Subscriber access provided by UNIV OF BARCELONA
Article
Misfolding of Human Islet Amyloid Polypeptide at Lipid Membrane Populates through #-Sheet Conformers without Involving #-Helical Intermediates Junjun Tan, Jiahui Zhang, Yi Luo, and Shuji Ye J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.8b08537 • Publication Date (Web): 08 Jan 2019 Downloaded from http://pubs.acs.org on January 8, 2019
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Journal of the American Chemical Society
Misfolding of Human Islet Amyloid Polypeptide at Lipid Membrane Populates through -Sheet Conformers without Involving -Helical Intermediates Junjun Tan, Jiahui Zhang, Yi Luo, Shuji Ye* Hefei National Laboratory for Physical Sciences at the Microscale and Department of Chemical Physics, and Synergetic Innovation Center of Quantum Information & Quantum Physics, University of Science and Technology of China, Hefei, 230026, China.
Supporting Information Placeholder transient forms due to their relatively modest time or low structural resolutions. For example, X-ray crystallography can provide high-resolution structures of fibrils,23 but does not have enough time resolution to probe the real-time aggregation process.3,7 Consequently, the basis of IAPP-induced cytotoxicity and the interaction between hIAPP and lipid membranes are not fully understood, and a wide range of mechanisms have been proposed. More specifically, the hIAPP structural details at the initial fibrillation stage are still controversial. The first controversy concerns the pathway of fibril formation. It is uncertain whether the oligomers are consumed during fibril growth. Some studies suggest that the oligomers are on-pathway intermediates that convert into fibrils,3,7 while others claim that the oligomers are off-pathway byproducts that are not on a productive route towards fibrils.10,11 The second controversy is associated with the intermediate conformation at the early oligomerization steps. Two conformational rearrangements have been proposed to initiate IAPP oligomerization,12,13 namely, the formation of -helical intermediates,14-17,22 and side by side hairpin dimers.5-7 In addition, it is also disputed whether the membrane-active oligomers are formed on the membrane from IAPP monomers or preassembled in solution before binding to the membrane.11,15 Resolving these controversies requires quickly capturing the time-dependent structural changes during the aggregation process. In this study, we verify experimentally that the combination of interface-sensitive chiral amide I, achiral amide II and amide III spectral signals of the protein backbone generated in sum frequency generation vibrational spectroscopy (SFG-VS) can provide a unique and powerful tool to capture the hIAPP intermediates at the interface during the aggregation process with sufficient structural and temporal resolutions. An experimental protocol with high-speed simultaneous measurement of chiral and achiral polarization combinations has been developed to reveal the real-time structural evolution of hIAPP at the lipid membrane surface. SFG-VS, a second-order nonlinear optical technique, is a powerful tool for the identification of interfacial molecular structures and dynamics in situ and in real time.24, 25 It has been applied to determine the structures of peptides and proteins at different interfaces by probing amide I vibrations.17-20, 26 -36
ABSTRACT: Amyloid formation has been implicated in many fatal diseases, but its mechanism remains to be clarified due to a lack of effective methods that can capture the transient intermediates. Here, we experimentally demonstrate that sum frequency generation vibrational spectroscopy can unambiguously discriminate the intermediates during amyloid formation at the lipid membrane in situ and in real time by combining the chiral amide I, achiral amide II and amide III spectral signals of the protein backbone. Such a combination can directly identify the formation of -hairpin-like monomers, and β-sheet oligomers, and fibrils. A strong correlation between the amide II signals and the formation of β-sheet oligomers and fibrils was found. With this approach, the structural evolution of human islet amyloid polypeptides (hIAPP) at negative lipid bilayers was elucidated. It was firmly confirmed that hIAPP populates through β-sheet conformers without involving α-helical intermediates. The membrane-associated assembly of hIAPP proceeds by assembling with a -hairpin-like monomer at the lipid bilayer surface, rather than by inserting the preassembled β-sheet oligomers in solution. This newly established protocol is ready to be utilized in revealing the mechanism of amyloid aggregation at the lipid membrane.
