Post-translational modifications are often detected in age-related diseases associated with protein misfolding such as cataracts from aged lenses. One of the major post-translational modifications is the isomerization of aspartate residues (L-isoAsp), which could be non-enzymatically and spontaneously occurring in proteins, resulting in various effects on the structure and function of proteins including short peptides. We have reported that the structure and function of an αA66–80 peptide, corresponding to the 66–80 (66SDRDKFVIFLDVKHF80) fragment of human lens αA-crystallin, was dramatically altered by the isomerization of aspartate residue (Asp) at position 76. In the current study, we observed amyloid-like fibrils of L-isoAsp containing αA66–80 using electron microscopy. The contribution of each amino acid for the peptide structure was further evaluated by circular dichroism (CD), bis-ANS, and thioflavin T fluorescence using 14 alanine substituents of αA66–80, including L-isoAsp at position 76. CD of 14 alanine substituents demonstrated random coiled structures except for the substituents of positively charged residues. Bis-ANS fluorescence of peptide with substitution of hydrophobic residue with alanine revealed decreased hydrophobicity of the peptide. Thioflavin T fluorescence also showed that the hydrophobicity around Asp76 of the peptide is important for the formation of amyloid-like fibrils. One of the substitutes, H79A (SDRDKFVIFL(L-isoD)VKAF) demonstrated an exact β-sheet structure in CD and highly increased Thioflavin T fluorescence. This phenomenon was inhibited by the addition of protein-L-isoaspartate O-methyltransferase (PIMT), which is an enzyme that changes L-isoAsp into Asp. These interactions were observed even after the formation of amyloid-like fibrils. Thus, isomerization of Asp in peptide is key to form fibrils of αA-crystallin-derived peptide, and L-isoAsp on fibrils can be a candidate for disassembling amyloid-like fibrils of αA-crystallin-derived peptides.
Technical progress and improved resolution of mass spectrometry (MS) has made it possible to trace tiny differences in chemically modified amino acid residues in proteins. Oxidation, phosphorylation, deamidation, and isomerization of amino acid residues have been identified and quantified by MS [[
Graph: Fig 1 Scheme for isomerization of Asp residues via L-succinimidyl reaction intermediate and repair mechanism of PIMT in protein tertiary structure.(A) Formation of L-Asp isomerization via a five membered-succinimide intermediate in proteins and peptides. (B) L-isoAsp can be converted into L-Asp by the enzyme, protein-L-isoaspartate O-methyltransferase (PIMT), through methyl esterification in the presence of SAM. The methyl ester undergoes spontaneous demethylation to form a L-succinimide intermediate. L-succinimide is hydrolyzed to L-isoAsp and L-Asp. The L-isoAsp product can render the methylation/demethylation cycle. SAM; S-(5′-adenosyl)-L-methionine and SAH; S-(5′-adenosyl)-L-homocysteine.
The predominant L-isoAsp was identified in lens tissues from age-related cataract patients [[
We have previously reported the contribution of the formation of L-isoAsp76 to the anti-chaperone function and secondary structure of αA66–80 [[
All 9-fluorenylmethyloxycarbonyl (Fmoc)-labeled amino acids and resin-bound amino acids were purchased from Watanabe Kagaku (Hiroshima, Japan). Piperidine for eliminating Fmoc during peptide synthesis was purchased from Tokyo Chemical Industry (Tokyo, Japan). N-Methylmorpholine and 1, 2-ethanedithiol were purchased from Nacalai Tesque (Kyoto, Japan). HPLC-grade acetonitrile, N, N-dimethylformamide, diethyl ether, trifluoroacetic acid, thioflavin T, and 4,4′-dianilino-1,1′-binaphthyl-5,5′-disulfonic acid (Bis-ANS) were purchased from Sigma Aldrich (St. Louis, MO, USA). Recombinant human protein-L-isoaspartate O-methyltransferase was purchased from NKMAX (Seongnam-si, Republic of Korea). All other analytical chemicals were purchased from FUJIFILM Wako Pure Chemical Industries (Osaka, Japan).
