Electron Spin Resonance

3 downloads 0 Views 1MB Size Report
macromolecules such as cell-wall polysaccharides (Irwin et al., 1985) and the catalytic mechanism of enzymes such as adenylate kinase (Kharatyan.
, Electron Spin Resonance

6

R. CAMMACK

1. Introduction to the ESR Method A. Systems That Can Be Detected by ESR B. Principles of ESR Spectroscopy C. Information from the ESR Spectrum D. Experimental Techniques E. Instrumentation II. Naturally Occurring Paramagnetic Centers A. Fn:e Radicals B. Tran!'ition Metals and Metalioproteins m. Applications to Plant Metabolism A. Photosynthesis B. Mitochondrial Respiratory Chain C. Applications of Spin Labels D. ESR Studies of Chloroplast Development IV. Concluding Remarks V. Glossary References

I. A.

INTRODUCTION TO THE ESR METHOD Sy~tems

That Can Be Detected by ESR

Electron spin resonance (ESR) is a spectroscopic technique for the study of paramagnetic molecules, that is, ones that contain unpaired electron spins. The principle is similar to nuclear magnetic resonance (NMR), which applies to nuclear spins. In comparison with NMR it is selective for a smaller range of SUbjects, and it is often much more sensitive. ESR spectroscopy The Biochemistry of Plants. Vol. 13 Copyright © 1987 by Academic Press, Inc. Al! rights of reproduction in any form reserved.

229

, 230

R. Cammack

does not detect the majority of organic molecules, or the commonest ions, since all of their electrons are paired in orbitals. However, ESR can readily detect one paramagnetic center in a high-molecular-weight species such as a protein. Therefore in metabolic studies, the technique has been used more extensively to study the properties of the enzymes than of their metabolites. Since most biological molecules are diamagnetic, it is possible to examine the paramagnetic species in complex biological systems such as whole chloroplasts or even whole plant tissue. The types of compounds that are studied may be divided into three types: (l) naturally occurring free radicals, which are organic compounds that contain an unpaired electron; (2) transition-metal ions such as Mn2 +, Fe 3 +, and Cu2'1-; and (3) synthetic free radicals, termed spin labels, that are added to "6' the biological syste"ffi as probes. Two alternative terms, electron spin resonance (ESR) and electron paramagnetic resonance (EPR) spectroscopy, are used for the same experimental technique. A distinction is sometimes drawn between ESR, as applied to compounds such as orgnic free radicals, where there is an identifiable unpaired electron spin, and EPR, as applied to compounds that have paramagnetism associated with a significant orbital moment, for example, transitionmetal ions. The same basic instrument is used for both types of sample, but there are differences in technique. For example, transition-metal ions, such as iron or copper, often require extremely low temperatures for their detection because their spectra are too broad at physiological temperatures. In this chapter, the term ESR will be used indiscriminately. The theory and practice of ESR spectroscopy have been described in numerous books and articles, such as the series edited by Berliner and Reuben (1981). Biological applications are included in a comprehensive series of surveys of the literature [the latest being by Symons (1986)J. An introduction to the practical aspects of ESR of biological materials is provided by Palmer (1967). The ESR spectrum of a paramagnetic species can provide much information about its mobility, orientation, coordination geometry, and interaction with nuclear and electron spins. In this chapter these features of the technique will be illustrated with reference to specific applications to plants. The emphasis will be on an overview of the types of compound that can be detected in whole tissue or in purified materials.

B.

Principles of ESR Spectroscopy

The following operational description is not intended to be rigorous, but to introduce some of the features of the ESR spectrum that can be exploited in the investigation of biological systems. All forms of spectroscopy observe the absorption of electromagnetic radiation by a system that is excited from one energy level to another. In ESR

, . 231

6. Electron Spin Resonance

spectroscopy, the two energy levels are induced by placing the electrons of the "sample in an applied magnetic field. Then, according to the laws of quantum mechanics, each electron must take up one of two possible states, corresponding to orientations aligned either with or against the field. These two states differ slightly in energy, and there will be a thermal population of electrons, with slightly more electrons in the lower energy level. At any particular value of the applied magnetic field H o, electrons in the lower level will become excited to the upper level by absorbing microwave energy of frequency v, determined by the resonance equation: hv = gf3Ho

(1)

In this equation hand g are constants. For a free-radical sample (g = 2.0023) and microwaVes in the X-band frequency (9 GHz), the magnetic field for resonant absorption is about 0.33 T (3.3 kG). The resonance process tends to equalize the popUlations of the two states, but electrons can return to the ground state by means of a relaxation process in which energy is given to the surroundings or "lattice" in the form of heat. The relaxation process may become a limiting factor at low temperatures and high microwave power, when the signal becomes smaller and is said to be saturated. C.

