Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A spectrophotometric investigation of the respiratory cytochromes of aerobically-grown Escherichia coli… Withers, Howard Keith 1989

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1990_A1 W58.pdf [ 18.03MB ]
Metadata
JSON: 831-1.0100499.json
JSON-LD: 831-1.0100499-ld.json
RDF/XML (Pretty): 831-1.0100499-rdf.xml
RDF/JSON: 831-1.0100499-rdf.json
Turtle: 831-1.0100499-turtle.txt
N-Triples: 831-1.0100499-rdf-ntriples.txt
Original Record: 831-1.0100499-source.json
Full Text
831-1.0100499-fulltext.txt
Citation
831-1.0100499.ris

Full Text

A Spectrophotometric Investigation of the Respiratory Cytochromes of  Aerobically-Grown Escherichia coli K-12 Howard Keith Withers B . S c , University of Birmingham, England, U.K., 1974, M . S c , University of British Columbia, Canada, 1979. A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE D E G R E E OF DOCTOR OF PHILOSOPHY in THE FACULTY OF G R A D U A T E STUDIES Department of Biochemistry We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA December 1989 © Howard Keith LUithers, 1989 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada Date DE-6 (2/88) {Abstract} ii ABSTRACT The cytochrome o and cytochrome d oxidase complexes provide twin termini for the branched respiratory chain of aerobically grown Escherichia coli. Combined use of mutant strains, modulated growth conditions and high resolution analytical techniques enabled cytochromes to be resolved, identified and partially characterized. The cytochrome complement of everted membrane vesicles and detergent extracts fractionated by liquid chromatography is more complex than previously recognised. Multiple type-6 cytochromes were resolved by potentiometry and by high resolution spectrophotometry in membrane vesicles from mutant strains lacking the cytochrome d oxidase complex and grown under conditions minimising respiratory chain diversity. Cytochrome o was identified with Em = +235 mV {vs. NHE) as were low potential cytochromes associated with dehydrogenases. Spectrally distinct components of the cytochrome d complex yielded Em values of +125 mV (cytochrome 6595) and +187 mV (cytochrome d). The latter displayed atypical redox behaviour with extreme hysteresis during potentiometric titrations. Several cytochromes 655 5 displaying single, symmetrical redox oc-bands at 77 K were resolved from detergent extracts of vesicles. Mutant strains identified one with Mr = 52 500 (gel filtration) and Em = +20 mV as the sdhC gene product, a component of succinate dehydrogenase. DL-lactate induced another while a hydroperoxidase, M r = 386 000 (gel filtration) with twin Em values of -2mV and -121 mV and a split Soret absorption band at 77 K ( A ? m a x = 426.0 nm + 434.0 nm) was produced under limited oxygen tension. The Triton-solubilized and purified cytochrome 0 complex exhibited Mx = 516 000 (gel filtration) with five component peptides of M r = 55 000, 32 000, 31 000, 21 000 and 16 000 (SDS-PAGE). It displayed mid-point potentials of -58 mV, +127 mV and +260 mV and three a-absorption maxima at 77 K : 554.5 nm, 557.0 nm and 563.5 nm. These components were reduced equivalently during poised-potential low temperature spectrophotometric analyses. Carbon monoxide binding changed the complex's redox a-absorption spectrum minimally but shifted the high potential Em to approximately +420 mV. Quinone analogues inhibited both reduction and reoxidation of the complex. Cytochrome o complex prepared from cloned sources contained a significantly greater proportion of the component with mid range electrochemical potential absorbing at 554.0 nm. These results are discussed in relation to possible structures of the complex, its respiratory interactions and the identity of cytochrome o itself. {Table of Contents} iii TABLE OF CONTENTS ABSTRACT: ; ii TABLE OF CONTENTS : iii LIST OF TABLES : vii LIST OF FIGURES : viii ABBREVIATIONS : xii ACKNOWLEDGEMENTS : xiv INTRODUCTION : A. Overview of Bioenergetic Metabolism 1 i. Cellular distribution of chemical energy 1 ii. Glycolysis and fermentative growth 3 iii. Electron transport phosphorylation 4 iv. Structure of respiratory electron transport systems 10 v. Structural properties of cytochromes 12 vi. Bacterial respiratory chains 15 B. Aerobic Electron Transport in Escherichia coli 16 i. Energetic diversity 16 ii. Environmental and ecological factors 16 iii. Respiratory control 17 iv. Aerobic respiratory chains — status of research prior to current study 19 v. Aerobic respiratory cytochromes — progress of the present investigation ... 21 C. Experimental Progress Concurrent with the Present Investigation 23 i. Overview 23 ii. Haem synthesis and incorporation into apocytochrome 24 iii. Immunologically based studies of cytochromes 26 iv. Genetically based studies of terminal oxidases 27 v. Purification and characterization of detergent solubilized terminal oxidases .. 28 vi. Molecular biological analyses of terminal oxidase genes 30 vii. Reconstution experiments and models of aerobic respiratory architecture 31 D. Synopsis of Research Aims and Strategy 33 MATERIALS & METHODS : a. Chemicals 35 b. Cell Types 35 C. Growth of cells 35 d. Crude membrane preparation 36 e. Inner membrane preparation 36 f. Solubilization of membranes 37 g. Fractionation of solubilized cytochromes 37 {Table of Contents} iv M A T E R I A L S & M E T H O D S (Continued) h. Purification of cytochrome o 38 i. Partial purification of the 'Peak III' respiratory cytochrome 39 j . Partial purification of the 'Peak I V hydroperoxidase 40 k. Preparation of nitrate reductase 41 1. Spectrophotometric analysis of cytochromes 41 m. Derivative analysis of spectrophotometric data 43 n. Redox kinetics of cytochromes 43 0. Potentiometric titrations 44 p. Spectrophotometry at selected electrochemical potentials 47 q. Stopped-flow spectrophotometry 48 r. Genetic manipulation of E. coli cells 49 S. Biochemical assays 51 t. Sodium dodecyl sulphate polyacrylamide gel electrophoresis 52 R E S U L T S & DISCUSSION : I. Reference Studies of Soluble Cytochromes. 1. Redox difference spectroscopy 53 a. Ambient temperature redox difference spectra 56 b. Low temperature redox difference spectra 56 C. Pyridine haemochromogen redox difference spectra 60 d. Fourth order derivatives of redox difference spectra 63 ii. Potentiometric titrations 69 a. Standard titrations and electrode calibration 69 b. Comparison of experimental data with literature values 74 C. Electrochemically poised high resolution spectrophotometric analysis .... 83 II. Membrane Studies. A. Aerobic Respiratory Type-6 Cytochromes 91 i. Redox difference spectroscopy 91 a. Visible range spectrophotometry of E. coli membrane cytochromes 91 b. High resolution ra-band cytochrome studies of respiratory cytochromes ... 100 1. Ct-bands of cell membrane type-£> cytochromes 100 2. Fourth order derivative spectra of cytochrome b a-bands 106 3. Multiple cytochrome'^550+' oc-band components I l l C Use of membranes prepared from cyd~ strains of E. coli 116 d. Mutant strains with an altered type-fc cytochrome complement 117 ii. Potentiometric titrations 121 a. Membrane cytochromes of cells grown on specific carbon/energy sources 121 b. Electrochemical characteristics of aerobic respiratory -^cytochromes .... 129 C Perturbation of the electrochemical response of certain cytochromes b .. 134 d. Poised potential high resolution redox difference spectrophotometry 137 e. Mutant strains with an altered type-ft cytochrome complement 143 iii. Reduction kinetics 147 a. Overview of technique 147 b. Membrane cytochrome analysis 153 {Table of Contents} v RESULTS & DISCUSSION (Continued B. Functional Pools of Respiratory Cytochrome 159 i . Respiratory cytochromes of cells grown aerobically on L-proline 159 a. Overview of investigative procedure 159 b. Titration of membranes from L-proline grown cells 160 C. Redox difference spectrophotometry of L-proline grown cells' membranes 160 d. Kinetic measurements of cytochrome reduction in cell membranes 160 C. Aerobic Terminal Oxidases 167 i . Membrane studies of cytochrome d 167 a. Spectroscopic properties of cytochrome d 167 b. Spectroscopic investigation of cytochrome d ligand binding 171 C. Photochemical degradation of the cytochrome d - CO complex 174 d. Potentiometric titration of the components of the cytochrome d complex . 177 e. High resolution redox difference spectrophotometry at poised potentials .. 184 f. Control of expression of the cytochrome d complex by growth substrate .. 185 ii. Membrane studies of cytochrome o 193 a. Spectroscopic properties of cytochrome o studied in cell membranes 193 b. Spectroscopic investigation of cytochrome o ligand binding 194 C. The cytochrome o-CO complex and low temperature photolysis 195 d. Potentiometric titration of components of the cytochrome o complex 202 e. High resolution redox difference spectrophotometry at poised potentials .. 203 HDL Purified Detergent Extracts. A. Cytochrome Solubilization and Fractionation 207 i . Membrane solubilization 207 a. Redox spectroscopy of extracted fractions 207 ii. Cytochrome fractionation by liquid chromatography 213 a. Ion-exchange chromatography of solubilized cytochromes 213 b. Spectral characteristics of fractionated cytochromes 221 B. Cytochrome Fractions from Ion-Exchange Chromatography 226 i. Peak I; cytochrome £556 associated with growth on succinate 226 a. Control of expression by growth substrate 226 b. Association with succinate dehydrogenasee 229 C. Synopsis of ion-exchange chromatographic datarelating to peak I 237 d. Biophysical parameters of 'succinate grown' peak I cytochrome ^555 239 ii. Peak II; cytochrome o complex 240 iii. Peak 111; cytochromes b associated with growth on L-proline 243 a. Control of expression by growth substrate 243 b. Partial purification and analysis of cytochrome components 246 iv. Peak IV; a cytochrome £1555 hydropcroxidase expressed at low p02 256 a. Control of expression 256 b. Redox spectrophotometry 259 C. Further purification of cytochrome 264 d. Spectrophotomctric properties of fractions from gel filtration 264 e. Stoicheiometry of the peak IV cytochrome complex 272 f. Potentiometric titration of the cytochrome complex 273 {Table of Contents} vi RESULTS & DISCUSSION (Continued") v. Peaks I and IV ; independence from frd genes and fumarate metabolism 281 a. Control of expression 281 b. Spectrophotometry of fractionated cytochromes 287 C. Fumarate reductase 290 C. Cytochrome o Terminal Oxidase Complex 293 i. Purification of the cytochrome o complex 294 a. Standard isolation procedure 294 b. Comparison with alternative procedures 300 ii. Spectrophotometric analyses of natural and cloned cytochrome o complex.. 306 a. Spectrophotometry at ambient temperature 306 b. High resolution spectrophotometry ' 306 C Perturbation spectrophotometry 315 d. Photolysis of the oxidase-carbon monoxide complex 320 iii. Kinetic analyses of reduction of the cytochrome o complex 320 a. Response to various reductants 320 b. Inhibition of of reduction and oxidation 323 C. Stopped-flow analysis of reduction kinetics 328 iv. Potentiometric titration of the cytochrome o complex 329 a. Titration of the complex isolated from natural sources 329 b. Titration of the cloned complex 339 v. Correlation of potentiometric and spectrophotometric analyses 347 a. High resolution spectrophotometry at kinetically poised potentials 347 b. Low resolution spectrophotometry at specific potentials 348 C. High resolution poised potential spectrophotometry 351 d. Comparison of poised potential spectrophotometry at 77 K and 305 K.... 354 vi. Synopsis of results relating to the solubilized cytochrome o complex 355 CONCLUSIONS : - 359 A. Cytochrome Pools in Membranes of Aerobically Grown E. coli 359 B. Aerobic Respiratory 'Cytochromes 6555' 363 C. The Cytochrome d Aerobic Terminal Oxidase 363 D. The Cytochrome 0 Aerobic Terminal Oxidase 365 BIBLIOGRAPHY : • 372 APPENDICES : 387 A. Spectral Characteristics of Cytochromes 387 B. Bacterial Strain Characteristics 389 C. Bacterial Growth Media 391 D. Buffer Solutions 393 {List of Tables} vii LIST OF TABLES TABLE I Spectrophotometer settings for optimal spectral resolution 42 TABLE II Electrochemical mediators for potentiometric titrations 46 TABLE III Spectral characteristics of standard cytochromes c, b$ and catalase 57 TABLE IV Comparison of potentiometric analyses of biological samples with previously published values 80 TABLE V Absorption maxima and extinction coefficients for respiratory cytochromes of cells grown to early exponential and stationary phases 97 TABLE VI Wavelengths of absorption maxima of cytochrome b components in exponential and stationary phase cell membranes 110 TABLE VII Wavelengths of absorption maxima of cytochrome b components in membranes exposed to terminal oxidase inhibitors 115 TABLE VIII Wavelengths of absorption maxima of cytochrome b components in membranes from classes of putative cyb~ mutants 120 TABLE IX Cytochrome b complement of aerobically-grown cell membranes determined by potentiometric titration 130 TABLE X Production of cytochrome d by cells of PLJ01 grown on specific carbon/energy sources 190 TABLE X I Absorption maxima of cytochrome fractions eluted from initial ion-exchange separation of solubilized membrane extracts 227 TABLE XII Lactate and succinate oxidase activities of resuspended membrane preparations from cells grown on different carbon/energy sources ... 238 TABLE XII! Major steps in the purification of the cytochrome o complex 301 TABLE XIV Absorption maxima of cytochrome o complex from natural and cloned sources 309 {List of Figures} viii LIST OF FIGURES Fig. 1: Models of the arrangement of respiratory components in electron transport pathways . 7 Fig. 2: Respiratory electron donors and acceptors available to facultative anaerobes 9 Fig. 3: Structures of protoporphyrin IX and haem prosthetic groups 14 Fig. 4: Visible range redox difference spectra of soluble reference cytochromes 55 Fig. 5: a-absorption bands from ambient and low temperature redox difference spectra of soluble reference cytochromes 59 Fig. 6: Redox difference spectra of pyridine hacmochromogen from types b andc soluble reference cytochromes 62 Fig. 7: Fourth-order finite difference spectTa of a-absorption bands from soluble reference cytochromes 65 Fig. 8: Absolute ultra-violet absorption spectrum and fourth-order finite difference spectra of soluble reference ferrocytochrome c 67 Fig. 9: Potentiometric titrations of standard cytochrome c in presence and absence of sucrose: direct plots 71 Fig. 10: Potentiometric titrations of standard cytochrome c in presence and absence of sucrose: Nemst plots 73 Fig. 11: Potentiometric titrations of soluble reference cytochromes 76 Fig. 12: Potentiometric titration of membrane cytochromes and nitrate reductase preparation from E. coli grown anaerobically in the presence of nitrate : direct plots 78 Fig. 13: Potentiometric titration of partially purified E. coli nitrate reductase: Nernst plots ... 82 Fig. 14: Potentiometric titrations of standard cytochrome c : low temperature poising and spectrophotometry at 77 K 85 Fig. 15: Potentiometric titrations of standard cytochrome c : effects of poising at low temperature 87 Fig. 16: Escherichia coli respiratory cytochromes: visible range spectra of cell membrane suspensions at ambient temperature 93 Fig. 17; Spectral identification of E. coli terminal oxidases following exposure of membranes to carbon monoxide 95 Fig. IS: Escherichia coli respiratory cytochromes: high resolution visible range redox difference spectra of cell membrane suspensions 99 Fig. 19: Absorbance shifts and improved spectral resolution of type-fc membrane cytochrome a-bands at low temperatures 102 Fig. 20: High resolution a-band spectra: variation of wild-type cells' type-t membrane cytochromes with aeration and carbon source 105 Fig. 21: High resolution a-band spectra : typc-fo cytochrome content of membranes from a selection of wild-type cell strains 108 Fig. 22: High resolution a-band spectra : absolute spectra of washed, resuspended membranes in the presence of respiratory inhibitors : 114 {List of Figures} ix Fig. 23: High resolution redox spectra of membrane preparations from putative cyb' mutants .. 119 Fig. 24: Differences in potentiometric titrations of membrane preparations from cells grown to stationary phase on either D-glucose or on DL-lactate 123 Fig. 25: Ambient temperature redox difference spectra and carbon monoxide binding spectra of wild-type cells grown on cither D-glucose or DL-lactate 126 Fig. 26: Potentiometric differences between membrane preparations from w+ and cydr cells grown to stationary phase 128 Fig. 27: Modification of membrane cytochrome potentials by cell growth without thiamine and by treatment of membrane preparations with ferricyanide 133 Fig. 28: Poised potential high resolution difference spectra of w+ and cyd' cells from the titrations illustrated in Figure 26 140 Fig. 29: Carbon monoxide binding spectra from cyb~ and putative cyo~ mutant isolates 146 Fig. 30: Potentiometric titration of mutant and parental strains with different complements of type-& cytochrome 149 Fig. 31: High resolution redox difference spectra of mutant strains with different complements of cytochrome 151 Fig. 32: Reduction kinetics of membrane cytochromes from cyd' cells grown aerobically to stationary phase on various carbon/energy sources 155 Fig. 33: Potentiometric titrations of membrane preparations from cydr cells grown to stationary phase on L-prolinc, D-glucose and DL-lactate 158 Fig. 34: High resolution redox difference spectra of membrane preparations from cells of cyd' phenotype grown to stationary phase on L-proline 162 Fig. 35: Membranes containing the cytochrome d complex : redox difference spectra after treating the reference material with cither oxygenating and/or oxidizing agents ... 170 Fig. 36: Scanning dual wavelength spectra of membranes containing the cytochrome d complex following carbon monoxide treatment and low temperature photodissociation 176 Fig. 37: Potentiometric titrations of the cytochrome 5^95 component of the cytochrome d complex in resuspended membrane preparations 179 Fig. 38: Potentiometric titrations of the cytochrome d component of the cytochrome d complex in resuspended membrane preparations 181 Fig. 39: Aerobic growth curves of strain PLJ01 : effects of increased aeration and dual carbon/energy sources 188 Fig. 40: Broad range difference spectra and photodissociation studies of the carbon monoxide derivatives of cytochromes in membrane preparations lacking cytochrome d 198 Fig. 41: Redox difference spectra of high potential cytochromes measured during potentio-metric titration of membrane preparations from strains with cyd' phenotype 205 Fig. 42: High resolution oc-band spectra of Triton X- l 14 detergent extracts from washed membranes of wild-type cells 210 Fig. 43: High resolution oc-band spectra of Triton X- l 14 detergent extracts from washed membranes of cyd' cells 212 Fig. 44: DEAE-BioGel.A clution profile of Triton membrane extracts from wild-type cells grown on glucose 215 {List of Figures} x Fig. 45: DEAE-BioGel.A elution profile of Triton membrane extracts from cydr cells grown on glucose 218 Fig. 46: DEAE-BioGel.A elution profiles of Triton membrane extracts from cydr cells grown on different carbon/energy sources 220 Fig. 47: Low temperature broad-range visible redox spectra of ion-exchange chromatographic fractions enriched in cytochrome : Peaks I, II and IV 223 Fig. 48: High resolution a-band spectra of ion-exchange chromatographic fractions enriched in cytochrome : Peaks I, II and IV 225 Fig. 49: High resolution a-band spectra and fourth order derivatives of resuspended membranes from strains GR19N and TK3D11 grown on DL-lactate and/or succinate 231 Fig. 50: DEAE-BioGel.A elution profiles of Triton membrane extracts from strain TK3D11 which lacks succinate dehydrogenase activity 234 Fig. 51: High resolution visible range redox difference spectra of partially purified 'Peak F cytochrome 236 Fig. 52: Potentiometric titration of partially purified'Peak I'cytochrome 242 Fig. 53: DEAE-BioGel.A and BioGel.HTP elution profiles of Triton membrane extracts from cells grown on L-proline 245 Fig. 54: High resolution visible range redox difference spectra of partially purified 'Peak IE' cytochrome : 248 Fig. 55: High resolution visible range redox difference spectra of partially purified 'Peak III' cytochrome fractionated on hydroxylapatite 250 Fig. 56: Broad range visible redox difference spectra: partially purified 'Peak III' cytochrome .. 253 Fig. 57: Sodium Dodecyl Sulphate-Polyacrylamidc Gel Electrophoresis of 'Peak 111' cytochrome 255 Fig. 58: DEAE-BioGel.A elution profiles of Triton membrane extracts from cells harvested at different phases of growth 258 Fig. 59: DEAE-BioGel.A elution profiles of Triton membrane extracts from cultures grown under different oxygen tensions 261 Fig. 60: High resolution visible range redox spectra of partially purified 'Peak I V cytochrome 263 Fig. 61: Gel filtration elution profiles of 'Peak I V cytochromes from Sephacryl S-200 and S-300 matrices 266 Fig. 62: High resolution redox difference spectra of 'Peak I V cytochrome subtractions separated by gel exclusion chromatography 268 Fig. 63: Broad range visible redox difference spectra of 'Peak I V cytochrome subfractions separated by gel exclusion chromatography 271 Fig. 64: Potentiometric titration of partially purified'Peak I V cytochrome 275 Fig. 65: Potentiomctrically poised low temperature redox difference spectra of 'Peak I V cytochrome 278 Fig. 66: Potentiometric titration of partially purified 'Peak I V cytochrome following treatment with carbon monoxide 280 Fig. 67: DEAE-BioGel.A elution profiles of Triton membrane extracts from cells carrying the fumarate reductase gene on plasmid pFRD84 284 {List of Figures} xi Fig. 68: DEAE-BioGel.A elution profiles of Triion membrane extracts from anaerobically-grown cells carrying the plasmid-borne fumarate reductase gene 286 Fig. 69: High resolution broad range visible redox difference spectra of partially purified solubilized cytochromes from cells grown anaerobically on glycerol and fumarate 289 Fig. 70: High resolution a-bands of redox difference spectra from partially purified solubilized cytochromes of cells grown anaerobically on glycerol and fumarate 292 Fig. 71: Gel filtration elution profiles : calibration of the Sephacryl S-300 matrix and elution of 'Peak IF cytochromes 297 Fig. 72: SDS-Polyacrylamide gel electrophoresis of samples from stages during the purification of cytochrome o 299 Fig. 73: SDS-Polyacrylamide gel electrophoresis of purified samples of uncloned and of plasmid encoded cytochrome o 305 Fig. 74: Absolute absorption spectra of cytochrome o measured at ambient temperature 308 Fig. 75: High resolution broad range redox difference spectra of uncloned and of plasmid encoded cytochrome o 312 Fig. 76: High resolution redox a-absorption spectra of uncloned and of plasmid encoded cytochrome o 314 Fig. 77: Reduction of cytochrome o by various reductants : high resolution redox a-absorption spectra 317 Fig. 78: Effects of HOQNO upon high resolution redox a-absorption spectrum of solubilized cytochrome o 319 Fig. 79: Ambient temperature carbon monoxide binding spectra of solubilized cytochrome o .. 322 Fig. 80: Kinetics of reduction of solubilized cytochrome o complex by various reductants 325 Fig. 81: Kinetics of reduction and reoxidation of solubilized cytochrome o complex 327 Fig. 82: Rapid kinetics of reduction of solubilized cytochrome o complex 331 Fig. 83: Potentiometric titration of solubilized cytochrome o : the effect of carbon monoxide .. 333 Fig. 84: Potentiometric titration of solubilized cytochrome o : the effect of ferricyanide 336 Fig. 85: Potentiometric titration of solubilized cytochrome o : perturbation effects 338 Fig. 86: Potentiometric titration of the high potential component of solubilized cytochrome o 341 Fig. 87: Potentiometric titration of solubilized cytochrome o from cloned sources 343 Fig. 88: Potentiometric titration of solubilized, cloned cytochrome o : the effect of carbon monoxide 346 Fig. 89: Response of redox a-absorption maximum of solubilized cytochrome o to the electrochemical potential of the titration buffer 350 Fig. 90: Poised potential high resolution redox difference spectra of solubilized cytcohrome o from uncloned and from plasmid encoded sources 353 Fig. 91: Model of respiratory cytochrome pools in the membrane of aerobically-grown E. coli 362 Fig. 92: Model of terminal oxidase action in E. coli aerobic respiratory chains 368 Fig. 93: Model of cytochrome o interaction with a protonmotive Q-cycle in E. coli membranes 370 {Abbreviations} xii ABBREVIATIONS ALA 5-aminolaevulinic acid ; 8-aminolaevulinic acid. DCBP 2,6-dichlorophenolindophenol. DMSO dimethyl sulphoxidc. DQ duroquinone ; 2,3,5,6-tetramethyl-l,4-bcnzoquinone. DQH2 duroquinol; 2,3,5,6-tctramethyl-l,4-benzoquinol. duroquinone 2,3,5,6-tetramethyl-l ,4-bcnzoquinonc. DW2c SLM/Aminco model DW2c dual wavelength double beam spectrophotometer, ec energy charge ; [ATP] + 0.5x[ADP] / [ATP] + [ADP] + [AMP] EDTA ethylenediaminetetraacctic acid. £j j ambient redox potential of the system with standard hydrogen half cell as reference. EM standard (mid-point) redox potential under experimental conditions. E'M standard oxidation-reduction (redox) potential at neutral pH. Eq standard oxidation-reduction (redox) potential at pH 0 and unit activities. EPR electron paramagnetic resonance. T Faraday constant (96 493 J V"1) GSH glutathione, reduced form ; y-L-glutamyl-L-cysteinylglycine. HOQNO 2-(n -heptyl)-4-hydroxyquinoline-/V -oxide. HPLC high performance liquid chromatography. mS milliSiemens (1.0 mS = 1.0 mmho). mV mV vs. NHE ; millivolt values are given relative to the Normal Hydrogen Electrode. 11 number of electrons transferred in an oxidation-reduction reaction. NAD + oxidized form of nicotinamide adenine dinucleotide. NADH reduced form of nicotinamide adenine dinucleotide. NPN iV-phcnyl-l-naphtlvylaminc. OCtylglucoside l-0-n-octyl-[j-D-glucopyTanosidc. PE-356 Perkin-Elmer model 356 dual wavelength double beam spectrophotometer. PMS phenazine methosulphatc. {Abbreviations} xiii PMSF phenylmethylsulphonylfluoride. R the gas constant, 8.31 J K"1 mol"1. RMS root mean square {estimate of statistical distribution}. fR-O] reduced minus oxidized (difference spectrum). [(R+CO) - R] reduced plus carbon monoxide minus reduced (difference spectrum}. Sarkosyl N-lauroylsarkosine ; N-dodecanoyl-N-mcthylglycine. SDS sodium dodecyl sulphate. SDS-PAGE sodium dodecyl sulphate polyacrylamide gel electrophoresis. spectra: CO-binding reduced plus carbon monoxide minus reduced difference spectrum, redox reduced minus oxidized difference spectrum T temperature above absolute zero in Kelvin. T M A O trimethylamine /V-oxide. T M B Z 3,3',5,5'-tetramethylbcnzidinc. T M P D N, N, N', N'-lelramcthyl-p-phcnylcnediamine dihydrochloride. Tris tris (hydroxymethyl) aminomethane. Triton : polyoxyethylenc p-t-ociy\ phenol; p-f-octylphenoxy polyethoxy ethanol derivative (trade mark, Rohm & Haas Co.). X-100 n = 10, A=0. X-114 n = l l , A =4. wild-type (bacterial strain}. AGp phosphorylation potential : eqn. (1), Introduction. A/Zpj+ proton electrochemical potential ; protonmotive force. A, wavelengtli. r mass-action constant of a reaction. [I conductivity (mS cm"') [i . bacterial growth rate, generations per hour: glucose as sole carbon/energy source, glc LL bacterial growth rate, generations per hour: lactate as sole carbon/energy source, lac (1 bacterial growth rate, generations per hour: succinate as sole carbon/energy source. {Acknowledgements} xiv ACKNOWLEDGEMENTS Lome Reid was the source of a much valuable advice and encouragement during the unreported, preliminary years of this work, his intimate knowledge of microcomputers and departmental politics being freely shared : Lome is truly the quintessential 'System Analyst'. I am grateful to Dr. Vladimir Palaty for encouragement and advice when initially contemplating these studies and to Dr. Robert C. Miller for bawling down a subterranean corridor all those years ago. Dr. Ross MacGillivray provided both financial assistance and an understanding ear at a critical time during the gestation of the thesis, and generously allowed me to learn techniques of molecular biology in his laboratory. Drs. R. B. Gennis, J. H. Weiner and W. Epstein made available a number of bacterial strains that were critical for these studies. Dr. P. D. Bragg created opportunities for me to teach myself a great variety of new skills. A small number of close friends assisted the progress of this undertaking with the support and practical advice that was vital for its conclusion, especially during the eleemosynary phases of the work : in particular Frank, Erin and 'Mac' Skelton and the proprietors of 'The Bank of Angela & Jim'. These investigations were supported in part by funds from the Medical Research Council of Canada. Major funding was also provided by Pacific BioConsultants (Vancouver, B.C.), by the University of British Columbia (Vancouver, B.C.) — as Teaching Assistantships and a Research Assistantship, by Simon Fraser University (Bumaby, B.C.) — a Sessional Lectureship, and by the British Columbia Provincial 'Challenge' Award programme. {Introduction} 1 INTRODUCTION "The Spectroscope seems likely to be of almost as great use in Medicine as it has already proved in terrestrial, solar and stellar Chemistry" {121} These words of Charles A. MacMunn (1852-1911), discoverer of the ubiquitous biological pigments that he named histohaematins, have been borne out by the multiplicity of spectroscopic and spectrophotometric techniques now applied to deciphering the complexities of respiratory chain structure, function and regulation {122). This is especially true of those respiratory components known collectively as cytochromes, the term coined by David Keilin upon his rediscovery of the 'histohaematins'in 1925 {93}. A . OV E R V I E W O F B I O E N E R G E T I C M E T A B O L I S M (i) Cellular distribution of chemical energy The energy required for cellular growth, for anabolic metabolism, nutrient accumulation and for electrical, mechanical and luminescent biological activities is most often distributed throughout the cell in the form of adenosine-5'-triphosphate, ATP. Cellular concentrations of this molecule are generally held away from equilibrium with those of the corresponding diphosphate, ADP, plus inorganic phosphate, P;. This status enables energetically or entropically unfavourable half-reactions to be 'driven' to metastable states displaced from equilibrium by enzymatic coupling to the dephosphorylation of ATP. Thus a net decrease is generated in the Gibbs (free) energy content of the coupled reactions, many of which proceed directly via the formation of a phosphorylated intermediate within the active site of the enzyme responsible for this linked catalysis. Consequently the energy state of a cell, its capability to carry out biochemical work, is frequently described in terms representing the displacement of the components of the ATP dephosphorylation reaction from equilibrium : 'energy charge' is equivalent to 0.5 ([ADP]+2[ATP] / [AMP]+[ADP]+[ATP]), and phosphorylation potential, which is often approximated as the ratio of [ATP] / ([ADP]+[Pj ]) {Introduction} 2 and designated AG p . This ratio, when expanded to include the participation of magnesium ions and protons is described by the mass-action constant of the ATP dephosphorylation reaction, r. Thus AGp is defined as in equation (1), where AG°' is the standard Gibbs energy change for this reaction and the 'prime' associated with each variable signifies the apparent, rather than standard system conditions: AGp = AG' = A G 0 ' + 2.303 RT\ogwr (1) The value of AG°' may also be expressed in units of millivolts by dividing throughout equation (1) by the Faraday constant, f (section I.ii.c). Concise reviews of these topics may be found in references 5, 64 & 145. The energy required for bacteria to generate ATP may be derived from many diverse reactions. Those pathways which a particular organism is capable of using are employed selectively in adaptation to environmental conditions. These reactions may be classified into two general processes for the conservation of energy: substrate level phosphorylation and electron transport phosphorylation. Many cellular activities, especially those of vectorial nature, are dependent upon membrane energization in the form of a proton electrochemical potential or protonmotive force (the combination of a proton gradient and electrical potential) across a sealed membrane {19,87}. ATP formed by substrate level phosphorylation may be used to generate such gradients, and conversely the proton motive force generated by electron transport processes is closely coupled to ADP phosphorylation (v./.). In eucaryotes the mitochondrial and chloroplast inner membranes are employed for this latter purpose; procaryotes, lacking intracellular organelles, use the cytoplasmic membrane. The proton electrochemical potential is the sum of electrical and chemical forces to which the protons are subjected, as described in equation (2): RT [H+]B RT A^ H+ = Al//+2.303 - log 1 0 = Ay/--2.303 ApH (2) r [ H + ] A r where A/ijj+ is the proton electrochemical potential, Av|/ is the electrical potential difference across the membrane, ApH is the pH difference between the two compartments created by the intact membrane and where the compartments'proton concentrations are [H+]A and [H+] .