Alanine substituents of the αA66–80 peptide, corresponding to 66–80 (
The iso-Asp residue in αA66–80 was confirmed by PIMT treatment and nanoscale reverse phase high performance liquid chromatography (Paradigm MS4, Michrom Bioresources, Auburn, CA) using a C18 column (L-column, 0.1 × 150 mm; Chemicals Evaluation and Research Institute, Tokyo, Japan). The αA66–80 (L-Asp76) and αA66–80 (L-isoAsp76) solutions were prepared at concentrations of 0.1 mg/mL. PIMT (0.02 mg/mL) was added to αA66–80 (L-isoAsp76), and then incubated for 24 h at 37°C. Each peptide (5 μg) was injected into the column and its elution profile was observed. A linear gradient of 5%–45% acetonitrile containing 0.1% formic acid was applied at a flow rate of 0.5 μL/min over 60 min. MS and MS/MS scan data were collected in positive mode with an ion trap system (LCQ Fleet, Thermo Fisher Scientific Inc., USA).
The sample solution was prepared as previously described [[
Far-UV circular dichroism (CD) was performed using a J-805 instrument (JASCO, Tokyo, Japan). Each αA66–80 peptide (0.1 mg) was dissolved in 1 mL of 10 mM phosphate buffer (pH 7.4). The solution was then diluted with distilled water to 16.6 mM phosphate. CD spectra were acquired in the range of 195–250 nm using a 1 mm path cylindrical quartz cell. All spectra were baseline corrected by subtracting the buffer-only spectrum. Each reported spectrum represents an average of five scans.
Bis-ANS assay was carried out using a Hitachi F-4500 fluorescence spectrophotometer with excitation at 360 nm and emission at 400–600 nm (Hitachi, Tokyo, Japan). A 20 μM Bis-ANS solution was prepared in 10 mM Tris buffer (pH 7.4). Next, 25 μL of each αA66–80 peptide solution (1.0 mg/mL) was added to 225 μL of Bis-ANS solution. After incubation for 10 s at room temperature, Bis-ANS fluorescence was measured. Each assay was conducted three times.
Thioflavin T assay was carried out using a Hitachi F-4500 fluorescence spectrophotometer with excitation at 450 nm and emission at 460–600 nm (Hitachi, Tokyo, Japan). A 50 μM Thioflavin T (ThT) solution was prepared in 10 mM phosphate buffer (pH 7.4). Next, 20 μL of each αA66–80 peptide solution (0.3 mg/mL) was added to 250 μL of ThT solution. After incubation for 60 s at room temperature, ThT fluorescence was measured. Each assay was conducted five times.
The protein-L-isoaspartate O-methyltransferase (PIMT) activity for the peptide was assessed by measuring the intensity of the ThT fluorescence signal using a Hitachi F-4500 fluorescence spectrophotometer with excitation at 450 nm and emission at 460–600 nm. ThT solution (50 μM) was prepared in 10 mM Tris buffer (pH 7.8). Each peptide (0.01 mg/mL) was incubated in 10 mM Tris buffer (pH 7.8) at 37°C for 24 h with PIMT at peptide to PIMT ratios of 1:1, 1:2, and 1:5 (w/w). The amount of S-(5′-adenosyl)-L-methionine (SAM) was the same as that of PIMT. Fifty microliter of the sample solution was mixed with 150 μL of ThT solution. After incubation for 60 s at room temperature, ThT fluorescence was measured. Each ThT assay was conducted five times.
Fig 2A shows the sequence of αA66–80 with the position of Asp76 in human lens αA-crystallin. It is located on the predicted β-sheet region of homologous bovine αA-crystallin (Fig 2B). This synthetic peptide containing Asp at the 76
Graph: Fig 2 L-isoAsp 76 containing human αA-crystallin 66–80 forms amyloid like fibrils.(A) αA66–80 peptide sequence. The Asp residues at 76 position in the sequence was replaced with the L-isoAsp residue. (B) The 3D structure of bovine αA-crystallin (PDB: 3L1F). The αA66–80 region of human αA-crystallin sequence is colored red in the structure. (C) Fluorescence spectra of ThT in the presence of L-Asp in the peptide after heat incubation (blue) and the amyloid fibrils formed from L-isoAsp-containing peptide (red). (D) TEM images of negatively stained fibrils of L-isoAsp containing αA66–80 after heat incubation. (E) TEM images of negatively stained fibrils of L-Asp containing αA66–80 after heat incubation.
