Structural Study of Caveolin-1 Intramembrane Domain by Circular Dichroism and Nuclear Magnetic Resonance Guanhua Yang,1 Zhe Dong,1 Haoran Xu,2 Chunyu Wang,1 Haichao Li,2 Zhengqiang Li,2 Fei Li1 1
State Key Laboratory of Supramolecular Structure and Materials, Jilin University, Changchun 130012, People’s Republic of China 2
Key Laboratory for Molecular Enzymology & Engineering, The Ministry of Education, Jilin University, Changchun 130012, People’s Republic of China Received 25 July 2014; revised 22 October 2014; accepted 30 November 2014 Published online 4 December 2014 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/bip.22597
ABSTRACT: Caveolin-1 is a main structural component of caveolae and essential for the invagination of caveolae by forming a hairpin-shaped structure in the membrane-inserting domain (residues, 102–122). In this article, we determined the tertiary structures of the peptides comprising residues 93–126 and 101–126 of caveolin-1 in
membrane-inserting domain and the segment Thr93– Arg101 flanking the membrane-inserting domain may play a role in the self-association of the caveolin-1 protein C 2014 Wiley Periodicals, Inc. at cellular membrane. V
Biopolymers (Pept Sci) 104: 11–20, 2015. Keywords: caveolin-1; intramembrane domain; structure; NMR; CD
1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) aqueous solution and sodium dodecyl sulfate (SDS) micelles, respectively, by nuclear magnetic resonance (NMR) study. The self-association of the peptides in SDS and dodecylphosphocholine (DPC) micelles was also studied by circular
This article was originally published online as an accepted preprint. The “Published Online” date corresponds to the preprint version. You can request a copy of any preprints from the past two calendar years by emailing the Biopolymers editorial office at [email protected].
dichroism (CD), NMR, and sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) techni-
INTRODUCTION
ques. Our results indicated that both peptides form a
aveolae are flask-shaped plasma membrane invaginations, being rich in cholesterol, sphingomyelin, and glycosphingolipids. They are prevalent in most terminally differentiated cell types, notably smooth muscle cells and adipocytes,1 where they exert a number of cellular functions or processes.2–6 The involvement of caveolae in a number of disease processes has been suggested.7–10 Caveolae contain a specific membrane protein, the caveolin1 (Cav-1). Cav-1 not only controls caveolae formation11 but also able to interact with a myriad of proteins in various signal transduction pathways and to modulate activity of the signaling proteins.12–18 Cav-1 is anchored to the membrane with a hydrophobic membrane-inserting segment (residues, 102–122) and three palmitoyl chains attached to Cys residues (residues, 133, 143, and 156) as well as a so-called scaffolding domain (residues, 80–101).19,20 The segment comprising residues 102–134 has
helix–break–helix structure with two helices spanning over Leu103–Phe107 and Ile117–His126 and a loop ranging over Gly108–Gly116. The longer peptide 93–126 showed a stronger propensity to aggregate than the shorter peptide 101–126 in the micelles. Our results suggested that the glycine residues at positions 108 and 116 are important for the break of the helical structure of the Correspondence to: Fei Li, State Key Laboratory of Supramolecular Structure and Materials, Jilin University, 2699 Qianjin Avenue, Changchun 130012, People’s Republic of China, Fax: 186-431-85193421, Tel.: 186-431-85168548; e-mail: [email protected] Contract grant sponsor: The NSFC Contract grant number: 20934002 C 2014 Wiley Periodicals, Inc. V
PeptideScience Volume 104 / Number 1
C
11
12
Yang et al.
MATERIALS AND METHODS Materials FIGURE 1 Schematic representation of Cav-1 protein and the domains interested in this study.
also been proposed as the membrane-inserting segment.21,22 The experimental evidence has indicated that the membraneinserting segment forms an unusual hairpin-shaped structure rather than a membrane-spanning a-helix that brings both the N- and the C-terminal domains of the molecule in the cytoplasm.21,23–25 Therefore, the membrane-inserting segment is also called as intramembrane domain (IMD). The hairpinshaped structure of the IMD is associated with the high degree of membrane curvature existed in caveolae.26,27 The recent NMR study of Cav-1 96–136 segment demonstrated a helix– break–helix conformation that could provide an evidence for the formation of a hairpin-shaped structure predicted for Cav1.28 The residue Pro110 in the IMD has been shown to be important for the formation of a hairpin-shaped structure of the protein in the membrane.20,29 The helical bend will disappear with the mutation of Pro110.28 The Cav-1 scaffolding domain (CSD) is not only responsible for membrane binding but also critical for oligomerization, protein interactions, and cholesterol recognition.30–35 The CSD is believed to be involved in caveolin–cholesterol interaction through the presence of a particular motif V94-T-K-Y-W-F-YR101 in the C-terminal part, which matches the cholesterol recognition amino-acid consensus (CRAC).36–38 The residues 82– 95 in the N-terminal part of the CSD have been evidenced to be able to mediate signaling protein inhibition.35 The two distinct functional parts are believed to have different structures based on the secondary structure prediction and experimental data. The former is an a-helix and the latter a b-strand.39 The residues 84–94 have been suggested to form a b-sheet hairpinshaped structure and play a role in the self-oligomerization of the Cav-1 protein at cell membrane.19 Although the structures of Cav-1 segments including IMD and CSD have been studied previously, no direct structural information at atomic level is available. In addition, as the Cterminal part of CSD and the linker of CSD with IMD, the CRAC segment may also play a role in the self-oligomerization of Cav-1. However, no relevant study is reported. Therefore, we undertook the studies of the structures and assemblies of the peptides corresponding to the residues 93–126 and 101– 126 of Cav-1 using CD, NMR, and SDS-PAGE techniques. Our results provide an insight into the structures of the functional domains of Cav-1 at atomic level and suggest a possible role of the CRAC motif in the aggregation of the Cav-1 protein.
The peptides from human Cav-1 93–126 and 101–126, named Cav1(93–126) and Cav-1(101–126) in this study (Figure 1), were synthesized and purified by GL Biochem (Shanghai). To increase the solubility of the peptides in the purification, three Lys residues were added on each terminal end of Cav-1(93–126) peptide, two Lys residues were added in the N-terminal end, and three Lys residues were added in the C-terminal end of Cav-1(101–126). The previous studies have demonstrated that tagging hydrophobic peptides with Lys residues does not affect the native characteristics of the peptides including insertion in membrane mimetic environments and fold as an a-helical structure, and even oligomeric states.40 The purity was assessed by high-performance liquid chromatography and mass spectroscopy to be >95%. The organic solvents 1,1,1,3,3,3-hexafluoro-2-propanol HFIP-d2 (>98%) and HFIP (>99.5%) were obtained from Acros Organics (Morris Plains, NJ). Deuterated SDS-d25 (>98%) and D2O (>99.8%) were purchased from Cambridge Isotope Laboratories (Tewksbury, MA). The detergents DPC (>99%) and SDS (>99%) were obtained from Sigma-Aldrich (St. Louis, MO). All chemicals were used as purchased directly without further purification.
Sample Preparation The NMR sample of 1 mM of Cav-1(93–126) in 60%HFIP-d2 aqueous solution was prepared by solubilizing appropriate amount of peptide in 300 mL of pure HFIP-d2 solvent and then adding deionized water to a total volume of 500 mL. The NMR samples of 1 mM of peptides in 120 mM SDS-d25 micellar solution were prepared by the previously used method.41 Briefly, the peptide and SDS-d25 were first solubilized in 200 mL of HFIP and 200 mL of deionized water, respectively. The HFIP solution of the peptide was then added into the SDSd25 solution. The mixture was further diluted with 1.6 mL of deionized water and lyophilized overnight under vacuum to remove the solvent completely. The resulting dry powder was rehydrated with a H2O/D2O (9/1 v/v) mixture to obtain a total volume of 500 mL. The CD samples were prepared using similar approaches as described above in the NMR sample preparation. For the samples of the peptides in 60%HFIP aqueous solution, the final concentration of the peptide was 24 mM. For the samples of the peptides in SDS or DPC micellar solution, the concentrations of the peptide and the detergent were 24 mM and 50 mM, respectively.
