The photobiology of cyanobacteriochrome Tlr0924

A thesis submitted to the University of Manchester for the degree of Doctor of Philosophy (PhD) in the Faculty of Life Sciences

2014

Anna Hauck

2 | P a g e

Table of Contents

Abbreviations ...... 6 List of figures ...... 10 List of tables ...... 12 Abstract ...... 13 Declaration ...... 14 Copyright statement ...... 14 Acknowledgements ...... 15 Preface to the alternative format ...... 16

Chapter 1 – Introduction ...... 17 1. Introduction ...... 18 1.1 Basis of light absorbance ...... 19 1.2 Photoreceptor structure ...... 25 1.3 Classes of photoreceptors ...... 29 1.3.1 Flavinoid photoreceptors ...... 29 1.3.1.1 Cryptochromes ...... 29 1.3.1.2 Blue light sensors using flavin adenine dinucleotide ...... 31 1.3.1.3 Light, oxygen or voltage sensors ...... 32 1.3.2 Isomerisation based photoreceptors ...... 33 1.3.2.1 Retinylidene proteins ...... 33 1.3.2.2 Xanthopsins ...... 36 1.3.2.3 Phytochromes ...... 38 1.3.2.4 Cyanobacteriochromes...... 40 1.4 Application ...... 48 1.5 Thesis aims and objectives...... 50

Chapter 2 – Experimental Set-up ...... 51 2. Spectroscopy ...... 52 2.1 Stationary absorption spectroscopy ...... 54 2.1.1 UV/Visible absorption spectroscopy ...... 55 2.1.2 Low-temperature UV/Visible absorption spectroscopy ...... 55 2.1.3 Infrared absorption spectroscopy ...... 57 2.2 Time-resolved absorption spectroscopy...... 59

3 | P a g e

2.2.1 Ultrafast visible absorption spectroscopy ...... 60 2.2.2 Ultrafast infrared absorption spectroscopy ...... 62 2.2.3 Flash photolysis ...... 63 2.3 Data analysis ...... 64 2.4 Conclusion…………………………………………………………………………………………………………….. 64

Chapter 3 – Heterologous expression and spectral characterisation of the cyanobacteriochrome Tlr0924 ...... 65 3.1 Abstract ...... 66 3.2 Introduction...... 67 3.3 Materials and methods ...... 69 3.3.1 Plasmid construction ...... 69 3.3.2 Protein expression ...... 70 3.3.3 Protein purification ...... 70 3.3.4 Spectroscopy ...... 70 3.4 Results ...... 72 3.4.1 Expression and purification of Tlr0924 full-length protein ...... 72 3.4.2 Stationary photoconversion properties ...... 74 3.4.3 Importance of a second thioether linkage ...... 76 3.4.4 Tlr0924 spectral sensitivities ...... 78 3.4.5 Chromophore dark reactions ...... 84 3.4.6 Temperature effects on Tlr0924 ...... 86 3.4.7 Cryobuffer compatibility ...... 88 3.5 Discussion ...... 92 Chapter 4 – The photoinitiated reaction pathway of full-length cyanobacteriochrome Tlr0924 ...... 94 4.1 Abstract ...... 95 4.2 Introduction...... 96 4.3 Experimental procedures ...... 99 4.3.1 Protein expression and purification ...... 99 4.3.2 Ultrafast transient absorption spectroscopy ...... 99 4.3.3 Laser flash photolysis ...... 100 4.3.4 LED flash photolysis...... 101 4.3.5 Cryotrapping ...... 101 4.3.6 Global analysis ...... 102

4 | P a g e

4.4 Results ...... 103 4.4.1 Ultrafast transient UV/Visible dynamics ...... 103 4.4.2 Ultrafast transient IR dynamics ...... 105 4.4.3 Slow, thermally-driven dynamics ...... 106 4.4.4 Cryotrapping of intermediate states ...... 112 4.5 Discussion ...... 114 Chapter 5 – Comprehensive analysis of the green to blue photoconversion of full-length CBCR Tlr0924 ...... 118 5.1 Abstract ...... 119 5.2 Introduction ...... 120 5.3 Materials and methods ...... 122 5.3.1 Protein expression and purification ...... 122 5.3.2 Ultrafast transient absorption spectroscopy ...... 122 5.3.3 Laser flash photolysis ...... 123 5.3.4 Cryotrapping ...... 123 5.3.5 Global analysis ...... 124 5.4 Results and discussion ...... 125 5.4.1 Ultrafast transient absorption ...... 125 5.4.2 Nanosecond to millisecond transient absorption ...... 129 5.4.3 Millisecond to second transient absorption ...... 131 5.4.4 Cryotrapping ...... 133 5.5 Conclusion ...... 135 5.6 Acknowledgements...... 136 5.7 Supporting Information ...... 137 5.8 Note on PCB Pr to Pb photoconversion ...... 143

Chapter 6 – Discussion ...... 145 6.1 Discussion ...... 146 6.2 Future Work ...... 153 6.3 Conclusion ...... 155 References ...... 156

Word count: 48 109

5 | P a g e

Abbreviations

A A Absorbance A/G cyc Adenylate/guanylate cyclase Asp Aspartate B BLUF Blue-light sensors using flavin adenine dinucleotide Bph Bacterial phytochrome BR Bacteriorhodopsin bZ Basic zipper C c Concentration c Speed of light in vacuum (2.998 x 108 m.s-1) CBCR Cyanobacteriochrome CBS Cystathionine β-synthase CCT Cryptochrome C-terminal domain c-di-GMP Cyclic dimeric guanosine monophosphate Cph1 Cyanobacterial phytochrome 1 Cry Cryptochrome Cryptochrome found in Drosophila, Arabidopsis, Synechocystis and Cry-DASH Homo Cys, C Cysteine D D Aspartate DFG Difference frequency generator E ε Extinction coefficient E Energy EAL Diguanylate phosphodiesterase EADS Evolution associated difference spectrum EM Electromagnetic EMR Electromagnetic radiation ESA Excited state absorbance eV Electron volt

6 | P a g e

F Ф Quantum yield f F box F Fluorescence F Phyenylalanine fad Photolyase alpha domain FAD Flavin adenine dinucleotide FMN Flavin mononucleotide FWHM Full width half maximum G GAF cGMP phosphodiesterase, adenylate cyclase, FhlA protein GGDEF Diguanylate cyclase GPCR G-protein coupled receptor GTP Guanosine triphosphate GSB Ground state bleach H H Planck constant (6.625 x 10-34 J.s) Histidine kinases, adenylate cyclases, methyl accepting proteins and HAMP phosphatises His Histidine HisK Histidine kinase (related) HO1 Heme oxygenase HY2 PФB synthase I I Light passed through sample

I0 Incident light IAM Iodoacetamide IC Internal conversion IPTG Isopropyl β-D-1 thiogalactopyranoside IR Infrared ISC Intersystem crossing IVR Inner vibrational relaxation J J Joule K

7 | P a g e

Kelch Kelch repeat

kp Rate of productive chemical pathways

kf Rate of fluorescence deexcitation pathways L λ Wavelength l Sampe cell path length LED Light-emitting diode LOV Light, oxygen or voltage sensor M MA-MCP Methyl accepting chemotaxis protein N ν Frequency N Number of atoms within molecule ND Neutral density Nd:YAG Neodymium-doped Yttrium Aluminium Garnet NOPA Noncollinear optical parametric amplifier O ω Harmonic OD Optical density OPA Optical parametric amplifier OPO Optical parametric oscillator P P Phosphorescence P Propionate PAS Per, Arnt, Sim Pb Blue light absorbing form PCB Phycocyanobilin PcyA Phycocyanobilin/ferredoxin oxidoreductase Pfr Far-red light absorbing form Pg Green light absorbing form Pho Photolyase α/β domain Phy Phytochrome PHY Phytochrome specific domain PMT Photomultiplier Tube Pr Red light absorbing form

8 | P a g e

PVB Phycoviolobilin PYP Photoactive Yellow Protein (domain) Q R Rho Rhodopsin RR Response regulator RSB(H+) Schiff base (protonated) S S Singlet

S0 Singlet ground state

S1 Singlet first electronically excited state

S2 Singlet second electronically excited state

Sn+1 Singlet electronically excited states SADS Species associated difference spectra SDS-PAGE Sodium dodecyl sulphate polyacrylamide gel electrophoresis SE Stimulated emission S/T K Serine/Threonine kinase Ser/Thr kinase T t Time delay T Triplet T Transmittance TM Transmembrane Trp Tryptophan Tyr Tyrosine U UV Ultraviolet UV/Vis Ultraviolet/visible V v Vibrational levels associated with electronic energy levels W X X Any amino acid Y Z

9 | P a g e

List of figures

Chapter 1 Figure 1.1 Jablonski diagram of energy levels for an organic molecule in solution ...... 20 Figure 1.2 Potential energy surface of a cis/trans isomerisation ...... 22 Figure 1.3 Chromophore reaction mechanisms in photosensory receptors ...... 24 Figure 1.4 Photoreceptor domain structures ...... 26 Figure 1.5 Crystal structures of representative photosensory domains of major photoreceptor classes...... 28 Figure 1.6 Flavin based photocycles of photolyase and cryptochrome ...... 30 Figure 1.7 BLUF photocycle ...... 31 Figure 1.8 LOV domain photocycle ...... 32 Figure 1.9 Retinylidene photocycles ...... 35 Figure 1.10 PYP photocycle ...... 37 Figure 1.11 Structures of linear tetrapyrrole compounds ...... 38 Figure 1.12 Phytochrome photocycle ...... 39 Figure 1.13 Photocycle of the red/green CBCR AnPixJg2 ...... 43 Figure 1.14 C10 thioether formation in dual-Cys CBCRs ...... 44 Figure 1.15 Insert-Cys CBCR photocycle ...... 45 Figure 1.16 DXCF CBCR photocycle of Tlr0924g ...... 47 Figure 1.17 Biotechnological use of photoreceptor modification ...... 49

Chapter 2 Figure 2.1 Spectral regions of electromagnetic radiation ...... 52 Figure 2.2 Schematic absorption spectroscopy set-up ...... 53 Figure 2.3 Principle of stationary absorption spectroscopy ...... 54 Figure 2.4 Cary Varian UV/Vis spectrophotometer ...... 55 Figure 2.5 Cryochamber for low-temperature experiments...... 56 Figure 2.6 Vibrational modes ...... 57 Figure 2.7 Vibrational modes and their energies ...... 58 Figure 2.8 Schematic diagram of a Bruker vertex80 FTIR spectrophotometer ...... 58 Figure 2.9 Schematic diagram of a Helios transient absorbance spectrophotometer ...... 60 Figure 2.10 UV/Visible absorbance difference spectrum ...... 61 Figure 2.11 Schematic diagram of a time-resolved IR spectrophotometer ...... 62

10 | P a g e

Figure 2.12 Schematic of a laser flash photolysis set-up ...... 63

Chapter 3 Figure 3.1 PCB and PVB chromophores in Tlr0924 ...... 68 Figure 3.2. Plasmid maps of pBAD-myc/HisB/Tlr0924 and pCOLADuet-1/HO1,PcyA...... 69 Figure 3.3 Lamp emission and bandpass filter transmission spectra ...... 71 Figure 3.4 Expression of plasmids 1 and 2 in BL21(DE3) ...... 72 Figure 3.5 2-step purification of Tlr0924 ...... 73 Figure 3.6 Absorbance spectra of purified Tlr0924 ...... 74 Figure 3.7 Tlr0924 difference spectra ...... 75 Figure 3.8 FTIR difference spectra under different illumination conditions...... 76

Figure 3.9 Blockage of thioether formation between C10 and Cys499 by IAM and H2O2. .... 77 Figure 3.10 Pb state photoconversion properties ...... 79 Figure 3.11 Pg state photoconversion properties ...... 81 Figure 3.12 PCB Pr and PVB Pg state photoconversion properties ...... 82 Figure 3.13 Thermal stability of the Pb state...... 84 Figure 3.14 Thermal stability of Pg and Pg/r states ...... 85 Figure 3.15 Temperature effects on Tlr0924 photoconversion ...... 87 Figure 3.16 Effects of MeOH on Tlr0924 photoconversion ...... 88 Figure 3.17 Effects of ethylene glycol on Tlr0924 photoconversion ...... 89 Figure 3.18 Effects of glycerol on Tlr0924 photoconversion ...... 89 Figure 3.19 Effects of sucrose on Tlr0924 photoconversion ...... 90

Chapter 4 Figure 4.1. Structures and related absorption spectra of the 15Z-PVB'Pb, 15E-PVBPg, 15Z-PCBPb, and 15E-PCBPr chromophores of Tlr0924 ...... 96 Figure 4.2. Ultrafast transient absorption spectra collected after excitation at 435 nm at selected time points ...... 104 Figure 4.3. Ultrafast transient IR absorption spectra ...... 105 Figure 4.4. Laser flash photolysis spectra (ms) ...... 107 Figure 4.5. Laser flash photolysis spectra (μs) ...... 108 Figure 4.6. LED flash photolysis spectra at selected time points (PCB+PVB) ...... 109 Figure 4.7. LED flash photolysis spectra at selected time points for PCB only ...... 111 Figure 4.8. Low-temperature stabilisation of reaction intermediates...... 112 Figure 4.9. Scheme showing suggested ground and excited state energy surfaces ...... 114

11 | P a g e

Figure 4.10. Suggested forward reaction pathway and lifetimes for PVB and PCB in Tlr0924 ...... 116

Chapter 5 Figure 5.1 Structures and absorption spectra of the PCB and PVB photostates ...... 121 Figure 5.2 Ultrafast visible transient absorption spectra at selected time points ...... 126 Figure 5.3 Ultrafast IR transient absorption spectra at selected time points ...... 127 Figure 5.4 Global analysis of the ultrafast visible and IR transient absorption data ...... 128 Figure 5.5 Transient absorption data of PVB Tlr0924 (μs) ...... 130 Figure 5.6 Transient absorption spectra of PVB Tlr0924 (ms) ...... 132 Figure 5.7 Tlr0924 photoconversion at cryogenic temperatures ...... 133 Figure 5.8 Suggested reaction scheme for the full PVB conversion cycle ...... 135 Figure S5.1 Ultrafast transient absorption spectra ...... 137 Figure S5.2 Cryotrapping experiments...... 137 Figure S5.3 Ultrafast visible transient absorption spectra ...... 138 Figure S5.4 The residual matrix from the global analysis ...... 139 Figure S5.5 Sequential global analysis of the ultrafast visible transient absorption data ... 140 Figure S5.6 Laser flash photolysis data ...... 141 Figure S5.7 Temperature dependence of selected wavelengths...... 142 Figure S5.8. ns to μs photoconversion of PCB Pr to PCB Pb ...... 143 Figure S5.9. ms PCB Pr to Pb photoconversion and global analysis ...... 144

Chapter 6 Figure 6.1 Proposed photocycles of PCB and PVB bound full-length Tlr0924 ...... 148

List of tables

Table 1.1. Typical photophysical loss processes in solution ...... 21 Table 1.2. Known CBCRs and CBCR photoactive GAF domains ...... 41 Table 3.1. Relative photon absorbance ...... 71

12 | P a g e

Abstract

The photobiology of cyanobacteriochrome Tlr0924

A thesis submitted to The University of Manchester for the degree of Doctor of Philosophy (PhD) in the Faculty of Life Sciences by Anna Hauck, 2014.

Several different classes of photoreceptors have evolved to support function of organisms in all kingdoms of life. Cyanobacteriochromes (CBCRs) are photoreceptors unique to cyanobacteria and have been an area of great interest in recent years. Analogously to the phytochrome superfamily, they harbour a covalently ligated, linearised bilin chromophore, which undergoes light-driven E/Z isomerisation around the C15=C16 double bond to convert between a ground state and a photoactive state. CBCRs have developed unique tuning mechanisms for their spectral sensitivities and collectively cover the entire visible region. The CBCR Tlr0924 of Thermosynechoccocus elongatus belongs to a blue/green light sensitive subgroup with a conserved cysteine in a DXCF motif. This residue is involved in the transient formation of a second thioether linkage with the chromophore, which is vital to the formation of the blue-absorbing state and hence the functionality of the protein.

Full-length Tlr0924 protein has been subjected to detailed spectroscopic characterisation in this thesis in order to provide a comprehensive understanding of the timescales and mechanisms of the molecular processes that take place during the photocycle of this CBCR.

Initial stationary spectroscopic techniques confirmed the presence of two chromophore populations with blue ground state absorbance (Pb) and green- (Pg, phycoviolobilin) or red light (Pr, phycocyanobilin) sensitive photoproduct states. Photoproduct states also absorbed blue light according to the S0 to S2 transitions. Pb, Pg and Pr showed high thermal stability with only low dark reversion rates of Pg and Pr to Pb, and slow phycocyanobilin (PCB) to phycoviolobilin (PVB) isomerisation. The importance of the second thioether linkage was demonstrated by chemical modification of the cysteine, which prevented Pb state formation. Instead, illumination revealed a photoisomerised PVB reaction intermediate at 564 nm. The photoconversions were subsequently studied in greater detail by employing time-resolved absorbance spectroscopy techniques in the visible- and infrared spectral range and low-temperature trapping techniques. Pb and Pg states were demonstrated to photoisomerise to their respective blue- and red-shifted intermediates within picoseconds. In the infrared region of the spectrum this transition was accompanied by a shift of the D-ring C=O stretching mode, representative of the altered hydrogen bonding environment after the D ring flip. These isomerisations proceeded at cryogenic temperatures implying only local structural changes are involved. In the Pb to Pg/r activation reaction no further intermediates were resolved across the ps to s time scales. The isomerised chromophore was observed to lose the second thioether linkage with a life- time of 937 ms for PVB and 3.1 s for PCB. Characterisation of the reverse reaction mainly focussed on the PVB chromophore due to low signal-to-noise ratios in PCB data. The photoreaction concluded faster (23.6 ms) than the forward reaction despite proceeding via an additional red-shifted intermediate (5.3 μs), which corresponded to the species that was chemically trapped previously. Since photoproduct formation for both chromophores was only possible above 220 K, it could be concluded that larger protein domain motions were required in the process.

This work therefore shows unprecedented detail in the characterisation of a CBCR photoreceptor. It fills gaps especially regarding secondary reactions whilst complementing data on the truncated protein.

13 | P a g e

Declaration

No portion of the work referred to in the thesis has been submitted in support of an application for another degree or qualification of this or any other university or other institute of learning.

Copyright statement

i. The author of this thesis (including any appendices and/or schedules to this thesis) owns certain copyright or related rights in it (the “Copyright”) and s/he has given The University of Manchester certain rights to use such Copyright, including for administrative purposes. ii. Copies of this thesis, either in full or in extracts and whether in hard or electronic copy, may be made only in accordance with the Copyright, Designs and Patents Act 1988 (as amended) and regulations issued under it or, where appropriate, in accordance with licensing agreements which the University has from time to time. This page must form part of any such copies made. iii. The ownership of certain Copyright, patents, designs, trademarks and other intellectual property (the “Intellectual Property”) and any reproductions of copyright works in the thesis, for example graphs and tables (“Reproductions”), which may be described in this thesis, may not be owned by the author and may be owned by third parties. Such Intellectual Property and Reproductions cannot and must not be made available for use without the prior written permission of the owner(s) of the relevant Intellectual Property and/or Reproductions. iv. Further information on the conditions under which disclosure, publication and commercialisation of this thesis, the Copyright and any Intellectual Property and/or Reproductions described in it may take place is available in the University IP Policy (see http://www.campus.manchester.ac.uk/medialibrary/policies/intellectual- property.pdf), in any relevant Thesis restriction declarations deposited in the University Library, The University Library’s regulations (see http://www.manchester.ac.uk/library/aboutus/regulations) and in The University’s policy on presentation of Theses.

14 | P a g e

Acknowledgements

Many people have happily given up their own time to help me improve my knowledge and skills over the past four years and thereby ensured this PhD project would be successful. I would like to express my sincere gratitude to all of them. I could not have asked for a better place or group of people to work with.

In particular, I would like to thank my supervisor, Nigel Scrutton, for welcoming me to his lab. I was lucky to enjoy a good balance of guidance and ready support and trust and freedom to pursue ideas. He gave me access to any machine or facility I could have asked for and enabled me to meet the tetrapyrrole community and present my research internationally. I also appreciated Steve Rigby’s willingness to act as scientific advisor to this project and the generous financial support of TgK Scientific and the BBSRC.

The immediate Tlr0924-team also deserves special mention: Sam, Roger, Derren. Sam Hardman not only contributed greatly to the content of this thesis, but also ensured it would be written on time with many interesting ways of encouragement. It was a pleasure working with and next to you. I’m sorry for the nerves I’ve cost you. I owe much of my understanding of photochemistry and spectroscopy to Roger Kutta, who has given me many new perspectives on my work. Thank you for the patience with all my stupid questions and your helpfulness at all times including your weekends! Despite being everyone’s contact point for problems of all kinds, Derren Heyes always found time to be proactive about the project and helped me not to lose the bigger picture. Team – you’ve been lots of fun and your kindness won’t be forgotten.

Besides the academic support, I have formed many friendships along the way with lab members past and present, which has made this work so enjoyable. I would not want to miss the hours we spent outside the lab and I will miss you guys! (For completion: Hanno, James, Mary – you’re the best ;))

My final thanks go to my family for all the love and support to help me get to this stage.

15 | P a g e

Preface to the alternative format

This thesis is presented in the alternative format with permission from the University of Manchester. In contrast to the traditional format, results chapters are here included as manuscripts for publication in peer-reviewed journals. Some elements have been formatted to ensure consistency throughout this work.

Results chapters and contributions are as follows:

Chapter 3: Heterologous expression and spectral characterisation of the cyanobacteriochrome Tlr0924 from Thermosynechococcus elongtus Authors: Anna F.E. Hauck, Roger J. Kutta, Derren J. Heyes, Nigel S. Scrutton Published in: Manuscript in preparation Contributions: This work was carried out by AH with guidance from DH, RK and NS.

Chapter 4: The photoinitiated reaction pathway of full-length cyanobacteriochrome Tlr0924 monitored over 12 orders of magnitude Authors: Anna F.E. Hauck, Samantha J.O. Hardman, Roger J. Kutta, Gregory M. Greetham, Derren J. Heyes, Nigel S. Scrutton Published in: The Journal of Biological Chemistry (2014) Vol. 289, pp. 17747-17757 Contributions: Molecular biology, laser flash photolysis and low-temperature measurements were carried out by AH. Ultrafast visible measurements were carried out by AH with the help of SH and ultrafast infrared measurements with the help of GG. RK supported the work with an alternative LED flash photolysis set-up. AH and SH carried out data analysis. SH analysed the data globally. AH and SH prepared the manuscript.

Chapter 5: Comprehensive analysis of the green to blue photoconversion of full-length cyanobacteriochrome Tlr0924 Authors: Samantha J.O. Hardman,* Anna F.E. Hauck,* Ian P. Clark, Derren J. Heyes, Nigel S. Scrutton *These authors contributed equally Published in: Biophysical Journal (2014) Vol. 107, pp. 2195–2203 Contributions: AH prepared samples and carried out laser flash photolysis and low- temperature studies. SH recorded ultrafast visible measurements and IC the corresponding infrared measurements. SH carried out global analysis. AH and SH analysed the data and prepared the manuscript.

16 | P a g e

Chapter 1

Introduction

- 1. 1 Basis of light absorbance

- 1.2 Photoreceptor structure

- 1.3 Classes of photosensory receptors

- 1.3.1 Flavinoid photoreceptors - 1.3.1.1 Cryptochromes - 1.3.1.2 BLUF - 1.3.1.3 LOV

- 1.3.2 Isomerisation based photoreceptors - 1.3.2.1 Retinylidene protein - 1.3.2.2 Xanthopsins - 1.3.2.3 Phytochromes - 1.3.2.4 Cyanobacteriochromes - 1.4 Application - 1.5 Thesis aims and objectives

17 | P a g e

1. Introduction

Light is quintessential to the evolution and continued existence of life on earth. It originally protected early organisms from extinction through ultraviolet (UV) light damage by photochemically producing the ozone layer from molecular oxygen.1 Light is also the major source of energy for carbon fixation by plants and cyanobacteria in photosynthesis. They in turn feed higher organisms and convert carbon dioxide to oxygen – supporting all aerobic life forms and maintaining our ecosystem.1

In order to maximally adapt to the external light conditions (e.g. increase light harvesting,2 enforce diurnal rhythm,3 avoid photodamage,4 vision5) organisms – both photosynthetic and non-photosynthetic – have developed a large array of photosensory receptors. Through photon absorbance and subsequent photochemical processes, photosensory receptors activate signalling cascades and thereby modulate the biological activity of the entire organism.

The question of the mode of action of these photoreceptors has challenged scientists of all different branches for decades and questions still remain to be answered. For example, what structural changes does photon absorption induce in the light-sensing moiety? On what time scale, and how efficiently? How are these changes transduced to activate protein structure? What are the effector domains and interaction partners? What signalling cascade might be induced and how does this affect the behaviour of the organism?

This knowledge becomes a powerful tool with regards to e.g. agriculture.6 It has also been actively applied to spectrally tune photoreceptors and alter their kinetics, to design novel fluorophores,7 control cell behaviour and to build novel photoreceptors in the field of biotechnology.8

In biology, there are six main classes of photoreceptors: cryptochromes (Cry), light, oxygen or voltage sensors (LOV), blue-light sensors using flavin adenine dinucleotide (BLUF), rhodopsins (Rho), xanthopsins, and phytochromes (Phy).9 Within the Phy branch, a new subfamily called cyanobacteriochromes (CBCR) was discovered fairly recently.10

This thesis describes the spectroscopic study of one such CBCR photoreceptor, specifically the blue/green photosensory receptor Tlr0924 of Thermosynechococcus elongatus.

18 | P a g e

1.1 Basis of light absorbance

Sunlight is the vital source of energy to many organisms ranging from plants to bacteria. A broad range of light is captured in light harvesting complexes, where accessory pigments such as chlorophyll b, xanthophylls and carotenoids (plants) or phycocyanobilin, phycoerythrobilin, phycoviolobilin and phycourobilin (e.g. algae and cyanobacteria) absorb and transfer light energy to chlorophyll (a) in the reaction centre.11 Light energy is stored by the reduction of NADP+ to NADPH and generation of a proton gradient across the thykaloid membrane, which drives ATP synthesis. In the light-independent reaction, NADPH and ATP feed the Calvin cycle to convert carbon dioxide to glucose and oxygen.2

However, light is versatile: Besides its energetic importance, it also encodes information in the form of different wavelength composition, direction, irradiance and periodicity, which fluctuate during the diurnal cycle and throughout the seasons.12 Photosensory receptors pick up these features and convert them into biological signals to adapt the organismal development, morphology and metabolism to the environment accordingly. The first step in the study of a photoreceptor is to consider the process of light absorbance by the sensory moiety of the protein. Most photoreceptors harbour an organic molecule, called a chromophore, for this purpose. This area of photoreceptor study is generally the domain of the biophysicist/photochemist.

From the biophysics point of view, light is electromagnetic radiation with both wave and particle properties. It covers the near ultraviolet to near infrared regions (350 nm to 750 nm), which is also the span of sunlight covered by biological photoreceptors. Particles of light are called photons and their energy is reciprocally dependent on the wavelength by the equation:

E = hc/λ where E is the energy in Joules (J), h is the Planck constant (6.625 X 10-34J.s), c is the speed of light in vacuum (2.998 X 108 m.s-1) and λ is the wavelength of radiation in meters.13 To demonstrate the process of photon absorption and the subsequent photophysical processes, the chromophore or any light-absorbing moiety can be represented in the form of a Jablonski diagram (Figure 1.1).

19 | P a g e

S3 T3 IVR S Rotational states 2 v4 IC v3 T2 v2 Energy ISC S1 v1 T hν 1 v0 F P v4 v3 v2 v1 S v 0 0

Figure 1.1. Jablonski diagram of energy levels for an organic molecule in solution Excited state species undergo a variety of radiative (straight line) and non-radiative

(wavy line) processes to dissipate the energy acquired by photon absorption. Energy transfers such as vibrational relaxation, internal conversion (IC) and intersystem crossing (ISC) occur non-radiatively. Return to the ground state by fluorescence (F) and phosphorescence (P) on the other hand will emit a photon of lower energy. Singlet (S) and Triplet (T) electronic energy levels contain vibrational (v) energy levels.

Here, an organic molecule in the singlet ground state (S0) in solution is depicted, showing also the higher electronic energy levels (Sn+1) and their associated vibrational levels (v) as discrete quantised states. The productive photochemical reactions that result in receptor activation will be discussed subsequently. Starting in the electronic ground state, absorbance can occur when the incident photon energy matches the energy gap to an electronically excited state (indicated as yellow line). All higher electronically excited states subsequently relax rapidly into the lowest level vibrational level of the first excited state

(S1). This process includes inner vibrational relaxation (IVR) within the vibrational states of one electronic state where energy is dissipated to surrounding molecules, and internal conversion (IC) where the system changes isoenergetically from the lowest vibrational state of a higher electronically excited state to an energetically equivalent higher vibrational state of the lower lying electronic state. In the relaxed lowest vibrational states of the S1 state the system has a longer lifetime due to the large energy splitting to the ground state.

20 | P a g e

As a consequence, two further processes become possible. The system can relax to S0 radiatively by release of a photon (fluorescence, F) or undergo inter system crossing (ISC), where the system isoenergetically changes its spin system from singlet (S) to triplet (T). In the triplet spin system the excited state can relax further into its lowest vibrational states via IVR. Since the ground state is a singlet state a direct conversion from the triplet system into the singlet ground state is spin forbidden. As a consequence the lifetime for the conversion of the triplet states back into the singlet ground state is much longer than for a

S1→S0 transition (μs to s). S1→T0 transitions can occur either by a release of a photon (phosphorescence, P) or by ISC and subsequent IVR.14

The loss processes described above occur on very fast time scales (Table 1.1) and compete with productive chemical reactions that will result in activation of photosensory receptors.

Process Time scale (s)

Photon absorbance sub-fs Internal vibrational relaxation ps

Internal conversion ns - μs

Fluorescence fs - ns Intersystem crossing ps - ns

Phosphorescence μs - 100s

Table 1.1. Typical photophysical loss processes in solution

Photoreceptors consequently have to undergo efficient primary photoreactions to minimise energy loss through de-excitation pathways and to maximise their signalling capacity. This relationship is commonly described by the quantum yield (Ф):

Ф = number of molecules in state of interest/number of absorbed photons.

Indeed, many important photobiological processes proceed on the fs to ps timescale (e.g. photosynthesis, vision).15 Productive primary photochemical reactions in photosensory receptors are efficient and perform specific elementary chemical reactions. They commonly include intersystem crossing (LOV16), electron transfer (BLUF17 and Cry18) and C=C isomerisation (xanthopsins,19 Phy/CBCR,20–23 Rho24). These types of reaction become feasible through the altered electronic arrangement in excited state molecules. The energy

21 | P a g e gained during photon absorption (40-60 kcal.mol-1 from a UV/Visible photon) is a means of overcoming activation barriers and can be sufficient to form/break bonds (LOV25 and some CBCRs26). During photoisomerisation of a C=C group for example, an electron is promoted from a bonding π-orbital to a π* anti-bonding orbital. This transition abolishes the fixed planar geometry of the C=C double bond and in the excited state a 90° rotation about this bond becomes the energetically favourable conformation. Relaxation to the ground state then forces the molecule back into a planar configuration which can be either in the cis or trans configuration.27 The simplified scheme for this reaction is shown in Figure 1.2 as a potential energy surface diagram.

Figure 1.2. Potential energy surface of a cis/trans isomerisation The potential energy surface describes the relationship between the energy of a molecule and its geometry. Energy minima represent stable chemical species and therefore the probable distribution of molecules. In this example the energetic barrier is overcome by relaxation from an electronically excited state.

The type of primary photochemical reaction undergone and the protein’s wavelength sensitivity is dictated by the structure of the chromophore. Chromophores of sensory photoreceptors are usually comparatively small organic molecules that can be associated with the protein covalently or non-covalently as prosthetic groups and serve as units of light absorbance. Their structure and photoactivity are variable. The only prerequisite of

22 | P a g e their functionality is a partly unsaturated chemical structure, enabling electron delocalisation across a conjugated π system. The more extended this conjugated system is, the smaller is the energy gap between S0 and S1, and the longer wavelength photons will induce the transition.28 Light-harvesting chromophores can often also chelate a metal ion. Absorbance of specific wavelengths and not others gives rise to characteristic spectral properties. As the structure changes during photoconversion reactions, so do the absorbance properties of the new species. These can be studied by a variety of spectroscopic techniques with regards to the identity and time constants of the reaction (see chapter 2). The portion of white light that is not absorbed is reflected or transmitted and determines the colour of the holoprotein. Figure 1.3 summarises the most commonly employed sensory chromophores and their mode of action. The main categories of chromophores in sensory photoreceptors are “aromatics” (A-C), which are blue-light absorbing flavin-derivatives, “polyenes” (D-E) such as rhodopsin and coumaric acid and “tetrapyrroles” (F), e.g. biliverdin and its derivatives.9 Tetrapyrrole ring structures with central metal ions form a more recently discovered subclass in sensory photoreceptors.29 Tetrapyrroles are the most varied in their absorbance properties and depending on the structure and mechanism can cover the entire visible region.

Protein aromatic amino acid side chains and polypeptide backbones also form conjugated π systems but absorb further into the UV. One exceptional photoreceptor, UVR8, takes advantage of this and absorbs 280-315 nm light without employing an external chromophore.30

The unique chromophore and photochemistry also form the basis for the classification of different photoreceptors: Light, oxygen or voltage sensors, blue-light sensors using flavinadenine dinucelotide (FAD), cryptochromes, rhodopsins, xanthopsins, phytochromes and cyanobacteriochromes. These will be described in more detail in the following subsections.

23 | P a g e

Figure 1.3. Chromophore reaction mechanisms in photosensory receptors Photoreactions of cryptochrome, BLUF and LOV flavin derivatives A-C include electron transfer mechanisms (Box1) and covalent bond formation/rupture (Box2). The rhodopsin and xanthopsin polyenes D-E and phytochrome and CBCR tetrapyrroles F undergo E/Z isomerisation (Box3). Adapted from 32

24 | P a g e

1.2 Photoreceptor structure

The light sensing moiety described above is rarely alone sufficient for photoreceptor function and a second signalling moiety generally forms part of the same protein. As the photocycle proceeds, the chromophore structural changes are transduced to the bulk protein structure to induce first localised changes within the chromophore binding pocket but subsequently also more long-range conformational transitions.31 These latter structural changes are light-independent and result in the formation of the meta-stable signalling states.32 In order to fully appreciate these motions it is important to have a structural basis for any predictions. This aspect of photoreceptor study often engages molecular and structural biologists. Structural information can be gathered by nuclear magnetic resonance, electron microscopy and X-ray crystallography. To then observe signal transduction between the ground and photoproduct states, more elaborate time-resolved or low-temperature techniques are required33–35 and the process itself is not very well understood to date.

Genome searches give a first indication of protein structure and function by assigning specific regions to certain domains with predicted fold and functionality. The N-terminal domain is usually the sensory domain binding the chromophore. Some photoreceptors harbour more than one chromophore (e.g. neochrome, phototropins) or additional ligand binding sites, conferring the potential to integrate more than one sensory input.32 The output domain will relay the information to downstream signalling partners through catalytic activity. Additional functional domains may e.g. act as sites of dimerisation or binding sites. Figure 1.4 shows the domain architecture for examples of the different photoreceptor classes.

25 | P a g e

Cryptochrome1 Pho fad CCT λ = blue Cryptochrome3 Pho fad

BLUF Tll0078 BLUF BLUF BlrP1 BLUF EAL λ = blue BLUF PACα BLUF A/G cyc BLUF A/G cyc

LOV Phototropin 1 LOV LOV S/T K λ = blue

LOV Neochrome 1 PAS GAF PHY LOV LOV S/T K λ = red/far-red, blue LOV f Kelch LOV ZEITLUPE λ = blue LOV Aureochrome bZ LOV λ = blue

Rhodopsin Rhodopsin λ = green

PYP PYP λ = blue

Ppr PYP PAS GAF PHY HisK λ = blue, red

C Plant Phytochrome PAS GAF PHY PAS PAS HisK λ = red/far-red C Cph1 PAS GAF PHY HisK λ = red/far-red C Bacterial BphP PAS GAF PHY HisK λ = red/far-red C Prokaryotic λ = red/far-red PAS GAF PHY HisK RR Fph/Dph C C λ = red/far-red, Cph2 GAF PHY GGDEF EAL GAF GGDEF blue/green C R/G CBCR SyCcaS GAF PAS PAS HisK λ = red/green C DXCF CBCR Tlr0924 CBS CBS CBS CBS GAF GGDEF λ = blue/green C DXCF CBCR TePixJ HAMP GAF HAMP MA-MCP λ = blue/green C Phy/Dual-Cys CBCR NpF1183 PAS GAF PHY HisK λ = violet/orange/red C C C C C C λ = UV/blue Dual-Cys CBCR NpF2164 GAF GAF GAF GAF GAF GAF GAF MA-MCP violet/orange C C C Dual-Cys CBCR NpR1597 GAF GAF GAF GAF HisK λ = blue/teal UV/blue

Figure 1.4 Photoreceptor domain structures Photoreceptor domain architecture is often very variable and provides a means of

combining different input and output domains to match a function to a certain spectral sensitivity. Since membranes are light permeable, most photoreceptors are soluble

cytoplasmic proteins. Transmembrane domains are represented as black bars.

