Renee D. JiJi

Simultaneous observation of peptide backbone lipid solvation and α-helical structure by deep-UV resonance Raman spectroscopy.

Despite a variety of methodologies aimed at improving membrane protein structure analysis, information about these proteins in their native membrane environments remains scarce. Currently, no structurally sensitive spectroscopic techniques are capable of co-determining ensemble structural content and localized lipid versus aqueous solvation information. Here, we describe the first deep-UV (lex<210 nm) resonance Raman (dUVRR) spectra of a model a-helical peptide embedded in a membrane-mimetic environment, confirming sensitivity to secondary structure content and revealing sensitivity of dUVRR to the lipid solvation of the peptide backbone. Analyses of membrane protein structural dynamics are hampered by the experimental difficulties associated with elucidating structural changes and correlating those changes to their respective solvation by the nonpolar lipid or surfactant versus the aqueous phases. No kinetically amenable spectroscopic techniques are capable of delineating subtle changes in protein structure while simultaneously reporting on that structure’s solvation without protein modification by deuterium exchange, isotope labeling, mutagenesis or post-translational spin/fluorophore labeling. Glimpses of the dynamics and stabilizing forces involved with protein folding and insertion into membranes have recently been gleaned by UV excited resonance Raman spectroscopy focused on excitation wavelengths specific for aromatic residues (lex>220 nm). Deep-UV (lex< 210 nm) excitation, which has been a valuable tool for analyzing the structure of soluble proteins by accessing the p!p* transition of the peptide backbone vibrational modes and their dynamics, has not been previously explored successfully for this class of hydrophobic proteins. The dUVRR protein spectral response consists of four peptide backbone related amide (Am) responses—I (C=O stretching), II (in phase C-H/N-H stretching/bending), III (out of phase C-H/N-H stretching/bending) and S (coupled C-H/N-H bending; alternately referred to as CaHb). [2a] The combinations of Am mode positions and intensities are strongly correlated to the constraints imparted by particular secondary structures with soluble proteins. Solvent interaction and its extent with the peptide backbone can also influence the Am mode spectral positions in dUVRR and IR and intensities in dUVRR alone. Theoretical calculations with Nmethylacetamide (NMA) in different solvent polarities have revealed that the solvent-dependent Am I intensity differences seen in the dUVRR spectra, but not the IR spectra are derived from the sensitivity of the former technique to the polarizability term of the C=O bond. Herein, we present evidence that a surfactant-solubilized protein region also has altered Am mode intensities, especially in the C=O stretching region. As a model for the common a-helical membrane-embedded protein domain, we have examined the de novo designed ME1 peptide, which contains a single hydrophobic a-helical segment encompassing roughly 75 % of the total peptide backbone. Like its parent protein, it is extremely insoluble in aqueous solvents and only forms stable a-helical homodimers within a micellar environment. The dUVRR spectrum using an excitation source of 197 nm of a dodecyl phosphocholine (DPC)-solubilized ME1 sample contains aromatic side chain-derived modes (1180–1210 and 1580–1620 cm ) arising from the single tyrosine and phenylalanine residues within the peptide sequence (Figure 1). Peptide backbone contributions can also be assigned for the Am I (1658 cm ), II (1546 cm ) and III (1260–1340 cm ) modes and a smaller feature where the Am S (1400 cm ) mode would be expected. The Am III mode’s position, coupled to the limited extent of the Am S contribu-

Resolution of localized small molecule-Aβ interactions by deep-ultraviolet resonance Raman spectroscopy.

The mechanism by which flavonoids prevent formation of amyloid-β (Aβ) fibrils, as well as how they associate with non-fibrillar Aβ is still unclear. Fresh, un-oxidized myricetin exhibited excitation and emission fluorescence maxima at 481 and 531 nm, respectively. Introduction of either Aβ(1-42) or Aβ(25-40) resulted in a fluorescence decrease, when measured at 481 nm, suggesting formation of a myricetin-Aβ complex. Circular dichroism (CD) and ultraviolet resonance Raman (UVRR) studies indicate that the association of myricetin with the Aβ peptide or its hydrophobic fragment, Aβ(25-40), leads to subtle changes in each peptides conformation. Aβ(25-40) formed amyloid fibrils at a similar rate, when compared to the full-length peptide, Aβ(1-42), using thioflavin T (ThT) fluorescence. Studies also indicated that myricetin was equally effective at preventing the formation of both Aβ(1-42) and Aβ(25-40) fibrils. Although ThT assays indicated that Aβ(1-16) did not form amyloid fibrils, CD studies of the hydrophilic fragment, Aβ(1-16), suggest possible interactions between myricetin and aromatic side chains. UVRR studies of the full-length peptide and Aβ(1-16) showed increases in the intensity of the aromatic modes upon introduction of myricetin. Our findings suggest that myricetin interacts with soluble Aβ via two mechanisms, association with the hydrophobic C-terminal region and interactions with the aromatic side chains.

