In this study we applied SFTIRM to screen out the protein misfolding in rat hippocampal tissue in response to both Al toxicity and the influence of LS in reversing these protein misfolding as proved earlier3,5,13,22 over the amide I and amide II range [1700-1500cm− 1]
Very slight changes are observed from the raw spectra obtained from Cont, AD, Cur and LS groups; the spectra are nearly superimposed (data not shown). For deeper spectral analysis, spectral second derivative and curve fitting were carried over the selected amide I &II regions from the computed averaged, normalized, smoothed and baseline corrected spectra obtained from each group. Figure 1(a, b) illustrated the second derivatives of Cont, AD, Cur and LS groups at the two tested time intervals [42d, 65d] respectively. The amide I band arises mainly to carbonyl C = O stretching vibrational modes of proteins, while the amide II band is due to N–H bending and C–N stretching of proteins 34–39. Only the marked changes in the amide II second derivatives bands are noticed in the early stage of the experiment at 42d Fig. (1a).
It seems that, at early stage of treatment [42d], the noticed spectral changes in the hippocampal proteins secondary structure beginning mainly in the NH bending of proteins i.e. in the amide II bands. This may be referred to the fact that, along the process of proteins aggregation, peptide bonds are formed between the NH2 and C = O of the amino acids as they both strongly contribute to this process.
Meanwhile, the most significant spectral changes in band intensities, HBW, bands position, disappearance and appearance of new bands are clearly shown in AD group at 65d rather than the Cont and the other LS treated groups [Cur & LS] for both the amide I and amide II bands ] Fig. (1b). This is logic, because as the time and the process of misfolding proteins and aggregation increase it should be accompanied by a dramatic alteration in the SFTIRM spectra.
The figure revealed a very broad wide amide I band in AD 2nd der. spectrum centred at ~ 1643 cm− 1 with doublet in the amide II band at ~ 1533&1546 cm− 1, that were shifted to lower frequency, and appearance of a new band at1585 cm-1 compared to the Cont 2nd der. spectrum [Fig. (1b)black arrows]. Cur spectrum showed several small bands in the amide II region compared to the AD 2nd der. spectrum [Fig. (1b) black stars]. It is also obvious from Fig. (1b) that the C = O stretching carbonyl bands in the AD group are depleted than the Cont and the other LS tested groups [black arrow]. This observation is in agreement with Catherine et al.,40 who stated the disappearance of the lipid carbonyl band at 1738 cm− 1 in the sFTIR spectrum of brillar Aβ42, aggregated plaque.
Figure 1b. Normalised, mean second derivative SFTIRM spectra from Cont, AD, Curative and LS groups over the amide I and amide II region After 65d, the black arrows pointed to the substantial changes in the AD 2nd der spectrum for both amide I and amide II compared to the control and the other LS tested groups [Cur & LS]. The depletion of the carbonyl C = O stretching bands is observed. The intensity, HBW and the position of the amide modes are both affected by Al and LS treatment.
Curve fitting analysis
For more qualitative and quantitative spectral analysis, curve fitting of the amide I&II region over 1700 − 1500 cm− 1 were carried from the previously protein processed map (Fig. 2). We choose the darkest red hot spectrum/spectra from each map, the resolved amide I and amide II sub-bands, the band position, the HBW and the percentage area are given in Table 1 together with their corresponding band assignments. It is clear from Table 1. That, at 42d of treatment, for the amide I, which is usually used to give all details about the protein secondary structure as C = O stretching vibrational mode is greatly affected by the hydrogen bonds between α-helix, β-sheets, random coil and β-turn structure41,42. Its sub-band around 1640 cm− 1 detected in the Cont peak resolved spectrum/spectra is shifted to higher wavenumber around 1645 − 1642 cm− 1 in the AD group. This band is mainly arise from the C = O stretching mode of vibration of proteins and is due to random coil secondary structure. It should be mentioned here that, the two different spectra taken for the curve fitting from AD [AD1, AD2] group gave two obviously different sub-bands. These is evident by the results obtained from both HCA and PCA [that will be discussed here later]. It seems that at early stage of Al intoxication, we have two different protein secondary structure; one is closely similar to the control while, the other is completely different.
Table 1
The sub-bands positions, HBW and percentage areas, of amide I sub-bands (1700–1500 cm− 1) of the rat brain hippocampal tissue for control, AL, Curative (Cur) and LS treated groups at 42d & 65 days.