Pho (Photolyase α/β domain), fad (Photolyase α domain), CCT (cryptochrome C-terminal domain), BLUF (blue-light sensor using FAD), EAL (diguanylate phosphodiesterase), A/G cyc (adenylate/guanylate cyclase), LOV (light-oxygen-voltage), S/T K (Serine/Threonine kinase), PAS (Per-Arnt-Sim), GAF (cGMP-specific

phosphodiesterases, adenylate cyclases and FhlA), PHY (phytochrome specific), f (F box), Kelch (Kelch repeat), bZ (basic zipper), PYP (photoactive yellow protein), HisK (Histidine kinase (related) domain), RR (response regulator), GGDEF (diguanylate cyclase), CBS (cystathionine-β-synthase), HAMP (Histidine kinases, adenylate cyclases, methyl accepting proteins and phosphatases), MA-MCP (methyl accepting chemotaxis 10,26,32,147,163 protein) Adapted from

26 | P a g e

Many of these domains are conserved in their structure and function during evolution across different organisms. Functional differentiation in proteins is achieved by “mixing and matching” of these different domains, so that the same input domain may have different output domains or a specific output domain may be activated by different stimuli. As these domains will fold independently (e.g. when genetically excised from the photoreceptor protein) and remain functionally viable in isolation, they are often characterised individually and studies can be related to the full-length protein and among different organisms.32 The domains are common to all signalling proteins and are equally found in ligand-activated chemoreceptors. The Per-Arnt-Sim (PAS) domain as found in the xanthopsin and phytochrome family for example, is a ubiquitous sensory transduction domain involved in ligand binding and protein interactions.36,37 The chromophore binding cGMP-specific phosphodiesterases, adenylate cyclases and FhlA (GAF) domain of mainly Phy and CBCR photoreceptors belongs to a large and functionally divergent superfamily.38,39 The acronyms of both these domains represent different protein classes in which they occur. Rhodopsin is an obvious exception to the domain architecture. Unlike other photoreceptors, it forms an integral membrane protein and has evolved from light harvesting proteins, which often generate energy by establishing an ion gradient across the membrane.32 Cryptochromes are also known to only act as one functional domain.

The single domain approach therefore has limitations, as it also disregards the covalent linkers between the individual domains and extensions, which are of functional importance in signal transduction.32 Furthermore, inter- and intra-protein interactions are largely lost. In order to gain a comprehensive overview of photoreceptor function, it is therefore essential to study the full-length protein.

The full structural picture is best described by crystal structures (Figure 1.5). The definite positioning of amino acids in the 3D fold enables predictions of the mode of action in the active site and rational design of site-directed mutagenesis to confirm them. On the downside, study of a tightly-packed crystal lattice at cryogenic temperatures does not permit the large-scale protein motion required for activity in signalling proteins.

27 | P a g e

A B

C D E

Cytoplasm

F G

Figure 1.5 Crystal structures of representative photosensory domains of major photoreceptor classes. Chromophores (green) are shown in their respective binding domains (blue). A Cryptochrome (1U3D) B BLUF (3GFZ) C LOV (2V0U) D Rhodopsin (1C3W) E Xanthopsin (2PHY) F Phytochrome (3C2W) G UV resistance locus 8 (4DNW). Additional domains are yellow or purple and the second subunit of homodimers is grey. Adapted from 32 by M.Ortmayer

28 | P a g e

1.3 Classes of photoreceptors

Photosensory receptors are found in all kingdoms of life from unicellular organisms right through to animals.40 As mentioned above, photoreceptor classification is based on the structure of the chromophore and the photochemistry of the primary photoreaction. Whereas historically they were classed as blue- or red-light sensitive,41 the plethora of photoreceptors known today highlights the importance of light-sensing of the full UV/Visible (UV/Vis) spectrum. The main classes will be described in some more detail in this section.

1.3.1 Flavinoid photoreceptors

In flavionoid photoreceptors, the light-sensing photochemistry is based around an isoalloxazine ring structure. Occurring in microorganisms, plants and animals they mediate a number of blue-light dependent adaptational functions (e.g. direction of growth, circadian rhythm42) by a light-induced electron and proton transfer mechanism.43 Representatives of this group are the cryptochromes (Crys) and photolyases,44 sensor of blue light using FAD domain (BLUF)45 and light, oxygen or voltage (LOV) domain proteins.46

It is noteworthy, that flavins are also commonly employed as redox cofactors with or without photosensory function.47

1.3.1.1 Cryptochromes

Cryptochromes (Cry) are regulators of plant growth and development,48 and the circadian rhythm.49 They have also been proposed as the magnetosensor in migratory birds.50 Three different types include the plant and animal signalling cryptochromes as well as the DNA- repairing Cry-DASH (found in Drosophila, Arabidopsis, Synechocystis and Homo)51 and photolyases. The shared features that led to the discovery of the first Cry52 include the N- terminal photoactive domain comprising two non-covalently bound chromophores (Figure 1.5A). In photoactivation, a (usually) pterin antenna cofactor transfers the absorbed excitation energy to a flavin adenine nucleotide (FAD) via Förster resonance transfer.53 The redox state of FAD depends on the protein class and species. In photolyases (Figure 1.6A), in vitro photoactivation generates fully reduced, enzymatically active FADH- from the semi- reduced neutral radical FADH•. Photoactivated FADH•* extracts an electron via a tryptophan (Trp) triad, three Trp residues bridging the gap between the cofactor and the

29 | P a g e protein surface, in ~30 ps. The final Trp•+ radical releases a proton prior to reduction by an extrinsic reductant.54 The fully reduced flavin is enzymatically active and transfers an electron directly to a pyrimidine dimer in the excited state to generate 2 monomeric pyrimidines. FADH- formation occurs with a quantum yield of 0.2 owing to the reversibility of the electron transport chain along the Trp triad and reduction of Trp•+ by FADH-.54 Following the FADH- to FADH• oxidation, the fully reduced flavin is restored by electron back transfer.55

Cryptochrome photoactivation (Figure 1.6B) proceeds analogously to the photolyase photoactivation mechanism via the Trp triad based photoreduction of flavin. The ground state flavin is however the oxidised FAD (FADox) cofactor proceeding to a semi-reduced •- • flavin (FAD in insects, FADH in A. thalania). The electron transfer to FADox* occurs in less than a picosecond and the slower steps are completed on a sub-nanosecond timescale. The flavin radical is readily reoxidised in aerobic environments.55 The magnetoreception is hypothesised to proceed through the radical pair formation between FAD and Trp in the avian retina.56 Since triplet states become energetically separated with increasing magnetic fields, spin state mixing is reduced and enhanced Cry activation can be observed.57,58 The

59 cryptochrome C-terminal extension is the putative interaction site for signalling partners.

Figure 1.6 Flavin based photocycles of A photolyase and B cryptochrome Both photocycles are thought to involve a light triggered electron transfer via a Trp triad to photoreduce the flavin chromophore. Reoxidation from this presumed signalling state concludes the photocycle. Adapted from 55 30 | P a g e

1.3.1.2 Blue light sensors using flavin adenine dinucleotide

Blue light sensors using flavin adenine dinucleotide (BLUF) domain proteins are most common to proteobacteria and cyanobacteria and mediate photophobic responses and gene regulation60 through a host of output domains (e.g. adenylate cyclases, photodiesterases) or interaction partners.61 BLUF domains bind a FAD cofactor non- covalently – held in place by a hydrogen-bonding network.62 During primary photochemical events (Figure 1.7), the photoexcited FAD cofactor abstracts an electron from a conserved tyrosine (Tyr) residue to form a short-lived radical pair63 and causing a rotation of a proximal glutamine side chain.62 Electron back transfer to Tyr results in the formation of a long-lived signalling state (Pred) with a 10 nm red-shift compared to the ground state. The chromophore is structurally identical to the ground state configuration but differs in the hydrogen bonding network with the protein (Figure 1.3B). The signalling state reverts back to the ground state thermally within seconds as the glutamine assumes its original orientation.63 An alternative mechanism has been proposed involving the tautomerisation of a glutamine side chain.64

P hv P* e- s ps

• ps FADH

Pred - e

Figure 1.7 BLUF photocycle Photoexcited FAD transiently undergoes radical pair formation with a conserved Tyr

residue before formation of the Pred signalling state. Pred thermally reverts back to the ground state (P).

Since FAD structural changes are minor, the photoreactions of BLUF domains are fast. Like cryptochromes, the signalling state is formed sub-nanosecond.

31 | P a g e

1.3.1.3 Light, oxygen or voltage sensors

Light, oxygen or voltage sensing (LOV) domains are the blue light photosensors of phototropins and related proteins (e.g. ZEITLUPE, aureochromes) in prokaryotes and eukaryotes.46 They mediate a variety of adaptational processes including phototropism,65 stomatal opening66 and chloroplast movement67,68 to improve photosynthetic efficiency. LOV domains belong to the subgroup of PAS domains36 and non-covalently bind an oxidised flavin mononucleotide (FMN) cofactor though hydrogen bonding.59 In phototropins LOV domains occur in tandem in the N-terminus, with the second LOV domain (LOV2) mediating the light-induced Serine/Threonine (Ser/Thr) kinase activity of the C-terminus. LOV1 attenuates LOV2 and assumes a role in receptor dimerisation.69 During the photocycle (Figure 1.8) FMN undergoes intersystem crossing from the excited singlet state to the excited triplet state, excited state proton transfer and finally covalent bond formation between the isoalloxazine ring C4 and a conserved Cysteine (Cys) residue on the

70 microsecond timescale (LOV390). In this signalling state the FMN planarity is disrupted – a structural change that is transmitted to the bulk protein structure and results in the disruption of a strongly conserved salt bridge. Thiol adduct formation usually is a reversible process but ground state recovery is slow (10-1 – 10-4 s-1).71

LOV H+ 447 hv s-min

S S* S-LOV 390 LOV

μs ns LOVT* H+ 660

Figure 1.8 LOV domain photocycle LOV domains undergo a unique photocycle that involves intersystem crossing from the excited singlet (S) state to the excited triplet (T) state and formation of a thioether linkage (S-). Adapted from 305

32 | P a g e

Signal transduction in LOV domain proteins mainly involves changes within the chromophore vicinity, which ultimately lead to the unfolding of a C-terminal α-helix extension. This movement is thought to regulate phototropin Ser/Thr kinase activity.72,73

Single LOV domain proteins are most commonly encountered in prokaryotes and fungi. The C-terminal effector domain can vary e.g. kinases,74,75 transcriptional regulators76,77 and phosphodiesterases.78

1.3.2 Isomerisation based photoreceptors

In isomerisation-based photoreceptors, the chromophores perform cis/trans or trans/cis isomerisation as initial light response. Members of this group are the green light absorbing retinylidene proteins,79 the blue light absorbing xanthopsins80 and the red light absorbing phytochromes.81 The phytochrome-related cyanobacteriochromes are a unique example, where many different types of the protein collectively absorb in the entire UV/Vis range.10

1.3.2.1 Retinylidene proteins

Retinylidene proteins are a class of proteins employing a retinal cofactor to confer light sensitivity.79 They are found throughout the microbial and animal kingdom. Structurally, they are part of the superfamily of 7 transmembrane (TM) receptors82 but can be divided into 2 distinct groups based on the protein sequence and retinal isomer employed, which also strongly reflects in their functionality. The retinal chromophore is attached to a lysine residue in the C-terminal TM α-helix via a functionally important Schiff base (RSBH+).83

Type 1 rhodopsins encompass the microbial rhodopsins, which generally undergo all-trans to 13-cis photoisomerisations. The output is variable and includes both energetic and signalling functions. Depending on several key amino acids within the chromophore binding pocket, microbial rhodopsins can act as proton (e.g. bacteriorhodopsin (BR)),84 chloride (e.g. halorhodopsin)85 or sodium (NQ rhodopsins) pumps,86 but also mediate phototactic or photophobic responses e.g. sensory rhodopsins,87,88 channelrhodopsins.89

Type 2 rhodopsins comprise the retinylidene photoreceptors of the animal kingdom, also referred to as opsins. Members of this family form a subgroup of the G-protein coupled receptor family and therefore mediate their signalling function through G-proteins.90 In

33 | P a g e contrast to microbial rhodopsins, the retinal photoisomerisation proceeds from 11-cis to all-trans. Besides the well-known scotopic visual receptor rhodopsin and its relatives, other less well-understood proteins also fall in this category but will not be discussed any further.91–94

For both classes of rhodopsin the primary photoreaction occurs on the femtosecond timescale and is one of the fastest and most efficient photobiological reactions known (200- 500 fs).24 The energetic barrier of retinal isomerisation is too high to be overcome in the ground state. The isomerisation has been proposed to occur to 90° on the excited state surface before relaxing back to either the photoproduct or the parent product at the conical intersection.95 For the slower BR reaction, a 3 state model with a low barrier prior to the energetic minimum was proposed.95 The chromophore assumes a strained conformation which subsequently relaxes via several intermediates, thereby driving changes within the protein conformation to accommodate the isomerised cofactor.96 The later stages of the photocycle deviate with microbial rhodopsins, which undergo a full photocycle including the thermal reisomerisation of retinal, whereas animal rhodopsins undergo a chromophore detachment reaction to recycle retinal back to the ground state.

The best studied protein of the microbial rhodopsin class is the bacteriorhodopsin proton pump (Figure 1.9A).97 The initial ultrafast photoisomerisation (J intermediate) is followed by several reaction steps with intermediates K, L, M, N and O.98 After the formation of the first stable photoproduct, K,99 all subsequent intermediates are formed in dark reactions through gradual energy dissipation of the twisted 13-cis retinal. The blue shifted L intermediate forms the precursor for the proton transfer from the RSBH+ to a primary carboxylic acceptor.100 The deprotonation and reprotonation of the Schiff base in intermediates M and N are vital steps in the BR function of pumping protons across the membrane from the cytosol to the extracellular environment.101 The M intermediate is particularly distinguishable due to a large blue shift upon deprotonation of the Schiff base nitrogen. The N intermediate brings about the largest protein structural change in the form of a tilting of the F helix.102 The O intermediate is red shifted and is the last step prior to resetting the ground state. The protein gradient generated is utilised in ATP synthesis. BR serves as model system for animal rhodopsin photoreceptors (Figure 1.9A).

Following the 11-cis to all-trans isomerisation of the retinal cofactor in type 2 rhodopsins, the primary, red-shifted photoproduct bathorhodopsin is formed from the rapid relaxation of photorhodopsin as mentioned above.103 Initial local side chain motions in the retinal

34 | P a g e binding pocket result in the formation of lumirhodopsin.104 The protein motions extend to more distant parts of the protein on the microsecond timescale upon formation of the metastates,105 which also involve deprotonation of the Schiff base (metarhodopsin II). The conformational changes permit interaction with the G-protein which activates a cyclic guanosine monophosphate messenger cascade to transduce the light signal.106 Metarhodopsin is therefore the active signalling state and characterised by a 100 nm blue shift caused by the Schiff base deprotonation. In order to regenerate the photoreceptor ground state, the retinal pigment dissociates from the opsin protein moiety by RSB hydrolysis. This can either proceed from metarhodopsin II or from metarhodopsin I via metarhodopsin III omitting the signalling state.107 Recombination of the opsin with 11-cis- retinal completes the photocycle (Figure 1.9B).

Figure 1.9 Retinylidene photocycles Bacteriorhodopsin (BR) and bovine rhodopsin undergo a semi-conserved photocycle, whereby the initial fs photoisomerisation results in a strained chromophore structure, and first drives local protein changes before larger helix motions occur on longer timescales. The retinal anchoring Schiff base has a crucial functional role in deprotonation and reprotonation steps to translocate protons in bacteriorhodopsin, and also undergoes deprotonation to form the active signalling state in the GPCR type rhodopsin. These initial intermediates can be considered equivalent. BR retinal can then thermally revert back to the ground state, whereas the Schiff base in rhodopsin is hydrolysed for external reisomerisation of the chromophore. The BR photocycle is hence completed a lot faster than the rhodopsin photocycle. Adapted from 296

35 | P a g e

1.3.2.2 Xanthopsins

Xanthopsins are blue light photoreceptors occurring in purple bacteria.108 The spectral sensitivity is mediated by a p-coumaric acid chromophore anchored to the protein via a thioester linkage109,110 and buried deeply in the hydrophobic protein core.

Xanthopsins can be divided into 3 subgroups based on sequence homology.111 Photoactive yellow protein (PYP)112 is the best characterised example. The protein is surprisingly small for a signal transducer, consisting of one PAS domain with an N-terminal extension only.113 This N-terminal domain is responsible for signal transduction, ultimately initiating a negative phototaxis movement to protect the organism from UV damage.114 Both other subgroups contain an additional bacteriophytochrome domain.

Signal transduction depends on the external light conditions as detected through the chromophore and mediated by activation and deactivation of the photocycle. In the pG ground state, the chromophore assumes a depronated trans-p-coumaric acid conformation.110 Upon blue light absorption the photocycle is initiated on a timescale of picoseconds.115,116 The chromophore photoisomerises within 3 ns to form a still deprotonated cis-p-coumaric acid pR state, which is red-shifted relative to the ground state.117 Transient absorbance measurements predict this process to proceed via at least one intermediate (I0) with a twisted cis-configuration, having just passed the trans/cis barrier, and some potential fluorescent states.118 Low-temperature studies for the same transition display features which may be attributable to artefacts of the technique.118 Slight wavelength shifts of the excitation light at cryogenic temperatures reversibly activate two different branches, passing either via hypsochromic or bathochromic intermediates to form the pR state.119 There are discrepancies in nomenclature between different studies, but the

118 first bathochromically shifted intermediate (PYPB) has been predicted to correspond to I0. This first isomerisation reaction from pG to pR shows very little structural change as the aromatic ring stays in position while the thiol-ester carbonyl rotates.120 pR1 and pR2 represent intermediates with the same absorbance features but localised and downstream changes of the protein.121 The chromophore isomerisation facilitates a proton transfer between Glutamate46 and the chromophore to form the next intermediate pB’ in a reversible manner.122 Whereas the initial isomerisation caused a red shift of the absorbance maximum, protonation causes a blue shift of the pB’ intermediate. During this step, the negative charge that was previously delocalised across the chromophore is passed onto a single residue generating free-energy stress within the protein.118 The former drives the

36 | P a g e subsequent large scale structural changes to the pB signalling state.123 The structural change in pB is a partial unfolding involving also the N-terminal signalling extension.124 The exact downstream signalling cascade remains unknown.

The signalling state is thermally unstable and will revert back to the ground state both in the presence and absence of a light stimulus.125 Both reaction pathways are much simpler than the activation process proceeding via only one apparent intermediate each. In the absence of light, the chromophore is deprotonated first (pBdeprot) and thereby induces localised changes that permit the reisomerisation and protein refolding required to form the pG ground state. Upon light activation, the ground state recovery first proceeds via cis/trans isomerisation (pBt) and is 3 orders of magnitude faster.125 A schematic of the photocycle can be found in Figure 1.10.

Figure 1.10 PYP photocycle as suggested by room temperature (yellow) and cryogenic (blue) studies

The p-couramic acid chromophore undergoes trans/cis isomerisation (pR1/PYPL) upon light activation. The structural change then facilitates the protonation of the

chromophore, introducing a negative charge at Glutamate46 (pB’). The main structural change within the protein occurs during neutralisation of this charge

(pB). Adapted from 118

37 | P a g e

1.3.2.3 Phytochromes

Phytochromes (Phys) have been extensively studied for several decades and are the best characterised photoreceptors behind the rhodopsins discussed earlier but full understanding was long compromised by the lack of crystal structures. Phys were discovered126 as the photoreceptors of plants mediating growth and development in a red- light-dependent manner.81,127,128 The term “phytochromes” today encompasses, besides the classical plant photoreceptors, phytochromes from a variety of eukaryotic organisms (e.g. fungi129–132, algae133–135) and bacteria136 (e.g. proteobacteria, cyanobacteria, deinococci, actinobacteria) irrespective of their ability to photosynthesise or not. 41,81,137–139 The linear tetrapyrrole chromophore conferring the photoactivity varies in its identity depending on the species. Bacterial phytochromes (Bphs) covalently bind biliverdin,136,140 which is a precursor to phycocanobilin (PCB) used in cyanobacterial phytochrome 1 (Cph1)141 and phytochromobilin (PФB) of plant phytochromes (Figure 1.11).

Figure 1.11. Structures of linear tetrapyrrole compounds Bilin chromophores are produced by the oxidative cleavage of heme. The cleavage product biliverdin is incorporated in bacterial and fungal phytochromes. Cyanobacterial and higher plant phytochromes use phycocyanobilin and phytochromobilin, respectively, which are produced by enzymatic reduction of biliverdin. Phytochromes can be reconstituted with non-physiological bilins in vitro through their auto-lyase activity. Adapted from 245

HO1 (Heme oxygenase 1), PcyA (phycocyanobilin/ferredoxin oxidoreductase), HY2 (PФB synthase)

38 | P a g e

The above chromophores photoconvert reversibly between two stable states in phytochromes - a red-absorbing state (Pr) and one absorbing near IR radiation (Pfr) (Figure 1.12). This dual light-sensitivity switching is a unique trait amongst photoreceptors.142 Pr is the dark adapted, typically physiologically inactive form.81 Photon absorption of ~660 nm causes Z to E isomerisation of the chromophore about the C15=C16 double bond143 to form the primary photoproduct (Lumi-R). The isomerisation process occurs on the picoseconds timescale with a quantum yield of ~0.15.144 The Lumi-R state decays thermally to the (light activated) Pfr state via several intermediates on the microsecond and millisecond timescale. This decay also involves a rotation about the C14-C15 bond and transient deprotonation. The deactivation reaction is less well understood because it is difficult to generate pure Pfr sample due to spectral overlap. Pfr reisomerisation to the chromophore Z conformer is initiated by absorption of a far-red photon. The quantum efficiency is even lower for this process (~0.06) than for the Z to E isomerisation and was proposed to also involve configurational and conformational isomerisation (15E,anti→15Z,syn).145 It is accepted that forward and reverse reactions proceed through different intermediates with different lifetimes. Pfr can also revert back to Pr thermally over hours in a process known as dark reversion.146

Figure 1.12. Phytochrome photocycle

The Pr dark state is triggered by red light to convert to the Pfr light-activated state, which in turn can be converted back by far-red light. Both processes are initiated by a

photoisomerisation reaction and proceed via several thermal intermediates on the electronic ground state surface. Pfr is meta-stable and will also thermally revert back

to Pr. Life times are shown in grey and thermal barriers in blue. The nomenclature is based on the rhodopsin system. Adapted from 145

39 | P a g e

Chromophore isomerisation brings about conformational changes within the protein structure to an energetically more favourable structure. On the domain level the photoreaction relies on an N-terminal photosensory tridomain typically composed of a PAS, GAF and phytochrome specific (PHY) domain.147 The chromophore in the GAF domain pocket assumes a C5-Z,syn, C10-Z,syn, C15-Z,anti configuration and is covered by a tongue feature protruding from the PHY domain. The PAS and GAF domains are connected by a unique and functionally significant figure of eight knot.148 Signal transduction is thought to occur along an extended α-helix to the C-terminal output domain, which usually has kinase activity: Histidine (His) kinase in Bphs and cyanbacterial phys149–152 and putative Ser/Thr kinase in plant phys.153 Phytochromes can undergo autophosphorylation or interact with further signalling partners. There are exceptions, e.g. BphGs (RSP4191 and RSP4111) contain a GGDEF-EAL module.154

1.3.2.4 Cyanobacteriochromes

Besides harbouring phytochromes, cyanobacteria were also found to encode more distantly phytochrome-related proteins. Genome sequencing studies of prokaryotic cyanobacteria155,156 have been a particularly active field of research in recent years and unearthed vast numbers of these so-called cyanobacteriochromes (CBCRs) (Table 1.2).147,157,158 CBCRs have been defined as photosensory proteins characterised by chromophore-binding GAF domain(s) homologous but distinct from the tetrapyrrole binding GAF domain of phytochrome.10 In contrast to Phys, the isolated GAF domain is sufficient for photoconversion159–162 in CBCRs and the spectral properties can range from the near UV to the red region 10,138,162–164 depending the bilin-protein interactions within the protein GAF domain. Multiple GAF domains can be contained within the same protein or in combination with phytochrome photosensory cores, theoretically providing a mechanism to integrate several signals. The N. punctiforme genome harbours the most Phys and CBCRs known to date, including the most complex CBCR, containing 6 CBCR GAF domains.163

40 | P a g e

Name Organism Function Photochromicity Chromophore Ref (Z/E)

AnPixJ Nostoc sp. PCC7120 Phototaxis red/green PCB 165 648/543

NpR6012g4 N. punctiforme unknown red/green PCB 166,167 652/538

NpF2164g6 N. punctiforme unknown red/green PCB 168 648/546

slr1393 Nostoc sp. PCC7120 unknown red/green PCB 169 650/535

CcaS Synechocystis sp. green/red PCB 161 PCC 6803 Chromatic 535/672 acclimation N. punctiforme 536/672 PCB 170 ATCC29133

171– RcaE F. displosiphon Chromatic green/red PCB acclimation 532/661 173

PixJ T. elongatus Bp-1 blue/green 174– Phototaxis 433/531 PCBPVB 176

Synechocystis 430/535 159,160 ,177,17 179,180 Tlr0924 T. elongatus unknown blue/green PCBPVB 8 436/532

Tlr1999 T. elongatus unknown blue/teal PCBPVB 164 418/498

UriS Synechocystis sp. Phototaxis UV/green PCBPVB 181,182 PCC 6803 382/534

NpF6001 N. punctiforme unknown blue/yellow PCB 26 426/578

NpR1597g1 N. punctiforme unknown blue/teal PCBPVB 26 420/498

NpR5113g1 N. punctiforme unknown green/teal PCBPVB 26 564/494

NpF2164g3 N. punctiforme unknown violet/orange PCB 23 ATCC 29133 400/600

NpF2164g2 N. punctiforme unknown UV/blue PCB 163 ATCC 29133 334,378/448

Table 1.2. Known CBCRs and CBCR photoactive GAF domains Different classes of CBCRs include red/green CBCRS (red box), green/red CBCRs (green box), DXCF CBCRs (blue box) and insert-cys (purple box). Many further N. punctiforme CBCR GAF domains of unknown function have been listed in 162,184,306.

41 | P a g e

CBCRs undergo the same primary photochemistry as Phys, photoisomerising about the C15=C16 double bond between a Z ground state and an E photoactivated state. The bilin chromophore for this class is either phycocyanobilin (PCB) or phycoviolobilin (PVB). Past the primary photochemistry, the photocycles must vary to account for the large spectral shifts observed in this class of proteins. To date, the characterisation often does not reach much further than basic spectroscopy and the full description of a photocycle is still missing. CBCRs have however, been subdivided into different classes40 according to their spectral properties and are described in more detail below: red/green CBCRs, green/red CBCRs, dual-Cys CBCRs (insert-Cys CBCRs and DXCF CBCRs). The latter use a second covalent linkage between the chromophore and the protein to break up the conjugation between the pyrrole B and C ring. Functionally, they are often involved in photochromatic adaptation and phototaxis.170–172,181,183

.

Red/green and green/red CBCRs

Cyanobacteriochromes of these classes switch between red and green absorbing photostates. Despite incorporating the same PCB chromophore, they exhibit opposite photocycles with the 15Z dark state being either red165 or green170 and vice versa for the 15E photoproduct. Examples of these classes include the green/red chromatic adaptation regulator SyCcaS161,170,171 and the red/green CBCR NpR6012g4184 and its orthologueAnPixJg2.165 Since study of CBCRs is still in its infancy, characterisation of these classes remains incomplete. NpR6012g4 has been studied on the ps to ns timescale and is the CBCR with the highest quantum yield known in the Phy superfamily (0.4).22,168,185,186 AnPixJg2 has been characterised more comprehensively, demonstrating a PCB photocycle with two intermediates each for the forward and reverse reaction, which both conclude in ~1 ms (Figure 1.13).187 The first intermediate corresponds to the photoisomerised chromophore conformation (nominated “Lumi” in phytochromes). The APr state and its first intermediate have been likened to the Cph1 photocycle but further similarities could not be drawn.166 Despite the conserved chromophore and domain architecture, the chromophore binding pocket differs markedly resulting in different protein-chromophore interactions and photocycle mechanisms. The proposed mechanism for red/green CBCRs revolves around the photoinduced hydrogen-bond switching of an Asp side chain from rings A-C (15Z) to ring D (15E). The repositioning could mediate signal transduction to the effector domain. A Trp residue was demonstrated to be important for generation of the Pr

42 | P a g e state. The green shift of the photoproduct was suggested to either be the result of a change in chromophore protonation or of A-ring rotation.183 A ring rotation would be possible in CBCRs but not Phys since the ring is solvent exposed. Furthermore, a deconjugated A ring in the PVB E-conformer of many DXCF proteins is known to form a Pg state. The molecular details of why green/red CBCRs undergo an inverse photocycle are speculative but would be a great advance in the characterisation of cyanobacterial photocycles. Green/red CBCRs lack the conserved Asp involved in hydrogen bonding to chromophore nitrogen and a conserved His stabilising the ring plane. Chromophore structure and protonation state are therefore likely to differ.

Figure 1.13 Photocycle of the red/green CBCR AnPixJg2 Pr photoexcitation results in the formation of a red-shifted R1 intermediate (680 nm). The second (R2) intermediate is green-shifted (610 nm) and followed by

formation of the Pg state. Pg photoexcitation proceeds through two consecutively red shifted intermediates (570 nm – 630 nm) until the Pr state is regenerated. Yellow arrows indicate fluorescence of the excited states. Adapted from 187

43 | P a g e

Dual-Cys CBCRs

In dual-Cys CBCRs, the dark adapted state has two thioether linkages between the protein and the bilin chromophore. The first covalent bond is formed between a conserved Cys residue of the GAF domain and the C3 ethylidene side chain of the bilin chromophore as common with plant Phys and CBCRs. The second thioether linkage is formed between another conserved Cys and the C10 and is responsible for creating two near-UV to blue absorbing conjugated π systems.163,179 Photoexcitation involves both the classical Z to E isomerisation but usually also the thermal elimination of the C10 adduct. The reverse process returns the chromophore to the double-linked Z isomer (Figure 1.14).

Figure 1.14. C10 thioether formation in dual-Cys CBCRs Insert-cys and DXCF CBCRs employ a second thioether in order to tune their ground state absorbance. The bond is formed transiently between the chromophore C10 and a conserved cysteine. P (Propionate)

There are currently two subgroups of dual-Cys CBCRs: Insert-Cys and DXCF, which are thought to have evolved independently.163 The common chromophore ground state configuration implies that they share a near UV to blue absorbing dark state. The photoproduct state absorbances are varied ranging from blue to red.

A unique protein with the complete Phy photosensory domain but exhibiting a dual-Cys ground state has also been discovered (NpF1183). It exhibits a violet/orange photocycle but thermally converts to a red absorbing species. 163

44 | P a g e

Insert-Cys

Insert-Cys CBCRs are most closely related to red/green CBCRs (AnPixJg2, Slr1393g3) and employ a non-conserved insertion loop for the second thioether linkage.163 The second conserved Cys is found within this loop region in a weakly conserved Cys-X-X- Arginine/Lysine motif. The regions around both Cys residues appear unstructured and have been suggested to facilitate the interaction with the PCB chromophore.188 Representatives of this class were discovered in the N. punctiforme genome, e.g. NpR1597 (UB1) and NpF2164g3 (VO1).163 UB1 exhibits a UV/blue absorbing photocycle with a constant double thioether linkage. VO1 has a labile thioether linkage and switches between a violet and an orange absorbing photostate. The VO1 photocycle has been characterised by time-resolved UV/Vis spectroscopy (Figure 1.15):23 The primary photoreactions were found to occur in ≤5 ps for both the forward and reverse isomerisation. The formation and elimination of the thioether linkage occurred post 1 ms. Three intermediates were proposed each for the forward and reverse reaction, whereby the intermediates of the forward reaction only differ in extinction coefficient and not in spectral position.23

There is currently no direct structural data available. However, the insert-Cys is predicted to be too far from C10 for adduct formation, meaning either the insertion loop has to move to the chromophore cleft or the thioether bond has to be intermolecular.183

Figure 1.15. Insert-Cys CBCR photocycle NpF2164g3 forward and reverse photoconversions proceed via three intermediates. The thioether formation and breakage are the final steps in each conversion. Adapted from 23

45 | P a g e

DXCF CBCRs

For DXCF CBCRs, the second conserved Cys is found in a motif of Aspartate (D) – any amino acid (X) – Cysteine (C) – Phenylalanine (F). DXCF CBCRs often possess autoisomerase activity tuning the PCB precursor chromophore to a phycoviolobilin (PVB), which is a unique property of this subgroup.175 Spectroscopically this mainly affects the photoproduct state, where the A ring is deconjugated in PVB but not PCB. In the doubly-linked ground state the Cys-C10 adduct restricts the main conjugated system to rings C and D which are structurally identical in PCB and PVB (Figure 1.14). The ground state is consequently sensitive to UV/violet/blue light, whereas the photoproduct absorbance is more varied, including teal, green, yellow, orange.26 Teal absorbance is achieved by a slightly different mechanism, where full chromophore isomerisation is prevented sterically.

Members of this group include SyPixJ, one of the first CBCRs characterised in vitro,159 its functional homologue, TePixJ189 and Tlr0924.179 DXCF GAF domains have also been found as components of multi-GAF proteins in the Nostoc punctiforme genome.162

The positive phototaxis regulator, SyPixJ, remains one of the few CBCRs ever purified from their host organism and was the protein in which PVB was first discovered as CBCR chromophore.159 Heterologously expressed TePixJ harbours both PVB and PCB and the GAF domain was ascribed autoisomerase activity.190 TePixJg is also the only CBCR with known structures for both photostates.176

Tlr0924 is the protein of interest in this thesis. While only one publication was available at the beginning of this work,179 our knowledge has progressively improved. Tlr0924 was also found to harbour two chromophores, the majority being PVB with some PCB bound. To what extent this incomplete isomerisation is intended or physiological is unknown. It appears to be fairly slow and potentially irreversible.162 Furthermore, the protein photostates showed temperature sensitivity. The significance of the second thioether linkage between C10 and Cys499 has been demonstrated.26

Tlr0924g is the only DXCF CBCR characterised on the ultrafast timescale.180 The PVB and PCB photocycles follow similar parallel reaction pathways, albeit PCB excited state relaxation is slower (Figure 1.16).180 Uniquely, Lumi-B intermediates are blue-shifted to the ground state, which is the first observation of this kind in the whole family. The secondary reactions were not time-resolved. However, the final intermediate of the reverse reaction can be trapped chemically by blocking thioether formation with iodoacetamide.26

46 | P a g e

Figure 1.16. DXCF CBCR photocycle of Tlr0924g Light triggered isomerisation results in the formation of “Lumi” primary photoproducts. The thermal cleavage or formation of the second thioether linkage from the “Meta” states account for the large red and blue shifts. The letter “b” indicates the intact thioether linkage. Adapted from 180

Based on the TePixJ GAF structure, the DXCF CBCR photocycle has been proposed to involve H-bond rearrangement from the pyrrole D-ring to aspartate (Asp) in the photoproduct state to rings B and C in the ground state. This Asp movement might represent the Lumi to Meta transition and could bring the Cys residue into proximity of C10 to form a covalent linkage. In photoactivation, the Asp and Cys movements would destabilise the covalent bond. The motions would induce signal transduction and initiate a signalling output.

47 | P a g e

1.4 Application

Photoreceptor research has many applications that can be, but are not necessarily related to the protein function. Some photoreceptor research has led to more fundamental insights into different protein classes: PYP served as the prototypical PAS domain.191 Rhodopsin provided the first ever membrane protein structure solved and the only G-protein coupled receptor (GPCR) structure available for many years192 – despite GPCRs being targeted by 50% of commercial drugs.193 Characteristic spectral properties and well-defined time resolutions have also warranted photoreceptors a place as model systems for general study of protein dynamics and functional conformational changes.9

Specifically, a better molecular understanding of phytochromes – potentially using cyanobacteriochromes as a model – is anticipated to provide the basis for improved crop yields in high density plantings e.g. by disabling the shade avoidance response.6 However, the main area of interest with more immediate impact is currently within the fields of synthetic biology,194 biotechnology and optogenetics.8,195

On the simplest level, photoreceptors can be engineered with respect to their most fundamental aspects, such as absorption and emission, kinetics or quantum yield (Figure 1.17A). Examples of such alterations are the spectral tuning of rhodopsin,196 chromophore substitutions,197 and altered kinetics through mutagenesis.198 Emission quantum yields have been purposefully increased through mutagenesis to generate new types of fluorophores to complement the most commonly used jellyfish fluorescent probes (Figure 1.17B).199 The principle behind these is to attempt to block the evolutionarily optimised, productive chemical pathways and force excited state relaxation through a radiative pathway (Figure 1.1). This change of functionality approach has been successfully demonstrated in LOV sensors200,201 and phytochromes.7,202–204 A patent has been filed for specifically bilin-type “phytofluors” demonstrating their commercial potential.205

Phys and CBCRs seem particularly suited given their broad spectral coverage, readiness to accept different bilin chromophores and ease of mutagenesis (e.g. point mutation from red/far-red to violet/orange163). Phys red-light sensitivity furthermore enables deeper tissue penetration.206 Another important factor shared between Phy, CBCR, xanthopsin, LOV and BLUF sensors is their architecture of individually functional domains, including a large variety of output domains. It has been demonstrated that photoreceptors can be designed artificially by recombination of photosensory input domains and effector domains

48 | P a g e

(Figure 1.17C), e.g. LOV blue-light regulation of a DNA binding protein,207 dihydrofolate reductase,208 histidine kinase,209 small GTPase (Rac1)210; bacterial phytochrome red-light regulation of histidine kinase194; phytochrome-controlled actin assembly.211 Given the varied spectral sensitivity of CBCR photosensory domains, it seems a realistic goal to use several artificial photoreceptors in parallel. Uniquely, Phys and CBCRs can be both activated and inactivated by light.

When light-sensitive proteins are expressed endogenously in non-parental cell lines or organisms, the field of optogenetics is entered (Figure 1.17D). This is possible because photoreceptor chromophores are often based on common cellular metabolites that are incorporated through photoreceptor autolyase activity. Optogenetics was established in 2005 with the discovery that light-activated channelrhodopsins could be used for selective polarity control in neuronal studies.189,212–214 This approach is reversible and less invasive, avoiding dissection for electrical stimulation or application of external neurotransmitter, while providing unprecedented spatio-temporal control. As another example, photoactivated adenylate cyclase has been used to rapidly modulate the levels of cyclic

215 adenosine monophosphate second messenger.

