Article Information

• PDF (187 kb)
• Full Text with Large Figures
• Supporting Material
• Cited by in Scopus (3)

PubMed

• Articles by Shadi Abu-Baker
• Articles by Gary A. Lorigan

Related Articles

• Polylysine-Induced H NMR-Observable Domains in Phosphatidyls...

  Polylysine-Induced H NMR-Observable Domains in Phosphatidylserine/Phosphatidylcholine Lipid Bilayers Biophysical Journal, Volume 81, Issue 6, 1 December 2001, Pages 3346-3362 Carla M. Franzin and Peter M. Macdonald Abstract The interaction of three polylysines, Lys (=5), Lys (=30), and Lys (=100), where is the number of lysine residues per chain, with phosphatidylserine-containing lipid bilayer membranes was investigated using H NMR spectroscopy. Lys and Lys added to multilamellar vesicles composed of (70:30) (mol:mol) mixtures of choline-deuterated 1-palmitoyl-2-oleoyl--glycero-3-phosphocholine (POPC)+1-palmitoyl-2-oleoyl--glycero-3-phosphoserine (POPS) produced two resolvable H NMR spectral components under conditions of low ionic strength and for cases where the global anionic lipid charge was in excess over the global cationic polypeptide charge. The intensities and quadrupolar splittings of the two spectral components were consistent with the existence of polylysine-bound domains enriched in POPS, in coexistence with polylysine-free domains depleted in POPS. Lys, however, yielded no H NMR resolvable domains. Increasing ionic strength caused domains to become diffuse and eventually dissipate entirely. At physiological salt concentrations, only Lys yielded H NMR-resolvable domains. Therefore, under physiological conditions of ionic strength, pH, and anionic lipid bilayer content, and in the absence of other, e.g., hydrophobic, contributions to the binding free energy, the minimum number of lysine residues sufficient to produce spectroscopically resolvable POPS-enriched domains on the H NMR millisecond timescale may be fewer than 100, but is certainly greater than 30. Abstract | Full Text | PDF (216 kb)

• Investigating Structural Changes in the Lipid Bilayer upon I...

  Investigating Structural Changes in the Lipid Bilayer upon Insertion of the Transmembrane Domain of the Membrane-Bound Protein Phospholamban Utilizing P and H Solid-State NMR Spectroscopy Biophysical Journal, Volume 86, Issue 3, 1 March 2004, Pages 1564-1573 Paresh C. Dave, Elvis K. Tiburu, Krishnan Damodaran and Gary A. Lorigan Abstract Phospholamban (PLB) is a 52-amino acid integral membrane protein that regulates the flow of Ca ions in cardiac muscle cells. In the present study, the transmembrane domain of PLB (24–52) was incorporated into phospholipid bilayers prepared from 1-palmitoyl-2-oleoyl--glycero-phosphocholine (POPC). Solid-state P and H NMR experiments were carried out to study the behavior of POPC bilayers in the presence of the hydrophobic peptide PLB at temperatures ranging from 30°C to 60°C. The PLB peptide concentration varied from 0mol % to 6mol % with respect to POPC. Solid-state P NMR spectroscopy is a valuable technique to study the different phases formed by phospholipid membranes. P NMR results suggest that the transmembrane protein phospholamban is incorporated successfully into the bilayer and the effects are observed in the lipid lamellar phase. Simulations of the P NMR spectra were carried out to reveal the formation of different vesicle sizes upon PLB insertion. The bilayer vesicles fragmented into smaller sizes by increasing the concentration of PLB with respect to POPC. Finally, molecular order parameters () were calculated by performing H solid-state NMR studies on deuterated POPC (-1 chain) phospholipid bilayers when the PLB peptide was inserted into the membrane. Abstract | Full Text | PDF (173 kb)

• Interaction of Alamethicin with Ether-Linked Phospholipid Bi...

  Interaction of Alamethicin with Ether-Linked Phospholipid Bilayers: Oriented Circular Dichroism, P Solid-State NMR, and Differential Scanning Calorimetry Studies Biophysical Journal, Volume 89, Issue 4, 1 October 2005, Pages 2434-2442 Paresh C. Dave, Emma Billington, Yeang-Ling Pan and Suzana K. Straus Abstract The arrangement of the antimicrobial peptide alamethicin was studied by oriented circular dichroism, P solid-state NMR, and differential scanning calorimetry in ether-linked phospholipid bilayers composed of 1,2-O-dihexadecyl--glycero-3-phosphocholine (DHPC). The measurements were performed as a function of alamethicin concentration relative to the lipid concentration, and results were compared to those reported in the literature for ester-linked phospholipid bilayers. At ambient temperature, alamethicin incorporates into the hydrophobic core of DHPC bilayers but results in more lipid disorder than observed for ester-linked 1-palmitoyl, 2-oleoyl--glycero-3-phosphatidylcholine (POPC) lipid bilayers. This orientational disorder appears to depend on lipid properties such as bilayer thickness. Moreover, the results suggest that alamethicin inserts into the hydrophobic core of the bilayers (at high peptide concentration) for both ether- and ester-linked lipids but using a different mechanism, namely toroidal for DHPC and barrel-stave for POPC. Abstract | Full Text | PDF (217 kb)

• …more

Copyright © 2007 The Biophysical Society. All rights reserved.
Biophysical Journal, Volume 93, Issue 10, 3480-3490, 15 November 2007

doi:10.1529/biophysj.107.107789

Membranes

Investigating the Interaction of Saposin C with POPS and POPC Phospholipids: A Solid-State NMR Spectroscopic Study

Shadi Abu-Baker^*, Xiaoyang Qi^† and Gary A. Lorigan^*, ,  

^* Department of Chemistry and Biochemistry, Miami University, Oxford, Ohio
^† Division and Program of Human Genetics, Cincinnati Children's Hospital Medical Center and University of Cincinnati College of Medicine, Cincinnati, Ohio

Address reprint requests to Gary A. Lorigan, Tel.: 513-529-3338; Fax: 513-529-5715.