Introduction Human islet amyloid polypeptide (hIAPP) is a 37-residue polypeptide hormone co-produced with insulin from pancreatic islet β cells. Its aggregation and misfolding have been implicated in the development of type 2 diabetes.1,2 Because of its clinical importance, hIAPP aggregation has been investigated by many techniques such as X-ray diffraction, NMR, EPR, and optical techniques.3-20 Although it is well established that hIAPP is spontaneously converted to fibrillar structures through a multistep misfolding process in which the monomer converts into various metastable oligomeric forms and ultimately forms fibrils,8,9 it is very challenging to identify the early formed intermediate species due to their transient nature,3,7,21 particularly the intermediates formed at the negative membrane because the aggregation process at such a surface is very fast.22 Therefore, many conventional methods fail to provide the specific structural details of the
1 ACS Paragon Plus Environment
Journal of the American Chemical Society 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Human islet amyloid polypeptide (hIAPP) (sequence: Lys-CysAsn-Thr-Ala-Thr-Cys-Ala-Thr-Gln-Arg-Leu-Ala-Asn-Phe-LeuVal-His-Ser-Ser-Asn-Asn-Phe-Gly-Ala-Ile-Leu-Ser-Ser-Thr-AsnVal-Gly-Ser-Asn-Thr-Tyr-NH2) with a purity of >97% was purchased from Sigma-Aldrich. The hIAPP molecule was not capped with other groups and a disulfide bond was formed to link the Cys2 and Cys7 residues. GP41 protein segment (FP23, sequence: Ala-Val-Gly-Ile-Gly-Ala-Leu-Phe-Leu-Gly-Phe-LeuGly-Ala-Ala-Gly-Ser-Thr-Met-Gly-Ala-Arg-Ser) with a purity of >95% was purchased from USA MyBioSource, Inc. Prion protein segment PrP106-126 (sequence: Lys-Thr-Asn-Met-Lys-His-MetAla-Gly-Ala-Ala-Ala-Ala-Gly-Ala-Val-Val-Gly-Gly-Leu-Gly) with a purity of >98% was ordered from Sigma-Aldrich. PrP118135 (sequence: Ala-Gly-Ala-Val-Val-Gly-Gly-Leu-Gly-Gly- TyrMet-Leu- Gly- Ser-Ala-Met-Ser) with a purity of > 98% was purchased from Shanghai Apeptide Co., Ltd. The lipid of 1palmitoyl-2-oleoyl- sn-glycero- 3-phospho- (1'-rac-glycerol) (sodium salt) (POPG) and 1,2-dipalmitoyl -sn- glycero- 3phospho-(1'-rac-glycerol) (sodium salt) (DPPG) were purchased from Avanti Polar Lipids (Alabaster, AL). 1,1,1,3,3,3-Hexafluoro2-propanol (HFIP) was ordered from Aldrich with a purity of >99.8%. Right-angle CaF2 prisms were purchased from Chengdu Ya Si Optoelectronics Co., Ltd (Cheng Du, China). All of the chemicals were used as received. PrP118-135 and FP23 were dissolved in HFIP and stored at 4 °C while other peptides were dissolved in methanol (purchased from Sinopharm Chemical Reagent Co., Ltd.) and stored at -20 °C. The phospholipids were dissolved in chloroform and methanol (purchased from Sinopharm Chemical Reagent Co., Ltd.) (with a volume ratio of 65:35) at a concentration of 1.0 mg/mL and kept at -20 °C. The concentration of hIAPP, PrP106126, PrP118-135, and FP23 was 0.1, 2.0, 4.0, and 1.5 mg/mL, respectively. The stock solution was not frozen prior to our experiments and no buffer was used both in the subphase and in the stock solution. The pH of the subphase was approximately 6.2 during the whole experiments because of the absorption of carbon dioxide from the air. The prism-cleaning and lipid monolayer/bilayer preparations were performed using a standard procedure given in supporting information. SFG-VS Experiments Details regarding SFG theories and instruments have been reported previously.24-30 The SFG spectra in Figure 1, Figure 3, Figure 4, Figure S3, Figure S7-S9, Figure S10A-S10B, Figure S12 and Figure S13 were collected using a newly developed highly-sensitive femtosecond SFG-VS system that was introduced in detail in the references.49,50 Briefly, the femtosecond SFG setup was constructed based on a high power regeneration amplifier (Spectra Physics, Spitfire Ace seeded by Mai-Tai SP), which offers 5.0 W output at 800 nm with 13 nm bandwidth, 100 fs pulse duration and 1 kHz repetition rate. A fraction of a pulse (2.0 W) of the output was used for the excitation of a commercial optical parametric amplifier (Spectra Physics, TOPAS Prime) and collinear difference frequency generation system (DFG, AgGaS2 crystal) to generate tunable broadband infrared pulses (2500nm~15000nm) for the SFG probe. A fraction of the output (1.0 W) was passed through a home-built 4-f pulse shaping system to produce a narrowband visible light with a bandwidth of 5 cm-1 for the SFG Vis probe. The broadband IR pulse and narrowband visible pulse were focused by using off-axis parabolic bare gold mirror (FL=150 mm) and plano-convex lens(FL=750 mm) with beam diameters at the sample surface of approximately 200 m and 250 m, respectively. The incident angles of IR and visible pulses were 45° and 60° relative to the surface normal,