Information from the ESR Spectrum

1. g Values

The value of g is a characteristic of the type of paramagnetic. center and should be a independent of the microwave frequency (though this is not strictly true if there are interactions with other nuclear or electron spins). For biochemical purposes, the ESR spectrum may be considered by analogy with the more familiar optical spectroscopy. The optical absorption spectrum is obtained by sweeping the optical frequency (and hence wavelength) and observing the absorption of radiation at these frequencies. Compounds are characterized in this spectrum by the characteristic wavelengths at which they absorb. In principle it would be possible to measure an ESR spectrum in the same way by sweeping the microwave frequency and detecting the absorption of microwave radiation. Each paramagnetic compound is identified by its characteristic g value. For instrumental reasons the spectrum is not obtained in this way. Instead the microwave frequency is held constant and the magnetic field is swept instead. It can be seen from Eq. (1) that increasing Ho will also have the effect of sweeping from higb to low g values of the paramagnetic species. The ESR spectrum is conventionally displayed as the first derivative of the microwave absorption spectrum versus applied magnetic field. As with an optical absorption spectrum, the shape and position of the spectrum are

.

.

, 232

R. Cammack

characteristic of the type of compound, and the intensity of the ESR signal is proportional to the amount of the compound present. In principle. it is possible to determine the absolute number of electron spins in any ESR spectrum by double integration, provided the signal is not saturated with microwave power. The ESR spectrum has additional properties that are not found in an optical spectrum and that greatly enhance its usefulness. It may be interpreted to obtain information about the structure of the paramagnetic species, its mobility, and its interaction with neighboring electron and nuclear spins. 2.

Anisotropy

In many moleculeA, particularly transition-metal ions, the energy of the ESR transition and hence the g value depend on the angle between the Hn field and the paramagnetic center. If the molecules are an aligned, as in a crystal, the ESR spectrum is a narrow line corresponding to a particular orientation. The position of the line will alter cyclically as the crystal is rotated. A few studies have been done on crystals of proteins, for example. plastocyanin (Penfield et a!., 1981), which give much more detailed information about the orientation of the g and A tensor axes. Biological samples usually consist of randomly oriented molecules. If the molecules are fixed, or moving slowly on the ESR time scale (see S~ction I,C,5) the resulting "powder" spectrum will consist of the sum of the spectra from all possible orientations. This leads to broadening of the ESR spectrallineshape. Further broadening in frozen samples is caused by strain effects, which are due to minor conformational differences between molecules. An example of such a spectrum is that of reduced ferredoxin (Fig. I b). The hyperfine and electron spin-spin interactions (Sections LC3 and LCA) also display anisotropy. In ESR spectra of radicals of low molecular weight in solution, the rapid tumbling of the molecules causes the g-value anisotropy to be averaged out so that the linewidth is usually narrow. The property of motional averaging is exploited in the spin-label method for investigating molecular motions (Section I,e,S). 3. HypeJjine Interactions

Some nuclei have a spin, and therefore a magnetic moment. If one of these is associated with an unpaired electron, it will influence the magnetic environment of the electron and hence the ESR spectrum. Specifically, if the spin on the nucleus has a value I, such as that of the hydrogen nucleus, lH (1 = i), its effect will be to split the ESR spectrum into (21 + 1) or, in this case, 2 lines. The magnitude of the splitting is known as the A value. In a nitroxide spin label the unpaired electron is associated principally with a nitrogen nucleus, of which thenatural isotope IS 14N (I = 1); therefore the spectrum consists of three lines. Organic radicals usually show extensive

