Solid-State NMR Studies of Biomineralization Peptides and Proteins

Chapter 4

Downloaded by UNIV OF ARIZONA on December 14, 2012 | http://pubs.acs.org Publication Date (Web): December 12, 2012 | doi: 10.1021/bk-2012-1120.ch004

Solid-State NMR Studies of Biomineralization Peptides and Proteins Adrienne Roehrich,1 Jason Ash,1,4 Ariel Zane,1 David L. Masica,2 Jeffrey J. Gray,2 Gil Goobes,1,3 and Gary Drobny*,1 1Department

of Chemistry, University of Washington, Box 351700, Seattle, Washington 98195 2Department of Chemical and Biomolecular Engineering, Johns Hopkins Univeristy, Baltimore, Maryland 21218 3Present Address: Department of Chemistry, Bar Ilan University, Israel 4Present Address: Center for Materials Science and Engineering, Merck, Rahway, New Jersey 07065 *E-mail: [email protected]. Phone: (206) 685-2052. Fax: (206) 685-8665

Nature has evolved sophisticated strategies for engineering hard tissues through the interaction of proteins, and ultimately cells, with inorganic mineral phases. The remarkable material properties of shell, bone and teeth thus result from the activities of proteins that function at the organic-inorganic interface. A better understanding of the biomolecular mechanisms used to promote or retard the formation of mineral-based structures could provide important design principles for the development of calcification inhibitors and promoters in orthopedics, cardiology, urology, and dentistry. In addition to investigating the molecular-level basis for the recognition of biomineral surfaces and the control of hard tissue growth by proteins, the development of materials using biomimetic principles has potential applications in catalysis, biosensors, electronic devices, chromatographic separations, to name only a few. Despite the high level of interest in elucidating and controlling the structure of proteins at material and biomineral interfaces, there is a decided lack of molecular-level structure information available for proteins at biomaterial interfaces in general, and in particular for mammalian proteins that © 2012 American Chemical Society In Proteins at Interfaces III State of the Art 2012; Horbett, T., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2012.


directly control calcification processes in hard tissue. The most fundamental questions regarding the secondary and tertiary structures of proteins adsorbed to material surfaces, how proteins catalyze the formation of biomineral composites, or how proteins interact at biomaterial interfaces, remain unanswered, largely due to a lack of methods capable of providing high resolution structural information for proteins adsorbed to material surfaces under physiologically relevant conditions (i.e. fully hydrated). In order to develop a better understanding of the structure and interactions of proteins in biomaterials, we have begun to utilize solid-state NMR techniques to determine the molecular structure and dynamics of proteins and peptides on inorganic crystal surfaces and within biomineral composites. In this review, we will highlight recent work that is providing insight into the structure and crystal recognition mechanisms of a salivary protein model system, as well as the structure and interactions of a peptide which catalyzes the formation of biosilica composites.

Introduction and Background Biomineralization is the process by which living organisms control the formation of inorganic materials like hydroxyapatite, calcite and silica into highly intricate and organized structures (1). Nature has evolved sophisticated strategies for engineering hard tissues through the interaction of proteins, and other biomolecules, with inorganic mineral phases. The remarkable material properties of bone, shells and teeth thus result in part from the activities of proteins that function at the organic-inorganic interface. A better understanding of the biomolecular mechanisms used to promote or retard mineral growth could provide important design principles for the development of calcification inhibitors and promoters in orthopedics, cardiology, urology, and dentistry (2, 3). A better understanding of how these proteins recognize and assemble in bioactive form on inorganic mineral phases could also aid in the development of surface coatings to improve the biocompatibility of implantable biomaterials and for hard tissue engineering and regeneration technologies. In addition to biomedical applications, current interest in biomineralization also derives from its potential applications in electronics, catalysis, magnetism, sensory devices, and mechanical design (4–12), where in contrast to anthropogenic synthesis of hard materials, which requires extremes in temperature, pressure and pH, biological organisms accomplish impressive feats of hard tissue engineering at ambient temperature and at physiological pH. 78 In Proteins at Interfaces III State of the Art 2012; Horbett, T., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2012.


