Tuning Bacterial Behaviour Judy Armitage University of Oxford Department of Biochemistry and Oxford Centre for Integrative Systems Biology StoMP 2009

E.coli chemotaxis-the best understood “system” in Biology E.coli has one constitutive chemosensory pathway. Biases swimming direction by regulating motor switching Not essential and phenotype obvious All components known, kinetics of all reactions, copy number of all proteins, structures of most Cells respond to ~2 molecules over 6 orders of magnitude Paradigm for 2 component pathways

E.coli chemotaxis 4 dedicated constitutive membrane spanning receptors (MCPs) plus Aer One sensory pathway via CheW (linker), CheA (histidine protein kinase), CheY (response regulator) Chemotaxis is via biasing a normally random swimming pattern Adaptation of MCPs via single CheB/R methylation system Mutations give either smooth swimming or tumbling phenotypes Unusual HPK pathway Termination of CheY-P through CheZ-not HPK phosphatase MCP  CheA  CheY/B Histidine protein kinase signalling

Rhodobacter sphaeroides Member of  -subgroup proteobacteria Heterotrophic, photoheterotrophic, anaerobic respiration, CO 2 - N 2 - fixation, hydrogenase, fermentation Quorum sensing, biofilm forming Membrane differentiation-aerobic vs photoheterotrophic Targeting-flagellum, cell division proteins, chemotaxis proteins

Chemotaxis in R.sphaeroides Single unidirectional flagellum (under lab conditions) Stopping involves a molecular brake 3 chemosensory operons Need transport and possibly partial metabolism for chemotactic response Why have 3 chemosensory pathways to control on flagellar motor? 4 CheAs 8 membrane spanning MCPs 4 cytoplasmic Tlps 6 CheYs 2 CheBs NO CheZ

R.sphaeroides uses a brake to stop

Activity of the chemotaxis proteins in vitro Is there “cross talk” between apparently homologous proteins encoded by the different operons? In vitro phosphotransfer measured between 4 CheA HPKs and the 6 CheY and 2 CheB RRs CheA has H on Hpt domain

Pattern of in vitro phosphotransfer

Kinase and Response Regulators CheA2 will phosphotransfer to all Che Response Regulators-wherever encoded (CheOp1, CheOp2 or CheOp3) CheA1 will only phosphotransfer to proteins encoded in own operon (CheOp1) CheA3/4 will only phosphotransfer to proteins encoded in its operon (CheOp3) How is discrimination achieved?

Chemotaxis: in vitro phosphotransfer Horribly complex!

Where are the gene products ? Do the genes encode proteins that make separate or cross-talking pathways in vivo ? G(C,Y)FP –(N and C terminal) fusions to all che genes; replaced in genome behind native promoters and tested for normal behaviour Confirmed by immuno-elecronmicroscopy

Pathways targeted to different part of cell Red: CheOp 2 Blue: CheOp 3. Cytoplasmic general: CheB1, CheB2, CheY3, CheY4, CheY6

Localisation Chemosensory proteins are physically separate in the cell CheOp2 encoded proteins with MCPs at poles and CheOp3 with Tlps in cell centre CheAs physically separate and therefore do not cross phosphotransfer in vivo ? What controls localisation? Why have 2 physically separate chemosensing pathways? Is this common? Does it only apply to taxis pathways? Would not have been identified without in vivo investigations

TlpT Putative cytoplasmic chemoreceptor Essential to chemotaxis to a range of organic acids Co-localises in the cytoplasm with CheA 3, A 4 and CheW 4, TlpC, TlpS PpfA (Slp) Homology to ParA family type 1 DNA partitioning proteins, contains “ Walker” type ATPase domain Deletion results in reduced taxis to a range of organic acids, but normal growth Localisation requires two CheOp3 proteins

PpfA regulates the number and position of cytoplasmic clusters Cephalexin treatedWS8N  ppfA

PpfA: a protein partitioning factor PpfA (Protein) signal for new cluster formation, and anchoring midcell, ¼ and ¾ positioning. ATP dependent (Walker box mutants=null) Partner/interactions? ParA (DNA) characteristic midcell, ¼ and ¾ positioning of plasmids Polymerisation? Oscillation? ATP/ADP ParA switch ParB and parC(S) partners

Cytoplasmic chemoreceptor TlpT TlpT :nucleating protein for cytoplasmic cluster?

How common is this protein segregating system? 53% of complete genomes in databases have more than one putative chemotaxis pathway (max 8) 60% of these have putative ppfA in one Che operon Of these 83% also have putative cytoplasmic chemoreceptor gene adjacent and all have disordered N-terminal domain

R.sphaeroides chemosensory pathway: the happiness centre? Metabolic state Kinase vs phosphatase CheY 6 -P A3A4A3A4 External world A2A2 CheY 3/4 -P CheA 3 is a kinase and specific phosphatase for CheY 6 Model prediction: phosphoryl groups originating from CheA 3 A 4 can end up on CheY 3 and CheY 4 using CheB 2 and CheA 2 as a phosphoconduit. His-asp-his-asp phosphorelay between clusters is route to integrating and balancing the signals from metabolism and the external environment. Dominant CheY6-P level regulated by CheA3 kinase:phosphatase activity CheB 2 -P

How do these pathways control the single motor? How is discrimination achieved?

What determines localisation Is it operon position on chromosome? Are there specific interaction domains?

Rhodobacter sphaeroides CheA Proteins CheA4 CheA3 CheA1 P1P2P3P4P5 P1P2P3P4P5 CheA2 P1P2P3P4P5 P1P2P3P4P5 P3P4P5 P3P4P5 P1P5 Swapped P1 domains and looked at phosphotransfer Swapped P5 domains and looked at localisation Created chimeras with same P1 domains in CheAs at both cell locations

Conclusions There is internal organisation in bacteria with apparent homologues targeted to specific sites in the cell (high throughput in vitro analysis may give misleading interaction patterns) Interaction between cognate HPK-RR depend on very few amino acids (motifs may allow engineering of novel interactions)

The people who did the work George Wadhams Steven Porter Mark Roberts Sonja Pawelczyk Mila Kojadinovic Kathryn Scott Nicolas Delalez Mostyn Brown David Wilkinson Christian Bell Yo-Cheng Chang Murray Tipping Gareth Davies Elaine Byles COLLABORATORS Dave Stuart Philip Maini Marcus Tindall Charlotte Deane Rebecca Hamer