, 6. Electron Spin Resonance

233 g VALUE

2.6

2.4

2.2

1.8

2

l" }

(b)

(c)

240

260

280

300

320

340

360

380

400

MAGNETIC FIELD (mT)

Fig. 1. Typicallineshapes of ESR spectra of transition metal ions. (a) Mn2~ ions in solution; (b) reduced ferredoxin from parsley, recorded at 30 K; (c) Cu/Zn superoxide dismutase, recorded at 77 K.

electron delocalization, so that their spectra are influenced by several nuclear spins. Hyperfine interaction with each additional nucleus causes each ESR line to be further split, and the spectra can become quite complex (Fig. 2). Such hyperfine patterns can be used to "fingerprint" a particular radical. Moreover, the pattern may be simulated by computer to determine the number of interacting nuclei and the extent of electron delocalization over the molecule (Edmondson, 1978). Hyperfine splittings are also observed in the spectra of transition metal ions, for example 55Mn2 + (1 = !), which gives six lines (Fig. la), or 63CU (1 = i), which gives four lines (Fig. Ic). In addition, if a transition-metal ion is coordinated to a ligand with a nuel.ear spin, such as 14N, the lines will be further split. This is sometimes referred to as a "superhyperfine" interaction. To identify the ion and its ligands, substitutions can be made with an

.

,

, 234

R. Cammack

(0)

3 Gauss

~-=-=~-+-.

H

Fig. 2. Semiquinone radicals of quinones in alkaline ethanol solution. (a) Juglone, from Pterocarya fraximfolia Spach; (b) plumbagin, from Drosera binata Labill; (c) plastoquinone. [From Pedersen (1978), with permission.]

isotope of different nuclear spin. For example, to identify an ESR spectrum of a protein containing iron, the organism might be grown on a medium containing 57Fe (I = i) instead of natural 56Fe (I = 0), causing a hyperfine splitting of the ESR spectrum into two lines. Similarly, interaction with exchangeable protons can be observed by substitution of deuterated water instead of H 2 0. In contrast to radioisotope tracer experiments, it is necessary to replace all or most of the natural isotope in these experiments. 4. Interactions with Other Electron Spins

As well as interactions with nuclear spins, the ESR spectrum is also influenced by the (much larger) magnetic moments of other unpaired electrons in the vicinity. In the case of a transition ion that has more than one unpaired electron, such as high-spin Fe (Ill) with five unpaired 3d electrons, the electrons will be strongly coupled together by quantum-mechanical interactions to produce several sets of electron energy levels. In general, only those ions with

, 235 .

6. Electron Spin Resonance 9 VALUE

10 8 7

6

5

4

2

3

IIII I I I

I J I I I I Ii" Iii

I I

J

I l" )

(b)

(c)

J

I

100

J

I

I

I I II I I I II

200

I

300

I

I

I

400

MAGNETIC FIELD (mT)

Fig. 3. Spectra of nitrite reductase from Cucurbita pepo: (a) spectrum of the [4Fe--4S] cluster cyanide derivative, reduced with dithionite; (b) nitrosyl siroheme derivative, formed by reduction in the presence of excess nitrite; (c) Fe(III)-siroheme, in the oxidized enzyme as prepared.

an odd number of unpaired electrons are likely to be detectable by ESR. The spectra are often highly anisotropic and have apparent g values that are widely separated from their "true" g values; for example the g values of high-spin Fe (III) heme are typically 6 and 2 (e.g., Fig. 3c), although the true g values are close to 2 (Palmer, 1985). Other effects may be produced by interactions with electron spins on neighboring molecules. These may cause broadening or splitting of the spectrum. If the sample of paramagnetic material is too concentrated, the spectrum may become undetectable. Fortunately, the paramagnetic centers associated with biological macromolecules fulfill the requirement of being "magnetically dilute." However, the spectra of spin labels can become wiped out if the sample is too concentrated, and this has provided the basis of methods for following spin-label diffusion. The presence of a high local concentration of paramagnetic salt such as ferricyanide in solution may broaden the spectrum so much that it is undetectable. Hence interactions with ions on one side of a membrane, for example, may be observed. When the paramagnetic centers are evenly spaced out, the electron spin-spin interactions can, in principle, provide information about distances between them (Coffman and Buettner, 1979).