To screen the specific interactions between L-isoAsp and other amino acids, individual alanine substitution was performed for each residue of the peptide. The alanine substituents of αA66–80 peptide, including L-isoAsp76, were synthesized and purified in the same manner as αA66–80. The sequences of the substituents are shown in S1 Table. To confirm whether each peptide was synthesized correctly, analytical LC-MS was performed. The theoretical and actual masses of the protonated precursor ions of the synthetic peptides are shown in S1 Table. The purity of the synthetic peptides was confirmed to be >95%. Fig 3 shows the CD spectra of the alanine substituents of αA66–80 (L-isoAsp76). In CD spectra of R68A, K70A, K78A, and H79A substituents of αA66–80 (L-isoAsp76), we observed that the β-sheet structures were retained. However, the other alanine substituents of αA66–80 peptides did not form a β-sheet structure. These results indicated that the positively charged amino acids were not important for forming the β-sheet structure of αA66–80 (L-isoAsp76).
Graph: Fig 3 Far-UV CD spectra of the alanine substituents of αA66–80 (L-isoAsp76) peptides.Black line shows the far-UV CD spectrum of the control αA66–80 (L-isoAsp76: Ctl) peptide. (A) Blue, red, orange, purple, green, yellow, and brown lines show the far-UV CD spectra of S66A, D67A, R68A, D69A, K70A, F71A, and V72A substituents, respectively. (B) Blue, red, orange, purple, green, yellow, and brown lines show the far-UV CD spectra of I73A, F74A, L75A, V77A, K78A, H79A, and F80A substituents, respectively. The spectra were recorded between 195 and 250 nm. Each spectrum represents an average of five scans.
Bis-ANS shows the increasing fluorescence intensity when binding to multiple sites through hydrophobic interactions. Therefore, the change of hydrophobicity of a peptide can be estimated by measuring the intensity of Bis-ANS fluorescence. In this experiment, the R68A, K70A, K78A, and H79A substituents of αA66–80 (L-isoAsp76) did not decrease the Bis-ANS fluorescence intensity much (Fig 4). The other alanine substituents of αA66–80 (L-isoAsp76) showed a reduction in the Bis-ANS fluorescence intensity, suggesting that alanine substitution except for positively charged amino acids decreases the hydrophobicity of αA66–80 (L-isoAsp76). The results were consistent with the CD spectra, suggesting that the change in secondary structure is caused by the change in hydrophobicity.
Graph: Fig 4 Emission spectra of Bis-ANS solution in the presence of the alanine substituents of αA66–80 peptides.Black lines show the emission spectrum of the Bis-ANS solution in the presence of normal αA66–80 (L-isoAsp76: Ctl). (A) Blue, red, orange, purple, green, yellow, and brown lines show the emission spectra of the Bis-ANS solution in the presence of S66A, D67A, R68A, D69A, K70A, F71A, and V72A substituents, respectively. (B) Blue, red, orange, purple, green, yellow, and brown lines show the emission spectra of the Bis-ANS solution in the presence of I73A, F74A, L75A, V77A, K78A, H79A, and F80A substituents, respectively. Each spectrum represents an average of three scans.
