Forskningscenter: inSPIN Center for Insoluble Protein Structures

Publiceret April 2014

It is well established that one way to understand function of biomolecules goes via determination of the structure and dynamics of the molecule in native or close to native conditions. This strongly motivates the development of tools enabling establishment of such information and in using these tools to obtain structural information on biological systems with atomic resolution. To increase our knowledge of complex biological systems, it is of interest to develop new techniques where it is possible to study the structure-function relationship as well as the biophysical properties of the involved proteins/biomolecules in their native functional environment. The aim is to obtain a higher degree of understanding on vital biological and physiological processes.

Structural Studies of "Insoluble" Proteins

The most widely used tools for obtaining high-resolution structures of biomolecules are X-ray crystallography and liquid-state NMR (nuclear magnetic resonance) spectroscopy. These methods, however, may have intrinsic difficulties in studying proteins in their native, functional, heterogeneous environments.

X-ray crystallography is dependent on the ability of the molecule (or molecular system) to form high-quality 3D crystals. This, however, is for some classes of proteins not a trivial task. In the same way, liquid-state NMR is dependent on fast molecular tumbling and the technique is only applicable if the protein of interest it is not too big (typically less than 30-100 kDa) and can be dissolved (and maintain its structure) in a suitable solvent at high concentration. To meet these demands both X-ray crystallography and liquid-state NMR spectroscopy most typically require the molecular system to be taken out of their native environment, with the potential consequences it may have on structure and function.

To increase our understanding of complex biological systems, it is of interest to develop new techniques where it is possible to study the structure-function relationship as well as the biophysical properties of the involved proteins/biomolecules in their native functional environment. With this aim, the research center inSPIN (center for inSoluble ProteIN structures) was established in 2005 and extended for a new 5-year period in 2010 supported by the Danish National Research Foundation. The idea of the center is to use a truly interdisciplinary set up exploiting the competences of four, then later five well-established research groups to face the challenge of contributing to the understanding of proteins in native, insoluble environments. It was envisaged that such understanding is not only of interest on the basic science level, but also may serve as inspiration to biological intervention, including rational drug design, development of biomarkers, and obtaining fundamental insight into the fascinating nanomachineries in Nature. All five groups are at present situated at Aarhus University, and are associated to the interdisciplinary nanoscience center (iNANO) while also being affiliated with the Department of Chemistry or Department of Molecular Biology and Genetics.

A new and rapidly evolving technique for solving structures of proteins in “solid” phase is solid-state NMR spectroscopy. This method is not dependent on either formation of high-quality 3D crystals, nor fast molecular tumbling. The interplay/complementarity between various methods for atomic resolution structure determination and potential targets is illustrated in Fig. 1. Relative to the much more widespread liquid-state NMR spectroscopy method, the solid-state variant faces challenges such as the presence of orientation-dependent (i.e., anisotropic) nuclear spin interactions. These interactions cause at one hand severe broadening of the NMR signals, thereby destroying the resolution and sensitivity, and on the other hand they may provide important structural information. The first problem may be coped with by “simulating” fast molecular motion by using so-called magic-angle spinning where the protein sample is spinning very fast (thousands of revolutions per second) around an axis inclined by 54.7 degrees relative to the magnetic field. Supplemented by synchronized radio-frequency field irradiation, one may now engineer advanced experiments that on our control turn the anisotropic interactions on and off and in this manner get high-resolution spectra (anisotropic interactions turned off) and structural information (anisotropic interactions turned on). Exploiting such techniques, solid-state NMR may open new avenues for detailed structure and dynamics determination, offering an enormous potential for applications in structural biology addressing complex heterogeneous samples. The number of potential targets (see Fig. 1) is huge considering that the so-called insoluble proteins (e.g., membrane proteins) encompass more than 1/3 of all proteins including many of the known primary drug targets. The task of making important contributions to development of this niche of structural biology cannot be taken up by solid-state NMR spectroscopy alone, but requires the outlined combination with protein chemistry, biophysics, organic chemistry, and molecular modeling.

Fig. 1: Structures of soluble proteins are solved in large numbers by X-ray crystallography and liquid-state NMR. However, very little is known on the structures of insoluble proteins, although these are of utmost importance for the pharmaceutical industry. Solid-state NMR is a very promising candidate for opening of systematically structure solution for insoluble proteins.

Since biological solid-state NMR is still a fairly new technique, one of the aims of inSPIN – alongside providing actual structural information on relevant biological systems – is to contribute to the general method development in the field in order to obtain tools for the study of proteins in their natural heterogeneous environment, and for analyzing protein interactions from the atomic level (sub-nanometers) to MRI (micro to millimeters) in vivo localization.