,

236

R. Cammack

5. Motion

The line shape of the ESR spectrum of a paramagnet in solution will also be influenced by any motion. As already mentioned in Section I,C,2, the ESR spectrum is broadened by anisotropy of g and A if the molecules of the sample are tumbling slowly in solution (for example protein molecules) or frozen. However, if the molecules are moving at rate that is comparable with the microwave frequency, the linewidth of the spectrum becomes narrower, at a position determined by the average values of g and A. This property has been particularly exploited in studies of nitroxide spin labels, where the spectra have been interpreted to estimate the rate of motion (see Watts, 1985). Rapid motion, such as that displayed by a small molecule in solution, will giye a narrow spectral linewidth. Slow motion, such as a protein in solution, will give a ''.powder'' spectrum and the spectral lines will be broad. It should be said that in this context, "slow" refers to a rotational correlation time of less than 10-6 sec, which would be very fast on the NMR time scale. More information about slower motions may be obtained from another ESR technique known as saturation transfer (Dalton et ai., 1985). In general, motion of paramagnetic molecules will also be anisotropic; for example, a lipid molecule in a membrane will tend to rotate about its long axis, perpendicular to the membrane, but will also undergo some precessional motion. More detailed analysis of the ESR spectral lineshape, by computer simulations, can be made to estimate the anisotropy of motion. The lineshape of a spin-label spectrum also contains information about the polarity (hydrophilic or hydrophobic) of its environment. 6.

Orientation

Another source of information in ESR spectra relates to the anisotropy of many signals. As already mentioned, the ESR spectrum of a single-crystal sample will vary as the sample is rotated in the cavity of the spectrometer. this will be true for any sample even if there is only a partial orientation of the paramagnetic centers. Such an orientation dependence has been observed in the ESR spectra of manganese in whole leaves from some plant species (McCain et aI., 1984). In organelles such as chloroplasts it is possible to produce samples with partially orientated membranes by carefully drying them on a surface (Prince et ai., 1980), or by viscous flow (Dismukes et a/.. 1978), or by a magnetic field (Dismukes and Sauer, 1978). The spectra of membrane-bound proteins may then show orientation dependence, which demonstrates that the molecules have a specific arrangement within the membrane plane. This has been observed, for example, in photo systems I and II of chloroplasts (Hootkins and Bearden, 1983; Rutherford, 1984). The results may be interpreted to show the direction of the g-tensor axes relative to the membrane plane. In the case of heme proteins, this will give structural information, since one of the g values, gz, lies perpendicular to the heme plane.

, 6. Electron Spin Resonance

237

D. Experimental Techniques 1.

Sample Preparation

The constraints on the preparation of biological samples for ESR work are that the samples should be (1) sufficiently concentrated-to examine a specific metal center or radical it may be necessary to use purified organelles or materials such as isolated proteins-and (2) free from contaminating organic or inorganic paramagnetic materials-dust or soil can be a rich source of unwanted ESR signals. For room-temperature work the sample size is limited by the dielectric microwave loss caused by water, so that very thin flat cells or capillaries are used. The requirement that the spectra of some transition-metal ions can only be observed at very low 1ymperatures means that the kinetics of enzymic reactions cannot be followed directly by ESR. However, techniques have been developed (Bray and George, 1985) whereby reactions can be initiated by rapid mixing and frozen within 5 msec in cold isopentane, and samples are prepared for ESR spectroscopy. 2.

Measurement of Redox Potentials

Many of the species observed by ESR are the products of electron-transfer reactions. ESR is a convenient way of observing the redox state, oxidized or reduced, of an electron carrier. It has therefore been used extensively in the estimation of midpoint potentials of metalloproteins such as ferredoxins (Cammack et aI., 1977), and more particularly for membranebound electron carriers that cannot be isolated in the pure state, for example, photosystem I (Evans et aI., 1974). The method involves bringing the electron-transfer system to equilibrium at a known redox potential and estimating the amount of oxidized and reduced species by ESR. If the species of interest is detectable at low temperatures, a sample is withdrawn and frozen for spectroscopic measurements. Practical details are provided by Dutton (1978).

