Acta Histochemica 116 (2014) 1359–1366
Contents lists available at ScienceDirect
Acta Histochemica journal homepage: www.elsevier.de/acthis
Fluorescence, aggregation properties and FT-IR microspectroscopy of elastin and collagen fibers Benedicto de Campos Vidal ∗ Department of Structural and Functional Biology, Institute of Biology, University of Campinas (Unicamp), Rua Monteiro Lobato 255, CEP 013083-862 Campinas, SP, Brazil
a r t i c l e
i n f o
Article history: Received 15 July 2014 Received in revised form 18 August 2014 Accepted 19 August 2014 Keywords: Elastin Collagen type I Collagen bundles Hydrophobicity Birefringence FT-IR
a b s t r a c t Histological and histochemical observations support the hypothesis that collagen fibers can link to elastic fibers. However, the resulting organization of elastin and collagen type complexes and differences between these materials in terms of macromolecular orientation and frequencies of their chemical vibrational groups have not yet been solved. This study aimed to investigate the macromolecular organization of pure elastin, collagen type I and elastin–collagen complexes using polarized light DIC-microscopy. Additionally, differences and similarities between pure elastin and collagen bundles (CB) were investigated by Fourier transform-infrared (FT-IR) microspectroscopy. Although elastin exhibited a faint birefringence, the elastin–collagen complex aggregates formed in solution exhibited a deep birefringence and formation of an ordered-supramolecular complex typical of collagen chiral structure. The FT-IR study revealed elastin and CB peptide N H groups involved in different types of H-bonding. More energy is absorbed in the vibrational transitions corresponding to CH, CH2 and CH3 groups (probably associated with the hydrophobicity demonstrated by 8-anilino-1-naphtalene sulfonic acid sodium salt [ANS] fluorescence), and to CN, ıNH and ωCH2 groups of elastin compared to CB. It is assumed that the ␣-helix contribution to the pure elastin amide I profile is 46.8%, whereas that of the B-sheet is 20% and that unordered structures contribute to the remaining percentage. An FT-IR profile library reveals that the elastin signature within the 1360–1189 cm−1 spectral range resembles that of Conex–Toray aramid fibers. © 2014 Elsevier GmbH. All rights reserved.
Introduction Interactions between collagen and elastic fibers have been proposed in reports using histological, histochemical and histophysical methods (Braga-Vilela and Vidal, 2006; Aldrovani et al., 2007) and electron microscopy (Gibson et al., 1997; Kielty et al., 2002). Based on the work of Gibson et al. (1997), Kielty et al. (2002) have suggested that a protein might link elastic fibers to collagen. Hard extraction methods and histochemical methods, accompanied by simultaneous analysis with polarized light microscopy, have suggested a tight collagen–elastic fiber relationship in porcine pericardium and aorta (Braga-Vilela and Vidal, 2006; Aldrovani et al., 2007). The amount of elastin isolated from some tissues has been surveyed with the following results: 30–43% in aorta, 2–5% in skin, 5–8% in ligamentum nuchae (Ayer, 1964), and ∼2 ± 1% in the dermis
∗ Corresponding author. E-mail address: [email protected] http://dx.doi.org/10.1016/j.acthis.2014.08.007 0065-1281/© 2014 Elsevier GmbH. All rights reserved.
of healthy human skin, as measured using morphometric techniques (Uitto et al., 1983). Regarding the anatomic structure of elastic fibers in the skin, three forms have been described: filaments, a natural network and 20 nm-beaded fibrils (Ayer, 1964). Images of morphological aspects along with biomechanical properties have been reported for elastic fibers and collagen in arteries in a study that combined second-harmonic generation (SHG) signals, two-photon excited fluorescence and multiphoton microscopy which showed the stress–strain distribution of these fibers in the arterial wall (Zoumi et al., 2004). One of the methods proposed for extraction of elastic fibers is based on boiling in diluted alkali. However, “by extending the extraction time it is found that considerable destruction of elastica occurs” (Ayer, 1964 – review). Another recommended method involves boiling the sample in dilute acetic acid; in this case, no changes in the fibrillar-meshed architecture of the elastica have been reported (Ayer, 1964; Eyre et al., 1984; Mecham, 2008). Some optical methods have indicated that elastic fibers can exhibit birefringence; this anisotropic optical property has been attributed to the oriented distribution of the filamentous elements
1360
B.d.C. Vidal / Acta Histochemica 116 (2014) 1359–1366
of elastin (Romhányi et al., 1975; Romhányi, 1986). A compositional complexity of molecules interacting with elastin includes type VIII collagen and emilins, as well as the microfibril-associated glycoprotein 1 (MAGP-1), which is an integral microfibril molecule that is important for the structural integrity of the elastic fibers (Kielty et al., 2002). These reports indicate that the composition and structure of the elastic fibers is complex, and that the biological role of many of their molecular components is far from being clear. A contribution to a comparative study of the macromolecular organization of elastin and collagen fibers and their complexes could be done by evaluating the optical anisotropic properties of these structures in separate and in the aggregates these materials could form in solution, under dialysis conditions. In addition, comparison of Fourier transform-infrared (FT-IR) microspectroscopic signatures between elastin and collagen bundles could probably provide information on differential frequencies of their most important vibrational groups. In the present study, pure elastin, collagen type I and elastin–collagen complexes, formed in solution, were studied using polarized light microscopy. Additionally, differences between elastin and collagen bundles were investigated at the level of their vibrational groups, as studied by FT-IR. Hydrophobicity by ANS fluorescence, which could add information for the interpretation of FT-IR results, was also searched in pure elastin. Material and methods
Elastin fluorescence analysis Some elastin smears were treated with 0.1% 8-anilino-1naphtalene sulfonic acid sodium salt (ANS) (T484 Kodak, Rochester, NY, USA) in a non-polar solvent, butanol, for 30 min in the dark; next, unbound ANS was removed from the smears using pure butanol (Vidal, 1978, 1980). Then, the preparations were cleared in xylene and mounted on slides in Nujol mineral oil (M3516; SigmaAldrich). The ANS-stained preparations were examined under a Carl Zeiss Axiophot 2 photomicroscope (Oberkochen/Munich, Germany) equipped for epifluorescence and using an HBO-mercury short-arc 103 W/2 as UV source and a 365 nm filter for the excitation light; FT395 and LP-420 nm filters were also employed. For image analysis and photomicrographs, a Zeiss Axiocam HRc camera and Kontron KS400-3 software were used. Optical anisotropy The identification of birefringence characteristics in unstained elastin fragments and elastin–collagen type I complexes was performed using an Olympus BMX 51 polarization microscope (Olympus Corp., Tokyo, Japan) connected to a computer and equipped with optical devices for differential interference contrast microscopy (DIC). Optical anisotropy documentation was carried out with Image-Pro-Plus 6.3 software (Media Cybernetics, Bethesda, MD, USA).