Two-Dimensional NMR Experiments and Structural Calculation All 1H NMR spectra were performed on a Bruker Avance 600 spectrometer (Bruker BioSpin, F€allanden, Switzerland) equipped with a cryoprobe at 25 C using sodium salt of 2,2-dimethyl-2-silapentane-5sulfonate-d6 as an internal standard. The TOCSY (total correlation spectroscopy) spectra were acquired using the MLEV-17 pulse sequence with a mixing time of 100 ms. The NOESY (nuclear Overhauser effect spectroscopy) experiments were performed with a mixing time of 50, 75, 100, 125, 150, and 200 ms and that with mixing time of 100 ms was used to determine the structures of the peptides. The WATERGATE technique was applied both in the TOCSY and in the NOESY experiments for water signal suppression. The data points
Biopolymers (Peptide Science)
Structural Study of Caveolin-1 Intramembrane Domain
13
Table I Statistics of 20 Structures With the Lowest Target Functions for Cav(93–126) in 60%HFIP-d2/40%H2O and Cav(101–126) in SDS-d25 at pH 5.56, 25 C
Average target functions (A˚2) Number of nonredundant distance restraints Intraresidual (|i 2 j| 5 0) Sequential (|i 2 j| 5 1) Medium (|i 2 j| 4) Long range (|i 2 j| > 4) Average sum of distance restraint violations (A˚) Average max. distance restraint violation (A˚) Average sum of torsion angle restraint violations ( ) Average max. of torsion angle restraint violation ( ) ˚) R.m.s. deviation from the mean structure (A All residues Backbone heavy atoms All heavy atoms Residues Backbone heavy atoms All heavy atoms Residues Backbone heavy atoms All heavy atoms Ramachandran plot statistics (for each helix span) Residues Residues in most favored region (%) Residues in additionally allowed region (%) Residues in generously allowed region (%) Residues in disallowed region (%) Residues Residues in most favored region (%) Residues in additionally allowed region (%) Residues in generously allowed region (%) Residues in disallowed region (%)
of 2 K in F2 dimension and increments of 512 in F1 dimension were collected with scans of 64 for each increment and a relaxation delay of 2 s. The two-dimensional (2D) NMR spectra were processed using standard Bruker software and analyzed using software SPARKY. The structural calculation was carried out using software CYANA (version 1.0.6)42 from 200 initial conformers with random torsion angle values. The 20 structures with the lowest target functions were visualized by MOLMOL program.43 The stereochemical qualities of the peptide structure were evaluated by PROCHECK-NMR software.44 The NMR restraints used in calculations and structural statistics are listed in Table I.
Far-UV CD Measurements The Far-UV CD spectra were recorded on a PMS-450 spectropolarimeter (Biologic, France) at room temperature using a 0.5-mm pathlength quartz cell. Data were recorded from 190 to 260 nm at a speed of 6 nm/min. The background signal was subtracted from all spectra. Final spectra were smoothed using a fast Fourier transform filter, and the intensity was expressed as mean residue ellipticity. The secondary structure contents were estimated by the CDPro software package
Biopolymers (Peptide Science)
Cav(93–126)
Cav(101–126)
0.34 6 0.11 292 162 85 43 2 2.0 6 0.5 0.25 6 0.06 1.0 6 1.4 0.76 6 1.12
0.79 6 0.28 249 135 70 44 0 3.6 6 0.8 0.39 6 0.14 0.3 6 0.7 0.26 6 0.71
7.69 6 1.87 9.12 6 1.76 14–18 1.18 6 0.38 2.25 6 0.50 28–37 0.71 6 0.31 1.75 6 0.51
5.19 6 1.53 6.79 6 1.50 5–9 0.62 6 0.30 1.66 6 0.53 19–28 0.73 6 0.43 1.77 6 0.67
14–18 31.0 45.0 22.0 2.0 28–37 67.0 31.5 1.5 0.0
5–9 76.0 22.0 1.0 1.0 19–28 78.5 20.5 1.0 0.0
using the program CONTIN/LL. A reference set of SMP56 including 56 proteins was used in the analyses of CD data.
SDS-PAGE Experiments A Tricine buffer system was employed, but increasing the SDS concentration to 0.5% (17 mM) in the gel, which is above the critical micelle concentration.45 The peptides incorporated with 100 mM of SDS in 10 mM of Tris buffer at pH 7.4 were prepared as described in Sample Preparation section. The peptide concentrations of 40 and 80 mM were used. The samples were diluted with 43 SDS loading buffer containing the tracking dye Serva Blue G, and were boiled for 5 min before loading on to a 16.5%-polyacryalamide/Tricine/SDS gel. The electrophoresis was performed at 4 C. A voltage of 30 V was used before the sample entered the separation gel and then a voltage of 130 V was used. After gel electrophoresis, the gel was placed in a solution containing 0.5% of gluteraldehyde and 30% of ethanol for 20 min prior to staining. The bands of peptides were visualized with Coomassie Brilliant Blue stain. The protein standards with ultra low molecular weight (3.3–20.1 kDa) were used to estimate the apparent molecular weights of the peptides studied.
14
Yang et al.
FIGURE 2 CD spectra of Cav-1(93–126) (A) and Cav-1(101–126) (B) in 60%HFIP aqueous solution (black), SDS micelles (red), and DPC micelles (green).
RESULTS Secondary Structure Analysis by CD Spectroscopy The CD spectrum of the peptide Cav-1(93–126) in 60%HFIP aqueous solution exhibited a positive peak at 195 nm and double minima at 210 and 225 nm (Figure 2). The CD spectra of the peptide Cav-1(93–126) in SDS and DPC micelles exhibited similar absorbance mode: a positive peak at 195 nm and double minima at 212 and 225 nm (Figure 2). Similar CD spectra were also observed for the peptide Cav-1(101–126) in the same media. This indicates that both peptides fold predominantly as a-helical structures in these conditions and the structures of the peptides in 60%HFIP aqueous solution are comparable with those of the peptides in SDS and DPC micelles. The ahelical contents obtained by the deconvolution of the CD spectra are summarized in Table II.
Self-Assembly Analysis It is worth noting that the two minima in the CD spectra of the micelle-bound Cav-1(93–126) are very close in the intensity ([h]225/[h]212 5 1), whereas the absorbance at 210 nm is Table II. Secondary structure data for the caveolin-1 peptides in different environments. Secondary Structure (%) Peptide
Environment
Helix
Strand
Turn
Unordered
Cav(93–126)
60%HFIP SDS DPC 60%HFIP SDS DPC
59.1 50.4 54.7 61.4 58.7 58.7
7.3 10.7 14.8 4.8 5.1 7.2
12.2 15.7 11.3 11.9 15.6 12.4
21.4 23.1 19.2 22 20.6 21.6
Cav(101–126)
clearly lower than that at 225 nm for the peptide in HFIP aqueous solution ([h]225/[h]210 5 0.9). It is likely that the peptide Cav-1(93–126) exists as an association state in the micelles.46,47 The oligomers may be disassociated in HFIP aqueous solution owing to the strong ability of HFIP to disassociate peptide/protein aggregates. In contrast, the CD spectra of the peptide Cav-1(101–126) in the three different environments showed similar absorbance modes with the absorbance at 210/212 nm evidently lower than that at 225 nm ([h]225/ [h]210/212 5 0.9). This suggests that the peptide Cav-1(93–126) in the micelles self-associates more easily than the peptide Cav1(101–126) in the micelles. The N-terminal CRAC-containing segment may play a role for the enhanced self-assembly propensity of Cav-1(93–126). One-dimensional 1H NMR spectra of the two peptides in SDS-d25 micelles were recorded. The 1H NMR spectrum of Cav-1(93–126) in SDS-d25 micelles was very poor (Figure 3). One-dimensional 1H NMR spectrum of the peptide Cav-1(93– 126) in DPC-d38 micelles was also recorded and a spectral appearance with very poor dispersion and low intensity of signals was also observed. Severe broadening and overlapping of resonance signals prevented us from obtaining high-resolution spectrum. However, the NMR spectrum was dramatically improved by solubilization of the peptide in 60%HFIP-d2 aqueous solution. This further indicates that the peptide Cav1(93–126) is self-associated in the micelles. In contrast, the 1H NMR spectrum of Cav-1(101–126) in SDS-d25 micelles was as good as that of Cav-1(93–126) in 60%HFIP-d2 aqueous solution, suggesting that the shorter peptide without the flanking CRAC-containing segment is likely monomeric or disaggregated to a less oligomeric state in the micelles. The self-assembly state of the peptides in SDS micelles was further tested by the SDS-PAGE experiment, which is often used to probe qualitatively the occurrence of transmembrane Biopolymers (Peptide Science)
Structural Study of Caveolin-1 Intramembrane Domain
FIGURE 3 One-dimensional 1HNMR spectra of Cav-1(93–126) in SDS-d25 micelles at pH 5.51 (A), in DPC-d38 micelles at pH 3.34 (B) and in 60%HFIP-d2/40%H2O (C) as well as that of Cav-1(101– 126) in SDS-d25 micelles at pH 5.56 (D).
helix oligomerization.48,49 The peptide concentrations of 40 and 80 mM were used in different lanes of the gel to probe whether the peptide concentration affects its aggregation behavior. As shown in Figure 4, the peptide Cav-1(101–126) displayed a clearly visible band at a position of 7.8 kDa, corresponding to a dimer (molecular mass of monomer, 3.6 kDa), whereas the peptide Cav-1(93–126) showed a broader band centered at a position of 14.4 kDa, corresponding to a trimer (molecular mass of monomer, 4.8 kDa). The peptide concentration has no considerable effect on the aggregate state. The electrophoresis results also suggest that Cav-1(93–126) is more prone to aggregate than Cav-1(101–126). If the same oligomeric states of the peptides were formed in the NMR samples, the two Cav-1(101–126) molecules in SDS-d25 micelles should be homogeneous in topology, whereas the three Cav-1(93– 126) molecules in the micelles should be topologically heterogeneous.