E. Instrumentation 1.

Continuous-Wave Spectrometers

These are the conventional type of spectrometer as commercially available. They are designed for high sensitivity. The ESR cavity in which the sample fits will accommodate a cryostat using nitrogen or helium flow to adjust the temperature. The cavity has a grid or slot through which samples may be irradiated with light for photosynthetic experiments. 2.

Kinetic ESR

Most ESR measurements are carried out relatively slowly, over the course of a few minutes. However, in the particular case of studying photo-

,

238

R. Cammack

synthetic primary reactions, the method can be extended to kinetic measurements to follow electron-transfer processes after a flash of light. Commercially available spectrometers incorporate magnetic field modulation, typically at 100 kHz, and electronic filters to enhance their signal: noise ratio. These features limit the response time to values of the order of milliseconds. This is sufficient to determine the rates of some recombination processes, for example, the decay in the dark of signal 1 in illuminated chloroplasts, but is quite inadequate for the rates of forward electron transfer, which are in the range of nanoseconds to microseconds. Adaptations of the spectrometer design, such as the use of direct detection instead of magnetic-field modulation, can reduce the response time down to 0.1-1 JLsec (McLaughlan, 1985). 3.

Other Types of Spectr'ifineters

For very rapid reactions an alternative approach is to use pulsed ESR spectrometers. Instead of irradiating the sample continuously with microwaves, the sample is irradiated by two or three short, high-energy pulses of microwaves. The signal is then detected as an "echo," in a way analogous to spin-echo NMR (Thurnauer et a!., 1979; Norris et aI., 1980; Nishi et al., 1980). The limit of time resolution of this technique is set by the time for the energy of the intense microwave pulse to decay in the cavity, on the order of 50 nsec. The penalty of very rapid techniques is a drastic reduction in sensitivity, and it is necessary to average the signal over many flashes by means of a transient recorder. As described in Section tC,4, hyperfine interactions and the study of spin-spin interactions with other paramagnetic centers have been invaluable in the study of the environment of paramagnetic center. The limitation on the measurement of A values in ESR spectra is set by the rather broad line shapes that are often observed and by the complexity of the spectra when there are several interacting nuclei. These may be overcome in part by making ESR measurements at other frequencies, such as Q-band (35 GHz) or S-band (4 GHz). Since the g values should be invariant with frequency, while the A values are invariant on a magnetic field scale, the effects of hyperfine interactions can be distinguished from the effects of g-value anisotropy. An example is the demonstration, by measurements at Q-band, that the complex spectrum of fully reduced photosystem I (see Fig. 6e) is due to two interacting iron-sulfur clusters (Aasa et ai., 1981). The effects can also be distinguished in rapid measurements by pulsed ESR. Spin-echo measurements of photosynthetic primary reactions have been made at Kband frequency, 24 GHz (Furrer and Thurnauer, 1983). From the instrumental point of view, changing frequency is not a simple matter, however; a significant change in frequency requires the reconstruction of much of the spectrometer. Instruments at frequencies other than X-band are generally less well optimized for signal: noise ratio.

.

t

, 6. Electron Spin Resonance

239

Another experimental approach to the measurement of hyperfine interactions is electron-nuclear double resonance (ENDOR). In principle at least, this method combines the resolution of NMR with the sensitivity of ESR. However, it has its own drawbacks, such as requiring critical temperatures for the signals to be detectable, and not being quantitative. ENDOR requires a more complex spectrometer in which the ESR spectrum is measured while irradiating the sample at a radio frequency in tune with the nuclear frequency. Each hyperfine interaction gives rise to a pair of lines in the ENDOR spectrum. In biological systems the most commonly observed hyperfine interactions are with protons from water or proteins. The ENDOR technique has been used in a number of cases, notably in photosystem II, where it was used to support the chlorophyll dimer model for the primary electron donor (Norris etiJl., 1974). A third instrumental approach to the study of hyperfine interactions is electron spin-echo envelope modulation (ESEEM). This method requires a pulsed ESR spectrometer (Section I,E,2) and involves analysis of the decay envelope after a microwave pulse (Mims and Peisach, 1979b). The method provides the same type of information as ENDOR, but the two methods are .somewhat complementary; for example, ESEEM may be better at detecting weak, long-distance hyperfine interactions. Both ESEEM and ENDOR have been applied to purified proteins from plants, notably the copper proteins (Mims and Peisach, 1979a~ Roberts et al., 1984) and ferredoxins (Peisach et al., 1977; Fritz et al., 1971).