A – optimised photoreceptors B – fluorescent photoreceptors

hv hv * *

kp kf kp kf

C – artificial light switches D – heterologous expression off

+ hv hv on

effector protein

Figure 1.17. Biotechnological use of photoreceptor modification

A Photoreceptor optimisation through spectral and kinetic tuning achieved by mutagenesis and chromophore substitutions B Fluorophore synthesis through

blocking of productive chemical pathways (kp) in favour of fluorescent de-excitation

pathways (kf) C novel photoreceptor design by linking photosensor and effector domains D Heterologous expression of photoreceptors to control certain aspects of cellular or organismal behaviour. Adapted from 32 49 | P a g e

1.5 Thesis aims and objectives

In contrast to many specialised labs around the world focussing on one particular aspect of photoreceptor chemistry, in this thesis a more holistic approach was used to study the photocycle of the CBCR Tlr0924, beginning with the light-dependent, elementary femtosecond processes through to coupled slower light-independent steps. This confers the distinct advantage that all measurements could be done under defined, consistent conditions permitting a comprehensive analysis.

Chapter 3 is concerned with the more basic stationary characterisation of the full-length Tlr0924 protein. It describes growth and purification conditions and UV/Visible characterisation.

Chapter 4 describes the Tlr0924 forward reaction in terms of intermediates by using time- and temperature-resolved techniques. The PVB and PCB chromophore populations are isolated, analysed and compared to each other.

Chapter 5 presents the reverse reaction with the same technical approach as employed in chapter 4. However, the PCB population could not be analysed satisfactorily by ultrafast techniques because of the very low signal intensity. The ns to ms data was not included in the publication for this reason but has been added as a note to chapter 5.

50 | P a g e

Chapter 2

Spectroscopy: Experimental set-up

- 2. 1 Stationary absorption spectroscopy - 2.1.1 UV/Vis absorption - 2.1.2 Low-temperature UV/Vis absorption - 2.1.3 Infrared absorption

- 2.2 Time-resolved absorption spectroscopy - 2.2.1 Ultrafast visible absorption - 2.2.2 Ultrafast infrared absorption - 2.2.3 Flash photolysis

- 2.3 Data analysis

- 2.4 Summary

51 | P a g e

2. Spectroscopy

Advanced spectroscopic techniques were used in this thesis to study the mechanistic aspects of the full-length CBCR Tlr0924. This chapter therefore briefly summarises the theoretical basis, the general experimental set-ups and methods for data analysis used in subsequent chapters of the thesis. This provides essential context to understand the relevance and meaning of the spectroscopic studies. Standard biochemical techniques used for protein expression and isolation are described in chapter 3.4.1.

Spectroscopy plays a fundamental role in scientific research of biological systems.216 By exploiting the interaction of electromagnetic radiation (EMR) with matter, it provides unique insights into molecular structure and light-response mechanisms. Spectroscopic techniques can be applied to samples of different consistency. Gas phase spectroscopy is used for very fine resolution and shows vibrational and rotational fine structures but is unsuitable for complex systems. Solid samples generally show better signal-to-noise ratios in both absorbance and scattering based techniques but prevent physiologically relevant dynamics.217 Liquid samples are most commonly studied, particularly in biology, and sometimes lack the detail achieved with gaseous or solid samples but permit the molecule to move freely, mimicking a more physiological environment. Spectroscopic techniques make use of all EMR available from nuclear magnetic resonance and electron paramagnetic resonance in the microwave region, via rotational techniques, Raman and infrared (IR) spectroscopy in the infrared region and UV/Visible spectroscopy in the small visible range of the EMR spectrum. X-ray and γ-ray spectroscopic techniques can also give detailed structural insights (Figure 2.1). This section will focus on the electronic absorption spectroscopy used in this thesis, which cover the mid-IR and visible region.

visible radiowaves microwaves infrared UV X-ray γ-ray

λ 1 m 1 mm 1 μm 1 nm 1 pm

ν 109 1012 1015 1018 1021 s-1 Photon energy 10-6 10-3 1 103 106 eV nuclear magnetism rotation vibration electronic nuclear

Figure 2.1. Spectral regions of electromagnetic radiation

The electromagnetic spectrum covers radiation of a wide range of wavelengths. The type of interaction with matter varies greatly with the wavelength and is indicated in 218 the bottom line. λ (wavelength), ν (frequency), eV (electron volt) Adapted from

52 | P a g e

Spectroscopic techniques employ, on the most basic level, a light source, the sample cell and a detector (Figure 2.2).

Light source Sample cell Detector

I0 I l

Figure 2.2. Schematic absorption spectroscopy set-up.

A sample is exposed to incident radiation (I0) from a light source. The fraction of light that passes through the sample (I) is quantified at a detector. Transmission depends

on the path length of the sample (l), its extinction coefficient and concentration. Depending on the set-up the light source may be monochromatic or wavelength

selection devices (e.g. prism, grating) may be introduced.

In absorbance spectroscopy, the fraction of light transmitted through the sample is monitored to draw conclusions about the species present and its concentration. The relationship is described by the Beer-Lambert law as

-ɛcl T = I/I0= 10

where T is transmittance, I0 is the incident radiation and I the radiation after transmission through the sample chamber. Transmittance is commonly converted to absorbance (A) to give a linear relationship to concentration (c), path length (l) and extinction coefficient (ɛ) in an instrument specific absorbance range.218

A = -log (I/I0) = ɛcl

To gain the most information possible, the sample is probed over a range of monochromatic wavelengths and absorbance plotted as a function of wavelength to generate a ‘spectrum’. Molecule specific extinction coefficients and known path lengths permit one to quantify the species studied, whereas the absorbance pattern provides information regarding the identity of the absorbing species.

53 | P a g e

2.1 Stationary absorption spectroscopy

When acquiring a static spectrum, consecutive monochromatic wavelengths are passed through the sample and the absorbance is recorded (Figure 2.3). Since, as described in chapter 1.1, transitions between different states have conserved associated energies, molecules absorb or emit light at fixed energies, and therefore wavelengths, resulting in characteristic spectra. Factors such as temperature, viscosity or cosolvents can be varied and their effects on the spectra quantified.

S3

S2 Energy S1

v4 v3 v2 v1 S0 v0 Absorbance

Figure 2.3. Principle of stationary absorption spectroscopy Spectrophotometers often progressively scan the entire spectral region. As the energy of the incident photons matches the energy gap between the molecule’s ground and electronically excited states, they are absorbed and consequently less light reaches the detector. The sum of all transitions gives rise to the molecule specific absorbance spectrum. The arrow width is intended to represent the most likely transitions of the

system.

54 | P a g e

2.1.1 UV/Visible absorption spectroscopy

Electronic transitions are measured with a UV/Visible (UV/Vis) spectrometer. The configuration of a commonly used spectrometer is described in Figure 2.4. The light source is a pulsed xenon lamp which covers the UV/Vis region by wavelength selection with a diffraction grating. Some of the light is diverted to a reference detector with a beam splitter while the remainder passes through the sample and is captured at a second silicon diode detector. The pulsed nature of the xenon lamp means photodamage is minimised as samples are only exposed to a certain wavelength for the duration of the recording.

Detector Sample

Grating Mirror Beam splitter

Figure 2.4. Cary Varian UV/Vis spectrophotometer Collimated white light is dispersed into its component wavelengths at a diffraction grating (blue). A specific wavelength is then focused onto the sample (pink). The absorbance of the sample is correlated with the wavelength.

2.1.2 Low-temperature UV/Visible absorption spectroscopy

In contrast to stationary UV/Vis spectroscopy at room temperature, low-temperature studies can be used to characterise reactions in more detail. Almost all biological processes are temperature dependent and low temperatures can be used to slow down molecular motions or fast enzymatic reactions.219 At the temperature of liquid nitrogen (77K) molecular motions are effectively arrested. The low temperature stabilises otherwise very transient intermediates so they can accumulate to a detectable level, also enabling further study. Gradual increase of the temperature permits the reaction to proceed to different stages, which allows for the consecutive detection of these intermediates.

55 | P a g e

Pure water crystallises at temperatures below 0°C, so miscible organic solvents are frequently used as “antifreeze” in these types of experiments.220 Ideally, non-viscous solvents are preferable, e.g. methanol, but these can be incompatible with many biological systems. Other options include DMSO or ethylene glycol. Glycerol and sugars have been found as natural freezing point depressants in poikilothermic organisms.221 Their physiological role as intermediates in lipid metabolism made them cryoprotectant of choice in biology and medicine.222 Cryoprotectants maintain samples in a liquid state down to temperatures of 200 K.223 Instead of freezing, the phase change is from liquid to glass, which remains transparent and permits spectroscopic studies. This temperature is termed the glass transition temperature and represents the point where large domain motions or conformational changes in proteins, which are coupled to the surrounding solvent, become forbidden. However, localised motions or light-driven primary reactions within chromophores / binding pocket can still occur below this transition temperature.224

The low-temperature technique is not without drawbacks or limitations. The addition of organic solvents can alter polarity and viscosity and therefore, can affect the potential energy surface and reaction pathways. Low-temperature studies may also permit the accumulation of physiologically irrelevant intermediates that would otherwise degrade more rapidly than they are formed. In an ideal scenario, both cosolvent and temperature should not affect structural or functional properties of the protein of interest. Experimentally, these measurements are realised by fitting a UV/Vis spectrophotometer with a cryochamber as the sample cell (Figure 2.5). This set up permits exchange and maintenance of the temperature of the sample via either a heat mantle or cooled nitrogen gas.

Figure 2.5. Cryochamber for low-temperature experiments The cryochamber has two regulators for temperature: a heat exchanger to raise it and valve-operated liquid nitrogen reservoirs for cooling. The device is maintained in a vacuum for insulation purposes. OVC (outer vacuum chamber)

56 | P a g e

2.1.3 Infrared absorption spectroscopy

As indicated in Figure 1.1, electronic states have associated rotational and vibrational states. These are energetically closer than electronic energy levels and therefore sensitive to longer, less energetic wavelengths of the infrared region of the electromagnetic (EM) spectrum (Figure 2.1). In molecular systems the vibrational energy levels correspond to bond vibrations of particular frequency. Different types of bond vibrations are possible, the number of which is determined by the number of atoms within the molecule (N) and its geometry by the relationship 3N – 5 for linear molecules and 3N – 6 for non-linear molecules (Figure 2.6).225 Molecules are only vibrationally active if they possess a dipole moment. Atoms and homomeric bonds do not show any vibrational levels.

Molecule

rocking Symmetric Stretching

Asymmetric stretching wagging

Bending

twisting

A B

Figure 2.6. Vibrational modes Heteromeric molecules undergo vibrational motions due to their dipole moments. Different types of motion can occur depending on the structure of the molecules with A possible in trimeric molecules but B confined to larger molecules as there is a displacement relative to the remainder of the molecule. Adapted from 225

IR techniques are frequently used to identify chemical compounds and groups. They can also be applied to samples that are spectroscopically silent in the visible region. The position of the absorbance features is indicative of the bond and vibration type induced in the molecule. Figure 2.7 summarises some of the vibrations observed and their respective occurrences. Positions are however not definite and isotopic labelling is frequently used for more accurate assignment. In this study IR was used to complement UV/Vis techniques

57 | P a g e with further information on structure, environmental interaction and electronic properties of the cofactor in its binding pocket.

O-H, N-H, C-H

- H = C, N, O = = C=O

=

S- H C=C, C=N - C-C, C-O, C-N -

3000 2000 1000 Wavenumber (cm-1)

Figure 2.7. Vibrational modes and their energies Infrared absorption frequencies can be used to aid identification of molecules. Some common types of molecular bonds are shown in this graph.

The vibrational transitions of the protein in solution in this study were recorded using a Bruker Vertex80 that covers the mid-IR region (Figure 2.8). Probe light from a broadband IR source passes through a Michelson interferometer, where the beam is split onto two mirrors, one of which is movable. When the two beams are recombined they interact constructively or destructively. Light absorbance by the sample generates a specific interferogram at the detector, which is subsequently Fourier transformed to generate a spectrum of absorbance versus wavenumber.

Beam splitter Mirror Sample

IR Detector

Figure 2.8. Schematic diagram of a Bruker vertex80 FTIR spectrophotometer The infrared output beam is split at a beam splitter and passes to a stationary and a movable mirror. The reflection is recombined and passes through the sample. The detector records an interferogram, which is converted to a spectrum by Fourier transformation.

58 | P a g e

2.2 Time-resolved absorption spectroscopy

Time-resolved spectroscopy is another means and major tool for characterising the mechanism of light-activated reactions. The advantage over low-temperature studies is that standard buffers can be used. Additionally, it provides the lifetimes of the intermediates detected with potentially excellent time-resolution. Time-resolution is achieved by initiating the photoreaction with a comparatively short light pulse, termed pump, and monitoring the absorbance changes with a different “probe” light at time delay (t).This can be done either with a white light probe measuring full spectra at fixed time points or using an array of monochromatic probes and recording kinetic data. The measurements generate a two dimensional data array A(t,λ), which contains spectral and kinetic information on the intermediates involved in the photoreaction.

In contrast to stationary measurements, in which data describe the species present, time- resolved measurements mostly generate difference spectra with the sample prior to the photoreaction being used as reference: ΔA(t,λ) = A(t,λ) - A(-∞,λ)

Timescales commonly covered range from femtoseconds to seconds and the time window measured determines the identity of the intermediates observed. The faster time resolutions have only become accessible with the advent of lasers, starting in the 1960s.226 Lasers can generate intense and narrow beams of light with enough energy to populate intermediate states to a detectable concentration. This opened up the possibility to detect transient, short lived species, such as reactive intermediates and excited states in real time and therefore, to characterise photochemical reactions in unprecedented dynamic detail.

The technique can be limited by photodegradation of the sample through high power laser pulses and can require large sample expenditures to study irreversible photoreactions. Time-resolved spectroscopy was introduced in 1949 under the name “flash-photolysis” by Norrish and Porter, who received the Nobel Prize for their discovery in 1967.227–229

59 | P a g e

2.2.1 Ultrafast visible absorption spectroscopy

Ultrafast spectroscopy is here defined to occur in the time window of femtoseconds to nanoseconds. This time resolution was first made possible with the development of femtosecond light pulses in the early 1980s230 and has been the basis of studying photosynthetic reactions and light-activated biological systems.231–234

The set up used in this work is a commercially available “Helios” system (Ultrafast Systems) as outlined in Figure 2.9.

Ti:Sapphire 802 nm Probe 100 mW Amplifier system 900 mW Pump NOPA Laser system 5 mW Delay stage Reference Sample Detector Detector 2 mW

CaF2

Mirror Beam splitter Chopper

ND Filter Shortpass Filter Sample

Figure 2.9. Schematic diagram of a Helios transient absorbance spectrophotometer The Ti:sapphire laser output is split to generate a pump and a probe beam. The 120 fs pump beam arrives at the sample in a stirred cuvette prior to the probe beam with a programmable delay of up to 3.3 ns. NOPA (noncollinear optical parametric amplifier), ND (neutral density)

A Ti:Sapphire amplifier system generates 800 nm pulses at 1 kHz frequency and an average power of 3.5 W. A small fraction of this power is split off for the probe beam. The vast majority is required to pump a noncollinear optical parametric amplifier (NOPA) to generate the visible pump beam. The frequency of the pump beam is reduced to 500 Hz using an optical chopper to allow for “pump on” and “pump off” readings from which the difference spectra are generated. The probe beam travels via a delay stage, the position of which determines the time delay of pump and probe at the sample and hence of the reading. To measure the full spectrum, the 800 nm probe beam passes through a CaF2

60 | P a g e crystal, which generates a white light continuum. Some of the probe light is diverted to a reference detector to correct for interpulse variation in amplitude. The overall time resolution of the system is 170 fs and the time range extends up to 3.3 ns determined by the optical delay line.

Data sets generated with this technique show bleaches where the ground state is depleted and positive features where new species are formed (Figure 2.10).

Figure 2.10. UV/Visible absorbance difference spectrum The spectral features contributing to absorbance spectra on the fs to ps timescale are ground state bleach (GSB), stimulated emission (SE) and excited state absorption (ESA). Adapted from 235

Following promotion of the sample to electronically excited states by the pump beam, the ground state population is comparatively depleted. This is seen as a negative spectral contribution, called the ground state bleach (GSB). The newly populated excited states absorb probe light, giving rise to positive excited state absorbance (ESA) features. These are often red-shifted based on the smaller energy gaps between higher electronic excited states. A red-shifted, negative feature also arises in the case of stimulated emission (SE), where the probe photon stimulates the release of an excited state photon.235

Within the time frame of these measurements (usually ps – ns), many different primary transitions can occur as described in chapter 1.1. These include photophysical processes (Table 1.1) such as rapid vibrational relaxation, internal conversion, intersystem crossing and fluorescence decay, but also photochemical conversions, e.g. proton transfer and cis/trans isomerisation.95 With some of these processes potentially occurring in parallel, sophisticated data analysis programs are required.

61 | P a g e

2.2.2 Ultrafast infrared absorption spectroscopy

The technique is analogous to UV/Vis transient absorbance, using the same pump wavelength to initiate the same photochemistry, but shifting the probe beam to the IR region of the EM spectrum. Spectra originate in this case from vibrational transitions, which are much lower in energy than the electronic transitions probed in the visible region. Energies are therefore expressed as wavenumbers rather than wavelengths.

The experimental setup used for the work in this thesis is located at the Central Laser Facility at the STFC Rutherford Appleton Laboratories236 and a schematic is shown in Figure 2.11.

Ti:Sapphire OPA Pump Amplifier system IR Probe OPA DFG Laser system Delay stage

Sample Spectrometer Detector

Reference Spectrometer Detector

Mirror Beam splitter Chopper Sample

Figure 2.11. Schematic diagram of a time -resolved IR spectrophotometer The output beam of a 10 kHz Ti:Sapphire laser is divided into a visible pump beam and an IR probe beam. The pump beam passes via a delay stage to generate the time resolution at the sample chamber. The sample is flowed and rastered to prevent damage. OPA (optical parametric amplifier) DFG (difference frequency generator)

As with the Helios transient absorbance system, an 800 nm output beam is generated by a

Ti:sapphire amplifier system (10 kHz) and split into a pump and a probe beam. The pump beam is generated by an optical parametric amplifier (OPA) and is routed to the sample via a delay line. The probe beam is also tuned by an OPA and then difference frequency mixed to generate the mid-IR probe beam. The setup is again fitted with a chopper to generate difference spectra of “pump on” vs. “pump off”. Shot-to-shot variability in terms of amplitude and spectral fluctuation was recorded and accounted for with a reference beam. In contrast to the Helios set-up where one detector array is used, the 500 cm-1 measuring window is generated by two separate detectors used in parallel and the data merged at the analysis stage.

62 | P a g e

2.2.3 Flash photolysis

Flash photolysis employs the same principle as the ultrafast techniques described above. A pump pulse is sent to the sample to initiate a photochemical reaction whilst a probe pulse monitors the change in absorbance. There are some fundamental differences between the techniques. Flash photolysis is not necessarily a laser-based technique and the pump pulse may equally be provided by a flash lamp or a light-emitting diode (LED). The pump light pulse width, probe light source and detector restricts the time resolution with this set-up to the nanosecond timescale upwards to seconds, minutes or hours if required. The disadvantage of this technique is that the probe pulse is monochromatic. In order to generate a comparable data matrix, the kinetics of the reaction need to be measured at many different wavelengths.

The experimental set up of the commercially available Applied Photophysics laser flash photolysis system used for this work is shown in Figure 2.12.

Laser system Monochromator Nd:YAG laser

OPO 2ω

Xe PMT Oscilloscope

Figure 2.12. Schematic of a laser flash photolysis set-up Visible pump pulses are generated by an Nd:YAG laser via either harmonics (ω) or an optical parametric oscillator (OPO) for more fine tuning. The probe beam of monochromatic light is generated by a xenon arc lamp and the transmitted photons counted at a photomultiplier tube (PMT).

A neodymium-doped yttrium aluminium garnet (Nd:YAG) laser generates a 1064 nm output beam, which is frequency doubled or tripled by the harmonics or tuned with an optical parametric oscillator (OPO). 12 ns pulses of this output beam are used to orthogonally excite the sample. The probe beam is generated by a 150W Xenon Arc flash lamp, which is pulsed for sub-millisecond time windows. The probe wavelength is selected by a monochromator. A second monochromator after the sample cell cuts out laser scatter before the kinetics of the absorbance change are measured using a photomultiplier tube detector and recorded by a digital oscilloscope.

63 | P a g e

To measure long time-windows, kinetic measurements can be recorded in a spectrophotometer with an orthogonally placed, pulsed LED exciting the sample after 10% of the time window to obtain a pre-trigger baseline. The advantage of this set up is that the spectrophotometer light source is much less powerful and pulsed to reduce photodamage to the sample.

Given the timescales of photophysical and photochemical processes, it can generally be assumed that the signals observed are no longer photophysical loss processes. They may show newly formed triplet states, intermediates and photoproducts of the reaction.

2.3 Data Analysis

In order to facilitate the interpretation of data matrices, they were subjected to global analysis.237 Especially with ultrafast measurements, where many processes occur at the same time, this is a means to extract time and/or species profiles and potentially suggest a model for the reaction. The analysis is based on the assumption that the data matrix can be described as the sum of the spectra of different species that only change in concentration as time progresses. The time profiles are so called evolutionary associated difference spectra (EADS). If the assumptions of the model are accurate, EADS correspond to the actual species spectra minus the ground state spectrum and are then called species associated difference spectra (SADS). EADS are otherwise linear combinations of the component SADS and still provide a useful description of the data.237

Spectra were also fitted to a linear combination of Gaussians to determine the likely position of all contributing species. Gaussian shapes were shown to be a good fit as can be expected for an inhomogeneous broadening mechanism.238

2.4 Conclusion

Even though not all techniques have been discussed in detail, each individual technique contributes a small piece of information to a complex mosaic when it comes to understanding a photoreaction. While the UV/Vis techniques report on the electronic transitions in the chromophore, time-resolved IR spectroscopy shows real time dynamic change of the chromophore but also proximal amino acids.

64 | P a g e

Chapter 3

Heterologous expression and spectral characterisation of the cyanobacteriochrome Tlr0924 from Thermosynechococcus elongatus

Authors: Anna F.E. Hauck, Roger J. Kutta, Derren J. Heyes and Nigel S. Scrutton

Affiliation: Manchester Institute of Biotechnology, Faculty of Life Sciences, The University of Manchester, Manchester M13 9PL, United Kingdom

Published in: Manuscript in preparation.

65 | P a g e

3.1 Abstract

Cyanobacteria rely heavily on light absorbance for photosynthesis and have a set of pigments to convey broad spectral sensitivity. A class of mainly soluble chromoproteins called cyanobacteriochromes was found to mediate chromatic adaptation and phototactic responses to presumably optimise light harvesting properties. At the molecular level these proteins are homologous to plant phytochromes and undergo E/Z isomerisations of their linearised bilin chromophores to switch between two photostates. Tlr0924 is an example of a cyanobacteriochrome in Thermosynechococcus elongatus and has been characterised biochemically in the present work. It was demonstrated that the heterologously expressed protein harboured two chromophore populations (phycoviolobilin and phycocyanobilin) with identical blue light-absorbing dark states (Pb), which converted to green- and red light- absorbing photoproduct states (Pg and Pr) upon light-induced isomerisation in a reversible manner. The basis of the large red-shift is the elimination of a thioether linkage between the chromophore C10 position and a conserved cysteine residue, reaction or oxidation of which prevented return to the dark state. In slow, light-independent processes, the phycocyanobilin population isomerised to phycoviolobilin and the Pr and Pg states underwent dark reversion to the Pb states. The protein showed some temperature sensitivity, but photoconversion appeared to be dominated by light exposure. Both photostates were sensitive to large portions of the visible spectrum, with the photoproduct states also converting back to the blue light-absorbing form upon exposure to blue light.

66 | P a g e

3.2 Introduction

Phytochromes (Phys) have long been known as photoreceptors mediating red and far-red light responses in plants.126 They were extensively studied for decades as mediators of many agriculturally important traits. Following the unexpected discovery of a first cyanobacterial Phy photoreceptor (Cph1)239 and the consequential surge of genome characterisation in this class of organisms, a related photoreceptor class termed “cyanobacteriochromes” (CBCRs), was defined.10 CBCRs were initially interesting as potential model systems for Phys to solve mechanistic and structural questions but have now gained importance as spectroscopically and functionally varied photoreceptors with applications in biotechnology.203,240 Phys and CBCRs share a conserved light-sensing mechanism, consisting of a linear bilin chromophore that reversibly photoisomerises around its C15=C16 double bond in response to mainly two particular wavelength ranges of light. In Phys these are restricted to red and far-red, whereas CBCRs are sensitive to many different visible spectral ranges. Based on this sensitivity, the CBCRs have been subdivided into at least four different families.40 Two groups absorb red and green light (green/red and red/green CBCRs), while the remaining two also form a second thioether linkage with the chromophore during the photocycle and are called insert-Cys and DXCF CBCRs.163

Tlr0924 is a hypothetical two component regulatory protein.241 It is named after its gene locus in the cyanobacterium Themosynechococcus elongatus, a unicellular thermophilic organism.242,243 The domain architecture includes the photoactive GAF domain that is common to all Phy and CBCR proteins, N-terminal cystathionine β-synthase multimerisation motifs and a C-terminal diguanylate cyclase domain,244 suggesting that cyclic diguanylate synthesis may be the output of the activated protein. Tlr0924 was amongst the first of the many CBCRs now identified and converts between a blue- (Pb) and a green light-absorbing state (Pg). The originally proposed mechanism in this class of protein involved chromophore tautomerisation from phycocyanobilin (PCB) to phycoviolobilin (PVB) and the B ring twisting out of plane in the 15Z Pb state.175 This view was revised when Cys499, which is in close proximity to the chromophore C10, was demonstrated to be essential for the formation of the Pb dark state.179 It was subsequently shown by chemical reaction with iodoacetamide or hydrogenperoxide and chromophore substitution techniques, that the role of the Cys was to form a labile thioether linkage with the chromophore at the electrophilic C10 position.26 The reaction cycle therefore proceeds between a blue light-absorbing dark state with an intact thioether linkage and a green light-absorbing photoproduct state where this

67 | P a g e bond has been hydrolysed. Evidence for this hydrolysis has also been provided by FTIR studies on the DXCF CBCR TePixJ, where a free thiol stretching mode was observed upon Pg formation.190 With a PVB chromophore this means the photoactive π conjugated system is extended from rings C and D to include ring B. However, it was also demonstrated by acid denaturation and sequential photoconversion that the PCB precursor chromophore was not isomerised completely in Tlr0924162 and contributed to the spectral features. PCB conjugation at C5 results in an extended π conjugated system absorbing red light in the photoproduct state (Pr) and an additional PCB ring A+B feature in the Pb ground state (Figure 3.1). The conversion efficiency is thought to be temperature dependent with extreme temperatures (5°C and 55°C) generating smaller turnovers.179

hv

hv

Phycocyanobilin 15Z Pb 15EPr

hv

hv

Phycoviolobilin 15Z Pb 15EPg

Figure 3.1 PCB and PVB chromophores in Tlr0924 PCB and PVB chromophores assume a double thioether-linked, 15Z dark state conformation, which confers blue light sensitivity to the Tlr0924 protein. Photoisomerisation to the 15E photoproducts induces cleavage of the C10 thioether to yield red and green light-absorbing states. P (propionate)

In addition to the photochemical aspects, Tlr0924 was also found to undergo dark reactions. Initially, photostates were believed to be thermally stable in the absence of light.179 Protein in the Pr state was however later found to revert to the Pb state in the dark quicker than protein in the Pg state, on the time scale of hours.162 This thermal reversion was also accompanied by a change in the bilin composition. PCB was found to isomerise to PVB, preferentially in the Pr state over the Pb state.162 Full-length Tlr0924 protein furthermore demonstrated light-independent temperature sensitivity: Pg-state protein was shown to revert to a Pb-like state in the dark at low temperature in a reversible manner. However, photochemically generated Pb appeared stable at higher temperature.179 This study was conducted in parallel to those summarised in the introduction and describes protein synthesis and basic photocycle characteristics of full-length Tlr0924.

68 | P a g e

3.3 Materials and Methods

3.3.1 Plasmid construction.

Cph1 and Tlr0924 genes were synthesised by GenScript USA Inc.. They were amplified by polymerase chain reaction, introducing 5’ and 3’ Nde1/Xho1 restriction sites. Double digests were performed to ligate the genes into pET-15b vectors (Novagen). The constructs have N-terminal hexahistidine tags followed by a thrombin cleavage site (MGSSHHHHHHSSGLVPAGSHM...). Cloning was verified by DNA sequencing.

Tlr0924 in pBAD-myc/HisB and pCOLADuet-1/HO1,PcyA were a generous gift from D. Heyes (University of Manchester). Plasmid maps are shown in Figure 3.2.

Figure 3.2. Plasmid maps of pBAD-myc/HisB/Tlr0924 and pCOLADuet-1/HO1,PcyA The Tlr0924 gene was inserted into pBAD-myc/HisB through Xho1/EcoR1 restriction sites. PcyA was introduced into pCOLADuet-1 multiple cloning site 1 via Nco1 and EcoR1; HO1 was introduced into pCOLADuet-1 multiple cloning site 2 via Nde1 and Kpn1.

69 | P a g e

3.3.2 Protein expression

Cph1 and Tlr0924 proteins were expressed recombinantly in the E. coli strain BL21(DE3) (New England Biolabs) that also contained a PCB chromophore expression plasmid (pCOLADuet-1/HO1,PcyA; Novagen). Cells were grown at 37°C in Luria-Bertani broth in the presence of 100 μg/ml ampicillin and 30 μg/ml kanamycin for 5 h and diluted 1:150 in the same medium for overnight culture. The overnight culture was diluted 1:100 and grown to an optical density (OD600) of 0.6. The enzymes required for PCB were expressed upon induction with 1 mM isopropyl-β-D-thiogalatopyranoside (IPTG) at 25°C. For pET-15b plasmids, this simultaneously induced target protein expression. pBAD-myc/HisB was otherwise induced 60 min later with 0.002% (w/v) L-arabinose. Cells were harvested by centrifugation (5000 g, 10 min, 4°C) after overnight incubation. Cell pellets were stored at -20°C. Cell culture components all came from Formedium with the exception of L-arabinose (Sigma).

3.3.3 Protein purification

Cell pellets were resuspended in lysis buffer (50 mM sodium-potassium phosphate pH 7, 300 mM NaCl) and homogenised by sonication (40% amplitude, 45 min, 33 cycles). Homogenates were clarified by centrifugation (20000 g, 30 min), filtered and loaded onto pre-equilibrated nickel resin (Nickel HisTrap, GE Healthcare Life Sciences). Columns were washed and the protein eluted with elution buffer (50 mM sodium-potassium phosphate pH 7, 300 mM NaCl, 200 mM L-histidine). The protein was desalted and purified further over a gel filtration column (e.g. HiLoad 26/600 Superdex 200 pg, GE Healthcare Life Sciences) and stored at -80°C in lysis buffer. Protein purity was confirmed by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) in a Laemmli buffer system. Buffer components were supplied by Fisher scientific.

3.3.4 Spectroscopy

Absorbance spectra were recorded on a Varian Cary 50 UV/Vis spectrophotometer (Figure 2.4) equipped with a Varian Cary single cell peltier temperature controller. Samples were irradiated from above to initiate photochemistry using a Schott KL1500 LCD cold source lamp fitted with the relevant 10 nm band pass filters (Andover and Thorlabs) (Figure 3.3).

70 | P a g e

Figure 3.3. Lamp emission and bandpass filter transmission spectra

A Emission of the Schott KL1500 LCD cold lamp source was recorded by scattering the emission in a fluorescence spectrometer. B Transmission spectra of bandpass

filters used in this study. All filters had a bandwidth of 10 nm.

λ (nm) Pb state Pg state Pg/r state The relative photon absorbance of each state 400 0.08 0.01 0.01 using different filters is calculated as an overlap 420 0.56 0.07 0.05 430 1.00 0.12 0.07 integral of the lamp spectrum, the filter 450 0.20 0.03 0.02 transmission and the absorbance of the 500 0.12 0.51 0.53 540 --- 1.00 1.00 respective states in the spectral region of 630 0.69 0.19 0.32 illumination (Table 3.1). 640 0.45 0.13 0.17 670 0.02 0.00 0.01 Table 3.1. Relative photon absorbance

For investigation of the second thioether linkage, iodoacetamide (Sigma) was added at a concentration of 21 mM and incubated for 1 h in the Pg/r state. 30% H2O2 (Sigma) was added to the Pb states in a ratio of 1:1 (v/v) and absorbance spectra recorded immediately.

For temperature experiments, samples were allowed to equilibrate to the temperature for 30 min. Photoconversion was induced for up to 30 min to ensure full turnover and the kinetics were monitored during this period.

Samples for cryobuffer testing were made up in 1 ml at constant protein concentration and subsequently fully converted to Pb and Pg/r. Chemicals were sourced from Fisher scientific except for ethylene glycol (Sigma).

FTIR measurements were recorded on a Bruker Vertex80 spectrophotometer. A CaF2 transmission cell with 100 μm path length was used and the sample was measured in the mid-IR region in H2O buffer.

71 | P a g e

3.4 Results

3.4.1 Expression and purification of Tlr0924 full-length protein

Heterologous Tlr0924 expression was attempted in a 2-plasmid system in E. coli cells. The first pCOLADuet-1/HO1,PcyA plasmid (plasmid 1) was required to synthesise the chromophore (Figure 1.11).245 The second plasmid carried the apoprotein gene. Both a pET15b and a pBAD-myc/HisB vector (plasmid 2) were tested with the aim of producing soluble holoprotein.

Figure 3.4A shows the expression trial for the pET15b vector. BL21(DE3) cells failed to produce chromophore-bound protein irrespective of the concentration of the IPTG inducer, as indicated by the lack of colouring. The cyanobacterial Phy (Cph1) expressed in parallel as a positive control showed the characteristic colouring (Figure 3.4B).

C A

B

0 0.05 0.1 0.4 1 IPTG (M) 1 1+2 2

Figure 3.4. Expression of plasmids 1 and 2 in BL21(DE3) A pET15b/Tlr0924 co-expressed with PCB synthesis plasmid fails to produce

significant quantities of holoprotein in contrast to the positive control. B pET15b/Cph1 co-expressed with PCB synthesis plasmid shows distinctive colouring

of chromophore-bound protein. C pBAD-myc/HisB/Tlr0924 coexpressed with plasmid 1 produced a green cell pellet containing holoprotein. Expression of either

plasmid in isolation did not give the distinctive colouring, demonstrating that free chromophore is unstable and that BL21(DE3) are unable to synthesise bilin chromophores.

Protein expression in the pBAD-myc/HisB plasmid added a second layer of control by permitting selective induction of the PCB synthesising apparatus and the cyanobacteriochrome itself. It was found that the PCB expression plasmid induced on its own did not produce any chromophore, indicating that PCB in an unbound state is unstable.

72 | P a g e

Delay times of 60 minutes between induction of plasmid 1 and plasmid 2 produced intensely green coloured pellets containing the desired holoprotein (Figure 3.4C). The optimum conditions were found to be induction of plasmid 1 at a relatively high OD (1) followed by plasmid 2 with a delay time of 20 minutes.

The CBCR, but not the PCB synthesising proteins, has an N-terminal hexahistidine tag permitting selective purification via metal affinity chromatography (Figure 3.5A). The standard imidazole elution buffer failed to elute the protein completely and appeared to lock the protein in a Pg-like state (pink aggregate). This problem was overcome by substituting the imidazole with L-histidine as the eluent. The second step routinely used was size exclusion chromatography (Figure 3.5B). The purified protein was green in colour.

A B 150 100 75 50 37 25 20 15 10 [Histidine] Volume kDa

Figure 3.5. 2-step purification of Tlr0924 Representative SDS-PAGE gels of a nickel affinity chromatography step eluting with a histidine gradient up to 200 mM A followed by a gel filtration step B. Molecular weight standards in kilodalton are labelled in red.

The protein identity and molecular weight (87 kDa) could be correctly identified by mass spectrometry (in house facility - Velos Pro, Thermofisher Scientific).

In an alternative protocol, Tlr0924 was expressed as an apoprotein by inducing only plasmid 2, purified as above and reconstituted with PCB extracted from Spirulina powder. This approach was more time-consuming and spectrally less clean and therefore not used in subsequent experiments.

73 | P a g e

3.4.2 Stationary photoconversion properties

The purified full-length Tlr0924 protein had a distinctive absorbance spectrum with apparent peaks at 431 nm, 542 nm and 625 nm (Figure 3.6). Selective illumination of the 3 peaks with green (B), blue (C) and red (D) light revealed photoactivity for each one of them, indicating more than one homogenous chromophore population. The shortest wavelength transition is denominated Pb state (B) and can be considered the dark state of the protein246 as the majority of the protein adopts this conformation during protein preparation. Upon depletion of Pb, two new states are formed, referred to as Pg (C) and Pr (D). This conversion is accompanied by an unusually large spectral shift as well as a change in spectral shape. The underlying molecular change in the chromophore is known to be a Z/E isomerisation about the C15=C16 double bond of the bilin chromophore in Phys and CBCRs.176,247,248

Figure 3.6. Absorbance spectra of purified Tlr0924 Absorbance spectrum of as-isolated full-length Tlr0924 protein A and following

saturating illumination with green B, blue C and red D light from a cold light source fitted with bandpass filters. Insets show the characteristic colouring of the photostates. E shows the overlaying absorbance spectra.

74 | P a g e

Data presentation in the form of difference spectra (D-B; C-D) allowed a better visualisation of the contributions of the individual chromophore populations and assignments of their peak positions (Figure 3.7A). Both populations (PCB and PVB) exhibited overlaying dark states at 436 nm and could be photoconverted to form their respective photoproduct states with a major band at 532 nm and a subpopulation at 588 nm. Denaturation studies are frequently used as an additional indicator to identify the chromophore(s) utilised by biliproteins.162,164,175,190 Upon denaturation of the protein, the chromophore-protein interactions are disrupted, to reveal the characteristic spectra of the free chromophores. For Tlr0924, denaturation with acidic urea revealed the presence of PCB and PVB chromophores with the following absorption maxima: PCB, 578 nm and 674 nm; PVB, 512 nm and 609 nm (Figure 3.7B).26 The blue-shifted spectra in the holoprotein compared to the free chromophore therefore indicate that the π conjugated system loses planarity upon protein binding through specific protein-chromophore interactions.