Sapocins (a four-member family, Saposin A, B, C, and D) are derived from the single precursor protein, prosaposin 1,2. These small (∼80 amino acids) heat stable glycoproteins are necessary for the hydrolytic activity of several lysosomal enzymes involved in the catalytic pathway of glycosphingolipids 1. The focus of this study is Saposin C (Sap C); a protein that can stimulate the catalytic activity of glucosylceramidase ((GCase) EC 3.1.2.45, a lysosomal enzyme) 3,4. GCase is essential for the hydrolysis of glucosylceramide to ceramide and glucose 3,4. Additionally, Sap C participates in the fusion and destabilization of negatively charged phospholipids 5,6. Membrane fusion is a key step in the processes of secretion, endocytosis, exocytosis, fertilization, and intracellular transport 7. Also, the destabilization ability of Sap C to the membrane surface is thought to promote the insertion of the lysosomal enzymes into the bilayer 5,8. Moreover, Sap C plays a significant role in lipids trafficking and transfer 2,9. Deficiency in either Sap C or GCase leads to different variant forms of Gaucher's disease 10,11. In Gaucher's disease as well as in other saposins related pathologies including Tay-Sachs, Sandhoff, and type I Niemann-Pick diseases, saposins were reported to accumulate in the tissues of lysosomal storage disease patients 1. In a previous review article, it was hypothesized that accumulation of saposins could be related to a stimulation in their synthesis by a compensatory mechanism that results from accumulation of either lipids or defective enzyme 1. Alternatively, it was suggested that codeposition of saposins with the accumulated lipids may explain the saposin accumulation 1.

The crystal structures of Sap C and D have recently been reported 12. Also, two recent reports have determined the three-dimensional structure of Sap C in solution and in a detergent environment via solution NMR spectroscopy 13,14. The amino acid sequence of Sap C is shown in Figure 1A. As shown in Figure 1B, Sap C was found to have five α-helices 13,14. Additionally, a recent solid-state NMR article has indicated that Sap C interacts with negatively charged phospholipid membranes under acidic conditions 15. The solid-state NMR data agrees with the structural model proposed by Qi and co-workers utilizing fluorescence quenching experiments 16. This model suggests that only the amphipathic amino terminus of helix 1 and the carboxyl terminus of helix 5 of Sap C are inserted into the membrane leaving the middle region of Sap C exposed to the aqueous phase 16.

Display large version of this figure

Figure 1

The amino acid sequence (A) and the previously reported three-dimensional solution NMR structure of Sap C (B).

Acidic conditions help to initialize the electrostatic binding between the Sap C protein and negatively charged phospholipids such as phosphatidylserine (PS) 5,13. The protonation of the negative charges of Glu residues of Sap C at acidic conditions eliminates the repulsive forces between Sap C and the negatively charged headgroups of POPS phospholipids 13. Additionally, Sap C has many positively charged Lys residues that are predominantly located at the amino terminal portion of the protein 16. The positively charged Lys residues are important in the recognition of the negative surface charges of phosphatidylserine-containing membranes for initial binding of Sap C 16. Finally, after Sap C recognizes the negatively charged membrane surface, only the terminus regions of helix 1 and helix 5 of Sap C insert into the negatively charged membrane 16.

The proton pumps in the lysosomes are key to making the interior of these intracellular digestive organelles acidic 17. Many lysosomal enzymes have acidic pH optimums for activity 18,19. Previously, we demonstrated with solid-state NMR spectroscopy that Sap C interacts with a mixture of negatively charged PS/PG phospholipids under acidic conditions 15. However, some lysosomal enzymes have neutral pH optimums 20,21. To make such enzymes active, the pH of the lysosome fluctuates from 4.5 to 7.0 or more 22,23. This variation helps the hydrolytic enzymes to efficiently degrade macromolecules 22. Additionally, some physiological disorders related to lipid accumulation have been connected with pH elevation in the lysosomes 24,25. To extend and complement our previous experiments, we designed new solid-state NMR experiments to probe the interaction of Sap C (initially protonated in a weak acid) with both the neutral 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) and negatively charged phospholipids such as 1-palmitoyl-2-oleoyl-sn-glycero-3- [phospho-L-serine] (POPS). Both randomly dispersed multilamellar vesicles (MLVs), as well as mechanically oriented bilayers were resuspended in a neutral buffer to mimic the less common neutral lysosomal pH conditions. For the first time, solid-state NMR data of both Sap C-POPS and Sap C-POPC MLVs, prepared under similar neutral pH conditions, were compared from the perspective of the lipids using ²H order parameter profiles, ³¹P chemical shift anisotropy (CSA) width and ³¹P T₁ relaxation times.

All synthetic phospholipids such as 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC); 1-palmitoyl-2-oleoyl-sn-glycero-3-[phospho-L-serine] (POPS) were purchased from Avati Polar Lipids (Alabaster, AL). The phospholipids were dissolved in chloroform and stored at −20°C; 2,2,2 triflouroethanol (TFE), N-[2-hydroxyethyl]piperazine-N′-2-ethane sulfonic acid (HEPES) and EDTA were obtained from Sigma-Aldrich (St. Louis, MO).

