New comprehensive biochemistry vol 11 modern physical methods in biochemistry part b

Time-resolved fluorescence spectroscopy Resolution of the emission spectrum of liver alcohol dehydrogenase Pulsed lasers for time-resolved fluorescence Frequency-domain resolution of

IN BIOCHEMISTRY, PART B

New Comprehensive Biochemistry

Modern Physical Methods in

Biochemistry Part B

Editors

London and Utrecht

1988 ELSEVIER AMSTERDAM * NEW YORK OXFORD

All rights reserved N o part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written permission of the Publisher, Elsevier Science Publishers B.V (Biomedical

Division), P.O Box 1527, loo0 BM Amsterdam, The Netherlands

No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a

matter of products liability, neghgence or otherwise, or from any use or operation of any methods; products, instructions or ideas contained in the material herein Because of the rapid advances in the medical sciences, the Publisher recommends that independent verification of diagnoses and drug dosages should be made

Special regulations for readers in the USA This publication has been registered with the Copyright Clearance Center, Inc (CCC), Salem, Massachusetts Information can be obtained from the CCC about conditions under which the photocopying of parts of this publication may be made in the USA All other copyright questions, including photocopying outside of the USA, should be referred to the Publisher

Sole distributors for the USA and Canada:

Elsevier Science Publishing Company, Inc

52 Vanderbilt Avenue

New York, NY 10017

USA

Library of Congress Cataloging-in-Publication Data

(Revised for volume 11 B)

Modem physical methods in biochemistry

(New comprehensive biochemistry; v 11 A, B)

Includes bibliographies and index

1 Spectrum analysis 2 Biochemistry-Technique

I Neuberger, Albert 11 Deenen, Laurens L.M van

QD415.N48 vol 11 A, etc 574.19’2 s [574.19’283] 85-4402

[QP519.9S6]

ISBN 0-444-80649-0 (v 11 A) 0-444-80968-6 (v 11 B)

Acknowledgment

Many illustrations and diagrams in this volume have been obtained from other publications In all cases

reference is made to the original publication ThejuN source can be found in the reference list Permission for the reproduction of this material is gratefully acknowledged

Printed in The Netherlands

Preface

In the former series of Comprehensive Biochemistry the contributions of physical methods to biochemistry were considered in volumes 1-4, a section which was devoted to the physicochemical and organic aspects of biochemistry In 1962 the series editors M Florkin and E.H Stotz emphasized the importance of these basic sciences for the future progress in the life sciences Since that time, the application

of physical methods to biological problems has solved many questions and opened new avenues of research

Volume 11, part A, of the present series contained chapters on protein crystallog- raphy, nuclear magnetic resonance spectroscopy, electron spin resonance, mass spectroscopy, circular dichroism and optical rotatory dispersion In this volume the range of spectroscopic techniques is extended to chapters on fluorescence and Raman spectroscopy One chapter deals extensively with neutron and X-ray solu- tion scattering techniques, and a choice of rapid reaction methods is discussed in a further chapter The use of electron microscopy has been another very important development in the biological sciences and the results are illustrated by a chapter with emphasis on biomembranes The New Comprehensive Biochemistry series contains a volume (8) devoted to separation methods This area is now supple- mented by a chapter in the present volume on high performance liquid chromatog- raphy of nucleic acids and a chapter on reversed phase HPLC of peptides and proteins The editors hope that the publication of this volume may serve the needs

of many biochemists and thus contribute to further research in the biological sciences

A Neuberger L.L.M van Deenen

J.R Lakowicz (Baltimore, MD, USA)

1 The phenomenon of fluorescence

2 Factors affecting the fluorescence emission

2.1 Solvent polarity and viscosity

2.2 Emission spectra of melittin

3 Time-resolved fluorescence spectroscopy

Resolution of the emission spectrum of liver alcohol dehydrogenase

Pulsed lasers for time-resolved fluorescence

Frequency-domain resolution of protein fluorescence

Anisotropy decays of protein fluorescence

4 Harmonic-content frequency-domain fluorometry

5 summary

Acknowledgements

References

Chapter 2

Raman and resonance Raman spectroscopy

P.R Carey (Ottawa, Ont., Canada)

1 Introduction

2 The units used in Raman spectroscopy

3 A model for Raman scattering based on classical physics

4 Raman and resonance Raman scattering: a quantum mechanical interpretation

5 Polarisation properties of Raman scattering

6 Basic experimental aspects

7 Raman studies on biological materials

7.1 Proteins

7.1.1 Amide I and amide I11 features

7.1.2 Side chain contributions to the Raman spectrum

7.2 Proteins containing a natural, visible chromophore

7.3 Resonance Raman labels

7.4 Nucleic acids

7.4.1 The purine and pyrimidine bases

7.4.2 Conformation of the (deoxy)ribose-phosphate backbone

7.4.3 Resonance Raman studies of nucleic acids

7.5 Viruses

7.6 Lipids and membranes

7.6.1 The C-C stretching region between 1050 and 1150 cm-'

7.6.2 The C-H stretching region between 2800 and 3000 cm-'

7.6.3 Deuterated lipids as selective probes

7.6.4 Lipid protein interactions and natural membranes

References

Chapter 3

Rapid reaction methods in biochemistry

Quentin H Gibson (Ithaca, NY, USA)

High performance liquid chromatography of nucleic acids

M Colpan and D Riesner (Dusseldorf, FRG)

1 Introduction

2 Techniques

2.1 Size exclusion chromatography

2.2 Anion-exchange chromatography

2.3 Reversed phase and hydrophobic interaction chromatography

2.4 RPC-5 and other mixed mode chromatography

2.5 Sample preparation and recovery

References 103

Chapter 5

Reversed phase high per,,rmance liquid chromatography of peptides and proteins

M.T.W Hearn and M.I Aguilar (Clayton, Vic., Australia) 107

111

120

126

2 Retention relationships of peptides in RP-HPLC

3 The relationship between peptide retention behaviour and hydrophobicity coefficients

4 Bandwidtb relationships of peptides in RP-HPLC

5 Dynamic models for interconverting systems

6 Conclusion

Acknowledgements

References

Chapter 6

X-ray and neutron solution scattering

S.J Perkins (London, UK)