^ Nichollshas provided a particularly lucid synopsis of these phenomena, both authenticated and proposed, aspects of which are described in detail below {145}. {Introduction} 3 (ii) Glycolysis and fermentative growth Relatively few substrate level phosphorylation reactions have been described, those associated with the central amphibolic pathways occurring in both of the major procaryotic glycolytic routes. Catabolism of organic substrates through either the Embden-Meyerhof-Pamas or the Entner-Doudoroff pathways generates certain common intermediates which are degraded with a sufficiendy large Gibbs energy change that enzymatic catalysis of these reactions couples them to the phosphorylation of ADP. All glycolytic enzymes are soluble and that they reside freely in the cytoplasm has recently been affirmed {124}. In the absence of an external electron acceptor reoxidation of the reduced nicotinamide adenine dinucleotide formed by glycolysis is achieved by reduction of available oxidized substrates: generally by conversion of pyruvate and other glycolytic products to one or more of ethanol, lactate, acetate or other short chain organic acids. It follows that growth upon one or more of these organic acids as the sole source of carbon is not possible by fermentative processes alone — i.e. without an external electron acceptor being present. These fermentations permit a continued supply of ATP from substrate level phosphorylations in the presence of a suitable carbon source. Hydrolysis of this ATP is coupled to the energisation of those membranes in which the proton translocating ATP hydrolase responsible for the process is incorporated. The principal molecular features are common to both eucaryotic and procaryotic forms of this complex enzyme. They comprise two major operational parts, the hydrophilic which effects the ATP hydrolase activity at the cytoplasmic surface of the membrane, and the hydrophobic FQ which spans the membrane. The latter component provides the proton translocating mechanism, usually symbolized as a pore or pump, by which H + may be extruded through the ion-impermeable membrane in conjunction with the ATPase activity of Fj at a rate now estimated to be 3H + per ATP hydrolysed. As alluded to above the complete vectorial reaction may approach physiological equilibrium from either direction when catalysed by the intact membrane enzyme complex. Nevertheless, the reaction kinetics and catalytic response to particular inhibitors differ between the hydrolysis and synthesis of ATP. Reviews of the structure and proposed mechanisms of the proton translocating ATPase, with emphasis upon that of E. coli, are provided in references {19,51,55, 139,145}. The proton electrochemical potential, A^H+, generated across the membrane by ATP hydrolysis may be utilized for active transport of ions and substrates across the cytoplasmic membrane of procaryotic cells, for transhydrogenation of nicotinamide adenine dinucleotide phosphate (NADP"1") to supply anabolic reductions, and if an electron transport pathway is present, for reverse electron transport which may generate reduced nicotinamide adenine dinucleotide (NADH) {71,87,145,196}. {Introduction} 4 Many fermentative bacteria are able to increase the efficiency with which they^ are able to generate a proton electrochemical potential by utilizing transport systems that export metabolic end-products down a concentration gradient into the surrounding environment. Carrier-mediated proton symport leads to an increase of transmembrane electrical potential and proton translocation to an increase in the pH gradient {78}. Most microorganisms and plants are capable of performing the reactions of the glyoxylate cycle which allows them to use short-chain organic acids, including acetate, or fatty acids as sole carbon source. By-passing the two tricarboxylic acid cycle reactions in which CO2 is evolved between the condensation of acetyl-CoA with oxaloacetate to form citrate and the reformation of oxaloacetate, the glyoxylate cycle permits replenishment of the metabolic constituents through generation of succinate from isocitrate. The glyoxylate formed in this reaction undergoes condensation with a second molecule of acetyl-CoA to form malate, and thence oxaloacetate. Consequently these cells are able to use the reactions of these central pathways for anabolic purposes, especially during growth on low molecular weight carbon sources such as TCA-cycle intermediates or precursors. However, these processes generate NADH that precludes the use of oxidized short chain substrates as sole carbon source by fermentation. (iii) Electron transport phosphorylation Oxidative phosphorylation describes ATP synthesis coupled to respiratory electron transport; photophosphorylation is that accompanying electron transport during photosynthetic energy transduction. Both processes employ molecular constituents described above in relation to the generation of a proton electrochemical gradient with a proton translocating ATP hydrolase : an intact, ion-impermeable membrane separating two compartments with a proton translocating ATPase attached to one membrane surface. An electron transport mechanism creates the proton electrochemical gradient across the membrane which then forces proton translocation through the F 0 portion of the ATPase resulting- in tightly coupled ATP synthetase activity at the Fj enzyme assembly. Consequently the ATPase reactions implemented during fermentation are reversed, albeit with some kinetic and inhibition response modifications as mentioned earlier. Without respiratory or photosynthetic electron transport to maintain the proton gradient synthesis of ATP by the enzyme complex is unable to proceed. The other facet of this coupling is that if ATP synthesis is blocked or prevented, as when substrate is exhausted, the established proton electrochemical potential will prevent further electron transport (v.i.). {Introduction} 5 Photosynthetic electron transfer causes the formation of a proton electrochemical potential which is exploited by a proton translocating ATP synthetase activity. In brief, light induces vectorial ejection of an electron at low potential through the reaction centre located within the membrane and the return of an electron to this reaction centre at higher potential is accomplished by a proton translocating electron transfer chain which thereby generates the proton electrochemical gradient. These essential principles are common to both bacterial (cyclic) and chloroplast (non-cyclic) photosystems {12, 44, 45, 145}. Light is also the energy source for the simplest known proton pump, the bacteriorhodopsin inserted through the energy transducing purple membrane of Halobacterium halobium. Photons bleach the pigment and cause protons to be released vectorially from the outer membrane surface; regeneration of the pigment is accomplished by protons from the inner surface, thereby generating a proton gradient. This proton electrochemical potential is used to drive a proton translocating ATPase and functional ATP synthetase activity may also be obtained in vesicles reconstituted with bacteriorhodopsin, phospholipid and the proton translocating ATPase from different sources, including that of higher animals {145,195}. As an alternative to light as an external energy source for the electron transport mechanisms generating a proton electrochemical gradient, a redox potential may be used for this purpose using reduced substrates that may be exogenous in origin or generated internally, as in the case of NADH. Oxidative phosphorylation also requires the presence of an external electron acceptor of greater potential than that of the molecule donating the reducing equivalents. Generally such a respiratory electron transport pathway requires that both donor and acceptor react at one face of the membrane. A series of enzyme complexes function as an alternating sequence of hydrogen carrier and electron carrier from one face of the membrane to the other in proton translocating loops connecting the donor indirectly to the acceptor (Fig. la). The number of loops, or proton translocating segments, that are possible is dependent upon the potential difference between donor and acceptor. Thus dioxygen, with a high mid-point potential for the O2I2H20 redox couple of +820 mV vs. NHE, is particularly suited to the role of acceptor and is the favoured molecule for this function in many species. Some of the alternatives available to facultative anaerobes are illustrated in Figure 2. Certain Clostridia species, growing anaerobically on carbon monoxide, generate a proton electrochemical potential used to drive L-alanine uptake which varies in nature from electrical potential plus proton gradient below pH 7.5 to a potential which is solely electrical in nature above this pH {86}. Refer to {15, 33, 59, 136, 137, 145 & 210} for explanations of proposed mechanisms of electron transport phosphorylation. In the past contradictory interpretations of a variety of experimental results were highly controversial, yet most evidence is now accepted as agreeing in principle with the functioning of a chemiosmotic process, originally described by Mitchell {136}. {Introduction} 6 Fig. 1: Models of the arrangement of respiratory components in electron transport pathways. Abbreviations: F M N & F P , flavoproteins; Q H 2 , ubiquinol; cyt., cytochrome; Fe/S, iron-sulphur proteins; Q , ubiquinone; e", electron. 2.. Generalised scheme proposed for the arrangement of respiratory components found in many aerobic bacteria. These respiratory chains translocate protons less efficiently than those incorporating cytochromes 0.03 and c which are thought to resemble the mitochondrial system. After Haddock & Jones, 1977 (71}. l l . i . The 'protonmotive Q-cycle' illustrated in general form with electrons passing from component x to component y. The terms cyt. ba and cyt. bb imply species or forms of cytochrome b accepting and donating an electron to quinol and quinone respectively. After Mitchell, 1975 (138). i i . Incorporation of the 'protonmotive Q-cycle' into a scheme for mitochondrial electron transport, illustrating greater efficiency of proton translocation than the scheme in ' a ' . After Haddock & Jones, 1977 (71}. {Introduction} 8 Fig. 2: Respiratory electron donors and acceptors available to facultative anaerobes. Mid-point potentials ( £ m ) of redox couples and standard Gibbs free energies (AG°') for the oxidation of NADH by each of three oxidants are indicated. Abbreviations: T M A , trimethylamine; D H A P , dihydroxacetone phosphate; T M A O , trimethylamine oxide; After Ingledew and Poole, 1984 {87}. {Introduction} 9 AG0' - - 4 0 0 mY -67 kJ mar1 - - 2 0 0 m Y glycerol/DHAP lactate/pyruvate - ±o mY -163kJ mol"1 -+200mY -218kJmol" , REDOX COUPLES formate/co 2 H ^ H * NADH/NAD* sucdnate/f umarate TMA/TMAO -+400 mY nitrite/nitrate +600 mY + 800 mY water/dl oxygen L+1000mY {Introduction} As described above, this proposes that electron transport chains of energy transducing membranes are coupled to ATP synthesis by a proton electrochemical gradient and that electron transport and ATPase activity are each integrated with reversible transmembrane 'proton pumps'. The proton electrochemical gradient generated across the ion-impermeable membrane by electron transport forces the ATPase catalysed reaction in the direction of ATP synthesis. Evidence for the broad distribution of these essential features has been accumulating for a number of years and itself suggests a unified fundamental mechanism when combined with the association of similarly sized proton electrochemical potentials across all functional energy transducing membranes at some 150 mV to 200 mV. Among other pertinent results in support of the chemiosmotic hypothesis is that the majority of 'uncoupler' compounds retain their lipid solubility in both protonated and anionic forms and act by increasing the proton conductance of natural and synthetic membranes — thereby collapsing the proton electrochemical potential — and that of the functional assembly of bacteriorhodopsin plus beef heart mitochondrial ATPase to create a 'coupled' photophosphorylation system. This generalised scheme has been elaborated with many specific hypotheses, notably that of Williams and coworkers that the initial charge separation leading to proton translocation is an intramembrane event and that the translocated protons remain associated with the membrane surface — a localized event in which the energy of hydration of the proton in the lipid phase is responsible for the production of phosphorylated product by the ATPase {209, 210}. (iv) Structure of respiratory electron transport systems Redox potentials of electron transport chains cover a range of about 1250 mV from the redox couples of V2H2 / H + and HCOOV C 0 2 with mid-point potentials at -420 mV and -440 mV respectively, to that of 0 2 / 2H 20 at +820 mV, although the potential spanned by chains present in a functional energy transducing membrane is unlikely to exceed the 1140 mV between the NAD+/NADH (-320 mV) and 0 2/2H 20 (+820 mV) redox couples {64,84,145}. This latter span is utilized by the electron transport pathway of the eucaryotic mitochondrion, a situation indicative of its efficiency when dioxygen is available as electron acceptor. Broader potential ranges are usually prevented by incompatibility of donor and acceptor for environmental or metabolic reasons — themselves often closely related to the function of procaryotic oxidative and photophosphorylation electron transport systems. For example, the pyruvate-formate lyase generating formate in fermentative E. coli is rapidly and irreversibly inactivated in the presence of oxygen {64}. Elements of procaryotic respiratory control are discussed below in the context of the thesis topic. {Introduction} The molecular components of a respiratory chain must act effectively as redox carriers equilibrating electrons within a particular segment of the chain so that there is a net transfer between the initial donor and terminal acceptor. In order to accomplish this, and maintain the critical reversibility of electron transport associated with oxidative phosphorylation, the various components must operate within the chain at a potential approximating that of their individual mid-point potentials thereby minimising free energy changes. This criterion influences the molecular structures functioning in electron transport chains at different potentials : hence particular types of redox component are generally associated with particular respiratory functions. The initial dehydrogenases function at relatively low potentials, with succinate dehydrogenase at the upper potential limit of these activities, the mid-point potential of the succinate/fumarate redox couple being +30 mV. The dehydrogenases are generally protein complexes with at least one subunit being a flavoprotein others containing iron-sulphur clusters and, in some dehydrogenases, another may be a cytochrome. Quinones transfer electrons from the dehydrogenases to the higher potential respiratory constituents although the mechanism of their action remains controversial. Cytochromes are components of these higher potential constituents, frequently as type-6 cytochromes functioning between lower potential components and the terminal oxidase. Since oxidized quinones and reduced quinones are electrically neutral but the semiquinone, half-reduced state is a highly reactive free radical it is generally assumed that the latter form is not released as a free entity in the hydrophobic energy transducing membrane during respiration. In order to account for the transfer of two electrons from quinol to higher potential cytochromes which can accept only a single electron, Mitchell has proposed a protonmotive quinone cycle or 'Q-cycle' mechanism for the mitochondrial system in which an electron is passed from the quinol to each of cytochrome bj and cytochrome Cj with release of protons at the outer face of the membrane {138}. Electron transfer within the 'Q-cycle' is thought to proceed from cytochrome bj to cytochrome and back to the oxidized quinone at the inner membrane surface in preparation for its complete reduction to quinol (Fig. lb). The precise role of quinol in reacting with the higher potential cytochrome components has not been determined in many procaryotic respiratory systems, including those of E. coli. Terminal oxidases probably display the greatest structural diversity between the groups of respiratory components, as might be expected from the range of possible electron acceptors. The terminal oxidases naturally have the highest mid-point potentials of any respiratory component within an electron transport chain and all appear to contain multiple cytochromes. This variation between oxidase types is of particular interest for evolutionary studies : an alteration of the efficiency of electron transport may have dramatic effects upon growth rate and oxidase-related proteins are often highly conserved. Moreover, {Introduction} environmental conditions at a particular evolutionary stage may be reflected direcUy in the respiratory competence of a particular procaryote. These different types of redox carrier and aspects of their functions will be discussed below with respect to the aerobic respiratory chains of Escherichia coli. (v) Structural properties of cytochromes Lemberg and Barrett have developed a practical definition of cytochromes as : 'haemoproteins whose principal biological function is electron and/or hydrogen transport by virtue of a reversible valency change of their haem iron between ferrous and ferric forms' {112} Reactive properties of the haem moiety are modified by the protein so that certain of them are magnified and others depressed. Haemoglobins and myoglobins are not oxidized following the binding of dioxygen and the bound intermediate is stable but dissociable. Hydroperoxidases decompose hydrogen peroxide extremely rapidly with concomitant oxidation of hydrogen donors in the case of true peroxidases. In comparison cytochromes are generally unreactive with dioxygen or hydrogen peroxide, the oxidation and reduction of the haem iron underlying their metabolic function, a notable exception being the reaction of aerobic terminal oxidases with dioxygen. Consequently the protein environment of the haem is particularly important for maintaining the appropriate chemical properties of a cytochrome as well as for its interactions with other respiratory components and metabolic control of its function. The initial classification of cytochromes has been based upon the structure of the haem prosthetic group which is generally indicated by the holoprotein's visible absorption spectrum. Structures of the four haem types — a ,b ,c and d — are illustrated in Figure 3. The b and d cytochromes are characterized by protohaem and dihydroporphyrin (chlorin) prosthetic groups respectively. Other haemoproteins may use these haems to generate very different properties such as the enzymatic activities of catalases and peroxidases or the ligand binding characteristics of haemoglobins and myoglobins. Cytochromes of type b and d are of particular relevance to the current work and both true prosthetic groups in that the haem is bound non-covalently. In many of the cytochromes b for which information is available both axial coordination sites of the haem iron are occupied by ligation to imidazole nitrogens of histidine residues in the peptide chain, although other amino acids may occupy one or other of these positions, notably the e-amino group of lysine (112). Aerobic terminal oxidases would require one free axial coordination site for binding dioxygen. {Introduction} Fig. 3: Structures of protoporphyrin IX and haem prosthetic groups. Protoporphyrin IX binds an iron atom to form haem b and is modified as shown at positions 2 + 8, 2 + 4 or 5 + 6 to generate haem a, c, or d respectively. a. Haem a, ]£. Protoporphyrin IX, C.. Haem c, with covalent linkage to peptide methionine residues indicated, fl. Haem d. {Introduction} {Introduction} Because of the presence of multiple type-6 cytochromes in respiratory chains they have been distinguished by the absorption maxima of their reduced minus oxidized difference spectra, in particular by the a-band absorbance maximum which is measured at 77K to increase resolution and which is denoted as a subscript {112}. These absorption maxima vary with temperature, and as this thesis demonstrates, even the improved resolution obtainable at liquid nitrogen temperatures may be insufficient when used as the sole method of distinguishing between cytochromes with similar spectral properties. Appendix 'A' provides an introduction to the spectral characteristics of cytochromes pertinent to this study. (vi) Bacterial respiratory chains Bacterial respiratory chains may comprise one or more segments of the total complement possible, the induction of particular components depending upon environmental and growth conditions. Additionally, investigation of the possible routes of electron transfer after growth in the presence of multiple terminal electron acceptors suggests that the alternative respiratory electron transport chains induced have components in common. Thus there is the possibility of branching between electron transport chains within the energy transducing membrane. The relatively large amounts of reducible quinone present in many of these bacterial membranes has been suggested to be the common 'sink' for reducing equivalents through which multiple electron transport chains may communicate (v./.). During growth in liquid media many enterobacteria produce one aerobic terminal oxidase in response to relatively high dissolved oxygen conditions and another under poor aeration. Intermediate or changeable conditions such as those developing in the later stages of simple aerated batch cultures lead to both oxidases being present and utilized simultaneously. In spite of these additional complexities the relative ease with which certain bacterial species may be manipulated genetically, their ability to grow rapidly under defined conditions on simple nutrients such that energy sources may be better controlled and the great resource of knowledge about the organization of the genome of Escherichia coli in particular has maintained this bacterium as an organism of choice for studying both chemical and physical properties as well as the biological consequences of electron transport and proton translocation. {Introduction} B. AEROBIC ELECTRON TRANSPORT IN ESCHERICHIA COLI (i) Energetic diversity Evolutionary constraints upon the genetic development of a bacterial species dictate the type of energy source it is capable of employing, phototrophs and chemotrophs employing electromagnetic or chemical energy sources respectively. Escherichia coli, a facultative anaerobe, uses only organic substrates for the dual role of electron donor and carbon source but may exploit a variety of electron acceptors including dioxygen, oxides of nitrogen and sulphur or several reducible organic compounds including fumarate, dimethylsulphoxide (DMSO) or trimethylamine oxide (TMAO). This Gram negative enterobacterium indulges in mixed acid fermentation in the absence of external electron acceptors, hence its nutritional classification as a C-heterotrophic chemoorganotroph. (ii) Environmental and ecological factors E. coli populates the human intestine shortly after birth: indeed, at the time of Keilin's pioneering work on cytochromes from a variety of biological sources it was known as Bacterium coli communis {93}. Capable of aerobic respiration under low oxygen tensions the bacterium effectively scavenges dioxygen thereby assisting in the creation of the anaerobic conditions necessary for the remainder of the normal flora to become established. After a few weeks the majority of the bacteria in the intestine are strict anaerobes and E. coli constitutes a miniscule proportion of the total population — one thousandth to one millionth of the total cell count. The efficiency with which aerobic culture conditions may be used to recover this minor organism from such a largely anaerobic environment after 10 4 to 105 bacterial generations demonstrates that its ability to respire aerobically serves a continued purpose throughout this period. It also indicates that that the organism itself, with this capability, functions to the advantage of the bacterial population as a whole within the human intestine — in addition, presumably, to that of the proprietor {21, 87}. A symbiotic aspect of the relationship between E. coli and its human host is that the former may synthesise menaquinone as a component of its respiratory chain. This compound, vitamin K, is required for synthesis of the clotting factor prothrombin in higher animals and may also used by them as a constituent of non-respiratory microsomal electron transport systems. {Introduction} The ability to consume oxygen and also to be capable of growth in the resulting anaerobic conditions may be particularly useful in the vicinity of epithelial cells with their generous blood supply. Other nutrients available in this environment are obviously variable and their diversity open to conjecture since much of the free carbohydrate and protein is absorbed in the upper digestive tract. Growth rate in the gut is also influenced significantly by competition from other bacterial species for nutrients, especially iron {87, 146}. The ability of E. coli to thrive on a limited supply of simple nutrients has been exploited in the laboratory as well as being of use to the organism in its normal habitat. Active transport systems enable the bacterium to grow on low concentrations of specific nutrients, the energy required being supplied by electron transport or fermentative substrate level phosphorylation depending on the growth conditions. Not only do these properties permit the organism to adapt to fluctuations in its natural environments but they make it particularly useful for studies of the mechanisms and the genetic and metabolic control of these various functions. (ni) Respiratory control Control of aerobic respiratory activity is able to respond to environmental alterations very rapidly, as shown by the 35 s to 120 s periodicity of oscillations in fluorescence intensity from flavins and pyridine nucleotides following interruption of the air supply to cultures of E. coli in late exponential phase {87}. Although the rate of electron transport is dependent upon supplies of both donor and acceptor and is subject to the constraint of forming a maximal proton electrochemical potential the mechanisms regulating respiratory activity are largely unknown and their elucidation will require further data regarding the structure and activity of individual respiratory components and the precise nature of proton translocation. It should be noted that under the most frequently employed laboratory culture conditions electron transport is the rate limiting activity for growth whereas this is certainly not the case for E. coli in its natural habitat {87}. Factors affecting longer term, transcriptional control are beginning to be understood as the genomic organisation of operons coding for certain complexes is revealed. The presence of electron acceptors with high mid-point potentials represses the expression of reductases reacting with acceptors of lower potential, each of these latter being able to repress fermentative activity to some degree {194}. The fnr operon is involved in the regulation of anaerobic respiratory chains and also that of purely fermentative enzymes, originally being detected by its control over induction of fumarate and nitrate reductases. The differential stimulation of these activities that was obtained by varying the gene dosage after cloning fnr indicates that a complex interaction with other regulatory factors must {Introduction} occur, while genetic sequence and other evidence implicates the catabolite gene-activator protein CAP and the nucleotide cofactor cyclic AMP in addition to the well established repressive effects of the highest potential electron acceptor, dioxygen {31, 153, 194}. Regulation of the two terminal oxidases reacting with dioxygen is much less well understood, partly because identification of their genomic locations has been made more recently: their structural genes being sited in the cyd and cyo operonsat 16.7 min and 10.2 min respectively {6,61}. Studies using individual and dual cyd and cyo mutants have shown that E. coli aerotaxis responds to alterations generated in the proton electrochemical gradient and hence to changes in the efficiency of aerobic electron transport under low partial pressures of oxygen {185}. Respiration in rich media with high oxygen levels is naturally efficient and results in low concentrations of respiratory components per cell, the cytochrome o complex being the sole terminal oxidase present in the early stages of growth. A decreased supply of oxygen causes the induction of the cytochrome d terminal oxidase and repression of the cytochrome o complex, an adaptation reflected in the values of Km for oxygen of the two oxidases : 0.38 mM and 2.9 mM for solubilized cytochromes d and o respectively {98}. This changeover is directly attributable to the altered concentration of dioxygen but the stage of growth of the culture may also influence terminal oxidase expression {56, 87}. Anaerobic growth with nitrate as terminal electron acceptor causes repression of the cytochrome d complex but this effect is less marked if fumarate is used in place of nitrate. It is unclear how these effects are mediated in the light of evidence that the fnr gene product is also required for cyd expression, while expression of fnr itself is independent of anaerobiosis, positively regulated by cAMP and negatively by its own gene product {54, 151}. Furthermore, under semianaerobic conditions cyd expression exhibits no requirement for the fnr gene product or CAP although it is transcriptionally regulated and induction occurs when the oxygen tension of a highly aerated culture decreases below a threshold level, as in the later growth phases of batch cultures {56}. While cAMP has no direct effect upon cyd induction it has been reported to enhance expression of the cytochrome d complex in several earlier studies {34,56}. The nature of the available carbon source may also influence cyd expression quantitatively, or in the case of certain mutants, absolutely {56, 89}. Poole has reported the persistence of cytochrome o under conditions of limited oxygen, albeit with modified spectral properties {160}. The majority of studies measuring synthesis of respiratory chain components through the cell cycle have indicated a continual increase in respiratory capability, although complex results obtained in certain investigations may reflect the influence of particular methods of cell cycle synchronization {87, 181}. {Introduction} (iv) Aerobic respiratory chains — status of research prior to the current study The extent of knowledge of aerobic electron transport in E. coli has expanded dramatically over the past ten years, and especially in the last five with the advent of certain cloned constituents. Nevertheless there is still controversy over some fundamental properties of these respiratory assemblies including the number and function of participating cytochromes. Earlier descriptions of the aerobic respiratory chain of E. coli were based upon functional analyses and recognised the presence of dehydrogenases which were associated with the flavoproteins and iron-sulphur proteins that had been detected, in addition to ubiquinone-8 (a benzoquinone) plus menaquinone-8 (a naphthoquinone), and four types of cytochrome identified by their spectral properties. Characteristic terminal oxidases were detected by spectral alterations following carbon monoxide binding and were identified as cytochromes d, a\ and cytochrome o {17,18, 24, 71}. Cytochrome o was the predominant terminal oxidase in cells harvested in early exponential phase while cytochromes d and were found in greater proportion in late exponential and stationary phase cells. Cytochrome was a significant constituent of respiratory chains at all stages of growth {17, 71}. Current views of the composition of the electron transport chains of aerobically grown E. coli are given below (section B.v). Analysis of the dehydrogenases was restricted by difficulties in isolating active complexes following membrane solubilization, the succinate dehydrogenase proving to be particularly refractory to purification in an active form {71}. The broad range of possible physiological reductants available to cells grown aerobically on rich media has necessitated strict control of the nutrients supplied in the growth medium in order to limit the number and to define the type of dehydrogenases present. It was known from purification studies that -^cytochromes were associated with membrane preparations containing certain dehydrogenase activities and that some of these preparations would reduce artificial and natural quinones {71}. Non-haem iron (iron-sulphur protein) and flavoprotein were associated with these activities. The quinol component of the aerobic respiratory chains contained ubiquinol-8 but the utility of menaquinol-8 under these growth conditions was equivocal. The latter had been shown to be required for cellular activity of the anaerobic fumarate reductase through the use of auxotrophic mutants, but either of the two quinols appeared to be capable of functioning under aerobic conditions, in which ubiquinol-8 was the predominant form {42, 71}. Not only was the existence of a 'Q-cycle' mechanism unproven for the transfer of electrons from the iron-sulphur and flavoprotein centres to the {Introduction} cytochromes but the order of interaction between the quinol present and the type-t cytochromes was unknown and disputed. Indeed, the presence of alternating segments of the respiratory chain transferring hydrogen atoms and electrons had not been proven in E. coli and so the method of creating a proton electrochemical potential in these aerobically grown procaryotes was not know to be equivalent to that in mitochondria, although other bacteria such as Paracoccus denitrificans are believed to possess a similar aerobic respiratory chain to that found in eucaryotes {88}. Spectral analyses of the cytochromes of the aerobic respiratory chains had permitted an initial classification used to identify the presumed haem content of each cytochrome : cytochromes and a^ apparently containing haem-a as prosthetic group, cytochrome b± containing haem-6 , and minor quantities of a cytochrome c with type-c haem as covalently-bound prosthetic group {49, 93}. Cytochrome was subsequenUy termed cytochrome d when shown to contain a chlorin or d-type haem and more recently cytochrome aj has been renamed cytochrome 6595 since its 'a-type' spectrum has been found to be caused by a type-Z> haem containing a high-spin iron atom {18, 71,112}. This high-spin 6-haem was responsible for the spectral characteristics previously attributed to the minor amount of cytochrome c, it having been shown that there is no cytochrome c in aerobically grown Escherichia coli {4, 87,118}. Cytochrome o had been named for its oxidase activity since it was thought to have minimal absorbance properties in the redox a-band — the major spectral characteristic used for cytochrome identification — but had later been shown to be a 6-type cytochrome since hemA' mutants required incubation with haematin and ATP to generate a functional electron transport system incorporating cytochrome o {71}. Apart from uncertainty as to the true arrangement of respiratory components in the aerobic electron transport pathways of E. coli the characteristics, function, and even the number of cytochromes present in aerobically grown cells was subject to dispute. Improvement in the sensitivity of spectrophotometric techniques had led to renewed interest in the multiple roles of cytochrome fcj which had been determined to be a composite of several cytochromes b with barely resolvable spectra, each being identified by the wavelength of its low temperature a-band absorption maximum. Thus cytochrome 6553 was known to be induced in proportion and simultaneously with cytochrome d as a component of the low aeration pathway, but the function of cytochrome aj was reported variously as that of a terminal oxidase, of a non-oxidase cytochrome and of an oxidase with low oxidase activity {17, 71}. The identity of cytochromes formed under conditions of high aeration was even more contentious. Since cytochrome 0 could not be identified directly by spectrophotometric means, its a-band absorbance being obscured by that of the other type-b cytochromes, the identities and sequence of cytochromes in the aerobic respiratory chains had been (Introduction} suggested in virtually every possible combination. Thus with three major type-fe cytochromes identified by their reduced minus oxidized (redox) low temperature a-absorption bands as cytochromes 6555, 6555 and £562* cytochrome o had been variously identified as cytochrome 6555 {95,97}, as cytochrome 6556 {162,174}, as cytochrome 6557 {173}, as cytochrome 6552 {69} and as a cytochrome with a split a-absorbance band identified as cytochrome 6555-6552 {1 1 7) • The possibility remained that cytochrome 0 was actually a high-spin 6-type cytochrome with minimal absorbance in the a-band region of the spectrum — i.e. it was none of the above candidates — as was suggested by the spectrum of its CO complex, but countered by electron paramagnetic resonance (EPR) evidence which had failed to detect the presence of high-spin iron in aerobically grown cell membrane preparations {159}. It should be noted that variations in spectrophotometric precision and accuracy may cause problems in interpreting results from different laboratories when the analyses are restricted to such a narrow wavelength range. That the spectral resolution required to distinguish between these cytochromes can only be obtained at liquid nitrogen temperatures precludes straightforward kinetic analyses and the methods used for these investigations included inhibitor studies, potentiometric techniques and spectral perturbation with carbon monoxide, each coupled with low temperature spectrophotometry. As the identity of cytochrome o was strongly disputed, so too was the sequence of electron transfer between the remaining cytochrome components of this respiratory chain. The isolation of a cytochrome o preparation solubilized with the ionic detergent Sarkosyl and identified by its quinol oxidase activity had been shown to contain both cytochromes 6555 and 6552 suggesting that cytochrome 6555 was located at an earlier site in the pathway, at lower potential than either of these 'oxidase-associated' cytochromes {97,103}. Early potentiometric studies not only suffered from the complicating effects of the presence of variable quantities of the cytochromes associated with each of the two aerobic terminal oxidases, but were also carried out in isolation from low temperature spectrophotometric analyses, so that the results obtained could not be attributed confidently to specific cytochromes. Thus the nature of the aerobic respiratory chains of E. coli was uncertain at the time the current studies were initiated. Indeed, the situation was accurately summarized some time earlier in a description of similar efforts relating to molecular analyses of a rather more fundamental nature : "But when I took the pains to impartially examine the Bodies themselves that are said to result from the blended Elements, and to torture them into a confession of their constituent Principles, I was quickly induc'd to think that the number of the Elements has been contended about by Philosophers with more Earnestness, than Success." {jg} {Introduction} Moreover, there were suggestions that the efficiency of electron transport was less than that observed in the mitochondrial system, and that of just two coupling sites, one was associated with the dehydrogenase activities and one with the terminal oxidases {71,112}. Determination of the proton translocation stoicheiometry of this respiratory system, the definition of the chemiosmotic redox loops active in these respiratory membranes and their organisation required that the overall architecture of the chains should be determined, in addition to the characterization of the individual respiratory components and their functions. (v) Aerobic respiratory cytochromes — progress of the present investigation The elucidation of the number of individual cytochromes functioning in the electron transport chains of aerobically grown Escherichia coli plus the determination of their individual properties was of importance to addressing many of the uncertainties described above. In consequence the purpose of the present study was to employ analytical techniques as combined procedures in an attempt to dissect the characteristics of the cytochrome components in membrane preparations. Performing combined spectral and potentiometric analyses was already being used with some success in studying anaerobic respiratory chains in the host laboratory, although it became apparent that these would be insufficient to delineate the properties of each of the aerobic cytochromes clearly. A program of refining the available methodologies was undertaken, large improvements being made in spectrophotometric and potentiometric measurements, partly through procedural modifications and partly through instrument upgrading. In addition the technique of poised potential trapping was investigated and mutant strains were used to minimise interference from alternate electron transport paths. In spite of the combination of these approaches data acquired from membrane preparations showed that definitive characterization of individual 6-cytochrome components in situ would not be possible, although useful information was obtained regarding minimum numbers of resolvable type-Z> cytochromes present in the cytoplasmic membrane under specific growth conditions, induction and spectral properties of cytochrome d, and the existence of cytochrome pools in these resuspended membranes with different electronic transfer rates. Solubilization of the respiratory cytochromes of E. coli was known to modify their properties and activity, and many attempts at fractionation of such solubilized preparations in the host laboratory had been unsuccessful (P. D. Bragg, personal communications). Nevertheless solubilization of the respiratory cytochromes was required to separate them sufficiently to allow adequate characterization, and as a preliminary step in the