Elastin and collagen type I preparation FT-IR microspectroscopy Pure elastin obtained from bovine nuchal ligament (SigmaAldrich, St. Louis, MO, USA) was used. Small fragments of this material were suspended in distilled water and smears were prepared from that suspension. The smears were dried following three stages: first, they were left in a refrigerator for 24 h, then they were left at 37 ◦ C for 3 h, and finally they were dried at 60 ◦ C for 2 h. Collagen type I was extracted from the rat tail tendons of the experimental animals previously used to study collagen bundle supraorganization in the skin (Ribeiro et al., 2013). The tails were washed extensively, and their skin was removed. The collagen bundles, which were visible in the external surface of the tail, were removed and washed in Milli-Q water from a Millipore Direct-Q3System (Molheim, France). After three washes, well-separated and clean collagen bundles were immersed in a 3% acetic acid solution. This solution remained in the refrigerator for 2–3 days for complete dissolution of the collagen fibers. Next, a careful filtration was performed. Subsequently, the collagen fibers were reconstituted by adding 20% NaCl solution to the collagen solution at final concentration of 10%. Immediately afterwards, visibly precipitated fibers, showing a transparent aspect, were left in the refrigerator; white floating reconstituted fibers were dissolved again in a 3% acetic acid solution and kept in the refrigerator until they were completely dissolved. Additional acetic acid solution was added when necessary. After complete fiber dissolution, a careful filtration was performed. In an attempt to form a complex between elastin and collagen, 150 mg of elastin powder was suspended in 3 mL of the collagen acetic acid solution and dialyzed in Milli-Q water in the refrigerator, until reconstituted collagen formed. In this case, reconstitution was performed without using NaCl. The dialysis was interrupted as soon as reconstituted collagen fibers appeared. At this step, a collagen–elastin complex was evident. All of the protocols involving animal care and use for collagen attainment were approved by the Committee for Ethics in Animal Experimentation of the University of Campinas (CEAE/IB/Unicamp) (protocol 2700-1) and were in accordance with the Guidelines of the Canadian Council on Animal Care.
FT-IR spectral profiles were obtained for sections of bovine collagen bundles (CB) and elastin smears. CB sections were prepared from bovine flexor tendons acquired from a slaughterhouse, as previously reported (Vidal, 2013). Briefly, tendon fragments cut parallel to the tendon axis and fixed in 4% paraformaldehyde at pH 7.4 under vacuum were embedded in Hystosec medium (Merck, Darmstadt, Germany) to obtain 10 m thick sections which were dewaxed in xylene, mounted on slides and dried at 37 ◦ C. Because an all-reflecting objective (ARO) was used for the microspectroscopic examination, the materials were mounted on the surface of thick gold-covered reflective glass slides (Mello and Vidal, 2014). The Illuminat IR IITM microspectrometer (Smiths Detection, Danbury, CT 06810, USA) equipped with a liquid-nitrogen-cooled mercury-cadmium-telluride detector, ARO (16×) and Grams/AI 8.0 software (Thermo Electric, Waltham, MA, USA), and coupled to an Olympus BM51 microscope was used for FT-IR microspectroscopy. The performance of the equipment was validated using a low signal-to-noise ratio (7929:1) (Vidal and Mello, 2011). The absorbances of the samples and background were obtained using 64 scans for each preparation with a spectral resolution of 4 cm−1 . The spectral absorption profiles were determined at frequencies ranging from 4000 cm−1 to 650 cm−1 . A plus-zero baseline was acquired, and the FT-IR spectrum was normalized using the highest absorbance peak as a reference. After 10 spectral profiles were obtained, an average profile was produced using the Grams software (Vidal and Mello, 2011; Vidal, 2013; Mello and Vidal, 2014). The area of the band peaks was determined using the numerical integral statistics of the Grams software. Peak fitting and peak fitting estimate procedures using a Gaussian function and low sensitivity level were also applied to specific spectral regions of absorption band peaks, using the Grams software, enabling the area of selected bands to be calculated (Mello and Vidal, 2014). An absorbance ratio, previously proposed to determine the integrity of the collagen triple helices or the relative amount of native collagen in a sample (Gordon et al., 1974; Goissis et al., 2001), was also estimated.
B.d.C. Vidal / Acta Histochemica 116 (2014) 1359–1366
1361
Fig. 1. ANS-stained elastin fragments with various dimensions exhibit blue fluorescence. Bar = 50 m.
Results
greater accessibility of the hydrophobic groups to ANS in isolated elastin.
Fluorescence Optical anisotropy The ANS-stained elastin fragments revealed a blue fluorescence (Fig. 1) which was deeper in comparison with that reported when ANS was used in histochemical studies (Vidal, 1978, 1980; Braga-Vilela and Vidal, 2006). This result is likely due to the
The smaller elastin fragments exhibited poor birefringence intensity at their surface. Birefringence slightly increased in the aggregates of fibrillar elastin (Fig. 2), due to fragment overlapping
Fig. 2. Elastin fragments examined under a polarized light microscope showed weak birefringence at their surfaces. When the fragments were stacked, their birefringence combined to increase the brilliance intensity. Bar = 50 m.
1362
B.d.C. Vidal / Acta Histochemica 116 (2014) 1359–1366
Fig. 3. Very thin collagen type I fibers reconstituted from acetic acid solutions, using dialysis in distilled water. In the segments with crossed fibers, the birefringence is internally compensated, and a black image results. The differences in brightness intensity are due to differences in fiber aggregation. Bar = 50 m.
and the relative thickness increase. After birefringence compensation in the elastin, it was possible to determine that the birefringence sign was positive and that this phenomenon was non-homogeneously distributed along the axis of the in tandem fibril aggregation. Pure reconstituted rat collagen thin fibers obtained from the collagen type I acetic acid solutions showed a deep positive
birefringence (Fig. 3), similar to that reported for bovine collagen (Vidal, 1995). Adding elastin to the collagen type I acetic acid solution increased the initial viscosity of the collagen solution. Microscopic examination of the elastin–collagen mixture revealed some regions containing only reconstituted fibers (Fig. 3) and others containing the elastin–collagen complex (Figs. 4 and 5). A visually
Fig. 4. Binding of elastin to acid-soluble collagen, which were dialyzed together in distilled water. After dialysis, a highly birefringent elastin–collagen complex formed. The areas of crossed fibers show internal birefringence compensation and dark areas. Collagen I reconstituted fibers are also seen. Bar = 30 m.
B.d.C. Vidal / Acta Histochemica 116 (2014) 1359–1366
1363
Fig. 5. Panel containing images of the collagen–elastin complex obtained at four different focal distances ((A)–(D)) using DIC. In (A) the image is shown at the most superficial focal level; (B) is the next level, and the focal depth continues to increase (C) until (D) which has the deepest focal distance. The helical structure of the supramolecular organization involved is shown. In this case, interference colors due to the birefringence phenomenon are observed because DIC is based on polarized light microscopy principles. Bar = 50 m.