The biologically active nonglycosylated recombinant form of Sap C 26,27 has been chosen in this study, since it has a high protein production yield. Preparation and purification of Sap C was performed according to procedures described previously in the literature 28. In brief, recombinant Sap C was overexpressed in Escherichia coli cells using an isopropyl-1-thio-β-D-galactopyranoside-induction pET system. Next, after Sap C was dialyzed and lyophilized, the dried protein was dissolved in 0.1% trifluoroacetic acid and further purified via HPLC C4 reverse-phase column. Then Sap C (with >98% purity) was lyophilized, the correct molecular weight was determined using MALDI-TOF, Western blot (data not shown), and Sap C concentrations were determined as previously reported 29. No lipids and glucosylceramidase were found in the Sap C preparation (GCase EC 3.1.2.45).

Two different MLV samples of POPC and POPS containing 0mol % and 2mol % of Sap C were prepared using a slightly modified protocol 30. The phospholipids (dissolved in chloroform) were mixed with Sap C, dissolved in a minimal amount of triflouroethanol ((TFE) ∼50μl), in a small test tube (total of 40mg lipids) and dried using a steady steam of N₂ gas. After placing the test tube in a vacuum desiccator overnight, the peptide/lipid sample was resuspended in 85μL of HEPES buffer (pH=7.0). The buffer was prepared using deuterium-depleted water from Isotec/Sigma-Aldrich (St. Louis, MO). Then, the test tube was shortly heated in a water bath at 50°C along with slight frequent sample agitation to avoid frothing of the suspension. After the phospholipid mixture was completely dissolved, the sample was transferred to a NMR sample tube. POPC-d₃₁ or POPS-d₃₁ (5mg) was added to the samples when performing the ²H NMR experiments.

The mechanically oriented bilayers were prepared using three different procedures, in which TFE, HEPES neutral buffer or methanol were used to dissolve Sap C. Weak acids such as TFE can protonate proteins 31. In this report, the term protonated Sap C will be used to describe Sap C that was dissolved in the TFE (weak acid). Conversely, the term unprotonated Sap C will be used to describe Sap C that was dissolved in methanol or directly in a neutral buffer.

In the first procedure, to insure that Sap C is incorporated within the bilayers, Sap C was dissolved in 50μl of TFE, mixed with POPS or POPC phospholipids (15mg), dissolved in chloroform, and the final volume was reduced to approximately one-half. Then the lipid sample was added via a syringe onto 15 glass plates (1mg/glass plate) and allowed to dry. The final Sap C concentration on each glass plate was 2mol % with respect to the lipids. Next, the glass plates were left in a desiccator overnight. The following day, 2μl of a neutral HEPES buffer (pH 7.0) was added to each glass plate and after 4–5min the glass plates were stacked and placed in a humidity chamber of saturated ammonium monophosphate at a relative humidity of 93% at 45°C for 2 days. Finally, the glass plates were wrapped in parafilm and placed inside polyethylene bags to minimize dehydration.

For the second procedure, the phospholipids dissolved in chloroform (15mg) were spotted onto 15 glass plates, air-dried, and left in the desiccator overnight. The following day, 2μl of the HEPES buffer (pH 7.0) was added. In this procedure, Sap C was completely dissolved in the buffer; therefore, Sap C would interact with the oriented bilayers only after the addition of buffer. Finally, for the third procedure, Sap C was dissolved in a minimal amount of methanol (∼50μl) and mixed with POPS phospholipids (15mg), dissolved in chloroform, and then the glass plates samples were prepared similar to the first procedure described above.

A Bruker Avance 500MHz WB solid-state NMR spectrometer was used to run all of the experiments. For the mechanically oriented bilayers samples, a static double-resonance flat coil probe from Doty Scientific (Columbia, SC) was used. The flat coil in the probe can be varied to adjust the angle at which the bilayer normal makes with respect to the magnetic field. The ³¹P NMR spectra were recorded with ¹H decoupling using a 5-μs π/2 for ³¹P and a 4-s recycle delay. For the ³¹P NMR spectra 512 scans were taken and the free induction delay was processed using 100Hz of line broadening. The spectral width was set to 250ppm. In addition, a Bruker 4mm triple resonance CP-MAS probe (Bruker, Billerica, MA) was used to record the ³¹P and ²H NMR spectra of MLVs. The ³¹P NMR spectra were recorded with ¹H decoupling using a 4-μs π/2 for ³¹P and a 4-s recycle delay. A 55-kHz proton decoupling field was uniformly applied for all static ³¹P NMR samples. For the ³¹P NMR spectra 1-k scans were taken and the free induction delay was processed using 300Hz of line broadening. The spectral width was set to 300ppm. All ³¹P NMR spectra were referenced by assigning the 85% H₃PO₄³¹P peak to 0ppm.

²H NMR spectra were collected at a frequency of 76.77MHz. A quadrupolar echo pulse sequence was employed using quadrature detection with complete phase cycling of the pulse pairs 32. The 90° pulse length was 3.8μs, the interpulse delay was 40μs, the recycle delay was 0.5s, and the spectral width was set to 100kHz. A total of 20,480 transients were averaged for each spectrum and processed using 300-Hz line broadening. ³¹P T_1Z longitudinal relaxation experiments were conducted using an inversion-recovery pulse sequence 180°-T-90°-acq with the sample spinning (4kHz) at the magic angle. After the 180° pulse, the delay time (T) was varied from 10.0ms to 6.0s. The recycle delay was set to 10s. To measure T₁, the change in the area under a particular peak was fitted to a single exponential function: I(t)=I(0)−Aexp(−t/T₁), where I(0) is close to 1 and A is ∼2 33.

