . The backbone plasmids pBbE8k, pBbE6k, pBbA8c, pdCas9 and pgRNA-bacteria has been supplied by Addgene (Cambridge, Massachusetts, USA) 24,41. All the sequences and plasmids were prepared and analysed using Vector NTI software (Thermo Fisher, Waltham, MA, USA) or Benchling web-based sequence designer (Benchling , San Francisco, CA, USA). All the protocols for transformations, PCR and DNA manipulation used in this work followed Sambrook or manufacturer manual, optimized if needed 42. Consumables and services NEBuilder master mix for Gibson assembly has been supplied by New England Biolabs (Ipswich, Massachusetts, USA). For PCR we used either iProof PCR kit supplied by Bio-Rad (Hercules, California, USA) or Q5 PCR kit supplied by New England Biolabs (Ipswich, Massachusetts, USA). Purification kits (PCR, Gel, Miniprep) were supplied by Axygen (Tewksbury, Massachusetts, USA) and Qiagen (Hilden, Germany). DNA concentrations were checked using Nanodrop ND-1000 spectrophotometer. Commercial M9 and LB broth and agar supplied by BD (Franklin Lakes, New Jersey, USA) were used in all the studies. Supplemented M9 (M9 salts, 1 mM thiamine hydrochloride, 0.8% glycerol, 0.2% casamino acids, 0.1 M MgSO4, 0.5 M CaCl2) or LB was used as the medium for the characterization. Kanamycin (50 mg/mL) and/or chloramphenicol (25 mg/mL) were added to the culture media for antibiotic selection where appropriate. All DNA sequencing services were supplied by 1st Base Company (Singapore). Microplate readings All the microplate readings were made using Biotek Synergy HT microplate reader. The protocol for all readings, unless with separately stated modifications, was as follows – overnight liquid LB culture of E. coli with tested plasmid supplemented with suitable antibiotic was diluted 75 to 100 times in supplemented M9 or LB and this diluted culture was left in an incubator for between 1.5 and 2.5 hours. Culture after reaching OD600 of between 0.1 – 0.2 was moved to 96-well, transparent, flat-bottom microplate and distributed in triplicates at 300 µL to each well. If inducer was needed it was added to the well before the culture or directly into tube with the culture. Microplate reading was continued with vigorous shaking for 6 hours, with OD600 and GFP fluorescence reading taken every 10 minutes. Relevant control and blank measurements were taken in each microplate session. Qrr4 promoter study Four plasmids have been prepared for this study, two