1 Introduction

Part A: Theoretical and Practical Aspects

2 Theory of X-ray and neutron scattering

2.1.3 Scattering angles, vectors and resolution

The scattering event and the Debye equation

Scattering densities and allowance for solvent

2.3.1 Concept of scattering densities

2.3.2 Scattering densities and volumes

2.3.3 The contrast difference A p

2.3.4 Mean macromolecular scattering densities p

2.3.5 Scattering density fluctuations pF(r)

The Guinier plot: Z ( 0 ) and R ,

2.4.1 The innermost scattering curve

2.4.2 Cross-sectional and thickness Guinier analyses

Analyses of I ( 0 ) values

Analyses of R, values

Non-uniform scattering densities and contrast variation

2.7.1 The Stuhrmann plot

2.7.2 Solvent penetration and exchange effects

2.7.3 Isomorphous replacement

2.7.4 Matchpoints of multicomponent systems

Label triangulation

Wide-angle scattering and modelling strategies

2.9.1 Spheres and ellipsoids

2.9.2 Scattering curves at large Q

2.9.3 Independent parameters from scattering

2.9.4 Debye curve simulations

3 Experimental practice and instrumentation

3.1 Sample preparation and measurement

3.1.1 Sample monodispersity and concentrations

3.1.2 Sample assays

3.1.3 Sample backgrounds

3.1.4 Sample holders

3.1.5 Instrumental calibration

3.2 Labelling techniques and deuteration

3.3 Sources of X-rays and neutrons

3.3.1 Anode sources

3.3.2 Synchrotron radiation

3.3.3 Reactor neutron sources

3.3.4 Spallation neutron sources

3.4 Scattering instrumentation

3.4.1 X-ray cameras

3.4.2 Neutron cameras

3.5 Data reduction

Part B: Biochemical Applications to Proteins, Carbohydrates, Lipids and Nucleic Acids

4 Applications of X-ray and neutron scattering

4.1 Introduction

4.2 X-ray studies on globular proteins

4.2.1 Relationship between R , and M ,

4.2.2 Comparison of crystal and solution structures

4.2.3 Conformational changes and ligand binding

4.2.4 AUostericism

4.2.5 Molecular modelling of proteins

4.2.6 Associative systems and time-resolved synchrotron radiation studies

4.2.7 Interparticle interference

4.2.8 X-ray contrast variation and anomalous scattering

4.2.9 Label triangulation of heavy metal probes

Neutron studies on globular proteins

4.3.1 Contrast variation studies

4.3.2 Label triangulation and deuteration

X-ray and neutron studies on glycoproteins

4.4.1 Plasma glycoproteins, proteoglycans and polysaccharides

4.4.2 Immunoglobulins

4.4.3 Components of complement

Lipids, detergents, membrane proteins and lipoproteins

4.5.1 Lipid vesicles and complexes with proteins

4.5.2 Detergent micelles and complexes with proteins

4.5.3 Lipoproteins

4.6.1 DNA studies by X-ray scattering

4.6.2 X-ray and neutron studies on transfer RNA

4.6.3 Protein-nucleic acid interactions by neutron scattering

4.6.4 Chromatin and chromosomes by X-rays and neutrons

4.6.5 Ribosomes and their constituents

CHAPTER 1

Fluorescence spectroscopy; principles

JOSEPH R LAKOWICZ

University of Maryland at Baltimore School of Medicine, Department of Biological Chemistry,

660 West Redwood Street, Baltimore, M D 21201, USA

1 The phenomenon of fluorescence

Luminescence is the emission of photons from electronically excited states Luminescence is divided into two types, fluorescence and phosphorescence In phosphorescence, the emission is from an excited triplet state to a ground state singlet Since this transition is forbidden the rate of return to the ground state is slow, which means the decay times are long (msec to sec) Fluorescence is the emission from excited singlet states, also yielding a ground state singlet These allowed transitions to occur rapidly, with rates near l o 8 sec-' Consequently, the decay times for fluorescence are typically near lo-' sec or 10 nsec In this chapter

we will discuss primarily fluorescence, but the concepts are also applicable to events

on a slower timescale if the phosphorescence is observed The nanosecond timescale

of fluorescence provides much of its usefulness in biophysical chemistry In solu- tions near room temperature, a variety of molecular events can occur within 10 nsec and alter the emission These events include rotational diffusion, collisions with quenchers, solvent reorientation, and energy transfer These events alter one or more

of the spectral observables, and can thus be detected by analysis of the emission Substances which display fluorescence are generally delocalized aromatic systems with or without polar substituents (Fig 1) It is difficult to predict which molecules will be fluorescent or non-fluorescent because exceptions can usually be found However, several general rules are generally true Rigid molecules are usually more fluorescent, or at least their fluorescence more predictable, than molecules with the possibility of internal rotation Hence, perylene and anthracene fluoresce with high efficiencies, whereas stilbene can be much less efficient In viscous solvents, in which rotational reorientation to cis-stilbene cannot occur, trans-stilbene is highly fluo- rescent In non-viscous solution stilbene is only weakly fluorescent This illustrates

an important aspect of fluorescence, which is that the excited states are involved,

* Dedicated to Professor Gregorio Weber on the occasion of his seventieth birthday

Perylene Anfhracene trans- Stilbene

I ndole P PO 2 - N a p h t h o l

Fig 1 Typical fluorescent molecules

and these states have a different electronic distribution which may alter their chemical properties In the excited state trans-stilbene (Fig 1) isomerizes to the non-fluorescent cis-stilbene The altered electronic distribution can also alter chem- ical reactivity For instance, the pKa of the hydroxyl group on naphthol decreases from 9 to 2 upon excitation, presumably as the result of transfer of electron density from the oxygen into the aromatic ring The emission from an aqueous solution of naphthol can be due to unionized naphthol, naphtholate, or both, depending upon

pH and the concentration of basic species available to accept the dissociated proton The presence of substitutes for carbon in the aromatic system generally alters the emission from the aromatic nucleus Insertion of oxygen or nitrogen into the ring system often results in good fluorescence Hence, indole, fluorescein, PPO, the rhodamines and similar substances are fluorescent (Fig 1) The presence of sulfur, nitro groups or heavy atoms like iodide generally result in quenching of fluores- cence

Biological systems contain a variety of intrinsic (natural) fluorophores (Fig 2) In proteins, tryptophan is the most highly fluorescent amino acid, accounting for 90%

of the emission from most proteins Emission from tyrosine residues is also observed, especially in proteins lacking tryptophan, in denatured proteins, or in those with a high ratio of tyrosine to tryptophan Tyrosine is highly fluorescent in solution, but its emission is often quenched in native proteins, due either to the quenching effects of hydrogen bonding to the hydroxyl group or because of energy transfer from tyrosine to tryptophan The emission of phenylalanine from proteins

is less studied

The nucleotides and nucleic acids are generally non-fluorescent However, some notable exceptions are known Phenylalanine transfer RNA from yeast (tRNAPhe) contains a single highly fluorescent base, called the Y-base, whch has an emission

maximum near 470 nm The presence of this intrinsic fluorophore has resulted in

numerous studies of tRNAPhe by fluorescence spectroscopy Regarding the “non- fluorescent” nucleic acids, it should be noted that they do fluoresce, but with very low yields and with short decay times

Fig 2 Intrinsic biological fluorophores For NADH and FAD we only showed the fluorescent part of the molecule

Other natural fluorophores include NADH and FAD, whose fluorescent moieties

are shown in Fig 2 In both cases the amount of fluorescence depends upon their

local environments For instance, the emission of NADH is usually increased about three-fold upon binding to proteins, whereas the emission of FAD is usually quenched