Figure 3.7. Tlr0924 difference spectra A Difference spectra under non-denaturing conditions reveal a major PVB chromophore population, which photoisomerises between 436 nm and 532 nm absorbing states and a minor PCB population, which photoisomerises between 436 nm and 588 nm absorbing states. B Protein denaturation of Tlr0924 with 8 M urea at pH 2 reveals the presence of 2 chromophore populations representing PCB and PVB. PCB (Phycocyanobilin), PVB (Phycoviolobilin)

75 | P a g e

3.4.3 Importance of a second thioether linkage

The large spectral shift is unique to dual-Cys CBCRs and arises due to thioether formation or breakage between a conserved Cys and the C10 position of the bilin. The importance of the linkage could be demonstrated by a variety of techniques. Static Fourier transform infrared (FTIR) spectroscopy of the S-H stretching region has previously been used to identify the free Cys thiol in the singly linked Pg state.190 In an analogous experiment (Figure 3.8), the samples were baselined on the Pb state (A) and illuminated with blue light only to generate the Pg/r states and also with blue light followed by red light to generate the Pg state. In both cases, illumination was expected to show peaks for the free thiol stretching mode. In similar experiments baselined on the Pg/r state (B) and Pg state (C), green light illumination would be expected to show bleaches in the thiol stretching region upon formation of the second thioether in the Pb states. In the Pg/r state red light illumination should show a similar effect, albeit with a much smaller amplitude due to the lower prevalence of the PCB chromophore in the sample. Blue light illumination of the Pg state represented the reverse reaction for the PCB chromophore and hence was expected to show a small positive feature.

Figure 3.8. FTIR difference spectra under different illumination conditions. The S-H stretching mode at 2562 cm-1 becomes visible upon photoproduct formation and disappears upon return to the ground state. This is evidence for the formation of a second thioether linkage during the photocycle.

76 | P a g e

Data acquired followed this exact trend accurately with a 2562 cm-1 feature possibly reporting on the presence or absence of a second thioether linkage. Further FTIR experiments would be required to confirm these data.

To obtain additional evidence, the conserved cysteine was modified chemically to prevent formation of the second thioether linkage upon conversion to the Pb states (Figure 3.9). Two chemicals used were iodoacetamide (IAM) and hydrogen peroxide. IAM alkylates cysteine residues249 and thereby prevents thioether linkage formation. Hydrogen peroxide oxidises thiols250 with the same effect. IAM added to the Pb state induced no spectral changes. In the Pg/r states IAM alkylated the cysteine and green light exposure no longer resulted in conversion to the dark state. Instead, red-shifted absorbance peaks appeared, which represent the isomerised Pg/r states prior to formation of the thioether linkage (15Z

PCB and 15Z PVB) and are intermediates in the photoconversion reaction. H2O2 added to Pb oxidised the thioether linkage and resulted in a spectrally similar, red-shifted intermediate, which is also likely to represent 15Z PCB and 15Z PVB. The apparent absorbance maxima of these singly linked Z intermediates are at 561 and 626 nm (IAM) and at 558 and 618 nm

(H2O2). Other solvent accessible Cys residues are likely to also be modified but are expected

to have comparatively small spectral effects.

Figure 3.9. Blockage of thioether formation between C10 and Cys499 by IAM and H2O2. IAM (21 mM, 1 h) prevents formation of the second thioether linkage and therefore the Pb dark state from Pg/r. Illumination results in the isomerised 15Z intermediates.

H2O2 (15%, 0 min) cleaves the thioether linkage in the Pb state, also revealing the absorbance spectrum of the singly-linked 15Z states.

During long experiments (12h), spontaneous oxidation of the conserved second cysteine residue was also observed in the Pg/r state with the same spectral characteristics of thioether prevention.

77 | P a g e

3.4.4 Tlr0924 spectral sensitivities

In order to further characterise the PCB and PVB populations and their sensitivity to different regions of the visible spectrum, a comprehensive conversion trial was carried out, to investigate the responses to different wavelengths of light. Selected bandpass filters (Figure 3.3: 400 nm, 420 nm, 430 nm, 450 nm, 500 nm, 540 nm, 630 nm, 670 nm) were used and light exposure times increased until photoequilibria were established. Pb states were sensitive to 400-450 nm illumination, which induced conversion to the PVB Pg and PCB Pr photoproduct states (Figure 3.10). Maximal Pg state formation occurred at 430 nm (Figure 3.10A) and decreased in the order 420 nm > 450 nm > 400 nm (Figure 3.10B). This order also reflects the relative speed of conversion at peak wavelengths (Figure 3.10C) and the photon absorbance (Table 3.1). The Pr state formation followed the same trend kinetically. However, the level of Pr state formed with 450 nm illumination was lower than with 400 nm illumination. There consequently must be a difference in the absorbance profile between PCB and PVB in the blue region of the spectrum despite the assumption that the Pb states are identical for the two choromphore populations. Possible explanations may be that the PCB Pb absorbance spectrum is slightly blue-shifted in comparison to the PVB Pb spectrum or that the Pr state also has an absorbance feature in the blue region representing the S0S2 transition.

If the first option is correct, the apparent rates of formation would also be expected to invert. It can however not be excluded that the lower photon numbers (x2.6) at 400 nm are the origin of this inversion.

78 | P a g e

Figure 3.10. Pb state photoconversion properties Full-length Tlr0924 in the Pb state was exposed to different visible light ranges from a cold light source fitted with bandpass filters. Illumination time was increased until the photostationary state was reached. Pg and Pr were formed maximally by exposure to 430 nm light over the time course of 20 minutes A. The photoequilibria reached varied especially upon illumination with blue light B. The kinetics of Pb decay and Pg/r formation are shown C.

79 | P a g e

In the reverse direction, the Pg/r photoproduct states were sensitive to virtually all wavelengths tested. Photoequilibria were awaited at all wavelengths but 400 and 420 nm, which were slow to convert. A maximal level of Pb formation was achieved by illumination with 540 nm (Figure 3.11A) radiation followed by 500 nm > 630 nm > 640 nm > 670 nm (>400 nm) > 450 nm (>420 nm) (Figure 3.11B), which also loosely represents the kinetic trend (Figure 3.11C). As 400 nm and 420 nm illumination initiated photoconversion from the Pg/r to the Pb state, this implies that Pg/r states also have absorbance features in the blue spectral region.

The Pb state is formed upon depletion of both Pg and Pr populations, which are converted to different proportions at different wavelengths. It was shown that 540 nm radiation photoconverts both the Pg and Pr states equally to their respective Pb states. Illumination at 500 nm fully converts the Pg to the Pb state but small levels of the Pr state remain. Further in the red region, 630 nm and 640 nm illumination show complete conversion of the Pr state and a small level of conversion of the Pg state to the respective Pb states. Illumination with 670 nm light appears to photoconvert only Pr to Pb.

In the blue region, similar phenomena are observed as previously in the late stages of the Pb to Pg/r conversion experiment. These confirm that the Pg state is more sensitive to bleaching at 400 nm than at 450 nm and that the opposite holds true for the Pr state. The amplitude of Pg state depletion proceeds in the order of 540 nm > 500 nm > 630 nm > 640 nm > 670 nm (>400 nm) > 450 nm (>420 nm) illumination. The amplitude of Pr state depletion follows the order 670, 640, 630 nm > 540 nm > 500 nm > 450 nm (>400 nm) > 420 nm illumination. Kinetically, Pg depletion at 532 nm follows a similar order initially but then deviates (540 nm > 500 nm > 630 nm > 640 nm > 450 nm > 670 nm > 400 nm > 420 nm). Pr kinetics are heavily affected by dominant overlapping Pg depletion. The order is 630 nm > 640 nm > 540 nm > 500 nm > 450 nm > 420 nm > 400 nm.

The fact that time of conversion and amplitude are not consistent in the sequence can also be explained by off peak illumination and consequently lower photon absorbance e.g. in the case of 670 nm to avoid Pg state contributions. The 400 nm excitation is therefore off- centre and red-shifted from the absorbance peak. In the previous experimental set-up (Pb illumination) 400 nm illumination also still showed a significant remaining Pb population, but was at this stage assumed to be caused by insufficient spectral overlap with the PVB Pb state. It can now however be conclusively said that both Pg and Pr states have associated blue absorbance peaks, which are rather the S0S2 transitions of the chromophores.

80 | P a g e

Figure 3.11. PCB Pr and PVB Pg state photoconversion properties Full-length Tlr0924 in the Pg/r state was exposed to different visible light ranges from a cold light source fitted with bandpass filters. Illumination time was increased until the photostationary state was reached. Pb was formed maximally by exposure to 540 nm light over the time course of 15 minutes. PCB Pb was selectively formed by 670 nm illumination for 38 minutes A. The photoequilibria reached at different wavelengths varied B. The kinetics of Pb formation and Pg/r decay are shown at appropriate wavelengths C.

81 | P a g e

Assuming that 670 nm illumination of the Pg/r state created “pure” PCB Pb and PVB Pg spectra, these were also tested for their spectral sensitivities in a final experiment (Figure 3.12).

Figure 3.12. Pg state photoconversion properties Full-length Tlr0924 in the PVB Pg/PCB Pb state was exposed to different visible light ranges from a cold light source fitted with bandpass filters. Illumination time was increased until the photostationary state was reached. Pb was formed maximally by exposure to 540 nm light over the time course of 2 minutes. Pr was formed maximally from Pb at 430 nm over the time course of 21 minutes A. The photoequilibria were variable and sometimes slow to establish. Dotted lines indicate spectra after 15 min of light exposure that had not yet reached stationary states B. The kinetics of Pb, Pg and Pr formation and/or decay are shown C.

82 | P a g e

The time course of spectral changes from PVB Pg to Pb and PCB Pb to Pr are shown in Figure 3.12A. PVB Pg was sensitive to 540 nm > 500 nm (> 630 nm > 640 nm) > 450 nm > 420 nm illumination. At 540 nm and 500 nm Pb is formed whereas at 450 nm and 420 nm Pr is formed from PCB Pb while the amplitude of Pg remains constant. Given the spectral overlap of Pg and Pr there hence must have been a small bleach of Pg. 630 nm and 640 nm illuminated samples had not reached the photostationary state yet. Pr formation from PCB Pb occurs in the blue region of the spectrum in the order 430 nm > 420 nm > 400 nm > 450 nm illumination (Figure 3.11B). PVB Pb formation is fastest at 540 nm> 500 nm and slow at 630 and 640 nm, which coincides with the depletion of Pg at 532 nm. However, 400 nm irradiation also initially depletes Pg faster than 640 nm light, again representing excitation of the S0S2 transition. PCB Pb is depleted most efficiently at 430 nm > 420 nm > 450 nm > 400 nm (Figure 3.11C). Coinciding Pr formation follows the same order. With 450 nm illumination the dominant process is the bleach of PCB Pb but a 450 nm induced Pg bleach also occurs. Unlike the Pg state associated ~400 nm S0S2 transition, which overlaps mainly with the Pb state, the Pr state associated ~450 nm S0S2 transition is at the border of both the Pg and the Pb state and will form heterogenous states of the same PVB chromophore.

In comparison, Pb depletion is faster for the PVB than the PCB chromophore. The same appears to hold true for the inactivation reaction. However, the reactions are less comparable due to the overlapping photoproduct states and differences in photons absorbed, based on the lamp profile and filter permeabilities at the respective wavelengths of illumination (Figure 3.3). These experiments do however demonstrate an incredible sensitivity to different wavelengths, which would theoretically permit a very gradual fine- tuning of the photoreceptor output reaction.

83 | P a g e

3.4.5 Chromophore dark reactions

The stability of the three photostates was investigated by incubation in the dark at 25°C for 24 h and showed that several dark reactions could be observed.

The Pb spectrum did not change markedly over the 24 h period (Figure 3.13). There was a small thermal reversion from photochemically inert features in the Pg/r spectral region to the Pb state. How these species reverted to the Pb state is unclear. The inert population was previously shown to depend on the concentration of oxidised Cys499 but spontaneous reduction of this residue is unlikely. It could potentially represent a chromoprotein population with unusual fold or aggregation properties, which forms the second thioether linkage very slowly.

Figure 3.13. Thermal stability of the Pb state

Tlr0924 in the Pb state was incubated in the dark at 25°C for 24 h and absorbance spectra recorded in 30 min intervals. The inset shows photoconversion prior to the experiment (dashed) and after the experiment (solid).

The photoactivity of the protein was compared before and after the 24 h period by converting the Pb state to the Pg and Pg/r states (Figure 3.13 inset). The Pg/r state showed a loss of amplitude in the Pr shoulder after 24 h whereas the Pg absorbance peak was increased. All spectra after 24 h are downshifted in amplitude according to the return of some of the inert population to the photochromic population. The decrease in absorbance of the Pr state in combination with the increase in absorbance of the Pg state is consistent with the conversion of PCB to PVB. When attempting to quantify this conversion, absorbance at 607 nm decreases by 29%. The increase of the Pg state at 532 nm is only 7%. The discrepancy would suggest that either not all PCB turned into PVB and there is another route of decay for the Pr or that overlapping spectral contributions of the Pr state decrease

84 | P a g e the apparent raise of Pg absorbance. A difference in spectral width of the two populations could also be an explanation for the difference in amplitude.

The Pg state when incubated in the dark at 25°C for 24 h showed little change (Figure 3.14A). In the red region there was a small decrease suggesting the reversion of the “inactive” PCB population observed in the Pb state was also occurring. The Pg state showed a small increase suggesting the rate of dark reversion was lower than the rate of PCB to PVB conversion. There was an increase in the Pb feature, which may have been caused by several contributing factors. A minor increase would be expected from the reversion of the “inactive” population. A general baseline drift is also observed. In addition, some dark reversion occurred, which can be demonstrated by generating a difference spectrum of the 24h reading and the final Pg state recording. The photostates were compared before and after the 24 h period by converting Pg to Pb and Pg/r (Figure 3.14A insert). After 24h, there were minor changes to the spectrum of the Pb state with a small decrease in the red and an increase in the green spectral region, which is attributed to oxidation of the Cys499 thiol (Figure 3.8). The Pg state showed an increased amplitude after 24 h demonstrating PCB was converted to PVB. The Pg/r state shows the concomitant decrease of the PCB Pr absorption.

Figure 3.14 Thermal stability of Pg and Pg/r states Tlr0924 Pg A and Pg/r B were incubated in the dark at 25°C for 24 h and spectral changes recorded in 30 min intervals. Insets show photoconversion before (dashed) and after the experiment (solid).

85 | P a g e

The Pg/r state was also incubated at 25°C for 24 h and showed a significant level of dark reversion (Figure 3.14B). There was clear formation of Pb state protein and an overall decrease of Pg and Pr, suggesting that the PCB chromophore is more likely to revert to the Pb state than the PVB chromophore. Comparison of the occupancy of the individual states before and after 24 h shows a decrease in Pr and an increase of Pg implying PCB to PVB chromophore isomerisation had also occurred (Figure 3.14B inset).

These data therefore confirm Pb as the dark state of Tlr0924 in accordance with most CBCRs harbouring a 15Z chromophore isomer in their dark state.26 Pg and Pr on the other hand exhibited significant dark reversion as previously observed for phytochrome photoproduct states with conversion of 5-10% within 12 h.149,251 The dark reversion rate is known to be increased in Cph1 when either truncated or reconstituted with a non- physiological bilin chromophore. Since Pr dark reversion is faster than Pg dark reversion, this may be an indication that PCB-bound Tlr0924 is indeed an artefact of heterologous expression. It was difficult to extract exact dark reversion rates because PCB was observed to isomerise to PVB in parallel. This process has also been previously described to occur non-linearly.162 Furthermore, Cys499 spontaneously oxidised and a photochemically inactive feature was observed to convert to Pb. These reactions are however comparatively minor on the timescales of the spectroscopic studies done subsequently and the photoproducts are considered thermally stable.

3.4.6 Temperature effects on Tlr0924

Thermosynechococcus elongatus is a thermophilic organism with optimal growth at 57°C.242

This reflects in the thermal stability of Tlr0924, which had a denaturing temperature (Tm) of 67°C. What is striking about the Tlr0924 protein is that cold and frozen protein assumes a green colour (Pb state) and when warmed up gradually turns pink (Pg/r state) implying also a temperature sensing function (Figure 3.14).179 Low temperatures might therefore have an inhibitory effect on protein activation.

To characterise the temperature effects on the photoconversion of Tlr0924 further, conversion trials were carried out in the temperature range of 5 to 45 °C. Samples were illuminated for a minimum of 15 minutes to ensure complete turn over. The overall conversion from Pb to Pg/r and vice versa was reproducible across all temperatures and is shown as difference spectra in Figure 3.15B. Only at high temperatures did the amplitude of spectral changes decrease, presumably as a result of protein denaturation caused by the

86 | P a g e long exposure to high temperature. The conversion generally appeared to proceed faster at higher temperature. It was difficult to extract accurate rate constants because of condensation effects on the cuvette at low temperature and evaporation of the sample at high temperatures, which caused scattering of the lamp light. This suggests that low temperature acts inhibitory to photoactivation of Tlr0924, which can however be fully overcome by long durations of light exposure.

Figure 3.15 Temperature effects on Tlr0924 photoconversion A Frozen Tlr0924 assumes a green colour. Defrosting causes a colour change to pink, suggesting Pb might be formed at low temperature and Pg at higher temperature. B Difference spectra recorded between 5 and 45°C in 5°C intervals. Samples were illuminated for at least 15 min at 430 nm or 540 nm to ensure complete turn over.

These data are in conflict with a previous study, which reported temperature effects on the yield of photoproduct formation at equilibrium. However, no specific illumination times are disclosed.179

87 | P a g e

3.4.7 Cryobuffer compatibility

In order to better characterise the photochemical behaviour of the protein, low- temperature studies requiring the buffer to form a glass were planned and are described in results chapters 2 and 3. Buffer additives suitable for low-temperature studies were tested for their compatibility with the protein. Glycerol, sucrose, ethylene glycol and methanol were investigated based on their availability and known properties as effective cryoprotectants. All samples were made up to the same concentration with the indicated concentration of additive and consecutively converted to the Pb and Pg/r state by illumination with 540 nm and 430 nm light for one minute.

Methanol immediately had an adverse effect on both the Pb and Pg/r state (Figure 3.16), decreasing the amplitude of both absorbance peaks and changing spectral features. The Pg/r state in particular was barely present at 30% methanol. The protein also precipitated at concentrations of 30% and above and was spun down prior to recording the absorbance spectrum. Pb and Pg/r spectra are identical above 40% suggesting protein is denatured and has released its chromophore. This clearly coincides with the formation of a new absorbance peak at 365 nm and a broad feature around 600 nm, which fits to the spectrum of free PCB chromophore.

Figure 3.16 Effects of MeOH on Tlr0924 photoconversion MeOH was added to Tlr0924 at different concentrations and the effects on photoconversion measured spectroscopically. The dotted black lines show the absorbance spectrum of free PCB.

88 | P a g e

These adverse effects were much less pronounced for ethylene glycol supplemented samples (Figure 3.17). The Pb state showed reasonable stability, whereas the Pg/r state was gradually prohibited. At 60% ethylene glycol, no Pg/r-like spectrum could be observed. Furthermore, the illumination time had to be increased from 1 minute to 5 minutes for full conversion. The same absorbance feature observed previously with methanol forms with an apparent absorbance peak at 370 nm.

Figure 3.17 Effects of ethylene glycol on Tlr0924 photoconversion

Different concentrations of ethylene glycol were added to Tlr0924 and the effects on photoconversion measured spectroscopically.

Glycerol slightly increased the amplitude of the Pb absorbance peak. The Pg state proved yet again the more sensitive with a ~50% decrease in amplitude at 60% glycerol. A new absorbance feature was also observed in the same position as previously with methanol and ethylene glycol (Figure 3.18).

Figure 3.18 Effects of glycerol on Tlr0924 photoconversion Glycerol was added to Tlr0924 at different concentrations and the effects on photoconversion measured spectroscopically.

89 | P a g e

Sucrose at concentrations up to 50% had very little effect on photoconversion (Figure 3.19). The most prominent change was an increase in absorbance of the Pb 627 nm feature.

Figure 3.19 Effects of sucrose on Tlr0924 photoconversion Sucrose was added to Tlr0924 at different concentrations and the effects on photoconversion measured spectroscopically.

The suitability of the buffer components tested therefore decreased in the order sucrose, glycerol, ethylene glycol, methanol. The last 3 display very similar behaviour with the appearance of a new 365 nm feature and a broad 600 nm peak. These were attributed to cleavage of the chromophore protein linkage based on the spectral similarity with free PCB chromophore. Failure to convert to the Pb state was previously demonstrated to be a consequence of the modification of the Cys499 (Figure 3.9). In this case, the Pg/r state red- shifted visibly upon illumination as the chromophore isomerised but did not form the C10 thioether. There is no red-shift in Figures 3.16-3.18 and Cys modification is therefore not an option. However, the reaction between the electrophilic C10 and methanol has been previously observed252 and would spectrally mimic the effect of a thioether linkage. The photochemically inert Pb state could therefore potentially be assigned to an ether of the chromophore. The decreased reactivity among the cosolvents could be based on the increased length of the carbon backbone and a combination of loss of nucleophilic strength and steric hindrance preventing access to the chromophore pocket. The inert Pb peak is also blue-shifted from the expected 436 nm position to 420 nm. This would correspond to the final intermediate of the activation reaction, because the electrophilic addition is only possible in the Pg 15E isomeric state, which when saturated at C10 represent a “double- linked” E isomer. The position previously assigned to 15E PVB in transient absorbance measurements was 415 nm.180 This reaction does however not appear the most feasible

90 | P a g e and less so with the longer chain alcohols. Other, more likely explanations would revolve around the fact that the cryoprotectants all denatured the protein and the spectral changes would be based on unknown alterations of the chromophore environment and potentially different spectral properties of the chromophore in an alcoholic buffer system. The peak shift of the blue absorbing state can also be attributed to the overlap of the Pb state with the free chromophore peak.

Based on these findings the buffer additives chosen for cryogenic measurements were glycerol and sucrose at the minimum concentration required for glass formation (30% and 48%, respectively). However in combination, they caused low Pg yield again.

91 | P a g e

3.5 Discussion

Tlr0924 was expressed heterologously in E. coli with a dual plasmid expression system to synthesise the holoprotein.245 Chromophore synthesised in the absence of the Tlr0924 apoprotein was unstable in the cell and given the inefficiency of the pET15b plasmid it was demonstrated that a time delay between induction of protein expression and PCB synthesis is crucial. This is presumably a fine balance between synthesis of the three different proteins required, the timing of chromphore production and resources available in the cell. The protein was shown to be unstable under the purification conditions described in the only other study of the full-length protein.179 Imidazole used as eluent was observed to precipitate the protein on the affinity column in a Pg-like state (pink colouring). This may have gone undetected on the talon resin used in the previous studies,179 which has the same colour as the inactive protein. The substitution of imidazole for histidine suggested in this study very likely represents an improvement to the current method.

The protein was fully photoactive permitting detailed characterisation of the Tlr0924 photocycle. As reported in previous studies on the GAF domain,26,162 Tlr0924 was demonstrated to harbour two chomophore populations (phycocyanobilin and phycoviolobilin) by sequential illumination and acid denaturation. PCB initially incorporated during grow-up or reconstitution was therefore converted to PVB by autoisomerase activity of the protein as frequently observed in the class of DXCF CBCRs.162,190 The photoreaction was shown to proceed from a doubly thioether-linked ground state (Pb) to a singly linked photoproduct state, Pg or Pr, by chemical modification of the cysteine residue as previously reported for the GAF domain.26 Furthermore, a potential thioether stretching signal at 2562 cm-1 was shown to be linked to the Pg/r states. This has previously only been demonstrated in one other DXCF family member, TePixJ.190 Blockage of thioether formation also revealed the red-shifted absorbance spectrum of the 15Z intermediates of the inactivation reaction.26,162 The 15E double-linked isomer was possibly recreated by reacting the Pg state chromophore with an alcohol to saturate the C10=C11 double bond. The 15E Pb isomer was therefore tentatively assigned a position at 420 nm. Previously, this state has only been visualised as a difference spectrum in time-resolved spectroscopy.180

This chromophore make-up confers the protein with spectral sensitivity across the entire visible range. Due to many overlapping absorbance peaks, different illumination patterns create different ratios of photoproduct versus dark state, which may present a way of “fine- tuning” activity of the protein. Additionally, two blue absorbing features associated with

92 | P a g e the photoproduct states and probably representing the S0S2 transition are reported for the first time.

It remains however debatable whether PCB is a physiological chromophore. As intriguing as the possibility of this further spectral tuning appears, to date the process of autoisomerisation has only been observed in one direction.162 Additionally, TePixJ purified from cyanobacteria harbours only PVB but contains a mixture of chromophores after heterologous expression.190 Given that PVB formation is a time and light-dependent process, this may also be associated with the different expression conditions. This study demonstrated again that PCB spontaneously isomerises to PVB in Tlr0924 in both the Pb and Pr form. Furthermore, the PCB Pr photoproduct state underwent faster dark reversion than the PVB Pg state, which was previously shown to be the case with non-native chromophores.149,251

In line with T. elongatus being a thermophilic organism, Tlr0924 showed a high thermal stability. Unlike previously reported,179 there was little effect of temperature on the yield of photoconversion under saturating light conditions, suggesting that irradiation is the dominant factor. It was difficult to extract accurate rates for the initial stages of photoconversion, but there appears to be a significant increase at higher temperatures. It seems plausible, that this represents a “shut-down mechanism” of the organism, which ceases to grow at room temperature but remains viable for days.242

Tlr0924 therefore presents itself as a complex chromoprotein, binding two alternate linearised bilin chromophores with overlapping absorbance features.

93 | P a g e

Chapter 4

The photoinitiated reaction pathway of full-length cyanobacteriochrome Tlr0924 monitored over 12 orders of magnitude

Authors: Anna F.E. Hauck†, Samantha J.O. Hardman†, Roger J. Kutta†, Gregory M. Greetham‡, Derren J. Heyes† and Nigel S. Scrutton†

Affiliation: †Manchester Institute of Biotechnology and Photon Science Institute, Faculty of Life Sciences, The University of Manchester, Manchester M13 9PL, United Kingdom and ‡Central Laser Facility, Research Complex at Harwell, Science and Technology Facilities Council, Harwell Oxford, Didcot OX11 0QX, United Kingdom

Published in: The Journal of Biological Chemistry (2014) Vol. 289, No. 25, pp. 17747-17757

94 | P a g e

4.1 Abstract

The coupling of photochemistry to protein chemical and structural change is crucial to biological light-activated signalling mechanisms. This is typified by cyanobacteriochromes (CBCRs), members of the phytochrome superfamily of photoreceptors that exhibit a high degree of spectral diversity, collectively spanning the entire visible spectrum. CBCRs utilise a basic E/Z isomerisation of the bilin chromophore as the primary step in their photocycle, which consists of reversible photoconversion between two photostates. Despite intense interest in these photoreceptors as signal transduction modules, a complete description of light-activated chemical and structural changes has not been reported. The CBCR Tlr0924 contains both phycocyanobilin and phycoviolobilin chromophores, and these two species photoisomerise in parallel via spectrally and kinetically equivalent intermediates before the second step of the photoreaction where the reaction pathways diverge; the loss of a thioether linkage to a conserved cysteine residue occurs and the phycocyanobilin reaction terminates in a red-absorbing state, while the phycoviolobilin reaction proceeds more rapidly to a final green-absorbing state. Here time resolved visible transient absorption spectroscopy (femtosecond to second) has been used, in conjunction with time resolved IR spectroscopy (femtosecond to nanosecond) and cryotrapping techniques, to follow the entire photoconversion of the blue-absorbing states to the green- and red-absorbing states of the full-length form of Tlr0924 CBCR. Our analysis shows that Tlr0924 undergoes an unprecedented long photoreaction that spans from picoseconds to seconds. We show that the thermally driven, long time scale changes are less complex than those reported for the red/far-red photocycles of the related phytochrome photoreceptors.

95 | P a g e

4.2 Introduction

Photon-driven E/Z isomerisation is the elementary chemical reaction which initiates the biological function of a large group of photoreceptors.9 Phytochromes and cyanobacteriochromes (CBCRs) form part of this family and mediate vital photomorphogenic processes in plants and photoadaptive behaviour in microorganisms.138,253 The two families share the very basic unit of light sensing: an open chain bilin chromophore covalently anchored via a thioether linkage to the protein, which is nestled in a GAF (cGMP-specific phosphodiesterase/adenylate cyclase/FhlA protein) domain. The bilin can exist in two different states, which are governed by a reversible photon-driven E/Z isomerisation of the methine bridge between rings C and D (Figure 4.1A).

Figure 4.1 A Structures and B related absorption spectra of the 15Z-PVB'Pb, 15E-PVBPg, 15Z-PCBPb, and 15E-PCBPr chromophores of Tlr0924.The PCB and PVB populations autoisomerise to yield a mixture of the two chromophores, which can be converted between photostates with the relevant wavelength of light.

96 | P a g e

This photocycle has been studied in detail in the structurally and spectrally conserved phytochromes.9,34,81,162,253 The exclusively red/far-red conversion of phytochromes involves an ultrafast photoisomerisation of the bilin cofactor followed by a number of slower, thermally-driven, conformational changes to trigger the signalling response.34 Photoactivity depends on the presence of flanking PAS (Per, Arnt, Sim) and PHY (phytochrome specific) domains.81 In contrast, CBCR GAF domains autonomously undergo efficient, reversible photoconversions in which different modifications to their covalently bound chromophore allow sensitivity to the entire UV/Visible spectrum.10,162 Modulation of the bilin conjugated system is achieved by different tuning mechanisms, such as structural variations in the chromophore and/or breakup of the delocalised π-electron system. However, the photochemical mechanisms that allow this photosensory flexibility are only beginning to be understood and a kinetic description of the complete photocycle of a CBCR is still lacking. Tlr0924 from Thermosynechococcus elongatus is a blue/green photoreceptor that belongs to the DXCF subgroup (containing the Asp-Xaa-Cys-Phe motif) of CBCRs.162 In a process that has been observed previously in similar CBCRs176,190 the intrinsic GAF isomerase activity in Tlr0924 converts the phycocyanobilin (PCB) precursor chromophore to phycoviolobilin (PVB) incompletely on a timescale of days, resulting in a heterogeneous population with a final ratio of around 1:4 (Figure 4.1A). The absorption spectrum is consequently composed of features originating from both the PVB and PCB chromophores (Figure 4.1B). The Tlr0924 photocycle has been studied by visible absorption spectroscopy and circular dichroism and was shown to interconvert between blue-absorbing (Pb) and green-absorbing (Pg) states upon absorption of the relevant wavelength of light.162,179 The large spectral shift cannot be achieved by isomerisation alone, but is brought about by a second thioether linkage formed between the bilin C10 and Cys499 in the protein DXCF motif.26,179,246 When the PVB and PCB chromophores are linked to the protein they become rubin-like structures with a saturated C10, referred to here as PVB' and PCB' respectively. The absorption maxima of the Pb ground states are spectrally identical for PCB' and PVB' in the visible region because the second thioether linkage restricts the photochemically relevant conjugated system to the C and D rings. Blue light absorption will therefore inevitably activate both photoconversions, resulting in green-absorbing PVB photoproduct and red-absorbing (Pr) PCB photoproduct, with their conjugated systems extended to ring B and rings A and B, respectively. The photoproducts still exhibit a significant spectral overlap and green light absorption returns both chromophores to their blue-absorbing states. However, far-red illumination will convert only PCB back to the Pb state. The individual spectral contributions can be isolated by this sequential photoconversion permitting further studies26,190 or the states without the

97 | P a g e second thioether linkage to the protein can be visualised by (permanent) acid denaturation.160,175,190 The ultrafast photoisomerisation reaction dynamics of the GAF domain of Tlr0924 have previously been investigated by exciting the two parallel photoconversions simultaneously, which has allowed the primary photoproducts to be identified.180 In the present work we have used a combination of time-resolved spectroscopies covering fs through to second timescales in the visible spectral region, fs to ns timescales in the IR spectral region, and cryotrapping techniques to characterise all steps in the forward photoreaction (Pb to Pg and Pr) of the full-length protein. The experimental isolation of the PVB and PCB populations allowed a detailed assignment of the intermediates formed. It is shown that the primary photoproducts are converted to the final Pg and Pr states by the removal of the second thioether linkage in a single, thermally driven step on the millisecond to second timescale. This detailed kinetic and spectroscopic characterisation demonstrates that the photoconversions of CBCRs are significantly different to the highly complex photocycles found in the related phytochrome proteins.23,187,254–256

98 | P a g e

4.3 Experimental procedures

4.3.1 Protein expression and purification

The gene for full-length Tlr0924 was synthesised (Genscript Inc.) and cloned into the pBAD- myc/HisB expression plasmid. Full-length Tlr0924 was expressed as a recombinant holoprotein by using a dual plasmid E. coli expression system.34,245 BL21(DE3) containing a PCB biosynthetic expression plasmid (pCOLADuet-1 (Novagen)/HO1, PcyA) for chromophore production was cotransformed with the pBAD-myc/HisB/Tlr0924 plasmid. The protein was purified by a 2-step method employing Nickel affinity chromatography followed by gel filtration in a phosphate based buffer system [100 mM sodium-potassium phosphate, 300 mM NaCl, pH 7] supplemented with 200 mM L-histidine for elution. Pure sample was flash-frozen and stored at -80 °C. The chromophore content was approximately 80% PVB and 20% PCB, as indicated by UV/Vis spectroscopy, and based on published conversion rates this ratio can be assumed to remain constant over the course of the experiments.162

4.3.2 Ultrafast transient absorption spectroscopy

The laser system used in the visible transient absorption measurements consists of a Ti:sapphire amplifier (hybrid Coherent Legend Elite-F-HE) pumped by a Q-switched Nd:YLF laser (Positive light Evolution-30) and seeded by a Ti:sapphire laser (Spectra Physics Mai Tai). The amplifier output (1 kHz repetition rate, 800 nm centre wavelength, ~120 fs pulse duration) was split to generate the pump and probe beams. A noncollinear optical parametric amplifier (Light Conversion TOPAS White) was used to generate the pump beam centred at 435 nm, with a full width at half maximum (FWHM) intensity of ca. 10 nm. Excitation energies of 0.75-1 μJ were used with a beam diameter of around 150 μm, which yielded pump fluences of 4.2-5.7 mJ/cm2. The accessible region of the spectrum was maximised by adjusting the polarisations of pump and probe to be perpendicular, and using a polariser before the detectors to eliminate a large proportion of the scattered pump light. Data collected with a depolarised pump beam yielded kinetics and spectra similar to those shown in Figure 4.2, although the intensity of the negative stimulated emission peak at ~510 nm varied between the samples, thus we assume any polarisation effects will not affect the model derived from these data. The probe beam consisted of a white light

99 | P a g e continuum generated in a rastered CaF2 crystal. The broad band pump-probe transient absorbance spectrometer ‘Helios’ (Ultrafast Systems LLC) had a time resolution of approximately 0.2 ps. Absorbance changes were monitored between 350 and 700 nm with data points collected randomly over the 3 ns time frame. Samples were contained in stirred 2 mm path length quartz cuvettes (absorbance at 535 nm = 0.5). During the measurements the samples were continuously illuminated using a cold light source (Schott KL1500) and the appropriate bandpass filter (Andover Corp). Illumination at 540 nm was used to regenerate the PVB' and PCB' Pb states from their corresponding PVB Pg and PCB Pr states, and 640 nm illumination used to regenerate the PCB' Pb state from the PCB Pr state. Time- resolved IR spectroscopy was carried out at the Ultra facility (CLF, STFC Rutherford Appleton Laboratory, UK), which uses a 10 kHz repetition rate laser and has a time

236 resolution of around 100 fs. Samples were flowed through a 100 μm CaF2 measurement cell and the sample holder rastered to avoid sample damage. In addition, the Pb states of the sample were regenerated by continuous sample illumination with a cold light source as described above. An excitation energy of 0.6 uJ at 435 nm was used, the beam diameter was around 150 μm, yielding a pump fluence of 4.3 mJ/cm2, and the excitation beam was set at the magic angle with respect to the IR probe beam. Data were collected for approximately half an hour per dataset. The spectral resolution was ~3 cm-1. Pixel to wavenumber calibration was performed as described previously.257

4.3.3 Laser flash photolysis

A Q-switched Nd:YAG laser (Brilliant B, Quantel) was used for sample excitation in single shot mode. 435 nm pump pulses were generated via an optical parametric oscillator. The pulse duration was 6-8 ns and energies of 5-30 mJ were used (depending on the set of measurements). The beam diameter was on the order of 1 cm, yielding pump fluences of 6.4-38 mJ/cm2. The detection system (Applied Photophysics Ltd) consisted of a 150 W Xenon Arc lamp and monochromators on either side of the sample holder, placed at right angles to the incident pump beam. Measurements on the ms to s scale were recorded using a photomultiplier tube. For faster timescale measurements the probe beam was pulsed and kinetic traces recorded on a digital oscilloscope (Agilent Technologies, Infiniium, 54830B). The 300-700 nm region was monitored by recording absorption transients in 5 nm steps, each data point being the average of ≥3 transients. The appropriate photostate was regenerated after each shot by illumination with a cold source lamp fitted with the relevant bandpass filters as described above. The sample was contained in a 1 cm path length quartz

100 | P a g e cuvette and maintained at 25 °C by a circulating water bath. The sample had an absorbance of 0.2-0.8 at the excitation wavelength depending on the species excited, samples were frequently replaced, and their quality monitored by UV/Visible Spectroscopy (Varian, Cary 50).

4.3.4 LED flash photolysis

Analogous to the laser flash experiment, which was used to cover the ns to ms dynamics of Tlr0924, a commercial UV/Vis spectrometer (Varian, Cary 50) was used to cover the dynamics in the millisecond to minute dynamics of PCB in Tlr0924. The sample was excited orthogonally by a pulsed high power LED (M455L3, Thorlabs) with a flash of 10 ms duration and 10 mJ pulse energy, which was collimated by an AR-coated aspheric lens (Thorlabs) to yield a pump fluence of around 5 mJ/cm2. A small sample volume of 0.5 ml with an absorbance of ~0.2 at the excitation wavelength was used, allowing excitation of the entire sample volume to avert diffusion effects on the long time scales measured. The absorbance changes were recorded at single wavelengths ranging from 305 to 675 nm in 10 nm steps with a time resolution of 12.5 ms in a time window of 30 s. Illumination at 640 nm was used to regenerate the PCB Pb state from the Pr state between datasets. The LED flash was manually triggered for each time trace and all data sets corrected to the same time zero during data processing. The raw data, consisting of 2400 data points, were reduced to 500 points by averaging according to a logarithmic time axis prior to analysis.