15
Ha(i) 2 HN(i 1 4) in the range of residues Trp17–His28 were obtained from the NOESY spectrum. However, the medium distance NOE connectivity was not observed in the range of residues Ile11–Trp17. This suggests that the peptide Cav1(101–126) forms a helix–break–helix structure in the micelles. The dynamic calculation based on the NOE connectivities demonstrated that the two helical domains comprising residues Leu5–Phe9 and Ile19–His28 (corresponding to 103–107 and 117–126 in the Cav-1 protein) were well defined and the region between the two helices was flexible (Figure 6). The 20 structures with the lowest target functions were well converged in each helix segment and most of the backbone dihedral angles u and w of the helices were distributed in the allowed region of the Ramachandran plot (Table I). The formation of the helix–break–helix structure was also predicted by the Ha chemical shift index (CSI). As shown in Figure 7, the Ha chemical shifts of residues both from Ile19 to the C-terminus and from Arg3 to Phe9 shifted upfield relative to the reference data50 and most of the difference values were more negative than 20.1 (indicating an a-helical structure according to the CSI50). In contrast, the consecutive upfield shifts were broken by several downfield shifts in the region of residues from Gly10 to Gly18 (corresponding to Gly108 and Gly116 in the Cav-1 protein). Although the downfield shift values of the dHa of the two Gly residues were less than 10.1, the calculated structure based on the NOE connectivities clearly demonstrated that the helix of the peptide was broken in the region between the two Gly residues. This suggests that the two Gly residues are crucial to the disruption of the helix of the peptide Cav-1(101–126). The important role of Pro12 (or Pro110 in Cav-1) in the disruption of the helix was implicated in a dramatic downfield shift of Ile11 (or Ile109 in Cav-1).
Tertiary Structure of Cav-1(93–126) in 60%HFIP-d2 Aqueous Solution The NMR structure of Cav-1(93–126) was studied in 60%HFIP-d2/40%H2O, but not in SDS-d25 micelles because of
Tertiary Structure of Cav-1(101–126) in SDS-d25 Micelles The 2D TOCSY and NOESY NMR spectra of the Cav-1(101– 126) peptide in SDS-d25 micelles were recorded and assigned. The fingerprint region of the NOESY spectrum with sequential assignment of residue is shown in Figure 5A and the NOE connectivities obtained from the NOESY spectrum are shown in Figure 5B. A series of NOE connectivities of Ha(i) 2 Hb(i 1 3) in the range of residues Leu4–Ile11 and a series of NOE connectivities of Ha(i) 2 HN(i 1 3), Ha(i) 2 Hb(i 1 3), and Biopolymers (Peptide Science)
FIGURE 4 SDS-PAGE data of the Cav-1 peptides at the concentrations of 40 mM (lanes, a and c) and 80 mM (lanes, b and d).
16
Yang et al.
FIGURE 5 Ha–HN region of 2D NOESY spectra (A) and NOE connectivities (B) for Cav-1(101– 126) in SDS-d25 micelles at 25 C.
the severe broadening of the 1H NMR signals of the peptide in the micelles (Figure 3). The fluoric alcohols (e.g., 2,2,2-trifluoroethanol and HFIP) have been widely used to simulate hydro-
phobic environments and to establish the in vivo conformations of many peptides that either reside within or are bound to membranes.51–55 The CD results have shown that
FIGURE 6 An ensemble of the 20 backbone structures with the lowest target function fitted over each helical region and a ribbon representation of the mean structure of the 20 structures for Cav1(101–126) in SDS-d25 micelles.
Biopolymers (Peptide Science)
Structural Study of Caveolin-1 Intramembrane Domain
17
DISCUSSION
FIGURE 7 The Ha chemical shifts of Cav-1(101–126) in SDS-d25 micelles relative to the reference chemical shifts from Ref. 47. The CSI values of 20.1 and 10.1 are indicated by dashed lines. The residues corresponding to positions 108, 110, and 116 in Cav-1 are shown in this figure.
the structure of the peptide in 60%HFIP/40%H2O is comparable with that of the peptide in SDS micelles (Figure 2). The 1H NMR chemical shifts of the peptide Cav-1(93–126) in 60%HFIP-d2 aqueous solution were assigned by the combinational analysis of TOCSY and NOESY spectra. Although there were crowded crosspeaks in the 2D NOESY spectrum, some of them were even overlapped, most of the signals were dispersive and could be unambiguously assigned (Figure 8A), and the NOE connectivities represented by the crosspeaks were obtained (Figure 8B). The NOE connectivities included nearly continuous HN(i) 2 HN(i 1 1) and Hb(i) 2 HN(i 1 1) both in the range from Arg12 to Gly19 and in the range from Trp26 to the C-terminus, several Ha(i) 2 HN(i 1 3) and Ha(i) 2 Hb(i 1 3) in the span over Tyr11–Phe18 and a series of Ha(i) 2 HN(i 1 3), Ha(i) 2 HN(i 1 4), and Ha(i) 2 Hb(i 1 3) in the range of Ile25–Leu36. The structural calculation based on these NOE connectivities demonstrated that the two helical domains from Leu14 to Phe18 and from Ile28 to His37 (corresponding to 103–107 and 117–126 in Cav-1, respectively), separated by a flexible segment comprising residues Gly19–Gly27 (corresponding to Gly108–Gly116 in Cav-1), were well defined (Figure 9). The CSI from the Ha chemical shifts confirmed the significant roles of Gly19, Gly27, and Pro21 in disrupting helix by the upfield shifts of the two Gly residues and Ile20 (Figure 10). Notably, the CSI method predicted a longer helix approximately from Lys7 to Ala16 in the N-terminal half of the peptide. However, only a short helix from Leu14 to Phe18 was obtained and the flanking segment Thr4–Arg12 (corresponding to Thr93–Arg101 in full Cav-1 protein) was poorly defined by the NOE restraints and the NOE-based structure calculation. Biopolymers (Peptide Science)
All the measurements of CD, NMR, and SDS-PAGE showed an evidence that the peptide Cav-1(93–126), corresponding to the intramembrane segment of Cav-1 flanked by a CRACcontaining segment, was more aggregated in SDS and DPC micelles than the peptide Cav-1(101–126), an intramembrane segment without the N-terminal CRAC motif. This indicates that the presence of the flanking CRAC-containing segment can promote the self-association of the intramembrane peptide and suggests that the CRAC-motif may be important for the aggregation of the peptide. The CRAC motif encompassing residues 94–101 is the C-terminal part of the CSD and is involved in the recognition and binding of cholesterol in caveolae. The residues 84–94 in the N-terminal part of CSD were suggested to form a b-sheet hairpin-shaped structure and play a role in self-oligomerization of the caveolin protein and interaction with other proteins at cell membrane.19 Our results suggest that the CRAC-motif may also play a role in the aggregation of Cav-1 induced by the CSD domain. The structures of both the residues 94–102 in DPC micelles and the residues 82–109 in lipid membrane containing cholesterol were studied previously by NMR, and the results showed an a-helix in the region 96–101.39,56 In our study, the CSI from the Ha chemical shifts of Cav-1(93–126) also indicated the formation of an a-helix in the region of 96–101. According to the a-helical contents obtained by CD, the peptide Cav-1(93–126) in 60%HFIP aqueous solution has the helix longer than the peptide Cav-1(101–126) in the same medium by approximately five residues. Most of the additional residues can be reasonably assigned to the residues in the CRAC motif of Cav1(93–126) if residues 101–126 in the two IMD-related peptides adopt the same structure. However, the medium range NOE crosspeaks from the flanking segment could not be unambiguously assigned in the NOESY spectrum likely owing to signal overlap or owing to too weak signal intensity. Thus, the helical structure in this region was not obtained based on the NOE data of Cav-1(93–126) in HFIP aqueous solution. Although there were uncertainty in the formation of the helical structure in residues 96–101, a helix–break–helix structure in the intramembrane segment encompassing residues 101–126 was obtained by 2D NMR measurements and dynamic calculation for Cav-1(93–126) in HFIP aqueous solution and for Cav-1(101–126) in SDS micelles. The two peptides in different media displayed coincident residue ranges in the helices and the break spans, that is, the residues 103–107 and 117–126 were involved in an a-helix and the residues between the two helical segments were flexible. The formation of similar structures in residues 101–126 in different media suggests that the IMD of Cav-1 alone is primarily responsible
18
Yang et al.
FIGURE 8 Ha–HN region of 2D NOESY spectra (A) and NOE connectivities (B) for Cav-1(93– 126) in 60%HFIP-d2/40%H2O at 25 C.
FIGURE 9 An ensemble of the 20 backbone structures with the lowest target function fitted over each helical region and a ribbon representation of the mean structure of the 20 structures for Cav1(93–126) in 60%HFIP-d2/40%H2O.
Biopolymers (Peptide Science)
Structural Study of Caveolin-1 Intramembrane Domain
FIGURE 10 The Ha chemical shifts of Cav-1(93–126) in 60%HFIP-d2/40%H2O relative to the reference chemical shifts from Ref. 47. The CSI values of 20.1 and 10.1 are indicated by dashed lines. The residues corresponding to positions 108, 110, and 116 in Cav-1 are shown in this figure.
for the hairpin conformation. The additional flanking segment in Cav-1(93–126) with respect to Cav-1(101–126) has no effect on the structure of the intramembrane segment although the presence of the flanking segment made the peptide more prone to aggregate. It was noted that the break between two helices was longer than that predicted by the chemical shift index for the Cav-1 peptides 96–136 incorporated with lyso-myristoyl phosphatidylglycerol micelles.28,57 The residues from 111 to 116 that were predicted to be helical were found in this study of the peptides Cav1(93–126) and Cav-1(101–126) to be unstructured. The previous studies have revealed that the amino acid P110 are critical for this helix–break–helix motif and G108 is also important for this structural fashion.28,29 We indicated in this study that the residue G116 is also important for the break of the intramembrane helix.