detectable increase in birefringence brightness was verified in the elastin–collagen complex fibers from acetic acid solutions spread on glass slides. Some elastin–collagen complex aggregates appeared with the supramolecular organization of intertwined helical fibers typical for collagen bundles, as demonstrated by DIC-polarized light microscopy analysis (Fig. 5). FT-IR microspectroscopy Differences in the distribution of band peak frequencies were detected in the FT-IR spectral signature of elastin in comparison to CB (Fig. 6). Visual inspection of the FT-IR spectra indicates a general shift of the peaks to a lower frequency in elastin compared to CB. The highlighted heights were more evident when considering a subtraction spectrum obtained from differences in absorbance values (Fig. 6, blue line). If using the numerical integral statistics provided by the Grams software in the 3669–3127 cm−1 spectral window, 292 area points were revealed for CB and 297 points were revealed for elastin, corresponding to areas of 213 and 212 units, respectively. Thus, despite the spectral profile appearance based on absorbance values with varying wave numbers (Fig. 6), the elastin did not show a significantly smaller area in this spectral window. This is because, for the numerical integral statistics calculation, the Grams software requires that a new baseline is considered, with the left and right edges at 3669 cm−1 and 3127 cm−1 , respectively, but at the height of the respective absorbance values. In the elastin case, the left edge corresponds to the absorbance value of 0.13
and the right edge corresponds to the absorbance value of 0.81, whereas in the collagen case, the left and right edges correspond to absorbance of 0.02 and 0.60, respectively. Five peaks were selected using Grams’ peak fitting procedure applied to both elastin and CB profiles at the spectral window of ∼3000–2940 cm−1 . For elastin, the total area comprised in the spectral profile from 2977 to 2951 cm−1 yielded a value of 5.20 units, and the maximal area, corresponding to the band peak at 2967 cm−1 , was equal to 1.15 units. For CB, the profile region between 3007 and 2938 cm−1 yielded a total area of 1.96 units, and the maximal area, corresponding to the band peak at 2959 cm−1 , was equal to 0.68 units. Thus, elastin showed noticeably higher total area and peak maximal area at this spectral window. Band peaks representative of amide A were revealed at 3286 cm−1 and 3325 cm−1 for elastin and CB, respectively. Regarding the amide I spectral region, it was found to be represented in the 1725–1589 cm−1 range for elastin and in the 1706–1657 cm−1 range for CB (Fig. 6). The numerical integral statistics obtained for this spectral window after peak fitting revealed an area of 34 units for elastin and 9 units for CB. After peak fitting, band peaks at 1656 cm−1 , 1648 cm−1 , 1639 cm−1 , 1628 cm−1 and 1620 cm−1 were detected for elastin. The amide I spectrum for CB yielded two visually distinct peaks (Fig. 6). After peak fitting procedure, four band peaks were obtained for CB, such that the two most representative ones appeared positioned at 1689 cm−1 and 1678 cm−1 and two smaller band peaks were positioned at 1665 cm−1 and 1658 cm−1 . The participation of each peak in the vibrational transition moment of the amide I was evaluated based
1364
B.d.C. Vidal / Acta Histochemica 116 (2014) 1359–1366
Fig. 6. FT-IR spectra obtained for collagen bundles (CB) (black line) and elastin (red line). The profile that represents the absorption difference between elastin and CB is also presented (blue line). X axis, wavenumbers in cm−1 ; Y axis, absorbances.
on the approximate fractional percentile of each band peak. Based on a previously reported study on FT-IR linear dichroism for the same CB material, one could hypothesize that the band peaks at 1689 cm−1 and 1678 cm−1 are related to C O (∼39%) and C N (∼33%) vibronic transition moments, respectively (Vidal, 2013). The band peaks positioned at 1665 cm−1 and 1658 cm−1 occupied areas corresponding to ∼14% and ∼6%, respectively, of the total area. Based on present spectral results for CB after peak fitting, a ratio equal to 0.75 was calculated from absorbances obtained at 1248 cm−1 /1453 cm−1 . For reconstituted collagen fibers (as those shown in Fig. 4) analyzed for FT-IR (spectrum not shown), such a ratio was equal to 1.10, that is, close to that reported for collagen in natura (1.38), evaluated at 1235 cm−1 /1450 cm−1 (Gordon et al., 1974). The lower ratio value found for CB in the present study (0.75) may be due to some degree of denaturation, introduced by paraffin embedding at 60 ◦ C (Mello and Vidal, 1973), a fact that is not supposed to occur in the process of preparing pure collagen to obtain reconstituted fibers (Vidal, 1995). Indeed, according to Gordon et al. (1974), the ratio obtained for gelatin cast at 60 ◦ C is lower (0.59) than that obtained for collagen in natura. The ratio evaluated from absorbance obtained at 1248 cm−1 /1453 cm−1 , comparable to that obtained at 1235 cm−1 /1453 cm−1 providing information on integrity of the collagen fiber triple helix (Goissis et al., 2001), would not be properly applicable to elastin because this molecule does not exhibit a triple helix conformation. Maybe, in the case of elastin, such a ratio value is promoted by ␣-helix and -sheet content contribution. Another IR spectral difference between elastin and collagen occurred in the range of 1350–1200 cm−1 , which is approximately the frequency range noted in the literature for CN, ıNH and ωCH2 groups (Fabian and Mäntele, 2002). Peak fitting revealed four peaks for elastin: from the center of the peak at 1331 cm−1 to the center of the peak at 1247 cm−1 , the numerical integral area was equal to 12
units, and the highest peak was positioned at the 1247 cm−1 band. Regarding CB, the numerically integrated area was equal to 9 units, and the highest partial area was positioned at the 1247 cm−1 band.
Discussion Under present experimental conditions, the elastin fibers revealed a predominantly unordered structure, which was assumed from the faint birefringence restricted to their surface. On the other hand, the intense birefringence in collagen fibers increased in elastin–collagen type I aggregates. In addition, these aggregates acquire a supramolecular organization of intertwined helical fibers, typical for collagen fibers. Present findings support the hypothesis that collagen type I can complex with elastin and induce elastin to adopt a supramolecular organization with a helical (chiral) structure like that previously reported for collagen (Vidal, 1995, 2003, 2013; Silva et al., 2013). An intimate relationship between CB and elastic fibers has already been suggested in porcine pericardium and aorta studies using ANS and dansylchloride staining (Braga-Vilela and Vidal, 2006). It is hypothesized here that collagen may bind to elastin through a lectin-type recognition reaction. Thus, basic differences between elastin and collagen fibers are not caused only by amino acid differences. As demonstrated by present optical anisotropy investigation, elastin does not display the same degree of molecular order as collagen fibers. This fact is an indication that their infrared spectral characteristics could also differ. It should be emphasized that the present work addresses very complex supramolecular architectures. Then, in the FT-IR microspectroscopy context, it is necessary to consider Eglinton’s (1970) words: In any case, the spectrum of such a complicated
B.d.C. Vidal / Acta Histochemica 116 (2014) 1359–1366
