A FRET-based method to study the activity of electron or oxygen transfer proteins and redox enzymes Zauner, G.

(1)

oxygen transfer proteins and redox enzymes

Zauner, G.

Citation

Zauner, G. (2008, October 23). A FRET-based method to study the activity of electron or oxygen transfer proteins and redox enzymes. Retrieved from https://hdl.handle.net/1887/13201

Version: Corrected Publisher’s Version

License: Licence agreement concerning inclusion of doctoral thesis in the Institutional Repository of the University of Leiden

Downloaded from: https://hdl.handle.net/1887/13201

Note: To cite this publication please use the final published version (if applicable).

(2)

Chapter I

Introduction to a FRET based approach applied to electron/oxygen transfer

proteins and enzymes

(3)

Abstract

This thesis centers around electron transfer (ET) proteins and enzymes for which we are introducing a novel fluorescence based method to monitor their redox activity. It takes advantage of the fact that the optical characteristics of the protein’s co-factor vary upon changing the redox state of the protein. This change can be translated into the fluorescence intensity of a label covalently attached to the protein surface on the basis of Förster Resonance Energy Transfer (FRET).

This chapter describes the general features of fluorescence and of ET proteins as an introduction to the material in subsequent chapters.

(4)

Introduction

Redox proteins and enzymes catalyze redox reactions in the living cell, which are essential for many biological processes. For their activity most of them depend on one or several prosthetic groups. We distinguish ET proteins, which shuttle electrons between two players in a cascade, as opposed to redox enzymes, which catalyze redox reactions by which one substance is converted into another. The kinetics of redox enzymes and the function of electron transfer chains can be studied by monitoring the redox state of the prosthetic groups.

One of the common ways of determining the redox state of a prosthetic group is via its optical absorption spectrum. Many prosthetic groups absorb in the UV or visible region; their spectral characteristics depend on their redox state. When measuring the concentration of a protein the sensitivity of absorption measurements is often not sufficient. Moreover, absorption by other components in the same sample solution might obscure the desired absorption spectra. Here we describe a method for sensing the changes in protein absorption spectra based on fluorescence by means of Förster Resonance Energy Transfer (FRET).

Fluorescence General introduction

The phenomenon of fluorescence has a long history. The first observation of fluorescence was already reported in 1845 by Sir John Frederick William Herschel.

When observing a quinine solution excited by the sunlight he noticed that the homogenous colourless solution activated colour at the surface (1). This invention started a revolution in the field of fluorescence research, which nowadays is still stimulating new fields of research due to the remarkable sensitivity of fluorescence

(5)

detection (2) as even single fluorescent molecules are visible against a dark background (3;4;5). This is exploited in recent methodological developments in the field of single molecule detection and spectroscopy (6;7).

Basically the fluorescence process can be described as an energy dissipation process of optically excited atoms or molecules. A useful approach to understand the details of the excitation and emission process is to render them in the form of a diagram first introduced by Alexander Jablonski in the 1930’s (Figure 1.1):

Exited Singlet States (S₁& S₂)

Ground State (S₀) Phosphorescence Fluorescence

Intersystem Crossing

2 1 0 S

S

Internal conversion

hv_A hv_A

hv_F Absorption

hv Exited Singlet States (S₁& S₂)

Exited Triplet State (T1)

2 1 0 S2

S1

hv_A hv_A hvhv_A_A

hv_F Absorption

P

2 1 0 S

S

hv_A hv_A

hv_F Absorption

2 1 0 S2

S1

hv_A hv_A hvhv_A_A

hv_F Absorption

P

2 1 0 S

S

hv_A hv_A hvhv_A_A

hv_F Absorption

2 1 0 S2

S1

hv_A hv_A hvhv_A_A

hv_F Absorption

P

Figure 1.1. Jablonski diagram (see text for description)

In this diagram the singlet ground and the electronically excited states are described by S0, S1 and S2. A molecule can occupy any of these electronic energy levels each of which is associated with a number of vibrational states, denoted by 0,1,2, etc.

Transitions from the S0 to the excited state (either S1 or S2) correspond with photon absorption and occur on a timescale of femtoseconds (10^–15s). Absorption of a photon typically occurs from molecules with the lowest vibrational energy in the ground state (S0) to a higher vibrational state in either S1 or S2 followed by an internal equilibration in these states. Generally, in the condensed phase, molecules relax very fast to the lowest vibrational level of S1. This phenomenon is called internal

(6)

conversion and takes place on a timescale of picoseconds (10^–12 s). Therefore, fluorescence emission is a consequence of a transition from a thermally equilibrated excited state (S1) back to the ground state (S0) where again a thermal equilibration takes place in picoseconds.