42 Day
|
Control
|
AD Group
|
Cur.
|
L.S
|
AD1
|
AD2
|
Center X
|
FWHH
|
% Area
|
Center X
|
FWHH
|
% Area
|
Center X
|
FWHH
|
% Area
|
Center X
|
FWHH
|
% Area
|
Center X
|
FWHH
|
% Area
|
1511.56
|
42.87
|
8.48
|
1511.62
|
43.18
|
7.80
|
1518.93
|
45.56
|
14.54
|
1529.278
|
53.209
|
15.6125
|
1521.736
|
48.855
|
16.7764
|
1535.19
|
44.40
|
17.51
|
1534.79
|
45.46
|
17.51
|
1542.89
|
41.7
|
13.51
|
1557.638
|
54.397
|
10.7216
|
1545.96
|
46.489
|
15.8532
|
1562.69
|
46.72
|
10.70
|
1561.83
|
48.41
|
9.73
|
1569.10
|
41.683
|
6.96
|
1610.975
|
53.076
|
6.4267
|
1571.473
|
44.751
|
5.8666
|
1609.60
|
50.98
|
10.36
|
1621.24
|
58.59
|
17.65
|
1612.83
|
49.655
|
11.43
|
1644.521
|
52.026
|
29.6072
|
1612.969
|
50.314
|
9.3758
|
1640.59
|
51.08
|
36.979
|
1645.94
|
52.88
|
39.49
|
1642.06
|
49.91
|
37.93
|
1671.494
|
50.92
|
8.7129
|
1642.395
|
51.538
|
37.0312
|
1669.15
|
53.468
|
15.94
|
1676.95
|
50.20
|
7.79
|
1668.83
|
53.04
|
15.60
|
-
|
-
|
-
|
1671.334
|
54.021
|
15.0965
|
64 Day
|
Control
|
AD1
|
Curative Group
|
L.S
|
Cur.1
|
Cur.2
|
Center X
|
FWHH
|
% Area
|
Center X
|
FWHH
|
% Area
|
Center X
|
FWHH
|
% Area
|
Center X
|
FWHH
|
% Area
|
Center X
|
FWHH
|
% Area
|
1538.954
|
55.966
|
3.5489
|
1533.677
|
62.646
|
2.8257
|
1514.012
|
23.218
|
6.1911
|
1511.004
|
38.496
|
6.4623
|
1510.485
|
31.13
|
4.5786
|
1556.406
|
49.892
|
0.3529
|
1576.894
|
63.849
|
0.1966
|
1531.579
|
21.638
|
6.2948
|
1536.076
|
41.11
|
10.7264
|
1528.2
|
29.772
|
4.9853
|
1588.88
|
49.229
|
0.0024
|
1611.909
|
65.445
|
0.7133
|
1546.657
|
23.799
|
7.2858
|
1555.055
|
43.421
|
2.7797
|
1545.917
|
31.524
|
7.5686
|
1603.074
|
49.115
|
1.519
|
1644.688
|
55.403
|
3.7068
|
1570.711
|
33.315
|
4.5493
|
1594.194
|
49.016
|
7.0204
|
1575.55
|
38.232
|
4.3931
|
1640.722
|
37.776
|
2.0096
|
1700.087
|
63.467
|
0.3296
|
1607.181
|
42.815
|
9.0204
|
1627.511
|
37.511
|
12.9508
|
1613.556
|
44.197
|
6.9779
|
1654.765
|
51.837
|
3.971
|
|
|
|
1633.809
|
36.347
|
16.5035
|
1649.373
|
33.977
|
13.2497
|
1639.679
|
38.677
|
15.963
|
1687.707
|
49.037
|
0.2147
|
|
|
|
1654.503
|
32.521
|
11.6542
|
1675.208
|
32.95
|
3.87
|
1654.653
|
34.479
|
7.3391
|
|
|
|
|
|
|
1679.157
|
27.622
|
2.9269
|
|
|
|
1679.511
|
34.993
|
4.3604
|
The average percentage area of the amide I sub-bands protein β-sheet (1620–1639 cm− 1), random coil (1640–1649 cm− 1) and alpha-helix (1650–1659 cm− 1) secondary structure, β-sheet (1620–1639 cm− 1), β-turn (1660–1670 cm− 1), parallel β-sheets protein secondary structure (1670–1689 cm− 1) and anti-parallel β-sheets (1690–1695 cm− 1). |
In addition to the noticed band shift, the appearance of a new bands in both AD1curve fitted spectrum around 1621, 1677 cm− 1 is also detected which correspond to β-turns and parallel β-sheets secondary structure respectively. This result Consistent with Lisa et al., 2006 [34] findings that detected an elevated β-sheet secondary structure in amyloid plaques indicated by the strong absorbance sFTIR band at 1625 cm− 1 compared to other surrounded unaffected area. β-sheets secondary structure as a shoulder of the dense core plaque at 1630 cm− 1 was also recorded 43.