At the level of fundamental science, it is important to note the paucity of molecular structure information available for biomineralization proteins in general, and in particular for mammalian proteins that directly control calcification processes in hard tissue. Even the most fundamental questions about how the proteins interact at the protein-biomineral interface are yet to be extensively addressed experimentally. These questions include: what is the general structure and orientation of proteins on mineral surfaces or within biomineral composites; what amino acid side chains or structural motifs do proteins use to interact with inorganic phases? In order to develop a better structure-function level understanding of protein-crystal molecular recognition, we have begun to utilize solid-state NMR (ssNMR) techniques to determine the molecular structure of proteins and peptides on calcium phosphate surfaces and within biosilica composites. In this review, we will highlight recent work that is providing insight into the structure and crystal recognition mechanisms of an acidic human salivary phosphoprotein statherin. Providing contrast to the interactions of statherin with hydroxyapatite crystal surfaces, we will also describe the application of solid state NMR to the study of the structure of basic peptides derived from the biosilicification protein silaffin and embedded within biosilica nanoparticles. Silaffin and peptides derived from silaffin catalyze formation of biosilica composites when added to solutions of silicic acid. Here the objective is to use ssNMR to determine the structure of proteins and peptides within biosilica composites, the interactions of these peptides with the surrounding silica, and ultimately the relationship between silaffin peptide secondary structure and biosilica composite morphology.

Principles of Solid-State NMR Solid state NMR spectra of spin-½ nuclei are dominated by two magnetic interactions, the direct nuclear dipolar interaction and the chemical shift anisotropy (CSA). The dipolar interaction has a straightforward structural interpretation because the dipolar coupling constant is proportional to the inverse cube of the internuclear distance. (See Figure 1, top). As discussed below, although the isotropic chemical shift interaction is valuable as an empirical probe of protein structure, in solid samples the chemical shift anisotropy makes the dominant contribution to the NMR line width. Although the dipolar coupling between two protons separated by a few Angstroms exceeds that proton chemical shift anisotropy, in systems composed of coupled 13C spins, the magnitude of the CSA may exceed the magnitude of the dipolar interaction, and under such circumstances the 13C NMR line shape will be a broad “powder pattern” (see Figure 1, bottom). Therefore the dipolar coupling constant may not be easily discerned in spectral data when the CSA exceeds the dipolar interactions by orders of magnitude. So it is necessary to use NMR methods which suppress one of these interactions thus enabling straightforward detection of the other.

79 In Proteins at Interfaces III State of the Art 2012; Horbett, T., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2012.


Figure 1. Top: The direct interaction between two spin magnetic dipole moments μ depends upon the angle θ between the internuclear vector and the direction of the magnetic field m and the inverse cube of the internuclear distance r. Bottom: The NMR “powder pattern” lineshape of a spin-½ nucleus where the dominant magnetic interaction is the chemical shift anisotropy (CSA). Physical rotation of the sample around a goniometer oriented at an angle θm = 54.74° relative to the magnetic field (i.e. Magic Angle Spinning (13), MAS, see Figure 2, left) coherently averages spatial parts of the spin magnetic interactions that transform as second rank tensors. If the rate of spinning is less than the anisotropy that underlies the broadened line shape, the powder pattern breaks up into spinning side bands, which appear in Figure 2 as the satellite lines about the isotropic chemical shift signals that are regularly spaced by intervals equal to the spinning rate. Since both the dipolar and the chemical shift Hamiltonians transform as second rank tensors in Cartesian space, MAS removes the effect of both interactions from solid-state NMR spectra. The objective of dipolar recoupling is to use MAS and radio frequency (r.f.) irradiation in synchrony, with the ultimate objective of suppressing the chemical shift while preserving the dipolar coupling. Because information on internuclear distances can be recovered using dipolar recoupling techniques, these pulse sequences are widely used in structural studies of molecular solids. Solid-state Rotational Echo Double Resonance (REDOR) NMR (14) is a high resolution technique employing MAS used to observe heteronuclear dipolar coupling. This method can be applied for either inter- or intramolecular interactions. Appropriate labels, in the case of the work described later in this chapter 13C{15N} and 13C{19F}, are placed either in i, i+4 residues or on selected side chains of amino 80 In Proteins at Interfaces III State of the Art 2012; Horbett, T., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2012.