The release of the absorbed energy by emission of a photon typically occurs at a longer wavelength than that of absorption, an effect which is known as the Stokes shift. Another consequence of the above described relaxation mechanism is that the fluorescence spectrum is independent on the wavelength of exciting radiation (8).

Fluorescence is a short-lived emission from the singlet state (S1) with the electrons spin-paired in an antiparallel fashion. Since the transition from S1 to the ground state (S0) is spin allowed, meaning that no change in multiplicity occurs, the probability for this process is high. Fluorescence decay times are usually in the order of nanoseconds (10^–9s).

On the other hand, molecules in the S1 state can also undergo a spin conversion to the first triplet state T1, which is called intersystem crossing. Emission from T1 is named phosphorescence and is mostly shifted to longer wavelength (lower energy) relative to fluorescence. In the case of phosphorescence the transition from T1 to the singlet ground state is spin-forbidden, and results therefore in a much lower probability for this process to occur as compared to fluorescence. Phosphorescence is a long lived emission (in the range of 10^-4 - 10¹ seconds) (8;9).

General observables of fluorescent molecules are the fluorescent lifetime, the fluorescence intensity as a function of wavelength, the quantum yield and the polarization. These observables are used in biochemical and biophysical studies to provide information, for example, about concentrations of solutes, the interaction of biomolecules or dynamic processes in solutions.

(7)

Natural fluorophores, fluorescent indicators and probes

The main fluorophores used in biochemical assays can be classified into natural fluorophores, fluorescence indicators and so called probes or dye molecules (10).

Only limited examples of natural (intrinsic) fluorophores exist such as oxidized flavins, the indole in tryptophane residues, tyrosins, phenylalanine, reduced nicotinamide adenine dinucleotide (NADH), pyridoxal phospate and chlorophyll (8). The emission of such fluorophores in a protein matrix can be employed to monitor several phenomena. For example fluorescence of tryptophan residues is used to study binding of ligands, protein folding and protein-protein associations. A common problem of this approach is the high sensitivity of the intrinsic fluorophores towards their local environment which often changes upon addition of ligands (8). Also protein substrates commonly contain aromatic residues which fluoresce themselves, thus hampering the precise observation of the desired intrinsic fluorophore.

Fluorescence indicators are fluorophores whose spectral properties are sensitive to a substance of interest. This class of fluorophores find their predominant application in analytical or clinical chemistry as the presently available indicators sense substances such as Na⁺, Ca²⁺, Mg²⁺, Cl^- and O2 (8), examples for applications can be found in the reviews by Rudolf et al (11) and Tsien et al (12). Even pH changes can be monitored by some indicators (13).

In the cases where the molecule of interest (i.e. a protein) is nonfluorescent, or where the intrinsic fluorescence is not sufficient or inadequate, one can label the molecule with an extrinsic probe. For proteins this has the main advantage that the interference of emission from multiple intrinsic residues can be avoided (8;14). In the sixties Shimomura et al made the groundbreaking discoveries of aequorin and GFP, both isolated from the medusa Aaquorea. These proteins are themselves intrinsically fluorescent (15). For example, GFP contains a highly fluorescent group within a protected region of the protein, which is formed spontaneously upon folding of the

(8)

polypeptide chain (16). Nowadays chimeric GFP fusions are still widely used, because they are advantageous due to the possibility of expression in situ by gene transfer into cells. The protein under investigation fused with the GFP can be directly obtained (17). This, in combination with the fact that fluorescence detection is non invasive, opened the possibility of direct in vivo studies (2).

In our studies we explicitly focus on commercially available (extrinsic) small organic fluorophores to label the proteins of interest. These fluorophores are available for almost any desired wavelength range (18-22). Labeling with a small sized molecule, at a position away from the active site of the protein, might be beneficial to minimize possible steric hindrance problems interfering with the protein function (17).

