Search

Accurate control of a biological process is essential for many critical functions in biology, from the cell cycle to proteome regulation. To achieve this, negative feedback is frequently employed to provide a highly robust and reliable output. Feedback is found throughout biology and technology, but due to challenges posed by its implementation, it is yet to be widely adopted in synthetic biology. In this paper we design a synthetic feedback network using a class of recombinase proteins called integrases, which can be re-engineered to flip the orientation of DNA segments in a digital manner. This system is highly orthogonal, and demonstrates a strong capability for regulating and reducing the expression variability of genes being transcribed under its control. An excisionase protein provides the negative feedback signal to close the loop in this system, by flipping DNA segments in the reverse direction. Our integrase/excisionase negative feedback system thus provides a modular architecture that can be tuned to suit applications throughout synthetic biology and biomanufacturing that require a highly robust and orthogonally controlled output.

Proteins, so small that one billion would fit on a full stop, carry out most of the vital activities in living cells; they drive chemical reactions, transport cargoes, communicate with the outside world and even segregate chromosomes. A novel approach now allows us to monitor single proteins in complicated molecular machines, and it seems that biological components wear out and get replaced just as they do in man-made machines.

The bacterial flagellar motor is one of the most complex protein machines found in nature and how it self-assembles and produces force are very much open questions. In this thesis, I study the constituent parts of the motor in vitro, with the focus being primarily on FliG, a protein which helps to form the motor’s cytoplasmic ring. The experimental approach employs a DNA scaffold to direct the in vitro formation of the FliG ring, with the aim of determining both the number of proteins in the complex and the means by which they associate to form a circular structure. One leading theory states that the FliG proteins interact through a process of domain-swap oligomerization and the primary goal of this work is to search for evidence of this process with the aid of in vitro DNA templates. To attach the FliG protein to the underlying DNA structures, I produced FliG-DNA conjugates through two different means. First, I bound both tris- and pentakis-NTA-modified DNA strands to the Histidine-tag of FliG and showed that these conjugates could be arranged on simple, linear DNA templates. In order for the proteins to be arranged with the spacing observed in the cytoplasmic ring however, a different conjugation strategy was required in which a maleimide-modified DNA strand was reacted with a single-cysteine FliG mutant. After issues relating to Histidine-tag-mediated FliG dimerization were resolved, these conjugates were able to be organized on both linear DNA strands and origami tiles with the desired spacing between the proteins. No interactions between either template-bound or free-floating FliG proteins could be observed however, meaning a new strategy was needed. One model of cytoplasmic ring assembly predicts that binding of FliG to the C-terminal domain of the membrane-bound FliF protein induces a conformational change in both partners which then triggers interactions between neighbouring FliG proteins. To test this, I designed a peptide corresponding to the FliFC domain of Salmonella and E. coli FliF and incorporated it into my system. Several biophysical assays indicated that the FliFC peptide stably bound to purified wild-type Salmonella FliG and that binding to the peptide increased the overall stability of the protein. This constitutes the first confirmation that the cytoplasmic domain of FliF is sufficient for FliG binding in enteric bacteria. Following this discovery, a maleimide-modified DNA strand was then conjugated to a single-cysteine version of the peptide and the peptide-DNA conjugates organized with various stoichiometries and spacing on DNA templates to create an array of ‘binding platforms’ for the FliG protein. FliG bound to the organized peptides in a one-to-one fashion but interactions between the organized FliG proteins were still not detected. In a final effort to stimulate FliG-FliG binding, a radially-symmetric DNA nanostructure was designed with dimensions and a geometry which matched those found in the cytoplasmic rings of bacteria. Gels and AFM microscopy proved that the 12 designed strands assembled correctly into the desired structure and that FliFC-DNA conjugates with FliG proteins bound could be organized on the circular arrangement of binding sites around the structure’s outer rim. Even in such an arrangement, interactions between neighbouring proteins could not be observed, and though higher concentrations of FliFC and FliG were tested, issues with purification prevented accurate characterization of these complexes. Thus, though the domain-swap model for FliG oligomerization was therefore neither confirmed or denied, the use of a DNA scaffold to organize protein molecules and study their interactions holds promise as a technique to investigate the proteins of other biological systems and will no doubt be employed to great effect in the coming years.