ACS Paragon Plus Environment

8

Page 9 of 12

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Synthetic Biology

with pQrr4 and two with tpQrr4, with both LuxO and LuxOt. Backbone for this plasmids has been pBbA8c. GFPmut3b was used as reporter protein. The four constructs have been tested in microplate with arabinose concentrations of: 0, 0.000005, 0.00001, 0. 0001, 0.0005, 0.001, 0.002, 0.005, 0.01, 0.02, 0.05, 0.075 and 0.125%. CRISPR matrix study Two sets of plasmids have been prepared for this study based on pdCas9 (for GFP and dCas9 expression) and pBbE6k (for gRNA expression) backbone plasmids. Four different plasmids based on pdCas9 has been prepared, each with GFPmut3b with one of the following Anderson promoters - J23101, J23109, J23110 and JS23115 added for constitutive GFP expression. The gRNA cassette has been copied from pgRNA-bacteria into the pBbE6k plasmid. The inducible promoter for dCas9 has been pLtetO-1 and for gRNA pLlacO-1. Two different gRNA sequences has been prepared per promoter targeting either -10 or -35 part of the promoter sequence. All of the gRNAs have been designed using the Benchling genome engineering tool (Benchling, CA, USA). It uses model derived by Doench et al. for on-target CRISPR cutting efficiency and second model created by Hsu et al. for off-target CRISPR cutting efficiency to present the user with a two scores (from 1 to 100 points) for each of the gRNAs, one for ontarget and second for off-target cutting efficiency. In both cases max score of 100 points is desired 43,44. Additionally, we run a BLAST run for each of the gRNAs against the MG1655 E. coli genome to check if there are any potential binding sites. Binding sites were identified based on several criteria, which all had to be achieved for any given gRNA for it to be considered possible off-target binding site. These criteria were – there must not be any mismatches between gRNA and the target sequence in the first seven bp (seed region). In addition, the homologous region could not be shorter than 12 bp 26. Each CRISPRi construct (with one of the pdCas9 plasmids and of the gRNA plasmids) has been tested in 2D matrix in microplate with ATc concentrations: 0, 0.1 nM, 0.25 nM and 0.5 nM and IPTG concentrations 0, 3 µM, 6 µM, 12 µM, 25 µM, 50 µM, 100 µM and 500 µM. The constitutive promoter for dCas9 expression is GabDP2 promoter and it has substituted the original ATc inducible promoter for dCas9. Full system integration study Plasmids used for the non-inverted sensor study consisted of sensor plasmid with constitutive (under p66) CqsS (BBa_B033 RBS) and LuxU (BBa_B032 RBS) and inducibly expressed LuxO (IPTG inducible promoter - pLlacO-1), BBa_B0032 RBS or BBa_B0034 RBS) in pBbE6k backbone. It was co-transformed with receiver plasmid with tpQrr4 in modified pBbA8c with GFPmut3b under its control. Constructs with BBa_B0032 LuxO RBS has been tested with 0, 100 or 500 uM IPTG and 0%, 5%, 30%, 50% and 90% of supernatants of V. cholerae El Tor, and E. coli K-12. Constructs with BBa_B0034 has been tested with 0, 25, 50, 100, 200, 300 or 500 uM IPTG and 0, 5%, 30%, 50% or 90% of supernatants of V. cholerae El Tor, and E. coli K-12 9,10,45. Supernatants have been prepared by culturing overnight respective bacteria in LB and then

removing the bacteria by centrifugation and filtering using 0.22 µm syringe filters. These tests followed the general microplate readings protocol with a modification in the preparation phase – the overnight cultures have been diluted 200 times in media already containing the intended supernatant and IPTG levels. For cultures containing 90% of either supernatant the media has been topped up with two times concentrated LB. In silico Model To build the kinetic model of the full system comprising the sensor-inverter-reporter, we chose ordinary differential equations (ODEs) to describe constitutive / regulated gene expression and phosphorylation. ODEs of constitutive transcription, translation and component degradation are reported in our unpublished work. Phosphorylation is simulated based on ultra-sensitivity analysis 46,47. We hypothesize that CRISPR based repression follows the mechanism of Hill function. Together with gene expressions and phosphorylation mentioned above, the final model consists of 27 equations, 39 parameters and 28 components (details in supporting information). The model was developed and analyzed using Matlab version 2012a (The MathWorks, Inc., MA, USA). Parameter values are obtained from either BioNumber and literature 49, or by parameter estimation performed in this study (supporting information Table S2). Each module is validated using experimental results measured from another genetic circuit(s), or in vitro study provided by other researches 21,50. 48

After model construction, we perform sensitivity analysis 51 to investigate the impact of variability in the parameters of the model on the outputs (details in supporting information).

ASSOCIATED CONTENT Supporting Information. Sequences of the genetic parts used, linear plasmid maps, additional figure and model equations can be found in supplementary materials available at http://pubs.acs.org/.

AUTHOR INFORMATION Corresponding Author * Chueh Loo Poh – [email protected] Author Contributions CLP and MBH conceived the idea of the system. MBH has designed, constructed and tested all the circuits. HJW has constructed the model used in the study. CLP and PJ have advised HJW and MBH on their respective parts. The manuscript was written through contributions of all authors.

ACKNOWLEDGMENT We would like to thank the financial support from Ministry of Education, Singapore. This work was funded under MoE Tier 2 grant (AcRF ARC43/13). Holowko M. B. would like to thank Agency for Science, Technology and Research (A*STAR) of Singapore for providing him with funding under Singapore International Graduate Award (SINGA). We would also like to thank A/Prof Diane McDougald from SCELSE,

ACS Paragon Plus Environment

9

ACS Synthetic Biology

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Singapore for helping us with obtaining the Vibrio cholerae strain used in this study.