To achieve an optimal use of an extrinsic chromophor, it should be inserted or covalently attached to the protein at a specific location. This can be achieved by applying different reactive moieties such as for example maleimide for sulfhydryl labeling (2;23) or succinimidylester to react with primary amines (23;18). The reactions can be controlled by the pH of the solution; NH groups of lysine side chains are reasonably good nucleophiles above pH 8.0 and can therefore react cleanly with a variety of reagents to form stable covalent bonds. D- amine groups (like the N- terminal NH2 group in a protein) are less basic than lysines giving the opportunity for a selective labeling reaction at this specific residue (23). Thiol groups are even more nucleophilic than amines and they are generally the most reactive functional group (in the form of an exposed cysteine residue) in a protein. Unlike most amines, thiols are reactive at neutral pH, allowing coupling to other molecules selectively even in the presence of amines. This feature makes the thiol group the attachment point of choice for covalent linkage of a dye molecule of interest to a protein (23).

(9)

Besides maleimide and succinimidylester other reagents for protein modification have been reported. Examples of them plus the reaction mechanisms yielding the covalent bonds are described by Brinkley et al (23).

Another rather innovative technique is based on labeling specific recombinant proteins with small organic fluorophores such as bioarsenic compounds even in living cells (17). Those dyes show a high specificity to a tetracysteine motif containing the rare sequence CCXXCC (C, cysteine; X, any residue), which can be genetically introduced into the protein sequence (24). Those labels are membrane-permeant and non-fluorescent, acquiring fluorescence only upon binding to the CCXXCC motif.

Therefore also this group of fluorophores has been introduced for in vivo imaging (25;26).

Förster Resonance Energy Transfer

Förster reported already in 1948 that a radiationless transfer occurs by a resonance interaction of the dipole pair between a fluorescent energy donor and a suitable energy acceptor (27). This application of fluorescence is based on the fact that energy absorbed by a chromophor can be transferred to another chromophor yielding an increase of its fluorescent intensity, depending on distance (10). The efficiency of this energy transfer (E) depends on the inverse sixth power of the distance between the donor and the acceptor and is given by:

E = R06/( R⁶ + R06), Eq.1.1

where R0 is the Förster radius, a parameter depending on the spectral properties of donor and acceptor and their relative orientation (10). Thus, if R0 is known, FRET efficiency can be used to estimate the donor-acceptor distance, R.

(10)

The Förster radius, R0, can be calculated from the equation (8)

R0 = 0.211(JN²n⁴ D)^1/6(Å). Eq.1.2

Here N² is an orientation factor, n is the refractive index, D is the fluorescence quantum yield of the donor, and J is the spectral overlap integral, defined as

J = FD()A()⁴d/FD()d, Eq.1.3

where FD() is the fluorescence intensity of the donor and A() is the extinction coefficient of the acceptor at wavelength , in M^-1cm^–1,and is expressed in nm.

If the donor- acceptor pairs are separated by a fixed distance the following two equations can be applied as well to measure the energy transfer efficiency:

E = 1 – (DA / D), Eq.1.4

where D is the fluorescence lifetime of the donor in absence and DA in presence of the acceptor.

Alternatively this equation can be expressed in terms of fluorescence intensities:

E = 1 – (FDA / FD), Eq.1.5

In the formula FD and FDA stand for the fluorescence intensity of the donor in absence and presence of the acceptor respectively.

For applications of FRET it is worthwhile mentioning that the transfer efficiency is mainly dependent on two parameters, the distance R and the Förster radius, R0. The

(11)

main application of FRET in biosciences is the so called ‘spectroscopic ruler’

approach (8;10;28;29), which assumes that the Förster radius between a donor acceptor pair remains constant throughout the experiment thus the changes in energy transfer between those two can be used to determine changes in the distance between the two molecules of interest (see equation Eq.1.1). Another application consists of attaching a donor and an acceptor to one molecule (for example to a protein) and then to use the observed energy transfer efficiency for measuring conformational changes in one molecule (30). Many additional applications have been reported for instance for monitoring protein- protein interactions (31).

Objective of our approach

Recently, a novel FRET based technique was developed (4;32) , which is based on the translation of the absorption changes of a redox protein into changes of the fluorescence of a covalently attached dye molecule. This method tremendously facilitates the detection of, for example, the redox state of the protein due to the sensitivity of fluorescence detection in comparison to conventional absorption measurements. According to the Förster theory, R0 depends on the spectral overlap between the donor fluorescence and acceptor absorbance (equation Eq.1.1) (8). We use the protein prosthetic group as the FRET acceptor. Therefore, if the absorption spectrum of the protein changes with its redox state, R0 will also depend on the redox state of the prosthetic group. This then modulates the fluorescence intensity of the attached dye molecule.