However, the protein secondary structure may contains another types of protein secondary structure. Many reports showed a conflicting views about the structure of tau proteins in AD brain, as it may be mainly α-helical structure to mainly β-sheet, or a mixture of mostly random coil and increased amount of β-structure44–46.
In our study, the hippocampal tissue at 42d for all groups are dominated with random coil, the band around 1640 cm− 1, over β-sheet structure respectively. On the other hand, the amide II sub-bands have nearly the same values in AD1 as the control group, but in AD2 they have marked bands shift towards higher frequencies than the control curve fitted group [Table 1] [from 1511, 1535, 1562 cm− 1 to 1518, 1542,1569 cm− 1] respectively. It may be suggested that, the beginning of misfolded proteins starts in the NH bending functional groups evident by the recorded increase in amide II sub-bands HBW in AD group compared to the control [Table 1]. This observation holds true for the curative group at 42d. By contrast, in LS treated groups, Cur & LS spectral analysis, showed nearly the same amide I sub-band with around 1610 cm− 1, 1621 cm− 1 respectively with the appearance of new band around 1671 cm− 1, that corresponds to β-sheet structure, and a band shift to higher frequency at 1645, 1642 cm− 1 for Cur & LS curve fitting data respectively.
Graphs of the parentage area of the amide I sub-bands, summarizing the total β-structure, random coil and alpha-helix content in rat hippocampal tissue at both tested time intervals are showed in Fig. 3(a, b). σ-helix protein structure is absent in 42d for all groups; the predominant protein secondary structure is random coil and in lower amount β-sheet at two different frequencies as seen in Table 1. Meanwhile, σ-helix structure is strongly present in all groups at different amounts after 65d and is completely absent in AD group. The majority protein structure in AD group is random coil (the band centred at1642 cm− 1 and a very little percentage of antiparallel β-sheet at 1694 cm− 1.
From [Table 1 & Fig. 3], at early stage of Al intoxication (42d), the formation of both extracellular soluble random coil and parallel β-sheets structure is evident by the noticeable increase in the % area of the sub-bands around 1642 cm− 1, 1621 &1674 cm− 1 respectively] than the control and LS treated groups. The 1621 cm− 1 amide I sub-band is only detected in AD1 peak resolved, which may suggested the misfolding of protein may started through parallel β-turn structure as it is disappeared in AD group at 65d. The protein misfolding after 65d of Al induction is well established and has formed the intracellular random coil (major) and anti-parallel (minor) insoluble Aβ proteins.
The detected increase in the random coil sub-bands in AD1 and AD2 peak resolved may also suggests another second transit phase through random coil structure as our and other previous works stated3,5,9,11. These findings are supported by the drastic increase in the sub-bands % area around 1642 cm− 1 in AD group from [39.49 & 37.93 in AD1, AD2 respectively at 42d to 47.39 at 65d].
The slight observed increase in the HBW of the random coil of the β-protein sub-bands around 1640 cm− 1 in AD1 group at 42d may be attributed to the fact that; the protein misfolding is a slow process that takes place through intermediate/or transit phase, which needs to break the hydrogen bonds and /or the reduction of the β-strand length, and hence proteins denaturation/ misfolding takes place 4,5,47–49.
The previous transit phase assumption of misfolded protein through β-sheets structure [sub-band at 1621 cm− 1] in AD1 group holds true for the renaturation of protein in Cur group after 65d of treatment, which predominantly occurs through β-sheets structure but in this time accompanied with marked band shift to higher frequencies at ~ 1633, 1627 cm− 1 in Cur1 and Cur2 after 65d respectively].