POPC-d₃₁ and POPS-d₃₁ were used to probe the changes in the order of the acyl chains of POPC and POPS MLVs, respectively, upon Sap C insertion when compared to the control (without Sap C). Powder-pattern ²H NMR spectra of the multilamellar dispersions were numerically deconvoluted (dePaked) using the algorithm of McCable and Wassall 15,34. The spectra were deconvoluted such that the bilayer normal was perpendicular with respect to the direction of the static magnetic field. S_CD order parameters were calculated from the expression previously described in the literature 15,30. The ²H peaks in the NMR spectra were assigned based upon dynamic properties of the individual CD₃ and CD₂ groups. The quadrupolar splitting of the CD₃ methyl groups at the end of the acyl chains are the smallest and closest to 0kHz and the ²H attached to the C-14 have the next smallest splitting followed by the ²H attached to the C-13, and so forth along the acyl chain. According to the literature, the quadrupolar splittings of the plateau region can be estimated by integration of the last broad peak in the ²H NMR spectra and the order parameters calculated for the CD₃ quadrupolar splitting were multiplied by three 15,34. Finally, the simulations of the static ³¹P NMR spectra were carried out using the DMFIT software program 35. The principle elements of the CSA tensors are represented according to the convention σ₃₃≥σ₂₂≥σ₁₁.

Previously, Sap C was studied using different biophysical techniques that includes Trp fluorescence and quenching experiments 16, fluorescent dye leakage 8, direct atomic force microscopy studies 36, solution NMR in a detergent environment 13, and fluorescence resonance energy transfer 6. In many of these studies, lipid vesicles or large unilamellar vesicles LUVs were used to study the effect of Sap C on the fusion, leakage, or distribution of these vesicles. Conversely, in this study, only the interaction of Sap C with POPC and POPS lipids as well as its perturbation effect on the lipid acyl chains and headgroups were investigated; therefore, MLVs (commonly used in solid-state NMR spectroscopic experiments) and not LUVs were chosen as a model membrane system.

The effect of Sap C on the order and dynamics of the acyl chains of both POPC-d₃₁ and POPS-d₃₁ bilayers was studied using ²H solid-state NMR spectroscopy in the absence (Figure 2AC) and in the presence (Figure 2BD) of 2mol % Sap C at 25°C. The addition of 2mol % Sap C to POPC does not significantly alter the ²H NMR spectra (Figure 2B) when compared to the control (Figure 2A). Inspection of (Figure 2D) clearly indicates that the addition of 2mol % Sap C to the POPS-d₃₁ bilayers alters the lineshape and the spectral resolution of the ²H NMR spectra. The loss in spectral resolution is manifested by the disappearance of sharp edges of the peaks of 2mol % Sap C in POPS bilayers. The changes in spectral resolution of the ²H NMR spectra confirm that Sap C interacts significantly with the POPS-d₃₁-containing MLVs and not with the POPC-d₃₁-containing MLVs. The central resonance doublet corresponds to the terminal CD₃ groups and the remaining overlapped doublets result from the different CD₂ segments of the acyl chain of the POPC-d₃₁ and POPS-d₃₁.

Display large version of this figure

Figure 2

²H NMR powder-pattern spectra of the both POPC and POPS bilayers in the absence (A and C) and in the presence (B and D) of 2mol % Sap C with respect to the lipids at 25°C.

In this study, POPC-d₃₁ and POPS-d₃₁ segmental S_CD order parameters describe the local orientation or dynamic perturbations of the C-D bond vector from its standard state due to perturbations of the POPC-d₃₁ and POPS-d₃₁ phospholipid conformations or dynamics as a result of the addition of Sap C to the lipid bilayers. The smoothed segmental C-D bond order parameters (S_CD) can be calculated by dePakeing the powder ²H NMR spectra represented in Fig. 2 for both POPC-d₃₁ and POPS-d₃₁. Fig. 3 reveals a characteristic profile of decreasing order (S_CD) with increasing distance from the glycerol backbone for the pure bilayer and 2mol % Sap C-bound bilayer at 25°C. In (Figure 3A), there is no significant difference in the order parameter profile of neutral phospholipids (POPC-d₃₁) in the presence of 2mol % of Sap C when compared to the control sample (without Sap C). Conversely, for the POPS sample (Figure 3B), there is a significant decrease in the order parameter profile values (mainly in the first half and not the second half of the acyl chain) upon interaction of negatively charged phospholipids (POPS) with 2mol % of Sap C when compared to the control sample (without Sap C). Therefore, the current ²H solid-state NMR data indicate that Sap C disrupts the negatively charged phospholipids (POPS) and not the neutrally charged (POPC) lipids.

Display large version of this figure

Figure 3

The smoothed acyl chains orientational order S_CD profiles calculated from the dePaked ²H NMR spectra (Fig. 2) of both POPC (A) and POPS (B) phospholipid bilayers with and without 2mol % Sap C with respect to the lipids at 25°C.

Additionally, the effect of temperature (25–45°C) on the order and dynamics of the acyl chains of POPC and POPS MLVs was investigated. The ²H solid-state NMR spectra indicate that as the temperature increases, the quadrupolar splittings decrease for all samples with and without the Sap C protein, implying that the mobility of the acyl chains increases as the temperature increases (Supplementary Material, Fig. 1S ).

To confirm the current ²H solid-state NMR spectroscopic results, additional ³¹P NMR spectroscopic studies were conducted as well.