Specially, the β-sheet structure is active in the amide I spectra with chiral-sensitive polarization measurements (such as psp and spp) whereas the -helical or unordered structures are both silent.27,28 SFG-VS is therefore an ideal method to characterize the transition between an -helix and a -sheet at the interface.27 Unfortunately, different -sheet structural types (including hairpin-like monomers, β-sheet oligomers, and fibrils) all contribute to chiral amide I signals. Here, -hairpin is the simplest -sheet structure and is comprised of two antiparallel hydrogenbonded -strands connected by a -turn or short loop.37 Accordingly, it is impossible to discriminate the intermediates such as -hairpin-like monomers and -sheet oligomers from the different -sheet structural types by analyzing the chiral amide I spectra alone. This problem in the characterization of -sheet structures can in principle be resolved when the amide II band (between 1500 and 1600 cm-1) is also considered. Amide II vibrations arise from the out-of-phase combination of the C-N stretch and the N-H in-plane deformation.38 They are either very weak or not observable at all in the conventional non-resonant Raman spectra.39 Because SFG signals rely on a tensor product of the IR transition dipole moment and the Raman polarizability tensor,24-36 the amide II SFG signals are too weak to be detected in the disordered, -helical structure, and -hairpin-like structure. Indeed, with the exception of one example of an amide II signal reported in intermolecular -sheet structure by Yan et al.,40 the amide II SFG spectra have not been applied to address the issue of amyloid formation when the SFG technique was applied to investigate the interfacial proteins.17-20,26-36 However, some recent studies have shown that upon formation of multistranded -sheet structures, long-range vibrational coupling from intermolecular sheet contacts can largely enhance the amide II signals,41-43 and make the amide II Raman spectra become visible,43-48 see Table S1. For example, the amide II Raman signals are absent in the disordered α- synuclein and appear when α- synuclein starts to aggregate to form oligomeric -sheet structures.44,45 It can be naturally anticipated that amide II SFG signals are active in sheet oligomers and fibrils. Thus, the combination of chiral amide I and achiral amide II spectral signals allows us to distinguish hairpin like monomers from other -sheet structural types and to clarify when -sheet oligomers start to form. On the other hand, although SFG-VS has been employed to investigate the structure of proteins and peptides at different interfaces,26-28 it often took more than ten minutes to acquire one spectrum.17 The long acquisition time may fail to capture the early formed intermediate species of amyloid formation. Furthermore, the reported experimental procedures only allow to collect the spectra with a polarization combination (for example, ssp) at a time,17, 26-28 which prevent to simultaneously measure the chiral and achiral spectra. Herein, we developed a highly-sensitive femtosecond SFG-VS system by employing several key technical improvements, including adopting a near-total-internal-reflection geometry (Figure S1) and utilizing two Glan-Laser polarizers (with an extinction ratio (Tp/Ts) of >100,000:1) to separate the polarization components (Figure S2). With these improvements, this SFG system is capable of acquiring the ssp and psp spectra simultaneously with a recording time of < 5 seconds.49 This fast recording time can effectively extract the specific structural details about intermediates with both high structural and fast temporal resolutions. Experimental Section Materials and Sample Preparation
2 ACS Paragon Plus Environment
Page 2 of 9
mode of parallel -sheet structures, respectively.17, 27,60 In contrast, the amide I signals in the ssp spectra decreased with time, indicating that hIAPP may undergo a structural transition from α-helix, loop, coil and turn-like structures into β-sheet structures. A new peak at 1540 cm-1 started to appear at t 20 min in the ssp spectra, and its intensity increased with time. The 1540 cm-1 peak is attributed to the amide II band.38,43-48 Earlier nonresonant Raman studies indicated that the appearance of amide II signals is associated with the formation of -sheet oligomers or fibrils.43-48 Therefore, Figure 1 indicated that -sheet structures mainly adopted -hairpin-like monomer at t < 20 min, while the -sheet oligomers and fibrils became detectable at t 20 min. The missing of amide II signals at t < 20 min could result from a very small number of -sheet oligomers and fibrils presented, which were not detectable due to the limited detecting sensitivity. Here, the -sheet structural type with chiral amide I signals but without amide II signals was assigned to the -hairpin-like monomer. This result was also confirmed by the SEM image of the aggregation at different times (Figure S4). SEM image at the aggregation time of t= 10 minutes looked similar to the image of pure DPPG/DPPG bilayer and no fibril was detected. However, at t = 20 min, few fibril was observed. At t = 60 min and 120 min, the image shown clear fibrils and the fibrils became more condensed as the aggregation time increased. The change trend in amyloid formation agreed well with the SFG amide II results.