There are numerous prosthetic groups; only the ones in the proteins used in our studies will be described in more detail:

(12)

Prosthetic groups in proteins Copper proteins

Copper is essential to a variety of proteins with functions such as electron transfer, oxygen transport and even organic chemistry like the insertion of oxygen into a substrate. Classically, copper proteins are categorized according to their spectroscopic properties into three different subdivisions (33). Due to the discovery of more copper protein structures over the last decades, the current classification extends to seven subgroups. For our purposes proteins from only two of these subdivisions have been used to demonstrate that our approach can be successfully applied to copper proteins:

Type 1: Type-1 copper sites are found mainly in small blue copper proteins (also named cupredoxins) such as azurin, plastocyanin, amicyanin and pseudoazurin, which are involved in electron transfer. Nevertheless also some enzymes harbour a Type-1 Cu site such as for example nitrite reductase or the multicopperoxidase laccase. The copper is bound to the protein matrix in a distorted tetrahedral arrangement of two histidines, a cysteine, and a methionine. In the protein’s oxidized state an intense absorption band around 600nm is present which gives it a strong blue colour (34;35).

This was first believed to arise from the charge transfer of the copper-cysteine bond in the active site, but now it has been assigned to a -* transition in the MO- scheme of the Cu site. Mainly involved are the dx2-y2 –orbital on the Cu and a 3p-orbital on the Cys sulfur of the ligand. In the reduced form Cu has a d¹⁰ electronic configuration, which does not lead to an optical absorption band in the 300-600nm region (36).

Type 3: The Type-3 site is found in proteins like tyrosinase and hemocyanin and consists of two copper ions each coordinated by three His residues. Molecular oxygen binds to the reduced and colorless [CuÎ-CuÎ] type-3 centre to yield the oxygenated [CuÎ-O2-CuÎ] species in which oxygen is bound in a Cu2 bridging ‘side-on’ geometry.

The type-3 copper is characterized by a strong absorbance at 345nm in the oxygen bound form (37).

(13)

Flavin containing proteins

Flavoproteins are proteins that contain a nucleic acid derivative of riboflavin: either the flavin adenine dinucleotide (FAD) or flavin mononucleotide (FMN).

Both of them are redox active compounds that have a three-ringed isoalloxazine group as their basic structure. Flavin mononucleotide is a derivate of the base structure with a phosphate group attached to the 5’ carbon atom of the ribose. The FMN is non-covalently bound to the protein backbone in all known cases (34). FMN can exist in three possible redox states: oxidized (quinone), one-electron reduced (semiquinone) and two-electron reduced (hydroquinone). Each state is spectroscopically distinct from the other as described by Klugkist et al (38). As an example from this family of proteins we have used a flavodoxin in our studies.

Heme proteins

Heme is the complex of iron and protoporphyrin IX, which is the common prosthetic group of all proteins known as cytochromes. Formally the change of redox state in these proteins involves a single electron and pertains to the reversible equilibrium between Fe(II) and Fe(III) states of the central iron ion (34). Also the Fe(IV) oxidation level of heme iron is found as a catalytic intermediate in some cases. The classes of heme- containing proteins are distinguished by the way of heme binding and their secondary structure (39). The subgroups in this family are numerous and varied due to the marked difference in function of the individual proteins. Out of the large variety of heme proteins two examples have been used in our studies:

1. Cytochrome c550, a small electron transfer (ET) protein, which belongs to the class I of c-type cytochromes and contains a covalently-bound haem, which is low-spin

(14)

both in the oxidized and reduced forms. Reduced cytochrome c550 shows an intense absorption band at 416 nm (Soret band), which is shifted and decreased in intensity in the oxidized form of the protein. It is also characterized by a sharp peak at 550 nm (D band) and a smaller band at 522 nm (E band) in the reduced state (40).

2. P450cam, a heme protein monooxygenase, belongs to the cytochrome P450 enzyme family named for the absorption band at 450nm of their carbon monoxide (CO) bound form. In this case the conversion from a low-spin form of the iron, characterized by an absorption band at 417nm, to a high-spin form (around 395nm) occurs upon substrate binding to the heme (41).

(15)

In figure 1.2 the different prosthetic group structures (42;43;44;45) involved in the experiments throughout this thesis are shown.

A B

C D

A B

C D

Figure 1.2. Some examples of active sites in biology A: A Type-1 copper site [4AZU.pdb, (42)].