In several cases, the lineshapes of the static ³¹P NMR spectra and its corresponding CSA values have been successfully used to study the perturbation effect induced by proteins on phospholipids 15,33,34. In this study, the static ³¹P NMR spectra of both POPC and POPS lipid bilayers prepared in the absence and presence of 2mol % of Sap C at 25°C are shown in Fig. 4. At 25°C, the motionally averaged powder pattern spectra are characteristic of MLVs in the liquid crystalline phase (L_α) and are expected for POPC and POPS bilayers at a temperature well above the different chain melting point transition temperatures (T_m). The T_m values are −2 and 14°C for POPC and POPS, respectively 37. For both POPS and POPC control samples (Figure 4AC), the phospholipid molecules in the MLVs can be characterized by an axially symmetric motion with fast reorientation about the long molecular axis (gives the sharp peak at the σ-perpendicular (σ_⊥) edge) and a slower reorientation rate perpendicular to that same axis 38. The POPC bilayers containing 2mol % Sap C (Figure 4B) have the same ³¹P lineshape as the pure POPC bilayers (Figure 4A). Moreover, the CSA width of pure POPC MLVs is 48±1ppm (Figure 4A) and the CSA width of POPC MLVs interacting with Sap C is 47±1ppm (Figure 4B), indicating that Sap C is not significantly interacting with the POPC bilayers. The error associated with measuring the CSA width was +/− 1ppm.

Display large version of this figure

Figure 4

³¹P NMR powder-pattern spectra of both POPC and POPS MLVs in the absence (A and C) and in the presence (B and D) of 2mol % Sap C with respect to the lipids at 25°C.

Conversely, the POPS MLVs containing 2mol % Sap C (Figure 4D) has a broader ³¹P lineshape than the pure POPS MLVs (Figure 4C). The decrease in the ³¹P spectral resolution (lineshape) of the Sap C-POPS MLVs (Figure 4D) indicates that the axially symmetric motion of the phospholipids has been altered upon protein incorporation 38. Additionally, the CSA width of the pure POPS MLVs is 55±1ppm (Figure 4C) and the CSA width of POPS MLVs interacting with Sap C is 51±1ppm (Figure 4D). The decrease in ³¹P CSA width of the POPS MLVs upon Sap C insertion could be due to several factors, including changes in orientation, conformation, or dynamics of the phospholipid headgroups. An increase in the axial rotational motion of the lipids as well as other wobbling motions of the phospholipid headgroups may also contribute to this decrease 15,33,34,39,40. Taking into consideration the observed decrease in the ²H order parameter profile values (see Fig. 3), the overall increase in motion and disorder of the upper half of the acyl chain of Sap C-POPS MLVs may be another reason for the decrease in the ³¹P CSA width of the POPS MLVs. The ³¹P lineshape and CSA width data indicate that the phospholipid headgroups are only perturbed by the insertion of Sap C into the negatively charged POPS bilayers and not into the neutral POPC lipids.

Additionally, the ³¹P powder pattern NMR spectra of both POPC and POPS MLVs, with and without 2mol % Sap C, were recorded at temperatures ranging from 25°C to 45°C (Supplementary Material, Fig. 2S ) and the chemical shift anisotropy width were measured by simulation using the DIMFIT software program (Supplementary Material, Fig. 3S ) 35. The data indicate that as the temperature increases, the CSA width of both POPC and POPS bilayers (with and without the Sap C) decreases, indicating that the axial rotational motion of the phospholipid headgroups increases with temperature.

To further probe the interaction of Sap C with the headgroups of POPC and POPS phospholipids, ³¹P MAS T₁ relaxation experiments were conducted (Fig. 5). ³¹P T₁ spin-lattice relaxation rates are sensitive to rapid conformational changes of lipid acyl chains and polar headgroups as well as to changes in the long axis rotation and diffusion of lipids upon peptide interaction 41. The ³¹P T₁ is directly related to the correlation time (τ_C) of fast molecular motion and the larmor frequency (ω_o) 42,43. This frequency (ω_o) is fixed when the same magnetic field is used in the whole study. The relation between T₁, τ_C, and ω_o can be summarized according to 1/T₁∝τ_C/(1+) 42,43. The relaxation mechanism of certain phospholipid membrane systems is efficient when the molecular correlation time (τ_C) fits the equation (ω_oτ_C≈1), which results in T₁ to be at its minimum 42,43. When τ_C moves away from 1/ω_o, the T₁ value increases 33,42,43. It has been previously reported that the temperature can also modulate τ_C42,43. The ³¹P MAS isotropic peak positions of POPC and POPS MLVs with and without 2mol % Sap C were recorded. POPC MLVs showed sharp isotropic peaks at −1.17 and −1.11±0.05ppm, with and without Sap C at 25°C, respectively. Conversely, POPS MLVs showed a sharp isotropic peak at −0.57±0.05ppm for the pure lipid sample and a broader isotropic peak at −0.63±0.05ppm for the sample with 2mol % Sap C (Supplementary Material, Fig. 4S ). At higher temperatures (35°C and 45°C), the peak positions slightly decreased (within 0.15ppm range; data not shown).

Display large version of this figure

Figure 5

³¹P longitudinal relaxation time (T₁) as a function of temperature (25–45°C) for both POPC (A) and POPS (B) phospholipid bilayers with and without 2mol % Sap C with respect to the lipids. The open squares (A) and triangles (B) represent POPC and POPS bilayer controls (without protein), respectively. The solid squares (A) and triangles (B) represent POPC and POPS bilayers with 2mol % Sap C, respectively. The error bars were obtained by averaging ³¹P T₁ values from two different samples.