purification of selected cytochromes. {Introduction} 2 3 Techniques of cytochrome solubilization, fractionation and purification were devised and developed for further investigation of several individual cytochromes or cytochrome complexes, including the cytochrome o complex, a cytochrome 6555 thought to be a component of the succinate dehydrogenase complex and another cytochrome 6555 possessing hydroperoxidase activity. The fractionation procedure was refined to permit routine resolution of cytochrome constituents solubilized from membrane preparations so that extracts of membranes from cells grown under different conditions could be screened for cytochromes 6555 associated with growth on specific substrates. Spectral, electrochemical and certain kinetic properties of the purified cytochrome o were characterized and compared to those of cytochrome o from a cloned source developed by R. B. Gennis. Continual adaptations have had to be made to the project's aims and research emphasis in order to prevent duplication of other groups' research interests and to maintain its relevance in an area of rapidly progressing knowledge. Consequendy the Results and Discussion sections of this thesis have been combined, thereby enabling the various sections of the work to be placed in context with greater clarity. The Conclusion of this thesis provides a summation and overall discussion of the project's achievements in the light of current understanding. C. EXPERIMENTAL PROGRESS CONCURRENT WITH THIS INVESTIGATION (i) Overview The vibrant interest in bacterial respiratory cytochromes over the past few years has dramatically expanded the extent of knowledge of procaryotic electron transport systems, especially those of aerobically-grown E. coli. Yet more interest has been stimulated as physiological parameters have gradually been defined and purer biochemical systems isolated to enable structural, kinetic and mechanistic properties of certain of these cytochromes and their reactions to be studied. Genetic manipulation of the respiratory cytochromes has been central to the advances in this field following location of their structural genes on the genome as have attempts to solubilize individual cytochromes or groups of cytochromes for investigations of their respective properties and functions in isolation from other components. Successes with these latter experiments have led to reconstitution attempts and the formation of model systems in which certain purified constituents of the electron transport chains have been reassembled in artificial liposomes for studying the conditions {Introduction} required to generate and maintain a proton electrochemical potential. The regulatory processes governing expression of the different respiratory cytochromes are now coming under scrutiny as molecular biological techniques are used to investigate the organisation ofgenes coding for the terminal oxidase complexes. At the commencement of these studies the arrangement of cytochromes in the aerobic respiratory chains of E. coli was disputed, as described above. A common factor among the various models proposed was that cytochromes closely associated with each terminal oxidase would interact with ubiquinol-8 in the membrane to provide quinol oxidase activity in the presence of oxygen. The models generally incorporated a split electron transport chain with the quinol component receiving electrons (or hydrogen atoms) from the dehydrogenases and delivering them to either of the two terminal oxidase complexes. Many models attempted to arrange the known cytochromes with one or more pools of quinol linking them in one or other sequence before the proposed 'branch point' to the oxidase complexes (see, for example, {42,95}). Thus the bacterium's energetic adaptability could be explained at a simplistic level by the interaction of multiple electron transfer routes which could be mobilised to the extent dictated by the concentrations of whichever electron donors and acceptors were available and yet also subject to exclusive transcriptional controls, such as that of oxygen over the anaerobic oxidases. Multiple electron transport routes connected by branch points suggested that a less rigid organisation of respiratory components might exist in the bacterial membrane than in the more specialised energy transducing membrane of the mitochondrion. Thus the bacterial system might require communication between cytochrome pools with different functional and/or topographical properties. The fifty-fold molar excess of ubiquinol over terminal oxidase in E. coli membranes, compared to seven-fold in the mitochondrion, made the small, lipophilic and therefore mobile quinol an excellent candidate for the redox mediator connecting segments of the respiratory chains {4}. However, this itself raised questions of whether the bacterial system could carry out 'Q-cycling' in such circumstances and if not, how might the quinols deliver two electrons to a cytochrome that would accept only one? The complexity of the oxidases, each associated with several distinct cytochromes, might illustrate a mechanism for solving this dilemma. (ii) Haem synthesis and incorporation into apocytochrome Recently published reports indicate that hetnA mutants, which are characterized by 5-aminolaevulinic acid auxotrophy, are due to lesions in the C5 pathway from the intact five carbon chain of glutamate via glutamyl-tRNA dehydrogenase {9, 114}. This pathway was {Introduction} previously thought to exist only in plants, algae and anaerobic bacteria whereas animals, fungi and facultative bacteria were believed to utilise the A L A synthase pathway which is now known to be absent from E. coli {114}. The incorporation of iron into protoporphyrin IX synthesised from 5-aminolaevulinic acid results in the formation of haem b and is catalysed by ferrochelatase in the mitochondria of eucaryotes. The procaryotic system probably requires enzymatic catalysis also, in view of the iron-limiting conditions in which the organism frequently exists and the extensive siderophore mechanisms developed to transport iron into the cell (section B.ii.) {146}. If hemA cells are grown aerobically without 5-aminolaevulinic acid no cytochromes are detectable by spectroscopic methods and oxidase activities are minimal {70}. When membrane preparations from these cells were incubated with haematin (haem b) plus ATP spectra typical of type-6 cytochromes were observed and NADH oxidase activity was detected. Since reconstitution of these properties was independent of protein synthesis the apocytochromes must have been synthesised in the cells and inserted into the membrane in the absence of haem, and the membrane associated apocytochromes' haem pockets must have been accessible for the haem insertion process, which required ATP {70,173}. More recently the expression of the two subunits of the cytochrome d complex has been shown to be dependent on the cydC gene, located separately from the cydAB operon containing the structural genes for cytochromes £558, 6595 and d. The polar mutation cydA prevents the production of either subunit of the oxidase complex {60, 62}. Growth of cydA' strains carrying the cydA gene on multicopy plasmids has shown that subunit I may be synthesised and inserted into the membrane independently of subunit II, as cytochrome 6558 {63}. Cells containing a cydC lesion lack cytochrome d terminal oxidase activity but synthesise low, possibly constitutive, levels of both subunit apoproteins, while haem d is completely absent {57}. The cydC strain carrying cydAB genes on a multicopy plasmid overproduces both subunit apoproteins and haem b is inserted at two sites to create cytochrome 6553 as subunit I and cytochrome 6595 which is thought to be formed by positioning of haem-6 between the two subunits {57, 62, 63}. This result implies that synthesis of haem b must be coordinated with apoprotein overproduced by the multicopy plasmid and that the cytochrome 6595 haem-binding site, previously found to be particularly sensitive to denaturing reagents, is reconstituted by the association of holosubunit I with aposubunit II which has sites for two haem d moieties {57, 62, 118}. The chlorin which forms haem d is thought to be derived from protoporphyrin IX although the biosynthetic pathway is unknown {57}. Found only as the oxygen and carbon monoxide binding moiety in the cytochrome d terminal oxidase complexes of certain bacteria the iron containing form of haem d is more labile than the unstable metal-free form {205}. Earlier {Introduction} suppositions that haem d was similar to haem d\ from bacterial dissimilatory nitrite reductases, in which the haem is a derivative of bacteriochlorin, and that haem d contained a spirolactone substituent have been modified to propose the structure shown in Figure 3 {112, 200, 205}. In the case of these type-d haems the visible spectra show many similarities in spite of substantial structural differences in their porphyrin substituents {200, 201, 205}. (iii) Immunologically based cytochrome studies The function of the small, soluble cytochrome 0552 of E. coli remains unknown although it has been purified and studied extensively to reveal a wealth of structural information. This haemoprotein was shown to be unrelated to any of the accessible domains of the membrane bound cytochromes by polyclonal antibody binding studies which also showed that the 'aerobic cytochrome 6555' was not formed in a constant ratio to cytochrome 0 , with which it had been consistently associated in previous studies, and was unrelated to the cytochrome 0555 of the anaerobic nitrate reductase (cytochrome onr) {106}. These immunological studies by R. B. Gennis and coworkers were continued with the generation of antibodies to the two aerobic terminal oxidase complexes following the isolation of mutant strains failing to produce one or other complex. These monoclonal and polyclonal antibodies were subsequently used to identify and characterize components of the oxidases during and after purification and to show that oxidases cross reacting with the cytochrome o complex or the cytochrome d complex exist in many other Gram-negative bacteria, oxidases closely related to the E. coli cytochrome d complex being particularly widely distributed {106,107,109}. Immunochemical studies of succinate dehydrogenase were used to characterise this unstable enzyme and later to aid its isolation as an active complex as discussed in relation to results from the current investigation under Results & Discussion {32, 91}. A cytochrome 0555 of Mx = 17 500 isolated by Kita et al. following solubilization of cytoplasmic membranes with a Sarkosyl/cholate mixture was thought to be the cytochrome 0555 associated with the respiratory chain terminating in cytochrome o {102}. It was reduced by D-lactate dehydrogenase from E. coli in the presence of menadione, as expected for a respiratory intermediate connecting the dehydrogenase and oxidase segments of the electron transport chain {102}. However, amino acid analyses and genetic mapping studies later identified it as the cytochrome b of the succinate dehydrogenase complex which had been investigated by means of protein chemistry by P. Owen and coworkers and by molecular biology in the laboratory of J. Guest {32, 143, 144, 215). A very recent collaboration has analysed the cloned complex after overexpression and solubilization in Lubrol PX which resulted in a high {Introduction} succinate dehydrogenase activity and a mid-point potential of the cytochrome 6555 similar to that obtained in the current study (+36 mV and +20 mV respectively) {101}. This preparation was able to generate succinate oxidase activity when reconstituted into phospholipid vesicles with ubiquinone-8 and purified cytochrome o terminal oxidase which has been proposed as evidence for a cellular electron transport chain of minimal complexity {101}. (iv) Genetically based studies of terminal oxidases The development of cyd' strains of the bacterium by R. B. Gennis and coworkers provided a major advance for such investigations just as the current study was undertaken. These strains were isolated by their failure to oxidize N,N,N\N'-tetramethyl-/?-phenylenediamine and are unable to induce the cytochromes of the 'low aeration' electron transport pathway terminating in cytochrome d {60}. Therefore they enable investigators to concentrate exclusively on those cytochromes associated with the alternative, 'high aeration' pathway which had previously been available only in low yield from early exponential phase cells. In the current study these strains have also been exploited in this way, enabling spectral and potentiometric analyses of greater resolution to be achieved and permitting fourth order finite difference spectra to be calculated with reproducible results. In the latter case, the qualitative response to the analytical method from samples of wild-type strains results in interference from disproportionate contributions to the analysis from constituents of 'minor' respiratory chains which are induced under imperfectly controlled conditions of growth and harvesting. One of the original cyd' mutants was subsequently shown to possess both regulatory and structural lesions, features of which are discussed in relation to results from this investigation (Results & Discussion) {69}. Subsequently the Gennis laboratory generated cyo' strains which fail to produce the cytochrome o complex, but grow adequately in aerobic conditions by expressing cytochrome d {8}. Potentiometric analyses of membrane suspensions from these oxidase deficient strains have indicated that cytochrome 0 and the cytochrome 6555 associated with it have mid-point potentials of +165 mV and +35 mV respectively, the values for cytochromes of the cytochrome d complex — which are spectrally distinct — being +260 mV {d), +180 mV (6558) and +150 mV (b595) {117}. By combining the techniques of potentiometric titration and spectrum deconvolution Stouthamer and coworkers analysed cytochromes b in membrane preparations of aerobically grown wild-type cells. Unfortunately no indication of the quantity of cytochrome d (and hence of cytochrome b^g) was given, although poised potential low temperature spectrophotometry provided values of +187 mV {Introduction} and +46 mV for cytochromes b with 77 K redox a-band absorbance maxima at 563.5 nm and 555.7 nm which may have corresponded to cytochrome o and cytochrome 6555 respectively {203}. Other in situ potentiometric titrations of type-o cytochromes from wild-type cells grown under high aeration had also indicated that the two major species had Em values of +175 mV to +200 mV and of +30 mV to +70 mV {66.80,170}. (v) Purification and characterization of detergent solubilized terminal oxidases Procedures used for the isolation of detergent-solubilized terminal oxidase complexes incorporating either cytochrome o or cytochrome d have now been reported by the laboratories of Y. Anraku, R. B. Gennis and H. R. Kaback as well as from the current study {99, 130, 135, 214}. Purification and characterization of the cytochrome d complex in Zwittergent solution showed that it comprised two peptides with MT = 57 000 (subunit I) and 43 000 (subunit II) with the three cytochrome constituents 0553, 0595 and d {118,135}. These two subunits could not be separated without denaturing the complex which possessed quinol oxidase activity as isolated or when reconstituted into membranes {104,135}. There are two moles of cytochrome d and one mole of each type-o cytochrome associated with each mole of the complex {4}. Subunit I corresponds to cytochrome 0553, which is transcribed from the cydA gene, is a transmembrane protein and has been overexpressed and characterized in isolation from subunit II {4,62,63}. Oxidation of quinol substrates by the complex is inhibited by antibodies to subunit I which has been shown to possess the quinol binding site and to monitor the redox steady state of the ubiquinol-8 pool in the aerobic respiratory chain {115,219}. Subunit II contains two type-d haems or chlorins which EPR studies indicate to be located close to the cytochrome 0595 {4, 74}. The cytochrome 0595 was renamed from cytochrome a\ when the cytochrome d complex was shown to lack haem a and when spectral deconvolution analyses revealed a 'peroxidase type' cytochrome b with high-spin haem iron and a small a-band absorption at 595 nm {118}. This cytochrome binds carbon monoxide weakly and is present in the purified complex but is not isolated with either subunit, suggesting that the haem may be inserted between the two peptides and therefore that it may be particularly susceptible to protein denaturation : this latter characteristic may account for earlier observations attributing CO binding properties of this cytochrome to oxidase activity {163}. Potentiometry of the purified cytochrome d complex yielded mid-point potentials of +232 mV (d ), +55 mV to +150 mV (0553) and +113 mV (0595), with each Em value being sensitive to pH and that of cytochrome 0553 particularly sensitive to the solubilizing detergent {105, 119}. Recent electron {Introduction} paramagnetic resonance and spectrophotometric studies have indicated that the cytochrome d is extremely stable in the oxygenated state which forms spontaneously when the membrane bound or solubilized forms are exposed to air, and which affects the EPR signal of cytochrome 6595 : the EPR studies also showed that oxidized cytochrome 6553 contains high-spin haem iron {4, 74,116}. Purification of the cytochrome o terminal oxidase complex in the laboratories of Y. Anraku and H. R. Kaback has resulted in preparations containing two copper (II) atoms and comprising either two {97} or four subunits {129}. The current study demonstrates that there are at least four subunits, none of which stains strongly for haem and one showing split bands in SDS polyacrylamide gel electrophoresis. The physical and spectral properties of this solubilized complex are described in detail with the high resolution results obtained for this thesis, low temperature redox difference spectroscopy exhibiting a complex absorption pattern with at least three absorption maxima in the a-band {214}. Three type-ft cytochrome components were observed in potentiometric titrations, the high potential component undergoing radical mid-point potential shift upon complexing with carbon monoxide and the two lower potential components being susceptible to moderate perturbations in Em when the reduced forms interacted with potassium ferricyanide. The complex has been shown to be a quinol oxidase, reducible by duroquinol, its activity inhibited by the quinol analogue HOQNO during both oxidative and reductive phases. The high potential cytochrome has a split redox a-band at 77 K which absorbs most strongly at 563.5 nm and also at 555 run and the other two components absorb at 555.0 nm and 557.0 nm. Other published spectra have demonstrated a split a-absorption band but at much lower resolution, work from the present laboratory suggesting that a cytochrome b component with an a-absorption maximum at 562 nm had a high mid-point potential {69,97}. The spectral identity of cytochrome o is still being disputed in the literature with Kaback and colleagues suggesting that of two type-6 cytochromes in the complex one is cytochrome o with a redox absorption maximum at either 558 nm or 562 nm, Anraku and coworkers identifying cytochrome 0 as cytochrome 6555 and the Gennis laboratory proposing that it is a split absorption cytochrome 6555/562 {3}. The results of the current study indicate possible reasons for this confusion, for there appears to be synchronous reduction of component cytochromes within the complex. Moreover, the current analysis of the complex expressed from the cyo operon cloned by R. B. Gennis shows spectral and potentiometric distinctions in the absence of gross structural differences from the cytochrome 0 complex prepared from wild-type cells. Evidence is presented suggesting that the complex may exist in different stable states which may have functional relevance. The EPR and resonance Raman data of Anraku which suggest that cytochrome 6555 reacts {Introduction} with carbon monoxide are in direct opposition with the titration and poised potential results from the current study which show the CO binding species to be cytochrome 6552 . also termed cytochrome 6553 as a result of a shift in a-absorption maximum upon solubilization {3, 202, 214}. Anraku's group has also suggested that the cytochrome 0 complex in both reduced and air-oxidized states contains a haem with a high-spin iron atom, in contrast to the earlier investigations which found no evidence for such a constituent of this terminal oxidase {159,202}. Working with intact, wild-type cells Poole and Chance have investigated the kinetics of photolysis of the CO adduct of cytochrome o in the presence of dioxygen at various subzero temperatures {160, 164}. The C O binding reaction is slower and more rapidly reversible by photolysis than that with dioxygen and the latter produces oxygen-bound intermediates with spectral characteristics similar to those of the C O adduct which may be trapped below 180 K (-98°C) for further analysis {164,165]. (vi) Molecular biological analyses of terminal oxidase genes Although genes for both terminal oxidases have now been cloned by the laboratory of R. B. Gennis and their products overexpressed in growing cells, no data has been published providing sequences for any of the components, nor for the number or arrangement of components in the cyo operon {6,61}. The cybB gene, encoding another aerobic respiratory cytochrome b, has also been cloned and its product isolated by solubilization in Sarkosyl or Triton X-100 followed by HPLC fractionation or standard ion-exchange chromatography respectively {100, 141]. Sequence data for this cytochrome also remains unpublished although purification studies have shown it to be a dihaem cytochrome 6 5 5 1 with redox a-absorption maxima at 555 nm and 561 nm {100, 142}. The mid-point potential of cytochrome 0 5 5 } is reported as +20 mV, although the value of n is not provided and the titration profile may constitute two individual haem responses that overlap substantially as has been found and discussed in the current study and in the cytochrome b™ of the E. coli respiratory nitrate reductase {67,142]. Membrane preparations from strains producing amplified quantities of cytochrome 0551 demonstrated aerobic steady state reduction of this cytochrome with substrates of NADH or D-lactate, indicating that it is a respiratory component preceding ubiquinone-8 in the electron transfer chain, although wild-type strains are reported to produce minor quantities of cytochrome 0551 {142}. The product of the cybA gene is a cytochrome 0 5 5 6 , which has been cloned as the sdhC gene and is described above in relation to succinate dehydrogenase activity (section C.ii) {215}. {Introduction} Both the DNA sequence and the amino acid sequence have been determined, but to date no indication has been provided of which residues bind the haem iron {144, 215}. Evidence is provided in this thesis to suggest that there are multiple cytochromes £555 in the membranes of aerobically grown E. coli although many other workers, notably those of the Anraku and Gennis laboratories have assumed that the succinate dehydrogenase cytochrome 6555 is the sole cytochrome present with these spectral characteristics {3,4}. Consequently earlier data indicating that a respiratory cytochrome &556 is associated with the cytochrome 0 branch of aerobic electron transport has been ignored when schemes have been proposed for the overall architecture of the respiratory systems. These topics are addressed in the Conclusion. (vii) Reconstitution experiments and models of aerobic respiratory architecture Since the current studies were initiated both of the two aerobic terminal oxidases have been purified by other laboratories and each oxidase complex has subsequently been reconstituted into liposomes in order to examine its mechanistic properties and interaction with other respiratory components. Initial experiments by Kaback and coworkers using membrane vesicles from mutants lacking D-amino acid dehydrogenase or D-lactate dehydrogenase demonstrated that an intact respiratory chain, functioning with the appropriate substrate, was reconstructed when one or other dehydrogenase had been reconstituted into either side of the membrane {72,148}. That the same effect was obtained with simultaneous reconstitution of the dehydrogenases into such vesicles suggested that they bound at dissimilar sites. As both activities generated a proton translocating step which preceded the oxidases it was proposed that this exists as a common reducible intermediate between the flavin-linked dehydrogenases and the cytochromes and that the standard chemiosmotic model of redox loops, as initially described for mitochondrial systems by Mitchell, is inappropriate for describing the first coupling site of the E. coli respiratory system {72,148}. Reconstitution of purified cytochrome d terminal oxidase complex plus purified pyruvate oxidase complex and ubiquinone-8 into pure phospholipid vesicles catalysed electron transport between pyruvate and oxygen with the generation of a 180 mV transmembrane potential which was sensitive to uncouplers {104}. When reconstituted into vesicles without the flavoprotein pyruvate oxidase the cytochrome d complex was shown to serve as a respiratory coupling site and to function as a quinol oxidase with specificity for ubiquinol-8, whereas it will also oxidize menaquinol-8 in detergent solubilized solutions {87,104}. Several reconstitution studies have been performed with the purified cytochrome o terminal {Introduction} oxidase complex in spite of the lack of consensus on the identity of cytochrome o itself. When sufficient phospholipid is present with the purified oxidase complex to form proteoliposomes, the fluorescence of the lipophilic probe N-phenyl-l-naphthylamine (NPN) will reflect the redox state of the complex {116}. This constitutes the uncoupler insensitive component of NPN fluorescence which requires sufficient lipid for the oxidase complex to be incorporated into vesicles and which occurs in addition to that observed with intact cells of E. coli in which the probe responds to energization of the cytoplasmic membrane {182}. Other experiments in which the cytochrome o complex was reconstituted into phospholipid vesicles or planar bilayers demonstrated that it would generate a membrane potential when supplied with artificial electron donors {73, 96}. These potentials were dissipated by protonophore uncouplers and inhibited by applied voltages of up to + 150 mV on the substrate side of a bilayer or by the oxidase inhibitors KCN and HOQNO, a quinol analogue. The cytochrome o oxidase was thus identified as a coupling site for oxidative phosphorylation. The earliest demonstration that membrane potentials generated in vitro could be used as models for cellular processes was provided by reconstitution of two purified complexes into phospholipid vesicles, one being the cytochrome o complex and the other the lac carrier protein from E. coli. By adding ubiquinol as electron donor a proton electrochemical potential was set up across the membrane, its magnitude being dependent upon the concentration of cytochrome o. The proteoliposomes would then transport a variable amount of lactose, dependent upon the potential created, against a lactose concentration gradient {128}. Construction of artificial electron transport chains has been attempted by reconstituting the purified cytochrome o complex into phospholipid vesicles in the presence of a purified dehydrogenase or bacterial photosynthetic reaction centres. In the former case electron transfer between D-lactate dehydrogenase and oxygen via cytochrome o was obtained in right-side-out and in everted membrane vesicles from aerobically grown cydr cells and also in vesicles reconstituted from phospholipid, D-lactate dehydrogenase and ubiquinone-1. The proton electrochemical potential generated by limited addition of either D-lactate or quinol indicated the existence of a single site of generation with a lH+/e" stoicheiometry, whereas that obtained on addition of NADH suggested the presence of two sites, one before and one after the quinone step {126}. Hybrid proteoliposomes containing the photosynthetic reaction centre plus the cytochrome o complex have been used to generate a proton electrochemical potential with associated oxygen consumption by means of steady illumination {140}. Flash illumination caused emission of single electron from each reaction centre, creating a pulse of quinol and transitory reduction of the oxidase. This reduction was sensitive to inhibition by quinol analogues and the reoxidation by dioxygen was {Introduction} sensitive to cyanide inhibition {140}. No spectral distinction could be made between the oxidase haems, even when a 'kinetic potentiometric titration system' was created by flashing at poised potentials. Thus the cytochrome o terminal oxidase appears to function through vectorial electron translocation and scalar proton transfer, creating a proton electrochemical potential when reduced within the lipid bilayer by ubiquinol-8 and reoxidized at the inner surface of the cytoplasmic membrane by dioxygen. These concepts will be addressed further in the Conclusion in the light of results from the current study. D. SYNOPSIS OF RESEARCH AIMS AND STRATEGY This work describes the resolution of E. coli aerobic cytochrome properties by combinations of several methodologies which had previously been employed independently. By determining the number of cytochrome species and elucidating their individual characteristics their precise role and their interaction with other electron carriers may be better understood. The initial results described relate to studies in which reference cytochromes were utilized to refine the spectrophotometric and potentiometric techniques for optimal sensitivity. Data produced by both of these methodologies now resolve the properties of individual cytochromes to a greater degree than those reported in the literature, whether the cytochromes are in solubilized preparations or beside in membrane vesicles. Although there should be less artifactual perturbation of spectral and potentiometric properties of membrane proteins in vesicle preparations than in solubilized solutions the complexity of the results obtained from the multiple type-6 cytochromes present in membranes prevented reliable interpretation. Attempts were made to generate mutant strains with aberrant or missing type-b cytochromes, including those of the cytochrome o complex of the 'high-aeration' respiratory chain. Several mutants were isolated and determined to contain an altered complement of respiratory cytochromes-^  when grown aerobically. Nevertheless their characterization remains unclear despite extended investigation. Protocols were developed for the solubilization and fractionation of E. coli aerobic respiratory cytochromes. The resolving power of these techniques exceeds that of related liquid chromatographic procedures published elsewhere. The fractionated cytochromes were subsequently investigated individually by the modified analytical techniques described above. The combination of {Introduction} improved techniques for fractionation and analysis permitted investigations of detergent-solubilized fractions of cytochrome 6555 and cytochrome o. The former comprised several similar cytochromes, each induced independendy. Cytochrome o was isolated and its spectral, electrochemical and certain kinetic properties characterized. Nevertheless the details of the interaction of cytochrome o with components preceding it in the aerobic respiratory chain remains open to speculation as indicated by the conflicting reports continuing to be published. Comparison of these proposals' predictions with the current data is provided in the Conclusion. {Materials & Methods} 3 5 M A T E R I A L S & M E T H O D S (a) Chemicals All reagents used in this study were of the highest quality available from commercial sources. Components of growth media were of ACS grade or higher. The cytochrome c used for reference studies was 'type VI' from equine heart, and the catalase was from bovine liver: both proteins were obtained from Sigma Chemical Co., St. Louis, MO, USA. Purified bovine hepatic microsomal cytochrome 65 (soluble trypsinized fragment) and triphenanthroline cobalt were prepared and generously provided by Dr. A. G. Mauk, Department of Biochemistry, U.B.C.. (b) Cell types The various strains of Escherichia coli K-12 used during this study are indicated in the text and listed in Appendix 'B' with their genotypes and sources. The constituents of bacterial growth media and of buffer solutions are described in Appendices ' C & 'D' respectively. (c) Growth of cells Cells were grown in batch culture at 37°C to the desired phase of growth on one of several minimal media as indicated in the text (see list in Appendix 'C'). Additions of nutritional supplements and antibiotics required to maintain the expression of desired phenotypic characters are described in detail in the text. Growth phase was estimated by optical density readings measured with a Perkin-Elmer 124 double beam spectrophotometer at 600 nm (OD60o)- Two standard incubation techniques were used, batches of up to 8 L being grown in 500 mL aliquots within 2 L Erlenmeyer flasks rotated at 300 rpm in a New Brunswick incubator/shaker and larger batches (to 22 L) being grown in a static vessel vigorously sparged with compressed air or a mixture of nitrogen/carbon dioxide (95:5 w/w) as indicated in Appendix ' C . Sparged cultures also received 200-500 u.L Antifoam Reagent A (Dow Corning Silicones Ltd., Downsview, ON.) which was added to the bulk medium with the final inoculum. Conditions were altered in some experiments to modify the degree of aeration of the cultures: details of these techniques are given in the text. Inoculum size was 10 % of the fresh medium volume. {Materials & Methods} 3 6 (d) Crude membrane preparation Cells ready for harvesting were cooled rapidly on ice and centrifuged at 10 000 x g for 25 minutes at 4°C in Beckman JA10 rotors, washed twice with TM buffer, then weighed and either used immediately or stored at 0°C for a maximum of 12 hours before use. All subsequent procedures were carried out at or below 4°C. After being resuspended in 1.5 volumes TM buffer the cells were disrupted by two passages through a French press (AMINCO, Inc; Silver Spring, MD, USA) at 1400 kg cm"2 (20 000 lb in"2) in the presence of a few crystals of calf thymus deoxyribonuclease type 1 (Sigma Chemical Co., St. Louis, MO, USA). Unbroken cells were removed by centrifugauon at 10 000 x g for 20 minutes in a Beckman JA20 rotor and the supernatant diluted three-fold before being centrifuged for 2.5 hours at 250 000 x g in Beckman Type 60Ti or 45Ti ultracentrifuge rotors. The resulting membrane pellets were washed by resuspension in TM buffer using a Thomas Teflon™-glass homogenizer and sedimentation as before to yield the 'crude membrane fraction'. Crude membranes were placed on ice as pellets and resuspended in the appropriate buffer immediately prior to further processing. (e) Inner membrane preparation Inner membranes were purified from crude membrane preparations by one of three methods as indicated in the text: (1) Crude membranes were resuspended in TDGA buffer, pH 7.5, and sedimentedby ultracentrifugation as described above. The pellet was then resuspended in TDGB buffer, pH 7.5, at a membrane protein concentration of 20.0 mg mL"1 and dialysed against 3 L of the same buffer for 16 hours. After further ultracentrifugation the soft, pigmented upper layer of the resulting pellet was readily removed with a spatula to provide the inner membrane preparation. (2) Crude membranes were resuspended in TM buffer containing 0.5 mM dithiothreitol to a final protein concentration of 10.0 mgmL"1. An equal volume of fresh 10 M urea containing 1.0 mM PMSF (pH 7.8, 20°C) was added slowly and the solution stirred for 15 min at 0°C. The urea-washed membranes were centrifuged for 3.0 hours at 250 000 x g causing the inner membrane to separate as a soft brown upper layer of the resulting pellet, distinct from the hard, white outer membrane layer below. The inner membrane layer was removed with a spatula and immediately washed by resuspension in TM buffer and recentrifugation to remove traces of urea. (3) Crude membranes were prepared in the presence of 0.1 mM dithiothreitol, then resuspended and centrifuged into a sucrose cushion as described below in section 'k'. {Materials & Methods} 3 7 (f) Solubilization of membranes Crude membranes and inner membranes were solubilized by resuspending in TX buffer to a protein concentration of 10.0 mg mL"1 using a Thomas Teflon-glass homogenizer. The homogenate was then stirred slowly at 0°C while an equal volume of TTX buffer was added one drop at a time, generating final extraction concentrations of 5.0 mg mL"1 membrane protein and 5.0 % Triton X - l 14. The mixture was stirred gently for 30 min at 0°C, pH 7.8 and centrifuged for 1.0 hour at 250 000 x g in Beckman Type 60Ti or 45Ti ultracentrifuge rotors to sediment unextracted material. The supernatant constituted the 'Triton-solubilized' crude membrane or inner membrane preparations used in this study. (g) Fractionation of solubilized cytochromes Solubilized membranes were used as a source of membrane cytochromes suitable for fractionation by a variety of liquid chromatography techniques. On a routine basis Triton solubilized membranes were diluted with an equal volume of T T E buffer to yield final concentrations of 2.5mgmL"1 membrane protein, 10.0 mM Tris-HCl, 5.0mMMgCl2, 0.5 mM EDTA, 2.5 % (w/v) Triton X - l 14 and 0.5 % (w/v) Triton X-100 at pH 7.8. This solution was stirred at 0°C for 45 min to permit equilibration of the detergents with the sample. Conductivity of the solution was checked to ensure that it was below 2.0 mS cm"1 before being loaded onto a (12.0 x 1.5)cm column