respectively. The SFG signal was sent to the polarization control module (Figure S2), which can separate S and P polarization signals to different heights on the ICCD. The energy profiles of the IR pulses were used to normalize the SFG spectra by measuring the SFG signals from the gold surface coated at the prism. The SFG spectra were collected immediately after injecting 10L hIAPP (0.1mg/mL) solution into the subphase (2.0 mL) of DPPG bilayer at t = 0 min. For the spectra in the amide I, amide II, and amide A regions, the acquisition time was 1 min. This means that the spectrum at N min was acquired from t= N-1 min to t= N min. For example, the spectrum at 1 min shown in Figure 1 was acquired from t= 0 min to t= 1 min. Because our SFG system was capable of acquiring the ssp and psp spectra simultaneously, one ssp spectrum and one psp spectrum in the amide I and amide II regions were collected at each minute. The 300 ssp spectra and 300 psp spectra were collected continuously for 5 hours. Here, the real time was defined as the actual time (with a time resolution of one minute) during which the interaction between hIAPP and lipid bilayers took place. For the spectra in the amide III region, the acquisition time was 2 min. The spectrum in the amide III region at N min was acquired from t= N-2 min to t= N min. The SFG spectra in Figure 2, Figure S5 and Figure S10CS10E were collected using a picosecond narrow-band frequency scanning SFG system, which is similar to that described in our earlier publication.51 All SFG spectra were averaged over 100 times at each point and normalized by the intensities of the input IR and visible beams. All SFG experiments were carried out at room temperature (24 C). IR beams were protected by a homebuilt chamber purged with dry gas (dry gas generator, Peak Scientific) to avoid IR energy loss due to water vapor absorptions.
A)
Results and Discussion The spectral evolution of hIAPP at the lipid membrane Figure 1 shows the psp and ssp spectra at some typical time interval after injecting 10 L hIAPP (0.1mg/mL) into the subphase (2.0 mL) of the DPPG lipid bilayer at t= 0. The spectra over the whole time are given in Figure S3. Upon adding hIAPP, two strong peaks at approximately 1660 and 1680 cm-1 were observed immediately at 1 min in the ssp spectra and a weak peak at 1640 cm-1 was detected in the psp spectra (Figure 1, 1min). Experiments and theoretical calculations both suggested that turnor bend-like structures show two amide I peaks near 1660-1670 cm-1 and 1680-1690 cm-1, respectively.52-57 In addition, the 1660 cm-1 peak can also be contributed by -helix, loop or coil structures.38 The chiral psp signal implied the formation of a sheet. Earlier IR studies indicated that the frequency of the looser, smaller and more disordered β-sheet structural type (for example, -hairpin-like monomer) extends from 1630 cm-1 to 1643 cm-1, while the longer, larger, and more rigid -sheets (for example, βsheet oligomers and fibrils) absorb below 1630 cm-1.38,58,59 The immediate observation of strong ~1660 and ~1680 cm-1 peaks in ssp spectra and a weak peak in the psp spectra indicated that hIAPP may predominantly adopt an α-helical, loop, coil and turnlike structures but with small β-sheet structures during the initial interaction between the hIAPP and the DPPG lipid bilayer. With increasing time, the intensity of the chiral psp peak increased (Figure 1B) and the frequency shifted to 1624 cm-1, implying that longer and larger -sheet structures were formed. Consistent with previous reports,17, 27,60 the amide I signals in the psp spectra exhibited one strong peak at 1624 cm-1 and a shoulder peak centered at 1670 cm-1, which corresponded to the B mode and A
10
0.8 0.4 0.0 0.4 0.0 0.3 0.0 0.2
10
15min
10
20min
5
50min
0.0 0.2
300min
0.0 1500
1600
B)
1min
1700
1800 -1
Wavenumber(cm )
0.01
SFG Intensity
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Journal of the American Chemical Society
SFG Intensity
Page 3 of 9
5
1min
