Engineering bacterial thiosulfate and tetrathionate sensors for detecting gut inflammation

Abstract There is a groundswell of interest in using genetically engineered sensor bacteria to study gut microbiota pathways, and diagnose or treat associated diseases. Here, we computationally identify the first biological thiosulfate sensor and an improved tetrathionate sensor, both two‐component systems from marine Shewanella species, and validate them in laboratory Escherichia coli. Then, we port these sensors into a gut‐adapted probiotic E. coli strain, and develop a method based upon oral gavage and flow cytometry of colon and fecal samples to demonstrate that colon inflammation (colitis) activates the thiosulfate sensor in mice harboring native gut microbiota. Our thiosulfate sensor may have applications in bacterial diagnostics or therapeutics. Finally, our approach can be replicated for a wide range of bacterial sensors and should thus enable a new class of minimally invasive studies of gut microbiota pathways.


Table of contents
Appendix Figure S1. S. typhimurium TtrS/TtrR domain layout and gene locus Appendix Figure S2. Identification of P phsA regulatory features Appendix Figure S3. Plasmid maps for inducible ThsSR production Appendix Figure S4. Shal_3129-induced activation of P phsA342 Appendix Figure S5. Optimization of protein expression using inducible promoters to give the highest dynamic for ThsSR Appendix Figure S6. Ligand inhibition assay for ThsSR Appendix Figure S7. Schild plot analysis of ThsSR tetrathionate inhibition Appendix Figure S8. Glucose sensitivity of thiosulfate-induced activation of ThsSR at P phsA Appendix Figure S9. Identification of P ttrB344 regulatory features Appendix Figure S10. Plasmid maps for inducible TtrSR production Appendix Figure S11. Sbal195_3858-induced activation of P TtrB344 Appendix Figure S12. Optimization of protein expression using inducible promoters to give the highest dynamic for TtrSR Appendix Figure S13. Glucose sensitivity of tetrathionate-induced activation of P ttrB185-269 Appendix Figure S14. TtrSR response in presence of 1 mM tetrathionate and 10 mM other TEAs Appendix Figure S15. Optimization of the ThsSR for use in the mammalian gut Appendix Figure S16. Optimization of TtrSR for use in the mammalian gut Appendix Figure S17. Plasmid maps for constitutive ThsSR production used in in vivo studies Appendix Figure S18. Plasmid maps for constitutive TtrSR production used in in vivo studies Appendix Figure S19. Fluorophore maturation time course of anaerobically grown Nissle bacteria in PBS+1 mg/mL chloramphenicol Appendix Figure S20. Flow cytometry profile before and after filtering through a 5 µM syringe filter Appendix Figure S21. ThsSR performance in ex vivo ligated colon explants. Appendix Figure S22. S. baltica TtrSR performance in ex vivo ligated colon explants. Appendix Figure S23. ThsSR and TtrSR sensor performance with and without 3% DSS Appendix Figure S24. Histograms of representative fecal samples from healthy control mice and DSS-treated mice Appendix Figure S25. 2D flow cytometry data Appendix Figure S26. Histologic scoring of all mice in the healthy (-DSS) and DSStreated cohorts for wild type and inactivated mutant ThsSR and TtrSR strains Appendix figure S27. Relationship between histologic score and fluorescence output of the ThsSR and ThsSR D57A sensors Appendix Figure S28 Appendix Figure S2. Identification of P phsA regulatory features. (A) Two potential operator sites, a 10 bp inverted repeat (light blue) and an 18 bp direct repeat (dark blue), were identified in the Shal_3129-regulated promoter (P phsA342 ). Each 18 bp repeat contains a 6 bp inverted repeat (black) separated by 2 bp, and therefore may be a complete operator. A consensus CRP binding site (purple) is located between the two 18 bp repeats. Truncation sites are indicated with arrows and the position in the promoter sequence. (B) Response of promoter truncations to thiosulfate. White bars are sfGFP fluorescence in the absence of ligand, black bars are with 5 mM thiosulfate, and red circles represent the fold induction (sfGFP fluorescence 5 mM/0 mM). Deletion of the first 150 bp of the 342 bp intergenic region (P phsA151-342 ) had no effect on thiosulfateinduced promoter activation, indicating the 10 bp inverted repeat is not the ThsR binding site. Removal of the first 18 bp direct repeat (dark blue) also had no effect on promoter activation. Deletion of the promoter through the predicted CRP site or mutagenesis of the CRP site diminished sensor output (P phsA182-342, P phsA342~CRP ; 25-fold to 4-fold), and deletion of the second 18 bp repeat abolished thiosulfate-induced signaling (P phsA219-342 ). Mutagenesis of either 6 bp inverted repeat in the second 18 bp element (P PhsA342~O1 , P PhsA342~O2 ) also resulted in loss of signaling, supporting this sequence as the likely operator site. Mutated sequences are indicated under the wild type promoter sequence.
No truncations were identified that gave enhanced promoter response, therefore the entire intergenic region was used as the output promoter.
Appendix Figure S3. Plasmid maps for inducible ThsSR production. Figure S4. Shal_3129-induced activation of P phsA342. Closed circles represent the wild type and open squares the phospho-accepting null mutant (D57A).
Appendix Figure S5. Optimization of protein expression using inducible promoters to give the highest dynamic for ThsSR. Data are shown for all aTc and IPTG conditions tested in the (A) absence of thiosulfate (0 mM), (B) presence of saturating thiosulfate (5 mM), and (C) the fold difference between the two (sfGFP fluorescence 5 mM/0 mM thiosulfate).