acids exposed to the surface. A specific pulse sequence with phase alternation and dependent upon the number of rotor cycles gives carbon signal intensity, S. A REDOR plot consists of S/S0 vs time, where S0 is obtained without dephasing pulses in the pulse sequence. Experimental data is fit with a simulation curve to determine distances.

Figure 2. Left: Geometric arrangements in the Magic Angle Spinning (MAS) experiment. An ensemble of randomly oriented dipole-coupled spin-½ nuclear pairs I and S has a NMR powder pattern line shape shown in Figure 1 if the CSA is non-axial and greater than the dipolar coupling If the spin pairs ensemble is spun rapidly around a goniometer whose axis is inclined at θm = 54.74° the powder pattern breaks up into side band peaks arranged at intervals equal to the spinning rate around a fundamental NMR frequency corresponding to the isotropic chemical shift. Right: Increasing the spinning rate increases the separation between the side bands and decreases the side band intensities. In the limit that the spinning rate exceeds the magnitude of the CSA, the side bands are completely suppressed and only the NMR signal appearing at the isotropic chemical shifts of the I and S spins remain.

An alternative strategy that has come into wide use in recent years is to use chemical shifts to empirically determine protein secondary structure. The isotropic chemical shifts of the 13CO (i.e. the backbone carbonyl 13C), 13Cα, and 13Cβ spins are known to be dependent upon local secondary structure of the protein. By comparing experimentally observed chemical shifts to an NMR database of known structures, software packages such as TALOS+ are able to predict the φ/ψ torsion angles of a given residue (15, 16). TALOS+ (Torsion Angle Likeliness Obtained from Shift and Sequence Similarity) is a program based on a 200+-protein database for which complete or nearly complete heteronuclear resonance assignments and high resolution X-ray coordinates are available in the protein database (PDB) from the research collaboratory for structural bioinformatics (RCSB) and the biological magnetic resonance bank (BMRB) (15, 16). The TALOS+ software package is based on the principle that homologous sequences give similar shifts (15, 16). The combination of solid-state NMR with backbone torsion angle assignments by the TALOS/+ program has been used to 13C


model several peptides including protein G (17–19), ubiquitin (20), amyloid β (21–23), and kaliotoxin (24). In the following sections we will show how these two types ssNMR measurements, dipolar couplings and chemical shifts, can be used to determine the structure of proteins at biomineral interfaces.


Solid-State NMR Studies of Salivary Statherin on HAP Crystal Surfaces Statherin is a small 5.68 kDalton phosphoprotein, and the only salivary protein that has been found to inhibit both the nucleation and the growth of hydroxyapatite (HAP) (25–30). Statherin is expressed at its active length and post-translationally phosphorylated at serines 2 and 3 (31). It is then excreted by the sublingual and submaxillary salivary glands where it travels to the mouth. Statherin is composed of 43 amino acid residues with primary sequence:

In addition to two phosphoserines (i.e. pS) statherin contains four aspartic/glutamic acid residues and four basic amino acids, almost all of which occur close to the N-terminus. Previous studies of statherin fragments by Nancollas and coworkers (32, 33) showed that the N-terminal 15 amino acid fragment (SN-15, H2N-DpSpSEEKFLRRIGRFG-COOH) is essential for binding to HAP crystal surfaces. SN-15 contains two phosphoserines, three acidic residues, and four basic residues. Removal of the DpSpSE moiety (i.e. resulting in SN-11, EKFLRRIGRFG) reduces the HAP binding affinity by a factor of 4.5 compared to SN-15. Mutation of both phosphoserines to simple serine reduces the HAP binding affinity by a factor of almost nine, while mutation from serine to aspartic acid restores the binding affinity to 70% that of SN-15, indicating that the phosphoserine residues and acidic amino acid side chains are important for the binding of statherin to HAP. Initial ssNMR studies of HAP-bound statherin measured a number of torsion angles and internuclear distances within the N-terminal 15 amino acid segment of native statherin, and within the 15 amino acid SN-15 peptide. These data indicate that the N-terminus in HAP-adsorbed statherin is an α-helix (34–36). Subsequent solid state NMR studies of labeled statherin molecules yielded measurements of backbone torsion angles (residues 33-34) and i-i+4 carbonyl-amide distances (residues 34-38) indicating that the C-terminus receptor site for bacterial adhesion adopts an α-helical conformation in statherin adsorbed to hydroxyapatite crystals (37). See Figure 3. In order to define tertiary folding, we also measured long range distances between carbonyl carbons in the C-terminus (residues 33-34) and a fluorine nucleus incorporated into a proline ring (residue 23) showing that this motif closes back onto the protein’s proline-rich region (16-28) through a series of backbone turns. These results elucidated the structure of the region in the protein that is recognized by pathogens and provided data on the overall fold of the protein when it binds to its natural solid substrate, hydroxyapatite. Some of these ssNMR data are summarized in Figure 3. 82 In Proteins at Interfaces III State of the Art 2012; Horbett, T., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2012.


Figure 3. Examples of ssNMR data used to constrain Rosetta simulations. and 13C-19F REDOR decay (S/S0) curves showing the carbon signal intensity as a function of the length of time the pulses are applied. a) 13C signal decay (S/S0) from recoupling of the 13C-15N dipolar couplings in a REDOR experiment. e) 13C signal decay (S/S0) from recoupling of the 13C-19F dipolar couplings in a REDOR experiment. The fits to these decay curves are used to extract the structural parameters. b) Graphic visualization of dipolar interaction parameters in the C(P33)-C(Y34)-N(Y38) spin triad. c) The contour plot of the 13C-15N



χ2(rC(P33)-N(Y38),α) function (middle) and d) graph of the χ2(rC(Y34)-N(Y38)) function showing values for which the CN REDOR data is minimized. f) Dipolar coupling parameters in the C(P33)-C(Y34)-F(P23) spin triad. Here, we denote the two carbons as C′ and C′′ since they are indistinguishable in the 13C spectrum. g) Plot of the χ2(rC′-F(P23),α) function and h) χ2(rC′′-F(P23)) function demonstrate values for which the CF REDOR data are fit by simulations.

Figure 4. Representative statherin structure from the final phase of ssNMR-constrained RosettaSurface refinement. Opacity represents statherin’s molecular shape, ribbons represent regions of helical structure, and individual amino acid side chains within the N-terminus that interact with the surface are shown as stick models. Statherin is shown adsorbed onto the {001} face of hydroxyapatite. The total number of distances between nuclei within statherin that reflect secondary and tertiary structure in the HAP-bound protein together with distances between nuclear spins in protein side chains 31P spins in the HAP surface that reflect surface proximity and orientation, is insufficient to fully constrain statherin’s surface structure. However, the body of ssNMR-derived structural measurements can be used to guide a molecular modeling computation which would provide an experimentally-constrained model of the HAP-bound structure of statherin. Professor Jeffrey Gray and coworkers at Johns Hopkins University have developed a novel algorithm (RosettaSurface.NMR) as part of the Rosetta molecular modeling package for modeling the interactions of proteins with HAP crystal surfaces (i.e.) (38). The algorithm uses experimentally derived ssNMR distance data to guide calculation of the structure and orientation of HAP-bound statherin. Using this combined computational-ssNMR approach, analysis wasmade at a sufficiently high resolution to begin understanding residueand atom-specific contributions to the process of biomineralization and hard tissue 84 In Proteins at Interfaces III State of the Art 2012; Horbett, T., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2012.