CONCLUSIONS In conclusion, we obtained the structure of the IMD of Cav-1 at atomic level by the NMR study of the peptides Cav-1(101– 126) in SDS-d25 micelles and Cav-1(93–126) in HFIP-d2 aqueous solution. We showed that the structures of both peptides are basically coincident in residues 101–126, characterized by a helix–break–helix motif, and both Gly108 and Gly116 are responsible for the break of helix in addition to Pro110. We suggested that the segment encompassing residues 93–100 (or CRAC motif comprising residues 94–101) may play a role in the intermolecular interaction of Cav-1 at cellular membrane.
REFERENCES 1. Williams, T. M.; Lisanti, M. P. Ann Med 2004, 36, 584–595. 2. Anderson, R. G. Annu Rev Biochem 1998, 67, 199–225.
Biopolymers (Peptide Science)
19
3. Shaul, P. W.; Anderson, R. G. Am J Physiol 1998, 275, L843– L851. 4. Razani, B.; Woodman, S. E.; Lisanti, M. P. Pharmacol Rev 2002, 54, 431–467. 5. Razani, B.; Lisanti, M. P. J Clin Invest 2001, 108, 1553–1561. 6. Smart, E. J.; Graf, G. A.; McNiven, M. A.; Sessa, W. C.; Engelman, J. A.; Scherer, P. E.; Okamoto, T.; Lisanti, M. P. Mol Cell Biol 1999, 19, 7289–7304. 7. Frank, P. G.; Woodman, S. E.; Park, D. S.; Lisanti, M. P. Arterioscler Thromb Vasc Biol 2003, 23, 1161–1168. 8. Hnasko, R.; Lisanti, M. P. Mol Interv 2003, 3, 445–464. 9. Gaudreault, S. B.; Dea, D.; Poirier, J. Neurobiol Aging 2004, 25, 753–759. 10. Simons, K.; Ehehalt, R. J Clin Invest 2002, 110, 597–603. 11. Parton, R. G.; Hanzal-Bayer, M.; Hancock, J. F. J Cell Sci 2006, 119, 787–796. 12. Galbiati, F.; Volonte, D.; Liu, J.; Capozza, F.; Frank, P. G.; Zhu, L.; Pestell, R. G.; Lisanti, M. P. Mol Biol Cell 2001, 12, 2229– 2244. 13. Fielding, P. E.; Russel, J. S.; Spencer, T. A.; Hakamata, H.; Nagao, K.; Fielding, C. J. Biochemistry 2002, 41, 4929–4937. 14. Gargalovic, P.; Dory, L. J Lipid Res 2003, 44, 1622–1632. 15. Gargalovic, P.; Dory, L. J Lipid Res 2003, 44, 11–21. 16. Le, P. U.; Guay, G.; Altschuler, Y.; Nabi, I. R. J Biol Chem 2002, 277, 3371–3379. 17. Chao, W. T.; Fan, S. S.; Chen, J. K.; Yang, V. C. J Lipid Res 2003, 44, 1094–1099. 18. Fu, Y.; Hoang, A.; Escher, G.; Parton, R. G.; Krozowski, Z.; Sviridov, D. J Biol Chem 2004, 279, 14140–14146. 19. Spisni, E.; Tomasi, V.; Cestaro, A.; Tosatto, S. C. E. Biochem Biophys Res Commun 2005, 338, 1383–1390. 20. Aoki, S.; Thomas, A.; Decaffmeyer, M.; Brasseur, R.; Epand, R. M. J Biol Chem 2010, 285, 33371–33380. 21. Schlegel, A.; Lisanti, M. P. J Biol Chem 2000, 275, 21605–21617. 22. Razani, B.; Woodman, S. E.; Lisanti, M. P. Pharmacol Rev 2002, 54, 431–467. 23. Parton, R. G.; Hanzal-Bayer, M.; Hancock, J. F. J Cell Sci 2006, 119, 787–796. 24. Ostermeyer, A. G.; Ramcharan, L. T.; Zeng, Y. C.; Lunlin, D. M.; Brown, D. A. J Cell Biol 2004, 164, 69–78. 25. Monier, S.; Parton, R. G.; Vogel, F.; Behlke, J.; Henske, A.; Kurzchalia, T. V. Mol Biol Cell 1995, 6, 911–927. 26. Li, S.; Song, K. S.; Lisanti, M. P. J Biol Chem 1996, 271, 568– 573. 27. Senes, A.; Engel, D. E.; DeGrado, W. F. Curr Opin Struct Biol 2004, 14, 465–479. 28. Lee, J.; Glover, K. J. Biochim Biophys Acta 2012, 1818, 1158– 1164. 29. Aoki, S.; Epand, R. M. Biochim Biophys Acta 2012, 1818, 12– 18. 30. Li, S.; Couet, J.; Lisanti, M. P. J Biol Chem 1996, 271, 29182– 29190. 31. Schlegel, A.; Schwab, R. B.; Scherer, P. E.; Lisanti, M. P. J Biol Chem 1999, 274, 22660–22667. 32. Fernandez, I.; Ying, Y. S.; Albanesi, J.; Anderson, R. G. W. Proc Natl Acad Sci USA 2002, 99, 11193–11198. 33. Couet, J.; Li, S.; Okamoto, T.; Ikezu, T.; Lisanti, M. P. J Biol Chem 1997, 272, 6525–6533.
20
Yang et al.
34. Epand, R. M.; Sayer, B. G.; Epand, R. F. J Mol Biol 2005, 345, 339–350. 35. Bernatchez, P. N.; Bauer, P. M.; Yu, J.; Prendergast, J. S.; He, P. N.; Sessa, W. C. Proc Natl Acad Sci USAm 2005, 102, 761–766. 36. Li, H.; Yao, Z. X.; Degenhardt, B.; Teper, G.; Papadopoulos, V. Proc Natl Acad Sci USA 2001, 98, 1267–1272. 37. Schlegel, A.; Schwab, R. B.; Scherer, P. E.; Lisanti, M. P. J Biol Chem 1999, 274, 22660–22667. 38. Epand, R. M.; Sayer, B. G.; Epand, R. F. J Mol Biol 2005, 345, 339–350. 39. Hoop, C. L.; Sivanandam, V. N.; Kodali, R.; Srnec, M. N.; van der Wel, P. C. A. Biochemistry 2012, 51, 90–99. 40. Cunningham, F.; Deber, C. M. Methods 2007, 41, 370–380. 41. Li, F.; Li, H.; Hu, L.; Kwan, M.; Chen, G.; He, Q. Y.; Sun, H. J Am Chem Soc 2005, 127, 1414–1423. 42. Koradi, R.; Billeter, M.; Wuthrich, K. J Mol Graph 1996, 14, 51– 55, 29–32. 43. Xiao, S. Y.; Li, J. T.; Wang, Y. X.; Wang, C. Y.; Xue, R.; Wang, S.; Li, F. Biochim Biophys Acta 2010, 1798, 1556–1564. 44. Laskowski, R. A.; Rullmannn, J. A.; MacArthur, M. W.; Kaptein, R.; Thornton, J. M. J Biomol NMR 1996, 8, 477–486. 45. Li, H.; Li, F.; Sun, H.; Qian, Z. M. Biochem J 2003, 372, 757–766.
46. Choy, N.; Raussens, V.; Narayanaswami, V. J Mol Biol 2003, 334, 527–539. 47. Dutta, K.; Alexandrov, A.; Huang, H.; Pascal, S. M. Protein Sci 2001, 10, 2531–2540. 48. Lemmon, M. A.; Flanagan, J. M.; Hunt, J. F.; Adair, B. D.; Bormann, B. J.; Dempsey, C. E.; Engelman, D. M. J Biol Chem 1992, 267, 7683–7689. 49. Li, E.; You, M.; Hristova, K. Biochemistry 2005, 44, 352–360. 50. Wishart, D. S.; Sykes, B. D.; Richards, F. M. Biochemistry 1992, 31, 1647–1651. 51. Gerig, J. T. Biophys J 2004, 86, 3166–3175. 52. Hirota, N.; Mizuno, K.; Goto, Y. J Mol Biol 1998, 275, 365–378. 53. Walgers R, L. T., Cammers-Goodwin A. J Am Chem Soc 1998, 120, 5073–5079. 54. Diaz, M. D.; Fioroni, M.; Burger, K.; Berger, S. Chemistry 2002, 8, 1663–1669. 55. Crescenzi, O.; Tomaselli, S.; Guerrini, R.; Salvadori, S.; D’Ursi, A. M.; Temussi, P. A.; Picone, D. Eur J Biochem 2002, 269, 5642–5648. 56. Le Lan, C.; Gallay, J.; Vincent, M.; Neumann, J. M.; de Foresta, B.; Jamin, N. Eur Biophys J 2010, 39, 307–325. 57. Rieth, M. D.; Lee, J.; Glover, K. J. Biochemistry 2012, 51, 3911–3918.