4.3.5 Cryotrapping

Samples were prepared at an absorbance at 435 nm of ~ 1 in cryogenic buffer (phosphate based buffer system [100 mM sodium-potassium phosphate, 300 mM NaCl, pH 7] made using 22% water, 30% glycerol and 48% sucrose) and cooled to 77 K in a cryochamber (Optistat DN liquid nitrogen cryostat, Oxford Instruments Inc.) in the desired state. The sample was then allowed to warm up in 10 K steps up to 327 K, illuminated at 430 nm for 10 minutes at each temperature point, and recooled to 77 K prior to recording a UV/Visible absorbance spectrum (Cary 50, Agilent Technologies). Photoconversion was achieved by illumination with a cold light source as described above.

101 | P a g e

4.3.6 Global analysis

The transient absorption and laser flash photolysis 3D datasets were analysed globally using the open-source software Glotaran.258 This procedure reduces the matrix of time, wavelength, and change in absorbance to one or more exponentially decaying time components, each with a corresponding difference spectrum. These spectral and lifetime components can then be used to identify individual photoproducts of the reactions. The data were fitted with a sequential, unbranched, unidirectional model, which shows the spectral evolution from one component to another, and yielded evolution associated difference spectra (EADS). It was assumed that all states of Tlr0924, including the intermediates in the photoconversion processes, have Gaussian shaped absorption profiles. Using this assumption the EADS were fitted with Gaussian functions in order to identify, and obtain accurate peak positions, of the individual intermediates. The visible and IR ultrafast transient absorption data were fitted from 0.3 and 0.2 ps, respectively, to avoid any contributions from coherent artefacts.259

102 | P a g e

4.4 Results

4.4.1 Ultrafast transient UV/Visible dynamics

Previous measurements have shown that the photoisomerisation of the bilin cofactor occurs on an ultrafast timescale for both phytochromes and the isolated GAF domain of CBCRs.20,167,180,254,256,260–264 Here, ultrafast transient absorption spectroscopy has been used to study the Z to E photoisomerisation process in the full-length Tlr0924 CBCR. Laser pulses at 435 nm were used to induce photoconversion of the 15Z-PVB'Pb and 15Z-PCB'Pb populations to the 15E-PVBPg and 15E-PCBPr populations, respectively. Continuous illumination with green light (540 nm) during data collection returned both of the ‘final’ 15E-PVBPg and 15E-PCBPr populations to their corresponding Pb states. Therefore excitation at 435 nm initiated the photoisomerisation of a mixture of both PCB' and PVB' Pb states. However, as Tlr0924 exists as a heterogeneous population of PVB and PCB in a ratio of approximately 4:1 the majority of the time-dependent absorption changes are dominated by PVB (approximately 80% of the signal). Further separation of the species was made possible by illumination with red light (640 nm) during data collection, which converted only the 15E-PCBPr population to the Pb state, leaving the PVB population in the Pg state. In this case, the majority of signal after excitation at 435 nm originated from PCB only. Using the conditions defined above, ultrafast transient absorption difference spectra were measured for both the predominantly PVB (hereafter termed ‘PVB’) and the predominantly PCB (hereafter termed ‘PCB’) forward reactions (Figures 4.2A and 4.2B). In both cases, there is a bleach of the main ground state absorption band (GSB) at ~436 nm, which is flanked by broad overlapping positive excited state absorption (ESA) signals. We also observed a small ~510 nm stimulated emission (SE) band at very early times, and a small negative feature at ~640 nm, which likely corresponds to the bleach of a previously observed ground state absorption peak attributed to a small amount of inactive or modified protein.179 On the sub-ps to 3 ns timescale monitored in these experiments the magnitude of the GSB is reduced and appears to slightly red-shift. The features in the red region of the spectrum disappear simultaneously and result in a broad positive feature spanning the entire 500-700 nm region at 3 ns. The data collected for the ‘PCB’ sample show very similar spectral features to the ‘PVB’ sample although the SE band at ~510 nm is more pronounced for the ‘PCB’ sample, and appears to be slightly red shifted. This may be due to contamination of the signal by a small amount of PVB Pg to Pb photoconversion, which would result in a bleach feature at 532 nm. The bilin structures of the two chromophores are broadly similar, with

103 | P a g e the conjugated systems being identical in PCB and PVB, hence photoexcitation is likely to result in correspondingly similar excited states and primary photoproducts.

Figure 4.2 Ultrafast transient absorption spectra collected after excitation at 435

nm at selected time points for A ‘PVB’ samples, where Pb states were constantly regenerated with green light, and B ‘PCB’ samples where the Pb states were constantly regenerated with red light. Global analysis of the ultrafast transient absorption data for the C ‘PVB’ and D ‘PCB’ samples yielded 3 evolution

associated difference spectra which sequentially interconvert.

The datasets were analysed globally using a sequential model to give rise to evolution associated difference spectra (EADS), which represent the spectral evolution as a function of sequential exponential time constants rather than actual spectra of populations (Figures 4.2C and 4.2D). In both the ‘PVB’ and ‘PCB’ datasets a good fit was achieved with 3 components and the resulting EADS were very similar for both samples. There are slight differences in the resulting lifetimes. This is almost certainly due to the small amount of signal contamination from the PVB Pg to Pb photoconversion in the ‘PCB’ dataset. The initial spectra, EADS1, show strong GSB and SE features, and a broad ESA, which rapidly decay in ~2 ps to the second spectra, EADS2. These spectra still contain a strong GSB feature and broad ESA, but no obvious SE. In the final component, EADS3, which grows in from EADS2 with a lifetime of ~10 ps, the GSB is still present, although red-shifted by ~5 nm compared to the previous components. This apparent shift is likely to be due to the

104 | P a g e appearance of a new positive feature at ~415 nm, corresponding to an isomerised, but still Cys-bound, intermediate.26,179

4.4.2 Ultrafast transient IR dynamics

Ultrafast time resolved IR spectroscopy was used to probe structural changes in the mid- infrared region of the electromagnetic spectrum after initiation of the photoisomerisation reaction by 435 nm laser pulses. The difference spectra were qualitatively similar for both samples although the amplitude was reduced by approximately a third for the ‘PCB’ sample, reflecting the lower relative concentration of PCB compared to PVB (Figures 4.3A and 4.3B). Consistent with visible ultrafast transient absorbance measurements, global analysis revealed that the data could be decomposed into 3 exponentially decaying species with lifetimes of τ= 3, 22, and ‘infinite’ ps and τ= 2, 17, and ‘infinite’ ps, for the ‘PVB’ and ‘PCB’ samples, respectively (Figures 4.3C and 4.3D).

Figure 4.3 Ultrafast transient IR absorption spectra, collected after excitation at 435 nm, at selected time points for A ‘PVB’ samples, where Pb states were constantly regenerated with green light, and B ‘PCB’ samples where the Pb states were constantly regenerated with red light. Global analysis of the ultrafast transient absorption data for the C ‘PVB’ and D ‘PCB’ samples yielded 3 evolution associated difference spectra which sequentially interconvert.

105 | P a g e

A number of previous studies on the related phytochrome proteins allow confident assignments to be made of the features in the difference spectra shown in Figure 4.3. The largest ground state bleach (negative) features occur at ~1624 and ~1640 cm-1, which correspond to C=C stretches in the C and D, and A and B rings, respectively.20,265 The small negative feature at ~1700 cm-1 is likely to correspond to the bleach of the C=O stretching mode from ring D.266,267 The other significant bleaches at ~1573 cm-1 and ~1467 cm-1 are likely due to C=O and C=N stretches, respectively, both coupled to a N-H rocking motion.268 The main differences between EADS1 and EADS2 are a general loss of intensity across the whole spectral window. Features that can be attributed to the isomerised intermediate can be observed in EADS3 and include the downshift of the D-ring C=O stretch from ~1700 to ~1686 cm-1, implying a more strongly hydrogen bonded environment in the 15E-PVB'Pb, compared to the 15Z-PVB'Pb state. Previous ultrafast IR studies have reported similar downshifts of this bond frequency upon isomerisation in Cph1261 and two bacteriophytochromes,266 but static measurements of phytochromes have reported the opposite effect,255,269 demonstrating the complexity of this family of photoreceptors. The C=N stretch at ~1467 cm-1 may also be downshifted by the isomerisation to form the small positive feature at ~1433 cm-1.268 The changes to the C=C stretches in the 1600-1650 cm-1 region are difficult to deconvolute, but it seems likely that the 1640 cm-1 feature originating from rings A and B will not change, whereas the ~1624 cm-1 feature from rings C and D may shift upon photoisomerisation.

4.4.3 Slow, thermally-driven dynamics

The second part of the Pb to Pg or Pr conversion involves the breakage of a thioether linkage between a cysteine residue and the methine bridge between pyrrole rings B and C.176,179 As this is likely to proceed on much longer timescales than the isomerisation, reaction intermediates were visualised on the sub-μs to s timescale by laser and LED flash photolysis. The Pb ground state was excited with a ~7 ns duration laser flash at 435 nm and time-dependent absorption changes were monitored in 5 nm intervals over the 300 nm to 700 nm wavelength range. Both PVB and PCB photoproducts were returned to the Pb state between laser shots by 540 nm illumination with a cold lamp source. The time-resolved difference absorption spectra were then assembled from these data (Figure 4.4A).

106 | P a g e

Figure 4.4 A Laser flash photolysis spectra after excitation at 435 nm at selected time points for a mixture of PVB and PCB Tlr0924, and B global analysis of the data showing the resulting EADS (black dots) fitted with a sum of Gaussian functions (red line). There are obvious features originating from the 15ZPb states 15E 15E-PVB (blue lines), the Pb states (cyan lines), and the Pg state (green lines). EADS1 converts to EADS2 with a lifetime of 937±1 ms.

After the initial photoisomerisation observed in the ultrafast measurements, no further significant changes in the absorption spectra were detected between 20 ns and ~1 ms (Figures 4.5A and 4.5B); the kinetics and difference spectra consistently show the bleach of the Pb ground states at ~450 nm as well as a small positive feature in the 340 to 390 nm region. On longer timescales there is an apparent blue-shift of the ground state bleach and the formation of a species at ~345 nm, together with the appearance of the final photoproduct at ~530 nm (Figure 4.4A).

107 | P a g e

Figure 4.5 A Laser flash photolysis spectra at selected time points between 0.7

and 450 μs for a mixture of PVB and PCB Tlr0924 after excitation at 435 nm. Laser flash photolysis kinetics recorded in two datasets, at selected wavelength

points between 20 ns and 2 ms for B a mixture of PVB and PCB of Tlr0924, and C PCB only Tlr0924 after excitation at 435 nm.

These changes become increasingly apparent at delay times of 500 ms onwards and are very similar to the static difference spectrum expected for the Pb to Pg (and Pr) photoconversion (dotted line). However, due to the slower reaction rate of the PCB chromophore the final spectrum measured at 3 s does lack some of the definition of the PCB photoproduct in the 600 nm region. Global analysis of the data using a sequential model shown in Figure 4.6B reveals an exponentially decaying component (EADS1) evolving into the final non-decaying component (EADS2) with a lifetime of 0.94 s (rate constant of 1.07 s-1).

108 | P a g e

Figure 4.6 LED flash photolysis spectra at selected time points for A PCB only ‘forward reaction’ after excitation at 455 nm and B a mixture of PVB and PCB ‘reverse’ photoconversion from Pg to Pb, after excitation at 530nm.

EADS1 can be fitted to the sum of three Gaussian functions (red dashed line): negative peaks at 322 and 436 nm (blue lines), representing the bleach of the 15ZPb ground state, and a positive peak 415 nm (cyan line), representing the blue-shifted, 15EPb, intermediate. EADS2 can be fitted to the sum of four Gaussian functions (red dashed line); the negative peaks at 322 and 436 nm remain, whilst the positive feature at 415 nm has been replaced by peaks at 340 and 532 nm (green lines), corresponding to the final 15E-PVBPg conformation. These peak assignments are entirely consistent with previous studies on the GAF domain of Tlr0924.26,162,179 The PCB population is not significantly distinct enough from the PVB population to be resolved here. In order to investigate the properties of the PCB forward photoreaction in more detail, the pure PCB' Pb state was regenerated by illuminating the 15E isomers (Pg and Pr) at 640 nm between each dataset collected. This converted only the PCB Pr population to the PCB' Pb population whilst the PVB population remained in the Pg state. As with the mixture of PVB and PCB, after the initial photoisomerisation observed in the ultrafast measurements there were no significant spectral changes between 20 ns and tens of ms (Figure 4.5C).

109 | P a g e

Due to the slow reaction dynamics and low concentrations of the ‘PCB’ sample LED flash photolysis was used to monitor the dynamic processes. Samples were excited with a 10 ms, 455 nm LED flash in a UV/Vis spectrophotometer to give the raw difference spectra shown in Figure 4.6A. However, it is apparent that there were overlapping signals in the dataset, which originated from both the forward and reverse photoreactions of the PVB chromophore. Subtraction of scaled datasets collected on the same wavelength and time domains for the photoconversion of PVB from Pb to Pg (shown in Figure 4.4A) and for the photoconversion of PVB from Pg to Pb (shown in Figure 4.6B) resulted in the difference spectra shown in Figure 4.7A, which should include components only from the PCB Pb to Pr photoconversion. Initial spectra show ground state bleaches at ~325, 450, and 640nm, and these spectra evolve over several seconds to a spectrum with a large bleach at ~435 nm, and a positive feature at ~590 nm. Global analysis of the data using a sequential model shown in Figure 4.7B reveals an exponentially decaying component (EADS1) evolving into the final non-decaying component (EADS2) with a lifetime of 3.1 s (rate constant of 0.33 s-1). EADS1 can be fitted to the sum of five Gaussian functions (red dashed line): in common with the data collected from a mixture of PVB and PCB there are negative peaks at 322 and 436 nm (blue lines), representing the bleach of the 15ZPb ground state, and a positive peak at 415 nm (cyan line), representing the blue-shifted 15EPb intermediate. There is also a very small amount of the 588 nm product peak, and an additional negative peak centred at 640 nm, which correlates well with the small negative feature at ~640 nm observed in the ultrafast measurements and may correspond to the bleach of a previously observed ground state absorption peak that was attributed to a small amount of inactive or modified protein.179 The negative feature at 640 nm remains in EADS2, which can be fitted to the sum of five Gaussian functions (red dashed line), with FWHM fixed from EADS1. In addition to the 640 nm feature the negative peaks at 322 and 436 nm remain, whilst the final 15EPr photoproduct is represented by peaks at 360 and 588 nm (dark red lines). There is also a small amount of the PVB Pg product, represented by a positive feature at 532 nm visible, although this is likely to be due to the fallibility of the method of subtraction of the scaled datasets from the PVB forward and reverse reaction.

110 | P a g e

Figure 4.7 A LED flash photolysis spectra at selected time points for PCB only Tlr0924 with overlapping PVB forward and reverse reactions subtracted, and B global analysis of the data showing resulting EADS (black dots) fitted with a sum of Gaussian functions (red line). There are obvious features originating from the 15Z-PCB'Pb state (blue lines), the 15E-PCB'Pb state (cyan lines), the 15E-PCBPr state (dark red lines), and the 15E-PVBPg state. There is an additional feature at 640nm (purple line) originating from inactive or modified protein. EADS1 converts to EADS2 with a lifetime of 3.1±0.1 s.

As with the mixture of PVB and PCB, the peak assignments are entirely consistent with previous studies on the GAF domain of Tlr0924.26,162,179

111 | P a g e

4.4.4 Cryotrapping of intermediate states

In order to confirm the intermediates observed by the time-resolved spectroscopy and to investigate the associated energetic or thermal barriers to these reaction steps in the Tlr0924 photocycle the photoconversion of the Pb to the Pg and Pr states was monitored by illuminating samples with blue light at temperatures ranging from 77 to 327 K. Absorbance difference spectra were recorded at 77 K by subtraction of the initial Pb ground state absorbance spectrum (Figure 4.8A).

Figure 4.8. Low-temperature stabilisation of reaction intermediates. A. Difference spectra, and B. change in absorption at 390 (cyan), 435 (blue), 535 (green), and 590 (red) nm after illumination at a range of temperatures between 77 and 327 K.

112 | P a g e

An initial light-dependent reaction that can be observed at temperatures below 200 K involves a bleach of the Pb ground state band at approximately 450 nm and the appearance of a new absorbance band at ~390 nm. These spectral features represent the primary 15EPb photoproduct that is formed upon photoisomerisation and are identical to those observed in the transient absorption measurements (see Figures 4.2, 4.4, and 4.7). The temperature dependence of this photochemical step was obtained by plotting the absorbance increase at 390 nm against the temperature at which the sample was illuminated. This reveals that it has reached completion at temperatures below 200 K (Figure 4.8B). After formation of the primary 15EPb photoproduct there is a further small decrease in the main Pb absorbance band at approximately 435 nm which can be observed at temperatures above 200 K, together with a shift of the small positive absorbance band at ~390 nm to ~345 nm (Figure 4.8A). This is accompanied by a simultaneous increase in the absorbance peaks at ~530 nm and ~590 nm (Figure 4.8B), which represents the Pg and Pr states of the protein, respectively.

113 | P a g e

4.5 Discussion

The combination of ultrafast visible and IR transient absorption data leads us to suggest the scheme as shown in Figure 4.9 for the photoisomerisation of the 15Z PVB' and PCB' species.

Figure 4.9.Scheme showing suggested ground and excited state energy surfaces and the processes, which occur after photoexcitation. FC (Franck-Condon region), SE (stimulated emission)

Immediately following photoexcitation to the Franck-Condon region very fast relaxation occurs within the time resolution of our detection method (~0.2 ps). From the Franck- Condon region an energetically excited region of the S1 potential energy surface is populated, from which SE can occur to the ground state. In ~2 ps this state has relaxed to the energy minima of the S1 surface, and on a timescale of ~10 ps the system relaxes to either the 15Z or 15E isomers. After this photoisomerisation step the 15EPb intermediate can then progress to the final 15E-PVBPg and 15E-PCBPr states. This scheme is less complex than, but similar to, those proposed in a previous study on the GAF domain of Tlr0924180 and for the corresponding isomerisation process in the related Cph1 phytochrome, both of which are suggested to involve multiple intermediates on both excited state and ground state energy surfaces.20,21,270 Previous studies on CBCRs involved only the GAF domain and it is possible that the full-length protein restricts the number of accessible states, resulting in the suggested simple mechanism. This would be a contrast to phytochrome systems, where

114 | P a g e even the full-length protein demonstrates multiple intermediates in the reaction pathway.138,260,271 The difference between the systems lies in the thioether linkage to the protein, which is lacking in phytochromes and may provide added stabilisation to the CBCR reaction intermediates. It may be that after the loss of the cysteine linkage further rearrangements occur, but if these processes are fast (sub-μs) they would not be resolved by the methods used here. The slower, thermally-driven dynamics monitored by flash photolysis measurements also present a less complex picture than that seen in many CBCR187 and phytochrome271 systems. It may be that there are further, spectrally identical, intermediates that we cannot distinguish using the techniques used here, however this seems unlikely as spectral shifts between intermediates are usually >10 nm in such systems.187,271 There appears to be only one step between the isomerisation and the breaking of the Cys bond, although unlike the isomerisation reaction the rate of this step appears to depend upon the properties of the chromophore. The PVB conversion completes nearly 3 times faster than that of PCB. Generally, proteins undergo a dynamic transition, termed the ‘glass transition’, at approximately 200 K, below which any large- scale structural changes in the protein become frozen out.34,223,272–274 As our cryotrapping measurements show, formation of the primary photoproduct can still proceed below 200 K so the photoisomerisation reaction is only likely to involve structural changes within the bilin molecule itself or minor, localised protein adjustment around the bilin cofactor. This is similar to the photoisomerisation reaction observed for the related Cph1 phytochrome.34 After photoisomerisation both PVB and PCB reaction pathways, which involve the breakage of the thioether linkage, proceed in a single step at temperatures well above the ‘glass transition’ temperature. Consequently, it is likely that there is a role for large-scale protein motions that become frozen out below the ‘glass transition’ temperature in order to eliminate the thioether linkage in Tlr0924. It is known that conformational changes and/or domain movements are crucial in the photocycle of the related Cph1 phytochrome. The formation of photoproduct occurs at much higher temperatures in Tlr0924 compared to Cph1,34 which is likely to represent the additional energy required to break the thioether linkage.

115 | P a g e

A detailed kinetic and spectroscopic analysis across a range of temperatures and timescales has provided a comprehensive understanding of the complete forward reaction dynamics during the photocycle of a full-length CBCR for the first time. By using a combination of transient spectroscopy techniques and cryotrapping measurements we propose the reaction pathway shown in Figure 4.10 for the photoconversion of the Pb state to the Pg and Pr states of the protein.

Figure 4.10. Suggested forward reaction pathway and lifetimes for PVB and PCB in Tlr0924. After photoexcitation both 15Z-PVB'Pb and 15Z-PCB'Pb relax to an excited state

minima within 2 ps, from which isomerisation to the 15E-PVB'Pb and 15E-PCB'Pb states can occur with a lifetime of ~10 ps. At this point the photoreactions diverge with the 15E-PVBPg and 15E-PCBPr states being formed with a lifetimes of ~0.9 and ~3.1 s, respectively.

116 | P a g e

The PVB' and PCB' chromophores in Tlr0924 isomerise on a ps timescale, similar to that observed for other CBCR and phytochrome proteins.20,21,254,270 Subsequently, the primary 15E PVB' and PCB' Pb photoproducts are converted to the final PVB Pg and PCB Pr states of the protein in slower and distinct, single-step reactions that involve the removal of a thioether linkage to a conserved cysteine residue in the protein. However, it is likely that some degree of protein conformational change is required to facilitate the removal of the thioether linkage in Tlr0924. This proposed reaction scheme is less complex than the multiple steps (on microsecond to millisecond timescales) that follow photoisomerisation in the phytochrome proteins, which are thought to involve a series of structural changes in the protein34 and at least 2 metastable intermediates.255,268 Studies of the GAF domain of CBCRs have also suggested multiple intermediates after photoisomerisation in each photoconversion step.23,187 The difference in conclusions drawn between the GAF domain only and full-length CBCRs may be explained as the full-length protein limiting the number of possible conformations for the active site. The cryotrapping measurements reported here demonstrate that protein conformational changes are involved in the loss of the second thioether linkage, giving credence to the theory that the overall protein structure affects the reaction dynamics. It is unclear whether, and to what extent, the Tlr0924 protein harbours PVB and PCB chromophores in the cell. Studies have suggested that the isomerisation of PVB to PCB is a fundamental property of the GAF domain,175 and there appears to be no difference between protein coexpressed in vitro (as we have done) or reconstituted with the chromophore after expression.162 If the naturally occurring protein does have a heterogeneous chromophore population it may be that PCB is isomerised to PVB on the basis of faster or more efficient reaction kinetics with the remainder of the PCB population providing additional spectral coverage.

In summary, this is the first analysis of the complete forward photoconversion of a full- length cyanobacteriochrome, which suggests an unprecedentedly slow and relatively simple reaction scheme for the thermally activated protein chemical/structural changes that follow the initial photochemical events.

117 | P a g e

Chapter 5

Comprehensive analysis of the green to blue photoconversion of full-length cyanobacteriochrome Tlr0924

Authors: Samantha J.O. Hardman†*, Anna F.E. Hauck†*, Ian P. Clak‡, Derren J. Heyes† and Nigel S. Scrutton†

Affiliation: †Manchester Institute of Biotechnology and Photon Science Institute, Faculty of Life Sciences, The University of Manchester, Manchester M13 9PL, United Kingdom and ‡Central Laser Facility, Research Complex at Harwell, Science and Technology Facilities Council, Harwell Oxford, Didcot OX11 0QX, United Kingdom

*These authors contributed equally

Published in: Biophysical Journal (2014) Vol. 107, pp. 2195–2203

118 | P a g e

5.1 Abstract

Cyanobacteriochromes are members of the phytochrome superfamily of photoreceptors and are central in biological light-activated signalling mechanisms. These photoreceptors are known to reversibly convert between two states in a photoinitiated process which involves a basic E/Z isomerisation of the bilin chromophore, and in certain cases the breakage of a thioether linkage to a conserved cysteine residue in the bulk protein structure. The exact details and timescales of the reactions involved in these photoconversions have not been conclusively shown. The cyanobacteriochrome Tlr0924 contains phycocyanobilin and phycoviolobilin chromophores, both of which photoconvert between two species, blue-absorbing and green-absorbing, and blue-absorbing and red- absorbing, respectively. Here we have followed the complete green to blue photoconversion process of the phycoviolobilin chromophore in the full-length form of Tlr0924 over timescales ranging from femtoseconds to seconds. Using a combination of time-resolved visible and mid-IR transient absorption spectroscopy and cryotrapping techniques we have shown that after photoisomerisation, which occurs with a lifetime of 3.6 ps, the phycoviolobilin twists, or distorts slightly with a lifetime of 5.3 μs. The final step, the formation of the thioether linkage with the protein occurs with a lifetime of 23.6 ms.

As an amendment to this thesis, laser flash photolysis data for the PCB reverse reaction is also included as a supplementary note.

119 | P a g e

5.2 Introduction

The photochemically versatile class of proteins known as cyanobacteriochromes (CBCRs) mediate a variety of molecular outputs via the reversible E/Z isomerisation of a sensory bilin chromophore in their cyanobacterial host organism.10 The general protein structure is divided into domains, always including the photosensory chromophore binding GAF (cGMP- specific phosphodiesterase/adenylate cyclase/FhlA protein) domain in the N-terminus and an output domain (e.g. histidine kinase or GGDEF domain) in the C-terminus.10,183 There is great functional versatility in different subgroups of the protein class. Modifications of the central chromophore also permit a collective spanning of the entire UV/Vis spectrum.10,163,171 CBCRs have attracted a great deal of attention as novel photoreceptors. Variations in domain architecture and chromophore structure permit creation of proteins of a specific function for applications such as fluorescence biomarking,181 optogenetics,240 and the production of biofuels.275 At the heart of these developments is a thorough understanding of the protein as a photoreceptor.

Photochemical reactions of biological molecules are frequently fast and efficient. CBCRs and the related phytochromes and bacteriophytochromes have increasingly been characterised on femtosecond to nanosecond timescales.34,254,256,260,264,276 In general terms, both the ‘forward’ Z/E isomerisation and ‘reverse’ E/Z isomerisation of the chromophore usually occur within several hundred ps.185,254,260 This induces further structural changes, manifested in the form of mostly unknown intermediate structures,176,187,267 ultimately leading to protein conformational changes to activate the output domain. The CBCR Tlr0924 from Thermosynechococcus elongatus is a photoreceptor from the DXCF subgroup (containing the Asp-Xaa-Cys-Phe motif) of CBCRs which has been the focus of a number of studies.23,179,180,277 Heterologously expressed Tlr0924 incorporates a red-absorbing (Pr) phycocyanobilin (PCB) chromophore through covalent linkage between Cys527 and the ethylidene side chain of ring A (Figure 5.1).179 On a timescale of days autoisomerase activity of the GAF domain subsequently converts ~80% of the PCB population to a phycoviolobilin (PVB) population.162,190 PVB is saturated at the C5 position, between rings A and B, and hence the π conjugated system is restricted to rings B-D to yield a green-absorbing species (Pg). The 15Z isomers of PVB and PCB can form a second covalent linkage between Cys499 and the C10 position, between rings B and C, to form species PVB' and PCB', respectively. This second Cys linkage shortens the π conjugation to rings C and D only in both chromophores, yielding identical blue-absorbing dark states (Pb). After photoconversion to

120 | P a g e the Pg or Pr states reversion to the dark Pb states typically occurs slowly, over several hours.162,171 The combination of chromophores means that Tlr0924 is photosensitive to virtually the entire visible spectrum (Figure 5.1). Characterisation of the various photostates involved has been carried out using acid denaturation and selective sequential photoconversion.162,179 If Tlr0924 containing a mixture of PVB' and PCB' is illuminated with blue light, conversion will occur from both (spectrally identical in the visible region) Pb states. When the sample is then illuminated with red light only the PCB Pr state will convert back to the Pb state. When green light is used both the PVB and PCB populations will be returned to their respective Pb states.

Figure 5.1 Structures and absorption spectra of the PCB and PVB photostates. The PCB and PVB chromophores can interconvert by autoisomerisation, the 15ZPb and 15E 15E Pg and Pr states can reversibly photoconvert.

A description of the complete ‘forward’ photoconversion for the full-length protein has recently been published.277 The only previous time-resolved study of the reverse photoreaction of Tlr0924 was of the fs-ns photoisomerisation dynamics of the GAF domain protein.180 Here we use a combination of time-resolved spectroscopies covering fs to s timescales, and cryotrapping measurements to comprehensively characterise the ‘reverse’ photoreaction of the dominant PVB chromophore in full-length Tlr0924.

121 | P a g e

5.3 Materials and methods

5.3.1 Protein expression and purification

Full-length Tlr0924 was expressed and purified as described previously.277 The sample was dissolved in a phosphate based buffer system (100 mM sodium-potassium phosphate, 300 mM NaCl, pH 7). UV/Vis spectroscopy was used to determine the relative PVB:PCB chromophore ratio as approximately 80% PVB and 20% PCB. This ratio was assumed to remain constant over the course of the experiments.81

5.3.2 Ultrafast transient absorption spectroscopy

The 1 kHz repetition rate laser system used to pump the broad band pump-probe visible transient absorbance spectrometer ‘Helios’ (Ultrafast Systems LLC, Sarasota, FL) has been previously described.277 The pump beam was centred at 530 nm with a full width at half maximum intensity of ~50 nm. Excitation energies of 0.6 μJ were used which yielded pump fluences of 3.4 mJ/cm2. To maximise the accessible region of the spectrum the polarisation of pump and probe were adjusted to be perpendicular, and a polariser before the detectors was used to eliminate a large proportion of the scattered pump light. Data collected with a depolarised pump beam (see Figure S5.1 in the Supporting Material) yielded kinetics and spectra similar to those shown in Figure 5.2, thus we assume any polarisation effects will not affect the model derived from these data. The time resolution of the experiment was ~0.2 ps, data points were collected randomly over the 3 ns time frame. Samples were contained in stirred 2 mm path length quartz cuvettes (absorbance at 535 nm = 0.5). During the measurements the samples were continuously illuminated through the appropriate bandpass filter (Andover Corp, Salem, NH) using a cold light source (KL1500, Schott, Stafford, UK). Illumination at 435 nm was used to regenerate the PVB and PCB Pr and Pg states from their corresponding Pb states, and simultaneous 640 nm illumination used to regenerate the PCB Pb state from the Pr state. This left only the PVB Pg state to be excited by the pump laser.

Time-resolved IR spectroscopy was carried out at the Ultra facility (CLF, STFC Rutherford Appleton Laboratory, UK), which uses a 10 kHz repetition rate laser and has a time

236 resolution of around 100 fs. Samples dissolved in D2O based phosphate buffer (at the same concentration and equivalent pD as the H2O buffer used in the visible measurements)

122 | P a g e were flowed through a 100 µm path length CaF2 measurement cell (Harrick Scientific, Pleasantville, NY) and the sample holder was rastered in the two dimensions orthogonal to the pump and probe beams to avoid sample damage. The sample was concentrated by a factor of around 20 times compared to that used in the visible measurements so that the absorbance at 535 nm was ~0.5. An excitation energy of 1 uJ at 530 nm was used, the pump beam diameter was around 150 µm, yielding a fluence of 7.2 mJ/cm2. The excitation beam was set at the magic angle with respect to the IR probe beam. The Pg state of the sample was regenerated by continuous sample illumination with a cold light source as described above. The spectral resolution was ~3 cm-1. Pixel to wavenumber calibration was performed as described previously.257

5.3.3 Laser flash photolysis

The laser flash photolysis experimental set up has been described in detail elsewhere.277 For measurements on sub-ms timescales the probe beam was pulsed and kinetic traces recorded on a digital oscilloscope (Infiniium, 54830B, Agilent Technologies, Santa Clara, CA). Measurements on longer timescales were recorded using a photomultiplier tube. The 532 nm pump pulse duration was 6-8 ns and energies of 26–100 mJ were used (depending on the set of measurements). The beam diameter was on the order of 1 cm, yielding pump fluences of 33–128 mJ/cm2. The 300–700 nm region was monitored by recording absorption transients in 5 nm steps, each data point being the average of ≥3 transients. The PVB Pg photostate was regenerated after each shot by illumination with a cold source lamp fitted with the relevant bandpass filters. Illumination at 435 nm was used to regenerate the PVB and PCB Pr and Pg states from their corresponding Pb states, and this was followed by 640 nm illumination which regenerated the PCB' Pb state from the Pr state. This left only the PVB Pg state to be excited by the pump laser. The sample was contained in a 1 cm pathlength quartz cuvette and maintained at 25°C by a circulating water bath. The sample had an absorbance of 0.5 at the excitation wavelength. Samples were frequently replaced and their quality monitored by UV/Visible spectroscopy (Cary 50, Agilent Technologies).

5.3.4 Cryotrapping

Samples were prepared in the PVB Pg photostate as described above, with an absorbance at 435 nm of ~ 1 in cryogenic buffer (100 mM sodium-potassium phosphate, 300 mM NaCl, pH 7 aqueous buffer system made using 30% glycerol and 48% sucrose) and cooled to 99 K

123 | P a g e in a cryochamber (Optistat DN liquid nitrogen cryostat, Oxford Instruments Inc., Abingdon, UK). After recording a UV/Visible reference spectrum (Cary 50, Agilent Technologies) at 99 K the sample was illuminated at 127 K for 10 minutes before being cooled back to 99 K. The sample was then warmed up to 297 K in 10 K steps, at each point the sample equilibrated for ~10 minutes before the sample was cooled to 99 K and a UV/Visible absorbance spectrum recorded. A temperature of 127 K was chosen as the illumination temperature because this was the point at which the ground state bleach feature reached a maximum intensity in a set of benchmarking experiments where the sample was warmed from 77 K in 10 K steps, illuminating at each temperature for 10 minutes before cooling to 77 K to record the spectrum (see Figure S5.2 in the Supporting Material).

5.3.5 Global analysis

The transient absorption and laser flash photolysis 3D datasets were analysed globally using the open-source software Glotaran.258 This procedure reduces the matrix of time, wavelength, and change in absorbance to one or more exponentially decaying time components, each with a corresponding difference spectrum. The visible and IR ultrafast transient absorption data were fitted from 0.3 and 0.2 ps, respectively to avoid any contributions from coherent artifacts.259

124 | P a g e

5.4 Results and discussion

5.4.1 Ultrafast transient absorption

To isolate the initial photoreaction of the PVB chromophore in Tlr0924 samples were illuminated simultaneously with constant blue and red light and excited with a ~530 nm pulsed laser as described in the methods section. The autoisomerisation of chromophores in Tlr0924, which occurs on a timescale of days, results in an approximate PVB:PCB ratio of 4:1.190 The comparatively small population of the PCB chromophore combined with the selective wavelength constant illumination allows almost complete isolation of the PVB photodynamics. The ultrafast transient absorption (TA) data collected for the initial steps of the PVB photoconversion are shown in Figure 5.2A. Within the time resolution of the experiment (0.2 ps) two major spectral features appear: a large negative feature at ~530 nm, which can be assigned to the ground state bleach (GSB) of the Pg state, and a positive feature at ~670 nm, which can be assigned to excited state absorption (ESA) of the Pg state. Within 5 ps these features have decayed almost completely, although they are all still distinguishable and a positive feature at ~550 nm appears (see Figure S5.3 in the Supporting Material). Global analysis was carried out to more accurately define the spectral intermediates and the timescales of interconversion between intermediate states (Figure 5.2B). A model with 4 sequentially converting spectral components, the evolution associated difference spectra (EADS), was found to be a good fit to the data (see Figure S5.4 in the Supporting Material) and broadly agreed with literature values for GAF domain only Tlr0924.180 There was also an obvious contribution from a negative feature at ~645 nm (see Figure S5.5 in the Supporting Material) which has previously been ascribed to inactive or modified protein.179,277

125 | P a g e

Figure 5.2 Ultrafast visible transient absorption spectra at selected time points A,

and the EADS resulting from a sequential global analysis of the data B. The sample was constantly illuminated during data collection with blue and red light so the

signals primarily originate from the Pg to Pb photoconversion of the PVB chromophore.

The ultrafast IR TA data were collected over the same time range, using the same time steps as were used for the visible TA experiments. The data, shown in Figure 5.3A, display a relatively simple picture. There are 3 major negative features at 1402, 1606, and 1685 cm-1, but no immediately apparent positive features. Similar phytochromes have been extensively studied in the mid-IR region, allowing assignment of the 1606 and 1685 cm-1 bleaches as C=C stretches in rings A and B,20,278 and C=O stretches in ring D,261,266 respectively. Global analysis of the dataset using a basic sequential model yields time constants comparable to those found from the visible TA data (Figure 5.3B). One fewer component is required to fit the data, and the resulting time constants are 3.6, 194, and ‘infinite’ ps.

126 | P a g e

Figure 5.3 Ultrafast IR transient absorption spectra at selected time points A, and the EADS resulting from a sequential global analysis of the data B. The sample was

constantly illuminated during data collection with blue and red light so the signals primarily originate from the Pg to Pb photoconversion of the PVB chromophore.

To more accurately model the ultrafast processes that occur upon photoisomerisation a more complex global analysis was performed (Figure 5.4) to produce species associated difference spectra (SADS). The model was selected after many iterations because it was the model which not only isolated the negative feature at ~645 nm, but also resulted in very comparable lifetimes between the visible and IR datasets. In this model SADS1 evolves into non-decaying component SADS2. Independently and in parallel to this process SADS3, representing the inactive protein, decays back to the ground state. In the case of the visible TA data an additional lifetime representing SADS1 decaying to the ground state was also fitted. The lifetimes produced by this model correlate extremely well between the visible and IR TA datasets. The SADS1 to SADS2 conversion lifetime was 3.6±0.1 ps in both datasets. The lifetime of SADS3 was fitted as 185±2 ps and 194±2 ps for the visible and IR TA data, respectively. In the analysis of the visible TA SADS1 was found to relax to the ground state with a lifetime of 0.6±0.1 ps. This component was not resolved in the IR TA data, likely due to the similarity of the excited state and ground state vibrational spectra, with the poor signal-to-noise ratio of these data playing a role. The poor signal-to-noise

127 | P a g e ratio was due to both the low volume sample and the air bubbles in the sample, unavoidable due to the high flow rate necessary to avoid sample damage during the measurements.