molecule is not interpretable in complete detail and one would not expect to arrive at the structure from the infrared spectrum alone. Nevertheless, regarding the 3700–3130 cm−1 IR spectral region, and considering the report by Popescu et al. (2010 – p. 1076), the following assignments can be proposed: 3560 cm−1 , stretching vibrations of free OH groups; 3390 cm−1 , stretching vibrations of H-bonded OH groups; and 3290 and 3195 cm−1 , free N H and H-bonded N H groups, respectively. The band related to vibrations of the OH groups is recognized by its broadness and low intensity. The OH groups are involved in two hydrogen-bonding modes. The occurrence of amide A at 3286 and 3325 cm−1 in the spectral profile of elastin and CB, respectively, indicated involvement of N H peptide groups in different H-bonding types. The changes currently found at the level of the OH and NH groups for amide I and II demonstrate the importance of H-bonding in the formation and stabilization of elastic structures. The assignments proposed for the peaks detected at the 2977–2951 cm−1 spectral range are CH, CH2 , and CH3 groups. In the case of elastin, these groups are mostly present in leucine, isoleucine and asparagine and caused the hydrophobicity that was demonstrated by the intense fluorescence due to the ANS binding. For the interpretation of results based on infrared band peak areas at this spectral window, the significance of infrared energy absorption was considered as “quantized energy into the fundamental vibration modes of the molecules under study” (Eglinton, 1970). There are reports establishing for different materials that the area under an absorption band peak in IR spectroscopy can be related to specific vibrational absorbed energy involved (Wilson and Wells, 1946; Eglinton, 1970; Yugami et al., 1996; Vidal et al., 2014). Present data thus support the hypothesis that the vibrational transitions of the CH, CH2 , and CH3 groups absorbed more energy in elastin than in CB. Similarly, regarding the spectral window of 1350 cm−1 to 1200 cm−1 , the vibrational transitions of the CN, ıNH and ωCH2 groups (Fabian and Mäntele, 2002) absorbed more energy in elastin compared to CB. According to Goormaghtigh et al. (2006), and considering several other authors, the process of “. . .curve fitting requires a series of subjective decisions that can dramatically affect both the result and the interpretation. Furthermore, curve fitting has a tendency to overestimate the -sheet content of primarily helical proteins and routinely finds 15–20% -sheet content for proteins that actually have none.” Moreover, “among the 3201 frequencies tested . . . only three frequencies contain all the information that is available to predict the secondary structure of the proteins” (Goormaghtigh et al., 2006). Despite these authors’ considerations, present peak fitting data allowed to associate five selected band peak areas to data from the literature that inform on the participation of ␣-helix and -sheet-type secondary structures as well as unordered molecular structures in the elastin amide I profile (Fabian and Mäntele, 2002; Fabian and Naumann, 2004; D’Souza et al., 2006). Based on 13 C NMR experiments, Kumashiro et al. (2006) have already reported that hydrophobic and ␣-helical cross-linking domains cooperate in the elastin folding and that “hinge” sequences (“unordered”?) occur in the middle of each cross-linking domain. Thus, in an attempt to quantify the ␣-helix and -sheet-type secondary structures in the elastin amide I IR spectral profile, it is assumed that the largest band peak area at 1655 cm−1 and the area under the band peak at 1648 cm−1 summed, correspond to a 46.8% contribution of ␣-helix. The area for the band peak positioned at 1639 cm−1 indicates a 20% participation of -sheet in this profile. The remaining area, corresponding to band peaks at 1628 cm−1 and 1620 cm−1 , is likely assigned to a contribution to this amide I profile of unordered structures that would also play a part in the poor birefringence phenomenon observed in the elastin fragments. When the Grams software library was consulted to find substances that presented FT-IR characteristics similar to those of
1365
elastin, 20 compounds were found, including 5 types of aramid fibers and 14 types of nylon fibers. The fiber spectrum that was most similar to that of elastin, within the 1360–1189 cm−1 range, was that of the Conex–Toray aramid. This polyamide contains a large amount of hydrogen bonds in the amide groups of its aromatic rings. No other particular spectral characteristics, besides those reflecting generic spectral polyamide properties, were found when comparing elastin, aramids and nylon types. On the other hand, the FT-IR spectral profile of CB has been previously found to be comparable to that of nylon 6 fibers, revealing vibronic molecular moments distributed parallel to their long axis and assigning CN and ıNH groups to the axis of collagen and nylon 6 fibers (Vidal and Mello, 2011; Vidal, 2013).
Conclusions Present data support the hypothesis that elastin can link to collagen type I, forming an ordered supramolecular complex. Despite the not so ordered structural organization in pure elastin as that in collagen fibers, when elastin–collagen aggregates are obtained after dialysis, a complex is generated with the supramolecular organization of intertwined helical fibers typical of collagen chiral structure. A detailed comparative FT-IR study of the elastin and CB revealed amide A at 3286 cm−1 and 3325 cm−1 , respectively, indicating that their peptide N H groups were involved in different types of H-bonding. The amide I band in elastin and CB is also identified at different IR frequency range (1756–1589 cm−1 , elastin; 1689–1678 cm−1 , CB). Based on FT-IR data for the 2977–2951 cm−1 spectral range, it is proposed that more energy is absorbed in the vibration transitions corresponding to CH, CH2 and CH3 groups in elastin in comparison to CB. These groups, mostly present in leucine, isoleucine and asparagine, are probably responsible for the hydrophobicity demonstrated in the elastin with ANS fluorescence. Vibrational transitions at the 1350–1200 cm−1 region, that are assumed to be assigned to the CN, ıNH and ωCH2 groups, are also proposed to absorb more energy in elastin compared to CB. The sum of the areas under the band peaks at 1655 cm−1 and 1648 cm−1 represents the ␣-helix contribution (46.8%) to the elastin amide I profile. The area under the band peak at 1639 cm−1 indicates the -sheet participation (20%) in this profile, and the areas corresponding to band peaks at 1628 cm−1 and 1620 cm−1 are assigned to the contribution of unordered structures to this profile, presumably correlated with the poor birefringence phenomenon in the elastin fragments. Consulting the FT-IR profile library provided by the Grams/AI microspectroscopic software, elastin keeps similarity to Conex–Toray aramid fibers within the 1360–1189 cm−1 IR spectral range.
Acknowledgments The author is indebted to Dr. Maria Luiza S. Mello for discussing and reviewing the manuscript. The support of Fundac¸ão de Amparo à Pesquisa do Estado de São Paulo (FAPESP), Brazil (grants no. 2003/04597-0, 2007/058251-8 and 2013/11078-0) is gratefully acknowledged.
References Aldrovani M, Guaraldo AMA, Vidal BC. Fluorescence, birefringence and confocal microscopy of the abdominal aorta from nonobese diabetic (NOD) mice. Acta Histochem 2007;109:248–54.
1366
B.d.C. Vidal / Acta Histochemica 116 (2014) 1359–1366
Ayer JP. Elastic tissue. In: Hall DA, editor. International review of connective tissue research, vol. 2. New York and London: Academic Press; 1964. p. 33–100. Braga-Vilela AS, Vidal BC. Identification of elastic fibers and lamellae in porcine pericardium and aorta by confocal, fluorescence and polarized light microscopy. Acta Histochem 2006;108:125–32. D’Souza AJ, Hart DS, Middaugh CR, Gehrke SH. Characterization of the changes in secondary structure and architecture of elastin–mimetic triblock polypeptides during thermal gelation. Macromolecules 2006;39:7084–91. Eglinton G. Application of infrared spectroscopy to organic chemistry. In: Scheinmann F, editor. An introduction to spectroscopy methods for the identification of organic compounds, vol. 1. Oxford: Pergamon Press; 1970. p. 129–41. Eyre DR, Paz MA, Gallop PM. Cross-linking in collagen and elastin. Ann Rev Biochem 1984;53:717–48. Fabian H, Mäntele W. Infrared spectroscopy of proteins. In: Chalmers JM, Griffiths PR, editors. Handbook of vibrational spectroscopy. Chichester: Wiley; 2002. p. 3399–425. Fabian H, Naumann D. Methods to study protein folding by stoppedflow FT-IR. Methods 2004;34:28–40. Gibson MA, Kumaratilake JS, Cleary EG. Immunohistochemical and ultrastructural localization of MP78/70 (ig-h3) in extracellular matrix of developing and mature bovine tissues. J Histochem Cytochem 1997;45:1683–97. Goissis G, Suzigan S, Parreira DR, Raymundo SRO, Chaves H, Hussain KMK. Malhas de polipropileno recobertas por colágeno polianiônico ou como dupla camada com poli(cloreto de vinila) para reconstruc¸ão da parede abdominal. Rev Bras Eng Biomed 2001;17:67–78. Gordon PL, Huang R, Lord RC, Yannas IV. The far-infrared spectrum of collagen. Macromolecules 1974;7:954–6. Goormaghtigh E, Ruyssharet J-M, Raussens V. Evaluation of the information content in infrared spectra for protein secondary structure determination. Biophys J 2006;90:2946–57. Kielty CM, Sherratt MJ, Shuttleworth CA. Elastic fibers. J Cell Sci 2002;115:2817–28. Kumashiro KK, Ho JP, Niemczura WP, Keeley FW. Cooperativity between the hydrophobic and cross-linking domains of elastin. J Biol Chem 2006;281:23757–65. Mecham RP. Methods in elastic tissue biology: elastin isolation and purifications. Methods 2008;45:32–41. Mello MLS, Vidal BC. Anisotropic properties of toluidine bluestained collagen. Ann Histochim 1973;18:103–22. Mello MLS, Vidal BC. Analysis of the DNA Fourier transform-infrared microspectroscopic signature using an all-reflecting objective. Micron 2014;61:49–52. Popescu M-C, Vasile C, Craciunesco O. Structural analysis of some soluble elastins by means of FT-IR 2D IR correlation spectroscopy. Biopolymers 2010;93:1072–83.