formation. Results of the study indicate that when adsorbed onto HAP crystals, statherin has a stable C-terminal helix, and a helical N-terminal HAP-binding domain with a local helix axis oriented more or less perpendicular to the local surface normal, see Figure 4. This domain interacts with the {001} HAP surface via both basic and acidic residues. Predicted by RosettaSurface (38–40) and confirmed by our ssNMR (41, 42) and calorimetric measurements (43–45), of three acidic and two phosphoserines within this binding helix, only pS3 and E5 interact directly with the HAP surface, while basic residues K6, R9, R10 and perhaps R13 interact directly with phosphate oxygen triads groups within the HAP surface (38–40). To further explore preferential face binding, RosettaSurface.NMR calculations were performed on each at five HAP crystal faces: {001}, {010}, {101}, and two differentially terminated {100} faces ({100}-T1 and {100}-T2), for a range of values of w, the parameter which weights ssNMR experimental constraints in the Rosetta Surface energy functional Econstraint. Resulting calculations, shown in Figure 5, indicate significantly greater congruency at three of five tested HAP crystal surfaces, suggesting some specificity. One of the preferred faces is {001}, the kinetically favored growth plane of HAP (46–49) and a face to which fluorescence microscopy experiments show HAP regulation proteins like statherin adhere (48). We have thus used a ssNMR/computational approach not only to determine the surface structure and orientation of statherin on HAP, but also to determine statherin’s face-binding specificity. With the advent of a combined computational-ssNMR approach, analysis can now be made at a sufficiently high resolution to begin understanding residue- and atom-specific contributions to the process of biomineralization and hard tissue formation. The results presented here indicate that statherin has a stable, folded HAp-binding domain, with the important role of surface recognition involving not simply acidic amino acid and phosphoserine side chains as commonly believed, but primarily the side chains of the basic amino acids. Also, the methods combined here show significantly greater congruency at three of five tested HAp crystal surfaces, suggesting some specificity.

Solid-State NMR Studies of Biosilicification Peptides Interest in the biomimetic approach to synthesis of silica-based materials derives from numerous technical applications as: catalysts, polymeric fillers, coatings, components in chemical and biological separations, sensors, photonic and electronic devices, bio-encapsulation, enzyme immobilization, bioimaging, drug delivery, insulators, coatings, and cosmetics (50–53). Biogenic silica (i.e. biosilica) is the most abundant biomineral with approximately six gigatons of silica produced per year by marine organisms alone. The dominant biosilica producing organism is the diatom, a microalgae characterized by an intricately decorated cell wall (i.e. the frustule) composed of organic material and silica (SiO2nH2O) (54). Diatoms take in silicon in the form of silicic acid (Si(OH)4), and catalyze its intracellular deposition as silica. As with most biogenic minerals, proteins play an important role in the regulation of silica formation in diatoms. 85 In Proteins at Interfaces III State of the Art 2012; Horbett, T., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2012.


The dominant protein fraction in the cell wall of the diatom Cylindrotheca fusiformis (see Figure 6) consists of a low molecular weight polycationic protein with high affinity for silica called silaffin (55–59). It is hypothesized that within the diatom, silaffins self assemble into a matrix which serves as a template for silica deposition. In vitro and in the absence of additional organizing components, addition of silaffins to silicic acid solutions results in the formation of silica nanospheres (55).

Figure 5. Minimization of Econstraint on five HAP surfaces as a function of weighting w. Preferential binding to various crystal faces is indicated.

The primary structure of the silaffin protein (i.e. sil1p) has been determined, and contains seven highly homologous peptide domains from residues 108 to 271 labeled R1-R7 (12, 55). These peptide domains, which are to the rich in serine and lysine, are associated with silica formation. Particular has focus been paid to the R5 sequence (SSKKSGSYSGSKGSKRRIL), as this peptide shows in vitro silica nanosphere formation activity (see Figure 7) that is similar native sil1p but occurs at neutral pH and without the need for phosphorylation of the serines and alkylation of the lysines, both commonmodifications in sil1p (60–62). Elucidating the structure of R5 within a biosilica composite is a valuable step toward understanding how small peptides direct the formation of silica structures and thus mimic the behavior of biosilicification proteins like silaffin. 86 In Proteins at Interfaces III State of the Art 2012; Horbett, T., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2012.