The method development is very important since biological solid-state NMR suffers in general from low sensitivity and resolution, implying that the applicability of the technique is dependent on the establishment of optimized methods offering the highest possible sensitivity and resolution. The biological solid-state NMR group headed by Prof. Niels Chr. Nielsen, who has been the director of inSPIN and also the interdisciplinary nanoscience center (iNANO), has for many years focused on the development of solid-state NMR spectroscopy. The NMR group also includes Prof. Thomas Vosegaard and Assoc. Prof. Frans Mulder.

Protein expression and purification are key elements in the inSPIN research, which are carried out in the laboratories headed by Prof. Jan J. Enghild and Assoc. Prof. Torsten Kristensen with access to both automated bioreactors and large-scale fermentors using a number of running expression systems. The expertise of these groups also includes full proteomics analysis and state-of-the-art mass spectrometry.

Synthetically produced peptides are also subject to extensive studies in inSPIN, with the underlying peptide synthesis carried out in the organic chemistry group headed by Prof. Troels Skrydstrup. Also, small organic compounds which act as ligands are of major interest, and such molecules are produced using sophisticated synthetic methods. Another area of focus is the use of new techniques for incorporation of different isotope labels suitable for NMR spectroscopy into the synthesized ligand. One of the techniques utilized and developed in the synthesis group is the use of metals as coupling reagents in the making of carbon-carbon bonds.

To complement the NMR experiments, and to identify both different kinds of molecular interactions as well as low-resolution structural details, a large number biophysical characterization methods are typically carried out, including circular dichroism, FT-infrared spectroscopy, fluorescence, electron microscopy, atomic force microscopy, light scattering etc. This is accomplished in the Biophysics group headed by Prof. Daniel E. Otzen.

Molecular dynamics (MD) simulations are highly relevant in the field of insoluble proteins in order to capture biological relevant dynamics, and to explain the function and physiology of these. Also molecular interactions on longer time scales may be studied using MD simulations based on experimental data, and MD is an integral part of many inSPIN projects. Prof. Birgit Schiøtt is head of the MD group.

Applications – examples

From the name of the center, it is clear that the main targets for inSPIN are proteins that in their native functional state are in an insoluble environment. To be more specific, the major part of the research is focused on peptides/proteins in three different classes: 1. Peptides/proteins that form amyloid fibrils, 2. Membrane bound peptides/proteins, 3. Proteins situated in the extracellular matrix.

Amyloid fibrils are signature of many very serious diseases and disorders, including Alzheimer’s disease, Parkinson’s disease, Huntington’s disease, various prion diseases, and pathological conditions such as diabetes 2. These diseases have in common that they involve so-called misfolding of proteins into insoluble amyloid deposits or plaques. The process of the misfolding has been subject to considerable attention during the past few years with respect to folding pathways, cytotoxicity of the various species in the folding process, and amyloid formation as a means of biological storage as obtained through biophysical characterisation, molecular level structures, and development of potential fibril markes and/or inhibitors.

To obtain a deeper understanding of the fibril formation process and general fibril properties, we have studied numerous amyloid systems relating to dementia, diabetes, corneal dystrophies, and functional bacterial amyloids. As an example, we have thoroughly analysed the ten-residue fragment amylin20-29 (SNNFGAILSS) which is known to be ‘the fibril core’ of amylin. The 37 amino acid peptide amylin forms fibrils that are deposited in the pancreas of patients suffering from diabetes 2. In a pioneering study, utilizing symmetry approaches in connection to the peak patterns in solid-state NMR spectra, the 3D structure of the amylin20-29 fibril was solved with very high resolution. The peptide folds into anti-parallel β-sheets with a so-called hetero zipper and it displays a slight twist around the fibril axis (see Fig. 2).

Fig. 2: Three different representations on the structure of amylin20-29 solved by solid-state NMR spectroscopy.

Fibrils are known to be stabilized by both hydrophobic and electrostatic side-chain interactions, however, also the regular hydrogen-bonding pattern of the β-sheets indicates important backbone interactions. To explore the role of such interactions as well as the well-known triggering effect from mutation, the inSPIN groups have synthesized a number of different amylin20-29 analogs with alterations in either backbone or side chains of the peptide. From thioflavin T fluorescence measurements – a technique exploiting that the dye molecule binds amyloid fibrils with the fluorescence signal intensity increasing upon formation of amyloid fibrils – it turned out that the peptides with altered backbone indeed did not cause amyloid fibril deposits. In contrast, amylin20-29 variants with only side chain modifications all demonstrated ability to form fibrils as revealed also by transmission EM pictures (Fig. 3), giving a macroscopic representation of various fibril types also associated with subtle structural changes on the molecular level. Such changes are here explored using solid-state NMR spectroscopy (with atomic resolution information) and AFM measurements to provide information about the assembly of the fibrils on a nano- to micrometer scale.