of DEAE-BioGel.A (Bio-Rad Laboratories, Richmond, CA, USA) previously equilibrated with TTE buffer. The loaded column was washed with two column volumes of TTE buffer in order to remove significant quantities of unbound cytochrome and elution was achieved by a linear gradient of (0-400)mM KC1 in ten column volumes of T T E buffer. Flow rates were approximately 20 mL hr"1 during loading and washing but a constant flow of less than or equal to 2.0 mL hr"1 was required during elution to ensure consistent resolution of cytochromes, the latter rate being maintained with a peristaltic pump utilizing a stepping motor. The protein content of each 2 mL fraction was estimated by a modified Lowry assay (v.j., section 'r'), the elution of visible chromophores was monitored at 412 nm in a Perkin-Elmer Lambda 3A double beam spectrophotometer (A412), and the relative ionic strength recorded from a Markson Model 10 conductivity meter. Those fractions potentially containing significant quantities of cytochrome, as shown by the A 4 1 2 measurements, were analysed by low-temperature reduced minus oxidized difference spectra as described below in section '1'. Resolution and recovery of cytochromes from a single DEAE-BioGel.A column operated under the stated conditions for 72 hours were superior to those obtained by fractionating samples upon multiple ion-exchange and gel filtration matrices over a similar period at higher flow {Materials & Methods} 3 8 rates. Significant variations of these techniques are described at appropriate junctures in the text. (h) Purification of cytochrome o The cells used for routine cytochrome o preparations were cyd' cells which had been grown to stationary phase although certain experiments, indicated in the text, required w+ cells grown to mid-exponential phase. Inner membranes were prepared from crude membranes by the urea-wash technique ('(e) 2' above). This preparation was then resuspended in T M buffer and a solution of 10.0 % (w/v) sodium cholate in T M buffer was slowly added to yield final concentrations of 4.0 mg mL" 1 membrane protein and 6.0 % (w/v) sodium cholate. The solution was stirred gently on ice for 30 min and then centrifuged for 60 min at 250 000 x g in a Beckman Type 60Ti or 45Ti ultracentrifuge rotor. After resuspending the pellets in T M buffer the washed, urea+cholate-stripped membranes were sedimented by recentrifuging for 2.5 hours at 250 000 x g. These membranes were then subjected to Triton X-l 14 extraction using T X and T T X buffers as described above (section T ) . Fractionation of cytochromes to purify cytochrome o was accomplished by modification and extension of the liquid chromatographic techniques described earlier, the fractions of each stage being monitored for cytochrome, protein and conductivity as indicated in section ' g ' above. An initial DEAE-BioGel.A column of (12.0 x 2.4)cm or (12.0 x 1.5)cm, depending upon the quantity of material being purified, was equilibrated with T T E buffer at pH 7.8, loaded with the sample, washed and then eluted with a (0-400)mM KC1 gradient with a constant 2.0 mLhr"1 flow-rate as previously described (section 'g ') . The fractions recognised as 'Peak II' were pooled and the cytochrome precipitated by adding slowly, with stirring, 361.0 mg (NH4)2S04 per millilitre of pooled sample: the solution was stirred slowly, on ice, for 10 min. The cytochrome was recovered as a floating precipitate when centrifuged at 12 000 x g in a Sorvall HB-4 swinging-bucket rotor and was resuspended in a minimal volume of T T E buffer (typically 500 uL) before loading onto a (37.5 x 1.5)cm column of Sephacryl S-300 'Superfine' (Pharmacia Inc., Uppsala, Sweden) equilibrated with T T E buffer, pH 7.8. The column was eluted with T T E buffer at a constant 2.0 mL hr"1. The major cytochrome peak was then pooled, checked to ensure that its ionic strength resulted in a conductivity value below 2.0 mS cm"1 and loaded onto a (12.0 x 1.5)cm or (8.0 x 0.8)cm DEAE-BioGel.A column equilibrated with T T E buffer. After washing the sample on the column with T T E buffer the cytochrome was eluted at a slow rate from the column by a pH-gradient created by pumping from the second of two linked reservoirs containing: (i) four column volumes of 1.0% (w/v) Triton X-100, 7.5 mM citric acid, 1.0 mM {Materials & Methods} 3 9 EDTA (bom of the latter being in free acid form), pH 4.0 with KOH, and (ii) four column volumes of T T E buffer, pH 7.8. Cytochrome remaining on the column at the end of the pH gradient was eluted with 200 mM KC1 in T T E buffer, pH 7.8. Both the pH gradient and salt-eluted fractions were shown to contain purified cytochrome o by spectrophotometric and gel electrophoretic techniques. Exchange of buffer and detergent in preparation for potentiometric titration was achieved by a three-fold dilution of the cytochrome o sample with D T E buffer, pH 7.8, and loading it onto a (7.5 x 50.0)mm column of DEAE-BioGel.A equilibrated with D T E buffer, pH 7.8 . The cytochrome was washed with ten column volumes of D T E buffer, pH 7.8 and then eluted in a concentrated fraction of approximately 2.5 mL with D T E buffer, pH 7.8, containing 300 mM KC1. This fraction was concentrated to approximately 500 u.L in an Amicon ultrafiltration cell using a PM-10 ultrafiltration membrane (Amicon Corp., Danvers, MA, USA) thereby restoring the Triton X-100 concentration to circa 1% (w/v). This 500 uL sample was loaded onto a (7.5 x 420)mm column of Sephacryl S-300 'Superfine' (Pharmacia Inc., Uppsala, Sweden) and eluted into 600 uL fractions with a running buffer of 0.2 % Triton X-100, 100 mM potassium phosphate, pH 7.0 at a flow rate of 1.0 mL hr-1. The Sephacryl column was calibrated under the standardized running conditions with 'molecular weight protein standards'. (i) Partial purification of the 'Peak HI' respiratory cytochrome The liquid chromatographic procedure for fractionating solubilized cytochromes described above in section (g) resulted in the preparation of a series of major cytochrome 'peaks', several of which were purified further. Recoveries of 'Peak I' cytochromes were adequate for subsequent analytical procedures but insufficient for preparative techniques. Fractions contributing to 'Peak II' were pooled and used for the isolation of cytochrome o (section (h)) and to separate 'Peak III' cytochromes when present (v./.). The 'Peak IV fractions were pooled for further investigations of their cytochrome content and the associated hydroperoxidase activity. The 'Peak III' respiratory cytochrome was observed as a component of a fused 'Peak II/III" when cells grown aerobically on L-proline were subjected to the standard procedures of membrane cytochrome extraction and fractionation on DEAE-BioGel.A. Fractions making up this fused peak were pooled and loaded onto an hydroxylapatite column of BioGel.HTP (2.5 cm diameter, 3.5 cm in length) (Bio-Rad Laboratories, Richmond, CA, USA) which had previously been equilibrated with T T E buffer, pH 7.8. The sample was washed with T T E buffer and eluted with a 200 mL gradient {Materials & Methods} 4 0 of (0-400)mM potassium phosphates, pH 7.8, in T T E buffer at pH 7.8. (J) Partial purification of the Teak IV hvdroperoxidase After the initial DEAE-BioGel.A separation of cytochromes from Triton-solubilized membranes (section 'g') the fractions recognised as 'Peak IV were pooled and either used directly for analyses or were prepared for potentiometric titration by buffer and detergent exchange procedures similar to those employed for the'Peak IF fraction (cytochrome d) described in section'h'. In the latter case pooled 'Peak IV fractions were diluted with an equal volume of D T E buffer, pH 7.8, and loaded onto a (5.0 x 30.0)mm column of DEAE-BioGel.A previously equilibrated with D T E buffer, pH 7.8 . The cytochrome was washed with ten column volumes of D T E buffer, pH 7.8 and then eluted in a concentrated fraction of approximately 2.5 mL with D T E buffer, pH 7.8, containing 400 mM KC1. This fraction was concentrated to approximately 500 uL in an Amicon ultrafiltration cell using a PM-10 ultrafiltration membrane thereby restoring the Triton X-100 concentration to circa 1% (w/v). The 500 (iL sample was loaded onto a (7.5 x 420)mm column of Sephacryl S-200 'Superfine' using T T E buffer, pH 7.8 as running buffer and eluting at a rate of 1.0 mLhr"1 into 600 u.L fractions which were monitored at 412 nm for cytochromes and for protein by the modified Lowry technique described below in section's'. An analogous ion-exchange plus ultrafiltration procedure was used to concentrate the major cytochrome peak from this first gel filtration column after which it was oaded onto a (7.5 x 420)mm column of Sephacryl S-300 'Superfine' using a running buffer of 0.2 % (w/v) Triton X-100, 100 mM potassium phosphate, pH 7.0 in preparation for subsequent potentiometric titration : fractions were collected and monitored as for the previous gel filtration step. Both Sephacryl columns were calibrated under the standardized running conditions with 'molecular weight protein standards' (Pharmacia Inc., Uppsala, Sweden). (k) Preparation of nitrate reductase Large batches of cells of strain RK4353 were grown on N R medium (Appendix'C') at 37°C in static culture and sparged with the (N2 + C0 2) gas mixture as described in 'c', above. When the cells had reached stationary phase the culture was cooled on ice for 45 min and harvested in the standard manner. The sample temperature was maintained at (Q-A)°C during all subsequent steps. Crude membranes were prepared as in'd', with the exception that all buffers were degassed and contained O.lmM dithiothreitol {28}. The upper, dark brown layer enriched in inner membranes was resuspended in 400 mL T M buffer containing O.lmM DTT and 25 mL aliquots were overlaid {Materials & Methods} 4 1 upon 2.5 mL cushions of 60 % (w/v) sucrose in TD buffer in preparation for ultracentrifugation for 2.5 hours at 250 000 x g in Beckman Type 60Ti or 45Ti ultracentrifuge rotors. The resulting dark brown sucrose solution was diluted to 11.0 mg mL"1 protein with TD buffer and a 20 % (w/v) solution of Triton X-100 in TD buffer was added dropwise as the preparation was stirred slowly on ice. The Triton extraction was continued for 60 min after which it was centrifuged at 250 000 x g for 1.5 hours. The supernatant was loaded onto a (2.5 x 31)cm column of DEAE-BioGel.A previously equilibrated with TD buffer containing 0.1% (w/v) Triton X-100. Fractions of 12.0 mL were collected as the column was washed with 2.5 volumes of equilibration buffer and eluted with a gradient of (0-300)mM NaCl in six column volumes of the equilibration buffer {67}. (I) Spectrophotometric analysis of cytochromes Two double beam, dual wavelength analytical spectrophotometers were available: an Hitachi/ Perkin-Elmer model 356 and an SLM/Aminco model DW2c with accompanying Midan II data processor and plotter. Both instruments are capable of measuring small optical absorbances in highly opaque samples such as aqueous suspensions of crude membranes. The DW2c is capable of spectral resolution of such samples to within 0.5 nm when adjusted optimally (v.j.). Pyridine haemochromogen estimation was performed by the method of Falk {49), with 'reduced minus oxidized' difference spectra and absolute reduced spectra being measured at ambient temperature on either of the two analytical spectrophotometers. Ambient temperature reduced minus oxidized difference spectra of cytochromes, membrane suspensions and solubilized material were each collected with a 10 mm pathlength and a bandwidth of 2.0 nm (PE-356) or 2.2 nm (DW2c) to cover the (400-700)nm spectral range: in many cases the (380-400)nm range also provided useful information. Potassium phosphate buffer, 100 mM, pH 7.0, was used to suspend and dilute samples before analysis. Electrochemical reduction and oxidation of the sample and reference solutions was achieved with a variety of reagents depending on the requirements of the experiment, but a few grains of Na2S204 and fresh 3.0 % (v/v) H 2 0 2 were routinely employed as reductant and oxidant respectively. Extended equilibration times at ambient temperatures are required for spectrophotometry of preparations containing Triton X-l 14 due to the 20°C cloud-point of this detergent; low temperature difference spectra are more suitable for such samples. Low temperature reduced minus oxidized difference spectra were obtained at liquid nitrogen temperature (-196°C, 77 K) in specialized sample chambers available as accessories for both PE-356 and DW2c spectrophotometers to provide increased resolution of the absorption bands of cytochrome {Materials & Methods} 4 2 spectra. Standard conditions include dilution of sample to double strength in 100 mM potassium phosphate buffer, pH 7.0 followed by dilution with an equal volume of the buffer containing 2.0 M sucrose {214}. The sample was split into two 1.0 mL aliquots, one being oxidized as the reference (normally with a drop of 3.0 % (v/v) H202) and the other being reduced (usually with a few grains of Na2S204) before placing each into the appropriate locations in the brass (PE-356) or aluminium (DW2c) sample holder. The holder was then carefully immersed in liquid nitrogen and left submerged to equilibrate for a minimum of 20 min after which it was positioned in the cryogenic chamber's Dewar containing the maximum volume of liquid nitrogen compatible with a stable signal {66,90,206,212}. Sample pathlengths were 1.0, 2.0 or 3.0 mm as indicated in the text; standard conditions are indicated in Table I. While the use of sucrose at 1.0 M dramatically increased both sensitivity and spectral resolution obtainable with the instruments devitrification was unnecessary with these samples and is generally reserved for cryogenic spectroscopy of glycerol solutions {206, 212}. Samples were poised at a variety of selected 'low temperatures' for studies of the photon induced relaxation of the complex formed between carbon monoxide and certain cytochromes in heir reduced state. These distinct temperature conditions were produced in the Dewar of the PE-356 cryogenic sample chamber by combinations of dry ice with ethanol (-75°C, 198 K), or wet ice with either an equal volume of ethanol (-30°C, 243 K) or with NaClto 2.0 M (-12°C, 261 K) {83}, the latter conditions being monitored with the remote temperature probe. Path Length Bandwidth Instrument: Data rendition Optimal scan speed Ambient Cryogenic Ambient Cryogenic PE-356 : Electromechanical 0.5 nras'' 10.0 mm 3.0 mm 2.0 nm 2.0 nm DW2c : Electronic immaterial 10.0 mm 2.0 mm 2.2 nm 0.8 nm Table I: Spectrophotometer settings for optimal spectral resolution. Resolution of the a-bands of ambient and low-temperature difference spectra was optimized by adjusting the bandwidth and scan speed of the instrument performing the spectral analysis and by averaging the results of up to nine successive spectral scans. These parameters varied between instruments and the critical values are indicated in Table I. Spectral calibration of the {Materials & Methods} 4 3 spectrophotometers was achieved by reference to the absorption spectrum of a standard holmium oxide filter (SLM/Aminco #A-0724) and to the visible emission maxima of each instrument's deuterium lamp at (486.0, 656.1 & 972.0)nm. Reduced plus carbon monoxide minus reduced difference spectra, or 'CO-binding' difference spectra — commonly referred to as 'carbon monoxide difference spectra', were routinely measured at ambient temperature. Instances of cryogenic measurement of such spectra for detailed analysis of their a-bands are indicated in the text. The sample was diluted to the required concentration in 100 mM potassium phosphate buffer, pH 7.0, reduced by adding a few crystals of Na2S2C"4 and carefully separated into two cuvettes without aeration. One of these reduced fractions was placed in the spectrophotometer's reference beam and the other was transferred to a fume-hood and had carbon monoxide bubbled through it gently from a fine nozzle for 90 s after which it was sealed with Parafilm™ and placed in the instrument's sample beam. The CO-treated sample was maintained under dark conditions throughout the gassing and loading procedures and exposure to the analytical light beams was kept to a minimum before spectral data were collected. Similar precautions were taken with low temperature measurements, the reduced, gassed samples being transferred to the pre-cooled cryogenic sample holder with a CO-filled Pasteur pipette, and frozen to the required temperature under a slow stream of carbon monoxide. Subsequent replacement of the cooling solution in the Dewar enabled temperature modulated relaxation studies to be performed as indicated above. (m) Derivative analysis of spectrophotometric data Derivative spectra were plotted after calculation of the fourth-order finite difference spectra of original data gathered by the DW2c. Data storage and manipulation was achieved with the Midan II processor installed with the spectrophotometer by the manufacturer. Interpretation of these derivative spectra is described in section I.i.d of the Results & Discussion {180, 186}. (n) Redox kinetics of cytochromes Dual-wavelength spectrophotometry was used to monitor the kinetics of cytochrome oxidation and reduction by means of the net absorption of the sample at (559.0-580.0)nm. Membrane suspensions were prepared at a protein concentration of 5.0 mg mL"1 in Hepes buffer and preoxidized in a standard cuvette with 1.0 uL of 3.0 % (w/v) aqueous H202 {123}. The addition of microlitre volumes of concentrated chemical or biological reductants plus rapid mixing initiated reduction of {Materials & Methods} 4 4 sample components, including dissolved oxygen, and established dynamic redox equilibria. The reduction state of the population of cytochromes was then followed spectrophotometrically. The effect of respiratory inhibitors and amphipathic electrochemical mediators was also tested as were the kinetic responses to reoxidation by chemical and biochemical reagents {214). (o) Potentiometric titrations Procedures described in this section are modifications of those developed by several investigators in this laboratory over an extended period {66, 69,167,170}. The potentiometric titrations were carried out using the PE-356 spectrophotometer in split beam mode and with its secondary sample chamber modified to accept a lateral stirring motor and a Dutton-style side-arm cuvette accommodating a top-mounted combination platinum electrode with internal reference (Fisher #13-639-82), remote temperature probe (YSI Tele-Thermometer Model 42SC, Yellow Springs Instrument Co., Yellow Springs, OH, USA) and nitrogen flushing lines {6, 66). A single side-arm sealed with a serum stopper provided a port for the addition of chemical oxidants and reductants. Nitrogen gas of 'prepurified' grade (minimal CO content) was passed through a two stage scrubbing system to remove oxygen and in which it was also water-saturated and brought to neutral pH. Stage one comprised sparging from a fritted glass inlet filter through a 500mL volume of zinc amalgam plus ammonium vanadate in HC1 prepared as described by Meites and Meites {67, 131}. This was followed by passage through two successive Fisher-Milligan gas-exchange chambers each containing 300 mL of the phosphate buffered solution of methyl viologen, proflavin and EDTA developed by Sweetser {67,198). The treated gas was passed at a slow rate across the surface of the sample in the modified Dutton cuvette during the equilibration period and throughout the the titration. The exhaust gas was passed through a small bubble chamber in order to monitor the exit flow rate and chamber pressure and to check that all seals were gas-tight. The sample was prepared by diluting it with potassium phosphate buffer (100 mM, pH 7.00) to a known concentration of protein in 25 mL such that, optimally, the a-band of the ambient temperature redox difference spectrum provided approximately 70 % deflection on the PE-356 at a full scale range setting of 0.1 A. This corresponded to a protein concentration of 5-10 mg mL"1 protein in typical membrane preparations and provided adequate sensitivity while minimizing interference by baseline shifts resulting from changes in mediator oxidation states. Moreover, by operating the spectrophotometer in split beam rather than in dual beam mode as traditionally used by other investigators {8, 81} a convenient yet accurate estimation was readily obtained of changes occurring in the spectral baseline as the potential of the sample altered {66,105, {Materials & Methods} 4 5 117,187}. A 2.0 mL fraction was removed from the sample preparation, oxidized in a cuvette with several crystals of K3Fe(CN)6, sealed with Parafilm™ and placed in the reference beam of the spectrophotometer. Mediators were immediately added to the remaining sample which was mixed by inversion and sealed in the Dutton-style cuvette (5.0 u.L each of 5.0 mg m L - 1 fresh mediator solutions as indicated in Table II, yielding final concentrations of approximately 20 u.M). The sample was allowed to equilibrate under scrubbed nitrogen, with stirring, for at least 60 minutes during which the temperature stabilized at 305 K. Monitoring of the potentiometric titrations was achieved by scanning the a-band of the ambient temperature reduced minus oxidized difference spectra at 1.0 runs"1 from 580.0 nm to 530.0 nm (650.0 nm to 530.0 nm for titrations of the cytochrome d complex) at approximately 5 mV intervals throughout the potential range of the sample. As the monochrometer scanned through the wavelength of maximum absorbance approximately 560 nm) the electrochemical potential was read from the millivoltmeter O^ isher 'Accumet' Model 325) connected to the platinum electrode. Collecting spectral data over the wavelength range indicated permitted an accurate baseline to be drawn on the plotted output and the peak height to be calculated at various wavelengths of interest; generally X = 558.0 nm, X m a x and A. = 563.0 nm for type-6 cytochromes. Monitoring was initiated during the autoreduction that was observed after the initial equilibration and continued throughout the subsequent reduction which was achieved at an extremely slow rate by injecting microlitre quantities of fresh phosphate-buffered NADH through the rubber septum sealing the side-arm of the Dutton cuvette. This slow reduction rate ensured full redox equilibration within the sample suspension. Monitoring of non-membranous sample solutions began with the NADH reduction step. Subsequent oxidation was achieved by injecting a fresh, phosphate-buffered, concentrated solution of either K^FetCN)^ or H2O2. Spectral and potentiometric data were gathered during the reductive phases of several reduction and oxidation cycles. Samples of membrane suspensions occasionally generated hysteretic and non-reproducible results when monitored during oxidation, presumably due to differences in the redox equilibration rates of certain sample components in comparison with those of the electrode. Consequently each oxidative phase was followed by an extended equilibration period of at least 30 minutes. The final measurements of each titration were taken at potentials below -300 mV following full reduction of the sample with Na2S2C"4 in order to observe the presence of any low potential cytochromes and to check that no significant sample degradation had occurred during the protracted experimental procedure. Titrations carried out in the presence of carbon monoxide were accomplished under forced ventilation with direct venting to the exterior of the building of the N2 and CO gas supply systems, {Materials & Methods} 4 6 Stock Reagent Em n solvent concentration 1 -monocarboxyferrocene +530 1 ethanol 5.0 mg mL"1 triphenanthroline cobalt III +370 1 buffer 5.0 mg mL"1 quinhydrone +270 1 buffer a 5.0 mg mL"1 2,6-dichlorophenolindophenol +224 2 buffer a 5.0 mg mL"1 1,2-naphthoquinone +157 2 ethanola 5.0 mg mL"1 phenazine methosulphate + 92 2 buffer b ' c 5.0 mg mL"1 phenazine ethosulphate + 55 2 buffer b - c 5.0 mg mL duroquinone ± 0 2 ethanol 5.0 mg mL"1 menadione - 50(?) 2 ethanola 5.0 mg mL"1 2-hydroxy-l,4-naphthoquinone - 139 2 ethanola 5.0 mg mL"1 anthraquinone-2-sulphonate -225 2 water a 5.0 mg mL"1 a Warming is required to dissolve the reagent at the indicated concentration. b Extremely light-sensitive in aqueous solution. c Possible neoplast. Table II: Electrochemical mediators for potentiometric titrations. These reagents were used collectively as electrochemical mediators in potentiometric titrations. The buffer used to dissolve the hydrophilic reagents was 100 mM potassium phosphate, pH 7.0 . Fresh stock solutions were made up immediately before each titration. {Materials & Methods} 4 7 the spectrophotometer sample chamber and the titration vessel exhaust gases. These titrations followed a standard reductive titration of the sample. A moderate flow of carbon monoxide was passed into the fully reduced sample through the stoppered side-arm of the titration vessel such that it bubbled slowly through the stirred liquid and was emitted from the nitrogen outlet. Nitrogen was not flushed over the sample surface during the fifteen minutes of carbon monoxide treatment, but was restarted immediately afterward at a minimal flow rate in order to maintain a positive internal gas pressure and prevent sample interaction with oxygen. Carbon monoxide was introduced through the side-arm periodically during the remainder of the procedure. Exposure of the sample to light was minimized throughout the treatment. Potentiometric data were modified by a simple FORTRAN program to generate data tables of standardized results and derived values in formats suitable for input to the BMD:P3R software for non-linear regression analyses (BMDP Statistical Software, Inc., Los Angeles, CA, USA) and into the TELL-A-GRAF two-dimensional graphics software (Computer Associates International, Inc., Garden City, NY, USA) running under the Michigan Terminal System on the University of British Columbia's Amdahl mainframe network (MTS-G, UBCnet). FORTRAN subroutines had been developed for one-to four-component curve fitting using BMD:P3R (40,66): additional subroutines were written to permit data-fitting analyses with up to six distinct theoretical components although it was determined empirically that the program was unable to resolve components differing in mid-point potential by less than 50 mV. Following tabulation, data from BMD:P3R analyses were converted to a format acceptable to TELL-A-GRAF by *TELLABANK software provided by the U.B.C. Computing Centre. (p) Spectrophotometry at selected electrochemical potentials Two methods of electrochemical poising were employed, one achieved kinetically and the other through potentiometric titration : both employed rapid freezing at 77 K as the method of voltage clamping during the extended spectrophotometric analyses. Kinetic poising of samples at specific potentials was either achieved by exploiting the temporary 'steady states' of reduction induced in the electron transport components of membrane suspensions by the addition of chemical oxidants or biological substrates to an aerated sample or by utilizing the partial reduction equilibria created during substrate reduction of solubilized cytochrome preparations. In both cases a preliminary experiment was required using timed dual-wavelength spectrophotometric studies to determine the duration of each phase of reduction at a specific temperature. Subsequent potentiometric poising could then be accomplished under equivalent {Materials & Methods} 4 8 conditions on the bench-top in the spectrophotometer's cryogenic sample holder which would be frozen rapidly in liquid nitrogen at the prescribed time. The frozen sample could then be subjected to standard low temperature spectrophotometry. Alternatively one spectrophotometer could be used to monitor a sample's reduction kinetics in dual-wavelength mode (PE-356) while the other (DW2c) was set for low temperature split-wavelength mode analyses of fractions withdrawn from the sample chamber of the first instrument and rapidly frozen under an inert atmosphere as described above. Thus the freezing to 77 K enabled the poised electrochemical potential of the sample to be maintained throughout the extended time period required for spectral determination as well as enabling accurate spectrophotometric analyses to be performed. During potentiometric titrations aliquots could be withdrawn through the septum sealing the side arm of the Dutton cuvette using a nitrogen-flushed 1.0 mL syringe fitted with a 2" 18-gauge stainless steel needle. The withdrawn sample could then be deposited into the sample holder of the cryogenic accessory under a stream of nitrogen in preparation for rapid freezing and low temperature spectrophotometry. The Results section describes control experiments in which the titrations were carried out in the presence and absence of 1.0 M sucrose in order to obviate the physical problems associated with mixing sucrose into the sample after the latter had been removed from the titration vessel. In these experiments the cryogenic accessory would be precooled with the reference chamber containing the frozen reference sample. It was possible to undertake such experiments with a single spectrophotometer by carrying out the titration itself in the Dutton cuvette within a light-proof box (thus protecting the light-sensitive mediators), withdrawing the sample in the manner described above and using the spectrophotometer exclusively for gathering cryogenic spectra. The decision of sampling time would be based upon the electrochemical potential values provided by the millivoltmeter. A superior technique was to monitor the titration spectrophotometrically at ambient temperature as well as potentiometrically, withdrawing samples from the 'titration' spectrophotometer (PE-356) as described above, and to transfer these under nitrogen to precooled cryogenic sample holders in preparation for immediate analysis in the second spectrophotometer (DW2c). (q) Stopped-flow spectrophotometry An Aminco/Morrow stopped-flow accessory (SLM/Aminco, Urbana, WI, U.S.A.) was utilized in conjunction with the SLM/Aminco DW2c spectrophotometer operated in dual wavelength mode for stopped-flow rapid kinetic analysis of the reduction of Triton-solubilized cytochrome o preparations. A fully oxidized sample of cytochrome in TTE buffer at a membrane protein {Materials & Methods} 4 9 concentration of 5.0 mg mL"1 was injected into the reaction chamber with an equal volume of fresh duroquinol (maintained under nitrogen) or Na2S2G*4, each at one of several concentrations. The ensuing reaction was monitored over times extending to several seconds at sample and reference wavelengths of 560 nm minus 575 nm. (r) Genetic manipulation of E. coli cells Transduction between E. coli strains with the generalized transducing bacteriophage P l v j r was performed following standard procedures {134}. Mutagenesis of E. coli strains was achieved using the nitrosoguanidine techniques described by Miller {134}. Survival curves were generated and used to determine exposure times necessary to produce 50 % killing under standard conditions: fresh mutagen at 50.0 mg mL"1 in 100 mM sodium citrate buffer, pH 5.5, 37°C containing 5 xlO8 freshly suspended exponential phase cells per millilitre. Treatments of 46 min and 15 min duration were required for strains PLJ01 and PLJ04 respectively. Exposure was terminated by sedimentation of the cells for 1.0 min in a microfuge followed by two washes in sterile 100 mM potassium phosphate buffer, pH 7.0, resuspension in the same buffer and immediate dilution for plating or for enrichment and selection procedures. Generally, P l V i r transduction was followed by straightforward enrichment and selection procedures exploiting the constructs' anticipated auxotrophic or antibiotic resistance characteristics {134}. Identification was achieved by screening for multiple phenotypic markers on additional selective media and, for cytochrome variants, by redox spectrophotometry {89 ]. When manipulating cytochrome genes the transduction or mutagenesis of cultures was followed by 'enrichment' growth under anaerobic conditions for 16-24 hours on minimal medium M9-K supplemented with glucose, a fermentable carbon-energy source. Although this shift to minimal medium caused these cultures to experience a significant lag phase it ensured that recovery of nutritional auxotrophs was minimized while providing the least growth disadvantage to cells undergoing alterations of genes governing the expression of aerobic respiratory components. Modifications of loci within the cyd operon were carried out by Plv ly cotransduction of the neighbouring nadA or sdh genes from Escherichia coli strains carrying the required characteristics of cyd gene expression, and thereby either modifying the recipient's auxotrophic status with respect to nicotinic acid or changing its ability to utilize succinate as sole carbon source {89}. Cells experiencing a positive change of status in either of these marker genes were obtained by direct selection on minimal agar plates containing the appropriate nutrient. Isolation of transductants carrying negative markers was more efficient after enrichment using penicillin-G at 50 mg mL"1 for {Materials & Methods} 5 0 several hours in minimal medium M9-K followed by sedimentation plus sterile washing as described above and replica plating onto minimal agar plates with and without the relevant nutritional supplement. Cells containing mutations associated with the cytochrome o respiratory pathway were obtained from strain PLJ01 in which the expression of the cyd operon is nutritionally dependent and from strain PLJ04 which is cyd' {89}. Mutagenised cultures of PLJ01 were exposed to two rounds of penicillin enrichment. Following 16 hours' anaerobic growth in 5.0 mL minimal medium M9-K with glucose they were sedimented in a Beckman JA-20 rotor at 12000 x g for 10 min at 22°C, resuspended in 5.0 mL glucose supplemented M9-K, partially degassed in a partly evacuated, uncharged desiccator and grown to stationary phase in standing culture over approximately nine hours. Glucose causes induction of the cyd operon under these conditions of low oxygen tension, consequently cyd' mutants and those unable to produce any functional cytochromes were rapidly outgrown by cyd+ cells. The cells were sedimented as before, and resuspended in 5.0 mL M9-K with lactate and succinate as carbon-energy sources. This culture was vigorously aerated with water-saturated, filter-sterilized air and after four hours another 5.0 mL M9-K was added, containing lactate plus succinate and also sufficient penicillin-G to yield a final concentration of 50 mg mL" 1. The airflow through this initial penicillin-enrichment culture was maintained for two more hours in order to kill cells using cytochrome o for energy production; cyd expression would have been repressed under these highly aerobic conditions rich in lactate and succinate, thereby preventing the growth of cyd+cyo' cells and promoting their survival. Since lactate and succinate are non-fermentable carbon sources cells unable to synthesise both aerobic terminal oxidases would also survive this enrichment procedure, as would all obligatorily fermentative cells: hence the importance of the primary outgrowth step described above. After two hours the cells were spun down as before and washed in 10.0 mL sterile 100 mM potassium phosphate buffer, pH 7.0 for twenty minutes in order to complete the lysis of penicillin-weakened cells. The cells were washed again with the buffer and resuspended in 5.0 mL M9-K medium with glucose as carbon-energy source. The culture was partially degassed as before and grown to stationary phase at 37°C for 15 hours in order to enrich or cyd* cells. Samples were removed for plating after this stage and also after the entire penicillin enrichment procedure had been repeated. Plating was carried out onto M9-K medium plus glucose followed by growth in a gas jar under low but significant oxygen tensions so that cyd + cells would grow rapidly, whether or not active cytochrome o was present. Colonies from these plates were subsequently picked onto gridded M9-K plates containing either glucose or a combination of lactate plus succinate and screened for ability to grow solely on the former under normal aerobic conditions at {Materials & Methods} 5 1 37°C. Thus cyo' cells were selected from wild type and from obligatorily fermentative mutants including those which are unable to produce any functional cytochromes, eg. unc~ and hem' strains, respectively. Another method used to generate cyo' strains from mutagenized cells of both PLJ01 and PLJ04 was that of Au et al.. This technique uses anaerobic conditions and a growth medium incorporating glycerol as sole carbon/energy source with nitrate as electron acceptor in order to select against obligatorily fermentative mutants {8}. Well aerated growth medium containing lactate plus succinate and supplemented with ampicillin at 40 mg mL"1 enables enrichment for cyo' cells. Survivors are grown under the established anaerobic conditions and tested for inability to grow under aerobic conditions on lactate and succinate. Under these environments both PLJ01 and PLJ04 would behave with a cyd' phenotype. Selection of Tets mutants from strains carrying TnlO was accomplished by the Davis implementation {37} of the technique of Bochner {13} in which induction of tetracycline resistance by non-toxic autoclaved chlortetracycline results in a sensitivity to growth inhibition by the ion chelating activities of fusaric and quinaldic acids. Plasmid DNA 'mini-preps' and full-scale CsCl purifications of plasmid DNA were carried out by the standard methods described in Maniatis' manual {123}. (s) Biochemical assays Protein concentrations were estimated by the method of Lowry et al. {29,120} as modified for greater sensitivity {23} and the presence of membrane lipid {125}. Bovine serum albumin fraction V was used as the reference (Sigma Chemical Corp., St. Louis, MO, U.S.A.). The