Ribeiro JF, dos Anjos EHM, Mello MLS, Vidal BC. Skin collagen fiber molecular order: a pattern of distributional fiber orientation as assessed by optical anisotropy and image analysis. PLoS One 2013;8:e54724. Romhányi G. Specific topo-optical reactions of connective tissue elements and their ultraestructural interpretation. Connect Tissue Res 1986;15:13–6. Romhányi G, Deák GY, Fischer J. Aldhyde-bisulfite-toluidine blue (PTB) staining as a topo-optical reaction for demonstration of linear order of vicinal OH groups in biological structures. Histochemistry 1975;43:333–48. Silva DFT, Gomes ASL, Vidal BC, Ribeiro MS. Birefringence and second harmonic generation on tendon collagen following red linearly polarized laser irradiation. Ann Biomed Eng 2013;41:752–62. Uitto J, Paul JL, Brockley K, Pearce RH, Clark JG. Elastic fibers in human skin: quantification of elastic fibers by computerized digital image analyses and determination of elastin by radioimmunoassay of desmosine. Lab Invest 1983;49: 499–505. Vidal BC. The use of 8-anilino-naphtalene sulfate (ANS) for collagen and elastin histochemistry. J Histochem Cytochem 1978;26:196–201. Vidal BC. Aorta and tendon collagen reactivity to 8-anilinonaphtalene sulfate (ANS) and dansylchloride. Cell Mol Biol 1980;26:583–8. Vidal BC. From collagen type I solution to fibers with a helical pattern: a self-assembly phenomenon. CR Acad Sci Paris, Sci de la Vie 1995;318:831–6. Vidal BC. Image analysis of tendon helical superstructure using interference and polarized light microscopy. Micron 2003;34:423–32. Vidal BC. Using the FT-IR linear dichroism method for molecular order determination of tendon collagen bundles and nylon 6. Acta Histochem 2013;115:686–91. Vidal BC, Mello MLS. Collagen type I amide I infrared spectroscopy. Micron 2011;42:283–9. Vidal BC, Ghiraldini FG, Mello MLS. Changes in liver cell methylation status in diabetic mice affect its FT-IR characteristics. PLoS One 2014;9:e102295. Wilson EB, Wells AJ. The experimental determination of the intensities of the infra-red absorption bands. 1. Theory of the methods. J Chem Phys 1946;14:578–80. Yugami H, Shibayama Y, Matsuo S, Ishigame M, Shin S. Proton sites and defect-interactions in SrZrO3 single crystals studied by infrared absorption spectroscopy. Solid State Ionics 1996;85:319–22. Zoumi A, Lu X, Kassab GS, Tromberg BJ. Imaging coronary artery microstructure using second-harmonic and two-photon fluorescence microscopy. Biophys J 2004;87:2778–86.
Contents lists available at ScienceDirect
Acta Histochemica journal homepage: www.elsevier.de/acthis
Fluorescence, aggregation properties and FT-IR microspectroscopy of elastin and collagen fibers Benedicto de Campos Vidal ∗ Department of Structural and Functional Biology, Institute of Biology, University of Campinas (Unicamp), Rua Monteiro Lobato 255, CEP 013083-862 Campinas, SP, Brazil
a r t i c l e
i n f o
Article history: Received 15 July 2014 Received in revised form 18 August 2014 Accepted 19 August 2014 Keywords: Elastin Collagen type I Collagen bundles Hydrophobicity Birefringence FT-IR
a b s t r a c t Histological and histochemical observations support the hypothesis that collagen fibers can link to elastic fibers. However, the resulting organization of elastin and collagen type complexes and differences between these materials in terms of macromolecular orientation and frequencies of their chemical vibrational groups have not yet been solved. This study aimed to investigate the macromolecular organization of pure elastin, collagen type I and elastin–collagen complexes using polarized light DIC-microscopy. Additionally, differences and similarities between pure elastin and collagen bundles (CB) were investigated by Fourier transform-infrared (FT-IR) microspectroscopy. Although elastin exhibited a faint birefringence, the elastin–collagen complex aggregates formed in solution exhibited a deep birefringence and formation of an ordered-supramolecular complex typical of collagen chiral structure. The FT-IR study revealed elastin and CB peptide N H groups involved in different types of H-bonding. More energy is absorbed in the vibrational transitions corresponding to CH, CH2 and CH3 groups (probably associated with the hydrophobicity demonstrated by 8-anilino-1-naphtalene sulfonic acid sodium salt [ANS] fluorescence), and to CN, ıNH and ωCH2 groups of elastin compared to CB. It is assumed that the ␣-helix contribution to the pure elastin amide I profile is 46.8%, whereas that of the B-sheet is 20% and that unordered structures contribute to the remaining percentage. An FT-IR profile library reveals that the elastin signature within the 1360–1189 cm−1 spectral range resembles that of Conex–Toray aramid fibers. © 2014 Elsevier GmbH. All rights reserved.
Introduction Interactions between collagen and elastic fibers have been proposed in reports using histological, histochemical and histophysical methods (Braga-Vilela and Vidal, 2006; Aldrovani et al., 2007) and electron microscopy (Gibson et al., 1997; Kielty et al., 2002). Based on the work of Gibson et al. (1997), Kielty et al. (2002) have suggested that a protein might link elastic fibers to collagen. Hard extraction methods and histochemical methods, accompanied by simultaneous analysis with polarized light microscopy, have suggested a tight collagen–elastic fiber relationship in porcine pericardium and aorta (Braga-Vilela and Vidal, 2006; Aldrovani et al., 2007). The amount of elastin isolated from some tissues has been surveyed with the following results: 30–43% in aorta, 2–5% in skin, 5–8% in ligamentum nuchae (Ayer, 1964), and ∼2 ± 1% in the dermis
∗ Corresponding author. E-mail address: [email protected] http://dx.doi.org/10.1016/j.acthis.2014.08.007 0065-1281/© 2014 Elsevier GmbH. All rights reserved.