In instances where nature has not provided an appropriate fluorophore, one can often add an extrinsic label The earliest probes include dansyl chloride [l] and ANS (Fig 3) Dansyl chloride can be covalently attached to macromolecules by reaction with amino groups ANS often binds spontaneously but non-covalently to proteins and membranes, probably by hydrophobic and electrostatic interactions The emis- sion of both molecules is sensitive to the polarity of the surrounding environment ANS is nearly non-fluorescent in water, but fluoresces strongly upon association with serum albumin, immunoglobulins and other proteins A wide variety of covalent and non-covalent probes are available [2,3]

Studies of cell membranes by fluorescence depends almost exclusively upon the use of extrinsic probes This is because most lipids are not fluorescent, and the emission from membrane-bound proteins is too heterogeneous for interpretation of the data The probe DPH (Fig 3) is typical of membrane probes, as is perylene (Fig 1) These non-polar molecules partition spontaneously into membranes And finally, Fig 3 shows several extrinsic probes for nucleic acids Addition of the etheno bridge

to ATP results in a highly fluorescent residue Unfortunately, this modification also disrupts the base pairing of the nucleotide Alternatively, nucleic acids can be labeled by spontaneous binding of planar cations such as ethidium bromide and acridine orange (Fig 3) Depending upon the structure, the fluorescence of the probe may be quenched or enhanced upon intercalation into DNA, and the emission may depend upon whether the intercalation site is adjacent to A-T or G-C pairs For instance, the fluorescence yield of ethidium bromide is enhanced about 30-fold upon intercalation into DNA Other intercalating dyes such as proflavin and 9-aminoacridine are quenched by their interactions with DNA [4]

2 Factors affecting the fluorescence emission

2 I Solvent polarity and viscosity

The variety of factors which can affect the fluorescence emission are illustrated by

the modified Jablonski diagram [5] shown in Fig 4 In this diagram we emphasize

fluorescence emission and quenching, and hence we have not included the higher electronic states or the triplet states Upon absorption of light the fluorophore arrives instantaneously in the first singlet state (Sl), usually with some excess

vibrational energy This excess energy is usually dissipated quickly in lo-'' sec by interaction with the solvent, resulting in a molecule in the lowest vibrational level of

S, The fluorophores remain at this level for the mean duration of the excited state,

which is typically 10 nsec Any process or interaction which occurs during this interval can alter the fluorescence emission These processes and interactions are the origin of much of the information available from fluorescence spectroscopy Fluorophores with polar groups are often sensitive to solvent polarity Interaction between the excited fluorophore and surrounding polar groups lowers the energy of the excited state, which shifts the emission to longer wavelengths The relative amounts of emission from the relaxed and the unrelaxed states depend upon the

\j S L a R e l a Y

Donor Emission

k, Energy Transfer

Acceptor absorption

X k i l h a

s o

Fig 4 Jablonski diagram for fluorescence emission and quenching

relative rates of depopulation of the excited state ( r + Cki) and that of solvent relaxation ( k R ) The rate of emission is r and the rate of return to the ground state exclusive of emission is Cki In fluid solvents near room temperature the rate of solvent reorientation is near 10-1'-10-12 sec Hence, this process is mostly com- plete prior to emission, so the observed emission is that of the relaxed state If the solution is cold or viscous, or if the probe is bound to a rigid site on a macromole- cule, then the rates of relaxation and emission can be comparable, so emission is seen from both the relaxed and the unrelaxed states If the solution is very viscous then solvent relaxation does not occur during the lifetime of the excited state, and the observed emission is from the higher energy (shorter wavelength) unrelaxed state

The effects of solvent polarity are best understood by specific examples To model the fluorescence emission of proteins we examine spectra for N-acetyl-L- tryptophanamide (NATA) This molecule is analogous to tryptophan in proteins It

is a neutral molecule, and its emission is more homogeneous than that of tryptophan itself In solvents of increasing polarity the emission spectra shift towards longer wavelengths (Fig 5) The emission maxima of NATA in dioxane, ethanol and water are 333, 344 and 357, respectively These solvents are non-viscous, so the emission is dominantly from the relaxed state (Fig 4) The spectral shifts can be used to calculate the change in dipole moment which occurs upon excitation [6] More typically, the emission spectrum for a sample is compared with that found for the same fluorophore in various solvents, and the environment judged as polar or non-polar While this approach is qualitative, it is simple and reliable, and does not involve the use of theoretical models or complex calculations

The timescale of the relaxation process also affects the emission This effect is illustrated for NATA in propylene glycol (Fig 6) At room temperature the

relaxation is mostly complete, and a red shifted spectrum is observed (348 nm) Lowering the temperature results in a progressive shift of the emission to shorter wavelengths, with an emission maximum of 329 nm at - 60 O C As the temperature

is decreasing the relaxation rate ( k R ) becomes slow relative to that of the decay rate

( r + Cki) Hence, an increasing proportion of the emission is from the unrelaxed and the intermediate states (Fig 4), which have higher energies and shorter emission

WAVELENGTH ,(nanometers)

Fig 5 Emission spectra of N-acetyl-L-tryptophanamide in various solvents of different polarities

wavelengths (Fig 6) In proteins it is probable that the emission maxima are affected by both the average environment of the tryptophan residues [7] and by the relaxation rates [8] Rather detailed data and analysis are needed for an unambigu- ous separation of these effects, but the average environment of the tryptophan residues seems to be the dominant determinant of the emission maxima

2.2 Emission spectra of melittin

Melittin is an amphipathic peptide component of bee venom which associates with cell membranes, enhances the phospholipase activity of venom and participates in the disruption of cell membranes This protein has been studied extensively by

WAVELENGTH [nanometers)

Fig 6 Emission spectra of in propylene glycol at 25 and 60 C

Fig 7 Emission spectra of melittin in the absence (- ) and presence (- - -) of 2 M NaCl The data are courtesy of N Joshi

fluorescence and other physical methods [9,10], and its X-ray structure is known [ll] In solution melittin exists as either a monomer or a self-associated tetramer

The self-association is driven by high salt concentration, which apparently shields the positive charges on the monomer from each other and allows the hydrophobic interactions to cause association The monomeric form of melittin is thought to be largely random coil with a high degree of segmental mobility In the tetrameric state the monomeric units are mostly a-helical

Melittin is an ideal protein to illustrate the effects of structure on the fluores- cence spectral properties Each monomer contains a singly tryptophan residue and

no tyrosine residues The X-ray structure of the tetrameric form shows that the tryptophan residues are buried in a non-polar pocket and are not directly exposed

to the aqueous phase

The emission spectra of melittin illustrate the effects of solvent exposure on the tryptophan emission (Fig 7) In the absence of salt, the emission maximum of 360

nm is comparable to that found for NATA in water In the presence of 2 M NaCl

the emission maximum is blue shifted by 12 nm to 348 nm This shift is a result of

shielding of the indole ring from the aqueous phase Hence, solvent relaxation proceeds to a lesser extent because there is less solvent available for interaction with the fluorophore