0.00 3 15min 0.01 0.00 3 20min 0.01 0.00 50min 0.01 0.00 0.04 300min 0.02 0.00 1550 1600 1650 1700 1750 -1
Wavenumber(cm )
Figure 1. The evolution of some typical SFG spectra after injecting 10L hIAPP (0.1mg/mL) into the subphase (2.0 mL) of DPPG bilayer. A) ssp amide I and amide II; B) psp amide I. Further confirming the relevance of amide II SFG signals and the formation of -sheet oligomers and fibrils The relevance of amide II signals with the formation of -sheet oligomers and fibrils can be further confirmed by the aggregation of the fragments of prion protein (PrP) and glycoprotein 41 (GP41) at the lipid bilayer surface. Although the structures of fulllength PrP and GP41 are very complex and largely unknown, the PrP fragment [106-126] (PrP106-126), PrP fragment [118-135] (PrP118-135), and fragment of GP41 (FP23) have been extensively studied in membrane-mimicking environments. It is evident that PrP106-126 forms oligomers and fibrils at the membrane, while PrP118-135 does not.61-65 Figure 2 shows the psp and ssp SFG spectra collected at t= 5 h after injecting 6 L PrP106-126 (2 mg/mL) and 50L PrP118-135 (4 mg/mL) solutions into the subphase (2.0 mL) of the POPG bilayer. The chiral psp spectra showed one resonant peak at 1630 cm-1 (Figure 2A), denoting the formation of a -sheet structure in both PrP106126 and PrP118-135. However, the achiral ssp spectra of PrP106126 and PrP118-135 (Figure 2B) were quite different. The ssp
3 ACS Paragon Plus Environment
Journal of the American Chemical Society
the loop and -helical structures were well separated in the amide III spectra which appeared below and above 1260 cm-1, respectively.30 In contrast, the amide III signals of coil and turnlike structures were very weak. Figure 3A shows the amide III ssp spectra after injecting 10L hIAPP (0.1mg/mL) into the subphase of the DPPG bilayer. The amide III signals were very weak at the beginning probably because hIAPP initially adopted coil and turnlike structures, as confirmed by ATR-FTIR spectra (Figure S6). With increasing time, the intensity increased and the spectra were dominated by the peaks at 1220 and 1240 cm-1, which were assigned to -sheet or loop structures. No signals above 1260 cm-1 were observed, indicating that no -helical structures formed.30,6973 Therefore, the 1660 cm-1 peak arose from coil and turn-like structures, excluding the generation of transiently -helical intermediates during the IAPP oligomerization. It needs to be noted that the absence of the signals above 1260 cm-1 in Figure 3A is not due to the lack of detection sensitivity or because the peptide adopted an unfavorable orientation. In Figure 1A, upon adding hIAPP, a very strong peak at 1660 cm-1 was observed immediately at 1 min in the ssp spectra. If the 1660 cm-1 peak arose from -helical structure, a characteristic peak would be detected at approximately 1290 cm-1 in the amide III band spectrum. However, this expectation is opposite to what we have observed. Earlier studies indicated that for the -helical structure, even if the amide I signal is much weaker than the intensity of the 1660 cm-1 peak (as shown in Figure 1A at 1 min), the amide III signals can still be detected.30,69-73 For example, in Figure 3B, even if the amide I signal of GP41 in the POPG bilayer at 1660 cm-1 is 5 times weaker than that of hIAPP in the DPPG bilayer, the amide III signal at 1290 cm-1 can be clearly detected for GP41; this is because the GP41 aggregation involves -helical intermediates.74 In addition, this detected signal is also supported by the controlled experiments on the interaction between rat-IAPP and the DPPG lipid bilayer. Rat-IAPP has a lower ssp amide I intensity, but its amide III spectra are dominated by three peaks at ~1230, ~1280, and ~1315 cm-1 (Figure S8). The peaks at ~1230 and ~1280 cm-1 were assigned to loop and -helical structures,30, 69-73 respectively, indicating that rat-IAPP adopts -helical and loop structures at the DPPG lipid bilayer interface, which is consistent with the NMR results.75 Furthermore, the absence of amide III