Diffusible signalling Factor (DSF) is a long chain cis-unsaturated fatty acid (link to Jess background?) produced by bacteria of the Xanthomonas genus, including the CF lung pathogen Stenotrophomonas maltophilia 1,2. To build our synthetic DSF detector, we chose to reconstitute the well-characterised DSF signal perception system from Xanthomonas campestris in our E. coli chassis. As can be seen in Fig 1, the RpfC/RpfG system, which perceives and responds to DSF, is more complex than the average two component system. DSF is detected at the periplasmic side of the membrane by the RpfC sensor protein, which autophosphorylates and then transfers its phospho group to the cytoplasmic response regulator RpfG. However, instead of acting as a direct transcriptional regulator phosphorylated RpfG acts as a phosphodiesterase cleaving the second-messenger molecule cyclic di-guanosine monophosphate (c-di-GMP) into 2 GMP molecules3,4. This reduction in cellular c-di-GMP relieves inhibition of the transcriptional activator Clp, which is then free to bind to the manA promoter and allow transcription of the downstream genes5. In our synthetic system we fused the manA promoter to promoter-less GFP (GFP biobrick hyperlink in here) to give a fluorescent output.

Chromosomal DNA from Pseudomonas Aeruginosa PA01 strain was kindly gifted to us by Robert Ryan and Shi-Qi An from The Division of Molecular Microbiology in the College of Life sciences at the University of Dundee. This was used as a template for the amplification of the pqsA promoter. The pqsA promoter region was cloned into the pSB1C3 plasmid (to give Biobrick (BBa_K1315001), and was then subcloned into pBluescript. Promoterless mCherry was amplified using BBa K562011 as a template and was cloned into pBluescript downstream of the pqsABCDE promoter. The PpqsA-mCherry construct was then subcloned into pUniprom. The pqsR gene (present in BBa_K1157001) which was designed and cloned by Lin Ang Chieh from iGEM13-NTU-Taida was used as a template for the amplification of pqsR with an influenza virus haemagglutinin (HA) tag coding sequence, which can be detected with commercial antibodies. This tag was added to the C-terminus of the protein to facilitate immunohistochemistry.

The plasmid was verified by sequencing.

A series of western blots were performed to test for the sequential production of each of the HA-tagged proteins. We transformed our plasmid series (encoding RpfC-HA, RpfC-HA + Clp-HA and RpfC-HA +Clp-HA + RpfG-HA) into E. coli strain MC1061 and blotted for recombinant protein production.

As shown in Fig 3, successful production of RpfC-HA (expected mass 81kDa), RpfG-HA (expected mass of 43kDa) and Clp-HA (expected mass 27kDa) was observed. We then went on to test our DSF detector by incubating E. coli producing the three HA-tagged proteins with increasing concentrations of pure DSF (obtained from Sigma) and monitoring for production of GFP from the manA promoter by Western blotting. As shown in Fig 4A, we were able to specifically detect GFP production in E. coli cells harbouring our DSF-detection plasmid. However we noted that GFP was being produced in the absence of exogenous DSF addition. We reasoned that this could be due to either: (i) the endogenous activation of the manA by an E. coli transcriptional regulator or (ii) the presence of DSF or a DSF-like compound in the cells and/or growth medium. To distinguish between these two possibilities we introduced pUniprom carrying the manA-gfp fusion alone (i.e. in the absence of the rpfC, rpfG and clp genes) into E. coli and blotted for GFP production. As shown in Fig 4B, GFP was produced by E. coli in the absence of the X. campestris DSF sensing proteins. We therefore conclude that the manA promoter is recognised by an E. coli transcription factor.

For the DSF sensing system to work we need to eliminate the background activation of PmanA-gfp. Since the X. campestris Clp protein is very similar to the E. coli cyclic AMP receptor protein (CRP) transcripton factor (46% identity, 71% similarity over the whole length of the protein), we next tested whether the production of GFP was dependent upon CRP. To this end we used a crp deletion strain from the E. coli Keio mutant collection6 and compared GFP production in this strain background with the cognate wild type strain (Fig 5). However GFP was still produced (although possibly at a lower level) than in the crp+ background.

We are continuing to investigate this issue by screening other E. coli mutants carrying deletions in genes coding for transcription factors which show similarity to Clp, for example the anaerobic regulator, FNR. We anticipate that it may be necessary to make combinatorial mutations to completely eliminate exogenous GFP production.