2.3 Quenching of fluorescence

Collisional quenching of fluorescence requires contact between the fluorophore and the quencher For quenching to occur the quencher must diffuse to and collide with the fluorophore in the excited state If this occurs the fluorophore returns to the

ground state without emission of a photon (Fig 4) Many small molecules act as

collisional quenchers of fluorescence [6,12] These include iodide, acrylamide,

halogenated hydrocarbons and occasionally amines and metal ions The excited state lifetimes provide ample opportunity for quenching For instance, acrylamide is known to be an efficient quencher of tryptophan fluorescence [12,13] Su.ppose its

diffusion coefficient is lop5 cm2/sec In 10 nsec an acrylamide molecule can diffuse

a distance of 44 A, as calculated using A x 2 = 2 0 7 , where A x is the distance, D is the diffusion coefficient and 7 is the fluorescence lifetime This distance is compara- ble to the diameter of many proteins Hence, we expect quenching to occur to a measurable degree and the extent of quenching to be sensitive to the average degree

of exposure of the tryptophan to the aqueous phase

Once again melittin illustrates the effect of protein structure on the fluorescence emission Acrylamide quenching data for melittin monomer and tetramer are shown

in Fig 8 Stern-Volmer plots are often used to present quenchmg data The Stern-Volmer equation is

FO

- = 1 + k.,[Q] = 1 + K [ Q ]

F

where Fo and F are the fluorescence intensities in the absence and presence of

quenching, respectively, T~ is the fluorescence lifetime in the absence of quenchmg,

[ Q ] is the concentration of quencher, k is the bimolecular quenching constant, and

K is the Stern-Volmer quenching constant The lifetime T,, is the reciprocal of the rates which depopulate the excited state From Fig 4,

If every collisional event results in quenching, the bimolecular rate constant can be estimated using the diffusion constants of the fluorophore ( D F ) and quencher ( DQ)

and the radius expected for contact ( R ) ,

where N is Avogadro's number and D = D, + DQ If the fluorophore is exposed to

the solvent we expect k to be near 0.5 X lo1' M-' sec-' If the residue is shielded

from collisional encounters this rate will be smaller This comparison is the basis for estimating the extent of exposure from quenching data The quenching constant measured for a protein is compared with that expected for a completely exposed fluorophore Typically, model compounds with no possibilities for shielding are studied to account for lack of precise knowledge of diffusion coefficients, and the possibility that the quenching encounter is not 100% efficient

The quenching data for both the monomeric and tetrameric forms of melittin indicate the tryptophan residues are accessible to acrylamide with the accessibility being greater in the monomeric state This conclusion is reached by comparison

with acrylamide quenching data for NATA At 25°C in water the acrylamide

quenching constant for NATA is 0.58 X 10" M-' sec-' [47] For the monomer the quenching constant is about one-third of this value, which is indicative of a rather fully exposed residue [13] The value of k for the tetramer is less, indicating

shielding of the tryptophan residue from the aqueous phase It should be noted that the relative shielding is only 40%, which probably indicates considerable penetration

of the tetramer by acrylamide In other more extensive studies Eftink and Ghiron showed that acrylamide quenching reflects the average degree of tryptophan ex- posure to the aqueous phase [13] The penetration of proteins by quenchers has been known for some time [14,15] For melittin tetramer the penetration by acrylamide is not unexpected since acrylamide is neutral and the tryptophans are located in a loosely packed non-polar region of the protein [ll]

2.4 Fluorescence energy transfer

Another process which can occur during the excited state is fluorescence energy transfer, which is the transfer of the excited state energy from a donor (D) to an acceptor (A) (Fig 4) The transfer is called radiation-less because it occurs without the appearance of a photon This process is strongly dependent upon distance because it is the result of dipole-dipole coupling between the donor and the acceptor [16] A requirement for energy transfer is that the emission spectrum of the donor overlaps with the absorption spectrum of the acceptor The rate of transfer ( k T ) is gven by

(4)

where R , is the distance at which 50% of the energy is transferred, and r is the

donor-to-acceptor distance The value of R o can be calcuated from the spectral

properties of donor and acceptor [6,16] The efficiency of energy transfer is given by the ratio of the rates of transfer to the total rate of depopulation of the donor Hence,

Usually, both the transfer efficiency ( E ) and R , are determined experimentally

Then, the donor-to-acceptor distance is calculated using

This method is widely used to measure the distance between sites on a macromole- cule, and has been the subject of considerable experimentation and discussion Energy transfer has been used to measure the self-association of melittin The melittin was labeled with a N-methylanthraniloyl (NMA) residue on one of the lysine residues This fluorophore serves as the energy acceptor for the single

tryptophan residue Only a small fraction (5%) of the melittin monomers was

labeled with NMA In the monomer there is only one tryptophan residue near the acceptor, whereas four such residues are present in the tetramer Hence, the extent

of tryptophan to NMA energy transfer should be sensitive to and increased by

melittin self-association In this experiment the intention is not to determine a distance, but rather to use the association-dependent energy transfer to determine the extent of self-association [20]

[ 17- 191

WAVELENGTH (nanometers)

Fig 9 Emission spectra of N-methylanthraniloyl-labeled melittin Spectra are shown for the monomer (0

M NaC1) and for the tetramer (2 M NaCI) From [20] The broken lines are the emission spectra of the unlabeled melittin

250 300 350 400

WAVELENGTH (nanometers 1

Fig 10 Excitation spectra of NMA-labeled melittin From [20]

Emission spectra of the labeled melittin are shown in Fig 9 Recall that only a small fraction of the melittin contains a NMA label Hence, the emission spectra are mostly characteristic of tryptophan, with shoulders at 430 nm due to the NMA emission In the presence of 2 M NaCl the NMA emission is enhanced, reflecting increased energy transfer from the additional tryptophan residues

Excitation spectra are often used to study energy transfer This is because energy transfer can be detected by enhanced emission from the acceptor when the excita- tion is centered at the donor absorption The effects of melittin self-association are evident from the excitation spectra (Fig 10) For these spectra the emission monochromator is centered on the NMA emission (430 nm) and the intensity recorded as the excitation monochromator is scanned through the absorption bands

of the NMA label (350 nm) and the tryptophan absorption (280 nm) Increasing salt concentrations result in increased intensity of the tryptophan excitation band (280 nm) This increase in energy transfer is due to the close proximity of the three additional donors to the NMA acceptor

2.5 Fluorescence anisotropy

The timescale of fluorescence emission is comparable to that of rotational diffusion

of proteins and the timescale of segmental motions of protein domains or individual amino acid residues The polarization or anisotropy of the emission provides a measure of these processes Suppose a sample is excited with vertically polarized light (Fig ll), and that the sample is viscous so that the fluorophores do not rotate during the lifetime of the excited state Then the emission is polarized, usually also

in the vertical direction This polarization occurs because the polarized excitation selectively excites those fluorophores in the isotropic solution whose absorption

m+

LIGHT SOURCE-

f /

DETEC~OR

Fig 11 Measurement of fluorescence anisotropies

moments are aligned vertically The extent of polarization is most conveniently defined by the anisotropy [6,40]

where I refers to the intensities, and the subscripts indicate the parallel (11) or perpendicular (I) component