II.

NATURALLY OCCURRING PARAMAGNETIC CENTERS

A.

Free Radicals

Free radicals are usually unstable species. They are produced by oneelectron redox reactions or by homolytic cleavage of a bond. These are uncommon processes in conventional metabolism but are of central importance in photosynthetic reactions. Homolytic cleavage requires considerable energy-for example, a photochemical process-and the products tend to recombine readily. Short-lived free-radical intermediates in chemical reactions may be detected by reaction with a compound known as a spin trap, which reacts to form a more stable paramagnetic species (Janzen, 1980). From the lineshape of the ESR spectrum of the resulting radical it is often possible to identify the radical, for example, hydroxyl or superoxide radicals (Harbour and Bolton, 1975; Green et af.. 1985). An example 'Jfthe application of the ESR of free radicals is a sensitive and rapid method for the identification of ortl1o- and para-quinones in crude extracts from a variety of plants (Pedersen, 1978). Quinones were extracted

, 240

R. Cammack

with ethanol and treated with alkali to stabilize the semiquinone radical species. The conditions could be adjusted so as to stabilize different quinone radicals and thus simplify the interpretation of superimposed spectra of mixed species. Only a few grams of leaf material were needed. The results were semiquantitative. The spectra coul,d be assigned by comparison of spectra with known standards or by computer simulation. Among the quinones that were unambiguously identified from their characteristic hyperfine splitting patterns were juglone, chlorogenic acid, and rosmarinic acid. In some cases, however, similar spectra were obtained from related compounds, such as chlorogenic acid and cynarin (I A-dicaffeylquinic acid) (Fig. 2). Fr~e-radical ESR signals have been observed in wood, pollen, and other dried vegetable matJe.r (Priestley et al., 1985). This has been related to the presence of lignin and other phenolic polymers. The degradation of wood by ligninases is also thought to involve free-radical intermediates, which have been investigated by ESR (Andersson et al., 1985). B.

Transition, Metals and Metalloproteins

The transition elements are characterized by having unfilled d electrons, and in certain oxidation states they are paramagnetic. Generally those states be detectable by ESR. These with an odd number of unpaired electrons states, and the nuclear spins of the elements of biological interest, are summarized in Table L

will

1.

Vanadium

The vanadyl ion YQ2+ is paramagnetic and characterized by a hyperfine splitting into eight lines by the sly nucleus. So far, only one vanadium enzyme has been identified in plants, a bromoperoxidase from a seaweed, Ascophyllum nodosum (De Boer et a!., 1986). The spectrum of vanadium has also been observed in the toadstool Amanita phalloides (Gillard and Lancashire, 1983). Manganese

2.

Mn2+ ions are unusual among the transition-metal ions in that they can be observed readily in solution at physiological temperatures. The ESR spectrum of the aquo ion typically consists of six hyperfine lines centered at g = 2 (Fig. la) and can be detected at concentrations of a few micromolar. This can be used to measure the concentration of free manganese in solution. In complexes-for example, with EDTA or with some proteins-the spectrum of Mn 2 + may become so broadened as to be undetectable. The EDTA effect may be exploited by adding EDTA to crude cell exracts, to remove signals from Mn2 + that would otherwise obscure the spectra of interest.

,

.

, .241

6. Electron Spin Resonance TABLE I Transition Elements of Biological Interesfi

Element

Nucleus

%NaturaI abundance

Nuclear spin

f

Vanadium Manganese Iron

51V 55Mn 57Fe

99.76 100 2.19

Cobalt Nickel Copper

59CO 61Ni 63CU 65CU 95Mo 97Mo

100 1.13 69.09 30.91 15.72 9.46

Molybdenum

"6.