of healthy human skin, as measured using morphometric techniques (Uitto et al., 1983). Regarding the anatomic structure of elastic fibers in the skin, three forms have been described: filaments, a natural network and 20 nm-beaded fibrils (Ayer, 1964). Images of morphological aspects along with biomechanical properties have been reported for elastic fibers and collagen in arteries in a study that combined second-harmonic generation (SHG) signals, two-photon excited fluorescence and multiphoton microscopy which showed the stress–strain distribution of these fibers in the arterial wall (Zoumi et al., 2004). One of the methods proposed for extraction of elastic fibers is based on boiling in diluted alkali. However, “by extending the extraction time it is found that considerable destruction of elastica occurs” (Ayer, 1964 – review). Another recommended method involves boiling the sample in dilute acetic acid; in this case, no changes in the fibrillar-meshed architecture of the elastica have been reported (Ayer, 1964; Eyre et al., 1984; Mecham, 2008). Some optical methods have indicated that elastic fibers can exhibit birefringence; this anisotropic optical property has been attributed to the oriented distribution of the filamentous elements
1360
B.d.C. Vidal / Acta Histochemica 116 (2014) 1359–1366
of elastin (Romhányi et al., 1975; Romhányi, 1986). A compositional complexity of molecules interacting with elastin includes type VIII collagen and emilins, as well as the microfibril-associated glycoprotein 1 (MAGP-1), which is an integral microfibril molecule that is important for the structural integrity of the elastic fibers (Kielty et al., 2002). These reports indicate that the composition and structure of the elastic fibers is complex, and that the biological role of many of their molecular components is far from being clear. A contribution to a comparative study of the macromolecular organization of elastin and collagen fibers and their complexes could be done by evaluating the optical anisotropic properties of these structures in separate and in the aggregates these materials could form in solution, under dialysis conditions. In addition, comparison of Fourier transform-infrared (FT-IR) microspectroscopic signatures between elastin and collagen bundles could probably provide information on differential frequencies of their most important vibrational groups. In the present study, pure elastin, collagen type I and elastin–collagen complexes, formed in solution, were studied using polarized light microscopy. Additionally, differences between elastin and collagen bundles were investigated at the level of their vibrational groups, as studied by FT-IR. Hydrophobicity by ANS fluorescence, which could add information for the interpretation of FT-IR results, was also searched in pure elastin. Material and methods
Elastin fluorescence analysis Some elastin smears were treated with 0.1% 8-anilino-1naphtalene sulfonic acid sodium salt (ANS) (T484 Kodak, Rochester, NY, USA) in a non-polar solvent, butanol, for 30 min in the dark; next, unbound ANS was removed from the smears using pure butanol (Vidal, 1978, 1980). Then, the preparations were cleared in xylene and mounted on slides in Nujol mineral oil (M3516; SigmaAldrich). The ANS-stained preparations were examined under a Carl Zeiss Axiophot 2 photomicroscope (Oberkochen/Munich, Germany) equipped for epifluorescence and using an HBO-mercury short-arc 103 W/2 as UV source and a 365 nm filter for the excitation light; FT395 and LP-420 nm filters were also employed. For image analysis and photomicrographs, a Zeiss Axiocam HRc camera and Kontron KS400-3 software were used. Optical anisotropy The identification of birefringence characteristics in unstained elastin fragments and elastin–collagen type I complexes was performed using an Olympus BMX 51 polarization microscope (Olympus Corp., Tokyo, Japan) connected to a computer and equipped with optical devices for differential interference contrast microscopy (DIC). Optical anisotropy documentation was carried out with Image-Pro-Plus 6.3 software (Media Cybernetics, Bethesda, MD, USA).
Elastin and collagen type I preparation FT-IR microspectroscopy Pure elastin obtained from bovine nuchal ligament (SigmaAldrich, St. Louis, MO, USA) was used. Small fragments of this material were suspended in distilled water and smears were prepared from that suspension. The smears were dried following three stages: first, they were left in a refrigerator for 24 h, then they were left at 37 ◦ C for 3 h, and finally they were dried at 60 ◦ C for 2 h. Collagen type I was extracted from the rat tail tendons of the experimental animals previously used to study collagen bundle supraorganization in the skin (Ribeiro et al., 2013). The tails were washed extensively, and their skin was removed. The collagen bundles, which were visible in the external surface of the tail, were removed and washed in Milli-Q water from a Millipore Direct-Q3System (Molheim, France). After three washes, well-separated and clean collagen bundles were immersed in a 3% acetic acid solution. This solution remained in the refrigerator for 2–3 days for complete dissolution of the collagen fibers. Next, a careful filtration was performed. Subsequently, the collagen fibers were reconstituted by adding 20% NaCl solution to the collagen solution at final concentration of 10%. Immediately afterwards, visibly precipitated fibers, showing a transparent aspect, were left in the refrigerator; white floating reconstituted fibers were dissolved again in a 3% acetic acid solution and kept in the refrigerator until they were completely dissolved. Additional acetic acid solution was added when necessary. After complete fiber dissolution, a careful filtration was performed. In an attempt to form a complex between elastin and collagen, 150 mg of elastin powder was suspended in 3 mL of the collagen acetic acid solution and dialyzed in Milli-Q water in the refrigerator, until reconstituted collagen formed. In this case, reconstitution was performed without using NaCl. The dialysis was interrupted as soon as reconstituted collagen fibers appeared. At this step, a collagen–elastin complex was evident. All of the protocols involving animal care and use for collagen attainment were approved by the Committee for Ethics in Animal Experimentation of the University of Campinas (CEAE/IB/Unicamp) (protocol 2700-1) and were in accordance with the Guidelines of the Canadian Council on Animal Care.
FT-IR spectral profiles were obtained for sections of bovine collagen bundles (CB) and elastin smears. CB sections were prepared from bovine flexor tendons acquired from a slaughterhouse, as previously reported (Vidal, 2013). Briefly, tendon fragments cut parallel to the tendon axis and fixed in 4% paraformaldehyde at pH 7.4 under vacuum were embedded in Hystosec medium (Merck, Darmstadt, Germany) to obtain 10 m thick sections which were dewaxed in xylene, mounted on slides and dried at 37 ◦ C. Because an all-reflecting objective (ARO) was used for the microspectroscopic examination, the materials were mounted on the surface of thick gold-covered reflective glass slides (Mello and Vidal, 2014). The Illuminat IR IITM microspectrometer (Smiths Detection, Danbury, CT 06810, USA) equipped with a liquid-nitrogen-cooled mercury-cadmium-telluride detector, ARO (16×) and Grams/AI 8.0 software (Thermo Electric, Waltham, MA, USA), and coupled to an Olympus BM51 microscope was used for FT-IR microspectroscopy. The performance of the equipment was validated using a low signal-to-noise ratio (7929:1) (Vidal and Mello, 2011). The absorbances of the samples and background were obtained using 64 scans for each preparation with a spectral resolution of 4 cm−1 . The spectral absorption profiles were determined at frequencies ranging from 4000 cm−1 to 650 cm−1 . A plus-zero baseline was acquired, and the FT-IR spectrum was normalized using the highest absorbance peak as a reference. After 10 spectral profiles were obtained, an average profile was produced using the Grams software (Vidal and Mello, 2011; Vidal, 2013; Mello and Vidal, 2014). The area of the band peaks was determined using the numerical integral statistics of the Grams software. Peak fitting and peak fitting estimate procedures using a Gaussian function and low sensitivity level were also applied to specific spectral regions of absorption band peaks, using the Grams software, enabling the area of selected bands to be calculated (Mello and Vidal, 2014). An absorbance ratio, previously proposed to determine the integrity of the collagen triple helices or the relative amount of native collagen in a sample (Gordon et al., 1974; Goissis et al., 2001), was also estimated.
B.d.C. Vidal / Acta Histochemica 116 (2014) 1359–1366
1361
Fig. 1. ANS-stained elastin fragments with various dimensions exhibit blue fluorescence. Bar = 50 m.
Results
greater accessibility of the hydrophobic groups to ANS in isolated elastin.