B: A Type-2 copper site [PDB 1JS8, (43)].

C: Flavin mononucleotide (FMN) [1YOB.pdb (44)].

D: A heme group [PDB 1DZ6, (45)].

See text for description of each of the structures.

(16)

Example for an application of the FRET based method

The structural example shown in Figure 1.3 is azurin, a type-1 copper protein, labeled with Cy5 fluorescent dye at the N-terminus.

-

e

^-

+

e

^-

-

e

^-

+

e

^-

Figure 1.3. Azurin from Pseudomonas aeroginosa labeled with Cy5 on the N-terminus of the protein;

Scheme below: Protein labeled with a fluorophore in the reduced state; upon oxidation FRET occurs and the fluorescence is quenched

(17)

The cartoon above is representing the basic idea for the experiments described in this thesis: A fluorescent dye molecule is attached to a redox protein, the absorption of which shows an overlap with the dye emission, in this particular case in the oxidized but not in the reduced state. The label fluorescence will be quenched in the oxidized but not in the reduced form of the protein (Figure 1.3; scheme).

The aim of the research presented here was to show different applications of this novel FRET based approach to demonstrate how widely applicable the method is.

The sensitivity of the approach even allows to go down to the single molecule level.

Chapter II reports the proof of principle that the technique can be applied to proteins harbouring different prosthetic groups and how it can be used to determine the individual midpoint potentials.

Chapter III describes a FRET based system to sense oxygen in solution. The method exploits the sensitivity of the endogenous fluorescence of type-3 copper proteins towards the presence of oxygen by translating the near-UV emission of the protein to label fluorescence in the visible range.

Chapter IV is an extension of chapter III in which the oxygen sensor in solution is immobilized in optically transparent silica matrices.

Chapter V reports an initial proof that the principle can be also applied to a P450 monooxygenase to monitor substrate binding to the active site of the protein.

Chapter VI reports the first single molecule study based on our method by means of copper containing nitrite reductase. It was investigated if and how the data obtained can be related to the enzymological parameters that have been reported for this enzyme in the literature.

Finally in ‘General discussion, conclusions and perspective’ the main conclusions are presented and the perspectives for using our approach for further investigations are discussed.

(18)

Reference List

1. Herschel F.W. (1845) Philosophical Transactions of the Royal Society of London.

2. Ernst, L. A., Gupta, R. K., Mujumdar, R. B., and Waggoner, A. S. (1989) Cytometry 3-10.

3. Steingart, K., Henry, M., Ng, V., Hopewell, P. C., Ramsay, A., Cunningham, J., Urbanczik, R., Perkins, M., Aziz, M. A., and Pai, M. (2006) Lancet Infectious Diseases 570-581.

4. Kuznetsova, S., Zauner, G., Schmauder, R., Mayboroda, O. A., Deelder, A. M., Aartsma, T. J., and Canters, G. W. (2006) Analytical Biochemistry 52-60.

5. Lichtman, J. W. and Conchello, J. A. (2005) Nature Methods 910-919.

6. Basche, T. (2001) Single Molecules 237-240.

7. Orrit, M. and Bernard, J. (1990) Physical Review Letters 2716-2719.

8. Lakowicz J.R. (1996) Principles of Fluorescence Spectroscopy, Kluwer Academic/Plenum Publishers, New York, Boston, Dordrecht, Moscow.

9. Amao, Y. (2003) Microchimica Acta 1-12.

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

11. Rudolf, R., Mongillo, M., Rizzuto, R., and Pozzan, T. (2003) Nature Reviews Molecular Cell Biology 579-586.

12. Tsien, R. Y. (1989) Annual Review of Neuroscience 227-253.

13. Wolfbeiss, O. and Marhold H. (1987) Fresenius Z Analytical Chemistry 347- 350.

14. Zauner, G., Lonardi, E., Bubacco, L., Aartsma, T. J., Canters, G. W., and Tepper, A. W. J. W. (2007) Chemistry-A European Journal 7085-7090.

15. Shimomura, O., Saiga, Y., and Johnson, F. H. (1962) Federation Proceedings 401-&.

16. Chalfie, M., Tu, Y., Euskirchen, G., Ward, W. W., and Prasher, D. C. (1994) Science 802-805.

(19)

17. Miyawaki, A., Sawano, A., and Kogure, T. (2003) Nature Cell Biology S1-S7.

18. Panchuk-Voloshina, N., Haugland, R. P., Bishop-Stewart, J., Bhalgat, M. K., Millard, P. J., Mao, F., Leung, W. Y., and Haugland, R. P. (1999) Journal of Histochemistry & Cytochemistry 1179-1188.