signals above 1260 cm-1 in Figure 3A is not because the peptide adopts an unfavorable orientation. Recent studies have found that amide III signals are detectable for -helical peptides with different orientations.30,69-73 For example, the peptides of GALA that either insert into or lie down on the lipid bilayer surface both show detectable amide III spectra.72 Fujii and coworkers also found that -helical keratin proteins show clear strong amide III signals even if the amide I signals almost disappear when the proteins lie down on the surface.73 In terms of these analyses, we can safely conclude that the strong peak at 1660 cm-1 is not contributed to the -helical structure. This result illustrates that hIAPP populates through the -sheet conformers, rather than the -helical intermediates. Indeed, the studies based on 2D IR and theoretical simulations also suggest that the early sheet oligomers are assembled from the -hairpin motif.3-7
spectra of PrP106-126 were dominated by the peaks at 1540, 1620, 1680 and 1730 cm-1. The 1540 cm-1 peak belongs to the amide II band. It was proven that the amide II band in the Raman optical activity (ROA) spectra of prion protein was contributed to the formation of multistranded -sheet structures (including oligomers and fibrils).43, 47,48 The 1620 cm-1 and 1680 cm-1 peaks arose from β-sheet and turn-like structures, respectively.27, 38 In contrast, the ssp spectra of PrP118-135 were dominated by two peaks at 1660 cm-1 (from α-helix, loop, or coil structures) and 1720 cm-1. Although the amount of PrP118-135 was more than 15 times larger than that of PrP106-126, no amide II signals were detected at 1510-1580 cm-1 for PrP118-135. This result exactly matched the reported structures of PrP106-126 and PrP118-135 at lipid membrane (Figure 2C). 61-65 Previous studies using NMR and ion mobility spectrometry-mass spectrometry revealed that PrP106-126 forms large -sheet oligomers and fibrils while PrP118-135 does not.61-65 This phenomenon was also found in the membrane-bound FP23 (Figure S5). A strong resonance peak centered at 1625 cm-1 was observed in the psp spectra as different amounts of FP23 solutions (1.5 mg/mL) were injected into the subphase of the POPG bilayers, indicating the formation of a βsheet structure.27 However, the amide II signal (1565 cm-1) was absent in the low concentration case (2 L), but appeared and became more intense as the FP23 amount increased. The appearance of the amide II signals indicated the formation of sheet oligomers and fibrils. This finding was highly consistent with the results obtained by the NMR measurements, which suggested that a larger dose of FP23 could lead to its aggregation by forming β-sheet oligomers and fibrils in aqueous and membrane lipid environments.66-68 All these results support that the amide II signals are sensitive to the formation of -sheet oligomers and fibrils and thus provide a molecular probe for the interfacial intermediates of aggregation processes. Therefore, the appearance of amide II signals in Figure 1A indicated the membrane-associated hIAPP -sheet oligomers and fibrils become detectable at t20 min and proceed by assembling with a -hairpin-like monomer on the bilayer,4-6 instead of inserting preassembled oligomers in solution.11 A) 0.3 PrP106-126 0.2 0.1 0.0 PrP118-135 0.1 0.0 1500
1600
SFG Intensity
SFG Intensity
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
1700
-1
Wavenumber(cm )
Page 4 of 9
B) 0.06 PrP106-126 0.03 0.00 PrP118-135 0.12 0.06 0.00 1500 1600 1700 1800 -1
Wavenumber(cm )
Figure 2. The SFG spectra at t= 5 h after injecting 6 L PrP106126 (2 mg/mL) and 50L PrP118-135 (4 mg/mL) solution into the subphase (2.0 mL) of POPG bilayer. A) chiral psp; B) achiral ssp. C) Scheme of PrP106-126 and PrP118-135 structures. The possibility of formation -helical intermediates It has been proposed that hIAPP oligomerization is initiated via the transiently populated -helical intermediates or the formation of -strand rich dimers. In Figure 1A, a very strong peak at 1660 cm-1 was observed in the ssp spectra at 1 min. However, all of the -helix, loop, coil and turn-like structures may contribute to this peak because their frequencies overlap at 1655~1660 cm-1. To discern them, we further measured the amide III signals. It has been verified that the SFG amide III signals are capable of directly distinguishing -helix, loop, coil and turn-like structures in proteins at interfaces.30,69-73 The spectral profiles of