Thioflavin T (ThT) shows specific binding to the amyloid state of the protein, and the change in ThT fluorescence intensity could be used to characterize the formation of amyloid fibrils. In this study, the ThT assay was carried out to determine whether alanine substituents of αA66–80 (L-isoAsp76) could form amyloid-like fibrils (Fig 5). Under the current experimental conditions, the R68A, K70A, and K78A substituents of αA66–80 (L-isoAsp76) did not alter the fluorescence intensity as αA66–80 (L-isoAsp76). The H79A substituent of αA66–80 (L-isoAsp76) showed highly intense fluorescence emission, suggesting that the H79A substituent of αA66–80 (L-isoAsp76) is more likely to form amyloid-like fibrils (Fig 5B). The H79A substituent of αA66–80 (L-Asp76) showed a little increase in the ThT fluorescence intensity (S2 Fig), indicating that it can form small amounts of amyloid-like fibrils. These results also suggest that the alanine substitution for histidine residue promotes the formation of amyloid-like fibrils of αA66–80 (L-isoAsp76) rather than of αA66–80 (L-Asp76). The negatively charged substituents, D67A and D69A, showed weak fluorescence similar to that of substituents of S66A of αA66–80 (L-isoAsp76) (Fig 5A). The hydrophobic residues decreased the fluorescence emission intensity of ThT at a wavelength near 480 nm. These results showed that R68A, K70A, K78A, and H79A mutant αA66–80 (L-isoAsp76) peptide generated amyloid-like fibrils, consistent with the results of CD analysis.
Graph: Fig 5 Emission spectra of the Thioflavin T (ThT) solution in the presence of the alanine substituents of αA66–80 peptides.Black lines show the emission spectrum of the ThT solution in the presence of normal αA66–80 (L-isoAsp76: Ctl). (A) Blue, red, orange, purple, green, yellow, and brown lines show the emission spectra of the ThT solution in the presence of S66A, D67A, R68A, D69A, K70A, F71A, and V72A substituents, respectively. (B) Blue, red, orange, purple, green, yellow, and brown lines show the emission spectra of the ThT solution in the presence of I73A, F74A, L75A, V77A, K78A, H79A, and F80A substituents, respectively. Each spectrum represents an average of five scans.
The fluorescence intensity of ThT when only using αA66–80 (L-isoAsp76) was too small to evaluate the PIMT activity. Therefore, we used the abovementioned H79A substituent of αA66–80 (L-isoAsp76), which forms large amounts of amyloid-like fibrils. After heat-induced amyloid formation of H79A peptide, excess amounts of PIMT and SAM were added to induce the backward reaction from L-isoAsp to L-Asp on amyloid-like fibrils. As shown in Fig 6A, the fluorescence emission intensities of ThT decreased according to the concentration of PIMT. TEM image supported the decreased formation of amyloid-like fibrils of the H79A substituent of αA66–80 (L-isoAsp76) after PIMT treatment (Fig 6B and 6C). The results indicated that PIMT could act on L-isoAsp on amyloid-like fibrils and convert it into Asp, thereby, reforming the non-amyloid-like structure of the peptide.
Graph: Fig 6 PIMT decreased the formation of amyloid-like fibrils of H79A alanine substituent of αA66–80 peptides.(A) Repair activity of PIMT was monitored as the intensity of ThT fluorescence. The mass ratio of H79A to PIMT was 1:0 (blue), 1:1 (red), 1:2 (green), and 1:5 (purple). Black lines show the emission spectrum of the ThT solution in the absence of peptide. Each spectrum represents an average of five scans. (B) TEM images of negatively stained fibrils of L-isoAsp containing H79A αA66–80 after heat incubation in the absence of PIMT. (C) TEM images of negatively stained fibrils of L-isoAsp containing H79A αA66–80 after heat incubation in the presence of PIMT.