Figure 6. C. fusiformis cell wall (transmission electron micrographs). A) Bar = 2um; B) bar = 0.2um. Reproduced with permission from reference (56). Copyright 2004 The American Society for Biochemistry and Molecular Biology. Control of silica morphogenesis by silaffin and R5 peptides has been studied at the macroscopic level (10, 59, 62–64) but atomic level insight into the structure of R5 within biosilica composites and the nature of peptide silica interactions is almost completely lacking. Questions we seek to address regarding peptide-silica interactions in biosilica composites include: • • • • •

What is the secondary structure of R5 in the biosilica matrix and to what degree does this structure differ from that of the free peptide? What amino acid side chains and functional groups interact with silica? How are these side chains arrayed? What secondary/tertiary structural motifs are used for this purpose? Do the small R5 peptides assemble into higher order structures that act as templates for silica morphogenesis?

In contrast to studies of the biomineralization proteins statherin and amelogenin (65) which control crystal growth by adsorbing onto hydroxyapatite surfaces, silaffin and silaffin-derived peptides are embedded within the biosilica matrix rather than being simply adsorbed onto the particle surface. The biosilica microspheres precipitated from silicic acid solutions by R5 (see Figure 7) are about 25-30% peptide by weight. This abundance of peptide within the biosilica composite makes possible the use of two dimensional solid state NMR techniques to obtain 13C and 15N chemical shift measurements from isotopically enriched R5 peptides embedded in biosilica composites. As explained in the section entitled “Principles of Solid State NMR,” chemical shifts obtained via two dimensional ssNMR experiments for the 13CO, 13Cα, 13Cβ, and the amide 15N spins are sensitive to local secondary structure. Once these chemical shifts are measured, the secondary structure of the peptide within the biosilica composite can be quantified in terms of the backbone φ/ψ angles using the TALOS+ software 87 In Proteins at Interfaces III State of the Art 2012; Horbett, T., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2012.


package. In addition, chemical shifts from 13C spins distally located in protein sides chains (i.e. 13Cγ, 13Cδ, 13Cε, etc.) while not directly correlated with the structure of the protein backbone, are known to be perturbed by proximity to silica interfaces (66, 67).

Figure 7. SEM image of the R5-silica co-precipitates, which shows the peptide-catalyzed formation of biosilica microspheres with diameters of 500-750 nm.

The strategy used in the two dimensional ssNMR studies of R5 peptides in biosilica composites starts with synthetic incorporation of uniformly 13C and 15N enriched amino acids into the 19 amino acid R5 peptide. To ensure that the 13CO, 13Cα, and 13Cβ NMR signals can be unambiguously identified with specific amino acids, at most 2-3 isotopically enriched amino acids are incorporated into a R5 peptide sample at a time, making necessary the preparation of seven isotopically enrich R5 samples for study. Where chemical shifts may overlap or have ambiguity in the one dimensional 13C MAS spectrum, 13C two dimensional (i.e.2-D) dipolar assisted rotational resonance (DARR) spectra can clarify spectral assignments (68, 69). DARR is a two dimensional ssNMR method which transfers magnetization from the 1H (protons) to the 13C nuclei, which in turn transfer magnetization to other 13C nuclei which are close in space. Magnetization transfers between 13C spins are indicated in a two dimensional spectrum by “cross peaks”, peaks that connect two 13C NMR peaks on the diagonal indicating the 13C spins are thus dipole-dipole coupled and close in space. Once isotopically enriched R5 peptides are produced, they are incorporated into biosilica composites. R5 was co-precipitated with silica from silicic acid solutions as follows. The peptide was dissolved in a 100 mM phosphate-citrate buffer, pH = 7.0, at 5 mg/mL. Silicic acid was prepared using 1 mL HCl and tetramethyl orthosilicate (TMOS) and was added to the peptide/buffer solution. The R5-silica complex precipitates out in 5-10 minutes. The solution is centrifuged and precipitate collected, then dried overnight under vacuum. 88 In Proteins at Interfaces III State of the Art 2012; Horbett, T., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2012.


