«Max-Planck-Institut für Biochemie Abteilung Strukturforschung Biologische NMR-Arbeitsgruppe Structural Investigations on Green Fluorescent Protein ...»
Technische Universität München
Institut für Organische Chemie und Biochemie
Max-Planck-Institut für Biochemie
Structural Investigations on Green Fluorescent
Protein Variants and the Adenylyl CyclaseAssociated Protein
Vollständiger Abdruck der von der Fakultät für Chemie der Technischen Universität
München zur Erlangung des akademischen Grades eines
Doktors der Naturwissenschaften
Vorsitzender: Univ.-Prof. Dr. W. Hiller
Prüfer der Dissertation:
1. apl. Prof. Dr. L. Moroder
2. Univ.-Prof. Dr. Dr. A. Bacher Die Dissertation wurde am 30.01.2003 bei der Technischen Universität München eingereicht und durch die Fakultät für Chemie am 20.03.2003 angenommen.
Parts of this thesis have been or will be published in due course:
Markus H. J. Seifert, Dorota Ksiazek, M. Kamran Azim, Pawel Smialowski, Nedilijko Budisa and Tad A. Holak Slow Exchange in the Chromophore of a Green Fluorescent Protein Variant J. Am. Chem. Soc. 2002, 124, 7932-7942 Markus H. J. Seifert, Julia Georgescu, Dorota Ksiazek, Pawel Smialowski, Till Rehm, Boris Steipe and Tad A. Holak Backbone Dynamics of Green Fluorescent Protein and the effect of Histidine 148 Substitution In press (Biochemistry) Dorota Ksiazek, Hans Brandstetter, Lars Israel, Gleb P. Bourenkov, Galina Katchalova, Hans D. Bartunik, Michael Schleicher and Tad A. Holak The Crystal Structure of the N-terminal Domain of the Adenylyl Cyclase-Associated Protein (CAP) from Dictyostelium discoideum submitted (Nature Structural Biology)
Lewinski K., Chruszcz M., Ksiazek D., Laidler P.
Crystallization and preliminary crystallographic analysis of a new crystal form of arylsulfatase A isolated from human placenta Acta Crystallogr D Biol Crystallogr. 2000, 56, 650-652 Litynska A, Przybylo M, Ksiazek D. Laidler P.
Differnces of alpha3beta1 integrin glycans from different human bladder cell lines Acta Biochim. Pol. 2000; 47, 427-34.
Publications Laidler P., Gil D., Pituch-Noworolska A., Ciolczyk D., Ksiazek D., Przybylo M., Litynska A.
Expression of beta1-integrins and N-cadherin in bladder cancer and melanoma cell lines Acta Biochim. Pol. 2000; 47, 1159-1170 Contents Contents
1. Introduction The work of this thesis has been carried out from October 1999 to December 2002 at the Department of Structural Research of the Max Planck Institute for Biochemistry. The scope of this thesis is to give a structural and dynamic characterization of fluorescent proteins and to determine the structure of the N-terminal domain of the adenylyl cyclase-associated N and 1H-15N nuclear magnetic protein (CAP-N) from Dictyostelium discoideum. The resonance (NMR) studies done on the green fluorescent protein (GFPuv) and its mutant His148Gly show a substantial conformational flexibility and a strong impact on fluorescence properties of GFPs and suggests the presence of two conformations in slow exchange on the NMR time scale in this mutant. The structure of the CAP-N determined in the present thesis is the first for an N-terminal domain of any CAP and can be useful for defining specific functions for these domains.
The family of fluorescent proteins is one of the most widely studied and exploited protein families in biochemistry and cell biology which represents basic tools for monitoring gene expression, protein localization, movement and interaction in living cells (Chalfie et al., 1994; Tsien, 1998; Garcia-Parajo et al., 2000). This kind of monitoring is minimally perturbing the cell under investigation. Fluorescent proteins provide also a system rich in photophysical and photochemical phenomena of which an understanding is crucial for the development of new and optimized variants of the GFP (Lossau et al., 1996). Its amazing ability to generate a highly visible, efficiently emitting internal fluorophore is intrinsically fascinating and tremendously valuable (Tsien, 1998; Ward, 1981). Up to now the family of fluorescent proteins comprises about 30 cloned and spectroscopically characterized proteins (Labas et al., 2002; Ormoe et al., 1996; Yang et al., 1996). High-resolution crystal structures of GFPs offer opportunities to understand and manipulate the relation between protein structure and spectroscopic function. But there is still a continuing effort to develop by mutagenesis and engineering new GFP variants with better properties and to open up new ways to monitor protein-protein interactions (Voityuket et al., 1998).
Cyclase associated proteins (CAPs) are multifunctional proteins with several structural domains, that are present in a wide range of organisms. Two domains are highly conserved;
one of them helps to activate the catalytic activity of the adenylyl cyclase in the cyclasebound state through interaction with Ras, which binds to the cyclase in a different region (Hubberstey & Mottilo, 2002). The second conserved domain of CAP can bind monomeric
actin and thus CAP has also a cytoskeletal function. CAP is involved in the Ras/cAMPChapter 1 Introductiondependent signal transduction and most likely serves as an adaptor protein translocating the adenylyl cyclase complex to the actin cytoskeleton. The CAP of Dictyostelium discoideum is involved in the microfilament reorganization at anterior and posterior plasma membrane regions. But the full CAP function still presents a mystery. CAP interaction with actin maybe controlled through phospholipid binding in a similar fashion as profilin (Gottwald et al., 1996). Phospholipid interactions may regulate the interaction between the amino- and carboxyl-terminal domains. The specific residues within CAP that interact with actin must be defined, and the possibility that actin binding may involve interaction between the amino and carboxyl termini cannot be excluded (Wesp et al., 1997).