REFERENCES (1) . WHO | 2 October 2015, vol. 90, 40 (pp. 517–544). (2) Harris, J. B., LaRocque, R. C., Qadri, F., Ryan, E. T., and Calderwood, S. B. (2012) Cholera 379, 2466–2476. (3) Huq, A., Haley, B. J., Taviani, E., Chen, A., Hasan, N. A., and Colwell, R. R. (2012) Detection, isolation, and identification of Vibrio cholerae from the environment. Curr Protoc Microbiol Chapter 6, Unit6A.5. (4) Focareta, A., Paton, J. C., Morona, R., Cook, J., and Paton, A. W. (2006) A recombinant probiotic for treatment and prevention of cholera. Gastroenterology 130, 1688–1695. (5) Duan, F., and March, J. C. (2010) Engineered bacterial communication prevents Vibrio cholerae virulence in an infant mouse model. Proc Natl Acad Sci U S A 107, 11260–11264. (6) Gupta, S., Bram, E. E., and Weiss, R. (2013) Genetically programmable pathogen sense and destroy. ACS Synth Biol 2, 715– 723. (7) Saeidi, N., Wong, C. K., Lo, T. M., Nguyen, H. X., Ling, H., Leong, S. S., Poh, C. L., and Chang, M. W. (2011) Engineering microbes to sense and eradicate Pseudomonas aeruginosa, a human pathogen. Mol Syst Biol 7, 521. (8) Ng, W.-L., and Bassler, B. L. (2009) Bacterial quorum-sensing network architectures. Annu Rev Genet 43, 197–222. (9) Ng, W.-L., Perez, L. J., Wei, Y., Kraml, C., Semmelhack, M. F., and Bassler, B. L. (2011) Signal production and detection specificity in Vibrio CqsA/CqsS quorum-sensing systems. Mol Microbiol 79, 1407–1417. (10) Miller, M. B., Skorupski, K., Lenz, D. H., Taylor, R. K., and Bassler, B. L. (2002) Parallel quorum sensing systems converge to regulate virulence in Vibrio cholerae. Cell 110, 303–314. (11) Higgins, D. A., Pomianek, M. E., Kraml, C. M., Taylor, R. K., Semmelhack, M. F., and Bassler, B. L. (2007) The major Vibrio cholerae autoinducer and its role in virulence factor production. Nature 450, 883–886. (12) Bolitho, M. E., Perez, L. J., Koch, M. J., Ng, W. L., Bassler, B. L., and Semmelhack, M. F. (2011) Small molecule probes of the receptor binding site in the Vibrio cholerae CAI-1 quorum sensing circuit. Bioorg Med Chem 19, 6906–6918. (13) Zhu, J., Miller, M. B., Vance, R. E., Dziejman, M., Bassler, B. L., and Mekalanos, J. J. (2002) Quorum-sensing regulators control virulence gene expression in Vibrio cholerae. Proc Natl Acad Sci U S A 99, 3129–3134. (14) Antonova, E. S., and Hammer, B. K. (2011) Quorum-sensing autoinducer molecules produced by members of a multispecies biofilm promote horizontal gene transfer to Vibrio cholerae. FEMS Microbiol Lett 322, 68–76. (15) Jang, J., Jung, K.-T., Yoo, C.-K., and Rhie, G.-E. (2010) Regulation of hemagglutinin/protease expression by the VarS/VarACsrA/B/C/D system in Vibrio cholerae. Microb Pathog 48, 245– 250. (16) Lenz, D. H., Miller, M. B., Zhu, J., Kulkarni, R. V., and Bassler, B. L. (2005) CsrA and three redundant small RNAs regulate quorum sensing in Vibrio cholerae. Mol Microbiol 58, 1186– 1202. (17) Lenz, D. H., and Bassler, B. L. (2007) The small nucleoid protein Fis is involved in Vibrio cholerae quorum sensing. Mol Microbiol 63, 859–871. (18) Freeman, J. A., and Bassler, B. L. (1999) A genetic analysis of the function of LuxO, a two-component response regulator involved in quorum sensing in Vibrio harveyi. Mol Microbiol 31, 665–677.