Figure 8. 13C-13C DARR spectrum of the R5 peptide SSKKSGSYSGSKGSKRRIL with uniformly enriched 13C and 15N amino acids incorporated at positions G10, S11, and K12. DARR spectra of the neat R5 peptide and the biosilica incorporated peptide are superimposed to emphasize chemical shift changes that occur due to conformational differences between the free peptide and the peptide in the biosilica composite.

Figure 8 shows the superimposed DARR spectra the neat and biosilica associated R5 peptide with the G10, S11, and K12 residues uniformly enriched with 13C and 15N. The neat peptide is a lyophilized sample while the second sample is precipitated from silicic acid as a silica-R5 complex. NMR peaks on the diagonal occur at the 13C isotropic chemical shift frequencies. Off diagonal peaks indicate the 13C spins are correlated via the dipolar interaction and thus are closely located in space. By following networks of dipolar-coupled 13C spins through the cross peaks in a DARR spectrum (shown as solid vertical and horizontal lines in Figure 8), the NMR peaks with known chemical shifts can be assigned to specific 13C spins within the R5 peptide. Selected assignments are shown for each spectrum. The neat sample shows the 13Cα of the lysine and glycine and the 13Cβ of the serine used to determine the exact peaks of each residue’s carbonyl (black inset.) One might note that the serine crosspeak indicated is the 13Cβ to the carbonyl, illustrating the through space contact of DARR. A 13Cα contact for the serine carbonyl and a 13Cβ to 13Cα serine contact are also present in the crosspeaks but not indicated explicitly in the figure. 89 In Proteins at Interfaces III State of the Art 2012; Horbett, T., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2012.


Besides showing how the cross-peaks in a 2-D spectrum helps elucidate silica sample is labeled to show the differentiation of the lysine sidechain. otherwise ambiguous chemical shifts, this overlay also shows there are perturbations in the chemical shifts occurring at some of the 13C and 15N spins within the silica-associated R5 peptide. Figure 9 shows the effect on the chemical shifts of the backbone 13C spins 13 ( CO, 13Cα, 13Cβ), amide 15N spins, and side chain 13C spins when the R5 peptide is co-precipitated with silica from silicic acid solutions. The histograms indicate chemical shift changes (i.e. ΔCS) obtained by subtracting the chemical shift of the 13C or 15N spin in the neat peptide from the chemical shift of the same spin in the R5-silica complex. Therefore a positive ΔCS indicates the chemical shift of a given spin has increased (relative to TMS) or been shifted downfield upon complexation with silica. A negative ΔCS indicates the opposite situation: that the 13C or 15N peak has been shifted toward TMS or upfield upon complexation with silica. Solid state NMR (70) and SFG (71) studies have shown that peptides interact with silica surfaces via positively charged side chain functional groups, including the NH3+ group of the lysine side chain or the guanidinium group of arginine. A recent 15N{29Si} REDOR study shows that monomeric amino acids with nonpolar side chains (i.e. alanine) similarly interact with silica via the NH3+ group (67). In 13C ssNMR studies of poly-lysine adsorbed onto silica the lysine side chain 13Cε spin’s chemical shift shifted by 2 ppm upfield upon adsorption onto silica, an effect attributed to proximity or the side chain of lysine to the negatively charged silica surface (70, 72, 73). Figure 9 shows that similar upfield shifts can be observed for some, but not all lysine side chains in R5 co-precipitated with silica. For example the 13Cγ, 13Cδ, and 13Cε spins of K3 are all shifted upfield by several ppm when R5 is co-precipitated with silica, while with the exception of 13Cδ, none of the side chain 13C chemical shifts in K4 are affected. Neither do the chemical shifts for the side chain 13C spins of K12 or K15 show significant changes upon co-precipitation with silica. These data indicate that while K3 and to a lesser extent K4 interact with silica, the absence of observable chemical shift changes indicate the side chains of K12 and K15 do not appear to be in proximity to silica. In Figure 9 it is also clear that the chemical shifts of many of the side chain 13C spins in the C-terminal RRIL moiety are perturbed upon co-precipitation with silica, indicating that there is an environmental change for this moiety as well. The ΔCS data in Figure 9 for R16, R17, and I18 (L19 was tethered to the resin and not labeled) show significant chemical shift changes in many of the side chain 13C spins, although the ΔCS trends indicate more complicated behavior than would be indicated by simple proximity to silica. In R16 the chemical shifts of 13Cγ, 13Cδ, and 13Cζ are moved upfield upon co-precipitation by silica, while R17 and I18 show downfield changes in side chain 13C chemical shifts, the exception being 13Cζ of R17 which shifts upfield upon co-precipitation with silica. The chemical shift changes observed for selected lysine side chains can be understood as resulting from varying degrees of interaction between positively charged amino acid side chains of aggregated peptides and the silica phase. Studies of silica precipitation by R5 mutants indicate the C-terminal RRIL moiety is necessary for silica formation activity (60, 61). In the presence of RRIL, R5 90 In Proteins at Interfaces III State of the Art 2012; Horbett, T., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2012.