Biopolymers (Peptide Science)
State Key Laboratory of Supramolecular Structure and Materials, Jilin University, Changchun 130012, People’s Republic of China 2
Key Laboratory for Molecular Enzymology & Engineering, The Ministry of Education, Jilin University, Changchun 130012, People’s Republic of China Received 25 July 2014; revised 22 October 2014; accepted 30 November 2014 Published online 4 December 2014 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/bip.22597
ABSTRACT: Caveolin-1 is a main structural component of caveolae and essential for the invagination of caveolae by forming a hairpin-shaped structure in the membrane-inserting domain (residues, 102–122). In this article, we determined the tertiary structures of the peptides comprising residues 93–126 and 101–126 of caveolin-1 in
membrane-inserting domain and the segment Thr93– Arg101 flanking the membrane-inserting domain may play a role in the self-association of the caveolin-1 protein C 2014 Wiley Periodicals, Inc. at cellular membrane. V
Biopolymers (Pept Sci) 104: 11–20, 2015. Keywords: caveolin-1; intramembrane domain; structure; NMR; CD
1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) aqueous solution and sodium dodecyl sulfate (SDS) micelles, respectively, by nuclear magnetic resonance (NMR) study. The self-association of the peptides in SDS and dodecylphosphocholine (DPC) micelles was also studied by circular
This article was originally published online as an accepted preprint. The “Published Online” date corresponds to the preprint version. You can request a copy of any preprints from the past two calendar years by emailing the Biopolymers editorial office at [email protected].
dichroism (CD), NMR, and sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) techni-
INTRODUCTION
ques. Our results indicated that both peptides form a
aveolae are flask-shaped plasma membrane invaginations, being rich in cholesterol, sphingomyelin, and glycosphingolipids. They are prevalent in most terminally differentiated cell types, notably smooth muscle cells and adipocytes,1 where they exert a number of cellular functions or processes.2–6 The involvement of caveolae in a number of disease processes has been suggested.7–10 Caveolae contain a specific membrane protein, the caveolin1 (Cav-1). Cav-1 not only controls caveolae formation11 but also able to interact with a myriad of proteins in various signal transduction pathways and to modulate activity of the signaling proteins.12–18 Cav-1 is anchored to the membrane with a hydrophobic membrane-inserting segment (residues, 102–122) and three palmitoyl chains attached to Cys residues (residues, 133, 143, and 156) as well as a so-called scaffolding domain (residues, 80–101).19,20 The segment comprising residues 102–134 has
helix–break–helix structure with two helices spanning over Leu103–Phe107 and Ile117–His126 and a loop ranging over Gly108–Gly116. The longer peptide 93–126 showed a stronger propensity to aggregate than the shorter peptide 101–126 in the micelles. Our results suggested that the glycine residues at positions 108 and 116 are important for the break of the helical structure of the Correspondence to: Fei Li, State Key Laboratory of Supramolecular Structure and Materials, Jilin University, 2699 Qianjin Avenue, Changchun 130012, People’s Republic of China, Fax: 186-431-85193421, Tel.: 186-431-85168548; e-mail: [email protected] Contract grant sponsor: The NSFC Contract grant number: 20934002 C 2014 Wiley Periodicals, Inc. V
PeptideScience Volume 104 / Number 1
C
11
12
Yang et al.
MATERIALS AND METHODS Materials FIGURE 1 Schematic representation of Cav-1 protein and the domains interested in this study.
also been proposed as the membrane-inserting segment.21,22 The experimental evidence has indicated that the membraneinserting segment forms an unusual hairpin-shaped structure rather than a membrane-spanning a-helix that brings both the N- and the C-terminal domains of the molecule in the cytoplasm.21,23–25 Therefore, the membrane-inserting segment is also called as intramembrane domain (IMD). The hairpinshaped structure of the IMD is associated with the high degree of membrane curvature existed in caveolae.26,27 The recent NMR study of Cav-1 96–136 segment demonstrated a helix– break–helix conformation that could provide an evidence for the formation of a hairpin-shaped structure predicted for Cav1.28 The residue Pro110 in the IMD has been shown to be important for the formation of a hairpin-shaped structure of the protein in the membrane.20,29 The helical bend will disappear with the mutation of Pro110.28 The Cav-1 scaffolding domain (CSD) is not only responsible for membrane binding but also critical for oligomerization, protein interactions, and cholesterol recognition.30–35 The CSD is believed to be involved in caveolin–cholesterol interaction through the presence of a particular motif V94-T-K-Y-W-F-YR101 in the C-terminal part, which matches the cholesterol recognition amino-acid consensus (CRAC).36–38 The residues 82– 95 in the N-terminal part of the CSD have been evidenced to be able to mediate signaling protein inhibition.35 The two distinct functional parts are believed to have different structures based on the secondary structure prediction and experimental data. The former is an a-helix and the latter a b-strand.39 The residues 84–94 have been suggested to form a b-sheet hairpinshaped structure and play a role in the self-oligomerization of the Cav-1 protein at cell membrane.19 Although the structures of Cav-1 segments including IMD and CSD have been studied previously, no direct structural information at atomic level is available. In addition, as the Cterminal part of CSD and the linker of CSD with IMD, the CRAC segment may also play a role in the self-oligomerization of Cav-1. However, no relevant study is reported. Therefore, we undertook the studies of the structures and assemblies of the peptides corresponding to the residues 93–126 and 101– 126 of Cav-1 using CD, NMR, and SDS-PAGE techniques. Our results provide an insight into the structures of the functional domains of Cav-1 at atomic level and suggest a possible role of the CRAC motif in the aggregation of the Cav-1 protein.
The peptides from human Cav-1 93–126 and 101–126, named Cav1(93–126) and Cav-1(101–126) in this study (Figure 1), were synthesized and purified by GL Biochem (Shanghai). To increase the solubility of the peptides in the purification, three Lys residues were added on each terminal end of Cav-1(93–126) peptide, two Lys residues were added in the N-terminal end, and three Lys residues were added in the C-terminal end of Cav-1(101–126). The previous studies have demonstrated that tagging hydrophobic peptides with Lys residues does not affect the native characteristics of the peptides including insertion in membrane mimetic environments and fold as an a-helical structure, and even oligomeric states.40 The purity was assessed by high-performance liquid chromatography and mass spectroscopy to be >95%. The organic solvents 1,1,1,3,3,3-hexafluoro-2-propanol HFIP-d2 (>98%) and HFIP (>99.5%) were obtained from Acros Organics (Morris Plains, NJ). Deuterated SDS-d25 (>98%) and D2O (>99.8%) were purchased from Cambridge Isotope Laboratories (Tewksbury, MA). The detergents DPC (>99%) and SDS (>99%) were obtained from Sigma-Aldrich (St. Louis, MO). All chemicals were used as purchased directly without further purification.
Sample Preparation The NMR sample of 1 mM of Cav-1(93–126) in 60%HFIP-d2 aqueous solution was prepared by solubilizing appropriate amount of peptide in 300 mL of pure HFIP-d2 solvent and then adding deionized water to a total volume of 500 mL. The NMR samples of 1 mM of peptides in 120 mM SDS-d25 micellar solution were prepared by the previously used method.41 Briefly, the peptide and SDS-d25 were first solubilized in 200 mL of HFIP and 200 mL of deionized water, respectively. The HFIP solution of the peptide was then added into the SDSd25 solution. The mixture was further diluted with 1.6 mL of deionized water and lyophilized overnight under vacuum to remove the solvent completely. The resulting dry powder was rehydrated with a H2O/D2O (9/1 v/v) mixture to obtain a total volume of 500 mL. The CD samples were prepared using similar approaches as described above in the NMR sample preparation. For the samples of the peptides in 60%HFIP aqueous solution, the final concentration of the peptide was 24 mM. For the samples of the peptides in SDS or DPC micellar solution, the concentrations of the peptide and the detergent were 24 mM and 50 mM, respectively.