Deep-UV resonance Raman analysis of the Rhodobacter capsulatus cytochrome bc₁complex reveals a potential marker for the transmembrane peptide backbone.

Classical strategies for structure analysis of proteins interacting with a lipid phase typically correlate ensemble secondary structure content measurements with changes in the spectroscopic responses of localized aromatic residues or reporter molecules to map regional solvent environments. Deep-UV resonance Raman (DUVRR) spectroscopy probes the vibrational modes of the peptide backbone itself, is very sensitive to the ensemble secondary structures of a protein, and has been shown to be sensitive to the extent of solvent interaction with the peptide backbone [ Wang , Y. , Purrello , R. , Georgiou , S. , and Spiro , T. G. ( 1991 ) J. Am. Chem. Soc. 113 , 6368 - 6377 ]. Here we show that a large detergent solubilized membrane protein, the Rhodobacter capsulatus cytochrome bc(1) complex, has a distinct DUVRR spectrum versus that of an aqueous soluble protein with similar overall secondary structure content. Cross-section calculations of the amide vibrational modes indicate that the peptide backbone carbonyl stretching modes differ dramatically between these two proteins. Deuterium exchange experiments probing solvent accessibility confirm that the contribution of the backbone vibrational mode differences are derived from the lipid solubilized or transmembrane α-helical portion of the protein complex. These findings indicate that DUVRR is sensitive to both the hydration status of a proteins peptide backbone, regardless of primary sequence, and its secondary structure content. Therefore, DUVRR may be capable of simultaneously measuring protein dynamics and relative water/lipid solvation of the protein.

Quantification of protein secondary structure content by multivariate analysis of deep-ultraviolet resonance Raman and circular dichroism spectroscopies

Determination of protein secondary structure (α-helical, β-sheet, and disordered motifs) has become an area of great importance in biochemistry and biophysics as protein secondary structure is directly related to protein function and protein related diseases. While NMR and X-ray crystallography can predict the placement of each atom in a protein to within an angstrom, optical methods (i.e. CD, Raman, and IR) are the preferred techniques for rapid evaluation of protein secondary structure content. Such techniques require calibration data to predict unknown protein secondary structure content where accuracy may be improved with the application of multivariate analysis. Here, a comparison of the protein secondary structure predictions obtained from multivariate analysis of ultraviolet resonance Raman (UVRR) and circular dichroism (CD) spectroscopic data using classical least squares (CLS), partial least squares (PLS), and multivariate curve resolution-alternating least squares (MCR-ALS) is made. Results of the multivariate analysis suggest that CD measurements provide more accurate prediction of protein α-helical content whereas UVRR more accurately predicts β-sheet content, an observation that is consistent with previous studies. Based on this analysis, it is suggested that the best approach to rapid and accurate protein secondary structure determination is to combine both CD and UVRR spectroscopic data.

Developing microwave-assisted ionic liquid microextraction for the detection and tracking of hydrophobic pesticides in complex environmental matrices

In the present study, we have developed a novel dispersive liquid–liquid microextraction (DLLME) based on microwave-assisted DLLME (MADLLME) using ionic liquids for the separation of environmentally-relevant pyrethroid pesticides from various aqueous milieux. High-performance liquid chromatography (HPLC) was employed for the detection and quantitative tracking of the pesticides. Six different ILs were preliminarily tested as extraction solvents against four representative model pyrethroids. The optimization of the current method was derived by consideration of the dispersal solvent, ionic liquid choice, extraction container material, aqueous-phase pH, and microwave conditions (particularly, the applied power and irradiation time). Optimal results were achieved using methanol as a dispersal solvent with trioctylmethylammonium bis(trifluoromethylsulfonyl)imide ([N8881][Tf2N]) as the extraction solvent at a microwave power of 200 W for 60 s. A number of spiked food samples (e.g., honey, milk, assorted fruits) were also tested using MADLLME, with excellent recoveries achieved from these complex matrices as compared to DLLME alone.

Bilayer surface association of the pHLIP peptide promotes extensive backbone desolvation and helically-constrained structures.

Despite their presence in many aspects of biology, the study of membrane proteins lags behind that of their soluble counterparts. Improving structural analysis of membrane proteins is essential. Deep-UV resonance Raman (DUVRR) spectroscopy is an emerging technique in this area and has demonstrated sensitivity to subtle structural transitions and changes in protein environment. The pH low insertion peptide (pHLIP) has three distinct structural states: disordered in an aqueous environment, partially folded and associated with a lipid membrane, and inserted into a lipid bilayer as a transmembrane helix. While the soluble and membrane-inserted forms are well characterized, the partially folded membrane-associated state has not yet been clearly described. The amide I mode, known to be sensitive to protein environment, is the same in spectra of membrane-associated and membrane-inserted pHLIP, indicating comparable levels of backbone dehydration. The amide S mode, sensitive to helical structure, indicates less helical character in the membrane-associated form compared to the membrane-inserted state, consistent with previous findings. However, the structurally sensitive amide III region is very similar in both membrane-associated and membrane-inserted pHLIP, suggesting that the membrane-associated form has a large amount of ordered structure. Where before the membrane-associated state was thought to contain mostly unordered structure and reside in a predominantly aqueous environment, we have shown that it contains a significant amount of ordered structure and rests deeper within the lipid membrane.