Fig. 3: A series of transmission electron microscopy pictures of amylin20-29 amyloid fibrils obtained using different modifications of the peptide sidechains and different fibrillation conditions and side chain modifications.

One of the major goals in inSPIN is to establish detailed structural information as described above to facilitate systematic development of compounds for inhibiting fibril formation and/or fibril decomposition, as well as “biomarker” compounds for early-stage disease detection/localization. This detection/localization can be carried out by magnetic resonance imaging (MRI) (see Fig. 4a; 16.4 T MRI image of the brain of a transgenic Alzheimer’s mouse) and positron emission tomography (PET). In such studies, it is essential to identify and structurally characterize the interactions between the fibrils (or early-stage protofibrils) and the potential binder/biomarker. For this purpose, the full competence span of the inSPIN center is brought into action, in the sense that advanced organic chemistry synthesis is required to synthesize and incorporate specific labels (e.g. 13C, 2H) into the binding compounds, biophysical methods are used to characterize binding stoichometries and kinetics, while molecular dynamics simulations supplement NMR structural analysis of the fibril structure to model the binding interaction over time. Taking this analysis as inspiration, solid-state NMR can be used to validate the models by measuring accurate binding distances between fibril and binder, and in that way determine the specific binding site in terms of structure and dynamics (see structural model in Fig. 4b). This information may then be used to inspire synthesis of novel organic compounds with improved binding characteristics.

Fig. 4: (a) MRI image of the brain of a transgenic mouse with Alzheimer's disease. (b) Molecular dynamics simulation showing the binding interactions between amylin20-29 and FSB, which is a strong candidate as a general fibril binder/marker.

inSPIN has many other targets than amyloid fibrils. For example, a major effort is devoted to the study of antimicrobial peptides which are small membrane proteins being part of the bacterial defense systems of numerous higher living forms. Such peptides may potentially offer interesting new possibilities in the battle against infections, being of increasing interest due to the risk of resistance towards existing small molecule antibiotics. One such example is given in Fig. 5a showing a structural model of the antimicrobial peptide alamethicin determined using liquid- and solid-state NMR spectroscopy in combination with molecular dynamics simulations. The membrane penetrating alamethicin multimer lyses the membrane (i.e., channels as shown in the right hand side of Fig. 5a punctures the bacterial cell membrane) and thereby kills the bacteria towards which the peptide has affinity. For the channel formation to happen (see Fig. 5b) a relatively high concentration of the antimicrobial peptide is required – thereby calling for large amounts of this “drug”, and with the quantity increasing the risk for undesired side effects also increases. To improve the effect at much smaller doses, we have chemically synthesized a “pre-assembled” nano-channel by attaching a number of alamethicin molecules to a cyclodextrin ring using so-called click chemistry. In this manner, the active concentration can be substantially lower and much longer opening times for the channels are ensured (see single-channel recordings in Figs. 5b and 5c, right side, where the lower and higher levels corresponds to open and closed channel, respectively).

Fig. 5: (a) Structural model of alamethicin as a monomer in a lipid bilayer (left) and as a multimeric channel (right). (b) Ion channel formed by alamethicin monomers (left) with single channel current recordings showing open and closed states of the channel (right). (c) Ion channel formed by alamethicin-cyclodextrin conjugates (left) with single channel current recordings showing open and closed states of the channel (right).

With a totally diff erent perspective, inSPIN also works with photoreceptors to get long-term inspiration to design novel nanoscale solar energy receiver systems. Such projects also challenge one of the goals of inSPIN, namely the establishment of detailed information about structure and dynamics for proteins (or complexes) residing in native, functional, and typically heterogeneous surroundings. The current target is the chlorosome of green sulphur bacteria. This photoreceptor system has an enormous capability of extracting energy from low intensity light sources, and in nature this ability is exploited by the green sulfur bacteria found in lakes at depths where only small amounts of light is available. Our specific interest is the baseplate antenna system, which in the chlorosome of Chlorobaculum tepidum is the monolayer attached to a water soluble Fenna-Matthew-Olsson (FMO) protein in a complex with pigment and carotenoid molecules (see Fig. 6a).

Aimed at simplifying solid-state NMR studies of the native baseplate system rather than the full chlorosome, a simplified mutant – the so-called carotenosome (Fig. 6b) – was expressed. The interior of this variant does not contain large amounts of the, in this context, uninteresting Bhcl c pigment, thereby significantly increasing the concentration of the baseplate in our samples which is important for sensitivity reasons. Furthermore, the protein complexity in the lipid monolayer (outside the baseplate) is reduced implying that more than 90% of the protein in the carotenosomes is the 59 amino acid protein CsmA located in the baseplate. In an ongoing work, the aim is to solve the high resolution structure of CsmA in this truly heterogeneous environment as well as the organization of the pigment molecules in the baseplate using solid-state NMR spectroscopy. As a primer to this investigation, the structure of isolated CsmA in solution was earlier determined using liquid-state NMR (Fig. 6e). This structure, being representative of most protein structure determinations, does, however, not necessarily reflect the details of the protein as located in the baseplate with different environment than an artificial organic solvent.