Page 10 of 12

(19) Lilley, B. N., and Bassler, B. L. (2000) Regulation of quorum sensing in Vibrio harveyi by LuxO and sigma-54. Mol Microbiol 36, 940–954. (20) Hunter, G. A., and Keener, J. P. (2013) Mechanisms underlying the additive and redundant Qrr phenotypes in Vibrio harveyi and Vibrio cholerae. J Theor Biol. (21) Wei, Y., Ng, W. L., Cong, J., and Bassler, B. L. (2012) Ligand and antagonist driven regulation of the Vibrio cholerae quorumsensing receptor CqsS. Mol Microbiol 83, 1095–1108. (22) Marraffini, L. A., and Sontheimer, E. J. (2010) CRISPR interference: RNA-directed adaptive immunity in bacteria and archaea. Nat Rev Genet 11, 181–190. (23) Peters, J. M., Silvis, M. R., Zhao, D., Hawkins, J. S., Gross, C. A., and Qi, L. S. (2015) Bacterial CRISPR: accomplishments and prospects. Curr Opin Microbiol 27, 121–126. (24) Larson, M. H., Gilbert, L. A., Wang, X., Lim, W. A., Weissman, J. S., and Qi, L. S. (2013) CRISPR interference (CRISPRi) for sequence-specific control of gene expression. Nat Protoc 8, 2180– 2196. (25) Cress, B. F., Toparlak, Ö. D., Guleria, S., Lebovich, M., Stieglitz, J. T., Englaender, J. A., Jones, J. A., Linhardt, R. J., and Koffas, M. A. (2015) CRISPathBrick: Modular Combinatorial Assembly of Type II-A CRISPR Arrays for dCas9-Mediated Multiplex Transcriptional Repression in E. coli. ACS Synth Biol 4, 987–1000. (26) Qi, L. S., Larson, M. H., Gilbert, L. A., Doudna, J. A., Weissman, J. S., Arkin, A. P., and Lim, W. A. (2013) Repurposing CRISPR as an RNA-guided platform for sequence-specific control of gene expression. Cell 152, 1173–1183. (27) Jia, H., Liang, T., Wang, Z., He, Z., Liu, Y., Yang, L., Zeng, Y., Liu, S., Tang, L., Wang, J., Chen, Y., and Xie, Z. (2014) Multistage Regulator Based on Tandem Promoters and CRISPR/Cas. ACS Synth Biol 3, 1007–1010. (28) Nielsen, A. A. K., and Voigt, C. A. (2014) Multi-input CRISPR/Cas genetic circuits that interface host regulatory networks. Mol Syst Biol 10, 763. (29) Cormack, B. P., Valdivia, R. H., and Falkow, S. (1996) FACSoptimized mutants of the green fluorescent protein (GFP). Gene 173, 33–38. (30) Svenningsen, S. L., Waters, C. M., and Bassler, B. L. (2008) A negative feedback loop involving small RNAs accelerates Vibrio cholerae’s transition out of quorum-sensing mode. Genes Dev 22, 226–238. (31) Morett, E., and Segovia, L. (1993) The sigma 54 bacterial enhancer-binding protein family: mechanism of action and phylogenetic relationship of their functional domains. J Bacteriol 175, 6067–6074. (32) Stanton, B. C., Nielsen, A. A. K., Tamsir, A., Clancy, K., Peterson, T., and Voigt, C. A. (2014) Genomic mining of prokaryotic repressors for orthogonal logic gates. Nat Chem Biol 10, 99– 105. (33) . Promoters/Catalog/Anderson - parts.igem.org. (34) Madar, D., Dekel, E., Bren, A., and Alon, U. (2011) Negative auto-regulation increases the input dynamic-range of the arabinose system of Escherichia coli. BMC Syst Biol 5, 111. (35) Ng, W. L., Wei, Y., Perez, L. J., Cong, J., Long, T., Koch, M., Semmelhack, M. F., Wingreen, N. S., and Bassler, B. L. (2010) Probing bacterial transmembrane histidine kinase receptorligand interactions with natural and synthetic molecules. Proc Natl Acad Sci U S A 107, 5575–5580. (36) Cecchini, F., Fajs, L., Cosnier, S., and Marks, R. S. (2016) Vibrio cholerae detection: Traditional assays, novel diagnostic techniques and biosensors. TrAC Trends in Analytical Chemistry 79, 199–209. (37) Dick, M. H., Guillerm, M., Moussy, F., and Chaignat, C. L. (2012) Review of two decades of cholera diagnostics--how far have we really come? PLoS Negl Trop Dis 6, e1845. (38) Gibson, D. G., Young, L., Chuang, R. Y., Venter, J. C., Hutchison, C. A., and Smith, H. O. (2009) Enzymatic assembly