19. Mujumdar, R. B., Ernst, L. A., Mujumdar, S. R., Lewis, C. J., and Waggoner, A.

S. (1993) Bioconjugate Chem 105-111.

20. Sauer, M. e. al. (1995) Journal of Fluorescence 247-261.

21. Kemnitzer, N. U. and et al. Fluorescent Dyes- Detection Methods of the Future. (2002) 242-243. GIT Verlag Gmbh&Co. Ref Type: Conference Proceeding

22. Buschmann, V., Weston, K. D., and Sauer, M. (2003) Bioconjugate Chem 195- 204.

23. Brinkley, M. (1992) Bioconjugate Chemistry 2-13.

24. Hoffmann, C., Gaietta, G., Bunemann, M., Adams, S. R., Oberdorff-Maass, S., Behr, B., Vilardaga, J. P., Tsien, R. Y., Eisman, M. H., and Lohse, M. J. (2005) Nature Methods 171-176.

25. Griffin, B. A., Adams, S. R., and Tsien, R. Y. (1998) Science 269-272.

26. Giepmans, B. N. G., Adams, S. R., Ellisman, M. H., and Tsien, R. Y. (2006) Science 217-224.

27. Förster, T. (1948) Annalen der Physik.

28. Wu, P. G. and Brand, L. (1994) Anal Biochem 1-13.

29. Selvin, P. R. (2000) Nature Structural Biology 730-734.

30. Kajihara, D., Abe, R., Iijima, I., Komiyama, C., Sisido, M., and Hohsaka, T.

(2006) Nature Methods 923-929.

31. Kenworthy, A. K. (2001) Methods 289-296.

32. Schmauder, R., Alagaratnam, S., Chan, C., Schmidt, T., Canters, G. W., and Aartsma, T. J. (2005) Journal of Biological Inorganic Chemistry 683-687.

33. Adman, E. T. (1991) in Advances in Protein Chemistry pp. 145-197, Academic press, Inc.

(20)

34. Palmer, G. and Reedijk, J. (1991) Biochimica et Biophysica Acta 599-611.

35. Gray, H. B., Malmstrom, B. G., and Williams, R. J. P. (2000) Journal of Biological Inorganic Chemistry 551-559.

36. Schmauder, R., Alagaratnam, S., Chan, C., Schmidt, T., Canters, G. W., and Aartsma, T. J. (2005) Journal of Biological Inorganic Chemistry 683-687.

37. Solomon, E. I., Sundaram, U. M., and Machonkin, T. E. (1996) Chemical Reviews 2563-2605.

38. Klugkist, J., Voorberg, J., Haaker, H., and Veeger, C. (1986) European Journal of Biochemistry 33-40.

39. Bendall, D. S. (1996) Protein Electron Transfer, BIOS Scientific Publisher, Oxford.

40. Lommen, A., Ratsma, A., Bijlsma, N., Canters, G. W., van Wielink, J. E., Frank, J., and Van Beeumen, J. (1990) European Journal of Biochemistry 653-661.

41. Peterson, J. A. (1971) Archives of Biochemistry and Biophysics 678-693.

42. Kolczak, U., Dennison, C., Messerschmidt, A., and Canters, G. W. (2001) in Handbook of Metalloproteins (Messerschmidt, A., Huber, R., Poulos, T., and Wieghardt, K., Eds.), pp. 1170-1194, John Wiley & Sons, Chichester.

43. Cuff, M. E., Miller, K. I., van Holde, K. E., and Hendrickson, W. A. (1998) Journal of Molecular Biology 855-870.

44. Alagaratnam, S., van Pouderoyen, G., Pijning, T., Dijkstra, B. W., Cavazzini, D., Rossi, G. L., van Dongen, W. M. A. M., van Mierlo, C. P. M., van Berkel, W. J.

H., and Canters, G. W. (2005) Protein Science 2284 - 2295.

45. Schlichting, I., Berendzen, J., Chu, K., Stock, A. M., Maves, S. A., Benson, D.

E., Sweet, B. M., Ringe, D., Petsko, G. A., and Sligar, S. G. (2000) Science 1615- 1622.

(21)