Figure 5.4 Global analysis of the ultrafast visible A-C and IR D-F transient absorption data showing resulting SADS (black). The visible TA SADS are fitted with a sum of Gaussian functions (dashed red). The features of SADS1 A and SADS2 B are assigned as: GSB of the PVB Pg state at 532 nm (green), ESA features at 440, 496, and 663 nm (lilac, cyan and orange), intermediate states at 550, 604, and 648 nm (pink), and pump scatter at 525 nm (grey). SADS3, representing inactive or modified protein has additional peaks at 641 nm (dark green) and 581 nm (purple).

A combination of Gaussian peaks was fitted to the SADS resulting from the visible TA data (Figure 5.4A-C). Although it has been shown that bilin chromophores do not have pure Gaussian lineshapes, the use of these functions can provide useful insight into which species are present in each spectrum.279,280 SADS1 is very similar in shape to the difference spectra collected at 0.3 ps after excitation. Features corresponding to the PVB Pg GSB centred at 532 nm, as well as ESA features at 440, 496, and 663 nm are clearly defined. The lack of any other negative features in this spectrum is consistent with excitation of only the Pg state, not of any intermediate states that may be present after 1 ms (the separation of pulses in the 1 kHz laser system). These features remain in SADS2 which displays an additional positive feature, which we ascribe to the first reaction intermediate at 555 nm. There are also less intense components at 525, 604, and 648 nm, which may be ‘real’

128 | P a g e features, but are more likely to be artefacts from scattered pump light, residual signal from the ‘inactive’ protein, and the very low signal levels. SADS3 decays in parallel to the SADS1 to SADS2 conversion and has features similar, although not identical, to SADS1 and SADS2 at 407, 438, 483, 554, 581 and 685 nm. The most significant difference is the large negative feature at 641 nm, corresponding to inactive or modified protein as observed previously.277 This feature does not correspond to any known state of PCB or PVB in Tlr0924,162 and various other PCB containing CBCRs show features in this spectral region,184,187 suggesting that it is a variation in the GAF domain which produces this red-shifted, inactive population of the protein. This population may account for some of the spectral density above 600 nm in the Pb absorption spectrum shown in Figure 5.1. The SADS resulting from analysis of the IR TA data support the species assignments of the visible TA analysis. SADS1 displays the 3 major bleach peaks observed in the raw data. In addition to the features observed in SADS1 there is a small, but significant positive feature at ~1711 cm-1 in SADS2. This correlates well with ultrafast IR TA measurements on the forward photoconversion of PVB' from the Pb state where, shortly after photoexcitation a downshift was observed in the same C=O stretching feature, from ~1700 to ~1686 cm-1, which is assigned to the isomerisation reaction.277 In parallel to the SADS1 to SADS2 conversion, SADS3 decays with a lifetime of 194 ps. This spectrum confirms the assignment of this independently decaying component as a structurally modified version of the protein. While SADS1 and SADS2 do appear similar SADS3 completely lacks the bleach at ~1402 cm-1, and the C=O stretching feature at ~1686 cm-1 in SADS1 appears to have downshifted to ~1680 cm-1.

5.4.2 Nanosecond to millisecond transient absorption

Laser flash photolysis was used to monitor the slower processes occurring in the photoinitiated reaction. By illuminating the entire sample with blue light, to produce the Pg and Pr states, then illuminating with red light it was possible to completely isolate the Pg species. To cover the ns to ms time range two sets of measurements were performed, from 20 ns to 8 μs, and from ~1 μs to 450 μs. From the data shown in Figure 5.5 (and Figure S5.6 in the Supporting Material) it is evident that there are both spectral and kinetic changes in the PVB data over these timescales.

129 | P a g e

Figure 5.5 Transient absorption data of PVB Tlr0924 after excitation at 532 nm collected over 1 – 450 μs A and the EADS resulting from global analysis (black dots) fitted with a sum of Gaussian functions (dashed red) B-D. The features are assigned as: GSB of the PVB Pg state at 532 and 340 nm (green), the first intermediate at 555 and 332 nm (pink), and the second intermediate at 564 and 335 nm (orange). Insets show expanded 600 – 700 nm region containing 650 nm intermediate (maroon).

130 | P a g e

The sub-μs dataset did not show any additional features so the 1 – 450 μs dataset was analysed using a sequential model (Figure 5.5B-D). Three components were needed to completely describe the data. The first component, EADS1, is qualitatively similar to the SADS2 derived from the ultrafast visible TA measurements (Figure 5.4B). The laser flash photolysis measurements extend further into the UV, allowing further characterisation of the various intermediates. The significant features present in both difference spectra are the ground state bleaches at 532 (and 340) nm, and the intermediate with a strong absorbance at 555 (and 332) nm. The less intense features do appear to be slightly different: 440, 496, 604 and 663 nm in the ultrafast data compared with 442 and 650 nm in the flash photolysis data. Any differences can be attributed to differences in the experimental techniques. Flash photolysis data will show all changes compared to a ‘true’ pre-excitation signal. In contrast the ultrafast visible and IR pump-probe measurements use a 1 or 10 kHz laser respectively, so data show changes compared to the sample 1 or 0.1 ms after excitation. EADS1 decays into EADS2 with a lifetime of 5.3 μs, and EADS3 grows in from EADS2 with a lifetime of 387 μs. The major spectral features of EADS2 and EADS3 are identical; the GSB features at 340 and 532 nm remain, but the intermediate features previously at 332 and 555 nm shift to 335 and 564 nm, respectively. The difference between EADS2 and EADS3 is in the lower intensity features in the red region of the spectrum, which are present in the EADS2, but not EADS3, which may be due to a small proportion of PCB Pr to Pb conversion occurring in parallel to the much more intense PVB processes. The final intermediate in these data, with an absorption maximum at 564 nm corresponds exactly with the intermediate previously observed in non time-resolved studies, assigned by those studies to the isomerised but not yet thioether linked chromophore.162,179

5.4.3 Millisecond to second transient absorption

The final set of laser flash photolysis measurements on millisecond timescales monitored the final step in the photoconversion reaction, when the isomerised intermediate forms a thioether linkage with the Cys499 residue in the protein. The PVB Pg population was isolated as described in the previous section. The data, shown in Figure 5.6A, initially have the same features as those described in the final EADS of the ns to ms dataset. Over the course of a few ms this converts to the final state with a large positive feature at ~430 nm. The data fit well to a sequential model with only 2 components, which interconvert with a lifetime of 23.6 ms. The first EADS is essentially identical to EADS3 in Figure 5.5D, with GSB

131 | P a g e of the Pg state at 340 and 532 nm, and the intermediate with features at 335 and 564 nm. EADS1 converts to the final difference spectra, EADS2, in which the GSB of the Pg state remain, and the final Pb state at 436 nm is apparent.

Figure 5.6 Transient absorption spectra of PVB Tlr0924 after excitation at 532 nm at selected time points over 0.2 – 199 ms A, and global analysis of the data B showing the resulting EADS (black dots) fitted with a sum of Gaussian functions (red line). There are features originating from the bleach of the Pg state (green lines), the second intermediate (orange lines), and the final Pb state (blue lines).

132 | P a g e

5.4.4 Cryotrapping

To determine which of the transitions observed in the time-resolved measurements were thermally activated cryotrapping measurements were carried out. In these experiments a sample in the PVB Pg state was illuminated with a cold light source fitted with a 530 nm bandpass filter at 127 K, then warmed to 297 K in 10 K steps, with spectra collected at 99 K between each temperature point. Difference spectra relative to the ‘dark spectrum’ at 99 K are shown in Figures 5.7A and B.

Figure 5.7 Tlr0924 photoconversion at cryogenic temperatures. Difference spectra at selected temperatures relative to ‘dark spectrum’ at 99 K A below 200 K and B above 200 K. C Temperature dependence of selected wavelengths over the 127 – 297 K temperature range.

133 | P a g e

The difference spectra in Figure 5.7A show a number of intermediates being formed. Throughout the measurement the bleach of the ground state at ~530 nm is visible. At the lowest temperatures there are positive features visible at ~583 nm and ~506 nm which are lost at 157 and 177 K, respectively. These two features are not observed in the time- resolved measurements, where the ground state bleach is significantly more intense due to the large excited state population. While the ~506 nm feature reduces in intensity as the sharp feature at ~560 nm appears, the ~583 nm peak does not seem to correlate with any other features. There are two most likely possible explanations for these features. They may be artefacts of the cryogenic techniques used; previous studies have shown that cryogenic temperatures can affect the energy landscape and possible conformational substates of proteins.281–283 This variation in protein structure could in turn affect the absorption properties of the chromophore. The other option is that these features are true intermediates in the photoreaction which were not resolvable in the ultrafast measurements because of either their very short lifetimes (<<0.6 ps) or low intensity relative to the very strong excited state features.

Above ~187 K (Figure 5.7B) the spectral features in the cryotrapping data more closely resemble the features observed in the time-resolved data. The temperature dependent spectral evolution at selected wavelength points is displayed in Figure 5.7C. It is not possible to resolve the 555 and 564 nm intermediates but the final step in the time- resolved data from the intermediate at 564 nm to the Pb state correlates very well with the final step observed in the cryotrapping measurements at 237 K. By fitting the temperature dependence at selected wavelengths (see Figure S5.7 in the Supporting Material) it is possible to put the temperature of formation of the 555/564 nm intermediate at around 179 K, and the formation of the final product state at around 232 K. At ~200 K it is known that proteins undergo a dynamic transition, termed the ‘glass transition’, below which any large-scale conformational changes in the protein that require solvent reorganisation become frozen out.34,223 As the formation of the 555/564 nm intermediate occurs close to this temperature it may be accompanied by some minor structural changes in the protein in order to accommodate the isomerised chromophore in the binding pocket. However, the final transition to form the Pb state can only proceed with formation of a second thioether linkage between chromophore and protein well above this ‘glass transition’ temperature, demonstrating that more large-scale protein motions are likely to be required for this stage of the photoconversion. This is a similar picture to that observed in the forward reaction of Tlr0924 and the photoreaction of the related Cph1 phytochrome.34,277

134 | P a g e

5.5 Conclusion

The time-resolved visible and IR transient absorption data described here present a clear picture of the PVB Pg to Pb photoconversion (Figure 5.8).

Figure 5.8 Suggested reaction scheme for the full PVB conversion cycle including details of both the Pg to Pb reaction from this work, and the Pb to Pg reaction.277

After photoexcitation of the PVB Pg state the majority of the excited state population relaxes back to the ground state with a lifetime of 0.6 ps. A small proportion of the excited population isomerises with a lifetime of 3.6 ps to 15Z PVB, with an absorption maximum at 555 nm. The final Pb state of PVB' is linked by a thioether linkage to a Cys in the bulk protein structure. Before the formation of this linkage another step, with a lifetime of 5.3 μs, occurs to form a further intermediate absorbing at 564 nm. The structure of this intermediate is not certain but it involves a red-shift in the absorbance maximum at temperatures close to the ‘glass transition’ temperature of proteins. Hence, it is possible that it may involve some twisting or other structural distortion of the chromophore, assisted by protein motions in the binding pocket, in which it becomes more planar, delocalising the π-orbital system, causing the red shift in absorbance maxima.187 Similar intermediates have been suggested for a number of related phytochrome and CBCR

135 | P a g e systems.160,187,269,284,285 The final step of the reaction progresses with a lifetime of 23.6 ms to the final Cys-bound 15Z Pb state and will likely involve some large-scale movement of the protein structure.

The time-resolved data presented here for μs to ms timescales are comparable to those reported for other CBCRs. Flash photolysis measurements have shown that the slower steps in the photoconversion process often proceed via 2-3 intermediates with spectrally distinct features. The precise nature of these intermediates is not confirmed, but isomerisation and changes in the localisation of the π-orbital system are both suggested. There is quite a variation in the lifetimes of conversion between these intermediates. The final two steps can occur with lifetimes of ~1 μs and ~920 μs (AnPixJ),187 750 ns and >1 ms (NpF2164g3),23 and 390 μs and 1.5 ms (Slr1393).286 None of these lifetimes are as long as the 23.6 ms observed for the final reaction step reported here, even in the case of NpF2164g3 which, as with Tlr0924, forms two thioether linkages with the protein. However, these previous studies were on the GAF domain only system, whereas here we have investigated the full- length protein, which may significantly extend the kinetics.286

The scheme suggested here is more complex than that suggested for the Pb to Pg forward reaction.277 In that case there are simply two steps, the fast photoisomerisation followed by the very slow breakage of the thioether linkage. Neither of those steps depends strongly on the geometry and configuration of the chromophore ring systems. In contrast, in the reverse reaction after the photoisomerisation the chromophore must move into a geometry favourable for the formation of the thioether linkage, but overall the reaction progresses faster and is completed in ms, compared to seconds.

5.6 Acknowledgements

The work was funded by the Engineering and Physical Sciences Research Council (EPSRC), references: EP/I01974X/1, and EP/J020192/1. NSS is an EPSRC Established Career Fellow and a Royal Society Wolfson Merit Awardee. The time-resolved infrared measurements were carried out through program access support of the UK Science and Technology Facilities Council (STFC). AFEH was funded by a Biotechnology and Biological Sciences Research Council Collaborative Award in Science and Engineering (BBSRC CASE) award supported also by TgK Scientific Ltd. (Bradford on Avon, UK).

136 | P a g e

5.7 Supporting Information

Figure S5.1 Ultrafast transient absorption spectra, collected using depolarized excitation with a wavelength of approximately 530nm, at selected time points for PVB Tlr0924 (A and B), global analysis of the ultrafast transient absorption data for PVB Tlr0924 (C and D)

Figure S5.2 Cryotrapping experiments where the sample was warmed from 77 K in 10 K steps, illuminating at each temperature for 10 minutes before cooling to 77 K to record the spectrum.

137 | P a g e

Figure S5.3 Ultrafast visible transient absorption spectra at time points between 20 and 3174 ps A, and the second two EADS resulting from global analysis of the data B. The sample was constantly illuminated during data collection with blue and

red light so the signals primarily originate from the ‘reverse’ photoconversion of the PVB chromophore.

138 | P a g e

Figure S5.4 The residual matrix from the global analysis of the ultrafast transient absorption data is deconvolved by singular value decomposition to the component times A and wavelengths B. The corresponding analysis performed for a 3- component fit to the data is shown in figures C and D. The lack of obvious structure in the time-component of the 4-component fit residual implies a good fit has been obtained.

139 | P a g e

Figure S5.5 Sequential global analysis of the ultrafast visible transient absorption data showing resulting EADS (black dots) fitted with a sum of Gaussian functions (dashed red lines). The features are assigned as: GSB of the PVB Pg state at 532 nm (green line), ESA features at 420, 495, and 667 nm (lilac, cyan and orange lines), inactive or modified protein at 645 nm (purple line), intermediate states at 580, 605, and 555 nm (pink and dark green), and pump scatter at 525 nm (grey line)

140 | P a g e

Figure S5.6 Laser flash photolysis data collected with PVB Tlr0924 after excitation at 532 nm: spectra at selected time points between 20 ns and 9 μs A, and 0.7 and 450 μs B, and kinetics at 510 and 570 nm C.

141 | P a g e

Figure S5.7 Temperature dependence of selected wavelengths over the 127 – 297 K temperature range fitted with Boltzmann distribution: A  A y  1 2  A where A1 is the initial ΔAbs value, A2 is the final 1 e(xx0 ) / dx 2 ΔAbs value, x0 is the centre temperature, and dx is the ‘time constant’. Values of x0 are: 179±2 K and 234±1 K for the 564 nm data, 228±2 K for the 532 nm data, and 233±3 K for the 435 nm data.

Low temperatures only permit localised structural changes. x0 values below the glass transition are therefore associated with the formation of reaction

intermediates with incomplete relaxation of the protein structure. x0 values above the glass transition are indicative of processes involving large scale protein motions, like the ones associated with the formation of the second thioether linkage in Tlr0924.

142 | P a g e

5.8 Note on PCB Pr to Pb Photoconversion

It was difficult to extract meaningful information on the ultrafast photoconversion of Tlr0924 PCB, because the concentration of PCB was much lower than that of PVB and the signal-to-noise ratio was accordingly low while the pump scatter was wider in the red than the blue spectral region. Laser flash photolysis measurements were carried out as described in section 5.3.3 on the ns to ms timescale. The protein in the Pg/r state was excited with a ~7 ns laser flash at 630 nm to selectively drive the PCB Pr population to the PCB Pb population. Time-dependent absorbance changes were recorded at 5 nm intervals across the visible range and assembled into 2D difference spectra. Between each laser shot, the sample was illuminated with blue light to reform the PCB Pr state.

The spectra show a bleach of the 15EPr state at ~575 nm and positive features at 620-660 nm and 340 nm. There is no significant spectral evolution on this μs timescale (Figure S5.8).

Figure S5.8. ns to μs photoconversion of PCB Pr to PCB Pb Panel 1 shows the visible difference spectrum of PCB Pb photoexcitation at 630 nm

in a microsecond time frame. Panel 2 shows kinetic changes at single wavelengths also covering the μs region. There is no spectral evolution on this time scale.

On the ms time scale there is an apparent red shift of the 15EPr ground state bleach as the 630 nm feature decays. Simultaneously, the PCB 15ZPb photoproduct forms at ~435 nm, while the 340 nm feature decays to sub-zero and blue shifts (Figure S5.9A). Sequential analysis of the data revealed two components (Figure S5.9B): EADS1, which decays into the final EADS2 with a lifetime of 37.9 ms (rate constant of 26.4 s-1). Both EADS could be described by 6 Gaussians for peak assignment: Wine coloured, negative Gaussians at 360 and 532 nm described the bleach of the Pr ground state population. The subsequently formed “630 nm” intermediate had in fact contributions from two species (622 nm and 640 nm), which also absorbed in the blue at 353 nm and 451 nm. In EADS2, the positive spectral

143 | P a g e contributions were located to 436 nm and 340 nm for Pb formation. Small contributions at 640 nm and 451 nm remain.

Figure S5.9. ms PCB Pr to Pb photoconversion and global analysis A Laser flash photolysis of PCB Pr after 630 nm excitation measured over 200 ms and B EADS resulting from global analysis and their respective fits to a sum of underlying Gaussian functions.

In contrast to the PVB Pg to Pb reaction, PCB Pr to Pb could be fitted to only 2 components rather than 3 chromophore specific spectra. However, PCB EADS 1 when fitted to underlying Gaussian peaks demonstrated two species as previously observed for PVB photoconversion. PCB could therefore potentially also undergo an analogous conversion of 622 nm  640 nm, which could not be resolved in a noisy data set. The final rate limiting step in the PCB reverse reaction is slightly slower than that of PVB (37.9 ms vs. 23.6 ms).

144 | P a g e

Chapter 6

Discussion

- 6.1 Discussion - 6.2 Future work - 6.3 Conclusion

145 | P a g e

6.1 Discussion

This thesis was concerned with the spectroscopic characterisation of the cyanobacterial photoreceptor Tlr0924 from Thermosynechococcus elongatus. In chapter 3, it was demonstrated that Tlr0924 could be produced heterologously in E. coli by co-expression with a PCB synthesising plasmid.245 Here, it was found that free PCB was unstable in the cell and the timing between induction of the two plasmids was of importance (Chapter 3.4.1). The isolated full-length protein was demonstrated to undergo blue/green and blue/red photochromicity based on the presence of two linear bilin chromophores, PVB and PCB (Chapter 3.4.2). These two photon-switching reactions are characteristic for the phytochrome and cyanobacteriochrome subclass of photoreceptors and the use of different bilin chromophores with slightly varied spectral properties has been demonstrated frequently.287 Whereas phytochromes undergo red/far-red photocycles, CBCRs are known to be much more varied162,288 and the Tlr0924 spectral characteristics are consistent with other proteins of the DXCF subgroup. The DXCF CBCR, TePixJ GAF, for example shows near identical spectral shapes and peak positions.190 The Pb, Pg and Pr states showed broad spectral sensitivities, including blue light-absorbing features in the Pg and Pr states according to the S0 to S2 transition (Chapter 3.4.4) and significant spectral overlap between states. Signals recorded will therefore often present the sum of several processes.

Like other members of the DXCF subgroup, Tlr0924 chromophores are covalently linked to the protein via the C3 of their A ring but also form a labile second thioether linkage between the chromophore C10 position and the conserved DXCF cysteine in the Pb state.142,176 This was suggested by a number of experiments (Chapter 3.4.3). FTIR measurements showed putative thiol stretching modes in the Pg and Pr states but not in the Pb state as previously observed in TePixJ.190 Furthermore, chemical modification of the DXCF cysteine by iodoacetamide or hydrogen peroxide oxidation prevented formation of the Pb state as previously reported.26 Instead, illumination of the Pg/r states gave rise to the formation of the red-shifted, photoisomerised intermediates. The only other known photoreceptor class to undergo transient cysteine adduct formation are the LOV domain proteins. In their case, the covalent bond is suggested to provide additional stability of the signalling state.9 In Tlr0924, the double-linked chromophore is however, associated with the dark state. It may serve a stabilising function, but its main role can certainly be assumed to be the spectral tuning from green/red- to blue light sensitivity.

146 | P a g e

The nature of the physiological bilin chromophore in Tlr0924 is unknown. Phycoviolobilin is not synthesised in cyanobacteria. It is isomerised from bilin precursors upon binding to the protein, sometimes with the help of other bilin isomerases.289 DXCF CBCRs are so far the only CBCR subclass known to convert the PCB precursor. This conversion is frequently incomplete in E. coli182,190 and with very few in vivo protein expression studies it remains unclear whether the isomerisation is a tuning mechanism162 or an artefact of heterologous expression.185 In Tlr0924, PCB conversion to PVB by autoisomerase activity is demonstrated in this work and previously.162 Furthermore, both PCB and PVB were found to undergo dark reversion reactions from their 15E to their 15Z isomers (Chapter 3.4.5). The rate of PCB reversion was faster than the rate of PVB reversion, which was previously demonstrated to be the case for non-pysiological chromophores.149 Both the rate of PCB to PVB conversion and the dark reversion were very slow, suggesting they would not affect the results of time- resolved photoexcitation studies.

The yield of photoconversion was previously suggested to be temperature dependent,179 which could not be confirmed in this study (Chapter 3.4.6). The conversion rates were affected by temperature and transient absorbance measurements were therefore maintained at 20°C.

Chapters 4 and 5 summarise the detailed kinetic and spectroscopic characterisation of full- length Tlr0924 across the fs to s timescale and covering cryogenic to room temperatures, permitting the construction of a photocycle model (Figure 6.1). This is the first comprehensive study in any phytochrome or cyanobacteriochrome system carried out in a single lab, eliminating factors such as different expression and buffer systems and experimental conditions. For the forward reaction a simple two step mechanism was suggested involving picosecond Z to E isomerisation of the PCB and PVB chromophore around the C15=C16 bond, followed by elimination of the protein chromophore thioether linkage at the C10 position within a few seconds (Chapter 4). The reverse reaction appeared slightly more complex. The return from the E to the Z state also occurred within ps but an additional intermediate was resolved on the microsecond timescale, which preceded the millisecond reformation of the C10 thioether linkage (Chapter 5). The two chromophore populations were separated experimentally and were shown to isomerise via spectrally and kinetically equivalent intermediates. The isomerisation reactions were independent of protein structural rearrangement whereas the subsequent steps were shown to proceed only above the glass transition temperature of proteins (Chapter 4.4.4 and 5.4.4).

147 | P a g e

hv hv 23.6 ms 15Z PVB’ 2 ps 37.9 ms 15Z PCB’ 2 ps 436 nm 10 ps 436 nm 10 ps 15Z PVB 15Z PCB 564 nm 15Z PVB’* 640 nm 15Z PVB’* 5.3 μs PVB 10 ps PCB 10 ps

15Z PVB 15Z PCB 3.6 ps 555 nm 622 nm 15E PVB’ 15E PCB’ 15E PVB* 415 nm 15E PVB* 415 nm 0.6 ps >0.2 ps 0.9 s 3.1 s hv hv 15E PVB 15E PCB 532 nm 588 nm

Figure 6.1 Proposed photocycles of PCB and PVB bound full-length Tlr0924 PCB and PVB bound Tlr0924 undergoes comparable photocycles. The forward reactions

proceed via a blue shifted isomer prior to elimination of the thioether. The reverse reactions proceed via two consecutively red shifted intermediates prior to formation of

the thioether. Lifetimes of the PCB photocycle are longer, at least for thermal steps. P (Propionate)

CBCR research has often focussed on the primary photoisomerisation reactions of the GAF domains only. Studies have been published for the red/green CBCRs NpR6012g422,185,186 and NpF2164g6,168 the NpF2164g3 insert-Cys23 and the Tlr0924g domain.180 Red/green CBCRs were described to undergo slower forward- (ps to ns) than reverse photoisomerisation kinetics (early ps).22,168,185,186 NpF2164g3 and Tlr0924g mainly isomerised on the picoseconds timescale. The present work on the full-length protein also shows photoisomerisation kinetics on the ps timescale and is therefore consistent with other studies. When comparing Tlr0924g to the full-length protein in particular, the transient spectra recorded appeared remarkably similar. However, the studies on the GAF domain used a mathematical model to separate the two chromophore populations,180 rather than the experimental approach used in the work presented here. Analysis therefore yielded different conclusions. In the Tlr0924g studies the chromophores were subdivided into heterogenous ground state populations with chromophore specific lifetimes. For the full- length protein, a less convoluted model was applied suggesting photoisomerisation of PVB and PCB Pb states with similar lifetimes of ~10 ps. The isomerised intermediate was blue-

148 | P a g e shifted in Tlr0924, which is not usually observed for Phys and CBCRs. The Pg photoisomerisation was observed with a lifetime of 3.6 ps. Definitive interpretation of Pr photoisomerisation was not possible because of low signals compared to the pump pulse. For forward and reverse reactions no further intermediates were detected within the 3 ns delay time of the instrument, which was consistent with Tlr0924g.

Different PCB and PVB isomerisation rates could potentially be expected based on the structural difference in the A-ring and C5-C6 saturation, which could lead to different interactions in the chromophore pocket and more rotational freedom in PVB than PCB. This is however likely to be more applicable to the Pg/r to Pb reaction than the Pb to Pg/r, where the conjugated systems are identical and restricted to rings C and D. Furthermore, the A and B rings in GAF domain only proteins are believed to be solvent exposed176 and might therefore be expected to have even less influence on photoisomerisation rates. The different chromophore environment could however be a rationale for some kinetic deviations. The heterogenous ground state populations regularly reported by this group23,167,180,185 may also be attributable to the freedom of the former half of the chromophore through lack of restriction by the full-length protein.

Overall, the GAF domain alone can still be considered a valid model for the full-length protein. The Tlr0924g stationary absorbance properties are practically identical to the full- length protein and the raw ultrafast transient absorbance data appear similar. This is a highly desirable property for artificial photoreceptor design and unlike e.g. LOV domains where photochemical properties are heavily reliant on the surrounding protein structure. The lifetimes of the Tlr0924 primary photoreactions are comparable to those of other isomerising cofactors, including the insert-Cys CBCR23 and phytochrome class290 and PYP.116 Rhodopsins isomerise on a faster, fs timescale.24,291

It has previously been shown that among the different classes of photoreceptors the quantum efficiency is lowest in the phytochrome family. However, the quantum yield of photochemistry in CBCR proteins have rarely been determined but have been assumed to be in the same 0.1-0.15 range180 with the exception of NpR6012g4, which was ascribed a special second chance isomerisation mechanism (Ф=0.4).166 Other isomerising cofactors routinely have quantum yields of 0.35 (PYP)292 and 0.65 (Rho/BR).293 Further analysis of the data acquired for this thesis is in progress to determine an exact value for Tlr0924.

Ultrafast time-resolved measurements were also recorded in the infrared region of the spectrum to further probe for light-induced chromophore changes and interactions in the

149 | P a g e chromophore pocket. These ps data represent the first of their kind on primary photodynamics in a CBCR protein. Kinetically they match the visible data extremely well and peak assignment could be carried out according to related phytochrome studies. The main finding was a downshift of the D ring C=O stretching mode in the Pb to Pg/r conversion, which corresponds to a stronger hydrogen bonding environment in the isomerised intermediate, as previously observed in Cph1 and BphPs.261,266 The direction of the shift has however not been universally agreed on.255,269 In the reverse reaction, formation of a positive feature at 1711 cm-1 coincided with the chromophore isomerisation reaction and was again assigned to changes to the D ring C=O hydrogen bonding network. Breakage of a hydrogen bond between a chromophore carbonyl and a cysteine residue is also observed following the isomerisation reaction of PYP and is associated with a shift of similar magnitude (20 cm-1).294

Furthermore, C=C stretching modes for rings A and B and rings C and D could be assigned to the 1650 to 1600 cm-1 region. C and D ring stretches were previously observed to shift upon isomerisation261 but in these data, a positive overlapping feature is believed to be the cause.

The slower thermal reactions of the CBCR photoactivation have so far often been neglected in studies of the overall photocycle. There have been indications for longer-lived species from ultrafast studies, where intermediates could be assigned to the μs timescale by recording a 1 ms difference spectrum but without assigning lifetimes or knowledge of the number of steps that might occur.23,180,295 Time-resolved spectral data on the ns to s timescales have only previously been measured for the red/green CBCR AnPixJg2187 and phytochrome systems.271

For the Tlr0924 PCB and PVB forward reactions no significant spectral evolution was observed on the ns to μs timescales. PVB then proceeded from the 415 nm 15E intermediate to the Pg state with a lifetime of 937 ms. PCB measurements were difficult to obtain. On the single wavelength detection system of the flash photolysis spectrometer, continuous probe light in the green spectral region converted the PVB Pg state to Pb and made data difficult to interpret. This problem was reduced by switching to a home-made LED-flash system with a Varian Cary spectrometer, which utilises a pumped probe light. Since the PCB Pb to Pr conversion occurred on the seconds timescale, this approach was possible. Spectral contributions of the PVB Pg to Pb conversion however still remained and

150 | P a g e had to be subtracted from the raw data to continue the analysis. The 415 nm intermediate was then found to decay with a life time of 3.1 s to form the Pr photoproduct.

In contrast, the PVB reverse reaction could be cleanly separated from the PCB contributions and was found to proceed via 2 consecutively red-shifted intermediates with lifetimes of 5.3 μs and 23.6 ms to form the thioether linkage of the Pb state. The PCB reaction spectra were compromised by lower signal-to-noise ratios. Two red-shifted intermediates at 622 and 640 nm could be fitted to the ms data but not resolved temporally. In line with PVB behaviour, the 640 nm intermediate would be assumed to form from the 622 nm intermediate. The final step of thioether breakage occured with a life-time of 37.9 ms.

This work therefore suggests that at least the thermal reactions of the PCB chromophore are slower than their PVB counterparts. This may be another indication that PCB is not the physiological chromophore and the reaction pathway has been optimised for PVB. Otherwise, PCB may possibly be employed to widen the spectral sensitivity of the protein.

There are no data available on UV/Vis studies at cryogenic temperatures in CBCR systems. It is however, a reliable means of assigning steps of the photocycle to either photochemical or thermal events and to determine the extent of protein conformational changes involved in the signal transduction mechanism within the protein. In Tlr0924, the isomerisation reactions were demonstrated to proceed at cryogenic temperatures where structural rearrangements are restricted to the chromophore pocket, consistent with observations in Cph1.34 The elimination and formation of the thioether linkage is dependent on long range motions as indicated by transition temperatures above 200 K. In the forward reaction, a blue-shifted 15E Pb intermediate was also seen with an apparent peak at 390 nm but overlapping heavily with the ground state bleach. The reverse reaction showed more spectral features as also observed in transient absorbance measurements - either because of the increased freedom of the chromophore without the C10 thioether or because of less spectral overlap and larger peak shifts than in the forward reaction. Red-shifting reaction intermediates were observed as in time-resolved techniques. Additionally, there are both hypsochromic and bathochromic shifts at the lowest temperatures as previously observed in PYP, where a branched low-temperature photocycle is attributed to the technique. The extended illumination times allows for species transiently formed from the ground state to be photoactivated as well, which may also be the origin of the additional features in Tlr0924. In general, peaks sharpen at low temperatures and may also shift slightly at cryogenic temperatures. The cryosolvents were also found to have some unassigned effects

151 | P a g e on the protein (Chapter 3.4.7). In comparison to Cph1, the Tlr0924 forward reaction requires higher temperatures to form photoproduct states. This has been attributed to the additional energy cost required to break a covalent bond and is consistent with longer life times in the time-resolved studies.

The PCB and PVB forward reactions therefore proceed via only one distinguishable isomerised intermediate with an intact thioether linkage. Breakage of the thioether linkage resulted in formation of the Pg/r photoproduct states. Compared to other photoreceptor systems this seems a very simplistic scheme, with rhodopsin signalling state formation proceeding via five intermediates,296 PYP via four118 and phytochromes145 and CBCRs23,166 via two or three. This was attributed to the presence of the second thioether linkage that was intact until the last step of the reaction, which is thought to restrict formation of many different states. It also affected the lifetime of this state. Isomerisation-based photoreceptors are known to form their signalling states on the μs to ms time scales whereas Pg formation in Tlr0924 took nearly a second and Pr formation 3 s. The reverse reaction was much faster and proceeded via two intermediates with μs and ms lifetimes as would be more commonly expected. Breakage of the thioether linkage therefore appears to be an energetically less favourable process than the formation. Spectrally silent, large scale protein motions may have to occur to induce enough strain on the bond to cause it to break. The formation of the thoiether linkage in the reverse reaction might precede and induce the large scale protein changes subsequently. This hypothesis is supported by preliminary wide-angle X-ray scattering measurements (data collected are currently being analysed by collaborators at the ESRF). The activation of the photoreceptor signalling cascade is presumably a comparatively slow process and a rapid onset may not be required. The pathway may be metabolically expensive, so a more rapid inactivation is desirable.

From the biotechnology point of view, Tlr0924 and other CBCRs offer a blue absorbing light switch, which extends the spectral sensitivities significantly. However, Tlr0924 also highlights possible issues of chromophore substitution reactions in this class, which may yield heterogenous populations and absorbance spectra and therefore cause a loss of specificity, which is certainly undesirable when using multiple photoswitches.

152 | P a g e

6.2 Future Work

With completion of the UV/Vis characterisation of the Tlr0924 photocycle a first milestone in the study of this photoreceptor has been achieved. From here, different routes and approaches for further study are possible:

One direction would be to focus further on the characterisation of the full-length protein with respect to structural analyses together with studies on the signal transduction and physiological function of the protein. Crystallographic studies in this photoreceptor family are however known to be challenging because of large protein sizes and spectral overlap between different photostates. Signalling proteins are generally flexible, which is essential to their function but results in heterogeneity and prevents crystallisation. To date, no Phy or CBCR full-length protein has been crystallised in both states. Crystal trials on Tlr0924 were not attempted in this PhD thesis due to polydispersed protein samples and a “very difficult” crystallisation rating (5) in Xtalpred.297 This prediction is based on several biochemical properties including e.g. protein length, hydrophobicity, and secondary structure. Structural information of the chromophore binding site could help determine a mode of action for the GAF domain autolyase and autoisomerase activity as well as the photoisomerisation reaction itself. This could be achieved by determining the role of certain amino acid side chains by site directed mutagenesis. Furthermore, temperature scan X-ray or Laue crystallography could resolve the structure of early photocycle intermediates35,99,298 with an indication as to how chromophore structural changes are translated to the bulk protein structure. Actual crystal structures or those of homologues would also be beneficial as starting point for computational modelling and X-ray scattering techniques. Small-angle X-ray scattering is a means of studying the size and shape of the native protein in solution. In a similar approach, wide-angle X-ray scattering gives a better resolution of the protein fold.299 We have already attempted to study the signal transduction mechanism of Tlr0924 further using a time-resolved wide-angle X-ray scattering technique in solution.300 Other techniques to observe protein motion could include electron microscopy301 and the introduction of spin labels or fluorescence labels. Spin labels permit studying of protein motion in terms of distance between each other in electron paramagnetic resonance studies,34 whereas fluorescence labels could add a time- resolved aspect to the motion between the sensory and effector domain further linking protein structural changes to the chromophore changes.302 Based on the Tlr0924 domain structure, the protein is predicted to have diguanylate cyclase activity. This activity could be

153 | P a g e studied by monitoring the guanosine triphosphate (GTP) turnover to cyclic diguanylate (c- di-GMP) by high performance liquid chromatography.154,303,304 In an initial test of Mant-GTP binding, we did not observe the expected increase in fluorescence read-out, which may be due to steric hindrance introduced by the fluorescent label. This approach would permit to confirm the activity in the two photostates and to definitvely assign the signalling state. It may also be beneficial to express the protein in a cyanobacterial host to confirm the physiological identity of the chromophore.

Alternatively, the focus could remain on the characterisation of cyanobacterial photocycles and similar or the same techniques could be applied to other DXCF CBCRs to confirm the mode of action is conserved across this subfamily and to other subgroups to highlight the mechanistic differences.

Another option would be to attempt to modify the protein on different levels. Chromophore substitutions and site-directed mutagenesis could be used to tune the photocycle properties. A bigger goal would be to reassemble a cyanobacterial photoreceptor of different domains to achieve various spectral sensitivities for different enzymatic activities. Different enzymes of biosynthetic pathways could in this way be selectively activated and inactivated to yield maximal product. Furthermore, disease models could be tested and onset and offset of the genotype controlled spatio-temporally. It would therefore be worthwhile to study the effect of deleting the N-terminal multimerisation motifs or attempt to change the signalling output all together. Prerequisite and knowledge gained in the process should be an understanding of the importance of the linker regions. The reconstitution of proteins with different domains or a change of the signalling output has already been demonstrated to be feasible.194,207,208 The photoreceptor design itself has also the potential to yield new insights into photoreceptor function.