In Figure 5A, the ³¹P T₁ value increases as the temperature increases from 25 to 45°C in the presence and absence of Sap C. This indicates that the POPC lipid's minimum T₁ value would be lower than 25°C with or without Sap C. The ³¹P T₁ value of pure POPC MLVs has been reported to increase with temperature from 25 to 45°C, which is typical for such hydrated phospholipid bilayers 33,42,43. This result confirms our static ³¹P solid-state NMR data, which indicate that Sap C is not significantly interacting with the POPC bilayers. The ³¹P T₁ value, ³¹P CSA width, and S_CD order parameters of pure POPC MLVs are very close (within 6% deviation) to what was reported previously 34.

Conversely, Figure 5B reveals that the ³¹P T₁ relaxation time of the 2mol % Sap C-POPS sample is considerably less than the POPS control at temperatures ranging from 25 to 45°C. At 25°C, the ³¹P T₁ relaxation value difference between the 2mol % Sap C-POPS sample and the POPS control is ∼250ms. In addition, Figure 5B indicates that this ³¹P T₁ gap increases as the temperature increases and that a minimum T₁ would occur at higher temperatures (when τ_C≈1/ω_o). In a recent article, the antimicrobial peptide G15 was studied using similar approaches 39. Opposite to Sap C-POPS MLVs, G15 induces an increase in the CSA width, as well as an increase in the T₁ relaxation times of the G15 containing MLVs 39. The increase in the T₁ relaxation times was explained as a result of a decrease in the axial rotational motion of the lipids leading to a less efficient relaxation 39. Similar to G15, the incorporation of another antimicrobial peptide (KIGAKI) to phospholipid MLVs has been reported to increase the ³¹P T₁ relaxation values 33. This less efficient relaxation was also attributed to a reduction in the fast axial rotation motion of the lipids caused by the interaction between the peptide and the lipid 33. Conversely, in the case of Sap C, the dramatic decrease in the ³¹P T₁ values of Sap C/POPS bilayers when compared to the Sap C/POPS control sample is obviously caused by the Sap C protein perturbing the headgroup region and increasing the axial rotational motion of the POPS negatively charged bilayers (more efficient relaxation). The decrease in the ³¹P T₁ values was not observed with the neutral POPC lipids/Sap C system. This agrees with the results of the static ²H and ³¹P solid-state NMR spectroscopic experiments discussed above.

To further probe the effect of Sap C on POPC and POPS phospholipids, mechanically oriented bilayers were prepared and investigated as well.

Solid-state NMR spectroscopic studies of mechanically oriented phospholipid bilayers/protein systems on glass plates can provide pertinent structural information on the protein embedded inside the membrane 44. In this study, this method was used to probe the specific interaction of Sap C with aligned phospholipid bilayers. Fig. 6 shows the ³¹P solid-state NMR spectra of oriented POPC and POPS phospholipid bilayers in the absence (A) and presence (B) of Sap C, when the bilayer normal is aligned parallel to the static magnetic field at 25°C. All the samples in Fig. 6 were prepared exactly under the same experimental conditions and repeated twice for consistency. Additionally, similar spectra were collected for the same samples when the bilayer normal was aligned parallel to the static magnetic field (data not shown).

Display large version of this figure

Figure 6

The ³¹P NMR spectra of oriented bilayer samples of both POPC and POPS bilayers in the absence (A and C) and in the presence (B–F) of 2mol % Sap C with respect to the lipids at 25°C. (B) Sap C was dissolved in a minimal amount of TFE and mixed with the lipids dissolved in chloroform. After application onto the glass plates and drying of the protein-lipid complex overnight, a neutral buffer (pH 7.0) was added at the top of the dried bilayers before stacking of the glass plates. Sample D was prepared exactly the same as sample B except that POPS was used instead of POPC phospholipids. (E) POPS alone was spotted (on the glass plates) and dried. Then the protein dissolved in a neutral buffer (pH 7.0) was added at the top of the dried phospholipids before stacking. Sample F was prepared exactly the same as samples B and D except that methanol was used to initially dissolve Sap C instead of TFE. All spectra were conducted with the bilayer normal of the sample aligned parallel to the static magnetic field.

In Figure 6A, the spectrum exhibits one predominant peak at ∼30±1ppm for the parallel orientation. Also, one predominant peak at approximately at −15±1ppm for the perpendicular orientation was observed (data not shown). The sharp peak and the lack of any powder components, indicate that the pure POPC phospholipids were well aligned. In Figure 6B, the Sap C (dissolved and protonated in TFE) and POPC (dissolved in chloroform) lipids were mixed together before spotting and drying on top of the glass plates, and the buffer used to stack the glass plates was added last. For this sample, the spectra reveal one predominant peak at ∼29±1ppm (Figure 6B) for the parallel orientation and one peak at about −15±1ppm for the perpendicular orientation (data not shown), indicating that all Sap C/POPC phospholipids bilayers are still well aligned. In addition, the ³¹P lineshapes of Sap C/POPC bilayers (Figure 6B) did not change when compared to the control sample (Figure 6A). Figure 6B indicates that Sap C did not induce any serious perturbations in the structure and alignment of the POPC bilayer at pH 7.0.