This intoduction is followed by Chapter 2, which provides a short introduction to nuclear magnetic resonance (NMR) spectroscopy and X-ray crystallography – the two most powerful methods of structural studies. The used materials and methods are described in Chapter 3. Chapter 4 of this thesis describes the work carried out on GFPs – dynamic studies and the crystal structure determination of DsRed. Chapter 5 deals with the structure determination of the N-terminal domain of CAP and specific functions of this domain and the whole protein.
2. Methods for Structural Studies
In this chapter the two most powerful techniques for structural studies are presented:
X-ray crystallography and nuclear magnetic resonance (NMR) spectroscopy. X-ray crystallography is the main method for elucidation of the three-dimensional macromolecular structures at the atomic level. NMR, which has the disadvantage of being more time consuming and restricted to smaller molecular weight proteins (up to 30 kDa), has, however, many advantages in comparison to X-ray crystallography, as it can be useful for dynamic studies and can provide many other useful information about the protein in solution (Sali, 1998).
2.1 X-ray Crystallography 2.1.1 General Background The central role of protein crystallography in structural analysis is illustrated by the increasingly high number of structures determined by X-ray diffraction techniques deposited in the Brookhaven Protein Data Bank (PDB). Until December 2002 a total of 19464 protein structures have been deposited, 16448 of them have been determined with the help of X-ray diffraction techniques and 3021 by NMR (Berman et al., 2000). Although crystallography, when compared with NMR gives a more static description of the macromolecular structures, there are no limits in the size of the molecule to be analyzed. This makes X-ray crystallography the method of choice for studying large macromolecular complexes at the atomic level. Otherwise, the structures determined by these two techniques are not much different. Differences arise when exposed protein regions are hindered by contacts in the crystalline lattice. In recent years, however, the advances in radiation detection and computing power have made it possible to study enzyme catalysis and associated conformational changes in the crystalline state. The so-called time-resolved crystallography overcomes this major disadvantage of X-ray protein crystallography (Moffat, 2001).
The main problem in X-ray crystal structure analysis is to find not only the amplitudes of all the diffracted X-rays (usually known as reflections), but also their phases. Knowledge of both, amplitudes and phases, allows the reconstitution of the electron density of the crystal.
The amplitudes can be deduced from the intensities of the diffracted X-rays but the phases cannot be directly measured. This is known as the “phase problem” (see below). To determine Chapter 2 Methods for Structural Studies proteins three-dimensional structure, one has to first obtain good diffracting crystals of the protein in what is mainly a trial-and-error process.
2.1.2 What is a Protein Crystal?
Crystals are regular, three-dimensional arrays of atoms, ions, molecules or molecular assemblies. Ideally, a crystal can be described as an infinite array in which the building blocks (the symmetric units) are arranged according to well-defined symmetries (forming one of 230 space groups) into unit cells that are repeated in three dimensions by translation. Proteins and nucleic acids do not crystallize in space groups with inversion symmetries because they are composed of enantiomers (L-amino acids and D-sugars, respectively), thus reducing the number of possible space groups to 65.
Why are protein crystals needed for X-ray three dimensional structure determination?
In practice, the reflection pattern of a single molecule cannot be observed, but only that of many in an ordered crystalline array. The maximum achievable resolution of any microscopic technique is limited by the applied wavelength. The radiation needed to analyze atomic distances lies within the spectral range of X-rays that are used in crystal studies because their wavelength (1.542 Å for copper Kα radiation) is comparable to the planar separation of atoms in a crystal lattice.
2.1.3 Crystal Growth
Growth of high quality single crystals is the basis of X-ray structure determination. It is also sometimes the primary difficulty in the determination of a macromolecular structure.
Proteins and nucleic acids are structurally dynamic systems, often micro-heterogenous, whose properties are influenced by environmental conditions such as pH, temperature, ionic strength and a number of other factors. Protein purity and homogeneity is essential to the growth of single protein crystals. Crystallization of macromolecules is a multiparametric process involving three main steps: nucleation, growth and cessation of growth. It is indispensable for crystallization to bring the protein to a supersaturated state (Fig. 2.1), which will force the macromolecules into the solid state - the crystal.
Herein, C0 is the solubility of the protein in water, I is the ionic strength (I=½ Σi cizi2 with concentration c and charge z of the ion), Z is the total charge of the protein, a is the sum of the radii of the protein and the salt ion, Ks is an empirical salting-out constant and the constants A and B depend on temperature and dielectricity. For high salt concentrations, the salting-out term will be dominant, and it can be derived that ions with a high charge density will have stronger influence on the solubility, as described in the Hofmeister series (Hofmeister, 1888).
Ks for anions: citratetartratesulfateacetatechloridenitrate Ks for cations: Li+K+NH4+Ca2+Sr2+Ba2+Al3+ In addition to salts, commonly used precipitants are polyethylene glycols or organic solvents like ethanol, isopropanol or methylpentane diol. At low ionic strength (low ionic concentration), the solubility of protein is higher if the amount of electrolytes is increased – termed “salting in”. At high ionic strength the ions start to compete with each other for water molecules, resulting in a decrease in solubility. This process is known as “salting out”. The crystallographer can shift the equilibrium to supersaturation by increasing or reducing the ionic strength of the protein solution. The Hofmeister series indicates that at a high ion concentration small ions with a high charge are generally most effective. For proteins, much larger and with complicated surface charge distributions, this theory is not sufficient to Chapter 2 Methods for Structural Studies explain the phenomenon of crystal formation. For instance, the difference between the free energies (∆G) of the solid and soluble states will indicate the favorable trend. In general the electrostatic interactions in crystals are much more favorable than the ‘‘interactions’’ in amorphous precipitates. As shown in Fig 2.1, crystal growth can be divided into two steps.