Bacteria use Two Component Systems (TCS) to sense and respond to changes in their external environment. TCS are used to navigate to nutrients or away from toxins (chemotaxis) and to adapt to changes in osmolarity (osomosensing). TCS are composed of a histidine protein kinase (HPK) which trans-autophosphorylates in response to environmental change, transferring the phosphoryl group to a cognate response regulator (RR). Phosphorylated RRs modulate an output response such as protein-protein interaction for chemotaxis, and transcription for osmosensing. RRs are composed of a conserved amino terminal REC domain, and where present a variable effector domain. CheY, the chemotaxis RR, contains only a REC domain, whilst OmpR, the osmosensing RR, also contains a DNA binding effector domain. Recently, TCS have been used in synthetic biology applications due to their modularity and conserved signalling mechanism. This thesis aimed to investigate whether it was possible to design a synthetic TCS composed of fused chemotaxis and osmosensing components. Synthetic RRs were designed, fusing the highly conserved REC domains of CheY and OmpR upstream of the OmpR effector domain. REC domains were fused across the α4-β5-α5 region, a region which transmits REC domain phosphorylation into effector domain activation. Synthetic RRs were designed to undergo phosphotransfer to their fused REC domains from the chemotaxis HPK, CheA, activate the attached OmpR effector domain and bind promoter DNA. Four chimeric RRs were created, although only three were structurally viable; F2, F3 and F4. Each fusion bound CheA, and F3 and F4 bound CheA with a significantly higher affinity than CheY. The chimeric RRs could all be phosphorylated byCheA-P; F4 and F3 were phosphorylated to wild-type levels. DNA binding affinitywas investigated with fluorescence anisotropy, hosphorylated and unphosphorylated F3 could not bind promoter DNA. F2 bound promoter DNA regardless of phosphorylation state. These data indicate that phosphorylation of the F2 REC domain does not lead to activation of the effector domain. F2 is likely to be constitutively active suggesting a previously unknown role for OmpR α5 as a mediator of effector domain activation. Furthermore, using a simple fusion approach to design RRs is not a viable method to create a synthetic TCS with a controllable output.

Motility and chemotaxis have been implicated in the process of biofilm formation in a wide range of species. Using a combination of microscopy and image analysis, genetics, microbiology and biochemistry, the initial approach of Rhodobacter sphaeroides cells to a solid surface has been characterised. Interestingly, these data suggest that for R. sphaeroides alterations in motility and swimming behaviour may result in differences in biofilm formation simply by changing the number of cells which reach the surface. This is in contrast to a few other well-studied species where the motility apparatus, the flagellum, has been shown to play an active role in surface sensing and the transition to biofilm growth. Tracking swimming cells and measuring surface attachment revealed that changes in motility affect the ability of cells to attach to a surface, with non-motile cells attaching least and mutants with frequent stops attaching less than smooth swimming cells with few stops. Tracking attaching cells and classifying their method of attachment revealed that flagellar tethering is not essential for R. sphaeroides attachment. Competition assays with fluorescently labelled strains showed that the initial imbalance between motile and non-motile cells remains as microcolonies develop over 48 hours,and the proportion of non-motile cells remains fairly constant. Development on a surface over 48 hours was similar for motile and non-motile strains, including aflagellate strains, once attached. Using parameters calculated by tracking swimming cells to calculate the effective diffusion coefficient in a simple model of cell movement suggested that motion alone could explain the differences in attachment without assuming different cell properties. In particular, aflagellate strains might be hindered from surface attachment by their reduced motility alone. This is interesting since some other bacterial species use the flagellum as a surface sensor.

For applications from biosensor generation to synthetic biology, the ability to accurately and quantitatively generate input-output data from biological systems over a wide dynamic range is becoming increasingly important. This study has demonstrated that a simple approach utilising multiple promoters, recognising the same transcriptional activator but with differing affinities, linked to compatible reporter genes can achieve this goal. This simple system highlights that even for complex promoters with multiple binding sites we can use the ratio of two reporters to obtain accurate and reproducible input-output data for reporters contained on a high copy number plasmid without the need for any background subtraction or accurate determination of cell number. It has also demonstrated that the precise promoter strengths are not too crucial to the design, provided they are significantly different, as taking the ratio of either the high affinity/low affinity or medium affinity/low affinity channels give similar improvements over any single channel alone. The modularity of this system means that it should be possible to exchange the IPTG inducible promoter and hence place the ntrC* gene under the regulation of any environmentally important promoter. It should also be possible to simply construct different strength promoters for the same transcriptional activator by standard DNA synthesis techniques, therefore allowing direct ratiometric detection of specific active transcription factors. Finally, there is considerable interest in the biosensor field in the use of immobilised cells for field based applications. In these cases, determining the exact number of viable cells present at the time of use may be problematic and obtaining high signal levels may also be essential. In this regard, the use of ratiometric reporters from constructs on high copy number plasmids could alleviate many of these potential problems and significantly simplify the required biosensor detection equipment.