For detailed solid-state NMR structural analysis the molecules need to be isotopically enriched with 13C and 15N (in nature, 99% of the carbon is 12C and 99.7% of the nitrogen is 14N; these isotopes are not suitable for NMR studies), being accomplished by growing the carotenosomes in bacteria fed with 13C labeled glucose and 15N labeled ammonium chloride. In this setting, all carbon atoms in the carotenosomes are 13C labeled, implying that the 13C signals from CsmA are accompanied by a large number of additional 13C signals from lipids, carotenoids etc. (see Fig. 6c). To circumvent the resulting heavy overlap of signals from wanted (baseplate components) and unwanted (carotenoids, lipids, etc.) components and to be able to identify the CsmA protein signals, the fact that CsmA also holds 15N labels in the protein backbone is exploited. Through transfer of magnetization from protons to 15N prior to further transfer to the 13C nuclei, it is ensured that only 13C resonances from the proteins are visible. This technique greatly simplifies the spectra (Fig. 6d), and opens for further NMR analysis. Such an analysis is carried out using a large number of 2D (Fig. 6d) and 3D solid-state NMR spectra and from these it is possible to uniquely identify almost all carbons and nitrogens in CsmA. The chemical shift values of a protein are quite precise measures of the secondary structure elements, and from the assigned chemical shift values clear indications of an alpha helical structure of CsmA are seen. The structure bear some resemblance to the liquid-state NMR structure (Fig. 6e), however, with important perturbations induced by intermolecular interaction between CsmA proteins as well at between CsmA and the pigment BChl a in the baseplate. Calculations on the fine details of the full CsmA structure and the pigment organization based on dipole-dipole interactions between 13C nuclei and also taking advantage of complementary data from cryo electron microscopy and oriented circular dichroism have been carried out. The quite spectacular results of this study will be published elsewhere shortly.

Fig. 6: (a) Chlorosome, (b) Carotenosome. Dark green: pigments and carotenoids. Orange: carotenoids. Black: lipids. Magenta: Proteins. Light green: CsmA binding pigment. Light blue: FMO proteins. Dark blue: reaction centers in the cytoplasmic membrane. (c) 13C cross-polarization magic-angle spinning (CP/MAS) spectrum of uniformly 13C labeled carotenosomes. (d) Overlaid 15N filtered (black) and non-filtered (cyan) 13C-13C correlation spectra. Boxes zoom on regions with cross peaks between Ca and C' (left) and Ca and aliphatic side chain carbons. (e) Solution state structure of CsmA.

The study of CsmA in the photo receptor is unique in the sense that this is one of the first studies where atomic-resolution structure information is obtained for a protein in a truly heterogeneous and functional environment, thereby forming an important stepping-stone to the next era of structural biology addressing structures of components in complex, heterogeneous, and functional biological systems. In these types of studies, it is of utmost importance to work in an interdisciplinary set up with access to a large number of complementary competences. As shown in the examples above inSPIN is well suited for such challenges, and the work on exploring the structural details of insoluble proteins in heterogeneous environments is under rapid development.


1. J. T. Nielsen, M. Bjerring, M. D. Jeppesen, R. O. Pedersen, J. M. Pedersen, K. L. Hein, T. Vosegaard, T. Skrydstrup, D. E. Otzen, and N. C. Nielsen, Unique identification of supramolecular structures in amyloid fibrils by solid-state NMR spectroscopy, Angew. Chem. Intl. Ed. 2009;48,2118-21.

2. M. Andreasen, S. Nielsen, T. Mittag, M. Bjerring, J.T. Nielsen, S. Zhang, E.H. Nielsen, M. Jeppesen, G. Christiansen, M. Dong, N.C. Nielsen, T. Skrydstrup, and D. Otzen, Modulation of Fibrillation of hIAPP Core Fragments by Chemical Modification of the Peptide Backbone, Biophys. Biochem. Acta 2012;1824,274-85.

3. N.V. Kulminskaya, M.Ø. Pedersen, M. Bjerring, J. Underhaug, M. Miller, N.-U. Frigaard, J.T. Nielsen, and N.C. Nielsen, In situ solid-state NMR of protein complexes in heterogeneous biological membranes: The baseplate antenna-complex of Chlorobaculum tepidum, Angew. Chem. Intl. Ed 2012;51,6891-95.

4. M.Ø. Pedersen, J. Underhaug, J. Dittmer, M. Miller, and N.C. Nielsen, The three-dimensional structure of CsmA: A small antenna protein from the green sulfur bacterium Chlorobium tepidum, FEBS Lett. 2008;582, 2869-74.