Conversely, Figure 6CF, shows the ³¹P solid-state NMR spectra of oriented POPS phospholipid bilayers with (D–F) and without (C) Sap C when the bilayer normal is aligned parallel to the magnetic field. In Figure 6C, the spectrum shows one predominant peak at ∼39±1ppm, implying that all POPS pure phospholipids are well aligned. Figure 6DF, represents three methods of Sap C addition to the POPS phospholipid bilayers. For comparison, it is important here to emphasize the presence of the protonated form of Sap C (to initialize its electrostatic binding with negatively charged phospholipids) in one sample (Figure 6D) as well as the presence of the unprotonated Sap C form in the other two samples (Figure 6EF). It is also important to distinguish between two types of mixing of the Sap C with the POPS bilayers, specifically the insertion (Figure 6DF) versus membrane association (Figure 6E). In the first method (Figure 6D), Sap C (dissolved and protonated in TFE) was added to the POPS phospholipids before addition to the glass plates. Then, after the protein-bilayer mixture was dried, the neutral buffer was added and the sample was placed into the humidity chamber. This first method of sample preparation is common in the literature and has been used before for the β-hairpin antimicrobial peptides studied using oriented samples 45. Also, the helix tilt angle of the M2 transmembrane peptide from the influenza A virus has been investigated using the protein and the lipid samples initially cosolubilized in TFE 46. Also, gramicidin D, a transmembrane peptide and ovispirin, a surface peptide were initially dissolved in TFE and used as model systems to determine the membrane peptide orientation using solid-state NMR spectroscopy 47.

Using a similar method, the spectrum in Figure 6D indicates that Sap C was able to significantly disrupt the alignment of the POPS bilayers (demonstrated by the splitting of the single peak of the control sample (Figure 6C) into several peaks (Figure 6D)). This splitting may indicate that the POPS bilayers were fragmented into different components as a result of Sap C insertion. Each component has a local bilayer normal that has a distinguished angle with respect to the magnetic field. These data indicate that after the neutral buffer addition, the Sap C-POPS suspension (and unlike Sap C-POPC suspension) was unable to form intact well-oriented bilayers. Thus, the POPS bilayer is significantly perturbed. Conversely, a different type of perturbation effect was observed in the ³¹P spectrum of the 2mol % Sap C-POPS MLVs sample (see Figure 4D). Only the ³¹P line shape and CSA widths of the POPS MLVs changed significantly when compared to the control (pure POPS MLVs). Unlike oriented samples (Figure 6D), no additional components or isotropic peaks were observed in Figure 4D. This inconsistency between oriented samples and MLVs has been observed in the literature 40. In a recent report, the inconsistency was attributed to differences in the method used to investigate the protein perturbation effect on the membranes 40. Additional factors such as the ability of the lipid to align in the presence of the protein as well as differences in hydration levels (70% and 30–50%) between MLVs and oriented bilayers may lead to this inconsistency 40. Taking together with the inconsistency in the ³¹P CSA values between MLVs and oriented bilayers, it can be proposed that the POPS oriented bilayers are exclusively more sensitive to perturbation effects induced by Sap C when compared to the corresponding POPS MLVs. Also, this suggests that the headgroups of the negatively charged POPS oriented bilayers have different conformation, order, architecture, and/or dynamics when compared to the headgroups of the corresponding POPS MLVs 48.

In the spectrum shown in Figure 6E, a sample was prepared using an alternative method, where Sap C was mixed with the neutral buffer and added to the dried phospholipids bilayers that were spotted at the top of the glass plates and the spectrum was recorded. Clearly, the spectrum show one predominant peak at ∼39±1ppm for the parallel orientation, implying that all POPS/Sap C phospholipids bilayers are still well aligned. Similarly, one ³¹P peak at −19±1ppm was observed for the perpendicular orientation (data not shown). Figure 6E indicates that the addition of 2mol % of Sap C to the negatively charged POPS bilayer did not significantly disturb the alignment of the POPS phospholipid bilayers. These data are expected because unprotonated Sap C has been reported not to interact significantly with PS lipids under neutral conditions 5,13. This alternative method is not commonly used to prepare protein-lipid oriented samples for solid-state NMR spectroscopic studies; however, it is commonly used to prepare Sap C samples studied using other techniques that include direct atomic force microscopy observation 36, solution NMR in a detergent environment 13, Trp fluorescence and quenching experiments 16, fluorescent dye leakage 8, and fluorescence resonance energy transfer 6. In all of these reports, Sap C was mixed with previously prepared lipid vesicles in the buffer at the desired pH. Then, the effect of Sap C association on the fusion, leakage, or distribution of these vesicles was measured.

Finally, a sample was prepared similar to the first method (Figure 6D) with a slight modification. In the modified method (Figure 6F), methanol was used to initially dissolve Sap C. Using methanol to dissolve proteins is common for preparing oriented samples for solid-state NMR spectroscopic studies and has been used to study the protein-lipid interaction of a magainin analog (MSI-78) and the synthetic lipopeptide MSI-843 49,50. In addition, the orientational behavior of phospholipid membranes with the toxic peptide mastoparan (isolated from wasp venom) was studied on oriented samples prepared from the peptide and DMPC initially cosolubilized in a methanol/chloroform mixture 51. Also, the target selectivity of the antimicrobial peptide G15 was investigated using an oriented sample of protein initially dissolved in a mixture of methanol and chloroform 39.

Similarly, this third method was used in this study and the spectra show one predominant peak at ∼39±1ppm for the parallel orientation (Figure 6F) and one at about −19±1ppm for the perpendicular orientation (data not shown), indicating that all POPS/Sap C phospholipid bilayers are well aligned. Additionally, the spectra in Figure 6F indicate that unprotonated Sap C was unable to significantly disrupt the alignment of the POPS bilayers when compared to the spectra shown in Figure 6D. This result is similar to the data of the second method (see Figure 6E) and is expected as discussed above 5,13.