These results consistent with that reported earlier1,7,13,14 which stated that nearly 90% of Aβ (40 amino acids) represents the soluble extracellular form while, 10% of Aβ (42) represents the hard intracellular hydrophobic core of Aβ form or NFTs.
In this study, 91% of the amide I protein present in AD hippocampal tissue at 65d is random coil (the soluble extracellular protein) and 9% is antiparallel β-sheets structure which now represents the hard intracellular hydrophobic core of NFTs.
The more neurotoxic protein structure that is associated with AD is believed to be the soluble extracellular Aβ that begin and continuously formed at the early stage of the disease pathogenesis and progression rather than tau proteins/NFTs in human, mice and rat brains50.
The drastic benefit of utilizing SFTIRM over the conventional FTIR on following the protein aggregation is that how the secondary structure of protein in AD group is restricted exactly to two entities; random coil and antiparallel β-sheets represented by only two sub-bands at 1642 &1694 cm-1 respectively, as the literatures compared to our previous studies 3,4 that gives several sub-bands for also rat hippocampal brain tissue.
Figure 3 (a-f) showed the above marked contribution of β-sheet, random coil and could be summarized the selected percentage area ratios and the total % area of σ-helix, β-sheet and random coil as follows
-
Total percentage area of σ-helix, β-structure and random coil after 42d and 65d in all tested groups; is an indicator of what kind of protein is there and how much (Fig. 3(a, b))
-
β-sheet/ Random coil represented by %area 1640-1/1669cm-1 & 1640/1687; is an indicator/biomarker for the transformation, [through transit phase between these proteins secondary structure during both Al and LS treatments after 42d and 65d (Fig. 3(c, d)).
- Amide I/Amide II is used as an indicator of the change composition of the protein pattern; the changes in the protein secondary structure folding and unfolding (Fig. 3(e, f)).39.
3.3. SFTIRM mapping
The protein profile of all tested groups at 42d and 65d are showed in Fig. 4 and were obtained by the peak around 1657 cm− 1. It seems that at early stage of either AL and LS induction, there are great activities in protein misfolding/refolding as mentioned above and SFTIRM, our previous reports 3,5,13,22 are also supporting these results.
3.4. Multivariate analysis
In order to validate and monitor the difference in biochemical structure of the control, Al-induced AD like and other tested groups in brain cortical tissue, we divided the IR spectra into three distinct regions namely; the membrane lipids and fatty acids of CH stretching over 3020 − 2810 cm− 1, the amide I and amide II of C = O stretching and N-H bending of proteins respectively over 1700 − 1500 cm− 1, and the Phosphate & nucleic acids1300-900 cm− 1 regions. Here in this paper we only focused on the amide region over 1700 − 1500 cm− 1. The other regions will be discussed later in another work.
Statistical analysis
We used both HCA and PCA multivariate analysis, over the amide I and amide II bands (1700 − 1500 cm− 1) of rat brain hippocampal tissues, to monitor how significantly the biochemical spectra among groups are different. These analysis were performed on the already processed sFTIRM spectra obtained of the absorbed amide I and amide II band’s intensities.
Figure 5(a) showed that, after 42d of treatment, both HCA and PCA could not differentiate between Cont and AD groups, at the same time, they clustered proteins for AD group into two entities. It seems that the proteins have two different secondary structure at the early stage of disease (42d), it fluctuated still between normal and misfolded types, in AD model, indicated by the higher heterogeneity between the two clustered groups of in HCA; greater than 430 [ black arrow (zone of interference (brown arrow)].
. Meanwhile, HCA and PCA between Cur and AD groups succeeded partially to differentiate between these groups over the same time of treatment Fig. 5(b).
On the other hand, for longer time of treatment after 65d, HCA and PCA between Cont and AD groups showed a significant and a complete segregation Fig. 5(c). The same complete segregation between the Cur and the AD groups are shown after 65d Fig. 5(d).
At the late stage of induced AD-like model after 65d, both Cont and AD groups established [obtained] their own protein secondary structure each of their main form and hence, HCA & PCA succeeded completely to segregate between them. Meanwhile, Cur group is the one that shows now two main protein secondary structure as the recovery and reversing the process of misfolding is not completed yet at that time of the experiment Fig. 5(d) black arrows (zone of interference (brown arrow)