B A
Appendix Figure S7. Schild plot analysis of ThsSR tetrathionate inhibition. (A) An increase in k 1/2 of the sensor was observed at higher concentrations of tetrathionate, however maximal response could not be achieved in the presence of tetrathionate. Thiosulfate concentrations >10 mM were not used because of an additional lowsensitivity response of ThsSR that would complicate analysis. (B) Schild plot of doseresponse data. Slope of the Schild plot was not 1 indicating tetrathionate is not a competitive antagonist or that the system is not in equilibrium.

A B
Appendix Figure S8. Glucose sensitivity of thiosulfate-induced activation of ThsSR at P phsA . The full-length promoter (P phsA342 ), a 5' truncated promoter removing the CRP site (P phsA182-342 ), and a promoter with CRP binding motif mutated (P phsA342CRP ) were tested in M9 minimal media with 0.4% glycerol or 0.4% glucose as the carbon source. White bars are in the absence of thiosulfate, black bars are in the presence of 5 mM thiosulfate, and red circles are the fold induction (sfGFP fluorescence 5 mM/0 mM). Appendix Figure S9. Identification of P ttrB344 regulatory features. (A) Potential operator sites based on direct or inverted repeat sequences are highlighted in a variety of colors. Truncation sites are indicated with arrows and the position in the promoter sequence. The predicted -35, -10 and +1 sites are based on sequence similarity to P ttrB from S. typhimurium. (B) Response of P ttrB promoter truncations to tetrathionate. The numbers indicated in the x-axis are the bps in the truncated sequence relative to the original 344 bp intergenic region. White bars are fluorescence in the absence of ligand and black bars are in the presence of 1 mM tetrathionate. Deletion of the first 184 bp of the 344 bp intergenic region (P ttrB185-344 ) had no effect on thiosulfate-induced promoter activation, indicating the majority of the highlighted sequences are not the operator site. Removal of the terminal 75 bp (P ttrB1-269 ) resulted in decreased basal fluorescence in the absence of tetrathionate and increased fluorescence in the presence of tetrathionate, relative to the full intergenic region. A minimal promoter with these combined truncations (P ttrB185-269 ) gave improved performance over the full-length promoter sequence. This truncated promoter also minimizes the potential for cross-talk from other transcription factors and was therefore used as the TtrSR output promoter in this study. (C) Additional promoter characterization of the minimal promoter P ttrB185-269 . Potential FixJ-like operator half sites are highlighted in black, green, and purple, and the FNR site in red. Promoter mutations are indicated below the wild type sequence and truncation positions are indicated with arrows. (D) P ttrB could be further truncated at the 5' end by two bp to P ttrB187-269 with no loss of function, however removal of the first inverted repeat highlighted in black decreased promoter output 14-fold (P ttrB192-269 ). No further truncation was tolerated at the 3' end, likely due to removal of the +1 transcriptional start site. (E) Mutagenesis of regions of interest within the truncated promoter P ttrB185-269 .
Appendix Figure S10. Plasmid maps for inducible TtrSR production.   Appendix Figure S15. Optimization of the ThsSR for use in the mammalian gut. 1) A library of varying strength constitutive promoters from the Anderson promoter library (indicated by identity J231XX) and designed synthetic ribosome binding sites (RBS) (if multiple were designed per promoter, they are designated with the letters a-d) were incorporated upstream of ThsS and ThsR.
2) The ThsR and ThsS plasmids were then combined together and screened for optimal expression in E. coli Nissle 1917 (fold induction +/-5 mM thiosulfate is indicated). Combinations that had significant growth defects are indicated in grey with a *. 3) In the best construct, the promoter of the constitutive mCherry marker was increased to allow for selection of the biosensor from the normal microbial floral and the construct was tested in anaerobic growth conditions. The final constructs used for in vivo studies are shown at the bottom.

Inducible promoters Domesticated strain Aerobic growth
Constitutive promoters Probiotic strain approved for human consumption Anaerobic growth Add strong mCherry marker Final Optimized for increased mCherry production and anaerobic growth Make library of SK and RR plasmids with varying strength constitutive promoters Screen library Increase mCherry production in best construct and test in anaerobic chamber Appendix Figure S16. Optimization of TtrSR for use in the mammalian gut. 1) A library of varying strength constitutive promoters from the Anderson promoter library (indicated by identity J231XX) and a designed synthetic RBS were incorporated upstream of TtrS and TtrR.
2) The TtrR and TtrS plasmids were then combined together and screened for optimal expression in E. coli Nissle 1917 (fold induction +/-1 mM tetrathionate is indicated). Combinations that had significant growth defects are indicated in grey with a *. 3) In the best construct, the promoter of the constitutive mCherry marker was increased to allow for selection of the biosensor from the normal microbial floral and the construct was tested in anaerobic growth conditions. This resulted in a significant decrease in sensor performance. Because the best construct from the initial screen was the weakest promoter, additional synthetic RBSs designed to be weaker were designed with this promoter and the strong mCherry promoter was incorporated into all RR plasmids. These plasmids were then combined and screened in anaerobic conditions. Although the best construct appears to be TtrR 09c and TtrS 14, the inactivating control plasmid (D55A) was not made in time for in vivo experiments. Therefore, the best available plasmid was used (indicated by a black box) which gave good dynamic range. The final constructs used for in vivo studies are shown at the bottom. Appendix Figure S18. Plasmid maps for constitutive TtrSR production used in in vivo studies. Appendix Figure S20. Flow cytometry profile before and after filtering through a 5 µM syringe filter. Data are only collected for cytometer counts demonstrating high mCherry fluorescence on the FL3 channel. The majority of counts from fecal and colon samples fall below this threshold and are not recorded by the cytometer. Data are then gated to only include counts with a similar FSC and SSC profile of E. coli (black ellipse), and with high fluorescence in both the FL2 and FL3 channels (to the right of the black bar).