Triton family of detergents interferes with several common methods of protein determination, including those of Barrett {178], Bartlett {11, 52} and Lowry et al. {120}. Consequently protein measurements of samples containing such detergents were executed in the presence of an equivalent concentration (w/v) of SDS in the assay reagent 'mix' as there was Triton added with the sample {125, 175}; this technique minimizes the errors due to the Triton but lowers the sensitivity of the of the assay and requires centrifugation of fractions high in potassium salts in order to remove finely suspended precipitates of potassium dodecyl sulphate. Colour development in the presence of SDS required approximately one hour and was stable for 20 hours in the dark at room temperature, during which time the fine precipitates settled to form a loosely packed pellet. Fumarate, nitrate and TMAO reductase activities were assayed in anaerobic cuvettes by the method of Jones & Garland {92} The oxidation of reduced benzyl viologen in the presence of an {Materials & Methods} 5 2 aliquot of sample membrane suspension by stock solutions of the appropriate substrate was monitored at 660 nm {66}. Oxidase activities were determined by dual wavelength kinetic analyses of cytochrome reduction (section 'n') indicating the rate at which dissolved oxygen was depleted from a freshly mixed sample and assuming an aqueous dioxygen saturation concentration of 260 uM. (t) Sodium dodecyl sulphate polvacrylamide gel electrophoresis Protein electrophoresis was carried out on denaturing polyacrylamide gels of either 10 % or 13 % (w/v) with a 4 % (w/v) stacking gel using protocols developed by Laemmli {110}. It was found that gel penetration and component resolution was superior if solubilized cytochrome o samples were not boiled in the SDS solution before loading onto the gels. When loading samples derived from fractions eluted off of ion-exchange columns with potassium salt gradients it was necessary to prevent precipitation of potassium dodecyl sulphate in the gel's loading well. This was accomplished by one of two techniques: either warm water at 50°C was circulated through the cooling system of the gel apparatus during the loading procedure and the initial ten minutes of electrophoresis or the sample was mixed into a sample buffer containing 0.5 % (w/v) SDS instead of the standard 2.3 % (w/v) SDS. The efficacy of each approach and the resulting resolution of peptide components depended upon the concentrations of potassium ions and protein in the samples. Fairbanks stain, incorporating Coomassie Brilliant Blue R250 was used for routine staining of SDS-PAGE gels {47} which was followed by destaining and fixing in 10 % (v/v) aqueous acetic acid. Photochemical silver staining of SDS-PAGE gels was also carried out when extra sensitivity was required {132}. SDS-PAGE of protein samples to be stained for haem was carried out with gels which had been 'pre-electrophoresed' in order to remove excess ammonium persulphate {199}. Staining for haemoproteins separated on these gels under reducing or non-reducing conditions was accomplished using these proteins' residual peroxidase activity in conjunction with the addition of hydrogen peroxide and either dimethylbenzidine {53} or 3,3',5,5'-tetramethylbenzidine {90, 133, 199}. {Results & Discussion} 5 3 R E S U L T S & D I S C U S S I O N H s R E F E R E N C E S T U D I E S O F S O L U B L E C Y T O C H R O M E S This initial section provides an overview of the techniques that were employed for the analysis of E. coli respiratory cytochromes. Three soluble cytochrome preparations were used as standard protein solutions when developing and calibrating the procedures and instruments utilized in this study : mitochondrial cytochrome c (equine heart), catalase (bovine liver) and the soluble fragment of trypsinized microsomal cytochrome b$ (bovine liver), each being obtained from the sources described under Materials & Methods. (i) Redox Difference Spectroscopy Appendix A provides a brief description of those spectral properties of cytochromes that are pertinent to the current investigation. Difference spectra were collected at ambient and liquid nitrogen temperatures in order to compare values obtained from standard cytochrome solutions against those recorded in the literature. These spectra included reduced minus oxidized (redox) difference spectra, reduced plus carbon monoxide minus reduced (CO-binding) difference spectra and reduced minus oxidized difference spectra of the pyridine haemochromogen derivatives of selected cytochrome samples. Analytical redox difference spectrophotometry was routinely performed at 77 K in order to exploit the greater resolution of cytochrome a-absorption bands obtainable under these conditions as is illustrated below. {Results & Discussion} 5 4 Fig. 4: Visible range redox difference spectra of soluble reference cytochromes. Ambient temperature (295 K) [reduced minus oxidized] difference spectra of soluble reference cytochromes in 100 mM potassium phosphate buffer, pH 7.0. Cytchrome c refers to equine heart cytochrome c, cytochrome b$ refers to the soluble tryptic fragment of bovine hepatic microsomal cytochrome 65. AA = 0.20. a . Cytochrome c, 12.5 ug mL"1 (Sigma-Aldrich, St. Louis, MO) plus cytochrome b$, 10.0 ug mL"1 (Dr. A. G. Mauk, U.B.C.). b . Cytochrome c, 25.0 ug mL"1. C . Cytochrome 65, 10.0 ug mL 1 . d . Baseline. {Results & Discussion} 5 6 Cryogenic accessories each incorporating a Dewar flask holding liquid nitrogen within the modular sample chamber of the respective high-resolution spectrophotometer were employed for these analyses as were the signahnoise enhancement techniques described under Materials & Methods. (a) Ambient temperature redox difference spectra Visible redox absorption spectra of standard mammalian cytochromes are shown in Figure 4. The major spectral features of each are the Soret, or y-band absorption peak at (420-430)nm, the (3-band absorbance at (520-535)nm and that of the oc-bands at (545-565)nm. Precise values for the absorption maxima are given in Table III. While each of the absorption peaks of the type-c cytochrome was at a lower wavelength than the corresponding peak of cytochrome b$ the combined solution showed a fused Soret peak without distinguishable characteristics of the individual components (Fig. 4). Although absorbance in the a-band region was less intense than that of the Soret (typically having a relative intensity at 295 K of one eighth the Soret size, and approximately one fifth at 77 K) this region of the combined spectrum displayed features of both individual spectra as is illustrated in greater detail in Figure 5. Analyses of P-band absorption patterns offered reduced intensity and resolution in comparison to those of the a-absorbance region. Although the cytochrome b$ a-band spectrum was clearly asymmetrical at ambient temperature, this was not the case for that of cytochrome c although the location of the type-c a-band absorption maximum at a shorter wavelength than that of the type-6 cytochrome was recognisable (Fig. 5a,c). Some characteristics of the a-bands of both component cytochromes were displayed by the ambient temperature a-band spectrum of the combined solution (Fig. 5e). (b) Low temperature redox difference spectra The improvement in resolution of spectral features of the reduced minus oxidized cytochrome a-bands when observed at low temperatures was demonstrated by the expanded mid-range spectra of Figure 5 (curves b, d & f). This technique provided narrower absorbance peaks with a shift of the maximal absorption to shorter wavelength and an increase in absorbance, although for practical purposes the increase in the extinction coefficient was offset by the requirement for a shorter pathlength due to the opacity of the sample below freezing temperatures and the need for a uniform sample cooling rate {90}. Both standard, soluble cytochromes dispayed a pronounced biphasic {Results & Discussion} 5 7 Reduced minus Oxidized Difference Spectra Temp. (K) max (nm) max (nm) max (nm) Reference # Cytochrome bs 295 423 526 556 {112} (bovine liver) (171) (26) {112} 295 423.5 527 556.3+560.0 {tunc} 77 — — 552+559 {112} 77 422.5 525.5+532.0 551.8+557.3 {nine} Cytochrome c 295 416 520.5 550.0 {112} (equine heart) (129) (15.9) (21) {112} 295 415.5 520.5 550.0 {hinc} 77 415+432 519 538+546+549 {112} 77 417+434 519.0 538+546+548.5 {hinc} Pyridine Haemochromogen Derivatives Reduced minus oxidized difference spectra Absolute Spectra (reduced form) Temp. (K) max (nm) Reference # Temp. (K) W max (nm) max (nm) max (nm) Reference # Haem b 295 557 {112} 295 419 526 557 {48,49} (20.7) {112} (191.5) (17.5) (34.4) {48,49} 295 556.5 {hinc} 295 419 525.0 556.5 {hinc} Haem c 295 522 551 {48,49} (18.6) (29.1) {48,49} 295 414 520.5 550.5 {hinc} Table III: Spectral characteristics of standard cytochromes c, Z?5 and their pyridine haemochrome derivatives. Absorption maxima and extinction coefficients are provided for the salient features of the dithionite reduced minus peroxide oxidized difference spectra of these cytochromes when observed at ambient (295 K) and liquid nitrogen (77 K) temperatures. Extinction coefficients are given in parentheses below the wavelength to which they apply; units = mM" 1; 'hinc' refers to the current study. {Results & Discussion} 5 8 Fig. 5: a-absorption bands from ambient and low temperature redox difference spectra of soluble reference cytochromes. Ambient temperature (295 K) [reduced - oxidized] difference spectra of soluble reference cytochromes in 100 mM potassium phosphate buffer, pH 7.00 and low temperature (77 K) reduced minus oxidized difference spectra of soluble reference cytochromes in 100 mM potassium phosphate buffer, 1.0 M sucrose, pH 7.0. Cytchrome c refers to equine heart cytochrome c, cytochrome b$ refers to the soluble tryptic fragment of bovine hepatic microsomal cytochrome b$. a. Cytochrome c, 25.0 ugmL- 1 ; 295 K ; A A = 0.025, b . Cytochrome c, 12.5 ugmL- 1 ; 77 K ; AA = 0.050, C . Cytochrome b5, 20.0 ug m l / 1 ; 295 K ; A A = 0.025, d . Cytochrome b5, 10.0 ugmL- 1 ; 77 K ; AA = 0.050, e . Mixed cytochromes c, 25.0 ug mL" 1, and b5, 20.0 ug mL" 1; 295 K ; A A = 0.025, f. Mixed cytochromes c, 6.75 ugmL" 1, and b$, 10.0 ugmL" 1; 77 K; AA = 0.050. {Results & Discussion} 6 0 spectrum at 77 K with contributions from each spectral discontinuity being clearly observable in the mixed solution. These 77 K measurements provided clear evidence of the multiple features of the a-band spectra in each case. As described in Appendix A the majority of individual cytochromes possess symmetrical a-band absorption peaks which undergo a blue shift at the low temperatures associated with high resolution spectrophotometry of these compounds {90, 112, 220}. Thus the chosen cytochrome standards were atypical in that they displayed biphasic absorbance characteristics. Although there have been no type-c cytochromes definitively associated with the respiratory chains of aerobically-grown Escherichia coli the distinct spectral properties of these standard solutions facilitated optimization and calibration of the analytical equipment (Table I) {section II.A.i.b.1,4, 87}. An example of a haemoprotein with a haem containing high-spin iron was provided by catalase, illustrated in Figure 6a;l, the a-band region being expanded in Figure 6b;l. The minimal a-band absorbance in relation to that of the Soret region is apparent, the spectrum of cytochrome c being shown for comparison (Fig. 6a;3, 6b;3). (c) Pyridine haemochromogen redox difference spectra The pyridine haemochromes are haem complexes formed by coordination with two molecules of a base. They may be prepared by denaturation of haemoproteins and solubilization of the haem in alkaline pyridine, their preparation from c-type cytochromes requiring cleavage of the covalent bond between haem and peptide (Materials & Methods) {48,49}. Curves 2 + 4 of Figures 6a and 6b show the reduced minus oxidized pyridine haemochrome spectra of samples prepared from catalase and cytochrome c respectively. It can be seen from Figure 6b that the iron atom in the the type-b haem of catalase relaxed into a low-spin form resulting in greater a-band absorbance by the haem than occurred in the original cytochrome with its 'high-spin' iron {49,158, 216}. Since they result from solubilized free haem the redox spectra of pyridine haemochromogens of all type-fe haems are equivalent. Consequently this technique provided a means of detecting the types of haem present in a complex mixture of cytochromes (Fig. 6b) even when the cytochrome samples were constituents of membrane complexes {48}. {Results & Discussion} 6 1 Fig. 6: Redox difference spectra of pyridine haemochromogen from types b and c soluble reference cytochromes. Ambient temperature (295 K) [reduced - oxidized] difference spectra of soluble reference cytochromes in 100 mM potassium phosphate buffer, pH 7.0 . Catalase refers to bovine hepatic catalase (Sigma-Aldrich, St. Louis, MO), cytchrome c refers to equine heart cytochrome c. a. Broad visible range spectra: 1. Catalase, 1000 u.g mL"1 ; 2. Pyridine haemochromogen derivative of catalase at 1000 ug mL" 1 ; 3. Cytochrome c, 10.0 |ig m L 1 ; AA = 0.200, AA = 0.200, AA = 0.040, 4. Pyridine haemochromogen derivative of cytochrome c at 10.0 ng mL"1 ; AA = 0.040. b_. Expanded a- and (3-bands of spectra in (a): 1. Expanded a-and P-band region of (a,l) 2. Expanded a-and P-band region of (a,2) 3. Expanded a-and P-band region of (a,3) 4. Expanded a-and P-band region of (a,4) AA = 0.040, AA = 0.016, AA = 0.004, AA = 0.004. (Results & Discussion} 6 2 {Results & Discussion} 6 3 (d) Fourth order derivatives of redox difference spectra Fourth order finite difference spectra were calculated as 'fourth order derivatives' from the a-band region of 77 K reduced minus oxidized difference spectra using upgraded firmware routines resident in the Midan II data processor interfaced with the SLM/Aminco DW-2c spectrophotometer. Excessive electronic noise prevented useful fourth order derivative analyses being obtained from the Soret region of the spectra when the instrument was adjusted for maximal resolution and sensitivity in the a-band range, a property of the optical design of this instrument. Figure 7 shows examples of fourth order derivative spectra calculated from redox spectrum a-bands of a combined preparation of mammalian cytochromes c and 65 measured at both 77 K and 295 K. Four major peaks may be observed in either case, these peaks having maximal absorbance at the same wavelengths as the dual peaks observable in individual preparations of these cytochromes at each temperature (Table III). Although the four distinct features of the combined spectrum are clearly apparent in the redox difference spectrum measured at 77 K the utility of determining component features from complex spectra is revealed by the ability of the fourth order derivative to distinguish four constituent peaks from the 295 K redox difference spectrumwhich displays a fused spectral aggregate in the a-absorption region (Fig. 7a,b). The individual contributing absorption maxima from the 77 K spectra correlated closely with values provided in the literature although such comparative analyses have not been presented for ambient temperature spectra (Table HI) {48,49}. Fourth order derivative studies of the reduced absolute ultraviolet absorption spectrum of equine heart cytochrome c by Dufiach and coworkers have been used to determine the influence of pH upon intramolecular movement of aromatic amino acid residues {46}. Thus a comparative calibration could be carried out to ensure the validity of wavelength values determined from derivative analyses in the current investigation (Fig. 8). It should be noted that in the current study the amplitude of peaks in the fourth order derivatizations provide only qualitative data since the height of each derivative peak is dependent upon the curvature of the original spectrum over the wavelength range described by that particular derivative peak. Consequently the amplitude of any peak corresponding to a specific discontinuity in the original spectrum is influenced by the proximity (in terms of wavelength) of neighbouring spectral features {22}. Dufiach exploited this characteristic of the technique to investigate small spectral shifts due to changes of environment undergone by specific amino acid residues which resulted in the fourth order derivatives of cytochrome c displaying significant alterations in amplitude in addition to small shifts in wavelength of their maxima {46}. The interpretation of variable peak height was not feasible for fourth order derivative analysis in the current {Results & Discussion} 6 4 Fig. 7: Fourth-order finite difference spectra of a-absorption bands from soluble reference cytochromes. Ambient temperature (295 K) reduced minus oxidized difference spectra of soluble reference cytochromes in 100 mM potassium phosphate buffer at pH 7.0, low temperature (77 K) reduced minus oxidized difference spectra of soluble reference cytochromes in 100 mM potassium phosphate buffer at pH 7.0 containing 1.0 M sucrose, and fourth-order derivatives of these spectra. Cytchrome c refers to equine heart cytochrome c, cytochrome 65 refers to the soluble tryptic fragment of bovine hepatic microsomal cytochrome 65. a. Mixed cytochromes c, 25.0 u.g m l / 1 , and b5, 20.0 u.g mL" 1; 295 K ; AA = 0.025. b . Fourth-order finite difference spectrum calculated from a-band of curve 'a'. C . Mixed cytochromes c, 6.75 u.g mL" 1 , and b5, 10.0 u.g mL" 1; 77 K ; AA = 0.050. d . Fourth-order finite difference spectrum calculated from a-band of curve 'C'. {Results & Discussion} 6 5 {Results & Discussion} 6 6 F i g . 8: Absolute ultra-violet absorption spectrum and fourth-order finite difference spectra of soluble reference ferrocytochrome c. Equine heart cytochrome c at 250 u-gmL,-1 in 100 mM potassium phosphate buffer, pH 7.1, reduced by the procedure of Duflach et al. {46}. a. Absolute absorption spectrum, AA = 0.100 . b . Fourth-order finite difference spectrum calculated from curve 'a' by four successive first-order derivatizations, A A = 0.100 . C. Fourth-order finite difference spectrum calculated from curve 'a' by two successive second-order derivatizations, A A = 0.020 . {Results & Discussion} 6 7 {Results & Discussion} 6 8 study because of the unknown number and nature of cytochromes present in the biological materials being investigated. Nevertheless, determining the wavelengths of maximal intensity of fourth order derivatives from low temperature redox difference spectra greatly facilitated the identification of individual components contributing to partially resolved a-bands in those spectra. Also relevant to these analyses was the requirement of optimizing the signaknoise ratio which degrades with each derivatization. This was achieved by means of moving average procedures accompanied by compensation for associated wavelength shifts. Two firmware procedures were provided for calculating the fourth order derivative spectra, one incorporating more substantial smoothing than the other, as illustrated in Figure 8. Although moving-average smoothing of spectra might be expected to lower the sensitivity of the technique the reproducibility of fine detail in derivative spectra obtained from diverse experimental samples over extended periods of time confirmed the validity of these analyses. Generally the option providing less noise has been illustrated for clarity, exceptions being noted in the figure legends. The possibility that the results might be influenced by instrument-induced spectral abberations was eliminated by the observation of similar spectral features in samples analysed by both types of high-resolution spectrophotometer that were available : DW2c and PE-356 (data not shown). Cautions relating to limitations of the technique and overinterpretation of fourth order derivative data have been presented by other investigators {90, 22}. The degree of confidence which can be placed in a single fourth order derivative trace is highly dependent upon the amount of noise associated with both the original signal and the final trace. For this reason most of the spectra illustrated in the manuscript are derived from summed and averaged scans and shown in the damped form, in which the salient features are readily apparent. Where less obvious features are to be discussed, they are demonstrated with representative traces plotted at higher sensitivity, the interpretation being based upon analyses of independent spectra collected in multiple experiments. When similar minor features have appeared reproducibly in samples prepared from different sources and treatments they have ultimately been recognised as significant. The repetitive nature of such features was a requirement for them to be distinguished from random noise and to be regarded as significant. As has been noted by others, with some experience fourth order derivative bands can usually be observed as peaks or small shoulders in the absorption spectrum {22}. That discontinuities in the original difference spectra causing minor fourth order derivative features as described above should be distinguishable in results from both types of high resolution, low temperature spectrophotometer supports this contention (data not shown). In sum, with the difficulty of determining from the parent {Results & Discussion} 6 9 spectra the precise wavelengths of those components causing such spectral features the advice of Butler & Hopkins is pertinent to spectral analysis by fourth order finite difference spectroscopy {22} : (1) "It is apparent that higher derivatives can extract information from spectral data that is not readily apparent in the original curves." (2) "Whenever possible additional evidence should be sought to confirm the validity of higher derivative bands." Appropriate optical and mechanical adjustments enabled reproducibility and accuracy of routine DW2c spectral wavelength determinations to be maintained within 0.5 nm and version 2.02 of the Midan n firmware enabled fourth order finite difference analyses to conform to these criteria. It was from the combined results of these spectrophotometric techniques, yielding mutually compatible data and calibrated against values published in the literature for the various standard cytochromes, that confidence could be expressed in the refined procedures that were instrumental in producing the highly detailed spectral data that were obtained from experimental samples. Thus overlapping spectral features could be distinguished and values estimated for their individual absorption maxima. (U) Potentiometric Titrations (a) Standard titrations and electrode calibration Spectroelectrochemical (potentiometric) titrations of standard solutions of equine heart cytochrome c were used to develop modifications and improvements to the procedures of Hackett {66, 69} as described under Materials & Methods and elsewhere {214}. In addition these standard titrations were used to calibrate the combination platinum electrodes incorporated into the apparatus. Data from typical standard titrations are shown in Figures 9 and 10, in each case a theoretical one-component curve being fitted to the data. Figure 9 illustrates the result of the BMD.P3R non-linear regression analysis of the original data and Figure 10 the corresponding Nernst plots with associated linear regression analyses. Figures 9b and 10b show the result of titrating the standard cytochrome c in the presence of 1.0 M sucrose in preparation for poised potential low temperature spectrophotometric analysis (described under Materials & Methods) in which no significant deviation from standard behaviour could be detected. Routine standard titrations provided duplicate mid-point potential values for cytochrome c that were within a 5 mV range for a particular {Results & Discussion} 7 0 Fig. 9: Potentiometric titrations of standard cytochrome c in presence and absence of sucrose : direct plots. The curves illustrate theoretical values for single electron transfer, one component fits. 2.. Equine heart cytochrome c at 30.0 ug m L - 1 titrated in 100 mM sodium phosphate buffer, pH 7.0, b_. Equine heart cytochrome c at 30.0 ug mL"1 titrated in 100 mM sodium phosphate buffer, 1.0 M sucrose, pH 7.0. {Results & Discussion} 7 1 (a) E h (mV vs NHE) {Results & Discussion} 7 2 Fig. 10: Potentiometric titrations of standard cytochrome c in the presence and absence of sucrose : Nernst plots. Linear regression curves indicate least squares fit: 3_. Nernst plot of data in Fig. 9 'a' ( £ m = 260.0 mVvs. NHE, slope = 56.0 mV, n = 1.05), Jb_. Nernst plot of data in Fig. 9 'b' (Em= 260.5 mVvs. NHE, slope = 59.0 mV, n = 1.00). {Results & Discussion} 7 4 platinum combination electrode, individual electrodes providing values within lOmV of each other. All absolute potential values are given relative to the normal hydrogen electrode (mVvs. NHE). Potentiometric titration and electrochemical characterization of the components of combined solutions of cytochromes was more complex, Figure 11 providing an example in which the standard cytochromes c and b$ are resolved into separately reducible species, described by a theoretical two-component fit of the data in panel 'a' and by Nernst curves in panel 'b\ It can be seen from Figure lib that interpretation of multicomponent Nernst curves in order to derive mid-point potentials and proportions of total reducible cytochrome becomes progressively more complex with an increasing number of components. Figure lib also illustrates a theoretical linear regression analysis of the data assuming that the cytochrome 65 is fully oxidized and the cytochrome c fully reduced at the 'plateau' potential of +100 mV. Although in this case the values for mid-point potentials derived from the Nernst analyses agree with published values {112,176,193} the technique is impractical for samples with greater numbers of cytochrome components and for those in which the reduction profiles overlap. In these cases, as illustrated in Figure 12a+b, computerized non-linear regression analyses of the original absorbance data was required to provide mid-point potential values for each component and the magnitude of its contribution to the total cytochrome a-band redox absorption {69, 214}. (b) Comparison of experimental data with published membrane cytochrome values Figure 12 presents examples of potentiometric titrations of biological samples. Curve 12a was obtained from the titration of a sample from an homogenized suspension of washed membrane vesicles. A detergent solubilized preparation of nitrate reductase was partially purified from this membrane suspension and titrated to yield curve 12 b. Curve 12a illustrates a five-component best fit of data obtained from titrating cytochromes present in a washed membrane preparation of E. coli wild-type cells of strain RK4353 grown anaerobically with nitrate as terminal electron acceptor. Mid-point potentials and the proportion of total cytochrome corresponding to each component are provided in the figure legend. These data and the derived values correlate well with those from the literature for similar analyses on such preparations, as indicated in Table IV {66, 67, 69}. Analysis of the data from the solubilized nitrate reductase preparation revealed the presence of three cytochrome components: equal quantities of two major cytochromes (43 % of total AA each, Em= +34 mV and +118 mV) and a minor, high potential species (vide infra). Nernst plots derived from the data of Figure 12b are presented in Figure 13. Panel 13a illustrates the treatment of the complete set of redox data with the slope, derived from equation (4), of {Results & Discussion} 7 5 Fig. 11: Potentiometric titrations of soluble reference cytochromes. Cytchrome c refers to equine heart cytochrome c, cytochrome 65 refers to the soluble tryptic fragment of bovine hepatic microsomal cytochrome 05. a.. Cytochrome c at 30.0 u.g mL"1 and cytochrome 65 at 30.0u.gmL"1 titrated in 100 mM sodium phosphate buffer, pH 7.0, h_. Nernst plot of data in 'a' : O - - O, total cytochrome. • cytochrome c ; £ m = +257.0 mVvj. NHE, slope=63.5 mV, n=0.93. cytochrome^ ; EM= +3.0 mVvs. NHE, slope=62.0 mV, n=0.95. {Results & Discussion} 7 6 {Results & Discussion} 7 7 Fig. 12: Potentiometric titration of membrane cytochromes and nitrate reductase preparation from E. coli grown anaerobically in the presence of nitrate : direct plots. a . Membranes of E. coli strain RK4353 grown anaerobically in the presence of nitrate were resuspended in degassed 100 mM potassium phosphate buffer, pH 7.0 at a protein concentration of 10.0 mg mL"1. The theoretical curve for a five component fit is illustrated. The mid-point potentials of these components and their individual contributions to the total total type-6 cytochrome are indicated +249.5 mWvs. NHE : 5.5 % total a-band X m a x absorbance. +117.0 mVvs. NHE : 24.5% total a-band X m a x absorbance, + 27.0 mVvs. NHE : 23.0 % total a-band X m a x absorbance, - 41.0 rnVvj. NHE : 26.0 % total a-band absorbance, -134.0 mVvs. NHE : 20.5 % total a-band absorbance, b . Nitrate reductase was solubilized and partially purified from membranes of E. coli strain RK4353 grown anaerobically in the presence of nitrate as described in M a t e r i a l s & M e t h o d s . The preparation was titrated in a solution of 100 mM potassium phosphate buffer, 0.2 mM dithiothreitol, 0.1% (w/v) Triton X-100 at a protein concentration of 0.88 mg mL"1 and at pH 7.0. An arrow indicates the potential below which the cytochrome of highest potential is fully reduced. The theoretical curve for a three component fit is illustrated. The mid-point potentials of these components and their individual contributions to the total total type-6 cytochrome are indicated: +231.5 mWvs. NHE : 14.0% total a-band X m a x absorbance, +118.0 mVvs. NHE : 43.0% total a-band absorbance, + 34.0 mVvs. NHE : 43.0 % total a-band X_.„v absorbance. {Results & Discussion} 7 9 [2.303 RT/nf = 120 mV] yielding the improbable value of 0.5 for 'n', the number of electrons involved in the transfer (section I.ii.c). Since the redox mediators effect electrochemical equilibration within the titration system this result indicates the presence of two contiguously titrating single electron components, and on close inspection a discontinuity at near 100 mV divides the data into two linear sets of points. Redox absorption spectrophotometry detected a sole, symmetrical cytochrome a-band in the sample, this being of type-6 with an absorption maximum of 556.0 nm at 77 K (data not shown) {67}. As described above, non-linear regression analysis of the titration data, utilizing an algorithm assuming single electron transfer, had provided theoretical values for two cytochromes of equal intensity with mid-point potentials approximately 90 mV apart: a separation that would provide a virtually contiguous titration curve. By choosing the half-reduction point of this curve, and treating each half independently the Nernst plots charted in Figure 13b were obtained. The value of 'n' approximates to unity for each of these curves confirming that results from these two methods of analysis were equivalent and suggesting that the solubilized nitrate reductase preparation contained two major cytochrome components of equal magnitude. This concept was originally proposed by Hackett & Bragg {67} as an interpretation of potentiometric titration data, for no spectroscopic, chromatographic or genetic evidence for more than a single form of cytochrome b™ had been found {14,67, 69}. Chaudhry & MacGregor erroneously challenged this view on the basis of spectrophotometry alone, actually confirming the proposition in the earlier publication {28}. The current study supports all the evidence put forward to date, suggesting that the nitrate reductase cytochrome b™ appears as a single type-6 cytochrome with a single absorption maximum of 556 run in reduced minus oxidized spectrophotometry at 77 K but that it may be resolved into two distinct cytochrome species by potentiometric analyses. More recently cytochromes C4 from several sources have been studied by spectroelectro-chemical redox titration and demonstrated a range of spectral and potentiometric responses with biphasic titration curves and either split or single a-bands (measured at ambient temperature) {ill}. The dihaem cytochromes C4 are found in Pseudomonads and other procaryotes where they appear to have a respiratory role in conjunction with cytochrome o. The authors suggest models in which the two haems may either function with individually defined mid-point potentials or in which they are equivalent in the fully oxidized state, the addition of a first electron occurring at either haem and causing a conformational modification such that the addition of the second electron is a less favourable process and takes place at a lower redox potential {111}. The spectral similarity of the two components has been attributed in each case to the dihaem cytochrome C4 having undergone gene duplication, as supported by amino acid analyses, each of the two resulting domains having similar {Results & Discussion} 8 0 {A} Cytochrome 0 5 : (tryptic fragment of bovine hepatic cytochrome 6 5 : Current Study Reference (112) Reference f 1761 Eh (mVvs. NHE) +3.0 Eh (mVvs. NHE) ±0, +20 Eh (mVvs. NHE) +6 (Under equivalent conditions) {B_} Washed Membranes: nitrate-grown cells of strain RK4353: Current Study Reference (204) * h (mVvs. NHEt +249.5 +117.0 + 27.0 - 41.0 -134.0 cytochrome b (% total) 5.5% 24.5 % 23.0 % 26.0 % 20.5 % (mVvs. NHE) +149.5 + 59.0 + 26.0 - 109.0 cytochrome b (% total) 29.0 % 15.0% 18.0% 15.0% {£} Solubilized. partially purified nitrate reductase preparation: Current Study (mVvs. NHE) +231.5 +118.0 + 34.0 cytochrome b (% total) 14.0% 43.0 % 43.0 % Reference 169) (mVvs. NHE) + 122.0 + 17.0 cytochrome b (% total) 50% 50% Table IV: Comparison of potentiometric analyses of biological samples with previously published values. Electrodes were calibrated by titration of standard solutions of equine heart cytochrome c, as described under M a t e r i a l s & M e t h o d s . Cell type, growth conditions and the number of components resolved will affect determinations of membrane cytochrome complement {B} although these factors are minimized in analyses of solubilized cytochrome preparations subjected to purification techniques {C}. {Results & Discussion} 8 1 g. 13: Potentiometric titration of partially purified E. coli nitrate reductase: Nernst plots. 2.. Nernst plot of data from Fig. 13 'b'. An arrow indicates the potential below which the cytochrome of highest potential is fully reduced. O—O ; total cytochrome : apparent £ m = +95 mVvs. NHE; slope = 120 mV ; n = 0.5 . b_. Nernst plots of data at potentials below discontinuity indicated by the arrow in Fig. 13 'b': • - • ; EM= +120.0 mVvs. NHE ; slope = 52.5 mV ; n=1.12, • - • ; EM= + 28.0 mVvs. NHE ; slope = 48.5 mV ; n = 1.22 . {Results & Discussion} {Results & Discussion} 8 3 haem environments. Further validation and calibration of the improved titration and curve fitting techniques was obtained by comparing mid-point potentials of cytochromes in membrane preparations from several bacterial strains with published, values, including those for the temperature sensitive polar chlC mutant TS9A at permissive and at non-permissive temperatures, governing formation of the chll gene product: a type-fe cytochrome implicated in nitrate reduction (data not shown) {39, 66, 67, 204}. Nevertheless, an empirically determined resolution limit of 50 mV between mid-point potentials was observed when using the BMD.P3R non-linear regression analyses : although this limited the number of possible components that could be fitted to an experimental curve, a practical threshold of 10 % of total sample cytochrome was used as a minimum significant quantity. (c) Electrochemicallv poised high resolution spectrophotometric analysis The coupling of analytical techniques promised to provide a correlation between the limited information available when each procedure was used in isolation. Such an approach has been attempted for some years in this laboratory, coupling high-resolution redox spectrophotometry with dual wavelength kinetic studies or with potentiometric poising and the latter combination has also been used in similar investigations by R. B. Gennis {69,117, 214}. Withdrawal of aliquots from a sample undergoing potentiometric titration and the subsequent manipulation and freezing of these aliquots under strictiy anaerobic conditions in order that their electrochemical potential remained constant presented significant mechanical and electrochemical problems. It has already been shown that there was no effect upon the potentiometric titration itself when the 1.0 M sucrose necessary for optimizing the 77 K redox spectra {188,206} was included in the titration buffer (Fig. 9+10). Figure 14 demonstrates that the refined procedure was capable of producing high resolution low temperature redox difference spectra that reflected the progression of the potentiometric titration from which the samples were taken. Alteration of the slope of the baseline as the potential of the sample was modified was due to differential redox absorbance by the electrochemical mediators in the titration buffer. Comparison of the proportion of cytochrome reduced as indicated by the poised potential low temperature redox difference spectra and those redox difference spectra collected at ambient temperature in the process of performing the titration shows clearly that the redox behaviour of the samples providing these spectra at the two temperatures was different (Figure 15). Curve 15a is the theoretical curve for a single component, single electron transfer reaction at the temperature of the {Results & Discussion} 8 4 Fig. 14: Potentiometric titrations of standard cytochrome c : low temperature poising and spectrophotometry at 77 K. The technique was performed on a sample of equine heart cytochrome c at 30.0 ug mL" 1 titrated in 100 mM sodium phosphate buffer, 1.0 M sucrose, pH 7.0 as described in Materials & Methods. The a-bands of the cytochrome c reduced minus oxidized difference spectra measured with the PE-356 at 77 K are shown after sampling and poising at the following potentials: a . +254.0 mVvs. NHE, b . +237.0 mVvs. NHE, C . +233.0 mVvs. NHE, d . +225.5 mVvs. NHE, e . +212.5 mVvs. NHE, f . +207.5 mVvs. NHE, g . +198.0 mVvs. NHE. Equivalent reduction estimates were obtained from absorbance changes measured at 550.5 nm (X^^ or at 548.0 nm, the wavelength of the a-band 'shoulder'. {Results & Discussion} {Results & Discussion} 8 6 Fig. 15: Potentiometric titrations of standard cytochrome c : effects of poising at low temperature. Nernst plots of data collected at 296 K and 77 K ; each data point is the average of two or more measurements. a . 