peptides are observed by light scattering to form aggregates about 825 nm in diameter. Formation of peptide aggregates is considered necessary for silica formation with R5 because when RRIL is omitted silica formation is reduced and no peptides aggregates are formed. The RRIL motif is believed to be involved in the self-assembly of the peptide where the pattern of arginine’s positively charged guanidinium groups in proximity to the hydrophobic leucine and isoleucine residues results in a micelle like assembly.

Figure 9. Perturbations of the chemical shifts of the 13CO, 13Cα, 13Cβ and 15N amide nitrogen spins (left) and the sidechain 13C spins (right) in R5 co-precipitated with silica referenced to the native peptide.Vertical scale is graduated in parts per million. Residues 7 and 8 were not labeled due to synthesis errors.

The pattern of chemical shift changes observed in R5 and portrayed in Figure 9, may therefore the result from the formation of micellar-like peptide aggregates where positively-charged amino acid side chains near the peptide N-terminus are exposed at the surface of the aggregate and interact with the surrounding silica matrix. This being the case the side chain of K3 would clearly be most affected, K4 shows less significant perturbations, while K12 and K15 are buried within the aggregate and are removed from the silica matrix. The RRIL chemical shifts are likely due to the effect of peptide-peptide interactions which are the basis for aggregate formation. Figure 9 also shows that numerous chemical shift changes occur for the 15NH, 13CO, 13Cα, and 13Cβ spins throughout the R5 peptide upon formation of biosilica composites. Interpretation of these chemical shift changes is potentially complicated by the fact that the backbone 13C chemical shifts are sensitive to 91 In Proteins at Interfaces III State of the Art 2012; Horbett, T., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2012.


conformational changes. However, given that biosilica composite formation is only observed in the presence of peptide aggregates, and given that the orientation of the peptides in these aggregates likely leaves amino acid side chains beyond K3 largely removed from the silica matrix, we assume the backbone 13C chemical shift changes observed beyond K3 are largely due to conformational changes. A preliminary evaluation of the carbon-only chemical shifts of CO, Cα and Cβ by TALOS+ shows that while most of the expected angles fall into β-sheet region (φ < 0°, ψ > 90°) for both the neat peptide and the peptide-silica composite, residues K4 goes from αL-helical (30°< φ

Solid-State NMR Studies of Biomineralization Peptides and Proteins

Recommend Documents