Spectroscopic detection of β -sheet structure in nascent Aβ oligomers.

Deep-UV resonance Raman (UVRR) spectroscopy and circular dichroism (CD) were employed to study the secondary structure of Aβ(1-42) in fresh samples with increasing fractions of oligomeric peptide. A feature with a minimum at ~217 nm appeared in CD spectra of samples containing oligomeric Aβ(1-42). UVRR spectra more closely resembled those of disordered proteins. The primary difference between UVRR spectra was the ratio of the 1236 cm(-1) to 1260 cm(-1) amide III peak intensities, which shifted in favor of the 1236 cm(-1) band as the fraction of oligomeric peptide increased.

On the freezing behavior and diffusion of water in proximity to single-supported zwitterionic and anionic bilayer lipid membranes

We compare the freezing/melting behavior of water hydrating single-supported bilayers of a zwitterionic lipid DMPC with that of an anionic lipid DMPG. For both membranes, the temperature dependence of the elastically scattered neutron intensity indicates distinct water types undergoing translational diffusion: bulk-like water probably located above the membrane and two types of confined water closer to the lipid head groups. The membranes differ in the greater width of the water freezing transition near the anionic DMPG bilayer compared to zwitterionic DMPC as well as in the abruptness of the freezing/melting transitions of the bulk-like water.

Role of Bilayer Characteristics on the Structural Fate of Aβ(1–40) and Aβ(25–40)

The β-amyloid (Aβ) peptide is derived from the transmembrane (TM) helix of the amyloid precursor protein (APP) and has been shown to interact with membrane surfaces. To understand better the role of peptide-membrane interactions in cell death and ultimately in Alzheimers disease, a better understanding of how membrane characteristics affect the binding, solvation, and secondary structure of Aβ is needed. Employing a combination of circular dichroism and deep-UV resonance Raman spectroscopies, Aβ(25-40) was found to fold spontaneously upon association with anionic lipid bilayers. The hydrophobic portion of the disease-related Aβ(1-40) peptide, Aβ(25-40), has often been used as a model for how its legacy TM region may behave structurally in aqueous solvents and during membrane encounters. The structure of the membrane-associated Aβ(25-40) peptide was found to depend on both the hydrophobic thickness of the bilayer and the duration of incubation. Similarly, the disease-related Aβ(1-40) peptide also spontaneously associates with anionic liposomes, where it initially adopts mixtures of disordered and helical structures. The partially disordered helical structures then convert to β-sheet structures over longer time frames. β-Sheet structure is formed prior to helical unwinding, implying a model in which β-sheet structure, formed initially from disordered regions, prompts the unwinding and destabilization of membrane-stabilized helical structure. A model is proposed to describe the mechanism of escape of Aβ(1-40) from the membrane surfaces following its formation by cleavage of APP within the membrane.

Insights into the aggregation mechanism of Aβ(25-40).

The hydrophobic fragment of the Alzheimers related β-amyloid (Aβ) peptide, Aβ(25-40), aggregates and forms insoluble amyloid fibrils at a rate similar to the full-length peptide. In order to gain insight into the fibrillization of Aβ(25-40) and the ability of the flavonoid myricetin to inhibit its aggregation, the isoleucine at position 32 (I32A) and the glycine at position 37 (G37A) in the full-length peptide were replaced with alanine. Thioflavin T assays indicate that substitution of isoleucine for alanine significantly reduces the rate and extent of fibrillization compared to the Aβ(25-40) and G37A peptides. Although all three peptides are fully disordered initially, circular dichroism studies suggest the structure of the I32A and G37A peptides are different from the parent peptide Aβ(25-40). Introduction of myricetin to the peptide samples results in modest structural changes for the Aβ(25-40) and G37A peptides but not the I32A peptide. Aβ(25-40) oligomers were predominantly tetramers, whereas I32A and G37A oligomers were a mixture of trimers and dimers. After 48h of incubation at 37°C, the amount of tetramers and trimers in solution dropped for the Aβ(25-40) and G37A peptides but remained similar for the I32A peptide. Incubation of Aβ(25-40) with myricetin increased the relative proportion of trimers to tetramers. Ultraviolet resonance Raman studies suggests that the I32A peptide may be more hydrated than the Aβ(25-40) and G37A peptides. Taken together, these data indicate the structural changes observed for the Aβ(25-40) and G37A peptides upon introduction of myricetin are localized around residue 32 and could arise from hydrophobic interactions between the peptide and the flavonoid or interference with the self-association of the peptide in this region. Substitution of isoleucine at position 32 with alanine had little effect on the peptides secondary structure but dramatically decreased the propensity of the peptide fibrillize.