A number of processes can result in the loss of anisotropy, the most common being rotational diffusion Melittin is expected to have rotational correlation times

near 2 and 8 nsec in the monomeric and tetrameric states, respectively The effect of

rotational diffusion on the anisotropy is described by the Perrin equation,

where r, is the anisotropy in the absence of rotational diffusion, r is the anisotropy,

7, is the lifetime and 6 is the correlation time The value of r, is usually measured in

a separate low-temperature experiment Its value depends upon the excitation

wavelength, and is typically in the range of 0.1 to 0.4 The r, value is a measure of

the angle between the absorption and emission transition moments of the fluoro-

phore For tryptophan the value of ro on the long wavelength side of the absorption

is near 0.32 Values of r, which are less than 0.1 are usually not useful because the

difference between 11, and I* will be small, and the precision of the measurements

will be decreased

When 7, and 8 are of similar magnitude then the measured anisotropy is dependent upon the correlation time Self-association of melittin is expected to increase its correlation time about four-fold Since the lifetime of melittin fluo-

resence is near 3 nsec we expect self-association to have a dramatic effect on the

old0 ' 0 4 ' ' 0.8 ' ' I .2 ' ' 1.6 ' ' 2.0 I '

[NaCII M

Fig 12 Fluorescence anisotropies of melittin [20] Anisotropies are shown both for the intrinsic tryptophan emission (A), and for that of the NMA label (0) Also shown is the effect of salt concentration on the extent of energy transfer (0)

anisotropy Anisotropy data for melittin at various salt concentrations are shown in

Fig 12 As the salt concentration is increased the anisotropy values increase and reach a plateau whch is characteristic of the tetramer Also shown in Fig 12 are the

anisotropy values of the NMA label These also increase with salt concentration, and reach a constant value at 1 M NaC1 The NMA anisotropy values are lower

than for tryptophan because the decay time of the NMA label is longer, near 8 nsec For comparison this figure also shows the extent of energy transfer All three measurements reflect the monomer to tetramer transition

Anisotropy measurements are generally useful for measuring any process which increases or decreases the rate or extent of rotational diffusion These processes

include domain motions of immunoglobulins [21], denaturation of proteins [22] and the association of proteins with membranes [lo] Additionally, there are numerous

applications of anisotropy measurements to membranes, in which the phase state and apparent fluidity are estimated from the anisotropy of probes which are bound

to the membranes [23,24]

3 Time-resolved fluorescence spectroscopy

The previous discussion and examples emphasized the use of steady-state fluores- cence data Steady-state data are measured with constant illumination of the sample The timescale of these measurements is slow relative to the fluorescence decay times Hence, the effects of the time-dependent processes are averaged to yield the average emission spectra, anisotropies, or extents of energy transfer Each measured steady-state quantity is the average of the time-dependent values of that quantity averaged over the time-dependent decay of the sample For instance, the

steady-state anisotropy is determined by its time-resolved decay ( r ( t )) and the time-dependent decay of the emission ( Z ( t ) )

If we assume that both r ( t ) and I ( t ) decay as single exponentials with time

constants of 1/8 and 1/~,,, respectively, then application of equation 9 yields the steady-state form, whch is the Perrin equation (8)

At present, there is widespread interest in directly measuring the time-dependent processes This is because considerably more information is available from the time-dependent data For example, the time-dependent decays of protein fluores- cence can occasionally be used to recover the emission spectra of individual tryptophan residues in a protein The time-resolved anisotropies can reveal the shape of the protein and/or the extent of residue mobility within the protein The time-resolved energy transfer can reveal not only the distance between the donor and acceptor, but also the distribution of these distances The acquisition of such detailed information requires high resolution instrumentation and careful data acquisition and analysis

There are presently two methods of obtaining the time-resolved data These are

by direct measurements in the time-domain [25,26] and by less direct measurements

in the frequency-domain [27,28] For time-domain measurements the sample is excited with a brief pulse of light (Fig 13) The time-dependent fluorescence intensities are used to estimate the decay time(s) of the sample In the frequency-do- main the sample is excited with intensity modulated light The frequency response (phase and modulation) of the sample are used to estimate the decay time(s) Both methods are rapidly evolving to take advantage of the increased time resolution obtainable using picosecond pulse lasers and faster detectors [29,30] The complex equipment and analyses necessary for time-resolved measurements has been the subject of numerous publications and monographs [25,26] In this article we will not describe the instrumentation, but will rather describe the results and their interpre- tation

The objective of either time or frequency-domain fluorometry is to determine the decay law of the sample For example, consider protein containing two tryptophan residues, and assume further that each residue has a single decay time The impulse response of the sample is the decay which would be observed with an ideal instrument following excitation with a &function light pulse For our hypothetical protein we expect a doubly exponential decay of intensity,

2

I( t ) = C aie-'/'~

i = l

In this expression the 7; values are the decay times of the individual residues and ai

values are the preexponential factors The contribution of each residue to the emission is

Suppose the data are measured at a number of wavelengths across the emission spectrum Then the data are described by

3.1 Resolution of the emission spectrum of liver alcohol dehydrogenase

The resolution and understanding of the emission from proteins is a difficult task This is because most proteins contain two or more tryptophan residues, and even

Fig 14 Time-resolved fluorescence intensity of liver alcohol dehydrogenase Left: single exponential fit,

x i = 3.02; right: double exponential fit, x i = 1.14 The faster decaying curves are the lamp profiles found using scattered light The upper panels are on a semilogarithmic scale, and the lower are on a linear scale The lower panels show the deviations, and the autocorrelation of the deviations as insets

the emission from proteins containing a single tryptophan is complex or multi-ex- ponential One of the most detailed studies is for liver alcohol dehydrogenase

(LADH) by Brand and co-workers [7] LADH consists of two identical subunits (dimeric) with a molecular weight of 80,000 daltons Each subunit contains two tryptophan residues, trp-15 and trp-314 One residue (trp-15) is exposed to the solvent, while t h s other (trp-314) is in a hydrophobic pocket at the dimer interface Time-resolved data for LADH are shown in Fig 14 These data for a single

emission wavelength, were obtained by the method of time-correlated single photon

counting [25,26] The light source is a flash lamp which fires repetitively at rates near 20 kHz The time-resolved decay is obtained by measuring the time interval between the lamp pulse and the first emitted photon The rate of detecting the emitted photons is kept near 2% of the lamp rate In this way the first photon represents a sampling of the entire time-resolved decay If many photons reach the detector, and only the first one is counted, then the measured decay does not represent the true decay The electronic circuits (in current use) which detect the arriving photons cannot accept and process multi-photon events, which limits the rate of data acquisition A second difficulty is the width of the lamp pulse Flash lamps are inexpensive, but the pulse widths are near 2 nsec, whch is not much shorter than the decay times of proteins Consequently, mathematical procedures, known as deconvolution or reconvolution, are used to account for the lamp pulse width Nonetheless, the necessary procedures are highly refined and reliable