ACS Paragon Plus Environment

10

Page 11 of 12

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Synthetic Biology

of DNA molecules up to several hundred kilobases. Nat Methods 6, 343–345. (39) . Ribosome Binding Sites/Prokaryotic/Constitutive/Community Collection parts.igem.org. (40) Benson, D. A., Karsch-Mizrachi, I., Lipman, D. J., Ostell, J., and Wheeler, D. L. (2005) GenBank. Nucleic Acids Res 33, D34– D38. (41) Lee, T. S., Krupa, R. A., Zhang, F., Hajimorad, M., Holtz, W. J., Prasad, N., Lee, S. K., and Keasling, J. D. (2011) BglBrick vectors and datasheets: A synthetic biology platform for gene expression. J Biol Eng 5, 12. (42) Sambrook, J., Russell, D. W., and Cold Spring Harbor, L. (2001) Molecular cloning : a laboratory manual / Joseph Sambrook, David W. Russell. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. (43) Hsu, P. D., Scott, D. A., Weinstein, J. A., Ran, F. A., Konermann, S., Agarwala, V., Li, Y., Fine, E. J., Wu, X., Shalem, O., Cradick, T. J., Marraffini, L. A., Bao, G., and Zhang, F. (2013) DNA targeting specificity of RNA-guided Cas9 nucleases. Nat Biotechnol 31, 827–832. (44) Doench, J. G., Hartenian, E., Graham, D. B., Tothova, Z., Hegde, M., Smith, I., Sullender, M., Ebert, B. L., Xavier, R. J., and Root, D. E. (2014) Rational design of highly active sgRNAs for

CRISPR-Cas9-mediated gene inactivation. Nat Biotechnol 32, 1262–1267. (45) Bassler, B. L., Wright, M., and Silverman, M. R. (1994) Sequence and function of LuxO, a negative regulator of luminescence in Vibrio harveyi. Mol Microbiol 12, 403–412. (46) Goldbeter, A., and Koshland, D. E. (1981) An amplified sensitivity arising from covalent modification in biological systems. Proc Natl Acad Sci U S A 78, 6840–6844. (47) Del Vecchio, D. (2012) Modularity in signaling systems. Phys Biol 9, 045008. (48) Milo, R., Jorgensen, P., Moran, U., Weber, G., and Springer, M. (2010) BioNumbers--the database of key numbers in molecular and cell biology. Nucleic Acids Res 38, D750–D753. (49) Saeidi, N., Arshath, M., Chang, M. W., and Poh, C. L. (2013) Characterization of a quorum sensing device for synthetic biology design: Experimental and modeling validation. Chem Eng Sci 103, 91–99. (50) Fuqua, W. C., Winans, S. C., and Greenberg, E. P. (1994) Quorum sensing in bacteria: the LuxR-LuxI family of cell density-responsive transcriptional regulators. J Bacteriol 176, 269–275. (51) Sahle, S., Mendes, P., Hoops, S., and Kummer, U. (2008) A new strategy for assessing sensitivities in biochemical models. Philos Transact A Math Phys Eng Sci 366, 3619–3631.

ACS Paragon Plus Environment

11

ACS Synthetic Biology

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 12

Table of Contents artwork

Biosensing Vibrio cholerae with genetically engineered Escherichia coli Maciej B. Holowko1, Huijuan Wang1, Premkumar Jayaraman1 and Chueh Loo Poh1*

ACS Paragon Plus Environment

1