Fluorescence Optical anisotropy The ANS-stained elastin fragments revealed a blue fluorescence (Fig. 1) which was deeper in comparison with that reported when ANS was used in histochemical studies (Vidal, 1978, 1980; Braga-Vilela and Vidal, 2006). This result is likely due to the
The smaller elastin fragments exhibited poor birefringence intensity at their surface. Birefringence slightly increased in the aggregates of fibrillar elastin (Fig. 2), due to fragment overlapping
Fig. 2. Elastin fragments examined under a polarized light microscope showed weak birefringence at their surfaces. When the fragments were stacked, their birefringence combined to increase the brilliance intensity. Bar = 50 m.
1362
B.d.C. Vidal / Acta Histochemica 116 (2014) 1359–1366
Fig. 3. Very thin collagen type I fibers reconstituted from acetic acid solutions, using dialysis in distilled water. In the segments with crossed fibers, the birefringence is internally compensated, and a black image results. The differences in brightness intensity are due to differences in fiber aggregation. Bar = 50 m.
and the relative thickness increase. After birefringence compensation in the elastin, it was possible to determine that the birefringence sign was positive and that this phenomenon was non-homogeneously distributed along the axis of the in tandem fibril aggregation. Pure reconstituted rat collagen thin fibers obtained from the collagen type I acetic acid solutions showed a deep positive
birefringence (Fig. 3), similar to that reported for bovine collagen (Vidal, 1995). Adding elastin to the collagen type I acetic acid solution increased the initial viscosity of the collagen solution. Microscopic examination of the elastin–collagen mixture revealed some regions containing only reconstituted fibers (Fig. 3) and others containing the elastin–collagen complex (Figs. 4 and 5). A visually
Fig. 4. Binding of elastin to acid-soluble collagen, which were dialyzed together in distilled water. After dialysis, a highly birefringent elastin–collagen complex formed. The areas of crossed fibers show internal birefringence compensation and dark areas. Collagen I reconstituted fibers are also seen. Bar = 30 m.
B.d.C. Vidal / Acta Histochemica 116 (2014) 1359–1366
1363
Fig. 5. Panel containing images of the collagen–elastin complex obtained at four different focal distances ((A)–(D)) using DIC. In (A) the image is shown at the most superficial focal level; (B) is the next level, and the focal depth continues to increase (C) until (D) which has the deepest focal distance. The helical structure of the supramolecular organization involved is shown. In this case, interference colors due to the birefringence phenomenon are observed because DIC is based on polarized light microscopy principles. Bar = 50 m.
detectable increase in birefringence brightness was verified in the elastin–collagen complex fibers from acetic acid solutions spread on glass slides. Some elastin–collagen complex aggregates appeared with the supramolecular organization of intertwined helical fibers typical for collagen bundles, as demonstrated by DIC-polarized light microscopy analysis (Fig. 5). FT-IR microspectroscopy Differences in the distribution of band peak frequencies were detected in the FT-IR spectral signature of elastin in comparison to CB (Fig. 6). Visual inspection of the FT-IR spectra indicates a general shift of the peaks to a lower frequency in elastin compared to CB. The highlighted heights were more evident when considering a subtraction spectrum obtained from differences in absorbance values (Fig. 6, blue line). If using the numerical integral statistics provided by the Grams software in the 3669–3127 cm−1 spectral window, 292 area points were revealed for CB and 297 points were revealed for elastin, corresponding to areas of 213 and 212 units, respectively. Thus, despite the spectral profile appearance based on absorbance values with varying wave numbers (Fig. 6), the elastin did not show a significantly smaller area in this spectral window. This is because, for the numerical integral statistics calculation, the Grams software requires that a new baseline is considered, with the left and right edges at 3669 cm−1 and 3127 cm−1 , respectively, but at the height of the respective absorbance values. In the elastin case, the left edge corresponds to the absorbance value of 0.13
and the right edge corresponds to the absorbance value of 0.81, whereas in the collagen case, the left and right edges correspond to absorbance of 0.02 and 0.60, respectively. Five peaks were selected using Grams’ peak fitting procedure applied to both elastin and CB profiles at the spectral window of ∼3000–2940 cm−1 . For elastin, the total area comprised in the spectral profile from 2977 to 2951 cm−1 yielded a value of 5.20 units, and the maximal area, corresponding to the band peak at 2967 cm−1 , was equal to 1.15 units. For CB, the profile region between 3007 and 2938 cm−1 yielded a total area of 1.96 units, and the maximal area, corresponding to the band peak at 2959 cm−1 , was equal to 0.68 units. Thus, elastin showed noticeably higher total area and peak maximal area at this spectral window. Band peaks representative of amide A were revealed at 3286 cm−1 and 3325 cm−1 for elastin and CB, respectively. Regarding the amide I spectral region, it was found to be represented in the 1725–1589 cm−1 range for elastin and in the 1706–1657 cm−1 range for CB (Fig. 6). The numerical integral statistics obtained for this spectral window after peak fitting revealed an area of 34 units for elastin and 9 units for CB. After peak fitting, band peaks at 1656 cm−1 , 1648 cm−1 , 1639 cm−1 , 1628 cm−1 and 1620 cm−1 were detected for elastin. The amide I spectrum for CB yielded two visually distinct peaks (Fig. 6). After peak fitting procedure, four band peaks were obtained for CB, such that the two most representative ones appeared positioned at 1689 cm−1 and 1678 cm−1 and two smaller band peaks were positioned at 1665 cm−1 and 1658 cm−1 . The participation of each peak in the vibrational transition moment of the amide I was evaluated based
1364
B.d.C. Vidal / Acta Histochemica 116 (2014) 1359–1366
Fig. 6. FT-IR spectra obtained for collagen bundles (CB) (black line) and elastin (red line). The profile that represents the absorption difference between elastin and CB is also presented (blue line). X axis, wavenumbers in cm−1 ; Y axis, absorbances.
on the approximate fractional percentile of each band peak. Based on a previously reported study on FT-IR linear dichroism for the same CB material, one could hypothesize that the band peaks at 1689 cm−1 and 1678 cm−1 are related to C O (∼39%) and C N (∼33%) vibronic transition moments, respectively (Vidal, 2013). The band peaks positioned at 1665 cm−1 and 1658 cm−1 occupied areas corresponding to ∼14% and ∼6%, respectively, of the total area. Based on present spectral results for CB after peak fitting, a ratio equal to 0.75 was calculated from absorbances obtained at 1248 cm−1 /1453 cm−1 . For reconstituted collagen fibers (as those shown in Fig. 4) analyzed for FT-IR (spectrum not shown), such a ratio was equal to 1.10, that is, close to that reported for collagen in natura (1.38), evaluated at 1235 cm−1 /1450 cm−1 (Gordon et al., 1974). The lower ratio value found for CB in the present study (0.75) may be due to some degree of denaturation, introduced by paraffin embedding at 60 ◦ C (Mello and Vidal, 1973), a fact that is not supposed to occur in the process of preparing pure collagen to obtain reconstituted fibers (Vidal, 1995). Indeed, according to Gordon et al. (1974), the ratio obtained for gelatin cast at 60 ◦ C is lower (0.59) than that obtained for collagen in natura. The ratio evaluated from absorbance obtained at 1248 cm−1 /1453 cm−1 , comparable to that obtained at 1235 cm−1 /1453 cm−1 providing information on integrity of the collagen fiber triple helix (Goissis et al., 2001), would not be properly applicable to elastin because this molecule does not exhibit a triple helix conformation. Maybe, in the case of elastin, such a ratio value is promoted by ␣-helix and -sheet content contribution. Another IR spectral difference between elastin and collagen occurred in the range of 1350–1200 cm−1 , which is approximately the frequency range noted in the literature for CN, ıNH and ωCH2 groups (Fabian and Mäntele, 2002). Peak fitting revealed four peaks for elastin: from the center of the peak at 1331 cm−1 to the center of the peak at 1247 cm−1 , the numerical integral area was equal to 12
units, and the highest peak was positioned at the 1247 cm−1 band. Regarding CB, the numerically integrated area was equal to 9 units, and the highest partial area was positioned at the 1247 cm−1 band.