The data for LADH (Fig 14) illustrate how the time-resolved data are analyzed The analysis starts by assuming the decay is a single exponential (equation 1 with

i = l), and calculation of the best fit This fitting procedure involved use of the lamp profile and a guess of the decay time to predict the data The numerical value of the single decay time is varied until the best fit is found Frequently, a single decay time model is not adequate, as was found for LADH The inadequacy of the fit is revealed by systematic deviations between the measured and calculated data (Fig

14, lower left), and an unacceptably high value for the goodness-of-fit parameter

xi = 3.02 For an acceptable fit xi should be near unity Greater values of xi can

be the result of either an inadequate model or systematic errors in the data The acceptable values of xi are determined, in principle, by statistical criteria [41], but individual judgements are necessary

The next step in the analysis is to fit the data using a more complex model The best fit for LADH for two decay times yields an improved match (Fig 14, right) The calculated and measured values are now in agreement, the deviations are small and random, and xi = 1.14 is acceptably close to unity The parameters (a, and 7,)

which yield this match are taken as the decay law of the sample The decay times (3.8 and 7.2 nsec) were taken to be due to tryptophan 314 and 15, respectively It must be emphasized that this result does not prove the decay is a double exponen- tial, but only shows that the double exponential model is adequate to explain the data If data were available with higher time resolution or statistical accuracy (more photons) then it may be necessary to use a more complex three-exponential decay model to explain the data

To resolve the emission spectra of each residue similar data were collected at closely spaced wavelengths across the emission The assumption was made that the decay time of each tryptophan residue was independent of emission wavelength, and constant across the emission The measured decay law was used to calculate the individual spectra using

3.2 Pulsed lasers for time-resolved fluorescence

The time-resolved emission of LADH was found to be complex due to the presence

of two tryptophans It is highly probable that the results in Fig 15 are only approximate, and that the decay of each tryptophan is multi-exponential Such complexity is hidden from view by the limited resolution of the available instrumen- tation The resolution of more complex decays for LADH or any protein requires higher resolution This is being accomplished by the use of pulsed laser sources One

of the most popular sources is shown in Fig 16 The main pump is shown as an Ar-ion laser, but the Nd-YAG lasers are used in a similar manner Normally, these lasers yield continuous output In this case the argon laser is mode-locked, which yields a repetitive train of pulses These pulses are spaced at the time interval for light to travel the length of the laser cavity, which is 83 MHz The argon laser output is then used to pump a dye laser, whose cavity length is identical to that of

the Argon laser This results in a 83 MHz train of pulses in the dye laser cavity

which has pulse widths near 5 psec These pulses are extracted at a slower rate using

a cavity dumper The cavity dumper is a small piece of quartz into which a time R F pulse is launched via a piezoelectric device This RF pulse sets up a diffraction grating which deflects the desired laser pulse The extracted pulse train is then

u

I I I

83 MHz

D Y E LASER CAVITY D U M P E R

285 - 305 nm

DOUBLER

5 p s Fourier - b 4MHz Transform 0

a

T i m e

Frequency (GHz)

Fig 16 Pulsed laser source for time-resolved fluorescence The lower panel shows the pulse train from

the laser source and its Fourier transform

frequency-doubled to yield the UV necessary to excite protein fluorescence Ad- ditional details are available in other books on this topic [42,43]

There are several significant advantages to the laser source over the more conventional flash lamp The laser source is more intense Its pulse width is near 5 psec, as compared for 2000 psec for a flash lamp And finally, the repetition rate can

be much higher, typically 800 kHz to 4 MHz Because of all these factors it is

possible to rapidly acquire data to a much higher level of statistical accuracy than with a flash lamp For example, a recent paper by Small and co-workers describes a multi-component resolution of a histone, which contains a single tyrosine residue [31] Because of the substantial increases in resolution, the laser sources are becoming more widely used in the biochemical applications of fluorescence, as illustrated by recent studies of the tryptophan emission from phospholipase A, [44] and hemoglobin [45]

3.3 Frequency-domain resolution of protein fluorescence

Multi-exponential decays of fluorescence can also be recovered by measurements in the frequency-domain This has only become practical within the past four years [27,28] The resolution of multi-exponential decays requires measurements over a wide range of light modulation frequencies Earlier instrumentation could operate at only one to three frequencies, and these limited data were not adequate to determine the four parameters in equation 10 (two a, and two T,) The new instruments

Fig 17 Frequency-domain data for the intrinsic fluorescence of S, Nuclease and melittin

operate over a wide range, typically 1 to 200 MHz, and some instruments now operate to 2000 MHz [37]

Typical frequency-domain data for two proteins are shown in Fig 17 The data consist of the phase angles and modulation, each measured over the widest possible range of frequencies This requirement illustrates the transform relationship between the time and the frequency-domain measurements In the time-domain, the most desirable excitation profile is the shortest obtainable pulse The Fourier transform

of a &function consists of all frequencies Hence, the experimental requirements are similar, short pulses or wide range frequencies For each protein (Fig 17) the phase angle increases and the modulation decreases as the frequency increases The data are analyzed in a manner analogous to the time-domain data That is, a decay law is assumed and the parameters varied until the best possible match is obtained between the measured ( 0 ) and calculated (-, - - -) values The adequacy of the fit is judged by the value of the xi, which is a weighted sum of the squared deviations between the measured and the calculated values If the value of xi is significantly larger than unity then the model can be rejected

Both melittin and S, Nuclease contain a single tryptophan residue The data illustrated the point that the emission from such simple proteins can be multi- exponential Even though only a single residue is responsible for the emission, it was not possible to fit the data using a single decay time This is shown by the failure of the single decay time model (- - -) to explain the data The decays are referred

to as being heterogeneous The decays of both nuclease and melittin are signifi-

cantly heterogeneous The decay of melittin is more heterogeneous as seen by the greater deviations of the data from the one decay time model, and the larger value

of xk For both proteins the data can be explained by more complex decay laws Two decay times are needed to fit the data for nuclease, and three are needed for melit tin

It is important to exercise clear thinking when fluorescence data in either domain are fitted to various models A poor fit can be used to reject a model The poor fit can be due to either an inadequate model or due to systematic errors in the data not known to the researcher If systematic errors occur a more complex model could be accepted to account for the errors, not because the model is appropriate for the sample Secondly, a good fit does not prove the model which yields the good fit is correct A good fit only shows that the model is adequate to explain the data Alternatively stated, the data which yield the good fit are not adequate to support a more complex model