Two-Dimensional NMR Experiments and Structural Calculation All 1H NMR spectra were performed on a Bruker Avance 600 spectrometer (Bruker BioSpin, F€allanden, Switzerland) equipped with a cryoprobe at 25 C using sodium salt of 2,2-dimethyl-2-silapentane-5sulfonate-d6 as an internal standard. The TOCSY (total correlation spectroscopy) spectra were acquired using the MLEV-17 pulse sequence with a mixing time of 100 ms. The NOESY (nuclear Overhauser effect spectroscopy) experiments were performed with a mixing time of 50, 75, 100, 125, 150, and 200 ms and that with mixing time of 100 ms was used to determine the structures of the peptides. The WATERGATE technique was applied both in the TOCSY and in the NOESY experiments for water signal suppression. The data points
Biopolymers (Peptide Science)
Structural Study of Caveolin-1 Intramembrane Domain
13
Table I Statistics of 20 Structures With the Lowest Target Functions for Cav(93–126) in 60%HFIP-d2/40%H2O and Cav(101–126) in SDS-d25 at pH 5.56, 25 C
Average target functions (A˚2) Number of nonredundant distance restraints Intraresidual (|i 2 j| 5 0) Sequential (|i 2 j| 5 1) Medium (|i 2 j| 4) Long range (|i 2 j| > 4) Average sum of distance restraint violations (A˚) Average max. distance restraint violation (A˚) Average sum of torsion angle restraint violations ( ) Average max. of torsion angle restraint violation ( ) ˚) R.m.s. deviation from the mean structure (A All residues Backbone heavy atoms All heavy atoms Residues Backbone heavy atoms All heavy atoms Residues Backbone heavy atoms All heavy atoms Ramachandran plot statistics (for each helix span) Residues Residues in most favored region (%) Residues in additionally allowed region (%) Residues in generously allowed region (%) Residues in disallowed region (%) Residues Residues in most favored region (%) Residues in additionally allowed region (%) Residues in generously allowed region (%) Residues in disallowed region (%)
of 2 K in F2 dimension and increments of 512 in F1 dimension were collected with scans of 64 for each increment and a relaxation delay of 2 s. The two-dimensional (2D) NMR spectra were processed using standard Bruker software and analyzed using software SPARKY. The structural calculation was carried out using software CYANA (version 1.0.6)42 from 200 initial conformers with random torsion angle values. The 20 structures with the lowest target functions were visualized by MOLMOL program.43 The stereochemical qualities of the peptide structure were evaluated by PROCHECK-NMR software.44 The NMR restraints used in calculations and structural statistics are listed in Table I.
Far-UV CD Measurements The Far-UV CD spectra were recorded on a PMS-450 spectropolarimeter (Biologic, France) at room temperature using a 0.5-mm pathlength quartz cell. Data were recorded from 190 to 260 nm at a speed of 6 nm/min. The background signal was subtracted from all spectra. Final spectra were smoothed using a fast Fourier transform filter, and the intensity was expressed as mean residue ellipticity. The secondary structure contents were estimated by the CDPro software package
Biopolymers (Peptide Science)
Cav(93–126)
Cav(101–126)
0.34 6 0.11 292 162 85 43 2 2.0 6 0.5 0.25 6 0.06 1.0 6 1.4 0.76 6 1.12
0.79 6 0.28 249 135 70 44 0 3.6 6 0.8 0.39 6 0.14 0.3 6 0.7 0.26 6 0.71
7.69 6 1.87 9.12 6 1.76 14–18 1.18 6 0.38 2.25 6 0.50 28–37 0.71 6 0.31 1.75 6 0.51
5.19 6 1.53 6.79 6 1.50 5–9 0.62 6 0.30 1.66 6 0.53 19–28 0.73 6 0.43 1.77 6 0.67
14–18 31.0 45.0 22.0 2.0 28–37 67.0 31.5 1.5 0.0
5–9 76.0 22.0 1.0 1.0 19–28 78.5 20.5 1.0 0.0
using the program CONTIN/LL. A reference set of SMP56 including 56 proteins was used in the analyses of CD data.
SDS-PAGE Experiments A Tricine buffer system was employed, but increasing the SDS concentration to 0.5% (17 mM) in the gel, which is above the critical micelle concentration.45 The peptides incorporated with 100 mM of SDS in 10 mM of Tris buffer at pH 7.4 were prepared as described in Sample Preparation section. The peptide concentrations of 40 and 80 mM were used. The samples were diluted with 43 SDS loading buffer containing the tracking dye Serva Blue G, and were boiled for 5 min before loading on to a 16.5%-polyacryalamide/Tricine/SDS gel. The electrophoresis was performed at 4 C. A voltage of 30 V was used before the sample entered the separation gel and then a voltage of 130 V was used. After gel electrophoresis, the gel was placed in a solution containing 0.5% of gluteraldehyde and 30% of ethanol for 20 min prior to staining. The bands of peptides were visualized with Coomassie Brilliant Blue stain. The protein standards with ultra low molecular weight (3.3–20.1 kDa) were used to estimate the apparent molecular weights of the peptides studied.
14
Yang et al.
FIGURE 2 CD spectra of Cav-1(93–126) (A) and Cav-1(101–126) (B) in 60%HFIP aqueous solution (black), SDS micelles (red), and DPC micelles (green).
RESULTS Secondary Structure Analysis by CD Spectroscopy The CD spectrum of the peptide Cav-1(93–126) in 60%HFIP aqueous solution exhibited a positive peak at 195 nm and double minima at 210 and 225 nm (Figure 2). The CD spectra of the peptide Cav-1(93–126) in SDS and DPC micelles exhibited similar absorbance mode: a positive peak at 195 nm and double minima at 212 and 225 nm (Figure 2). Similar CD spectra were also observed for the peptide Cav-1(101–126) in the same media. This indicates that both peptides fold predominantly as a-helical structures in these conditions and the structures of the peptides in 60%HFIP aqueous solution are comparable with those of the peptides in SDS and DPC micelles. The ahelical contents obtained by the deconvolution of the CD spectra are summarized in Table II.
Self-Assembly Analysis It is worth noting that the two minima in the CD spectra of the micelle-bound Cav-1(93–126) are very close in the intensity ([h]225/[h]212 5 1), whereas the absorbance at 210 nm is Table II. Secondary structure data for the caveolin-1 peptides in different environments. Secondary Structure (%) Peptide
Environment
Helix
Strand
Turn
Unordered
Cav(93–126)
60%HFIP SDS DPC 60%HFIP SDS DPC
59.1 50.4 54.7 61.4 58.7 58.7
7.3 10.7 14.8 4.8 5.1 7.2
12.2 15.7 11.3 11.9 15.6 12.4
21.4 23.1 19.2 22 20.6 21.6
Cav(101–126)
clearly lower than that at 225 nm for the peptide in HFIP aqueous solution ([h]225/[h]210 5 0.9). It is likely that the peptide Cav-1(93–126) exists as an association state in the micelles.46,47 The oligomers may be disassociated in HFIP aqueous solution owing to the strong ability of HFIP to disassociate peptide/protein aggregates. In contrast, the CD spectra of the peptide Cav-1(101–126) in the three different environments showed similar absorbance modes with the absorbance at 210/212 nm evidently lower than that at 225 nm ([h]225/ [h]210/212 5 0.9). This suggests that the peptide Cav-1(93–126) in the micelles self-associates more easily than the peptide Cav1(101–126) in the micelles. The N-terminal CRAC-containing segment may play a role for the enhanced self-assembly propensity of Cav-1(93–126). One-dimensional 1H NMR spectra of the two peptides in SDS-d25 micelles were recorded. The 1H NMR spectrum of Cav-1(93–126) in SDS-d25 micelles was very poor (Figure 3). One-dimensional 1H NMR spectrum of the peptide Cav-1(93– 126) in DPC-d38 micelles was also recorded and a spectral appearance with very poor dispersion and low intensity of signals was also observed. Severe broadening and overlapping of resonance signals prevented us from obtaining high-resolution spectrum. However, the NMR spectrum was dramatically improved by solubilization of the peptide in 60%HFIP-d2 aqueous solution. This further indicates that the peptide Cav1(93–126) is self-associated in the micelles. In contrast, the 1H NMR spectrum of Cav-1(101–126) in SDS-d25 micelles was as good as that of Cav-1(93–126) in 60%HFIP-d2 aqueous solution, suggesting that the shorter peptide without the flanking CRAC-containing segment is likely monomeric or disaggregated to a less oligomeric state in the micelles. The self-assembly state of the peptides in SDS micelles was further tested by the SDS-PAGE experiment, which is often used to probe qualitatively the occurrence of transmembrane Biopolymers (Peptide Science)
Structural Study of Caveolin-1 Intramembrane Domain
FIGURE 3 One-dimensional 1HNMR spectra of Cav-1(93–126) in SDS-d25 micelles at pH 5.51 (A), in DPC-d38 micelles at pH 3.34 (B) and in 60%HFIP-d2/40%H2O (C) as well as that of Cav-1(101– 126) in SDS-d25 micelles at pH 5.56 (D).
helix oligomerization.48,49 The peptide concentrations of 40 and 80 mM were used in different lanes of the gel to probe whether the peptide concentration affects its aggregation behavior. As shown in Figure 4, the peptide Cav-1(101–126) displayed a clearly visible band at a position of 7.8 kDa, corresponding to a dimer (molecular mass of monomer, 3.6 kDa), whereas the peptide Cav-1(93–126) showed a broader band centered at a position of 14.4 kDa, corresponding to a trimer (molecular mass of monomer, 4.8 kDa). The peptide concentration has no considerable effect on the aggregate state. The electrophoresis results also suggest that Cav-1(93–126) is more prone to aggregate than Cav-1(101–126). If the same oligomeric states of the peptides were formed in the NMR samples, the two Cav-1(101–126) molecules in SDS-d25 micelles should be homogeneous in topology, whereas the three Cav-1(93– 126) molecules in the micelles should be topologically heterogeneous.