Finally, most of the spectra shown in Fig. 6 have additional small peaks besides the predominant component of the most intense peak (other than those discussed for Figure 6D). The smaller peaks represent additional minor components that were imperfectly aligned or minor powder-pattern components that were not aligned at all.

It is clear based upon the results of the ²H order parameter profiles, ³¹P chemical shift anisotropy width, and ³¹P T₁ relaxation times of the MLVs, as well as the ³¹P NMR data of the oriented bilayers, that protonated Sap C perturbs the POPS bilayers and not the POPC when compared to the control (pure MLVs) under the conditions presented in this study. Additionally, since protonated Sap C does not interact strongly with the neutral POPC lipids as indicated by the ³¹P and ²H solid-state NMR data, POPC can be used as a control model membrane. The initial acidic binding conditions between Sap C and POPS lipids was provided using the weak Bronsted acids (TFE) which have been reported to be a good H-bond donor and to have a protonation effect on proteins 31. The measured pH of TFE used in this report was ∼5.3. This pH is enough to initiate the Sap C interaction with the negatively charged lipids 13. Acidic conditions have been reported to be necessary to initialize the electrostatic interaction between the Sap C protein and negatively charged phospholipids 5,13. Chloroform is more volatile than TFE; therefore, the protein-lipid suspension was given more time under the steady steam of N₂ gas to completely dry the sample when compared to the controls (no protein or TFE). Interestingly, although the pH of the HEPES buffer added to the dried protein-phospholipid complex was neutral (pH 7.0), the perturbation effect induced by Sap C on the POPS MLVs was observed using all our solid-state NMR experiments. Finally, to confirm that the perturbation effect is caused by Sap C and not by any residual TFE used to initially dissolve the protein, another control sample was prepared. In this second procedure, the phospholipids (dissolved in chloroform) were mixed with ∼50μl of TFE. Then, the sample was dried using a steady steam of N₂ gas, placed in a vacuum desiccator overnight, and resuspended in HEPES buffer (pH=7.0) as described previously. The pure MLVs prepared in the absence or presence of TFE (as described above) showed similar static ³¹P NMR spectra (Supplementary Material, Fig. 5S ).

Sap C not only facilitates the phospholipids lysosomal digestion but also participates in lipid trafficking as a transfer protein 2,9. However, whether the Sap C-lipid complex is capable of transferring lipids to their targets through different pH environments (without the dissociation back to lipids and protein) is still unclear in the literature. In a previous report, de Alba and co-workers utilized solution NMR spectroscopy to study the association of Sap C with vesicles prepared with L-α-phosphatidylecholine from egg and L-α-phosphatidyleserine (using a high protein/lipid molar ratio (1:1) or 0.5:0.5mM protein/total lipid mixture) 13. When the pH of this system (pH=5.4) was increased back to the original state (pH=6.8), the authors noticed that the NMR signal intensity recovers to 91% of its original value implying that the majority of the protein gets released. The authors concluded that the binding of Sap C to phospholipid vesicles is a pH-controlled reversible process under the conditions described in their report 13. Moreover, this conclusion was based upon the assumption that vesicles that have a molecular mass larger than 100kDa tumble very slow, and therefore, the NMR signal that was observed in the presence of vesicles was considered to belong only to the Sap C that is free in solution 13.

Conversely, in this study, we probed the effect of Sap C insertion on both the headgroups’ mobility and the acyl chains disorder of the POPC and POPS phospholipids. Using this approach, the data cannot confirm directly whether or not Sap C was partially released from the Sap C-POPS complex after its resuspension in a neutral buffer and whether or not the binding process is reversible. However, our data can clearly confirm using several solid-state NMR experiments that a significant protein-lipid perturbation effect was observed after the exposure of protonated Sap-C/POPS complex to a neutral pH under the current experimental conditions (low protein/lipid molar ratio (1:50); or 2mol % protein with respect to lipid). A similar observation has been made in the case of the binding of Sap C to PS liposomal vesicles. Salt concentrations have been known to affect the binding of Sap C to PS membranes 52. The insertion of Sap C into PS membranes is blocked at 0.4M sodium chloride (NaCl). However, the Sap C-PS complex is very stable at salt concentrations up to 1M 52.

This study indicates that ²H order parameter profiles, ³¹P CSA widths, and ³¹P T₁ relaxation times has been successfully used to probe the interaction of the protonated Sap C with neutral (POPC) and negatively charged (POPS) phospholipids at neutral pH conditions. Additionally, oriented phospholipids on glass plates were used to investigate the perturbation effect of Sap C inserted into both POPS and POPC using ³¹P solid-state NMR spectroscopy. The Sap C, protonated in a weak acid, was unable to produce well-oriented POPS bilayers. From the prospective of the lipids’ acyl chain and headgroup dynamics, our new solid-state NMR spectroscopic spectra indicate that Sap C (initially protonated in a weak acid) perturbs the negatively charged (POPS) and not the neutrally charged lipid headgroups (POPC). Unexpectedly, the protonated Sap C was able to conserve its perturbation effect on the POPS membrane even after the Sap C-POPS system was resuspended in neutral buffer. This suggests that Sap C inserted into POPS bilayers initially form a stable complex that lasts even after its resuspension in neutral buffer. Similar complexes have been reported to be stable even at high salt concentrations up to 1M 52. Finally, as expected, unprotonated Sap C is unable to directly associate with the POPS membrane under neutral pH conditions.