154 | P a g e

6.3 Conclusion

Cyanobacteriochrome research is an exciting field with many potential applications. This thesis covered the fundamental mechanistic aspects of one such CBCR, Tlr0924, which are the basis for the rational design of novel fluorescent probes and photoswitchable domains for use in optogenetics or control of biosynthetic pathways.

It is conceivable that the Tlr0924 GAF domain could be used as blue/green light switch fused to target genes to spatio-temporally control their activity. In synthetic biology this would allow selection of product formation over toxic products and metabolic burden to the cell to produce high-value chemicals sustainably, greener and cheaper. The same light- switch could also be applied in optogenetics to study neuronal circuits and disease models. Unlike phytochromes however, Tlr0924 spectral sensitivity is less suitable for deep tissue penetration. Advantageous over other photoreceptors used in biotechnology is the fact that cyanobacteriochrome GAF domains are comparatively small to phytochromes and also can be rapidly converted between active and inactive conformations.

155 | P a g e

References

(1) Wayne, C. E., and Wayne, R. P. (1996) Photochemistry in nature, in Photochemistry, pp 67–77. Oxford University Press, Oxford.

(2) Croce, R., and van Amerongen, H. (2014) Natural strategies for photosynthetic light harvesting. Nat. Chem. Biol. 10, 492–501.

(3) Oike, H., Oishi, K., and Kobori, M. (2014) Nutrients, Clock Genes, and Chrononutrition. Curr. Nutr. Rep. 3, 204–212.

(4) Haupt, W. (2001) Photomovement: past and future, in Comprehensive Series in Photosciences Volume 1 (Haeder, D.-P., and Jori, G., Eds.), pp 1–14. Elsevier, Amsterdam.

(5) Imamoto, Y., and Shichida, Y. (2014) Cone visual pigments. Biochim. Biophys. Acta 1837, 664–73.

(6) Committee on Biosciences Research in Agriculture. (1985) Plant Science, in New Directions for Biosciences Research in Agriculture: HIgh-Reward Opportunities, pp 54–80. National Academy Press, Washington, D.C.

(7) Auldridge, M. E., Satyshur, K. A., Anstrom, D. M., and Forest, K. T. (2012) Structure-guided engineering enhances a phytochrome-based infrared fluorescent protein. J. Biol. Chem. 287, 7000–9.

(8) Christie, J. M., Gawthorne, J., Young, G., Fraser, N. J., and Roe, A. J. (2012) LOV to BLUF: flavoprotein contributions to the optogenetic toolkit. Mol. Plant 5, 533–44.

(9) Van der Horst, M. A., and Hellingwerf, K. J. (2004) Photoreceptor proteins, “star actors of modern times”: a review of the functional dynamics in the structure of representative members of six different photoreceptor families. Acc. Chem. Res. 37, 13–20.

(10) Ikeuchi, M., and Ishizuka, T. (2008) Cyanobacteriochromes: a new superfamily of tetrapyrrole- binding photoreceptors in cyanobacteria. Photochem. Photobiol. Sci. 7, 1159–67.

(11) Berg, J. M., Tymoczko, J. L., and Stryer, L. (2002) Light reactions of photosynthesis, in Biochemistry (Berg, J. M., Tymoczko, J. L., and Stryer, L., Eds.) 5th ed. W H Freeman, New York.

(12) McClung, C. R. (2006) Plant circadian rhythms. Plant Cell 18, 792–803.

(13) Kohen, E., Santus, R., and Hirschberg, J. G. (1995) The nature of light, in Photobiology, pp 3–22. Academic Press, San Diego.

(14) Kohen, E., Santus, R., and Hirschberg, J. G. (1995) Pathways of molecular excitation and deactivation, in Photobiology, pp 73–96. Academic Press, San Diego.

(15) Petrich, J. W., and Fleming, G. R. (1984) ULTRAFAST PROCESSES IN BIOLOGY. Photochem. Photobiol. 40, 775–780.

(16) Kennis, J. T. M., Crosson, S., Gauden, M., van Stokkum, I. H. M., Moffat, K., and van Grondelle, R. (2003) Primary reactions of the LOV2 domain of phototropin, a plant blue-light photoreceptor. Biochemistry 42, 3385–92.

156 | P a g e

(17) Mathes, T., van Stokkum, I. H. M., Bonetti, C., Hegemann, P., and Kennis, J. T. M. (2011) The hydrogen-bond switch reaction of the Blrb Bluf domain of Rhodobacter sphaeroides. J. Phys. Chem. B 115, 7963–71.

(18) Giovani, B., Byrdin, M., Ahmad, M., and Brettel, K. (2003) Light-induced electron transfer in a cryptochrome blue-light photoreceptor. Nat. Struct. Biol. 10, 489–90.

(19) Perman, B., Srajer, V., Ren, Z., Teng, T., Pradervand, C., Ursby, T., Bourgeois, D., Schotte, F., Wulff, M., Kort, R., Hellingwerf, K., and Moffat, K. (1998) Energy transduction on the nanosecond time scale: early structural events in a xanthopsin photocycle. Science 279, 1946–50.

(20) Dasgupta, J., Frontiera, R. R., Taylor, K. C., Lagarias, J. C., and Mathies, R. a. (2009) Ultrafast excited-state isomerization in phytochrome revealed by femtosecond stimulated Raman spectroscopy. Proc. Natl. Acad. Sci. U. S. A. 106, 1784–9.

(21) Kim, P. W., Pan, J., Rockwell, N. C., Chang, C.-W., Taylor, K. C., Lagarias, J. C., and Larsen, D. S. (2012) Ultrafast E to Z photoisomerization dynamics of the Cph1 phytochrome. Chem. Phys. Lett. 549, 86–92.

(22) Kim, P. W., Freer, L. H., Rockwell, N. C., Martin, S. S., Lagarias, J. C., and Larsen, D. S. (2012) Second-chance forward isomerization dynamics of the red/green cyanobacteriochrome NpR6012g4 from Nostoc punctiforme. J. Am. Chem. Soc. 134, 130–3.

(23) Gottlieb, S. M., Kim, P. W., Corley, S. C., Madsen, D., Hanke, S. J., Rockwell, N. C., Chang, C.-W., Martin, S. S., Lagarias, J. C., and Larsen, D. S. (2014) Primary and Secondary Photodynamics of the Violet/Orange Dual-Cysteine NpF2164g3 Cyanobacteriochrome Domain from Nostoc Punctiforme. Biochemistry.

(24) Schoenlein, R. W., Peteanu, L. A., Mathies, R. A., and Shank, C. V. (1991) The first step in vision: femtosecond isomerization of rhodopsin. Science 254, 412–5.

(25) Ko, W.-H., Nash, A. I., and Gardner, K. H. (2007) A LOVely view of blue light photosensing. Nat. Chem. Biol. 3, 372–4.

(26) Rockwell, N. C., Martin, S. S., and Lagarias, J. C. (2012) Mechanistic insight into the photosensory versatility of DXCF cyanobacteriochromes. Biochemistry 51, 3576–85.

(27) Wayne, C. E., and Wayne, R. P. (1996) Chemical change, in Photochemistry, pp 18–38. Oxford University Press.

(28) Turro, N. J. (1991) Radiative Transitions - The Absorption and Emission of Light, in Modern Molecular Photochemistry, pp 76–152. University Science Books, Sausalito, CA.

(29) Ortiz-Guerrero, J. M., Polanco, M. C., Murillo, F. J., Padmanabhan, S., and Elías-Arnanz, M. (2011) Light-dependent gene regulation by a coenzyme B12-based photoreceptor. Proc. Natl. Acad. Sci. U. S. A. 108, 7565–70.

(30) Jenkins, G. I. (2014) The UV-B photoreceptor UVR8: from structure to physiology. Plant Cell 26, 21–37.

(31) Baker, A. W., and Forest, K. T. (2014) Structural biology: Action at a distance in a light receptor. Nature 509, 174–5.

(32) Möglich, A., Yang, X., Ayers, R. A., and Moffat, K. (2010) Structure and function of plant photoreceptors. Annu. Rev. Plant Biol. 61, 21–47.

157 | P a g e

(33) Takala, H., Björling, A., Berntsson, O., Lehtivuori, H., Niebling, S., Hoernke, M., Kosheleva, I., Henning, R., Menzel, A., Ihalainen, J. A., and Westenhoff, S. (2014) Signal amplification and transduction in phytochrome photosensors. Nature 509, 245–8.

(34) Heyes, D. J., Khara, B., Sakuma, M., Hardman, S. J. O., O’Cualain, R., Rigby, S. E. J., and Scrutton, N. S. (2012) Ultrafast red light activation of Synechocystis phytochrome Cph1 triggers major structural change to form the Pfr signalling-competent state. PLoS One 7, e52418.

(35) Yang, X., Ren, Z., Kuk, J., and Moffat, K. (2011) Temperature-scan cryocrystallography reveals reaction intermediates in bacteriophytochrome. Nature 479, 428–32.

(36) Taylor, B. L., and Zhulin, I. B. (1999) PAS domains: internal sensors of oxygen, redox potential, and light. Microbiol. Mol. Biol. Rev. 63, 479–506.

(37) Möglich, A., Ayers, R. A., and Moffat, K. (2009) Structure and signaling mechanism of Per-ARNT- Sim domains. Structure 17, 1282–94.

(38) Ho, Y. S., Burden, L. M., and Hurley, J. H. (2000) Structure of the GAF domain, a ubiquitous signaling motif and a new class of cyclic GMP receptor. EMBO J. 19, 5288–99.

(39) Aravind, L., and Ponting, C. P. (1997) The GAF domain: an evolutionary link between diverse phototransducing proteins. Trends Biochem. Sci. 22, 458–9.

(40) Rockwell, N. C., Ohlendorf, R., and Möglich, A. (2013) Cyanobacteriochromes in full color and three dimensions. Proc. Natl. Acad. Sci. U. S. A. 110, 806–7.

(41) Losi, A., and Gärtner, W. (2008) Bacterial bilin- and flavin-binding photoreceptors. Photochem. Photobiol. Sci. 7, 1168–78.

(42) Cashmore, A. R. (2003) Cryptochromes: enabling plants and animals to determine circadian time. Cell 114, 537–43.

(43) Laan, W., van der Horst, M. A., van Stokkum, I. H., and Hellingwerf, K. J. (2003) Initial characterization of the primary photochemistry of AppA, a blue-light-using flavin adenine dinucleotide-domain containing transcriptional antirepressor protein from Rhodobacter sphaeroides: a key role for reversible intramolecular proton transfer f. Photochem. Photobiol. 78, 290–7.

(44) Fortunato, A. E., Annunziata, R., Jaubert, M., Bouly, J.-P., and Falciatore, A. (2014) Dealing with light: The widespread and multitasking cryptochrome/photolyase family in photosynthetic organisms. J. Plant Physiol.

(45) Masuda, S. (2013) Light detection and signal transduction in the BLUF photoreceptors. Plant Cell Physiol. 54, 171–9.

(46) Crosson, S., Rajagopal, S., and Moffat, K. (2003) The LOV domain family: photoresponsive signaling modules coupled to diverse output domains. Biochemistry 42, 2–10.

(47) Senda, T., Senda, M., Kimura, S., and Ishida, T. (2009) Redox control of protein conformation in flavoproteins. Antioxid. Redox Signal. 11, 1741–66.

(48) Li, Q.-H., and Yang, H.-Q. Cryptochrome signaling in plants. Photochem. Photobiol. 83, 94–101.

(49) Gauger, M. A., and Sancar, A. (2005) Cryptochrome, circadian cycle, cell cycle checkpoints, and cancer. Cancer Res. 65, 6828–34. 158 | P a g e

(50) Gegear, R. J., Foley, L. E., Casselman, A., and Reppert, S. M. (2010) Animal cryptochromes mediate magnetoreception by an unconventional photochemical mechanism. Nature 463, 804–7.

(51) Pokorny, R., Klar, T., Hennecke, U., Carell, T., Batschauer, A., and Essen, L.-O. (2008) Recognition and repair of UV lesions in loop structures of duplex DNA by DASH-type cryptochrome. Proc. Natl. Acad. Sci. U. S. A. 105, 21023–7.

(52) Ahmad, M., and Cashmore, A. R. (1993) HY4 gene of A. thaliana encodes a protein with characteristics of a blue-light photoreceptor. Nature 366, 162–6.

(53) Saxena, C., Wang, H., Kavakli, I. H., Sancar, A., and Zhong, D. (2005) Ultrafast dynamics of resonance energy transfer in cryptochrome. J. Am. Chem. Soc. 127, 7984–5.

(54) Byrdin, M., Lukacs, A., Thiagarajan, V., Eker, A. P. M., Brettel, K., and Vos, M. H. (2010) Quantum yield measurements of short-lived photoactivation intermediates in DNA photolyase: toward a detailed understanding of the triple tryptophan electron transfer chain. J. Phys. Chem. A 114, 3207– 14.

(55) Chaves, I., Pokorny, R., Byrdin, M., Hoang, N., Ritz, T., Brettel, K., Essen, L.-O., van der Horst, G. T. J., Batschauer, A., and Ahmad, M. (2011) The cryptochromes: blue light photoreceptors in plants and animals. Annu. Rev. Plant Biol. 62, 335–64.

(56) Ritz, T., Adem, S., and Schulten, K. (2000) A model for photoreceptor-based magnetoreception in birds. Biophys. J. 78, 707–18.

(57) Solov’yov, I. A., Chandler, D. E., and Schulten, K. (2007) Magnetic field effects in Arabidopsis thaliana cryptochrome-1. Biophys. J. 92, 2711–26.

(58) Ahmad, M., Galland, P., Ritz, T., Wiltschko, R., and Wiltschko, W. (2007) Magnetic intensity affects cryptochrome-dependent responses in Arabidopsis thaliana. Planta 225, 615–24.

(59) Crosson, S., and Moffat, K. (2001) Structure of a flavin-binding plant photoreceptor domain: insights into light-mediated signal transduction. Proc. Natl. Acad. Sci. U. S. A. 98, 2995–3000.

(60) Masuda, S., and Bauer, C. E. (2002) AppA is a blue light photoreceptor that antirepresses photosynthesis gene expression in Rhodobacter sphaeroides. Cell 110, 613–23.

(61) Gomelsky, M., and Klug, G. (2002) BLUF: a novel FAD-binding domain involved in sensory transduction in microorganisms. Trends Biochem. Sci. 27, 497–500.

(62) Anderson, S., Dragnea, V., Masuda, S., Ybe, J., Moffat, K., and Bauer, C. (2005) Structure of a novel photoreceptor, the BLUF domain of AppA from Rhodobacter sphaeroides. Biochemistry 44, 7998–8005.

(63) Gauden, M., van Stokkum, I. H. M., Key, J. M., Lührs, D. C., van Grondelle, R., Hegemann, P., and Kennis, J. T. M. (2006) Hydrogen-bond switching through a radical pair mechanism in a flavin-binding photoreceptor. Proc. Natl. Acad. Sci. U. S. A. 103, 10895–900.

(64) Khrenova, M. G., Nemukhin, A. V, and Domratcheva, T. (2013) Photoinduced electron transfer facilitates tautomerization of the conserved signaling glutamine side chain in BLUF protein light sensors. J. Phys. Chem. B 117, 2369–77.

(65) Salomon, M., Zacherl, M., and Rudiger, W. (1997) Asymmetric, blue light-dependent phosphorylation of a 116-kilodalton plasma membrane protein can be correlated with the first- and second-positive phototropic curvature of oat coleoptiles. Plant Physiol. 115, 485–91.

159 | P a g e

(66) Kinoshita, T., Doi, M., Suetsugu, N., Kagawa, T., Wada, M., and Shimazaki, K. (2001) Phot1 and phot2 mediate blue light regulation of stomatal opening. Nature 414, 656–60.

(67) Kagawa, T. (2002) Blue Light-Induced Chloroplast Relocation. Plant Cell Physiol. 43, 367–371.

(68) Wada, M., Kagawa, T., and Sato, Y. (2003) Chloroplast movement. Annu. Rev. Plant Biol. 54, 455–68.

(69) Nakasako, M., Zikihara, K., Matsuoka, D., Katsura, H., and Tokutomi, S. (2008) Structural basis of the LOV1 dimerization of Arabidopsis phototropins 1 and 2. J. Mol. Biol. 381, 718–33.

(70) Salomon, M., Christie, J. M., Knieb, E., Lempert, U., and Briggs, W. R. (2000) Photochemical and mutational analysis of the FMN-binding domains of the plant blue light receptor, phototropin. Biochemistry 39, 9401–10.

(71) Swartz, T. E., Corchnoy, S. B., Christie, J. M., Lewis, J. W., Szundi, I., Briggs, W. R., and Bogomolni, R. A. (2001) The photocycle of a flavin-binding domain of the blue light photoreceptor phototropin. J. Biol. Chem. 276, 36493–500.

(72) Harper, S. M., Christie, J. M., and Gardner, K. H. (2004) Disruption of the LOV-Jalpha helix interaction activates phototropin kinase activity. Biochemistry 43, 16184–92.

(73) Harper, S. M., Neil, L. C., and Gardner, K. H. (2003) Structural basis of a phototropin light switch. Science 301, 1541–4.

(74) Swartz, T. E., Tseng, T.-S., Frederickson, M. A., Paris, G., Comerci, D. J., Rajashekara, G., Kim, J.- G., Mudgett, M. B., Splitter, G. A., Ugalde, R. A., Goldbaum, F. A., Briggs, W. R., and Bogomolni, R. A. (2007) Blue-light-activated histidine kinases: two-component sensors in bacteria. Science 317, 1090– 3.

(75) Purcell, E. B., Siegal-Gaskins, D., Rawling, D. C., Fiebig, A., and Crosson, S. (2007) A photosensory two-component system regulates bacterial cell attachment. Proc. Natl. Acad. Sci. U. S. A. 104, 18241–6.

(76) Nash, A. I., McNulty, R., Shillito, M. E., Swartz, T. E., Bogomolni, R. A., Luecke, H., and Gardner, K. H. (2011) Structural basis of photosensitivity in a bacterial light-oxygen-voltage/helix-turn-helix (LOV- HTH) DNA-binding protein. Proc. Natl. Acad. Sci. U. S. A. 108, 9449–54.

(77) Ballario, P., Vittorioso, P., Magrelli, A., Talora, C., Cabibbo, A., and Macino, G. (1996) White collar-1, a central regulator of blue light responses in Neurospora, is a zinc finger protein. EMBO J. 15, 1650–7.

(78) Cao, Z., Livoti, E., Losi, A., and Gärtner, W. (2010) A blue light-inducible phosphodiesterase activity in the cyanobacterium Synechococcus elongatus. Photochem. Photobiol. 86, 606–611.

(79) Spudich, J. L., Yang, C. S., Jung, K. H., and Spudich, E. N. (2000) Retinylidene proteins: structures and functions from archaea to humans. Annu. Rev. Cell Dev. Biol. 16, 365–92.

(80) Kort, R., Hoff, W. D., Van West, M., Kroon, A. R., Hoffer, S. M., Vlieg, K. H., Crielaand, W., Van Beeumen, J. J., and Hellingwerf, K. J. (1996) The xanthopsins: a new family of eubacterial blue-light photoreceptors. EMBO J. 15, 3209–18.

(81) Rockwell, N. C., Su, Y.-S., and Lagarias, J. C. (2006) Phytochrome structure and signaling mechanisms. Annu. Rev. Plant Biol. 57, 837–58.

160 | P a g e

(82) Henderson, R., and Unwin, P. N. T. (1975) Three-dimensional model of purple membrane obtained by electron microscopy. Nature 257, 28–32.

(83) Nakanishi, K. (1991) 11-cis-retinal, a molecule uniquely suited for vision. Pure Appl. Chem. 63.

(84) Oesterhelt, D., and Stoeckenius, W. (1973) Functions of a new photoreceptor membrane. Proc. Natl. Acad. Sci. U. S. A. 70, 2853–7.

(85) Schobert, B., and Lanyi, J. K. (1982) Halorhodopsin is a light-driven chloride pump. J. Biol. Chem. 257, 10306–13.

(86) Inoue, K., Ono, H., Abe-Yoshizumi, R., Yoshizawa, S., Ito, H., Kogure, K., and Kandori, H. (2013) A light-driven sodium ion pump in marine bacteria. Nat. Commun. 4, 1678.

(87) Takahashi, T., Yoshihara, K., Watanabe, M., Kubota, M., Johnson, R., Derguini, F., and Nakanishi, K. (1991) Photoisomerization of retinal at 13-ene is important for phototaxis of Chlamydomonas reinhardtii: Simultaneous measurements of phototactic and photophobic responses. Biochem. Biophys. Res. Commun. 178, 1273–1279.

(88) Bogomolni, R. A., and Spudich, J. L. (1982) Identification of a third rhodopsin-like pigment in phototactic Halobacterium halobium. Proc. Natl. Acad. Sci. U. S. A. 79, 6250–4.

(89) Harz, H., and Hegemann, P. (1991) Rhodopsin-regulated calcium currents in Chlamydomonas. Nature 351, 489–491.

(90) Fredriksson, R., Lagerström, M. C., Lundin, L.-G., and Schiöth, H. B. (2003) The G-protein-coupled receptors in the human genome form five main families. Phylogenetic analysis, paralogon groups, and fingerprints. Mol. Pharmacol. 63, 1256–72.

(91) Ebrey, T., and Koutalos, Y. (2001) Vertebrate photoreceptors. Prog. Retin. Eye Res. 20, 49–94.

(92) Matsuyama, T., Yamashita, T., Imamoto, Y., and Shichida, Y. (2012) Photochemical properties of mammalian melanopsin. Biochemistry 51, 5454–62.

(93) Terakita, A. (2005) The opsins. Genome Biol. 6, 213.

(94) Tsukamoto, H., and Terakita, A. (2010) Diversity and functional properties of bistable pigments. Photochem. Photobiol. Sci. 9, 1435–43.

(95) Sundström, V. (2008) Femtobiology. Annu. Rev. Phys. Chem. 59, 53–77.

(96) Liu, R. S., and Asato, A. E. (1985) The primary process of vision and the structure of bathorhodopsin: a mechanism for photoisomerization of polyenes. Proc. Natl. Acad. Sci. U. S. A. 82, 259–63.

(97) Lanyi, J. K. (2004) Bacteriorhodopsin. Annu. Rev. Physiol. 66, 665–88.

(98) Balashov, S. P., and Ebrey, T. G. (2001) Trapping and spectroscopic identification of the photointermediates of bacteriorhodopsin at low temperatures. Photochem. Photobiol. 73, 453–62.

(99) Edman, K., Nollert, P., Royant, A., Belrhali, H., Pebay-Peyroula, E., Hajdu, J., Neutze, R., and Landau, E. M. (1999) High-resolution X-ray structure of an early intermediate in the bacteriorhodopsin photocycle. Nature 401, 822–6.

161 | P a g e

(100) Royant, A., Edman, K., Ursby, T., Pebay-Peyroula, E., Landau, E. M., and Neutze, R. (2000) Helix deformation is coupled to vectorial proton transport in the photocycle of bacteriorhodopsin. Nature 406, 645–8.

(101) Braiman, M. S., Bousché, O., and Rothschild, K. J. (1991) Protein dynamics in the bacteriorhodopsin photocycle: submillisecond Fourier transform infrared spectra of the L, M, and N photointermediates. Proc. Natl. Acad. Sci. U. S. A. 88, 2388–92.

(102) Xiao, W., Brown, L. S., Needleman, R., Lanyi, J. K., and Shin, Y. K. (2000) Light-induced rotation of a transmembrane alpha-helix in bacteriorhodopsin. J. Mol. Biol. 304, 715–21.

(103) Yoshizawa, T., Shichida, Y., and Matuoka, S. (1984) Primary intermediates of rhodopsin studied by low temperature spectrophotometry and laser photolysis. Vision Res. 24, 1455–1463.

(104) Nakamichi, H., and Okada, T. (2006) Local peptide movement in the photoreaction intermediate of rhodopsin. Proc. Natl. Acad. Sci. U. S. A. 103, 12729–34.

(105) Ye, S., Zaitseva, E., Caltabiano, G., Schertler, G. F. X., Sakmar, T. P., Deupi, X., and Vogel, R. (2010) Tracking G-protein-coupled receptor activation using genetically encoded infrared probes. Nature 464, 1386–9.

(106) Hargrave, P. A., Hamm, H. E., and Hofmann, K. P. (1993) Interaction of rhodopsin with the G- protein, transducin. Bioessays 15, 43–50.

(107) Bartl, F. J., and Vogel, R. (2007) Structural and functional properties of metarhodopsin III: recent spectroscopic studies on deactivation pathways of rhodopsin. Phys. Chem. Chem. Phys. 9, 1648–58.

(108) Hellingwerf, K. J., Hendriks, J., and Gensch, T. (2003) Photoactive Yellow Protein, A New Type of Photoreceptor Protein: Will This “Yellow Lab” Bring Us Where We Want to Go? ‖. J. Phys. Chem. A 107, 1082–1094.

(109) Baca, M., Borgstahl, G. E., Boissinot, M., Burke, P. M., Williams, D. R., Slater, K. A., and Getzoff, E. D. (1994) Complete chemical structure of photoactive yellow protein: novel thioester-linked 4- hydroxycinnamyl chromophore and photocycle chemistry. Biochemistry 33, 14369–77.

(110) Hoff, W. D., Düx, P., Hård, K., Devreese, B., Nugteren-Roodzant, I. M., Crielaard, W., Boelens, R., Kaptein, R., van Beeumen, J., and Hellingwerf, K. J. (1994) Thiol ester-linked p-coumaric acid as a new photoactive prosthetic group in a protein with rhodopsin-like photochemistry. Biochemistry 33, 13959–62.

(111) Briggs, W. R., and Spudich, J. L. (2005) Handbook of Photosensory receptors. Weinheim, Germany.

(112) Meyer, T. E. (1985) Isolation and characterization of soluble cytochromes, ferredoxins and other chromophoric proteins from the halophilic phototrophic bacterium Ectothiorhodospira halophila. Biochim. Biophys. Acta 806, 175–83.

(113) Borgstahl, G. E., Williams, D. R., and Getzoff, E. D. (1995) 1.4 A structure of photoactive yellow protein, a cytosolic photoreceptor: unusual fold, active site, and chromophore. Biochemistry 34, 6278–87.

(114) Sprenger, W. W., Hoff, W. D., Armitage, J. P., and Hellingwerf, K. J. (1993) The eubacterium Ectothiorhodospira halophila is negatively phototactic, with a wavelength dependence that fits the absorption spectrum of the photoactive yellow protein. J. Bacteriol. 175, 3096–104.

162 | P a g e

(115) Devanathan, S., Pacheco, A., Ujj, L., Cusanovich, M., Tollin, G., Lin, S., and Woodbury, N. (1999) Femtosecond spectroscopic observations of initial intermediates in the photocycle of the photoactive yellow protein from Ectothiorhodospira halophila. Biophys. J. 77, 1017–23.

(116) Baltuška, A., van Stokkum, I. H. M., Kroon, A., Monshouwer, R., Hellingwerf, K. J., and van Grondelle, R. (1997) The primary events in the photoactivation of yellow protein. Chem. Phys. Lett. 270, 263–266.

(117) Imamoto, Y., Kataoka, M., Tokunaga, F., Asahi, T., and Masuhara, H. (2001) Primary Photoreaction of Photoactive Yellow Protein Studied by Subpicosecond−Nanosecond Spectroscopy †. Biochemistry 40, 6047–6052.

(118) Van der Horst, M. A., Hendriks, J., Vreede, J., Yeremenko, S., Crielaard, W., and Hellingwerf, K. J. (2005) Photoactive Yellow Protein, the Xanthopsin, in Handbook of Photosensory Receptors (Briggs, W., and Spudich, J., Eds.), pp 391–416. Wiley-VCH, Weinheim, Germany.

(119) Imamoto, Y., Kataoka, M., and Tokunaga, F. (1996) Photoreaction cycle of photoactive yellow protein from Ectothiorhodospira halophila studied by low-temperature spectroscopy. Biochemistry 35, 14047–53.

(120) Brudler, R., Rammelsberg, R., Woo, T. T., Getzoff, E. D., and Gerwert, K. (2001) Structure of the I1 early intermediate of photoactive yellow protein by FTIR spectroscopy. Nat. Struct. Biol. 8, 265– 70.

(121) Takeshita, K., Imamoto, Y., Kataoka, M., Mihara, K., Tokunaga, F., and Terazima, M. (2002) Structural change of site-directed mutants of PYP: new dynamics during pR state. Biophys. J. 83, 1567–77.

(122) Xie, A., Hoff, W. D., Kroon, A. R., and Hellingwerf, K. J. (1996) Glu46 donates a proton to the 4- hydroxycinnamate anion chromophore during the photocycle of photoactive yellow protein. Biochemistry 35, 14671–8.

(123) Xie, A., Kelemen, L., Hendriks, J., White, B. J., Hellingwerf, K. J., and Hoff, W. D. (2001) Formation of a new buried charge drives a large-amplitude protein quake in photoreceptor activation. Biochemistry 40, 1510–7.

(124) Van Brederode, M. E., Hoff, W. D., Van Stokkum, I. H., Groot, M. L., and Hellingwerf, K. J. (1996) Protein folding thermodynamics applied to the photocycle of the photoactive yellow protein. Biophys. J. 71, 365–80.

(125) Hendriks, J., van Stokkum, I. H. M., Crielaard, W., and Hellingwerf, K. J. (1999) Kinetics of and intermediates in a photocycle branching reaction of the photoactive yellow protein from Ectothiorhodospira halophila. FEBS Lett. 458, 252–256.

(126) W. L. Butler, K. H. Norris, H. W. Siegelman, S. B. H. (1959) Detaction, assay, and preliminary purification of the pigment controlling photoresponsive development of plants. Proc. Natl. Acad. Sci. U. S. A. 1703–1708.

(127) Franklin, K. A., Larner, V. S., and Whitelam, G. C. (2005) The signal transducing photoreceptors of plants. Int. J. Dev. Biol. 49, 653–64.

(128) Quail, P. H. (2002) Photosensory perception and signalling in plant cells: new paradigms? Curr. Opin. Cell Biol. 14, 180–188.

163 | P a g e

(129) Blumenstein, A., Vienken, K., Tasler, R., Purschwitz, J., Veith, D., Frankenberg-Dinkel, N., and Fischer, R. (2005) The Aspergillus nidulans phytochrome FphA represses sexual development in red light. Curr. Biol. 15, 1833–8.

(130) Corrochano, L. M. (2007) Fungal photoreceptors: sensory molecules for fungal development and behaviour. Photochem. Photobiol. Sci. 6, 725–36.

(131) Brandt, S., von Stetten, D., Günther, M., Hildebrandt, P., and Frankenberg-Dinkel, N. (2008) The fungal phytochrome FphA from Aspergillus nidulans. J. Biol. Chem. 283, 34605–14.

(132) Froehlich, A. C., Noh, B., Vierstra, R. D., Loros, J., and Dunlap, J. C. (2005) Genetic and molecular analysis of phytochromes from the filamentous fungus Neurospora crassa. Eukaryot. Cell 4, 2140–52.

(133) Wu, S. H., McDowell, M. T., and Lagarias, J. C. (1997) Phycocyanobilin is the natural precursor of the phytochrome chromophore in the green alga Mesotaenium caldariorum. J. Biol. Chem. 272, 25700–5.

(134) WU, S.-H., and LAGARIAS, J. C. (1997) The phytochrome photoreceptor in the green alga Mesotaenium caldariorum: implication for a conserved mechanism of phytochrome action. Plant, Cell Environ. 20, 691–699.

(135) De Riso, V., Raniello, R., Maumus, F., Rogato, A., Bowler, C., and Falciatore, A. (2009) Gene silencing in the marine diatom Phaeodactylum tricornutum. Nucleic Acids Res. 37, e96.

(136) Davis, S. J. (1999) Bacteriophytochromes: Phytochrome-Like Photoreceptors from Nonphotosynthetic Eubacteria. Science (80-. ). 286, 2517–2520.

(137) Karniol, B., Wagner, J. R., Walker, J. M., and Vierstra, R. D. (2005) Phylogenetic analysis of the phytochrome superfamily reveals distinct microbial subfamilies of photoreceptors. Biochem. J. 392, 103–16.

(138) Rockwell, N. C., and Lagarias, J. C. (2010) A brief history of phytochromes. Chemphyschem 11, 1172–80.

(139) Auldridge, M. E., and Forest, K. T. (2011) Bacterial phytochromes: more than meets the light. Crit. Rev. Biochem. Mol. Biol. 46, 67–88.

(140) Quest, B., Hübschmann, T., Sharda, S., Tandeau de Marsac, N., and Gärtner, W. (2007) Homologous expression of a bacterial phytochrome. The cyanobacterium Fremyella diplosiphon incorporates biliverdin as a genuine, functional chromophore. FEBS J. 274, 2088–98.

(141) Lamparter, T., Esteban, B., and Hughes, J. (2001) Phytochrome Cph1 from the cyanobacterium Synechocystis PCC6803. Purification, assembly, and quaternary structure. Eur. J. Biochem. 268, 4720–30.

(142) Ulijasz, A. T., Cornilescu, G., von Stetten, D., Cornilescu, C., Velazquez Escobar, F., Zhang, J., Stankey, R. J., Rivera, M., Hildebrandt, P., and Vierstra, R. D. (2009) Cyanochromes are blue/green light photoreversible photoreceptors defined by a stable double cysteine linkage to a phycoviolobilin-type chromophore. J. Biol. Chem. 284, 29757–72.

(143) Song, C., Psakis, G., Lang, C., Mailliet, J., Gärtner, W., Hughes, J., and Matysik, J. (2011) Two ground state isoforms and a chromophore D-ring photoflip triggering extensive intramolecular changes in a canonical phytochrome. Proc. Natl. Acad. Sci. U. S. A. 108, 3842–7.

164 | P a g e

(144) Van Thor, J. J., Mackeen, M., Kuprov, I., Dwek, R. A., and Wormald, M. R. (2006) Chromophore structure in the photocycle of the cyanobacterial phytochrome Cph1. Biophys. J. 91, 1811–22.

(145) Tu, S.-L., and Lagarias, J. C. (2005) The Phytochromes, in Handbook of Photosensory Receptors (Briggs, W. R., and Spudich, J. L., Eds.), pp 121–150. Wiley-VCH, Weinheim, Germany.

(146) Butler, W. L., Lane, H. C., and Siegelman, H. W. (1963) Nonphotochemical Transformations of Phytochrome in Vivo. Plant Physiol. 38, 514–9.

(147) Montgomery, B. L., and Lagarias, J. C. (2002) Phytochrome ancestry: sensors of bilins and light. Trends Plant Sci. 7, 357–366.

(148) Wagner, J. R., Brunzelle, J. S., Forest, K. T., and Vierstra, R. D. (2005) A light-sensing knot revealed by the structure of the chromophore-binding domain of phytochrome. Nature 438, 325–31.

(149) Yeh, K. (1997) A Cyanobacterial Phytochrome Two-Component Light Sensory System. Science (80-. ). 277, 1505–1508.

(150) Lamparter, T., Michael, N., Mittmann, F., and Esteban, B. (2002) Phytochrome from Agrobacterium tumefaciens has unusual spectral properties and reveals an N-terminal chromophore attachment site. Proc. Natl. Acad. Sci. U. S. A. 99, 11628–33.

(151) Karniol, B., and Vierstra, R. D. (2003) The pair of bacteriophytochromes from Agrobacterium tumefaciens are histidine kinases with opposing photobiological properties. Proc. Natl. Acad. Sci. U. S. A. 100, 2807–12.

(152) Bhoo, S. H., Davis, S. J., Walker, J., Karniol, B., and Vierstra, R. D. (2001) Bacteriophytochromes are photochromic histidine kinases using a biliverdin chromophore. Nature 414, 776–9.

(153) Yeh, K. C., and Lagarias, J. C. (1998) Eukaryotic phytochromes: light-regulated serine/threonine protein kinases with histidine kinase ancestry. Proc. Natl. Acad. Sci. U. S. A. 95, 13976–81.

(154) Tarutina, M., Ryjenkov, D. a, and Gomelsky, M. (2006) An unorthodox bacteriophytochrome from Rhodobacter sphaeroides involved in turnover of the second messenger c-di-GMP. J. Biol. Chem. 281, 34751–34758.

(155) Kaneko, T., Sato, S., Kotani, H., Tanaka, A., Asamizu, E., Nakamura, Y., Miyajima, N., Hirosawa, M., Sugiura, M., Sasamoto, S., Kimura, T., Hosouchi, T., Matsuno, A., Muraki, A., Nakazaki, N., Naruo, K., Okumura, S., Shimpo, S., Takeuchi, C., Wada, T., Watanabe, A., Yamada, M., Yasuda, M., and Tabata, S. (1996) Sequence analysis of the genome of the unicellular cyanobacterium Synechocystis sp. strain PCC6803. II. Sequence determination of the entire genome and assignment of potential protein-coding regions. DNA Res. 3, 109–36.

(156) Kaneko, T., Tanaka, A., Sato, S., Kotani, H., Sazuka, T., Miyajima, N., Sugiura, M., and Tabata, S. (1995) Sequence analysis of the genome of the unicellular cyanobacterium Synechocystis sp. strain PCC6803. I. Sequence features in the 1 Mb region from map positions 64% to 92% of the genome. DNA Res. 2, 153–66, 191–8.

(157) Ohmori, M., Ikeuchi, M., Sato, N., Wolk, P., Kaneko, T., Ogawa, T., Kanehisa, M., Goto, S., Kawashima, S., Okamoto, S., Yoshimura, H., Katoh, H., Fujisawa, T., Ehira, S., Kamei, A., Yoshihara, S., Narikawa, R., and Tabat, S. (2001) Characterization of genes encoding multi-domain proteins in the genome of the filamentous nitrogen-fixing Cyanobacterium anabaena sp. strain PCC 7120. DNA Res. 8, 271–84.

165 | P a g e

(158) Herdman, M., Coursin, T., Rippka, R., Houmard, J., and Tandeau de Marsac, N. (2000) A new appraisal of the prokaryotic origin of eukaryotic phytochromes. J. Mol. Evol. 51, 205–13.

(159) Yoshihara, S., Katayama, M., Geng, X., and Ikeuchi, M. (2004) Cyanobacterial phytochrome-like PixJ1 holoprotein shows novel reversible photoconversion between blue- and green-absorbing forms. Plant Cell Physiol. 45, 1729–37.