15
Ha(i) 2 HN(i 1 4) in the range of residues Trp17–His28 were obtained from the NOESY spectrum. However, the medium distance NOE connectivity was not observed in the range of residues Ile11–Trp17. This suggests that the peptide Cav1(101–126) forms a helix–break–helix structure in the micelles. The dynamic calculation based on the NOE connectivities demonstrated that the two helical domains comprising residues Leu5–Phe9 and Ile19–His28 (corresponding to 103–107 and 117–126 in the Cav-1 protein) were well defined and the region between the two helices was flexible (Figure 6). The 20 structures with the lowest target functions were well converged in each helix segment and most of the backbone dihedral angles u and w of the helices were distributed in the allowed region of the Ramachandran plot (Table I). The formation of the helix–break–helix structure was also predicted by the Ha chemical shift index (CSI). As shown in Figure 7, the Ha chemical shifts of residues both from Ile19 to the C-terminus and from Arg3 to Phe9 shifted upfield relative to the reference data50 and most of the difference values were more negative than 20.1 (indicating an a-helical structure according to the CSI50). In contrast, the consecutive upfield shifts were broken by several downfield shifts in the region of residues from Gly10 to Gly18 (corresponding to Gly108 and Gly116 in the Cav-1 protein). Although the downfield shift values of the dHa of the two Gly residues were less than 10.1, the calculated structure based on the NOE connectivities clearly demonstrated that the helix of the peptide was broken in the region between the two Gly residues. This suggests that the two Gly residues are crucial to the disruption of the helix of the peptide Cav-1(101–126). The important role of Pro12 (or Pro110 in Cav-1) in the disruption of the helix was implicated in a dramatic downfield shift of Ile11 (or Ile109 in Cav-1).
Tertiary Structure of Cav-1(93–126) in 60%HFIP-d2 Aqueous Solution The NMR structure of Cav-1(93–126) was studied in 60%HFIP-d2/40%H2O, but not in SDS-d25 micelles because of
Tertiary Structure of Cav-1(101–126) in SDS-d25 Micelles The 2D TOCSY and NOESY NMR spectra of the Cav-1(101– 126) peptide in SDS-d25 micelles were recorded and assigned. The fingerprint region of the NOESY spectrum with sequential assignment of residue is shown in Figure 5A and the NOE connectivities obtained from the NOESY spectrum are shown in Figure 5B. A series of NOE connectivities of Ha(i) 2 Hb(i 1 3) in the range of residues Leu4–Ile11 and a series of NOE connectivities of Ha(i) 2 HN(i 1 3), Ha(i) 2 Hb(i 1 3), and Biopolymers (Peptide Science)
FIGURE 4 SDS-PAGE data of the Cav-1 peptides at the concentrations of 40 mM (lanes, a and c) and 80 mM (lanes, b and d).
16
Yang et al.
FIGURE 5 Ha–HN region of 2D NOESY spectra (A) and NOE connectivities (B) for Cav-1(101– 126) in SDS-d25 micelles at 25 C.
the severe broadening of the 1H NMR signals of the peptide in the micelles (Figure 3). The fluoric alcohols (e.g., 2,2,2-trifluoroethanol and HFIP) have been widely used to simulate hydro-
phobic environments and to establish the in vivo conformations of many peptides that either reside within or are bound to membranes.51–55 The CD results have shown that
FIGURE 6 An ensemble of the 20 backbone structures with the lowest target function fitted over each helical region and a ribbon representation of the mean structure of the 20 structures for Cav1(101–126) in SDS-d25 micelles.
Biopolymers (Peptide Science)
Structural Study of Caveolin-1 Intramembrane Domain
17
DISCUSSION
FIGURE 7 The Ha chemical shifts of Cav-1(101–126) in SDS-d25 micelles relative to the reference chemical shifts from Ref. 47. The CSI values of 20.1 and 10.1 are indicated by dashed lines. The residues corresponding to positions 108, 110, and 116 in Cav-1 are shown in this figure.
the structure of the peptide in 60%HFIP/40%H2O is comparable with that of the peptide in SDS micelles (Figure 2). The 1H NMR chemical shifts of the peptide Cav-1(93–126) in 60%HFIP-d2 aqueous solution were assigned by the combinational analysis of TOCSY and NOESY spectra. Although there were crowded crosspeaks in the 2D NOESY spectrum, some of them were even overlapped, most of the signals were dispersive and could be unambiguously assigned (Figure 8A), and the NOE connectivities represented by the crosspeaks were obtained (Figure 8B). The NOE connectivities included nearly continuous HN(i) 2 HN(i 1 1) and Hb(i) 2 HN(i 1 1) both in the range from Arg12 to Gly19 and in the range from Trp26 to the C-terminus, several Ha(i) 2 HN(i 1 3) and Ha(i) 2 Hb(i 1 3) in the span over Tyr11–Phe18 and a series of Ha(i) 2 HN(i 1 3), Ha(i) 2 HN(i 1 4), and Ha(i) 2 Hb(i 1 3) in the range of Ile25–Leu36. The structural calculation based on these NOE connectivities demonstrated that the two helical domains from Leu14 to Phe18 and from Ile28 to His37 (corresponding to 103–107 and 117–126 in Cav-1, respectively), separated by a flexible segment comprising residues Gly19–Gly27 (corresponding to Gly108–Gly116 in Cav-1), were well defined (Figure 9). The CSI from the Ha chemical shifts confirmed the significant roles of Gly19, Gly27, and Pro21 in disrupting helix by the upfield shifts of the two Gly residues and Ile20 (Figure 10). Notably, the CSI method predicted a longer helix approximately from Lys7 to Ala16 in the N-terminal half of the peptide. However, only a short helix from Leu14 to Phe18 was obtained and the flanking segment Thr4–Arg12 (corresponding to Thr93–Arg101 in full Cav-1 protein) was poorly defined by the NOE restraints and the NOE-based structure calculation. Biopolymers (Peptide Science)
All the measurements of CD, NMR, and SDS-PAGE showed an evidence that the peptide Cav-1(93–126), corresponding to the intramembrane segment of Cav-1 flanked by a CRACcontaining segment, was more aggregated in SDS and DPC micelles than the peptide Cav-1(101–126), an intramembrane segment without the N-terminal CRAC motif. This indicates that the presence of the flanking CRAC-containing segment can promote the self-association of the intramembrane peptide and suggests that the CRAC-motif may be important for the aggregation of the peptide. The CRAC motif encompassing residues 94–101 is the C-terminal part of the CSD and is involved in the recognition and binding of cholesterol in caveolae. The residues 84–94 in the N-terminal part of CSD were suggested to form a b-sheet hairpin-shaped structure and play a role in self-oligomerization of the caveolin protein and interaction with other proteins at cell membrane.19 Our results suggest that the CRAC-motif may also play a role in the aggregation of Cav-1 induced by the CSD domain. The structures of both the residues 94–102 in DPC micelles and the residues 82–109 in lipid membrane containing cholesterol were studied previously by NMR, and the results showed an a-helix in the region 96–101.39,56 In our study, the CSI from the Ha chemical shifts of Cav-1(93–126) also indicated the formation of an a-helix in the region of 96–101. According to the a-helical contents obtained by CD, the peptide Cav-1(93–126) in 60%HFIP aqueous solution has the helix longer than the peptide Cav-1(101–126) in the same medium by approximately five residues. Most of the additional residues can be reasonably assigned to the residues in the CRAC motif of Cav1(93–126) if residues 101–126 in the two IMD-related peptides adopt the same structure. However, the medium range NOE crosspeaks from the flanking segment could not be unambiguously assigned in the NOESY spectrum likely owing to signal overlap or owing to too weak signal intensity. Thus, the helical structure in this region was not obtained based on the NOE data of Cav-1(93–126) in HFIP aqueous solution. Although there were uncertainty in the formation of the helical structure in residues 96–101, a helix–break–helix structure in the intramembrane segment encompassing residues 101–126 was obtained by 2D NMR measurements and dynamic calculation for Cav-1(93–126) in HFIP aqueous solution and for Cav-1(101–126) in SDS micelles. The two peptides in different media displayed coincident residue ranges in the helices and the break spans, that is, the residues 103–107 and 117–126 were involved in an a-helix and the residues between the two helical segments were flexible. The formation of similar structures in residues 101–126 in different media suggests that the IMD of Cav-1 alone is primarily responsible
18
Yang et al.
FIGURE 8 Ha–HN region of 2D NOESY spectra (A) and NOE connectivities (B) for Cav-1(93– 126) in 60%HFIP-d2/40%H2O at 25 C.
FIGURE 9 An ensemble of the 20 backbone structures with the lowest target function fitted over each helical region and a ribbon representation of the mean structure of the 20 structures for Cav1(93–126) in 60%HFIP-d2/40%H2O.
Biopolymers (Peptide Science)
Structural Study of Caveolin-1 Intramembrane Domain
FIGURE 10 The Ha chemical shifts of Cav-1(93–126) in 60%HFIP-d2/40%H2O relative to the reference chemical shifts from Ref. 47. The CSI values of 20.1 and 10.1 are indicated by dashed lines. The residues corresponding to positions 108, 110, and 116 in Cav-1 are shown in this figure.
for the hairpin conformation. The additional flanking segment in Cav-1(93–126) with respect to Cav-1(101–126) has no effect on the structure of the intramembrane segment although the presence of the flanking segment made the peptide more prone to aggregate. It was noted that the break between two helices was longer than that predicted by the chemical shift index for the Cav-1 peptides 96–136 incorporated with lyso-myristoyl phosphatidylglycerol micelles.28,57 The residues from 111 to 116 that were predicted to be helical were found in this study of the peptides Cav1(93–126) and Cav-1(101–126) to be unstructured. The previous studies have revealed that the amino acid P110 are critical for this helix–break–helix motif and G108 is also important for this structural fashion.28,29 We indicated in this study that the residue G116 is also important for the break of the intramembrane helix.