0--0 ; Equine heart cytochrome c at 30.0 ug mL" 1 titrated and analysed in 100 mM sodium phosphate buffer, 1.0 M sucrose, pH 7.0 at 296 K. The curve illustrates theoretical values for single electron transfer, one component fit. (EM= +260.0 mVvs. NHE; slope = 59.0 mV ; n=1.00). b . Recalculation of theoretical curve 'a' using a temperature coefficient of T= 270 K (freezing point of sample) and arbitrarily assigned E M = +260.0 mVw. NHE. (Slope = 15.3 mV ; n=1.00). C . • — • ; Equine heart cytochrome c at 30.0 ug mL" 1 titrated in 100 mM sodium phosphate buffer, 1.0 M sucrose, pH 7.0 at 296 K and poised samples analysed spectrophotometrically at 77 K. The linear regression curve is illustrated and indicates theoretical values of E M = +213.0 mVvs. NHE; slope = 42.6 mV. The value of 'n' corresponds to unity at a temperature of 215.5 K {-57.5°C}. d . Recalculation of theoretical curve 'c' using a temperature coefficient of T= 77 K and arbitrarily assigned E M = +213.0 mVvs. NHE. (Slope = 15.3 mV ; n=1.00). {Results & Discussion) 8 7 {Results & Discussion} 8 8 titration and it correlates with reduction values obtained from the ambient temperature reduced minus oxidized difference spectra collected during titration. The intercept has been positioned at 260 mV, the value obtained by linear regression analysis of these data. Curve 15c demonstrates similar data obtained from high resolution redox difference spectrophotometry at 77 K of samples poised at specific potentials during the titration. Theoretical curves corresponding to single component, single electron transfer at 270 K (the freezing point of the buffered sucrose solution) and at 77 K (the temperature of spectral measurement) are illustrated as curves 15b and 15d respectively. The intercepts for curves 15b and 15d have been chosen arbitrarily since insufficient data is available to describe the behaviour of cytochrome mid-point potentials at freezing temperatures. Electron equilibration between half cells such as a redox couple and the normal hydrogen electrode may be described by equation (3), first published by Peters and generally refered to as the Nernst Equation {43,213,217}. RT [Aox] Eh = EQ + . In (3) X \ T [Ared] where is the redox potential difference between the sample and reference half cells at the ambient temperature with the subscript ' h' denoting the reference half cell to be the standard hydrogen half cell; EQ is the standard redox potential of redox couple A (the £ h at which [Aox] = [Ared ] with both factors maintained at unit activities at pH = 0); R is the gas constant (8.31 J K"1 mol"1); T is the ambient temperature in Kelvin; n is the number of electrons transfered in the reaction and T the Faraday constant (96 493 J V"1). Modifications of these standard conditions to those appropriate for biological samples, where activities and often concentrations are unknown and pH values may have to be restricted to physiological ranges, are represented by a similar equation, (4), in which E0 is replaced by the mid-point potential at the ambient pH of the determination, Em {43,45,213}. RT [Aox] Eh = Em + 2.303 . Iog10 (4) [Ared] In the current study all potentiometric titrations were buffered with 100 mM potassium phosphate buffer at pH 7.0 and £j, was derived from the potential measured at the calibrated {Results & Discussion} 8 9 platinum combination electrode as each spectrum was recorded during the course of the titration. The values of R and T are constant and that of T is governed by the conditions of the titration and poising procedures. It is assumed that for type-fc cytochromes n = 1 since the haem iron alternates between Fe11 and F e m redox states {43,45}. From the equations it is apparent that the the Em value of the molecule under study will dictate the ordinate intercepts of the Nernst curves in Figure 15 — the mid-point potential being a characteristic molecular property under a defined set of applied conditions. The apparent mid-point potential exhibited by those samples which were rapidly cooled and then analysed at liquid nitrogen temperatures was 47 mV lower than that derived from those samples measured by the standard procedure at ambient temperature (curves 15 a+c). Consequently the redox phenomena being observed must have been more complex than simple poising of the low temperature samples at the potential of the system at which they were withdrawn. Although this phenomenon might have been caused by a small degree of oxidation during the transfer and freezing steps, and this might also explain the change in slope of the Nernst plot, the regularity of the data argues against such an arbitrary cause for either of these effects. Mid-point potential values for type-6 cytochromes are known to be particularly susceptible to environmental changes in terms of proton activity (response to pH), specific interactions (especially preferential ligand binding to one redox form), ionic strength, temperature and intramolecular modification {45, 156, 176}. In addition to the effects of the temperature drop the consequences of freezing upon the Em are unknown, although it has been reported that freezing phosphate buffered solutions may cause pH shifts as large as -3 pH units in sodium phosphate buffers and -1 pH unit in potassium phosphate buffers {149}. Acidification of methaemoglobin has been observed upon cooling to cryogenic temperatures, although in the current investigation the presence of 1.0 M sucrose in the titration buffer — added in order to optimise the signalmoise ratio and reproducibility of the high resolution spectra — would be expected to decrease pH changes and abberations due to the formation of large ice crystals {90, 94,149}. (It has been shown earlier, section I.ii.a, that the sucrose had no effect upon the results of standard titrations.) Factors relevant to the alteration of redox characteristics in the frozen samples may include localised concentration effects brought about by the physical phase change (this may cause preferential precipitation of one of the buffer salts) or disruption of the equilibrium with the redox mediators in the sample. The change in redox state of the cytochrome as the applied potential was modified was also different between the two sets of data, and is reflected in the difference of the slopes of Figure 15, curves a+c. As defined in equation (1) (see Introduction), if R, T and n are assumed to be {Results & Discussion} 9 0 constants and the spectrophotometric measurements are indicative of the proportion of sample reduced (as is supported by the linearity of the data) then a sample that was not truly 'poised' at the potential of the system during withdrawal and rapid freezing would be expected to reflect a dependence upon the temperature, T. Yet the low temperature samples did not behave in the manner expected of samples 'poised' at the temperature of the titration itself, at the temperature of the physical phase change of the sample (measured independently at -3°C), or at their temperature during analysis (77 K), but as if they were responding to a specific applied temperature of 215.5 K (-57.5°C). The consistency of the data with respect to this temperature suggests that while the factors described above may be responsible, these observations may also correspond to the energetic trapping of a mechanistic intermediate in the transfer of electrons between sample and mediators, the original equilibrium being modified upon cooling. In practical terms these results imply that one should not assume that low temperature spectra recorded from 'potentiometrically poised' samples are representative of their precise sampling potentials, as has previously been the case {69, 117) but that if a series of such poised potential, low temperature spectra is analysed it will provide a representative, high-resolution display of spectral phenomena occurring in sequence over a broad, albeit currently ill-defined range of electrochemical potentials. Thus the technique, performed with all the appropriate precautions, may be deemed as qualitatively valid, accompanied by the reservations stated above. These observations merit further investigation. Steady state poising has been employed in dual wavelength investigations of cytochrome kinetics in studies associated with the current and other investigations {42, 69, 95, 214). Not only may these reactions be carried out within the cryogenic sample holder with a minimum of manipulation and very rapidly frozen at the appropriate time but the stability of the steady states due to the presence of relatively large pools of oxidant or reductant create a system in which the gross electrochemical potential is electrochemically buffered, although trapping of particular intermediates reflecting specific intramolecular modifications may still be important phenomena at these low temperatures {25). A similar situation is described by Chance's investigations of mammalian cytochrome oxidase in which low temperatures were used to trap intermediates in the binding and release of oxygen and carbon monoxide within the 'oxygen pocket' of the oxidase {38}. {Results & Discussion} 9 1 HH 2 M E M B R A N E S T U D I E S {A} Aerobic Respiratory Tvpe-/> Cytochromes (i) Redox Difference Spectroscopy (a) Visible range spectrophotometry of E. coli membrane cytochromes Absolute reduced and absolute oxidized spectra of washed membranes from E. coli wild-type strain GR17N, grown aerobically on a defined medium to stationary phase, are shown in Figure 16; a+b. One function of recording redox difference spectra from biological samples was to eliminate background absorption from all material failing to undergo spectral changes with an alteration of oxidation state: the reduced minus oxidized difference spectrum from curves 16(a-b) is shown in curve 16c, all three spectra having been collected at 295 K (ambient temperature). As described in the Introduction stationary phase cells contain the components of both aerobic respiratory chains and Figure 16c is marked with the approximate locations of the standard absorption patterns caused by a-, p- and y-bands of the b- and d-type cytochromes. The absorption bands of cytochrome 6595, which contains a high-spin type-6 haem {118} are not marked and will be discussed in detail below; nevertheless a minor absorption band is visible at 595 nm in Figure 16c as is the Soret band shoulder at 440 nm which is associated with the cytochrome d complex containing cytochrome 6595. Mutants expressing a cytochrome d complex deficient in cytochrome 6595 have only recently been isolated {communication from R.B. Gennis} and at present it is still uncertain whether the Soret shoulder is caused by the cytochrome 6595 or by cytochrome d itself. This question is of importance in determining the identity of the terminal oxidase of the cytochrome d respiratory chain, since carbon monoxide binding to the terminal oxidases causes perturbations of the reduced spectrum as shown in Figure 17. Recent publications have affirmed that cytochrome d is the terminal oxidase and that it is responsible for the Soret shoulder and for spectral shifts upon CO binding {87,105,135}. Absolute reduced spectra, measured at ambient temperature, {Results & Discussion} 9 2 Fig. 16: Escherichia coli respiratory cytochromes : visible range spectra of cell membrane suspensions at ambient temperature. Wild-type (w+) strain GR17N grown on glucose to stationary phase: cells were grown to stationary phase on CYD minimal medium containing glucose; crude membranes were prepared from them and washed as described in Materials & Methods. These membranes were resuspended in 100 mM potassium phosphate buffer, pH 7.0 to a protein concentration of 5.0 mg mL"1. a. b. c. Dithionite reduced absolute spectrum obtained at ambient temperature, A A = 0.40, Hydrogen peroxide oxidized absolute spectrum at ambient temperature, A A = 0.40, Dithionite reduced minus peroxide oxidized difference spectrum obtained at 295 K. {Results & Discussion} 9 3 {Results & Discussion} 9 4 Fig. 17: Spectral identification of E. coli terminal oxidases following exposure of membranes to carbon monoxide. Cells were grown on CYD minimal medium containing glucose; crude membranes were prepared from them and washed as described in Materials & Methods. These membranes were resuspended in 100 mM potassium phosphate buffer, pH 7.0 to a protein concentration of 5.0 mg mL" 1. a. Dithionite reduced absolute spectrum obtained at ambient temperature, AA = 0.40. Strain GR17N grown on glucose to stationary phase. b . Dithionite reduced absolute spectrum obtained at ambient temperature, AA = 0.40. Strain GR17N (w+) grown on glucose to early exponential phase. C . Dithionite reduced plus carbon monoxide minus dithionite reduced difference spectrum obtained at 295 K. Strain GR17N (w+) grown on glucose to stationary phase. d . Dithionite reduced plus carbon monoxide minus dithionite reduced difference spectrum obtained at 295 K. Strain GR17N (vv+) grown on glucose to early exponential phase. {Results & Discussion} 9 5 (Results & Discussion) 9 6 were obtained from washed membranes of cells grown to stationary and early exponential phases (Fig. 17 a+b). Ambient temperature 'reduced plus carbon monoxide minus reduced difference spectra' ('carbon monoxide spectra') for each of these samples are illustrated in curves 17c and 17d respectively, demonstrating the alterations in the spectral features of these reduced samples. Carbon monoxide has been shown to bind to the free coordination site of several terminal oxidases, including the mitochondrial terminal oxidase cytochrome a.a-^ {38, 112} and it acts similarly in those procaryotic terminal oxidases that react with oxygen {87, 112, 216}. Thus the major absorbance troughs in the Soret region of Figure 17c+d indicate that in stationary phase cells carbon monoxide binds to terminal oxidases with Soret absorbances at 430 nm and 440 nm whereas in early exponential phase only the 430 nm oxidase is present. In addition curve 17c demonstrates the major effect of carbon monoxide upon the isolated a-band of cytochrome d with a shift from 623 nm to 645 nm. No spectral features of the cytochrome d complex are visible in absorption spectra of membranes from early exponential phase cells, curve 17d. Table V provides precise data for these various spectral parameters., High resolution, low temperature redox difference spectra of membranes from stationary and early exponential phase cells are shown in Figure 18, curves a and b respectively. These spectra provide the extra detail and sharpness of absorption bands noted earlier in those from the standard soluble cytochromes (Fig. 5, 7). Inclusion of 1.0 M sucrose in the low temperature sample buffer was found to increase the extinction coefficient of the total membrane cytochrome preparations by a factor of five (data not shown) as predicted and discussed by Jones & Poole {90}. As the cells progressed from exponential to stationary phase the Soret absorbance indicated that the total cytochrome complement per unit of membrane protein doubled. There was also an increase from 4.0 to 6.4 in relative maximal absorbance between the type-6 Soret and a-bands suggesting that cytochromes with distinct Soret bands and superimposed a-bands had been replaced with those exhibiting the reverse characteristics. Examination of the a-band regions of curves 18a and 18b reveals that there was a complete absence of cytochrome d in the early exponential phase cell membranes (and, by inference, the other components of the cytochrome d complex with less clearly defined spectra were also absent: cytochromes 6553 and 6595) whereas a pronounced type-6 a-band shoulder is visible at (560-565)nm, indicative of the cytochrome 0 respiratory chain. The characteristics of the cytochrome d complex described above were more distinct at 77 K than at ambient temperature (Fig. 18a and Fig. 16c). Moreover the stationary phase membranes displayed greater absorbance due to iron-sulphur protein and flavoprotein, as may be seen from the absorbance trough between 400 nm {Results & Discussionl 9 7 {A} Reduced minus oxidized difference spectra: cytochrome d Growth Phase Temperature cyt. b Vmax gyt.fr ^ m a x b ^ a m a x Aamax_;_Aa (K) (nm) (nm) (nm) (nm) (nm) Exponential 295 430 . 529 562 77 428 526+532 556, (563) — — Stationary 295 430, (440) 528 560 626 650 77 428, (438) 526 558 624-626 648 {B} Reduced plus carbon monoxide minus reduced difference spectra: Growth Phase Temperature oxidase A^ cytochrome o ^ m m cytochrome d (K) (nm) (run) (nm) Exponential 295 416 430 — Stationary 295 416 428-432 440 fC> Extinction coefficients at 295 K : Sample type Spectrum Wavelength Pair £ Reference Total membrane cyt. b Reduced minus oxidized 560/580 17.5 (104} Purified cyt. d complex Reduced minus oxidized 562/580 10.5 {104,135} (using cyt. b55S) Purified cyt. o complex Reduced minus oxidized 560/580 18.7 {97} Purified cyt. o complex Reduced+CO minus reduced 416/430 145.0 {97} Purified cyt. o complex Reduced+CO minus reduced 416/430 80.0 {129} Table V: Absorption maxima and extinction coefficients for respiratory cytochromes of cells grown to early exponential and stationary phases. Values in parentheses indicate incompletely resolved absorption peak 'shoulders'. (Results & Discussion! 9 8 Fig. 18: Escherichia coli respiratory cytochromes : high resolution visible range redox difference spectra of cell membrane suspensions. Cells were grown on CYD minimal medium containing glucose; crude membranes were prepared from them and washed as described in Materials & Methods. Membrane samples were resuspended in 100 mM potassium phosphate buffer, 1.0 M sucrose, pH 7.0 to a protein concentration of S.OmgmLr1. a. 77K[R-0]DS : Dithionite reduced minus peroxide oxidized difference spectrum obtained at 77 K. Strain GR17N (w+) grown on glucose to stationary phase, AA = 0.40. b. 77K[R-0]DS : Dithionite reduced minus peroxide oxidized difference spectrum obtained at 77 K. Strain GR17N (vv+) grown on glucose to early exponential phase, AA = 0.20. C. Fourth-order finite difference spectrum calculated from curve 'a' by four successive first-order derivatizations, AA = 0.08. d. Fourth-order finite difference spectrum calculated from curve 'b' by four successive first-order derivatizations, A A = 0.08. m o u l t s & Discussion! 9 9 a b c d 1 {Results & Discussionl 100 and 500 nm. In both exponential and stationary phase membrane preparations the 6-type cytochrome a-bands were asymmetrical and shown to contain at least five distinguishable features by their fourth order derivatives, curves 18c+d. Distinct features of the cytochrome P-bands are also visible in both redox spectra and their derivatives and at least nine fourth order derivative bands are observed in the Soret region. Because of the complexity and incomplete resolution of these P- and v-band derivatives they will not be considered further. Of particular interest was the lack of response of the fourth order derivative analysis to the complex a-bands of cytochrome d, suggesting that absorbance features of both the oxidized and the reduced forms (which overlap to some extent at approximately 630 nm) are neither Gaussian nor Laurentzian in nature but comprise absorption bands displaying a constant fourth order rate of change of absorbance with unit change in wavelength. Unless stated otherwise, further discussion of reduced minus oxidized difference spectra of cytochromes from membranes or solubilized preparations will be restricted to those collected at 77 K. (b) High resolution a-band cytochrome studies of E. coli membrane cytochromes 1* a-bands of cell membrane type-fr cytochromes Four type-fr a-band profiles from redox difference spectra of membranes prepared from wild-type cells grown to stationary, early exponential or late exponential phase are illustrated in Figure 19a, curves 1, 2 and 3, respectively. A progressive change of form of the a-band profile occurred as the cell population matured and their respiratory cytochrome complement altered. The stationary phase cell membranes contained two predominant type-fr cytochromes with absorption maxima at 556 nm and 559 nm, corresponding to a cytochrome £555 and to the cytochrome £553 of the cytochrome d complex respectively. Minor shoulders at 548 nm and 565 nm may also be observed in these spectra. The cause of the former has been suggested to be the broad p-band absorption of cytochrome 6595 while the latter is probably associated with the cytochrome o produced during exponential phase {97,118,129}. Late exponential phase cell membranes clearly contained a smaller proportion of cytochrome £553 and the spectral shoulder at 565 nm is slightly more pronounced than in spectra from stationary phase membranes, suggesting that significantly less cytochrome d complex was present and somewhat more of the cytochrome o complex. This situation was more extreme in the early exponential phase membranes where a broad cytochrome £555 band was a major spectral component and the 'red' shoulder associated with cytochrome o was large {Results & Discussion) 101 Fig. 19: Absorbance shifts and improved spectral resolution of type-fe membrane cytochrome a-bands at low temperatures. Cells were grown on CYD minimal medium containing glucose; crude membranes were prepared from them and washed as described in Materials & Methods. Membrane samples were resuspended in 100 mM potassium phosphate buffer, 1.0 M sucrose, pH 7.0 to a protein concentration of 5.0 mg mL"1. a. Dithionite reduced minus peroxide oxidized difference spectra obtained at 77 K. 1. Strain GR17N (w+) grown to stationary phase, AA = 0.20, 2. Strain GR17N (w+) grown to late exponential phase, AA = 0.20, 3. Strain GR17N (w+) grown to mid exponential phase, AA = 0.16, 4. Strain PLJ01 (cyd') grown to stationary phase, AA = 0.16. b_. Fourth-order finite difference spectra calculated from corresponding curves 'a' by four successive first-order derivatizations. 1. Strain GR17N (w+) grown to stationary phase, 2. Strain GR17N (w+) grown to late exponential phase, 3. Strain GR17N (w+) grown to mid exponential phase, 4. Strain PLJ01 (cyd') grown to stationary phase. {Results & Discussion.} 10 2 {Results & Discussion) 103 enough to indicate the presence of three possible features and to obscure the contribution from any cytochrome 6553 that may have been present These features at the various growth stages correspond well to the postulated alterations in cytochrome complement of aerobically grown Escherichia coli cells during the maturation of the culture, as described in the Introduction. In Figure 19a curve 4 shows the a-band profile of membranes from the cyd' strain PLJ01 grown to stationary phase on the same defined medium as the other examples in Figure 19. This sample displayed a-band characteristics which closely resemble those of the exponential phase wild-type cell membranes. Thus there was partial resolution of type-6 membrane cytochrome a-bands at 77 K and this was capable of indicating a variation of 6-cytochromes with the change of respiratory chain constituents expressed at different phases of aerobic growth. Further analyses of these spectra will be discussed after illustrating the types of phenomena that affected the cytochrome 6 a-band profiles and, by implication, the cytochrome b complement of the cell membranes. Figures 20 and 21 illustrate variations in respiratory cytochrome b complement under different growth conditions and between various strains of cells each grown under identical conditions. Panel 18a, curves 3 and 4 show that wild type strain MR43L grown to stationary phase on defined media with succinate as sole carbon/energy source generated b cytochrome a-bands indicating a greater proportion of cytochrome 6555 to 6553 under conditions of high aeration than when grown in poorly aerated medium. The proportion of cytochrome d a-band absorbance to maximal cytochrome b a-band absorbance reflected the increased proportion of cytochrome d complex to total cytochrome b content indicated by the partial resolution of the cytochrome 6553 a-band (data not shown). Earlier studies have presented evidence to suggest that the cause of the cytochrome variations with growth phase (Fig. 19) is the reduction in the dissolved oxygen concentration of the liquid medium during the later phases of growth {87, 172, 204). This would provide an explanation for the similarity of cytochrome b a-bands when cells were grown upon a defined medium with glucose as sole carbon/energy source under either high or low aeration conditions (Fig.20a;l+2), since the greater rapidity of growth on CYD medium supplemented with glucose rather than succinate would enable cultures of moderate to high population density to deplete the oxygen supplied by either aeration rate (M-gic= 1.5 ; Hsuc~ 0-5) {172}. It is noteworthy that the total yield of type-6 cytochrome per mg of membrane protein was approximately equal under the two sets of aerated conditions although when grown to stationary phase the succinate-nourished cells produced 2.4 times the total quantity of type-6 cytochrome (measured at 556 nm) compared to those {Results & Discussion! 104 Fig. 20: High resolution a-band spectra : variation of wild-type cells' type-6 membrane cytochromes with aeration and carbon source. Cells of w+ strain GR17N were grown on CYD minimal medium supplemented with the carbon-energy sources stipulated below; crude membranes were prepared from each culture and washed as described in Materials & Methods. Membrane samples were resuspended in 100 mM potassium phosphate buffer, 1.0 M sucrose, pH 7.0 to a protein concentration of S.OmgmL"1. a. Aeration effects indicated by dithionite reduced minus peroxide oxidized difference spectra obtained at 77 K. 1. Strain MR43L (w+) grown to stationary phase on glucose under low aeration, AA= 0.10 , 2. Strain MR43L (w+) grown to stationary phase on glucose under high aeration, AA= 0.10 , 3. Strain MR43L (w+) grown to stationary phase on succinate under low aeration, AA= 0.24 , 4. Strain MR43L (w+) grown to stationary phase on succinate under high aeration, AA= 0.24 . b_. Substrate effects indicated by dithionite reduced minus peroxide oxidized difference spectra obtained at 77 K . 5. Strain GR17N grown to mid-exponential phase on glucose, AA= 0.10, 6. Strain GR17N grown to mid-exponential phase on succinate, AA= 0.12, 7. Strain GR17N (w+) grown to mid-exponential phase on lactate, AA= 0.40. {Results & Discussion) 10 5 {Results & Discussion) 106 grown on glucose. Figure 20b displays membrane cytochrome type-6 a-bands from cells of wild-type strain GR17N grown to mid-exponential phase under standardized conditions on defined media containing sole carbon/energy sources of glucose, succinate and lactate (curves 20b; 5,6 & 7 respectively). In all three spectra the pronounced 'red' shoulder is indicative of the presence of the cytochrome o complex in the membranes although both total and relative amounts of cytochrome o associated type-fr cytochrome and the b cytochromes absorbing at (555-556)nm vary with the carbon/energy source supplied. Growth on glucose stimulated the greatest production of cytochrome o associated cytochromes b in these mid-exponential phase cell membranes. Growth on lactate produced by far the lowest complement of total cytochrome b (one third of the concentration produced by growth on glucose). Growth to mid-exponential phase on succinate resulted in membranes with a total cytochrome content similar to that from membranes of cells grown on glucose although there was much greater absorbance in the (555-556)nm region of the a-band in the succinate-grown samples with much less absorbance corresponding to material associated with the cytochrome o complex. Figure 21 shows the variety of cytochrome b a-band profiles exhibited by membranes from a selection of wild-type cells grown to mid-exponential phase under standard conditions. The major feature of each spectrum is the maximal absorption at (555-556)nm although it is clear that there are a number of features contributing to the shape of the 'red' shoulder, and that the relative proportion of each of these features is variable. The partially resolved a-band spectra of type-fe cytochromes illustrated in Figures 19, 20 and 21 indicate that the complement of these cytochromes in cell membranes was dependent upon the cell strain, the carbon/energy source upon which the cells were grown, the abundance of dioxygen provided during aerobic growth and the phase of growth in which the cells were harvested. Although curve deconvolution techniques for quantitating these partially resolved bands were not available (203,204}, fourth order derivative analyses were used to aid component peak identification and will be discussed in relation to the phenomena described above. 2u Fourth order derivative spectra of cell membrane cytochrome b a-bands Curves 1 to 4 of Figure 19b depict the fourth order finite difference spectra corresponding to the redox difference spectra illustrated in panel 'a' of Figure 19. These fourth order derivatives illustrate that there was a progression in the change of membrane cytochrome a-band features as the harvesting-time of the cells is moved forward from stationary phase to early exponential phase. The (Results & Discussion! 107 Fig. 21: High resolution a-band spectra : type-b cytochrome content of membranes from a selection of wild-type cell strains. Cells of several w+ strains were grown to mid exponential phase on CYD minimal medium containing glucose; crude membranes were prepared from each culture and washed as described in Materials & Methods. Membrane samples were resuspended in 100 mM potassium phosphate buffer, 1.0 M sucrose, pH 7.0 to a protein concentration of S.OmgmL"1. a.. Dithionite reduced minus peroxide oxidized difference spectra obtained at 77 K. 1. Strain GR17N (w+) AA = 0.20, 2. Strain PA2-18 (w+) AA = 0.20, 3. Strain MR43L (w+) AA = 0.16, 4. Strain ML308-225 (w+) AA = 0.16. b_. Fourth-order finite difference spectra calculated from corresponding curves 'a' by four successive first-order derivatizations. 1. Strain GR17N (w+), 2. Strain PA2-18 3. Strain MR43L (w+), 4. Strain ML308-225 (w+). (Results & Discussion} 1 0 8 {Results & Discussionl 109 major features of the spectra from stationary phase cell membranes are absorption maxima at 556.0 nm and 559.0 nm, the latter corresponding to cytochrome 6553 of the cytochrome d complex (Fig. 19b;l). Curve 2 shows that late exponential phase cell membranes contain fr-type cytochromes with absorption maxima at 556.0 nm, (559-560)nm and 565.0 nm. There is an indication that there may have been other components present also, these less-well resolved constituents being illustrated more clearly by curves 3 & 4, representing spectra from membranes of early exponential phase cells and from membranes of cyd~ cells respectively. In these samples there are contributions to the type-ft cytochrome a-band profile from components with absorption maxima at 554.5 nm, (556-557.5)nm, 562.5 nm and 565.0 nm. Although the maxima of the cytochromes Z?556 and £553 (at 556.0 nm and 559.0 nm) associated with the cytochrome d complex are visible in the early exponential phase sample (Fig. 19b;3) they are clearly absent from membranes of the cyd' strain which is known to be incapable of expressing components of the cytochrome d complex. All the membranes appear to contain a cytochrome component with an absorption maximum at approximately 552 nm, although reference to the original spectra in panel 'a' shows that this makes only a minor contribution to the total cytochrome spectrum, and it may be caused by a secondary feature of one or more of the cytochromes discussed above. Table VI shows the absorbance maxima of these membrane preparations and compares them to previously published estimates. The type-fe cytochromes associated with each of the two aerobic respiratory pathways are distinct with the electron transport chain terminating in cytochrome o containing a larger number of -^cytochrome species. Many dehydrogenases incorporate type-Z? cytochromes and, depending on the growth conditions, these may be expressed in association with either of the aerobic respiratory chains {4, 87}. By growing cells on minimal media supplemented with different carbon/energy sources it was anticipated that these might be distinguished in the spectrophotometric and electrochemical analyses. Evidence will be provided in subsequent sections linking certain dehydrogenases with multiple cytochromes 6555. The identification of absorption peaks in the range 555 nm to 556 nm is therefore complicated and the significance of the participation of one or more cytochromes £55^ in either respiratory chain, not directly associated with dehydrogenase activity, was unclear from these studies. Partial purification of the cytochrome o complex and various inhibition studies have linked a cytochrome 6555 with the cytochrome 0 respiratory chain : experiments relevant to the conclusions drawn from these investigations will be described below {95, 117, 128, 143, 180}. The traces in panel 'b' of Figure 21 provide greater detail of the components associated with the cytochrome o complex having absorption maxima above 560 nm and were derived from (Results & Discussion! 110 Strain Growth Substrate/Phase a-Band Absorption Maxima GR17N (glucose-grown) Exponential 554.5 556.0 — — 560.5 562.0 565.0 Stationary — 556.0 557.5 559.0 — — — PLJ01 (glucose-grown) Exponential 554.5 556.0 — — 560.5 562.0 565.0 Stationary 554.5 556.0 — — 560.5 562.0 565.0 GR17N (succinate-grown) Exponential 554.5 556.0 — — 560.5 562.0 565.0 Stationary 554.5 556.0 557.5 559.0 560.5 562.0 565.0 PLJ01 (succinate-grown) Exponential 554.5 556.0 — — 560.5 562.0 565.0 Stationary 554.5 556.0 — — 560.5 562.0 565.0 Table VI: Wavelengths of absorption maxima of cytochrome b components in exponential and stationary phase cell membranes. Low temperature absorption maxima of 6-cytochrome a-absorption bands from reduced minus oxidized spectra of membranes from cells grown to exponential or stationary phase on D-glucose or disodium succinate. Wavelength values are expressed in nanometres and derived from fourth order derivative analyses of the original partially resolved high resolution a-band spectra (Fig. 19). (Results & Discussion! I l l membranes of four wild-type strains harvested in exponential phase. Although the 77 K redox difference spectra of these cells' membranes displayed markedly different profiles in the 'red' shoulder region of the a-band, between 560 nm and 565 nm (panel'a') the fourth order derivative spectra revealed that components with the same absorption maxima are creating the profile in each case. As illustrated in Figure 19b;3, the sensitivity of the fourth order derivative analysis resulted in contributions from the cytochrome 6553 associated with cytochrome d being observed even when it was present in very low concentrations in exponential phase wild-type cells (Fig. 21b). Further spectrophotometric analysis of this central region of the a-band required removal of interference by cytochrome 6553. A series of cyd' strains was developed for this purpose from strain GR19N which fails to express any of the components of the cytochrome d complex (vide infra ) {60, 61}. i . Multiple cytochrome '6550+' a-band components Demonstration by fourth order derivative analyses that the three 'cytochrome 6560+' cytochromes were present in each of the several strains portrayed in Figure 21, coupled with the observation that the original a-band profile between 560 nm and 565 nm was of radically different shape in each sample indicates that these a-band discontinuities cannot be the result of a single cytochrome with multiple absorption maxima. If the latter situation were the case, the intensity of the three observable features of the 'red' shoulder would be expected to have differed between samples in constant mutual proportion rather than with the spectral fluctuations that are observed in relative absorbance at 561.0 nm, 562.5 nm and 565.0 nm. Fluctuations of relative absorbance between the cytochromes '6560+' w e r e a l s o observed when single strains of cells were grown on a range of media supplemented with different carbon/energy sources (Fig. 20b). Since these three features always appeared together and their presence was associated with those conditions known to induce cytochrome o, and as they were observed through high resolution analysis of the 'cytochrome 6553' component of the type-6 cytochrome a-band spectrum it is proposed that that these three components are associated with the cytochrome 0 complex itself. It is possible that they may be the result of a single cytochrome being exposed to three distinct environments within the respiratory system, the proportions of cytochrome in each environment, or the availability of each environment, varying in each of the strains illustrated. Such environments might be functional in nature or could be the result of perturbation during sample preparation. Additional evidence of the linkage between these features of the 6-cytochrome a-band observed between 560 nm and 565 nm is provided by the action of certain respiratory inhibitors known to interact with terminal oxidases. Figure 22b, curve 5 demonstrates that addition of {Results & Discussion! 