4 ACS Paragon Plus Environment
Page 5 of 9
B)
300 min 100 min 50 min 10 min 2 min
1
0 1150 1200 1250 1300
-1
Wavenumber(cm )
1.0
GP41 hIAPP
0.5 0.0 1500 2.0 1.5 1.0 0.5 0.0 1150
1600
1700
GP41 hIAPP
1200
SFG Intensity
2
A)
SFG Intensity
3
SFG Intensity
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Journal of the American Chemical Society
1800
-helical
structure
1250
1300 -1
Wavenumber(cm )
0.4 2min 0.2 0.0 15min 0.2 0.0 20min 0.2 0.0 50min 0.2 0.0 0.6 300min 0.3 0.0 3100 3200 3300 3400 3500 -1
Wavenumber(cm )
Figure 3. A) The evolution of ssp amide III SFG spectra after injecting 10L hIAPP (0.1mg/mL) into the subphase (2.0 mL) of DPPG bilayer. B) The ssp amide I and amide III spectra of the hIAPP in DPPG bilayer (1 min case) and GP41 in POPG bilayer.
Figure 4. The evolution of chiral psp N-H SFG spectra after injecting 10L hIAPP (0.1mg/mL) into the subphase (2.0 mL) of DPPG bilayer.
It is worth mentioning that the Yan group reported that the protein secondary structures at interfaces could be distinguished by combining the chiral N−H stretch SFG signals of protein backbones with the chiral amide I signals.17,27 They found that a parallel β-sheet exhibited a distinct chiral signal only for the amide I mode, an α-helix exhibited a signal only for the N-H stretch mode, and an antiparallel β-sheet exhibited signals for both modes.27 Using these empirical vibrational signatures, they monitored the aggregation of hIAPP at a lipid/water interface and observed an intermediate with the chiral N−H signal but without the chiral amide I signal. They assigned that intermediate to helix.17 Later on, they found that hIAPP aggregates exhibited a significantly different chiral optical response: the chiral N-H signal was absent at the lipid/water interface, while it appeared on a glass slide. They interpreted that this difference was because the chiral N-H signal is highly dependent on the orientation of the parallel β-sheet.19 Being consistent with the new finding of the Yan group,19 we observed both chiral N-H and chiral amide I signals. The chiral N-H peak showed a peak center near 3285 cm-1 (Figure 4 and Figure S9), near the common frequency of many βsheet structures (Table S3). The intensity of chiral N-H peak increased with time and showed similar time-dependent behavior with the chiral amide I signals (Figure S9), indicating that the chiral N-H signals were originated from the β-sheet structure, rather than the -helical structure. This conclusion was further supported by the achiral and chiral spectra of a series of lipid membrane-bound peptides including WALP23, rat-IAPP, melittin, pardaxin and LK14 (Figure S10). These peptides all adopted predominantly α-helical and loop structures without βsheet structures in lipid bilayer environments.30, 75-77 It was evident that no detectable chiral N-H and chiral amide I signals were observed in these -helical peptides under current experimental conditions.
The structural evolution of hIAPP at the lipid membrane To quantitatively analyze the change in the intensities of the amide I and amide II spectra, we fitted the psp and ssp spectra (Figure S3) using Eq.S1. We defined the effective peak strength ( (2) ) as the fitting strength A normalized by . The effective peak strength( (2) ) of the ssp amide I signals (Figure 5A, from the coil and turn-like structures) decreased with time, while the psp amide I signal and the ssp amide II signal (Figure 5B, from the -sheet structure) increased with time. The time-dependent (2) change of each peak all exhibited two kinetic processes: a fast process at t