What is the origin of the multi-exponential decays found for single tryptophan proteins? Surprisingly, this is an unanswered question which is the focus of current research Clearly, the protein matrix provides a unique but asymmetric environment for each tryptophan residue The spectral properties (emission maximum, lifetimes, anisotropy and yield) are determined by this environment The protein environ- ments can now be conveniently examined using the X-ray coordinates and modem computer graphics, Additionally, it is becoming increasingly easy to replace individ- ual amino acid residues using the techniques of molecular biology These capabili- ties, and the increased resolution available from state-of-the-art instrumentation, should allow the linkage to be established between structural data and fluorescence spectral parameters

3.4 Anisotropy decays of protein fluorescence

There is considerable interest in measuring the rotational diffusion and the dynamic properties of proteins The rates of rotation diffusion can reveal the size and shape

of the protein Also, proteins are known to undergo structural fluctuations, a topic which has been broadly studied by both experimentation and computer simulations

[32-361 The time-resolved experiments are often directed towards a comparison of the measurable dynamics of proteins with the calculated dynamics One promising approach is to use anisotropy data from intrinsic protein fluorescence If such data are available with picosecond resolution then such a comparison should be possible

In the time-domain the anisotropy decay is obtained from the time-resolved decays of the parallel and perpendicular polarized components of the emission More specifically, one measures the time-resolved decays of the parallel ( I and the perpendicular ( I ) components of the emission, and calculates the time-resolved anisotrop y,

Generally, the anisotropy decay is multi-exponential

In this expression r( t ) is the time-dependent anisotropy, Bi the correlation times

and g, the fraction of the total anisotropy ( r o ) whch decays with this correlation

time In general we expect one component (8,) due to rotational diffusion of the protein, and one due to torsional motions of the tryptophan residue, if such motions are significant In proteins which contain more than a single fluorescent residue there can be energy transfer among the residues, whch can appear as a component

in the anisotropy decay The timescale of energy transfer depends upon the distance and orientation between the residues, but there is little information on the timescale

of energy transfer between intrinsic fluorophores in proteins

The measurements are different in the frequency-domain In this case we measure the phase shift between the parallel and perpendicular components of the emission, and a frequency-dependent anisotropy, which is analogous to the steady state anisotropy These two types of data are used to determine the decay law for the anisotropy (equation 15)

Melittin and S, Nuclease illustrate how the anisotropy decay is reflected in the

frequency-domain data From earlier studies it was known that the single tryptophan residue in S, Nuclease was mostly rigid [36], so that its anisotropy decay should display a single correlation time for rotational diffusion near 11 nsec In contrast, melittin monomer is thought to be disordered in aqueous solution, so that a rapid anisotropy decay is expected due to local tryptophan motions

The frequency-domain anisotropy data for nuclease and melittin are shown in Fig 18 The data for nuclease are nearly Lorentzian and centered near 30 MHz, which is expected for a single correlation time near 11 nsec In contrast, the differential phase data for melittin show no such maximum, and the phase angles increase up to the 200 MHz limit This is characteristic of a subnanosecond anisotropy decay

For both proteins a multi-exponential anisotropy decay was necessary to explain the data, and in both cases a short correlation time ( < 1 nsec) was indicated by the data In the case of nuclease only 12% of the anisotropy decays by this rapid process, indicating that the torsional motions have a limited amplitude In contrast, 75% of the melittin anisotropy decays by the rapid process, which indicates considerable free motion of the tryptophan residue

To illustrate the nature of the anisotropy decays the equivalent time-dependent anisotropies are shown as an insert These were calculated from the frequency-do- main data For S, Nuclease the plot of log r ( t ) versus time is mostly linear with a

slope of (12 nsec)-' This is the portion of the anisotropy decay due to overall rotational diffusion of the protein The rapid component in the nuclease anisotropy decay is seen only near the t = 0 origin The anisotropy decay of melittin is much more rapid, which reflects the greater motional freedom of the tryptophan residue in this disordered polypeptide Because of the segmental motions which depolarize the

O 5 10 20 50 100 200 f

2

FREQUENCY (MHz)

Fig 18 Frequency-domain resolution of the anisotropy decay of S, Nuclease and melittin monomer

Melittin: r( t ) = 0.24 exp (- t / 0 2 6 ) + 0.08 exp ( - r/3.04) Nuclease: r( t ) = 0.04 exp ( - r/0.20) + 0.28 exp ( - r/12.18)

emission the rotational diffusion of melittin is barely evident in the anisotropy decay

4 Harmonic-content frequency -domain j7uorometry

The frequency-domain data for melittin (Fig 18) revealed the need for still higher modulation frequencies Resolution of the anisotropy decay parameters is decreased

if the phase angle maximum is not reached This is perhaps analogous to the data

obtained with flash lamps (Fig 14), in which the width of the pulse was comparable

to the decay times of the emission Very recently, this laboratory developed a hybrid instrument which uses components typical of both time-domain and frequency-do-

main fluorometers [37] The instrument uses a 4 MHz train of 5 psec pulses from a

cavity-dumped dye laser, whch is the same source as is used for time-correlated single photon counting (Fig 16) However, the pulses are not used to perform time-domain measurements The pulse train possesses intrinsic modulation to many

GHz, which is shown by the Fourier transform in Fig 16 This source can be used directly as the modulated light source, an idea proposed originally for pulsed laser

excitation by Merkelo et al [46] and expanded to use the higher harmonics of

pulsed synchrotron radiation by Gratton and Degado [48,49] When used with a fast detector the frequency-domain measurements now extend to 2 GHz [37]

This development is recent, and harmonic-content data are not yet available for the proteins described above However, the potential of the 2 GHz measurements is illustrated by the data for oxytocin, which is a cyclic nona-peptide containing a single tyrosine residue The frequency-response for the intensity decay is shown in Fig 19 The mean decay time for oxytocin is near 0.7 nsec Even with this short decay time the entire frequency response was measured, as seen by phase angles which extend to 70" and modulations which decrease to 20% These data are adequate to support a three exponential analysis The apparent decay times are 80,

359 and 927 psec Once again, we find that even the decay of a single tyrosine residue can be complex

The 2 GHz data provide resolution of complex anisotropy decays on the picosecond timescale Data for the anisotropy decay of oxytocin are shown in Fig

20 It was not possible to fit the data using a single correlation time (- - -,

xi = 292) In contrast, a two-correlation time model provides a good fit, which is not improved by the use of a third correlation time We believe the correlation times

of 29 and 454 psec reflect local tyrosyl motions and overall rotational diffusion, respectively It is important to note that the measurements to 2 GHz provide considerable information beyond the data to the previous 200 MHz limit Data to

200 MHz would not display the shoulder seen at 600 MHz, which represents the transition from rotational diffusion to segmental motions

5 Summary

The phenomenon of fluorescence can provide information about the physical properties of proteins and other macromolecules The information content results from the sensitivity of the spectral properties to the average and dynamic properties

of the environment surrounding the fluorescent residues In general, more detailed information is obtainable from time-resolved data than from steady-state measure- ments However, the steady-state measurements are considerably easier to perform