(160) Yoshihara, S., Shimada, T., Matsuoka, D., Zikihara, K., Kohchi, T., and Tokutomi, S. (2006) Reconstitution of blue-green reversible photoconversion of a cyanobacterial photoreceptor, PixJ1, in phycocyanobilin-producing Escherichia coli. Biochemistry 45, 3775–84.

(161) Hirose, Y., Shimada, T., Narikawa, R., Katayama, M., and Ikeuchi, M. (2008) Cyanobacteriochrome CcaS is the green light receptor that induces the expression of phycobilisome linker protein. Proc. Natl. Acad. Sci. U. S. A. 105, 9528–33.

(162) Rockwell, N. C., Martin, S. S., Gulevich, A. G., and Lagarias, J. C. (2012) Phycoviolobilin formation and spectral tuning in the DXCF cyanobacteriochrome subfamily. Biochemistry 51, 1449– 1463.

(163) Rockwell, N. C., Martin, S. S., Feoktistova, K., and Lagarias, J. C. (2011) Diverse two-cysteine photocycles in phytochromes and cyanobacteriochromes. Proc. Natl. Acad. Sci. U. S. A. 108, 11854– 9.

(164) Enomoto, G., Hirose, Y., Narikawa, R., and Ikeuchi, M. (2012) Thiol-based photocycle of the blue and teal light-sensing cyanobacteriochrome Tlr1999. Biochemistry 51, 3050–8.

(165) Narikawa, R., Fukushima, Y., Ishizuka, T., Itoh, S., and Ikeuchi, M. (2008) A novel photoactive GAF domain of cyanobacteriochrome AnPixJ that shows reversible green/red photoconversion. J. Mol. Biol. 380, 844–55.

(166) Kim, P. W., Freer, L. H., Rockwell, N. C., Martin, S. S., Lagarias, J. C., and Larsen, D. S. (2012) Femtosecond photodynamics of the red/green cyanobacteriochrome NpR6012g4 from Nostoc punctiforme. 1. Forward dynamics. Biochemistry 51, 608–18.

(167) Kim, P. W., Freer, L. H., Rockwell, N. C., Martin, S. S., Lagarias, J. C., and Larsen, D. S. (2012) Femtosecond photodynamics of the red/green cyanobacteriochrome NpR6012g4 from Nostoc punctiforme. 1. Forward dynamics. Biochemistry 51, 608–18.

(168) Chang, C.-W., Gottlieb, S. M., Kim, P. W., Rockwell, N. C., Lagarias, J. C., and Larsen, D. S. (2013) Reactive Ground-State Pathways Are Not Ubiquitous in Red/Green Cyanobacteriochromes. J. Phys. Chem. B.

(169) Chen, Y., Zhang, J., Luo, J., Tu, J.-M., Zeng, X.-L., Xie, J., Zhou, M., Zhao, J.-Q., Scheer, H., and Zhao, K.-H. (2012) Photophysical diversity of two novel cyanobacteriochromes with phycocyanobilin chromophores: photochemistry and dark reversion kinetics. FEBS J. 279, 40–54.

(170) Hirose, Y., Narikawa, R., Katayama, M., and Ikeuchi, M. (2010) Cyanobacteriochrome CcaS regulates phycoerythrin accumulation in Nostoc punctiforme, a group II chromatic adapter. Proc. Natl. Acad. Sci. U. S. A. 107, 8854–9.

(171) Hirose, Y., Rockwell, N. C., Nishiyama, K., Narikawa, R., Ukaji, Y., Inomata, K., Lagarias, J. C., and Ikeuchi, M. (2013) Green/red cyanobacteriochromes regulate complementary chromatic acclimation via a protochromic photocycle. Proc. Natl. Acad. Sci. U. S. A. 110, 4974–9.

166 | P a g e

(172) Terauchi, K., Montgomery, B. L., Grossman, A. R., Lagarias, J. C., and Kehoe, D. M. (2004) RcaE is a complementary chromatic adaptation photoreceptor required for green and red light responsiveness. Mol. Microbiol. 51, 567–77.

(173) Kehoe, D. M., and Grossman, A. R. (1996) Similarity of a chromatic adaptation sensor to phytochrome and ethylene receptors. Science 273, 1409–12.

(174) Ishizuka, T., Shimada, T., Okajima, K., Yoshihara, S., Ochiai, Y., Katayama, M., and Ikeuchi, M. (2006) Characterization of cyanobacteriochrome TePixJ from a thermophilic cyanobacterium Thermosynechococcus elongatus strain BP-1. Plant Cell Physiol. 47, 1251–61.

(175) Ishizuka, T., Narikawa, R., Kohchi, T., Katayama, M., and Ikeuchi, M. (2007) Cyanobacteriochrome TePixJ of Thermosynechococcus elongatus harbors phycoviolobilin as a chromophore. Plant Cell Physiol. 48, 1385–90.

(176) Burgie, E. S., Walker, J. M., Phillips, G. N., and Vierstra, R. D. (2013) A photo-labile thioether linkage to phycoviolobilin provides the foundation for the blue/green photocycles in DXCF- cyanobacteriochromes. Structure 21, 88–97.

(177) Yoshihara, S., Suzuki, F., Fujita, H., Geng, X. X., and Ikeuchi, M. (2000) Novel putative photoreceptor and regulatory genes Required for the positive phototactic movement of the unicellular motile cyanobacterium Synechocystis sp. PCC 6803. Plant Cell Physiol. 41, 1299–304.

(178) Bhaya, D., Takahashi, A., and Grossman, A. R. (2001) Light regulation of type IV pilus- dependent motility by chemosensor-like elements in Synechocystis PCC6803. Proc. Natl. Acad. Sci. U. S. A. 98, 7540–5.

(179) Rockwell, N. C., Njuguna, S. L., Roberts, L., Castillo, E., Parson, V. L., Dwojak, S., Lagarias, J. C., and Spiller, S. C. (2008) A second conserved GAF domain cysteine is required for the blue/green photoreversibility of cyanobacteriochrome Tlr0924 from Thermosynechococcus elongatus. Biochemistry 47, 7304–16.

(180) Freer, L. H., Kim, P. W., Corley, S. C., Rockwell, N. C., Zhao, L., Thibert, A. J., Lagarias, J. C., and Larsen, D. S. (2012) Chemical inhomogeneity in the ultrafast dynamics of the DXCF cyanobacteriochrome Tlr0924. J. Phys. Chem. B 116, 10571–81.

(181) Song, J.-Y., Cho, H. S., Cho, J.-I., Jeon, J.-S., Lagarias, J. C., and Park, Y.-I. (2011) Near-UV cyanobacteriochrome signaling system elicits negative phototaxis in the cyanobacterium Synechocystis sp. PCC 6803. Proc. Natl. Acad. Sci. U. S. A. 108, 10780–5.

(182) Narikawa, R., Suzuki, F., Yoshihara, S., Higashi, S., Watanabe, M., and Ikeuchi, M. (2011) Novel photosensory two-component system (PixA-NixB-NixC) involved in the regulation of positive and negative phototaxis of cyanobacterium Synechocystis sp. PCC 6803. Plant Cell Physiol. 52, 2214–24.

(183) Narikawa, R., Ishizuka, T., Muraki, N., Shiba, T., Kurisu, G., and Ikeuchi, M. (2013) Structures of cyanobacteriochromes from phototaxis regulators AnPixJ and TePixJ reveal general and specific photoconversion mechanism. Proc. Natl. Acad. Sci. U. S. A. 110, 918–23.

(184) Rockwell, N. C., Martin, S. S., and Lagarias, J. C. (2012) Red/green cyanobacteriochromes: sensors of color and power. Biochemistry 51, 9667–77.

(185) Kim, P. W., Freer, L. H., Rockwell, N. C., Martin, S. S., Lagarias, J. C., and Larsen, D. S. (2012) Femtosecond photodynamics of the red/green cyanobacteriochrome NpR6012g4 from Nostoc punctiforme. 2. reverse dynamics. Biochemistry 51, 619–30.

167 | P a g e

(186) Kim, P. W., Freer, L. H., Rockwell, N. C., Martin, S. S., Lagarias, J. C., and Larsen, D. S. (2012) Femtosecond photodynamics of the red/green cyanobacteriochrome NpR6012g4 from Nostoc punctiforme. 2. reverse dynamics. Biochemistry 51, 619–30.

(187) Fukushima, Y., Iwaki, M., Narikawa, R., Ikeuchi, M., Tomita, Y., and Itoh, S. (2011) Photoconversion mechanism of a green/red photosensory cyanobacteriochrome AnPixJ: time- resolved optical spectroscopy and FTIR analysis of the AnPixJ-GAF2 domain. Biochemistry 50, 6328– 39.

(188) Lim, S., Rockwell, N. C., Martin, S. S., Lagarias, J. C., and Ames, J. B. (2013) (1)H, (15)N, and (13)C chemical shift assignments of cyanobacteriochrome NpF2164g3 in the photoproduct state. Biomol. NMR Assign.

(189) Ishizuka, T., Kakuda, M., Araki, R., and Yawo, H. (2006) Kinetic evaluation of photosensitivity in genetically engineered neurons expressing green algae light-gated channels. Neurosci. Res. 54, 85– 94.

(190) Ishizuka, T., Kamiya, A., Suzuki, H., Narikawa, R., Noguchi, T., Kohchi, T., Inomata, K., and Ikeuchi, M. (2011) The cyanobacteriochrome, TePixJ, isomerizes its own chromophore by converting phycocyanobilin to phycoviolobilin. Biochemistry 50, 953–61.

(191) Pellequer, J. L., Wager-Smith, K. A., Kay, S. A., and Getzoff, E. D. (1998) Photoactive yellow protein: a structural prototype for the three-dimensional fold of the PAS domain superfamily. Proc. Natl. Acad. Sci. U. S. A. 95, 5884–90.

(192) Palczewski, K. (2000) Crystal Structure of Rhodopsin: A G Protein-Coupled Receptor. Science (80-. ). 289, 739–745.

(193) Lundstrom, K. (2009) An overview on GPCRs and drug discovery: structure-based drug design and structural biology on GPCRs. Methods Mol. Biol. 552, 51–66.

(194) Levskaya, A., Chevalier, A. A., Tabor, J. J., Simpson, Z. B., Lavery, L. A., Levy, M., Davidson, E. A., Scouras, A., Ellington, A. D., Marcotte, E. M., and Voigt, C. A. (2005) Synthetic biology: engineering Escherichia coli to see light. Nature 438, 441–2.

(195) Drepper, T., Krauss, U., Meyer zu Berstenhorst, S., Pietruszka, J., and Jaeger, K.-E. (2011) Lights on and action! Controlling microbial gene expression by light. Appl. Microbiol. Biotechnol. 90, 23–40.

(196) Lin, J. Y., Knutsen, P. M., Muller, A., Kleinfeld, D., and Tsien, R. Y. (2013) ReaChR: a red-shifted variant of channelrhodopsin enables deep transcranial optogenetic excitation. Nat. Neurosci. 16, 1499–508.

(197) Mathes, T., Vogl, C., Stolz, J., and Hegemann, P. (2009) In vivo generation of flavoproteins with modified cofactors. J. Mol. Biol. 385, 1511–8.

(198) Christie, J. M., Corchnoy, S. B., Swartz, T. E., Hokenson, M., Han, I.-S., Briggs, W. R., and Bogomolni, R. A. (2007) Steric interactions stabilize the signaling state of the LOV2 domain of phototropin 1. Biochemistry 46, 9310–9.

(199) Tsien, R. Y. (2009) Constructing and exploiting the fluorescent protein paintbox (Nobel Lecture). Angew. Chem. Int. Ed. Engl. 48, 5612–26.

(200) Chapman, S., Faulkner, C., Kaiserli, E., Garcia-Mata, C., Savenkov, E. I., Roberts, A. G., Oparka, K. J., and Christie, J. M. (2008) The photoreversible fluorescent protein iLOV outperforms GFP as a reporter of plant virus infection. Proc. Natl. Acad. Sci. U. S. A. 105, 20038–43.

168 | P a g e

(201) Drepper, T., Eggert, T., Circolone, F., Heck, A., Krauss, U., Guterl, J.-K., Wendorff, M., Losi, A., Gärtner, W., and Jaeger, K.-E. (2007) Reporter proteins for in vivo fluorescence without oxygen. Nat. Biotechnol. 25, 443–5.

(202) Fischer, A. J., and Lagarias, J. C. (2004) Harnessing phytochrome’s glowing potential. Proc. Natl. Acad. Sci. U. S. A. 101, 17334–9.

(203) Shu, X., Royant, A., Lin, M. Z., Aguilera, T. a, Lev-Ram, V., Steinbach, P. a, and Tsien, R. Y. (2009) Mammalian expression of infrared fluorescent proteins engineered from a bacterial phytochrome. Science 324, 804–7.

(204) Murphy, J. T., and Lagarias, J. C. (1997) The phytofluors: a new class of fluorescent protein probes. Curr. Biol. 7, 870–6.

(205) Lagarias, J. C., and Murphy, J. T. (2000) Phytofluors as fluorescent labels.

(206) Stolik, S., Delgado, J. ., Pérez, A., and Anasagasti, L. (2000) Measurement of the penetration depths of red and near infrared light in human “ex vivo” tissues. J. Photochem. Photobiol. B Biol. 57, 90–93.

(207) Strickland, D., Moffat, K., and Sosnick, T. R. (2008) Light-activated DNA binding in a designed allosteric protein. Proc. Natl. Acad. Sci. U. S. A. 105, 10709–14.

(208) Lee, J., Natarajan, M., Nashine, V. C., Socolich, M., Vo, T., Russ, W. P., Benkovic, S. J., and Ranganathan, R. (2008) Surface sites for engineering allosteric control in proteins. Science 322, 438– 42.

(209) Möglich, A., Ayers, R. A., and Moffat, K. (2009) Design and signaling mechanism of light- regulated histidine kinases. J. Mol. Biol. 385, 1433–44.

(210) Wu, Y. I., Frey, D., Lungu, O. I., Jaehrig, A., Schlichting, I., Kuhlman, B., and Hahn, K. M. (2009) A genetically encoded photoactivatable Rac controls the motility of living cells. Nature 461, 104–8.

(211) Leung, D. W., Otomo, C., Chory, J., and Rosen, M. K. (2008) Genetically encoded photoswitching of actin assembly through the Cdc42-WASP-Arp2/3 complex pathway. Proc. Natl. Acad. Sci. U. S. A. 105, 12797–802.

(212) Bi, A., Cui, J., Ma, Y.-P., Olshevskaya, E., Pu, M., Dizhoor, A. M., and Pan, Z.-H. (2006) Ectopic expression of a microbial-type rhodopsin restores visual responses in mice with photoreceptor degeneration. Neuron 50, 23–33.

(213) Boyden, E. S., Zhang, F., Bamberg, E., Nagel, G., and Deisseroth, K. (2005) Millisecond- timescale, genetically targeted optical control of neural activity. Nat. Neurosci. 8, 1263–8.

(214) Li, X., Gutierrez, D. V, Hanson, M. G., Han, J., Mark, M. D., Chiel, H., Hegemann, P., Landmesser, L. T., and Herlitze, S. (2005) Fast noninvasive activation and inhibition of neural and network activity by vertebrate rhodopsin and green algae channelrhodopsin. Proc. Natl. Acad. Sci. U. S. A. 102, 17816–21.

(215) Schröder-Lang, S., Schwärzel, M., Seifert, R., Strünker, T., Kateriya, S., Looser, J., Watanabe, M., Kaupp, U. B., Hegemann, P., and Nagel, G. (2007) Fast manipulation of cellular cAMP level by light in vivo. Nat. Methods 4, 39–42.

(216) Kohen, E., Santus, R., and Hirschberg, J. G. (1995) Spectroscopy and Instrumentation, in Photobiology, pp 43–69. Academic Press, San Diego.

169 | P a g e

(217) Richards, F. M., and Richmond, T. (1978) Solvents, interfaces and protein structure, in Ciba Foundation Symposium 60 - Molecular Interactions and Activity in Proteins (Porter, R., and Fitzsimons, D. W., Eds.), pp 23–45. John Wiley & Sons, Ltd., Chichester, UK.

(218) Atkins, P., and de Paula, J. (2002) Spectroscopy 1: Rotational and vibrational spectra, in Atkins’ Physical Chemistry 7th ed., pp 481–537. Oxford University Press, Oxford.

(219) Fink, A. L., and Geeves, M. A. (1979) Cryoenzymology: The study of enzyme cataysis at subzero temperatures, in Methods in enzymology. Vol.63, Enzyme kinetics and mechanism. Part A : Initial rate and inhibitor methods (Colowick, S. P., Kaplan, N. O., and Purich, D. L., Eds.), pp 336–370. Academic Press, London.

(220) Douzou, P. (1977) Properties of Mixed Solvents as a Function of Temperature, in Cryobiochemistry, pp 11–90. Academic Press, London.

(221) Ramløv, H. (2000) Aspects of natural cold tolerance in ectothermic animals. Hum. Reprod. 15 Suppl 5, 26–46.

(222) Fuller, B. J. Cryoprotectants: the essential antifreezes to protect life in the frozen state. Cryo Letters 25, 375–88.

(223) Vitkup, D., Ringe, D., Petsko, G. A., and Karplus, M. (2000) Solvent mobility and the protein “glass” transition. Nat. Struct. Biol. 7, 34–8.

(224) Ringe, D., and Petsko, G. A. (2003) The “glass transition” in protein dynamics: what it is, why it occurs, and how to exploit it. Biophys. Chem. 105, 667–680.

(225) Hollas, J. M. (2004) Vibrational spectroscopy, in Modern Spectroscopy 4th ed., pp 137–198. John Wiley & Sons, Ltd, Chichester.

(226) MAIMAN, T. H. (1960) Stimulated Optical Radiation in Ruby. Nature 187, 493–494.

(227) Porter, G. (1967) Flash photolysis and some of its applications. Nobel Lect.

(228) Norrish, R. G. W. (1967) Some fast reactions in gases studied by flash photolysis and kinetic spectroscopy. Nobel Lect.

(229) NORRISH, R. G. W., and PORTER, G. (1949) Chemical Reactions Produced by Very High Light Intensities. Nature 164, 658–658.

(230) Fork, R. L. (1981) Generation of optical pulses shorter than 0.1 psec by colliding pulse mode locking. Appl. Phys. Lett. 38, 671.

(231) Busch, G. E., Applebury, M. L., Lamola, A. A., and Rentzepis, P. M. (1972) Formation and decay of prelumirhodopsin at room temperatures. Proc. Natl. Acad. Sci. U. S. A. 69, 2802–6.

(232) Netzel, T. L., Rentzepis, P. M., and Leigh, J. (1973) Picosecond kinetics of reaction centers containing bacteriochlorophyll. Science 182, 238–41.

(233) Kaufmann, K. J., Dutton, P. L., Netzel, T. L., Leigh, J. S., and Rentzepis, P. M. (1975) Picosecond kinetics of events leading to reaction center bacteriochlorophyll oxidation. Science 188, 1301–4.

170 | P a g e

(234) Rockley, M. G., Windsor, M. W., Cogdell, R. J., and Parson, W. W. (1975) Picosecond detection of an intermediate in the photochemical reaction of bacterial photosynthesis. Proc. Natl. Acad. Sci. U. S. A. 72, 2251–5.

(235) Berera, R., van Grondelle, R., and Kennis, J. T. M. Ultrafast transient absorption spectroscopy: principles and application to photosynthetic systems. Photosynth. Res. 101, 105–18.

(236) Greetham, G. M., Burgos, P., Cao, Q., Clark, I. P., Codd, P. S., Farrow, R. C., George, M. W., Kogimtzis, M., Matousek, P., Parker, A. W., Pollard, M. R., Robinson, D. A., Xin, Z.-J., and Towrie, M. (2010) ULTRA: A Unique Instrument for Time-Resolved Spectroscopy. Appl. Spectrosc. 64, 1311–9.

(237) Van Stokkum, I. H. M., Larsen, D. S., and van Grondelle, R. (2004) Global and target analysis of time-resolved spectra. Biochim. Biophys. Acta 1657, 82–104.

(238) Hollas, J. M. (2004) Electromagnetic radiation and its interaction with atoms and molecules, in Modern Spectroscopy 4th ed., pp 27–39. John Wiley & Sons, Ltd, Chichester.

(239) Hughes, J., Lamparter, T., Mittmann, F., Hartmann, E., Gärtner, W., Wilde, A., and Börner, T. (1997) A prokaryotic phytochrome. Nature 386, 663.

(240) Pathak, G. P., Vrana, J. D., and Tucker, C. L. (2013) Optogenetic control of cell function using engineered photoreceptors. Biol. Cell 105, 59–72.

(241) http://genome.microbedb.jp/cyanobase/Thermo/genes/tlr0924.

(242) Yamaoka, T., Satoh, K., and Katoh, S. (1978) Photosynthetic activities of a thermophilic blue- green alga. Plant Cell Physiol. 19, 943–954.

(243) Nakamura, Y., Kaneko, T., Sato, S., Ikeuchi, M., Katoh, H., Sasamoto, S., Watanabe, A., Iriguchi, M., Kawashima, K., Kimura, T., Kishida, Y., Kiyokawa, C., Kohara, M., Matsumoto, M., Matsuno, A., Nakazaki, N., Shimpo, S., Sugimoto, M., Takeuchi, C., Yamada, M., and Tabata, S. (2002) Complete genome structure of the thermophilic cyanobacterium Thermosynechococcus elongatus BP-1 (supplement). DNA Res. 9, 135–48.

(244) http://prosite.expasy.org/cgi-bin/prosite/ScanView.cg.

(245) Gambetta, G. a, and Lagarias, J. C. (2001) Genetic engineering of phytochrome biosynthesis in bacteria. Proc. Natl. Acad. Sci. U. S. A. 98, 10566–71.

(246) Cornilescu, C. C., Cornilescu, G., Burgie, E. S., Markley, J. L., Ulijasz, A. T., and Vierstra, R. D. (2013) Dynamic Structural Changes Underpin Photoconversion of a Blue/Green Cyanobacteriochrome Between its Dark and Photoactivated States. J. Biol. Chem.

(247) Thümmler, F., and Rüdiger, W. (1983) Models for the photoreversibility of phytochrome. Tetrahedron 39, 1943–1951.

(248) Rüdiger, W., Thümmler, F., Cmiel, E., and Schneider, S. (1983) Chromophore structure of the physiologically active form (P(fr)) of phytochrome. Proc. Natl. Acad. Sci. U. S. A. 80, 6244–8.

(249) Smythe, C. V. (1936) The reaction of iodoacetate and iodoacetamide with various sulfhydryl groups, with urease, and with yeast preparations. J. Biol. Chem. 601–612.

(250) Luo, D., Smith, S. W., and Anderson, B. D. (2005) Kinetics and mechanism of the reaction of cysteine and hydrogen peroxide in aqueous solution. J. Pharm. Sci. 94, 304–16.

171 | P a g e

(251) Litts, J. C., Kelly, J. M., and Lagarias, J. C. (1983) Structure-function studies on phytochrome. Preliminary characterization of highly purified phytochrome from Avena sativa enriched in the 124- kilodalton species. J. Biol. Chem. 258, 11025–31.

(252) Chen, L., Kinoshita, H., and Inomata, K. (2009) Synthesis of Doubly Locked 5Zs15Za-Biliverdin Derivatives and Their Unique Spectral Behavior. Chem. Lett. 38, 602–603.

(253) Lamparter, T. (2004) Evolution of cyanobacterial and plant phytochromes. FEBS Lett. 573, 1–5.

(254) Müller, M. G., Lindner, I., Martin, I., Gärtner, W., and Holzwarth, A. R. (2008) Femtosecond kinetics of photoconversion of the higher plant photoreceptor phytochrome carrying native and modified chromophores. Biophys. J. 94, 4370–82.

(255) Foerstendorf, H., Mummert, E., Schäfer, E., Scheer, H., and Siebert, F. (1996) Fourier-transform infrared spectroscopy of phytochrome: difference spectra of the intermediates of the photoreactions. Biochemistry 35, 10793–9.

(256) Heyne, K., Herbst, J., Stehlik, D., Esteban, B., Lamparter, T., Hughes, J., and Diller, R. (2002) Ultrafast dynamics of phytochrome from the cyanobacterium synechocystis, reconstituted with phycocyanobilin and phycoerythrobilin. Biophys. J. 82, 1004–16.

(257) Jones, A. R., Russell, H. J., Greetham, G. M., Towrie, M., Hay, S., and Scrutton, N. S. (2012) Ultrafast infrared spectral fingerprints of vitamin B12 and related cobalamins. J. Phys. Chem. A 116, 5586–94.

(258) Snellenburg, J. J., Laptenok, S. P., Seger, R., Mullen, K. M., and van Stokkum, I. H. M. (2012) Glotaran: a Java -based graphical user interface for the R-package TIMP. J. Stat. Softw. 49, 1–23.

(259) Lorenc, M., Ziolek, M., Naskrecki, R., Karolczak, J., Kubicki, J., and Maciejewski, A. (2002) Artifacts in femtosecond transient absorption spectroscopy. Appl. Phys. B Lasers Opt. 74, 19–27.

(260) Bischoff, M., Hermann, G., Rentsch, S., and Strehlow, D. (2001) First Steps in the Phytochrome Phototransformation: A Comparative Femtosecond Study on the Forward (Pr → Pfr) and Back Reaction (Pfr → Pr) †. Biochemistry 40, 181–186.

(261) Van Thor, J. J., Ronayne, K. L., and Towrie, M. (2007) Formation of the early photoproduct lumi- R of cyanobacterial phytochrome cph1 observed by ultrafast mid-infrared spectroscopy. J. Am. Chem. Soc. 129, 126–32.

(262) Schumann, C., Gross, R., Wolf, M. M. N., Diller, R., Michael, N., and Lamparter, T. (2008) Subpicosecond midinfrared spectroscopy of the Pfr reaction of phytochrome Agp1 from Agrobacterium tumefaciens. Biophys. J. 94, 3189–97.

(263) Van Wilderen, L. J. G. W., Clark, I. P., Towrie, M., and van Thor, J. J. (2009) Mid-infrared picosecond pump-dump-probe and pump-repump-probe experiments to resolve a ground-state intermediate in cyanobacterial phytochrome Cph1. J. Phys. Chem. B 113, 16354–64.

(264) Gottlieb, S. M., Kim, P. W., Rockwell, N. C., Hirose, Y., Ikeuchi, M., Lagarias, J. C., and Larsen, D. S. (2013) Primary photodynamics of the green/red-absorbing photoswitching regulator of the chromatic adaptation E domain from Fremyella diplosiphon. Biochemistry 52, 8198–208.

(265) Mroginski, M. A., Kaminski, S., von Stetten, D., Ringsdorf, S., Gärtner, W., Essen, L.-O., and Hildebrandt, P. (2011) Structure of the chromophore binding pocket in the Pr state of plant phytochrome phyA. J. Phys. Chem. B 115, 1220–31.

172 | P a g e

(266) Toh, K. C., Stojkovid, E. a, Rupenyan, A. B., van Stokkum, I. H. M., Salumbides, M., Groot, M.-L., Moffat, K., and Kennis, J. T. M. (2011) Primary reactions of bacteriophytochrome observed with ultrafast mid-infrared spectroscopy. J. Phys. Chem. A 115, 3778–86.

(267) Piwowarski, P., Ritter, E., Hofmann, K.-P., Hildebrandt, P., von Stetten, D., Scheerer, P., Michael, N., Lamparter, T., and Bartl, F. (2010) Light-induced activation of bacterial phytochrome Agp1 monitored by static and time-resolved FTIR spectroscopy. Chemphyschem 11, 1207–14.

(268) Andel, F., Lagarias, J. C., and Mathies, R. A. (1996) Resonance raman analysis of chromophore structure in the lumi-R photoproduct of phytochrome. Biochemistry 35, 15997–6008.

(269) Foerstendorf, H., Benda, C., Gärtner, W., Storf, M., Scheer, H., and Siebert, F. (2001) FTIR studies of phytochrome photoreactions reveal the C=O bands of the chromophore: consequences for its protonation states, conformation, and protein interaction. Biochemistry 40, 14952–9.

(270) Kim, P. W., Rockwell, N. C., Freer, L. H., Chang, C.-W., Martin, S. S., Lagarias, J. C., and Larsen, D. S. (2013) Unraveling the Primary Isomerization Dynamics in Cyanobacterial Phytochrome Cph1 with Multi-pulse Manipulations. J. Phys. Chem. Lett. 4, 2605–2609.

(271) Chizhov, I., Zorn, B., Manstein, D. J., and Gärtner, W. (2013) Kinetic and thermodynamic analysis of the light-induced processes in plant and cyanobacterial phytochromes. Biophys. J. 105, 2210–20.

(272) Durin, G., Delaunay, A., Darnault, C., Heyes, D. J., Royant, A., Vernede, X., Hunter, C. N., Weik, M., and Bourgeois, D. (2009) Simultaneous measurements of solvent dynamics and functional kinetics in a light-activated enzyme. Biophys. J. 96, 1902–10.

(273) Heyes, D. J., Ruban, A. V, Wilks, H. M., and Hunter, C. N. (2002) Enzymology below 200 K: the kinetics and thermodynamics of the photochemistry catalyzed by protochlorophyllide oxidoreductase. Proc. Natl. Acad. Sci. U. S. A. 99, 11145–50.

(274) Heyes, D. J., Ruban, A. V, and Hunter, C. N. (2003) Protochlorophyllide oxidoreductase: “dark” reactions of a light-driven enzyme. Biochemistry 42, 523–8.

(275) Zhou, J., and Li, Y. (2010) Engineering cyanobacteria for fuels and chemicals production. Protein Cell 1, 207–10.

(276) Linke, M., Yang, Y., Zienicke, B., Hammam, M. A. S., von Haimberger, T., Zacarias, A., Inomata, K., Lamparter, T., and Heyne, K. (2013) Electronic transitions and heterogeneity of the bacteriophytochrome Pr absorption band: An angle balanced polarization resolved femtosecond VIS pump-IR probe study. Biophys. J. 105, 1756–66.

(277) Hauck, A. F. E., Hardman, S. J. O., Kutta, R. J., Greetham, G. M., Heyes, D. J., and Scrutton, N. S. (2014) The Photoinitiated Reaction Pathway of Full-length Cyanobacteriochrome Tlr0924 Monitored Over 12 Orders of Magnitude. J. Biol. Chem. 289, 17747–57.

(278) Salewski, J., Escobar, F. V., Kaminski, S., von Stetten, D., Keidel, A., Rippers, Y., Michael, N., Scheerer, P., Piwowarski, P., Bartl, F., Frankenberg-Dinkel, N., Ringsdorf, S., Gärtner, W., Lamparter, T., Mroginski, M. A., and Hildebrandt, P. (2013) Structure of the biliverdin cofactor in the Pfr state of bathy and prototypical phytochromes. J. Biol. Chem. 288, 16800–14.

(279) Mirkovic, T., Doust, A. B., Kim, J., Wilk, K. E., Curutchet, C., Mennucci, B., Cammi, R., Curmi, P. M. G., and Scholes, G. D. (2007) Ultrafast light harvesting dynamics in the cryptophyte phycocyanin 645. Photochem. Photobiol. Sci. 6, 964–75.

173 | P a g e

(280) Zehetmayer, P., Kupka, M., Scheer, H., and Zumbusch, A. (2004) Energy transfer in monomeric phycoerythrocyanin. Biochim. Biophys. Acta 1608, 35–44.

(281) Anderson, S., Šrajer, V., and Moffat, K. (2007) Structural Heterogeneity of Cryotrapped Intermediates in the Bacterial Blue Light Photoreceptor, Photoactive Yellow Protein¶. Photochem. Photobiol. 80, 7–14.

(282) Frauenfelder, H. (2010) Proteins, supercooled liquids, and glasses: A micro-review. Phys. E Low- dimensional Syst. Nanostructures 42, 662–665.

(283) Lubchenko, V., Wolynes, P. G., and Frauenfelder, H. (2005) Mosaic energy landscapes of liquids and the control of protein conformational dynamics by glass-forming solvents. J. Phys. Chem. B 109, 7488–99.

(284) Matysik, J., Hildebrandt, P., Schlamann, W., Braslavsky, S. E., and Schaffner, K. (1995) Fourier- transform resonance Raman spectroscopy of intermediates of the phytochrome photocycle. Biochemistry 34, 10497–507.

(285) Rockwell, N. C., Martin, S. S., Gulevich, A. G., and Lagarias, J. C. (2014) Conserved phenylalanine residues are required for blue-shifting of cyanobacteriochrome photoproducts. Biochemistry 53, 3118–30.

(286) Xu, X.-L., Gutt, A., Mechelke, J., Raffelberg, S., Tang, K., Miao, D., Valle, L., Borsarelli, C. D., Zhao, K.-H., and Gärtner, W. (2014) Combined mutagenesis and kinetics characterization of the bilin- binding GAF domain of the protein Slr1393 from the Cyanobacterium Synechocystis PCC6803. Chembiochem 15, 1190–9.

(287) Mukougawa, K., Kanamoto, H., Kobayashi, T., Yokota, A., and Kohchi, T. (2006) Metabolic engineering to produce phytochromes with phytochromobilin, phycocyanobilin, or phycoerythrobilin chromophore in Escherichia coli. FEBS Lett. 580, 1333–8.

(288) Sharrock, R. A. (2008) The phytochrome red/far-red photoreceptor superfamily. Genome Biol. 9, 230.

(289) Frankenberg-Dinkel, N., and Lagarias, J. C. (2003) Biosynthesis and Biological Functions of Bilins, in The Porphyrin Handbook (Kadish, K. M., Smith, K. M., and Guilard, R., Eds.), pp 211–236. Academic Press, San Diego, CA.

(290) Iii, F. A., Hasson, K. C., Gai, F., Anfinrud, P. A., and Mathies, R. A. (1997) Femtosecond Time- Resolved Spectroscopy of the Primary Photochemistry of Phytochrome.

(291) Peteanu, L. A., Schoenlein, R. W., Wang, Q., Mathies, R. A., and Shank, C. V. (1993) The first step in vision occurs in femtoseconds: complete blue and red spectral studies. Proc. Natl. Acad. Sci. U. S. A. 90, 11762–6.

(292) Van Brederode, M. E., Gensch, T., Hoff, W. D., Hellingwerf, K. J., and Braslavsky, S. E. (1995) Photoinduced volume change and energy storage associated with the early transformations of the photoactive yellow protein from Ectothiorhodospira halophila. Biophys. J. 68, 1101–9.

(293) Kim, J. E., Tauber, M. J., and Mathies, R. A. (2001) Wavelength Dependent Cis-Trans Isomerization in Vision †. Biochemistry 40, 13774–13778.

(294) Heyne, K., Mohammed, O. F., Usman, A., Dreyer, J., Nibbering, E. T. J., and Cusanovich, M. A. (2005) Structural evolution of the chromophore in the primary stages of trans/cis isomerization in photoactive yellow protein. J. Am. Chem. Soc. 127, 18100–6.

174 | P a g e

(295) Gottlieb, S. M., Corley, S. C., Madsen, D., and Larsen, D. S. (2012) Note: a flexible light emitting diode-based broadband transient-absorption spectrometer. Rev. Sci. Instrum. 83, 056107.

(296) Ernst, O. P., Lodowski, D. T., Elstner, M., Hegemann, P., Brown, L. S., and Kandori, H. (2014) Microbial and animal rhodopsins: structures, functions, and molecular mechanisms. Chem. Rev. 114, 126–63.

(297) Slabinski, L., Jaroszewski, L., Rychlewski, L., Wilson, I. A., Lesley, S. A., and Godzik, A. (2007) XtalPred: a web server for prediction of protein crystallizability. Bioinformatics 23, 3403–5.

(298) Bourgeois, D., Schotte, F., Brunori, M., and Vallone, B. (2007) Time-resolved methods in biophysics. 6. Time-resolved Laue crystallography as a tool to investigate photo-activated protein dynamics. Photochem. Photobiol. Sci. 6, 1047–56.

(299) Cammarata, M., Levantino, M., Schotte, F., Anfinrud, P. A., Ewald, F., Choi, J., Cupane, A., Wulff, M., and Ihee, H. (2008) Tracking the structural dynamics of proteins in solution using time- resolved wide-angle X-ray scattering. Nat. Methods 5, 881–6.

(300) Westenhoff, S., Malmerberg, E., Arnlund, D., Johansson, L., Nazarenko, E., Cammarata, M., Davidsson, J., Chaptal, V., Abramson, J., Katona, G., Menzel, A., and Neutze, R. (2010) Rapid readout detector captures protein time-resolved WAXS. Nat. Methods 7, 775–6.

(301) Li, H., Zhang, J., Vierstra, R. D., and Li, H. (2010) Quaternary organization of a phytochrome dimer as revealed by cryoelectron microscopy. Proc. Natl. Acad. Sci. U. S. A. 107, 10872–7.

(302) Pudney, C. R., Khara, B., Johannissen, L. O., and Scrutton, N. S. (2011) Coupled motions direct electrons along human microsomal P450 Chains. PLoS Biol. 9, e1001222.

(303) Ryjenkov, D. A., Tarutina, M., Moskvin, O. V, and Gomelsky, M. (2005) Cyclic diguanylate is a ubiquitous signaling molecule in bacteria: insights into biochemistry of the GGDEF protein domain. J. Bacteriol. 187, 1792–8.

(304) Savakis, P., De Causmaecker, S., Angerer, V., Ruppert, U., Anders, K., Essen, L.-O., and Wilde, A. (2012) Light-induced alteration of c-di-GMP level controls motility of Synechocystis sp. PCC 6803. Mol. Microbiol. 85, 239–51.

(305) Bogomolni, R. A., and Swartz, T. E. (2005) LOV-domain Photochemistry, in Handbook of Photosensory Receptors (Briggs, W. R., and Spudich, J. L., Eds.), pp 305–322. Wiley-VCH, Weinheim, Germany.

(306) Ma, Q., Hua, H.-H., Chen, Y., Liu, B.-B., Krämer, A. L., Scheer, H., Zhao, K.-H., and Zhou, M. (2012) A rising tide of blue-absorbing biliprotein photoreceptors: characterization of seven such bilin-binding GAF domains in Nostoc sp. PCC7120. FEBS J. 279, 4095–108.

175 | P a g e