CONCLUSIONS In conclusion, we obtained the structure of the IMD of Cav-1 at atomic level by the NMR study of the peptides Cav-1(101– 126) in SDS-d25 micelles and Cav-1(93–126) in HFIP-d2 aqueous solution. We showed that the structures of both peptides are basically coincident in residues 101–126, characterized by a helix–break–helix motif, and both Gly108 and Gly116 are responsible for the break of helix in addition to Pro110. We suggested that the segment encompassing residues 93–100 (or CRAC motif comprising residues 94–101) may play a role in the intermolecular interaction of Cav-1 at cellular membrane.
REFERENCES 1. Williams, T. M.; Lisanti, M. P. Ann Med 2004, 36, 584–595. 2. Anderson, R. G. Annu Rev Biochem 1998, 67, 199–225.
Biopolymers (Peptide Science)
19
3. Shaul, P. W.; Anderson, R. G. Am J Physiol 1998, 275, L843– L851. 4. Razani, B.; Woodman, S. E.; Lisanti, M. P. Pharmacol Rev 2002, 54, 431–467. 5. Razani, B.; Lisanti, M. P. J Clin Invest 2001, 108, 1553–1561. 6. Smart, E. J.; Graf, G. A.; McNiven, M. A.; Sessa, W. C.; Engelman, J. A.; Scherer, P. E.; Okamoto, T.; Lisanti, M. P. Mol Cell Biol 1999, 19, 7289–7304. 7. Frank, P. G.; Woodman, S. E.; Park, D. S.; Lisanti, M. P. Arterioscler Thromb Vasc Biol 2003, 23, 1161–1168. 8. Hnasko, R.; Lisanti, M. P. Mol Interv 2003, 3, 445–464. 9. Gaudreault, S. B.; Dea, D.; Poirier, J. Neurobiol Aging 2004, 25, 753–759. 10. Simons, K.; Ehehalt, R. J Clin Invest 2002, 110, 597–603. 11. Parton, R. G.; Hanzal-Bayer, M.; Hancock, J. F. J Cell Sci 2006, 119, 787–796. 12. Galbiati, F.; Volonte, D.; Liu, J.; Capozza, F.; Frank, P. G.; Zhu, L.; Pestell, R. G.; Lisanti, M. P. Mol Biol Cell 2001, 12, 2229– 2244. 13. Fielding, P. E.; Russel, J. S.; Spencer, T. A.; Hakamata, H.; Nagao, K.; Fielding, C. J. Biochemistry 2002, 41, 4929–4937. 14. Gargalovic, P.; Dory, L. J Lipid Res 2003, 44, 1622–1632. 15. Gargalovic, P.; Dory, L. J Lipid Res 2003, 44, 11–21. 16. Le, P. U.; Guay, G.; Altschuler, Y.; Nabi, I. R. J Biol Chem 2002, 277, 3371–3379. 17. Chao, W. T.; Fan, S. S.; Chen, J. K.; Yang, V. C. J Lipid Res 2003, 44, 1094–1099. 18. Fu, Y.; Hoang, A.; Escher, G.; Parton, R. G.; Krozowski, Z.; Sviridov, D. J Biol Chem 2004, 279, 14140–14146. 19. Spisni, E.; Tomasi, V.; Cestaro, A.; Tosatto, S. C. E. Biochem Biophys Res Commun 2005, 338, 1383–1390. 20. Aoki, S.; Thomas, A.; Decaffmeyer, M.; Brasseur, R.; Epand, R. M. J Biol Chem 2010, 285, 33371–33380. 21. Schlegel, A.; Lisanti, M. P. J Biol Chem 2000, 275, 21605–21617. 22. Razani, B.; Woodman, S. E.; Lisanti, M. P. Pharmacol Rev 2002, 54, 431–467. 23. Parton, R. G.; Hanzal-Bayer, M.; Hancock, J. F. J Cell Sci 2006, 119, 787–796. 24. Ostermeyer, A. G.; Ramcharan, L. T.; Zeng, Y. C.; Lunlin, D. M.; Brown, D. A. J Cell Biol 2004, 164, 69–78. 25. Monier, S.; Parton, R. G.; Vogel, F.; Behlke, J.; Henske, A.; Kurzchalia, T. V. Mol Biol Cell 1995, 6, 911–927. 26. Li, S.; Song, K. S.; Lisanti, M. P. J Biol Chem 1996, 271, 568– 573. 27. Senes, A.; Engel, D. E.; DeGrado, W. F. Curr Opin Struct Biol 2004, 14, 465–479. 28. Lee, J.; Glover, K. J. Biochim Biophys Acta 2012, 1818, 1158– 1164. 29. Aoki, S.; Epand, R. M. Biochim Biophys Acta 2012, 1818, 12– 18. 30. Li, S.; Couet, J.; Lisanti, M. P. J Biol Chem 1996, 271, 29182– 29190. 31. Schlegel, A.; Schwab, R. B.; Scherer, P. E.; Lisanti, M. P. J Biol Chem 1999, 274, 22660–22667. 32. Fernandez, I.; Ying, Y. S.; Albanesi, J.; Anderson, R. G. W. Proc Natl Acad Sci USA 2002, 99, 11193–11198. 33. Couet, J.; Li, S.; Okamoto, T.; Ikezu, T.; Lisanti, M. P. J Biol Chem 1997, 272, 6525–6533.
20
Yang et al.
34. Epand, R. M.; Sayer, B. G.; Epand, R. F. J Mol Biol 2005, 345, 339–350. 35. Bernatchez, P. N.; Bauer, P. M.; Yu, J.; Prendergast, J. S.; He, P. N.; Sessa, W. C. Proc Natl Acad Sci USAm 2005, 102, 761–766. 36. Li, H.; Yao, Z. X.; Degenhardt, B.; Teper, G.; Papadopoulos, V. Proc Natl Acad Sci USA 2001, 98, 1267–1272. 37. Schlegel, A.; Schwab, R. B.; Scherer, P. E.; Lisanti, M. P. J Biol Chem 1999, 274, 22660–22667. 38. Epand, R. M.; Sayer, B. G.; Epand, R. F. J Mol Biol 2005, 345, 339–350. 39. Hoop, C. L.; Sivanandam, V. N.; Kodali, R.; Srnec, M. N.; van der Wel, P. C. A. Biochemistry 2012, 51, 90–99. 40. Cunningham, F.; Deber, C. M. Methods 2007, 41, 370–380. 41. Li, F.; Li, H.; Hu, L.; Kwan, M.; Chen, G.; He, Q. Y.; Sun, H. J Am Chem Soc 2005, 127, 1414–1423. 42. Koradi, R.; Billeter, M.; Wuthrich, K. J Mol Graph 1996, 14, 51– 55, 29–32. 43. Xiao, S. Y.; Li, J. T.; Wang, Y. X.; Wang, C. Y.; Xue, R.; Wang, S.; Li, F. Biochim Biophys Acta 2010, 1798, 1556–1564. 44. Laskowski, R. A.; Rullmannn, J. A.; MacArthur, M. W.; Kaptein, R.; Thornton, J. M. J Biomol NMR 1996, 8, 477–486. 45. Li, H.; Li, F.; Sun, H.; Qian, Z. M. Biochem J 2003, 372, 757–766.
46. Choy, N.; Raussens, V.; Narayanaswami, V. J Mol Biol 2003, 334, 527–539. 47. Dutta, K.; Alexandrov, A.; Huang, H.; Pascal, S. M. Protein Sci 2001, 10, 2531–2540. 48. Lemmon, M. A.; Flanagan, J. M.; Hunt, J. F.; Adair, B. D.; Bormann, B. J.; Dempsey, C. E.; Engelman, D. M. J Biol Chem 1992, 267, 7683–7689. 49. Li, E.; You, M.; Hristova, K. Biochemistry 2005, 44, 352–360. 50. Wishart, D. S.; Sykes, B. D.; Richards, F. M. Biochemistry 1992, 31, 1647–1651. 51. Gerig, J. T. Biophys J 2004, 86, 3166–3175. 52. Hirota, N.; Mizuno, K.; Goto, Y. J Mol Biol 1998, 275, 365–378. 53. Walgers R, L. T., Cammers-Goodwin A. J Am Chem Soc 1998, 120, 5073–5079. 54. Diaz, M. D.; Fioroni, M.; Burger, K.; Berger, S. Chemistry 2002, 8, 1663–1669. 55. Crescenzi, O.; Tomaselli, S.; Guerrini, R.; Salvadori, S.; D’Ursi, A. M.; Temussi, P. A.; Picone, D. Eur J Biochem 2002, 269, 5642–5648. 56. Le Lan, C.; Gallay, J.; Vincent, M.; Neumann, J. M.; de Foresta, B.; Jamin, N. Eur Biophys J 2010, 39, 307–325. 57. Rieth, M. D.; Lee, J.; Glover, K. J. Biochemistry 2012, 51, 3911–3918.
Biopolymers (Peptide Science)