TEM images showed the presence of amyloid-like fibrils (Fig 2D). Short and sharp shapes possibly indicate the seeds of long-length amyloid-like fibril formation, the short shape formed due to low concentrations of peptide (~0.3 mg/mL). Since the morphology of amyloid-like fibrils depends on the concentrations of peptide, an increase in the concentrations of peptide including L-isoAsp corresponded to an increase in extensions observed. However, in TEM images of amyloid-beta 15–28 peptide containing L-isoAsp at the same concentration and pH, we observed shorter structures than those observed in other reports [[
The current data also suggested that the hydrophilicity of peptides is important for fibril formation. The alanine substituent for positively charged residues did not alter the formation of secondary structure, hydrophobicity, and amyloid-like fibrils (Figs 3–5). Although the side chain of L-isoAsp has a negative charge and interacts with positively charged residues, the data indicated that there was no interaction between L-isoAsp and positively charged residues in amyloid-like fibrils. Nevertheless, it has been suggested that the isomerization of Asp residue at position 76 of αA66–80 is essential for the formation of secondary structure and amyloid-like fibrils (S3 Fig). This implied the significance of L-isoAsp at position 76 of this peptide. These also emphasized the significance of the hydrophobic amino acid network for the formation of amyloid-like fibrils. Previously, the region from F71–V77 residues of αA66–80 has been reported as the core region of the amyloid-forming αA66–80 [[
The enzymatic treatment using PIMT decreased amyloid-like fibrils of L-isoAsp containing H79A peptide (Fig 6). This treatment was performed after the formation of amyloid-like fibrils; thus, PIMT could recognize L-isoAsp on the surface of the short amyloid-like fibrils, inhibiting fibril-fibril interaction, or loosening fibrils in each unit. As shown above, presumably, the main chain of L-isoAsp76 works as a part of the hydrophobic network in amyloid-like fibrils, but the exposed side chain of L-isoAsp was recognized by PIMT. Here, the concentration of PIMT was higher than that of the peptide. Thus, bulky PIMT inhibition inhibits further extension of amyloid-like fibrils. It is still unclear whether L-isoAsp76 on the H79A mutant αA66–80 (L-isoAsp76) was converted into L-Asp via a five-ring succinimide intermediate. Further studies are needed to confirm the details of the reaction and generality for other peptides, including L-isoAsp. However, as far as we know, this is the first report that PIMT can function as a degrading enzyme for amyloid-like fibrils in vitro.
In vivo, the presence of peptide αA66–80 and isomerization of Asp76 was confirmed in aged lenses [[
It should be discussed whether the isomerization of Asp76 occurs before or after truncation. Isomerization ratio has been reported to be faster in proteins than in random coiled structures [[
In vivo, the isomerization of Asp76 functions to acquire the chaperone-like activity of αA66–80 (L-isoAsp76) to suppress the aggregation of other lens crystallins. Therefore, the presence of PIMT in lens would be unsuitable. In actuality, the expression levels of PIMT are high in the brain, but low in lens [[
S1 Fig. LC-MS chromatogram of the αA66–80 peptide containing L-isoAsp76.
(A) LC spectrum of peptide, with a molar mass of 623.0, that eluted from αA66–80 (L-isoAsp76). (B) LC spectrum of peptide, with a molar mass of 623.0, that eluted from αA66–80 (L-Asp76). (C) LC-MS chromatogram of the PIMT-treated αA66–80 (L-isoAsp76). There were two peaks in the LC spectrum on (C). The late elution peak had the same elution time as αA66–80 (L-Asp76). The early elution peak did not show the identical elution time from both, but showed 14 Da mass difference from αA66–80 (L-Asp76) and αA66–80 (L-isoAsp76). This mass difference was derived from methyl group during PIMT interaction with αA66–80 (L-isoAsp76). These showed the activity of PIMT for L-isoAsp76 containing peptide in current experiments.
(TIFF)
S2 Fig. Emission spectra of ThT solution in the presence of the H79A αA66–80 (L-Asp76) and H79A αA66–80 (L-isoAsp76) peptides.
(TIFF)
S3 Fig. Comparison of the secondary structure of αA66–80 (L-isoAsp76) and D76A αA66–80 peptides.
(A) Far-UV CD spectra of αA66–80 (L-isoAsp76) (black) and D76A αA66–80 (red). (B) The fluorescence spectra of ThT in the presence of αA66–80 (L-isoAsp76) (black) and D76A αA66–80 (red).
(TIFF)
S1 Table. Molar masses of each alanine substituents.
(TIFF)
By Kosuke Magami; Naomi Hachiya; Kazuo Morikawa; Noriko Fujii and Takumi Takata
Reported by Author; Author; Author; Author; Author