At present, the ability to recover time-resolved spectral data is rapidly improving, primarily because of advances in instrument design The newer instruments may possess resolution adequate to correlate experimental data with the structural or dynamic properties of macromolecules

Acknowledgements

This work was supported by grant DAAG29-85-G-0017 from the Army Research Office, grants DMB-8511065 and DMB-08502835 from the National Science Foundation and grants GM-39617 and GM-29318 from the National Institutes of Health The author wishes to especially acknowledge the National Science Founda- tion for providing support to construct the frequency-domain fluorometer, at a time when there were doubts about their design and usefulness

References

1 Weber, G (1952) Biochem J 51, 155-167

2 Haugland, R.P (1983) In: Excited States of Biopolymers (R.F Steiner, ed.) pp 29-58, Plenum Press,

3 Beddard, G.S and M.R West, eds (1981) Fluorescent Probes Academic Press, New York

4 Steiner, R.F and Y Kubota (1983) In: Excited States of Biopolymers (R.F Steiner, ed.) Plenum

5 Jablonski, A (1935) Z Phys 94, 38-46

6 Lakowicz, J.R (1983) Principles of Fluorescence Spectroscopy Plenum Press, New York

7 Ross, J.A.B., C.J Schmidt and L Brand (1981) Biochemistry 20, 4369-4377

8 Lakowicz, J.R and H Cherek (1980) J Biol Chem 255, 831-834

9 Talbot, J.C., J Dufourcq, J DeBong, J.F Faucon and C Lussan (1979) FEBS Lett 102, 191-193

New York

Press, New York

10 Faucon, J.F., J Dufourcq and C Lussan (1979) FEBS Lett 102, 187-190

11 Terwillinger, T.C and D Eisenberg (1982) J Biol Chem 257, 6016-6022

12 Eftink, M.R and C.A Ghiron (1981) Anal Biochem 144, 199-227

13 Eftink, M.R and C.A Ghiron (1976) Biochemistry 15,672-680

14 Lakowicz, J.R and G Weber (1973) Biochemistry 12,4161-4170

15 Lakowicz, J.R and G Weber (1973) Biochemistry 12, 4171-4179

16 Forster, Th (1948) Ann Phys (Leipzig) 2, 55-77 Translated by R.S Knox

17 Stryer, L (1978) Annu Rev Biochem 47, 819-846

18 Stryer, L and R.P Haugland (1967) Proc Natl Acad Sci USA 58, 719-726

19 Dale, R.E., J Eisinger and W.E Blumberg (1979) Biophys J 26, 161-194

20 Hermetter, A and J.R Lakowicz (1986) J Biol Chem 261, 8243-8248

21 Hanson, D.C., Y Yguerabide and V.N Schumaker (1981) Biochemistry 20, 6842-6852

22 Brand, J.C and R.H Cagen (1977) Biochim Biophys Acta 497, 178-187

23 Dale, R.E., L.A Chen and L Brand (1977) J Biol Chem 252, 2163-2169

24 Kawato, S., K Kinosita and A Ikegami (1977) Biochemistry 16, 2319-2324

25 OConnor, D.V and D Phillips (1984) Time-Correlated Single Photon Counting Academic Press,

26 Demas, J.N (1983) Excited State Lifetime Measurements Academic Press, New York

27 Lakowicz, J.R and B.P Maliwal (1985) Biophys Chem 21, 61-78

28 Gratton, E and M Limkemann (1983) Biophys J 44, 315-324

29 Visser, A.J.W.G., ed (1985) Anal Instrum 14, 193-566

30 Lakowicz, J.R., G Laczko and I Gryczynski (1986) Rev Sci Instrum 57, 2499-2506

31 Libertini, L.J and E.W Small (1985) Biophys J 47, 765-772

32 Gurd, F.G and T.M Rothgeb (1979) Adv Prot Chem 33, 73-165

33 Karplus, M and J.A McCammon (1981) CRC Crit Rev Biochem 9, 243-249

34 Lakowicz, J.R (1986) Methods in Enzymology, in press

35 Lakowicz, J.R., G Laczko, I Gryczynski and H Cherek (1986) J Biol Chem 261, 2240-2245

36 Munro, I., I Pecht and L Stryer (1979) Proc Natl Acad Sci USA 76, 56-60

37 Lakowicz, J.R., G Laczko and I Gryczynski (1986) Rev Sci Instrum 57, 2499-2506

38 Lakowicz, J.R., G Laczko and I Gryczynski (1986) Biophys Chem 24, 97-100

39 Lakowicz, J.R and Balter, A (1982) Photochem Photobiol 36, 125-132

40 Jablonski, A (1960) Bull Acad Pol Sci 8, 259-264

41 Bevington, P.R (1969) Data Reduction and Error Analysis for the Physical Sciences McGraw-Hill,

42 Svelto, 0 (1982) Principles of Lasers Plenum Press, New York

43 Visser, A.J.W.G., ed (1985) Analytical Instrumentation 14 Marcell Dekker, New York

44 Ludescher, R.D., Volwek, J.R., DeHaas, G.H and Hudson, B.S (1985) Biochemistry 24, 7240-7249

45 Albana, J., Albert, B., Krajcarski, D.T and Szabo, A.G (1985) FEBS Lett 182, 302-304

46 Merkelo, H., Hartman, S.R., Mar, T., Singhal, G.S and Govindjie (1969) Science 164, 301-302

47 Eftink, M and Ghiron, C.A (1976) J Phys Chem 80, 486-493

48 Gratton, E and Degado, L.R (1980) Nuovo Cimento 56, 1110-1124

49 Gratton, E and Delgado, L.R (1979) Rev Sci Instrum 50, 789-790

New York

New York

0 1988 Elsevier Science Publishers B.V (Biomedical Division)

of atoms within a molecule Since the vibrational spectrum is sensitive to the molecule’s conformation and environment the Raman spectrum is used to probe the detailed chemistry of the biochemical molecule under study [2-41

The physical origin of Raman scattering lies in inelastic collisions between the molecules composing the liquid and photons which are the particles of light making

up the light beam An inelastic collision means that there is an exchange of energy between the photon and the molecule with a consequent change in energy, and

hence wavelength, of the photon (Fig 1) Moreover, since total energy is conserved

during the scattering process the energy gained or lost by the photon must equal an energy change within the molecule It follows that by measuring the energy gained

or lost by the photon we can probe changes in molecular energy The changes in the molecule’s energy are called transitions between molecular energy levels As already mentioned, in biochemical studies Raman spectroscopy is concerned primarily with the molecule’s vibrational energy level transitions although the resonance Raman effect also provides detailed information on electronic energy levels

Although present-day Raman spectroscopy uses high technology-based instru- mentation the experimental technique is simple in conception and can usually be depicted as in Fig 2 A monochromatic laser beam of wavelength, A, is focussed into the sample to produce a high photon density and the resulting scattered light, which includes the Raman spectrum, is analysed for wavelength and intensity The Raman effect is extremely weak and only a minute proportion of the incident

* Published as NRCC Number 28, 789