112 cyanide to a dithionite reduced washed membrane preparation abolished the entire red shoulder of the a-band spectrum observed in absolute spectra, with a shift of the a-band absorption maximum to 557 nm. Carbon monoxide, however, had a minimal effect upon the -^cytochrome a-band region of membrane suspensions reduced with sodium dithionite suggesting that the cytochrome or cytochromes contributing to the cytochrome b^Q+ shoulder may not correspond to cytochrome o itself, but rather to other components of the cytochrome o complex (Fig. 22b; 2+3). Carbon monoxide is known to interact with both of the aerobic terminal oxidases of E. coli {157, 216} and to perturb the Soret region of the absorbance spectrum as shown in Figure 17. The other respiratory inhibitors used in these experiments are also known to interact with the aerobic terminal oxidases through ligand interaction at the free sixth coordination site of the oxidases' haem groups which is reserved as the site of reaction with dioxygen {87, 112, 157]. Haems of other respiratory cytochromes have all six iron coordination sites occupied by linkage to amino acids of the apoprotein backbone as described in detail in the Introduction. Thus perturbation of reduced and/or oxidized cytochrome absorption spectra may be caused by direct interaction with a terminal oxidase, as is generally assumed to be the case with carbon monoxide {216), or by an allosteric effect of such an interaction as may be the explanation for the observed influence of cyanide upon those components of the cytochrome o complex with reduced minus oxidized a-bands in the 560 nm to 565 nm spectral range. A pertinent comparison is provided by the reaction of stationary-phase wild-type membrane suspensions to respiratory inhibitors acting upon the cytochrome d complex (Table VII). In these membranes the cytochrome o complex was largely absent, as shown by the lack of material absorbing between 560 nm and 565 nm in the low temperature dithionite reduced absolute spectrum of Figure 22a; 5. Upon treating such a reduced sample with carbon monoxide, cytochrome d itself responded with a pronounced red shift of its a-absorbance band from 626 nm to 633 nm, although the cytochrome 6553 component of the complex showed no apparent change in absorption characteristics, its low temperature absorption maximum remaining at 559 nm (Figure 22a; 4+5). On the other hand the actions of azide and cyanide resulted in the partially resolved a-band peaks of the type-fr cytochromes moving together to form a sharp, overlapping peak with an absorption maximum between 557 nm and 558 nm, and although each of these inhibitors caused a blue shift of the cytochrome d a-band cyanide virtually abolished the dithionite reduced a-band absorbance spectrum of cytochrome d whereas azide had little effect on its intensity (Figure 22a; 5,6,7). At present it is unknown whether these inhibitors have more than one site at which they may act upon the cytochrome d complex although azide has been shown to have multiple effects upon respiration and oxidative phosphorylation, each with a characteristic kj value {103). High resolution a-band (Results & Discussion) 113 Fig. 22: High resolution a-band spectra : absolute reduced spectra of washed, resuspended membranes in the presence of respiratory inhibitors. Cells of w+ strain GR17N and its cyd' derivative GR19N were grown to stationary phase on M9K minimal medium containing DL-lactate; crude membranes were prepared from each culture and washed as described in Materials & Methods. Membrane samples were resuspended in 100 mM potassium phosphate buffer, pH 7.0 to a protein concentration of 2.9 mg mL"1 (GR17N) or 5.2 mg mL"1 (GR19N). Samples were reduced with a few crystals of Na2S2C"4 before addition of the respiratory inhibitor. The latter was added as a few crystals of KN0 3 , NaNG ,^ NaN3 or NaCN or as CO bubbled gently through the reduced sample for 90 seconds. The CO-treated samples were then loaded into the cryogenic cuvette and frozen under a stream of CO. All sample preparation was carried out under dark conditions and absolute spectra were recorded at 77K as the average of three scans. Subsequent illumination of the CO-treated samples was performed by exposing the cry genie cuvette to the light (and warmth) from a 150 W incandescent bulb at a distance of 5 cm for 180 s intervals before refreezing and rescanning. The glutathione treatment illustrated was by addition of a few crystals to a freshly prepared membrane suspension which was subsequently reduced with Na2S204-Dithionite reduced controls ± NaCl crystals were used to ensure that dissolving and mixing inhibitors into the samples did not cause sample oxidation by introducing dioxygen. a.. Cytochrome content of membranes of strain GR17N, AA = 0.20 : 1. dithionite reduced + nitrite, 2. dithionite reduced + nitrate, 3. dithionite reduced + carbon monoxide, illuminated for a total of 540 s, 4. dithionite reduced + carbon monoxide, unilluminated, 5. dithionite reduced, 6. dithionite reduced + azide, 7. dithionite reduced + cyanide, 8. endogenous + glutathione, then dithionite reduced, 9. endogenous, (oxidized, oxygenated). b_. Cytochrome content of membranes of strain GR19N, AA = 0.20 : 1. dithionite reduced + carbon monoxide, illuminated for a total of 180 s, 2. dithionite reduced + carbon monoxide, unilluminated, 3. dithionite reduced, 4. dithionite reduced + azide, 5. dithionite reduced + cyanide, 6. endogenous, (oxidized, oxygenated). {Results & Discussion} 11 4 {Results & Discussion! 115 Strain Treatment cytochrome b X* cytochrome b Xa (nm) (nm) cytochrome d Xa (nm) OR17N dithionite reduced 426 556.5 (558) 628 dithionite + nitrite 426 556.0 (558) 640 dithionite + nitrate 426 556.0 (558) (635+642) dithionite + carbon monoxide 425 556.0 (557) 636 dithionite + CO + hva 424 556.0 (558) 635 dithionite + azide 424 556.5 (560) 622 dithionite + cyanide 425 557.0 — 624 glutathione, then reduced 426 555.0 (559) (636+642) fJR19N dithionite reduced 426 555.0 (560+) — dithionite + cabon monoxide 425 555.0 (560+) — dithionite + CO + hvb 425 555.0 (560+) — dithionite + azide 424 556.0 (565) — dithionite + cyanide 425 557.0 — a Illumination for 540 seconds as described in Figure 22. b Illumination for 180 seconds as described in Figure 22. Table VII: Wavelengths of absorption maxima of cytochrome b components in membranes exposed to terminal oxidase inhibitors. Low temperature absorption maxima of reduced minus oxidized spectra of ^ -cytochromes in membranes from cells grown to stationary phase on D-glucose. Wavelengths are expressed in nanometres, values in parentheses indicate absorption peak 'shoulders'. (Results & Discussion} 116 studies may prove a fruitful source of information regarding functions of specific terminal oxidase components in addition to the generally accepted ambient temperature spectrophotometric analyses of the overall Soret-band absorption characteristics of each complex. Detailed descriptions of the effects of respiratory inhibitors upon the terminal oxidases of relevance to this study are provided below : sections Il.C.i and n.C.ii describe membrane studies of cytochromes d and o respectively and sections m.C.v+vii relate to investigations of solubilized and purified cytochrome o preparations. (cj Use of membranes prepared from cyd strains of E. coli A number of benefits of using cyd" strains have been described and illustrated above. These strains, unable to express the components of the cytochrome d complex under the conditions of growth promoted resolution of the overlapping a-bands of 77 K reduced minus oxidized difference spectra because they lack a cytochrome 6553 (Fig. 18, 22). In addition spectral analysis following interaction of cytochromes with carbon monoxide or other respiratory inhibitors was simplified for there was no interference with the absorption spectra of cytochrome o in the Soret region by cytochrome d (Fig. 17, 22). Moreover, as detailed above, these strains provided a stable model of cell membranes containing a cytochrome complement usually observed under early exponential phase growth conditions and they enabled large quantities of cells and membranes to be prepared in this state. Not only were the large quantities of cells required for the physical studies of cytochromes described in this study difficult to prepare as coordinated, adequately aerated early exponential phase cultures, but delays inherent in the cooling and harvesting procedures inevitably resulted in some anaerobiosis of the samples and consequent initiation of induction of the cytochrome d respiratory pathway by wild-type cells. Consequently the cydA strains provided by the laboratory of R. B. Gennis during the course of this study and cyd' derivatives generated from these have been used throughout these investigations both independently and in comparison with their wild-type parental strains. The investigation of mechanisms controlling the expression of the components of the cytochrome d terminal oxidase will be discussed in relation to other cytochrome d results, in section Il.C.i.f. (Results & Discussion! 117 (r|) Analysis of mutant strains with an altered tvpe-6 cytochrome complement Additional attempts to improve resolution and characterization of type-6 cytochrome a-band characteristics utilised membranes from a series of putative cy6 mutants which had been generated as described under Materials & Methods (Fig. 23). Each strain illustrated contained an aberrant complement of type-6 cytochromes. It is unclear from the reduced minus oxidized difference spectra alone whether these mutants contained one or more altered cytochromes or if the spectral modifications were caused by altered proportions of cytochromes normally present in such membranes. Table VIII lists the spectral properties of a number of such mutants which represent classes of strains which contained similar cytochrome profdes : precise wavelengths were determined with the aid of fourth order derivative analyses. A selection of representative mutant strains were subjected to further analysis. Besides the potentiometric and kinetic studies which are described in subsequent sections, spectrophotometric analyses suggested that certain of the selected mutants, e.g. KW107, may have been deficient in a cytochrome component with an a-band absorption maximum at 556.0 nm and others, e.g. most of the 'KW400' series mutants, appeared to be lacking a component absorbing at 554.5 nm. These mutants were selected on the basis of their inability to utilize the cytochrome o respiratory pathway as described under Materials & Methods. Certain of them are described in sections II.A.ii.e and Il.C.ii.b+c with respect to cytochrome interactions with carbon monoxide. The spectra portrayed in Figure 23 indicate the difficulties encountered with interpretation of the mutants' a-bands, for the blue shoulder observed at 554 nm in membrane suspensions of KW107 could have been caused by the over-production of a component absorbing light at this specific wavelength or by the lack of material absorbing strongly at 556 nm in control samples. Nevertheless, high resolution spectral analyses of the putative cy6" mutant preparations reinforced the earlier categorization and tentative spectral identification of type-6 cytochrome components of the aerobic respiratory pathways on the basis of their absorption maxima, for specific components were consistently absent from the fourth order derivative analyses of redox a-band spectra of membrane suspensions from many of these strains (Table VIII). Indications that some of the cytochrome 6550+ features could be found in specific mutant strains without the appearance of other 6560+ features was not taken as evidence that each of these spectral characteristics was caused by a separate component An equally feasible result of an induced mutation would be the prevention of the formation of one or more specific molecular environments affecting distinct forms or states of a single cytochrome responsible for generating all of these related features. These strains provide a {Results & Discussion! 118 Fig. 23: High resolution redox spectroscopy of membrane preparations from putative cyb~ mutants. Cells of several mutant strains were grown to mid exponential phase on CYD minimal medium containing succinate; crude membranes were prepared from each culture and washed as described in Materials & Methods. Membrane samples were resuspended in 100 mM potassium phosphate buffer, 1.0 M sucrose, pH 7.0 to a protein concentration of lOmgmL"1. The cyd-notation refers to phenotypic character under these culture conditions. a.. Dithionite reduced minus peroxide oxidized difference spectra obtained at 77 K. 1. Strain PLJ01 -2. Strain KW102 3. Strain KW105 4. Strain KW107 5. Strain KW183 (cyct) AA = 0.150 AA = 0.150 AA = 0.120 AA = 0.120 AA = 0.125 1'. fourth order finite difference spectrum. 2'. fourth order finite difference spectrum. 3'. fourth order finite difference spectrum. 4'. fourth order finite difference spectrum. 5'. fourth order finite difference spectrum. b_. Dithionite reduced minus peroxide oxidized difference spectra obtained at 77 K. 1. Strain PLJ04 2. Strain KW401 3. Strain KW407 4. Strain KW417 5. Strain KW420 (cyd+) AA = 0.250 AA = 0.012 A A = 0.001 AA = 0.125 AA = 0.250 1'. fourth order finite difference spectrum. 2'. fourth order finite difference spectrum. 3'. fourth order finite difference spectrum. 4'. fourth order finite difference spectrum. 5'. fourth order finite difference spectrum. (Results & Discussion} 11 9 (Results & Discussion) 120 Strain C/E source cyt. d cytochrome b a-band absorption maxima GR17N D-glucose ++ — 556.0 557.5 559.0 — — — PLJ01 D-glucose — 554.5 556.0 — — 560.5 562.0 565.0 KW102 D-glucose ++ 556.0 557.5 559.0 trace KW105 D-glucose ++ — 556.0 557.5 559.0 trace — — KW107 D-glucose + trace — 557.5 trace — 562.0 565.0 KW110 D-glucose ++ — 556.0 — 559.0 trace — — KW131 D-glucose ++ — 556.0 557.5 trace — 562.0 — KW144 D-glucose ++ — 556.0 — 559.0 560.5 trace — KW183 D-glucose — 554.5 556.0 557.5 — — 562.0 565.0 KW201 D-glucose ++ 556.0 559.0 565.0 KW203 D-glucose ++ trace 556.0 — — 560.5 562.0 565.0 KW205 D-glucose ++ trace 556.0 — trace 560.5 562.0 565.0 KW401 D-glucose 554.5 556.0 557.5 560.5 562.0 565.0 KW407 D-glucose ++ 554.5 trace — 559.0 — trace — KW417 D-glucose — 554.5 — 557.5 — — 562.0 565.0 KW420 D-glucose ++ 554.5 — 557.5 trace 560.5 trace 565.0 KW424 D-glucose ++ 554.5 556.0 trace — 560.5 trace 565.0 KW425 D-glucose ++ 554.5 — 557.5 — 560.5 — 565.0 PLJ01 succinate 554.5 556.0 560.5 562.0 565.0 KW102 succinate — — 556.0 557.5 — 560.5 trace 565.0 KW105 succinate — — 556.0 557.5 — 560.5 562.0 565.0 KW107 succinate — trace — 557.5 — trace trace 565.0 KW183 succinate — trace 556.0 — — — 562.0 565.0 KW401 succinate — 554.5 556.0 — — 560.5 562.0 565.0 KW407 succinate — 554.5 — 557.5 — — trace 565.0 KW417 succinate — 554.5 556.0 557.5 trace — 562.0 trace KW420 succinate — 554.5 556.0 trace — 560.5 562.0 565.0 Table VIII: Wavelengths of absorption maxima of cytochrome b components in membranes from classes of putative cyb~ mutants. Low temperature absorption maxima of 6-cytochrome a-absorption bands from reduced minus oxidized spectra of membranes from cells grown to stationary phase on D-glucose or succinate. Wavelength values are expressed in nanometres and derived from fourth order derivative analyses of the original partially resolved high resolution a-band spectra. {Results & Discussion} 121 useful resource for future investigations of these phenomena. The case of KW107 was particularly intriguing for the interpretation that it lacked cytochrome 6555 was supported by its failure to utilize the respiratory chain terminating in cytochrome o while retaining the ability to express and use that terminating in cytochrome d. Yet at the time this mutant was isolated the order of electron transfer between the cytochrome components of the former pathway was still being disputed by the laboratories of Anraku, Gennis and Poole as described in detail in the Introduction {95, 117, 128, 180}. Consequently this strain was subjected to further analyses and its respiratory cytochromes were solubilized and fractionated by liquid chromatography (section III, A+B). Investigations into the identity of 'cytochrome 6555' and the relationship of this fraction to cellular dehydrogenase activities are also described in section III.B. (ID Potentiometric Titrations (a) Membrane cytochrome complement of cells grown on specific carbon-energy sources Potentiometric analyses of cells' respiratory response to growth on a variety of specific carbon-energy sources was detected by redox difference spectrophotometry of membrane suspensions in the absorption range of the type-fr cytochromes' a-band in the presence of a platinum electrode as described under Materials & Methods. It was impractical to use the more sensitive Soret region due to spectral interference from the multiple amphipathic electrophilic mediators that were required to bring the cytochromes present in the sample membranes into equilibrium with the electrode. However, by using the a-band region of the absorption spectrum to determine the proportion of cytochrome reduced at selected potentials additional information was gathered at specified wavelengths which were representative of certain partially resolved a-bands even at the temperature of 305 K at which the titrations were performed. Thus a comparison could be made between reduction potentials of those cytochromes displaying absorption maxima from 562 nm to 565 nm and those cytochromes absorbing between 558 nm and 560 nm at the titration temperature. Titrations of resuspended, washed membranes from cells grown aerobically on minimal medium showed distinct profiles which were dependent upon the strain of cells being used, upon the carbon-energy source available and upon the phase of growth at which the cells were harvested {66, 69,117}. Therefore the basis of comparison for this study was a series of cyd' cell lines derived from wild-type strain GR17N grown on a minimal, defined medium (CYD) in the presence of the {Results & Discussion} 122 Fig. 24: Differences in potentiometric titrations of membrane preparations from cells grown to stationary phase on either D-glucose or on DL-lactate. Wild-type strain GR17N was grown aerobically to stationary phase on CYD minimal medium supplemented with either 2.. D-glucose (•—•—•), or b_. DL-lactate (0--0—0). Membranes were harvested, washed and resuspended in degassed 100 mM potassium phosphate buffer, pH 7.0 at a protein concentration of (a) 25.9 mg mL"1, (b) 14.6 mg mL"1. {Results & Discussion} 12 4 non-fermentable carbon-energy source DL-lactate. These strains have been used by the laboratory of R. B. Gennis for numerous investigations of aerobic respiratory pathways in Escherichia coli and the published results include several potentiometric titrations {8, 62,117}. DL-lactate has been used as a carbon-energy source by the majority of laboratories studying aerobic respiratory pathways of E. coli {60, 100, 102, 130} although the current investigations demonstrated that there were disadvantages associated with its use when preparing membrane suspensions for electrochemical studies (v.i.). Figure 24 shows potentiometric titration profiles for type-6 cytochromes of membranes from wild-type GR17N cells grown to stationary phase on either DL-lactate or on D-glucose. The redox difference spectra and CO-binding spectra in Figure 25 indicate that under these growth conditions the respiratory pathway terminating in cytochrome d predominated and that relatively little cytochrome o was present, especially in the glucose-grown cells (see also Fig. 16 to 20). Nevertheless the two titration profiles deviate substantially at both low and high potentials as did the spectral determinations of the cytochrome 6 complement of each set of membranes (Fig. 24, 25). Although evidence will be presented in subsequent sections suggesting that certain low-potential characteristics of each profile were probably related to differences in dehydrogenase complement caused by adaptation to the carbon-energy source provided, the high-potential, endogenous response of freshly resuspended membranes from cells grown on DL-lactate displayed an anomalous quantity of reduced cytochrome: approximately 25 % of the total. Furthermore there was a rapid rate of spontaneous reduction of these 'lactate-grown' samples which stabilised at a potential of approximately +25 mV. In order to obtain a complete reduction profile the reduced samples were routinely reoxidized with K3Fe(CN)6 as illustrated in Figure 26b. The two panels of Figure 26 show that detectable differences in type-6 cytochrome complement between wild-type strain GR17N and its cyd' derivative GR19N were limited to high potential cytochromes when both strains were grown to stationary phase under aerobic conditions. This result may be surprising in light of the distinct features of the low-temperature redox difference spectra provided by similarly prepared membrane samples from these strains which indicated that under such conditions the predominant respiratory chain produced by the parent strain was that of the cytochrome d complex while the cyd' derivative relied on that terminating in cytochrome o (Fig. 25). It is suggested that only a limited number of the type-6 cytochromes resolvable by high resolution difference spectroscopy were distinguishable by potentiometric titration of resuspended membranes as implemented in this study. This restriction may have been a result of a limited number of those spectrally resolvable cytochromes being major contributors to the a-band absorption spectrum of the membranes, or the major contributors to the spectrum may have had mid-point {Results & Discussion} 12 5 Fig. 25: Ambient temperature redox difference spectra and carbon monoxide binding spectra of wild-type cells grown on either D-glucose or DL-lactate. Cells of w+ strain GR17N were grown to stationary phase on CYD minimal medium supplemented with carbon-energy sources of either D-glucose or DL-lactate; crude membranes were prepared from each culture and washed as described in Materials & Methods. Membrane samples were resuspended in 100 mM potassium phosphate buffer, pH 7.0 to protein concentrations of21.7mgmL"1 (DL-lactate), 21.0 mgmL"1 (D-glucose). 3,. Dithionite reduced plus carbon monoxide minus dithionite reduced difference spectra obtained at 295 K. 1. Strain GR17N grown on DL-lactate to stationary phase, AA = 0.10. 2. Strain GR17N grown on D-glucose to stationary phase, A A = 0.03. b_. Dithionite reduced minus peroxide oxidized difference spectra obtained at 295 K. At wavelengths greater than 500 nm, AA = 0.10 ; below 500 nm, AA = 0.33. 1. Strain GR17N grown on DL-lactate to stationary phase. 2 . Strain GR17N grown on D-glucose to stationary phase. {Results & Discussion} 12 6 {Results & Discussion} 12 7 Fig. 26: Potentiometric differences between membrane preparations from w and cyd' cells grown to stationary phase. Washed membranes were prepared from E. coli strains GR17N (w+) and GR19N (cyd') after aerobic growth to stationary phase in CYD minimal medium in the presence of: a. D-glucose : ( 0 - - 0 - - 0 ) , GR17N ; (+-+-+), GR19N ; or b_. DL-lactate : ( 0 - - 0 - - 0 ) , GR17N ; (+-+-+), GR19N. The washed membranes were resuspended in degassed 100 mM potassium phosphate buffer, pH 7.0 at protein concentrations of: a- GR17N, 25.9 mg mL"1 , GR19N, 33.4 mg mL"1 ; AA = 0.050, ii. GR17N, 14.6 m g m L ' 1 , GR19N, 22.8 mg mL"1 ; AA = 0.040. Arrows refer to potentials at which samples were removed and rapidly frozen for high resolution spectrophotometric analysis (see Fig. 28). {Results & Discussion} 12 8 5 < 6 o u o o -a 3 J . O X " i **** AA GR19N base l ine GR17N base l ine o -300 -100 +100 , +300 Eh (mV vs NHE) < a E c u o c 4-1 o GR19N base l ine GR17N base l ine o 1--300 -100 +100 v , „ +300 E h (mV vs NHE) {Results & Discussion} 129 potentials that were grouped such that under the electrochemically equilibrated conditions of the titration procedure brought about by the influence of the mediators they acted as pools of -^cytochrome and appeared as a smaller number of electrochemically distinguishable entities. Note that part of the spectral evidence for the greater number of type-fr cytochromes being present in the membrane comes from fourth order derivative analyses which are known to be qualitative in nature and extremely sensitive so that minor components, undetected by other analytical methods, may have been disproportionately emphasised (section I.i.d) {22,90}. Thus it is proposed that the differences in the low potential type-fr cytochrome complement of the cells grown on glucose or lactate were caused by distinct cytochromes associated with the dehydrogenases induced under each set of growth conditions (Figure 26a vs. 26b). Similarities between the two strains' mid-potential cytochromes (-25 mV to +125 mV) suggest that the two aerobic respiratory chains may share -^cytochromes with common characteristics, one candidate being cytochrome 6555 which was found to be ubiquitous under aerobic conditions (Figure 25 a+b) {4, 87, 180, 186}. When the two strains were grown on the same substrate, whether glucose or lactate, differences between their type-fr cytochromes were restricted to the high potential range (+125 mV and above), implying that this was the region in which differences between the two terminal oxidases were being observed (Figure 26 a,b). (b) Electrochemical characteristics of the aerobic respiratory tvpe-fe cytochromes The cyd' strains were essential for investigating electrochemical characteristics of the high-aeration respiratory pathway terminating in cytochrome 0 : harvesting cells in early exponential phase by the procedures available would have provided insufficient material for analysis by potentiometric titration. The type-fr cytochromes detected by this procedure and associated with cytochrome o respiration are shown to possess mid-point potentials of approximately +220 mV by comparison of the curves in both panels of Figure 26. Table IX displays Eh values for -^cytochromes determined from titrations of membrane suspensions derived from several strains, these values being within the range published by other laboratories {157, 216}. The high potential type-6 cytochrome present in membranes of GR17N wild-type cells grown to stationary phase on either substrate may have been the cytochrome £553 of the cytochrome d complex : it was not revealed by titrations of cyd' membranes, it demonstrated an absorption maximum of approximately 559nmat 305 K and the investigations of R. B. Gennis have shown it to be particularly sensitive to environmental manipulation {119}. In these and related experiments {Results & Discussion} 13 0 Growth on DL-lactate : GR17N GR19N GR19N (+thiamine; n=3) (+thiamine; n=4) (-thiamine; n=4) E_h % total E H % total E H % tptal +363 (±6) 21 (± 5) — — +379 (±34) 13 ( ± 7 ) +211 (±13) 25 ( ± 8 ) +202 (±22) 26 (±5) +201 (±40) 18 (±5) + 90 (±20) 36 (± 2) + 69 (±17) 29 (± 13) + 70 (± 31) 45 (± 6) - 38 (±32) 10 (± 3) - 27 (±31) 28 (± 11) - 59 (± 18) 16 (± 6) -232 (±26) 8 ( ± 1 ) -173 (±11) 15 (±6) -214 (±46) 9 ( ± 3 ) Growth on D-glucose GR17N (+thiamine) £ h % total +262 +129 + 30 - 118 46 20 23 11 GR19N (+thiamine) £h +214 +101 + 37 -109 % total 16 32 34 18 PLJ01 (+thiamine) £ h % total +275 +103 + 24 - 71 27 28 34 14 PLJ04 (+thiamine) E H % total +263 28 +178 18 + 76 38 - 8 15 PLJ07 (+thiamine) E H % total +232 25 +162 15 + 58 37 - 82 14 - 224 8 KW107 (+thiamine) E H % total +223 37 +144 43 + 33 20 Table IX: Cytochrome b complement of aerobically-grown cell membranes determined by potentiometric titration. Mid-point potential values are provided as mVvs. NHE with the percentage contribution to total type-6 cytochrome measured in the cytochromes during each titration. Cells were grown to stationary phase on CYD minimal medium supplemented with the carbon/energy sources indicated; values in parentheses are standard deviations. Four component analyses are taken as the basis for comparison, as these are the most generally applicable under the constraints described in Materials & Methods; three or five component analyses are shown where necessitated because a four component fit was inadequate. {Results & Discussion} 13 1 the complex was titrated in membranes and also after solubilization and purification in different detergent solutions, showing the mid-point potential of cytochrome 6553 to vary from +61 mV in octylglucoside solution to +180 mV in membrane suspensions {8,105,117,119). Although both of these values are much lower than those determined in the current study, the cytochrome d complex exhibits radically different behaviour in an oxygenated or partially oxygenated state, so sample preparation procedures may be influential (v./.). Alternatively, the high potential titration feature may be related to the P-band of cytochrome 6595 which absorbs at 562 nm in reduced minus oxidized difference spectra at ambient temperatures and possesses a mid-point potential of +113 mV {105, 117). Again, although this is substantially less than the value of +235 mV depicted in this titration, Fig. 27, it has not yet been determined whether cytochrome 6595 interacts directly with molecular oxygen : an oxygenated form may exhibit distinct properties. Analysis of the properties of the cytochrome d complex in greater detail (v.i.) suggested that although free oxygen had been rigorously excluded from these membrane titrations it is possible that this terminal oxidase, which has a particularly low A:m for dioxygen {4, 112, 157), retained bound oxygen and behaved as an oxygenated complex (section II.B.ii) {80, 105, 157}. Multiple wild-type strains were available which generated different patterns of type-6 cytochromes in their a-band redox difference spectra, indicating that they contained varying quantities of similar cytochromes when grown under comparable conditions (Fig. 21). The incomplete resolution of type-6 cytochrome a-bands achieved by low temperature redox difference spectrophotometry prevented accurate quantitative estimation of the individual 6-cytochromes present. When coupled with fourth order derivatization the techniques provided accurate spectral identification of those type-6 cytochromes but remained solely qualitative. The potentiometric titration data characterized a limited number of discernable type-6 cytochromes both quantitatively as their percentage contribution to the total cytochrome 6 a-band absorbance and qualitatively by electrochemical identification in terms of mid-point potential. Consequently unambiguous correlation between cytochromes detected in each of these two analytical methods was impossible, even when the aerobic respiratory chains were investigated independently by the selection of genetic and growth conditions described above. Attempts to achieve such correlations were made by means of several procedures. Investigations were conducted into observed perturbations of the electrochemical response of certain type-6 cytochromes by manipulating conditions of cell growth or of the titration itself. Samples were removed from titrations poised at known potentials for low temperature spectrophotometric analysis. Strains containing putative cytochrome 6 mutations in addition to the cyd' characteristics {Results & Discussion} 13 2 Fig. 27: Modification of the titration characteristics of specific membrane cytochromes by cell growth without thiamine and by treatment of membrane preparations with ferricyanide. Washed membranes were prepared from cyd' strain GR19N after aerobic growth to stationary phase in CYD minimal medium supplemented with DL-lactate. The washed membranes were resuspended in degassed 100 mM potassium phosphate buffer, pH 7.0. a. Cells grown in the presence (+-+-+) or absence (O-O-O) of thiamine.HCl, which was added to a concentration of 10 u.g mL"1 under standard conditions. b_. Successive reductive titrations of a membrane suspension, the sample being reoxidized with minimal quantities of titration buffer saturated with K3Fe(CN)6 between each titration : (-+-+-+) initial reduction profile, (-0--0-) secondary titration following reoxidation, (-B-*) tertiary titration following second reoxidation. {Results & Discussion} 13 3 -300 -100 +100 ' +300 Eh (mV vs NHE) {Results & Discussion} 13 4 were generated and analysed in order to obtain additional simplification of the cytochromes expressed. The following sections describe these studies. (c) Perturbation of the electrochemical response of certain tvpe-fc cytochromes Perturbation of the electrochemical response of certain type-6 cytochromes as determined by potentiometric titration could be achieved by several means. Apart from adding ligands that would bind to the terminal oxidase, a technique discussed in detail below (sections II.B.i+ii, III.B.iv.f & Ill.C.iv.a), the growth of cells without a thiamine supplement in the medium and the addition of ferricyanide to reduced titration samples both dramatically affected the ^ -cytochrome titration profile of membrane suspensions. The effects of the latter two procedures are illustrated by Figure 27 in which panel 'a' depicts representative titrations of cells grown under equivalent conditions other than the presence or absence of a thiamine supplement to a final concentration of 10 ug mL"1 and panel 'b' shows the effect of successive reoxidation of a reduced membrane suspension with ferricyanide ion on subsequent reductive titrations. All strains of Escherichia coli used in these investigations were K-12 derivatives and as such are categorised as thiamine auxotrophs. Nevertheless these cells will grow, albeit slowly, on minimal media in the absence of added thiamine indicating that the requirement is not absolute and suggesting that the thi-1 lesion itself is 'leaky'. In both procaryotes and eucaryotes thiamine pyrophosphate is the cofactor for decarboxylases acting on a-keto acids, transketolases and other reactions involving transfer of groups derived from a ketone {36}. The pyruvate decarboxylase complex is of central metabolic importance to an actively respiring bacterium growing aerobically in a defined, minimal medium with either glucose or lactate as sole carbon/energy source, as are pyruvate oxidase and the oc-ketoglutarate dehydrogenase complex, two other enzymes that require thiamine pyrophosphate {64}. An inadequate supply of the cofactor required for activity of these enzymes would disrupt intracellular metabolic pools and the dynamic equilibria between reduced and oxidized forms of the pyridine nucleotides and between the adenosine phosphates (phosphorylation potential 'AG p ' , or energy charge 'ec') {5, 64}. A feasible consequence of such a situation might be an adaptation of the electron transport pathways involving an alteration of cytochrome components. However, whereas quantitative or qualitative changes in the expression of dehydrogenases might be expected in these circumstances, revealed by a low potential modification of type-fc cytochromes, a more general {Results & Discussion} 13 5 response was observed. As indicated in Figure 27a the entire cytochrome-6 profile of membrane suspensions from cells grown in the absence of thiamine is shifted to higher potential in comparison with controls. This general displacement of some 100 mV is difficult to explain without postulating the complete replacement of respiratory pathways. An alternative interpretation of the data is that the effect of growth without thiamine caused a respiratory adaptation in which the cells failed to express the low potential type-6 cytochromes. Since the data are illustrated as the proportion of total cytochrome reduced the latter situation would cause the high potential cytochromes to be emphasised in the titration profile which would consequently appear to undergo a gross shift to higher potential. The concentrations of type-6 cytochrome per milligram of membrane protein were not comparable between cells grown with and without thiamine, and so a direct test of this phenomenon cannot be made until there are accurate means of identifying which cytochromes are being reduced at each stage of a potentiometric titration. Nevertheless a metabolic cause for the loss of dehydrogenase expression is relatively simple to postulate if thiamine is absent, since lactate utilization via pyruvate would be slower as would the passage of metabolites through the early steps of the tricarboxylic acid cycle resulting in lower concentrations of a number of potential dehydrogenase substrates. Thus the means by which thiamine influences respiration and cytochrome expression merit further investigation. Further evidence of a major change in respiratory complement was obtained by membrane solubilization with Triton detergents and anion-exchange chromatographic fractionation of the extracts. In cells grown without thiamine these extracts contained significant quantities of a cytochrome with reduced minus oxidized absorption maxima of 423.0 nm, 524.0 nm and 553.5 nm, indicative of a c-type cytochrome, the original membranes having a pronounced a-band absorption shoulder at approximately 550.0 nm. (data not shown). Fourth order derivative analysis indicated that the a-band peak comprised two components, one of 552.0 nm and the other of 554.5 nm which may be indicative of two cytochromes or may be a feature of a respiratory cytochrome c (c.f. section I.i.b ; Fig 5). The quantity of this material was difficult to estimate because of spectral overlap and incomplete chromatographic fractionation, but this cytochrome was resistant to both urea and cholate stripping of the membrane and total Soret absorbance of the cytochrome c fractions from the extract was approximately one quarter of that of the pooled cytochrome o complex, with which it partially overlapped (data not shown). The appearance of this novel cytochrome under such conditions of nutritional stress is curious because of the lack of c-type cytochrome in E. coli K-12 cells grown aerobically with an adequate thiamine supply, and it may be related to the perturbation observed in the potentiometric membrane titration (Fig. 27). A second form of electrochemical perturbation of cytochromes was discovered during the t {Results & Discussion} 13 6 course of potentiometric titrations of membranes from lactate-grown cells. The differential effect of ferricyanide was subsequendy exploited to distinguish certain type-6 cytochromes within membrane suspensions and solubilized preparations. As described above, freshly resuspended membranes prepared from cells grown on DL-lactate tend to auto-reduce, possibly as a result of endogenous substrates being present even after extensive washing of the membrane preparation with aqueous buffer in repeated cycles of homogenization and centrifugation. In order to obtain data points at the high potential range of the titration profile it was necessary to reoxidize the sample. When oxygenating reagents were used for this purpose, such as small aliquots of concentrated solutions of (NH4)2S20g or H2O2, an extended period was required before the sample re-equilibrated with the electrode as indicated by inconsistent data and considerable hysteresis on comparing reducing and oxidizing curves. Following full reoxidation with these reagents and adequate equilibration times of 45 min to 60 min subsequent reductive titrations closely followed the original pattern (data not shown). However, if oxidation of the reduced sample was carried out with K3Fe(CN)6 a displacement of part of the titration curve was observed while most of it remained constant. The modification indicated that a cytochrome generally comprising about 20% of the total and with a potential of approximately +25 mV was undergoing a dramatic shift in its electrochemical properties such that after the ferricyanide treatment it behaved with a new mid-point potential of -175 mV. The clarity with which this effect could be observed depended upon the source of the membrane sample : the proportions of individual cytochromes and the resolution of components in any specific section of a titration profile was dependent upon the bacterial strain and the growth conditions used. In a few cases the original cytochrome was detectable from the initial reductive titration profile whereas in the majority of samples it appeared as part of a larger, unresolved portion of the total cytochrome having a mid-point potential of approximately +45 mV. The appearance of the displaced component was clearly resolved in either situation after the ferricyanide treatment since there were no cytochromes with mid-point potentials interfering in the -200 mV region of the titrations. No significant alterations in the complement of type-fr cytochromes were apparent from the a-bands of the redox difference spectra performed at 305 K as part of the potentiometric procedure nor from spectrophotometric analysis of poised potential samples taken during the course of the titrations (section II.A.ii.d). This was an additional indication that neither spectrophotometric nor potentiometric analytical methods were capable of resolving all the type-6 cytochromes present in a typical membrane sample from aerobically-grown cells. This phenomenon complicated the collection of high-potential titration data from membrane {Results & Discussion} 13 7 samples derived from cells grown on DL-lactate. These samples also auto-reduced and were difficult to maintain in the oxidized state required for the spectroscopic reference cell during the extended titration procedure. Therefore disodium succinate was routinely used as a non-fermentable carbon-energy source for not only did it eliminate these procedural difficulties it also provided higher yields of respiratory cytochrome. The selective perturbation of cytochrome mid-point potentials by ferricyanide suggested the possibility of its use as a technique for resolving and identifying certain type-6 cytochromes. However, the inability to correlate spectrophotometric and potentiometric data, even when incorporating the available cyd' strains and variations of growth conditions into the experiments, indicated that this procedure would have to be used in conjunction with additional methods of separating the complex mixtures of cytochromes present in membrane preparations. (d) Poised potential high resolution redox difference spectrophotometry In spite of the limited resolution provided by the redox difference spectra collected at 305 K as part of the potentiometric titration procedure it was apparent that in membranes from cyd' strains and in those from wild-type cells grown to exponential phase the cytochromes being reduced at the highest potentials displayed a biphasic a-band. It could not be determined whether one or multiple cytochromes caused these features, but a broad absorption peak at 565 nm was accompanied by a slightly smaller broad absorption peak at 558-560 nm. The lack of spectral resolution at these temperatures and the small absorption change observed in the a-band region due to the reduction of these components prevented more detailed analysis and suggested that potentiometric poising coupled with high-resolution spectrophotometric analyses would be fruitful. As described in section I.ii.c the procedure of high resolution spectrophotometry of samples poised at specific potentials offered an opportunity to simplify the number of cytochromes contributing the a-band redox difference spectrum of a membrane sample. By taking samples for high resolution spectrophotometry which had been poised at a range of potentials throughout the titration of a membrane preparation increasing proportions of the constituent cytochromes would be electrochemically reduced. High resolution spectrophotometry could then be employed at 77 K to indicate the combined spectral characteristics of those cytochromes reduced at each 'poised' potential. It was anticipated that subtraction of selected digitised spectra would yield high-resolution spectral data for those cytochromes undergoing reduction between each pair of 'poised' potentials. It should be {Results & Discussion} 13 8 remembered that low temperature spectrophotometric data from samples poised at cited potentials cannot be taken as representing the state of reduction of similar samples maintained at ambient temperature at the same potential. The poising technique only provides a valid direct comparison between similarly treated samples although it appears to be indicative of the sample components' relative sensitivity to reduction if low temperature analyses are performed on a series of samples poised at a progression of potentials (section I.ii.c). Figure 28 portrays low temperature spectra of samples removed and poised at several potentials during the wild-type and cyd' strain membrane titrations illustrated in Figure 26. These high resolution spectra show the a- and P-bands of the type-fr cytochromes and the a-band region of cytochrome d. When each of the two cell strains had been grown to stationary phase the low potential spectra of the wild-type, GR17N, showed that it contained very significant quantities of cytochrome d by the combined peak and trough between 600 nm and 700 nm, but there was no detectable Z>-cytochrome a-band shoulder between 560 nm and 565 nm which would have been attributable to the cytochrome o complex. The cyd' strain GR19N displays minimal cytochrome d but a substantial red shoulder on the 6-cytochrome a-band, confirming that these two preparations are representative models for investigating the two alternative aerobic terminal oxidases. (The small amount of cytochrome d complex expressed by GR19N when grown to stationary phase upon DL-lactate has been commented upon elsewhere by Johnson & Bragg {89}. This phenomenon was not observed when these cells were grown on succinate, nor when the derivative strain PLJ01 was grown on either of these non-fermentable carbon sources.) However, the presence of other cytochromes, especially those of low potential, in membranes containing the cytochrome o terminal oxidase which have been derived from cyd~ cells grown to stationary phase may not be representative of cytochromes normally associated with the cytochrome o complex. Under the batch conditions of this study, stationary phase cyd' cells had necessarily been grown under abnormal conditions, for in wild-type strains the cytochrome o respiratory chain would have been replaced by that of cytochrome d as soon as the dissolved oxygen concentration fell below a critical level, as described in the Introduction. Thus the cytochrome o complex may not be functioning effectively in cyd' cells approaching stationary phase and components of other respiratory chains may be expressed as the cells adapt to the carbon and energy sources resulting from sub-optimal utilization of aerobic respiratory metabolism of the nutrients in the medium supplied. Fourth order derivative analyses of the poised potential spectra are shown in Figure 28b and their peak positions identify the ^ -cytochromes reduced at each potential. The red a-band shoulder associated with cytochrome o that was present in the cyd' cell membranes at +98.5 mV was {Results & Discussion} 139 Fig. 28: Poised potential high resolution difference spectra of w+ and cyd' cells from the titrations illustrated in Fig. 26. High resolution reduced minus oxidized difference spectra of poised potential samples removed during the titrations illustrated in Figure 26 and measured at 77 K. Strains GR17N (w+) and GR19N (cyd") had been grown aerobically to stationary phase in C Y D minimal medium in the presence of D-glucose. Washed membranes were prepared and resuspended in degassed 100 mM potassium phosphate buffer, pH 7.0 at protein concentrations of 25.9 mg m l / 1 (GR17N) and 33.4 mg ml / 1 (GR17N) for the titrations. Potential values indicate the potential at which the sample was removed from the titration vessel. 2.. GR17N membranes, AA = 0.020, b_. GR19N 1. -337.5 mVvs.NHE, versus reference reoxidized with H2O2. 2. -337.5 mVv5.NHE, 2'. fourth order finite difference spectrum. 3. -111.5 mVv .^NHE, 3'. fourth order finite difference spectrum. 4. -26.0 mVv .^NHE, 4'. fourth order finite difference spectrum. 5. +185.0 mVv .^NHE, 5'. fourth order finite difference spectrum. membranes, AA = 0.020, 1. -348.5 mVvs.NHE, 1'. fourth order finite difference spectrum. 2. -90.5 mVvj.NHE, 2'. fourth order finite difference spectrum. 3. -25.5 mVvs.NHE, 3'. fourth order finite difference spectrum. 4. +98.0 mVvs.NHE. 5. +264.5 mVw.NHE. {Results & Discussion} 1 4 0 {Results & Discussion} 141 accompanied by a distinct, albeit smaller, peak at 556 nm reflecting the observations obtained from the 305 K spectra. This peak increased in size as the potential was lowered until it became the dominant feature of the spectrum. Because of the overlap between these two absorption maxima and the change in size of the one peaking at lower wavelength it was not possible to determine whether the higher wavelength peak was also increasing in amplitude as the potential was decreased. While any increase in the size of the red shoulder was significantly less than that of the 556 nm peak an accurate determination of relative amplitude would have required curve deconvolution analyses of the spectra — a technique that has been used by other laboratories for determination of component contributions to ambient temperature spectra but one that was not available for the current investigation {203, 204}. The fourth order derivative analyses indicated that those cytochromes undergoing reduction at potentials below +100 mV had redox absorption maxima of 556 nm at 77 K. The cytochromes that were reduced after adjustment of the potential to +100 mV, which included those associated with the cytochrome o complex, had low temperature absorption maxima at 555.0 nm, 557.5 nm, 562.0 nm and 565.0 nm. The two higher wavelength peaks are components of the red shoulder, there being insufficient absorption by these poised potential samples to determine whether or not a third component of the red shoulder was present, as observed in fully reduced minus fully oxidized membrane suspensions. At lower potentials the fourth order derivative spectra exhibited decreased resolution of these components, which may have been caused by an increase in concentration of a component absorbing at 564.5 nm or simply by the influence of the increasing quantities of electrochemically reduced cytochrome £556- The latter explanation is favoured due to the results from high resolution fully reduced minus fully oxidized difference spectra of membrane preparations of this type (Fig. 19, 21). The wild-type membranes were shown to have contained a simpler fc-type cytochrome complement in addition to cytochrome d (Fig. 28a). At +190.5 mV the cytochrome d was fully reduced, suggesting that the large inflection in the potentiometric titration profile (Fig. 28) was caused by absorption by one or both of the ^ -cytochrome components of the cytochrome d complex, cytochromes 6553 and 6595. The fourth order derivative spectra shown in Figure 28b indicate that there is an absorption peak at 559 nm, the wavelength associated with absorption by cytochrome 5^58 although the original 77 K redox difference spectra suggested that between 50 % and 70 % of this cytochrome was not reduced at this high potential. However, because of the overlap of the two cytochrome a-bands the 'total' absorption at 558 nm at low potential would have incorporated a substantial contribution from cytochrome 6556. implying that the large, high potential inflection in the titration profile corresponded to the full reduction of cytochrome 6553 which itself is associated {Results & Discussion} 142 with the terminal oxidase (Fig. 27, 28). Thus a precise estimation of percentage reduction of individual components could not be provided without deconvolution analysis since there was interference from the adjacent absorption maximum at 556 nm. Additionally the contribution to the combined peak from the P-band of cytochrome 6595 could not be gauged accurately : the broad, low amplitude absorption peak caused by the a-band of cytochrome 6595 was too small for accurate quantitation {118}. Nevertheless there was little change in amplitude of either the small a-band attributed to the latter cytochrome's high-spin haem or the small shoulder at approximately 550 nm which has been suggested to be caused by its P-band {118} indicating that a substantial portion, if not all of the cytochrome 6595 was reduced at this high potential, as was cytochrome d itself. As the potential was lowered there was a gradual shift in the absorbance maximum of the lower wavelength component, changing from 554.0 nm at +190.5 mV to 555.5 nm at +532.0 mV. It is uncertain whether this change was significant, although examination of the titration profile in Figure 26 indicates that several electrochemically distinct type-6 cytochromes were reduced over this potential range. In spite of this, there are few alterations visible in the poised potential redox difference spectra (Fig. 28b). The 6-cytochrome a-band increased in height by a factor of 2.5 between +190.0 mV and -20.5 mV. Cytochrome 6553 was responsible for a significant portion of this change as described above, but an increasing contribution was due to one or more type-6 cytochromes absorbing between 555 nm and 556 nm. Moreover, although there was no apparent alteration of the cytochrome d a-band at potentials above -20.5 mV at each of the lower potentials one observed a progressive decrease in amplitude of both the 626 nm 'reduced' peak and of the 648 nm 'oxygenated' trough and a shift of the former from 626 nm to 635 nm. The cause of this absorbance shift is unknown, although the sensitivity of the reduced peak of cytochrome d to perturbation by the addition of certain ligands was demonstrated in Figure 22: the decrease in amplitude of the adjacent peak and trough may simply be a consequence of greater overlap between these two spectral features as the peak shifts to higher wavelength. The results illustrated in Figure 28 indicate that although the poised potential technique for obtaining high resolution redox difference spectra at distinct potentials provided some additional information about the components of the two terminal oxidase complexes it was of little assistance in resolving the complexities of the other 6-type cytochromes with lower mid-point potentials. {Results & Discussion} 143 (e) Analysis of mutant strains with an altered type-fr cytochrome complement Generation of mutants was intended to simplify the cytochrome-fr content of relevant cell membranes in order that the remaining cytochromes might be resolved more thoroughly upon subsequent spectrophotometric or potentiometric investigation. The advantages of this approach were evident from the utility of the cyd' strains used throughout these studies which removed interference by cytochromes associated with the cytochrome d terminal oxidase. However, further simplification of the membranes' cytochrome content was required if the type-6 cytochromes of the central and early segments of the electron transport chains were to be analysed, as has been shown in previous sections by the inability of the applied methodologies to accomplish adequate resolution of these components. In addition it was anticipated that if ^ -cytochrome mutants could be prepared they might also serve as keys with which to correlate the information provided by each of these analytical techniques. Two types of mutants were sought: those with defective expression of constituents of the cytochrome o complex which would be used in an analogous manner to the cyd' strains that were already available (cyo' strains), and mutants which expressed abnormal type-fr respiratory cytochromes or that failed to express one or more of these cytochromes, useful in studies of the order of electron flow between respiratory components (cyb~ strains). Selection procedures were developed to promote the isolation of mutants defective in the activity of the respiratory chain terminating in cytochrome o yet spectroscopic techniques were unable to make positive identification of any strains failing to express the cytochrome o complex (see, however, section Il.C.ii.b). However, certain mutant strains were observed to generate abnormal 'reduced plus carbon monoxide minus reduced' spectra with minimal features in the Soret region of cytochrome o itself. Following nitrosoguanidine mutagenesis enrichment and selection procedures for potential cyo' mutants were performed as described under Materials & Methods. Strains KW420, KW424 and KW425 were each distinguished by the lack of a clear absorbance trough at 430 nm in CO-binding spectra indicating the possibility that cytochrome o was absent from their membranes (Fig. 29). Interpretation of these reduced plus carbon monoxide minus reduced difference spectra was complicated by the overlapping absorbance shifts resulting from CO binding to the cytochrome d complex, which was induced under the growth conditions provided. These strains were maintained and subcultured under anaerobic conditions in order to minimize reversion, but as found with the cyo',cyd' strains when competing respiratory chains were eliminated the reversion rate was sufficient to induce enough cytochrome o revertants to dominate the culture. This would even occur in cultures maintained under strict anaerobic conditions until the final aerobic {Results & Discussion} 144 batch culturing of the cells for analysis of the aerobic oxidases, respiring revertants rapidly outgrowing the fermentative mutant cultures (v.i.). Thus maintaining and using pure cultures of strains carrying point mutations in the cytochrome o terminal oxidase was not possible unless an alternative respiratory chain was able to be induced under the relevant growth conditions : in the case of this study the alternative chain was that terminating in cytochrome d, effectively negating attempts to simplify the cytochrome content of the respiratory apparatus for spectrophotometric or other analyses. The absence of cytochrome o was unable to be confirmed from the CO-binding spectra of membrane preparations from the putative cyo',cyd* strains KW420, KW424 and KW425 because of the overlap between the Soret bands of the two aerobic terminal oxidase complexes, coupled with that from their CO derivatives (Fig. 29). Nevertheless, these analyses showed that if present, any cytochrome o that existed in these membranes was in greatly reduced quantities when compared to preparations from the parental strain PLJ04. A 560 nm CO-binding spectral trough was still visible in these spectra, as was the red shoulder on the redox a-band between 560 nm and 565 nm (Fig. 29). It is therefore possible that these spectral features are only related to cytochrome o indirecdy, and may be caused by another cytochrome component of the oxidase. Since considerable progress was being reported by the laboratory of R. B. Gennis in locating the chromosomal position of the cyo operon in addition to the generation of cyo' strains as a preparatory step in cloning the gene products responsible for this oxidase activity, attempts at creating and characterizing additional cyo' mutants in the current study were abandoned. R. B. Gennis generously provided cyo',cyd' strain RG167 which had been produced from strains with single site polar mutations in each of these two operons but under aerobic conditions the selection pressure for respirationally competent back mutations was sufficiently great to prevent a pure culture from being maintained. Subculturing plus spectroscopy of carbon monoxide derivatives showed that even single revertant strains were unstable, generating a series of respiratory phenotypes in which features of the membranes' CO-binding spectra could be predicted from certain characteristics of each strain's colony morphology (data not shown). The fc-cytochrome mutants generated and isolated in this study were classified into groups on the basis of their parentage and any cytochrome abnormalities detected by low temperature redox difference spectrophotometry coupled with fourth order derivative analysis. Spectral characteristics of the major groups of putative cyb~ strains are listed in Table VIII. Strain KW107 is a representative example of those mutants which appeared to have, from low temperature spectrophotometry of membrane suspensions, a minimal cytochrome 6555 content (Fig. 23). Electrochemical properties {Results & Discussion} 14 5 Fig. 29: Carbon monoxide binding spectra from putative cyb" and potential cyo' mutant isolates. Ambient temperature reduced plus carbon monoxide minus reduced difference spectra of washed membranes of strains grown aerobically to stationary phase on CYD minimal medium supplemented with D-glucose. Membranes were resuspended at the stated protein concentrations in 100 mM potassium phosphate buffer, pH 7.0. Procedures for preparation and gassing of samples are described under Materials & Methods. 1 . Parental strain PLJ04 , 2. Derivative strain KW420 , 3 . Derivative strain KW424 , 4 . Derivative strain KW425 , 2.5 mg mL"1, AA = 0.050. l . O m g m L - 1 , AA = 0.005. 1.0 mg mL"1, AA = 0.005. l .OmgmL" 1 , AA = 0.005. {Results & Discussion} 146 {Results & Discussion} 14 7 of KW107 are listed in Table IX and illustrated in Figure 30 in comparison to titration profiles of membranes from GR17N & GR19N grown under the same conditions. It is apparent that membranes of cyd' strain KW107 contained virtually no cytochromes with mid-point potentials lower than 0 mV when it was grown on minimal medium supplemented with glucose. This strain is a derivative of PLJ01 and consequendy expresses cytochrome d components when grown on glucose, but not when grown on succinate or DL-lactate. It grows relatively slowly under aerobic conditions which would be expected to delay the oxygen-dependent onset of expression of the cytochrome d complex: a phenomenon demonstrated by the spectra in Figure 31a which show the mutant to have levels of both terminal oxidases that are intermediate between the exclusive reliance on either the cytochrome d complex or on the cytochrome o complex displayed respectively by GR17N and GR19N when each of the three strains is grown to stationary phase. (Cytochromes associated with the cytochrome o complex absorb in the 560 nm to 565 nm range of these redox difference spectra and cytochrome d a-band absorption is visible between 615 nm and 695 nm). If the high potential cytochromes are those associated with the terminal oxidase complexes, the identity and spectroscopic properties of those cytochromes of KW107 shown to be reduced over the range of 0 mV to +100 mV by Figure 31b are pertinent to the elucidation of organization of the aerobic respiratory chains, especially if these cells do indeed lack the cytochrome £55 5. These phenomena are addressed below, as is the difficulty of identifying 'cytochrome 6555' unambiguously. Solubilization and fractionation of membrane cytochromes from strain KW107 resulted in essentially 'normal' elution profiles from standard anion exchange chromatographic separations, including significant quantities of cytochrome with redox a-band absorption maxima of 556 nm at 77 K suggesting that the cytochromes 6555 solubilized from cell membranes by the standard procedure were not related to the 'central respiratory cytochrome ^ 555' of the cytochrome o electron transport chain (sections III.A.i+ii). (iii) Reduction Kinetics (a) Overview of technique Utilization of dual wavelength spectrophotometry for investigating the reduction kinetics of cytochromes enabled small changes of absorbance to be detected in the cytochromes' Soret or a-band regions even when the sample generated a high degree of incident beam light scattering, such as that {Results & Discussion} 14 8 Fig. 30: Potentiometric titration of mutant and parental strains with different complements of type-b cytochrome. Washed membranes were prepared from parental strain GR17N (w+) and derivative strains GR19N (cydr) and KW107 (cytochrome d complex under the control of growth substrate and possible cytochrome b modifications) after each had been grown aerobically to stationary phase in CYD minimal medium in the presence of D-glucose. The washed membranes were resuspended in degassed 100 mM potassium phosphate buffer, pH 7.0 at the following protein concentrations : GR17N, 25.9 mgmL-1, ( ); GR19N, 33.4 mgmL"1, ( ); KW107, 15.7 mgmL"1, ( • • • ) . Data points for the GR17N and GR19N curves are provided in Figure 26 but have been omitted here for clarity. {Results & Discussion} 15 0 Fig. 31: High resolution redox difference spectra of mutant strains with different complements of cytochrome. Washed membranes were prepared and resuspended for potentiometric titration from strains GR17N, GR19N and KW107 grown aerobically to stationary phase in CYD minimal medium in the presence of D-glucose (Fig. 30). Spectrophotometry was performed at 77 K following dilution of the samples to final concentrations of 100 mM potassium phosphate buffer, 1.0 M sucrose, pH 7.0, at the protein concentrations indicated: 3.. Dithionite reduced minus peroxide oxidized redox difference spectra. 1. GR17N, 13.0 mgmL" 1; 2. KW107, 7.9 mgmL- 1 ; 3. GR19N, 16.7 mgmL' 1 . Fourth order finite difference spectra calculated from the corresponding redox difference spectra in panel 'a'. {Results & Discussion} 1 5 1 500 • 600 X(nm) 700 I i I I I 1 I I 1 i r— 5^0 ' ' ' 560 ~ ~ n ' 5^ 0 A (nm) {Results & Discussion} 15 2 caused by membrane suspensions {90}. In practice the a-band absorbance change was most often used to indicate the proportion of cytochrome in the reduced state to circumvent interference from the coloured mediator phenazine methosulphate which was frequendy used as an intermediary for the electronic equilibration between respiratory components and the experimental electron donor. Addition of a moderate excess of substrate to the single cuvette containing the resuspended sample results in an absorbance change indicative of the steady state level of reduction of the cytochromes present. Electrons flowing into the respiratory chains contained within the sample membranes pass through the various components and are transferred to dioxygen by the terminal oxidase or oxidases until all the oxygen dissolved in the sample has been consumed. When the suspension becomes anaerobic the loss of electrons from the terminal oxidases will cease while electron flow into the respiratory chains continues until all components susceptible to reduction have been reduced. Several complications are immediately apparent. Without addition of an electrochemical mediator only certain biological substrates are capable of transferring electrons into the respiratory chains : this may be used as the basis of assays for specific dehydrogenase activities. Notable exceptions are quinol analogues, many of which are soluble in the membrane lipid phase and interact with cytochrome quinol binding sites; other exceptions are strong redox reagents such as Na2S2C>4 and K3Fe(CN)6- The sample membranes are impermeable to some of these reagents (e.g. (NH4)2S20g) which may thus be used for topological studies. The relative rates of electron transfer into and out of each component of the respiratory chains present will affect the size and duration of the steady states observed. If the respiratory systems in the resuspended membranes are intact transfer of electrons into electron transport chains from a biological substrate through a dehydrogenase will lead to reduction of only those cytochromes functionally linked to the dehydrogenase. This topic of functional linkage between cytochromes required clarification at the time of these studies, since it had been proposed that natural quinol analogues in the membranes of Escherichia coli (ubiquinol-8, menaquinol-8) would create an equilibrium between all respiratory components at the potential generated by the biological reductant supplied, should an appropriate dehydrogenase be present. If the fr-cytochrome components of the respiratory chains could be distinguished one might determine whether this was the case or whether restricted populations of cytochromes were preferentially reduced by specific substrates, while others of equivalent potential remained isolated : were the multiple respiratory chains of E. coli functionally or structurally distinct or did electronic equilibration occur between them? Sequential addition of reductants and oxidants could be used to determine the amounts of cytochrome sensitive to each of these reagents after specific pretreatment, and respiratory inhibitors were available to provide blocks in respiratory {Results & Discussion} 153 pathways enabling functional groups of cytochromes to be determined. As implemented the method provided absorbance data at a specific wavelength relative to that at an isosbestic point. Consequently the variation in quantity of reduced cytochrome was determined over time as a proportion of the the total absorbance change due to fully reduced cytochrome at the selected wavelength pair, although the identity of those cytochromes being reduced at any specific stage of the experiment was unknown. Steady state poising experiments were therefore linked with the procedure, as described under Materials & Methods, in order to obtain high resolution redox difference spectra from stages in the redox manipulation of cytochromes. These redox states were attained in the cryogenic sample holder at predetermined intervals after reagent addition. (b) Membrane cytochrome analysis When using an analytical procedure with so many variables meaningful interpretation of the results requires initial simplification of the system in order that the individual components may be recognised and distinguished. To this end cyd' strains were used for dual wavelength studies of the effect on the total cytochrome content of membranes from cells grown on different carbon/energy sources when succinate is added to the suspension (Fig. 32). Following an initial steady state of cytochrome reduction there was a rapid and a subsequent slow phase of dynamic cytochrome reduction, neither of which exhibited a first order rate. The changes in the rate of cytochrome reduction in each of these traces confirmed the results from other analyses that there were multiple respiratory type-6 cytochromes present in cells grown under conditions causing them to channel electrons from a single substrate, chosen from several carbon/energy sources tested, to the cytochrome o terminal oxidase. With all four membrane preparations resuspended at an equivalent protein concentration the total amounts of type-6 cytochrome in these variously-grown cells are seen to be different, the concentrations achieved, expressed as nmol (mg membrane protein)"1, being approximately 0.17 (glucose), 0.20 (succinate), 0.20 (DL-lactate) and 0.30 (L-proline) (49,97). The absorbance change attained at the initial steady state is indicative of the proportion of total cytochrome reduced, which itself is dependent upon the relative rates of electron influx and efflux through individual respiratory cytochromes as the substrate is oxidized and electrons transferred to dioxygen. Since the reductant used to create each trace in Figure 32 was disodium succinate the steady state absorbance reflects the dynamic reduction level of 6-cytochromes associated with succinate oxidase activity. This level of steady state reduction varies considerably between the various {Results & Discussion} 15 4 F i g . 32: Reduction kinetics of membrane cytochromes from cyd' cells grown aerobically to stationary phase on various carbon/energy sources. Washed membranes were prepared from cells of cyd' strain GR19N grown aerobically to stationary phase on CYD minimal medium supplemented with the carbon/energy sources indicated below. Dual wavelength analyses were carried out upon membranes resuspended in 100 mM potassium phosphate buffer, pH 7.0 at 4.0 mgmL" 1, as described under Materials & Methods, sections 1+n. Substrates were added to the final concentrations shown as indicated : S, disodium succinate (5.0 mM); N, NADH (0.8 mM); D, sodium dithionite (A slight excess, freshly saturated in buffer solution to provide complete reduction of sample). Maximal succinate reduction was obtained over the time course indicated with membranes from cells grown on DL-lactate or L-proline : it required considerably longer than 30 min in the other cases and the level of reduction achieved is indicated by the pointer S 1. DL-lactate (sodium), 2. succinate (disodium), 3 . L-proline, 4. D-glucose. {Results & Discussion} 15 5 {Results & Discussion} 156 preparations : 49.4 % (glucose), 40.0 % (succinate), 44.7 % (DL-lactate) and 3.0 % (L-proline) indicating that the rate of electron flow through respiratory cytochromes from this substrate is markedly different in the proline grown cells. Among parameters affecting the steady state reduction level are the succinate dehydrogenase activity (supplying electrons to the respiratory chain), susceptibility of respiratory cytochromes to reduction by succinate (relative activities of succinate dehydrogenase and 'succinate oxidase'), and the ability of other respiratory cytochromes to withdraw electrons from those components of the succinate oxidase chain ('ancillary oxidase activities', secondary cytochrome pools). The duration of the initial, aerobic steady state demonstrates how rapidly oxygen dissolved in the preparation is consumed and thus measures the efficiency of electron transfer through the available pathways from succinate to dioxygen. Oxidase activities expressed by these preparations were, in nmol min"1 (mg protein)"1, 0.022 (glucose), 0.063 (succinate), 0.137 (DL-lactate) and 0.032 (L-proline). The total amount of type-6 cytochrome that was reduced by succinate was 80.2 % (glucose), 93.6 % (succinate), 92.6 % (DL-lactate) and 73.2 % (L-proline), reflecting the results of potentiometric titrations which demonstrated that cells grown on glucose or proline had greater quantities of low potential cytochromes than those grown on lactate (Table IX, Fig. 33), the mid-point potential for the succinate/fumarate redox couple being +31mV {145}. These succinate oxidase activities were competitively inhibited by malonate and oxaloacetate (data not shown). Potentiometric titrations illustrate the inherent electrochemical properties of all the type-6 cytochromes present in a sample by progressively reducing each cytochrome with all components kept in electronic equilibrium by the added mediators. In contrast the absorbance changes observed in dual wavelength kinetic studies are the result of dynamic processes observed in progress and may be caused by partial reduction of several cytochromes simultaneously. Thus even when comparative data was available for the action of certain reductants upon membranes derived from cells grown under different nutritional conditions it was still not possible to identify the redox status of specific cytochrome species. Further simplification of the cytochrome complement of sample membranes was one response to this dilemma, using the putative cyb~ strains that had been generated in order to identify kinetic abnormalities which could be related to lesions affecting cytochromes and identified from spectral and potentiometric studies, as described elsewhere and discussed in the following section {214}. {Results & Discussion} 15 7 Fig. 33: Potentiometric titrations of membrane preparations from cyd' cells grown to stationary phase on L-proline, D-glucose or DL-lactate. Washed membranes were prepared from cells of cyd' strain GR19N grown aerobically to stationary phase on CYD minimal medium supplemented with L-proline, D-glucose or DL-lactate as indicated below. Error bars show the standard deviation of mid-point potential and percentage contribution to total type-6 cytochrome, in 'n' sample preparations, each cytochrome b being resolved by theoretical fits of the data by curves with the stated number of components. 2,. D-glucose supplement, n = 2, three component fit. b_. DL-lactate supplement, n = 3, four component fit. C_. L-proline supplement, n = 3, four component fit. d_. L-proline +Mo/Se supplements, n = 3,