Discussion Under present experimental conditions, the elastin fibers revealed a predominantly unordered structure, which was assumed from the faint birefringence restricted to their surface. On the other hand, the intense birefringence in collagen fibers increased in elastin–collagen type I aggregates. In addition, these aggregates acquire a supramolecular organization of intertwined helical fibers, typical for collagen fibers. Present findings support the hypothesis that collagen type I can complex with elastin and induce elastin to adopt a supramolecular organization with a helical (chiral) structure like that previously reported for collagen (Vidal, 1995, 2003, 2013; Silva et al., 2013). An intimate relationship between CB and elastic fibers has already been suggested in porcine pericardium and aorta studies using ANS and dansylchloride staining (Braga-Vilela and Vidal, 2006). It is hypothesized here that collagen may bind to elastin through a lectin-type recognition reaction. Thus, basic differences between elastin and collagen fibers are not caused only by amino acid differences. As demonstrated by present optical anisotropy investigation, elastin does not display the same degree of molecular order as collagen fibers. This fact is an indication that their infrared spectral characteristics could also differ. It should be emphasized that the present work addresses very complex supramolecular architectures. Then, in the FT-IR microspectroscopy context, it is necessary to consider Eglinton’s (1970) words: In any case, the spectrum of such a complicated
B.d.C. Vidal / Acta Histochemica 116 (2014) 1359–1366
molecule is not interpretable in complete detail and one would not expect to arrive at the structure from the infrared spectrum alone. Nevertheless, regarding the 3700–3130 cm−1 IR spectral region, and considering the report by Popescu et al. (2010 – p. 1076), the following assignments can be proposed: 3560 cm−1 , stretching vibrations of free OH groups; 3390 cm−1 , stretching vibrations of H-bonded OH groups; and 3290 and 3195 cm−1 , free N H and H-bonded N H groups, respectively. The band related to vibrations of the OH groups is recognized by its broadness and low intensity. The OH groups are involved in two hydrogen-bonding modes. The occurrence of amide A at 3286 and 3325 cm−1 in the spectral profile of elastin and CB, respectively, indicated involvement of N H peptide groups in different H-bonding types. The changes currently found at the level of the OH and NH groups for amide I and II demonstrate the importance of H-bonding in the formation and stabilization of elastic structures. The assignments proposed for the peaks detected at the 2977–2951 cm−1 spectral range are CH, CH2 , and CH3 groups. In the case of elastin, these groups are mostly present in leucine, isoleucine and asparagine and caused the hydrophobicity that was demonstrated by the intense fluorescence due to the ANS binding. For the interpretation of results based on infrared band peak areas at this spectral window, the significance of infrared energy absorption was considered as “quantized energy into the fundamental vibration modes of the molecules under study” (Eglinton, 1970). There are reports establishing for different materials that the area under an absorption band peak in IR spectroscopy can be related to specific vibrational absorbed energy involved (Wilson and Wells, 1946; Eglinton, 1970; Yugami et al., 1996; Vidal et al., 2014). Present data thus support the hypothesis that the vibrational transitions of the CH, CH2 , and CH3 groups absorbed more energy in elastin than in CB. Similarly, regarding the spectral window of 1350 cm−1 to 1200 cm−1 , the vibrational transitions of the CN, ıNH and ωCH2 groups (Fabian and Mäntele, 2002) absorbed more energy in elastin compared to CB. According to Goormaghtigh et al. (2006), and considering several other authors, the process of “. . .curve fitting requires a series of subjective decisions that can dramatically affect both the result and the interpretation. Furthermore, curve fitting has a tendency to overestimate the -sheet content of primarily helical proteins and routinely finds 15–20% -sheet content for proteins that actually have none.” Moreover, “among the 3201 frequencies tested . . . only three frequencies contain all the information that is available to predict the secondary structure of the proteins” (Goormaghtigh et al., 2006). Despite these authors’ considerations, present peak fitting data allowed to associate five selected band peak areas to data from the literature that inform on the participation of ␣-helix and -sheet-type secondary structures as well as unordered molecular structures in the elastin amide I profile (Fabian and Mäntele, 2002; Fabian and Naumann, 2004; D’Souza et al., 2006). Based on 13 C NMR experiments, Kumashiro et al. (2006) have already reported that hydrophobic and ␣-helical cross-linking domains cooperate in the elastin folding and that “hinge” sequences (“unordered”?) occur in the middle of each cross-linking domain. Thus, in an attempt to quantify the ␣-helix and -sheet-type secondary structures in the elastin amide I IR spectral profile, it is assumed that the largest band peak area at 1655 cm−1 and the area under the band peak at 1648 cm−1 summed, correspond to a 46.8% contribution of ␣-helix. The area for the band peak positioned at 1639 cm−1 indicates a 20% participation of -sheet in this profile. The remaining area, corresponding to band peaks at 1628 cm−1 and 1620 cm−1 , is likely assigned to a contribution to this amide I profile of unordered structures that would also play a part in the poor birefringence phenomenon observed in the elastin fragments. When the Grams software library was consulted to find substances that presented FT-IR characteristics similar to those of
1365
elastin, 20 compounds were found, including 5 types of aramid fibers and 14 types of nylon fibers. The fiber spectrum that was most similar to that of elastin, within the 1360–1189 cm−1 range, was that of the Conex–Toray aramid. This polyamide contains a large amount of hydrogen bonds in the amide groups of its aromatic rings. No other particular spectral characteristics, besides those reflecting generic spectral polyamide properties, were found when comparing elastin, aramids and nylon types. On the other hand, the FT-IR spectral profile of CB has been previously found to be comparable to that of nylon 6 fibers, revealing vibronic molecular moments distributed parallel to their long axis and assigning CN and ıNH groups to the axis of collagen and nylon 6 fibers (Vidal and Mello, 2011; Vidal, 2013).
Conclusions Present data support the hypothesis that elastin can link to collagen type I, forming an ordered supramolecular complex. Despite the not so ordered structural organization in pure elastin as that in collagen fibers, when elastin–collagen aggregates are obtained after dialysis, a complex is generated with the supramolecular organization of intertwined helical fibers typical of collagen chiral structure. A detailed comparative FT-IR study of the elastin and CB revealed amide A at 3286 cm−1 and 3325 cm−1 , respectively, indicating that their peptide N H groups were involved in different types of H-bonding. The amide I band in elastin and CB is also identified at different IR frequency range (1756–1589 cm−1 , elastin; 1689–1678 cm−1 , CB). Based on FT-IR data for the 2977–2951 cm−1 spectral range, it is proposed that more energy is absorbed in the vibration transitions corresponding to CH, CH2 and CH3 groups in elastin in comparison to CB. These groups, mostly present in leucine, isoleucine and asparagine, are probably responsible for the hydrophobicity demonstrated in the elastin with ANS fluorescence. Vibrational transitions at the 1350–1200 cm−1 region, that are assumed to be assigned to the CN, ıNH and ωCH2 groups, are also proposed to absorb more energy in elastin compared to CB. The sum of the areas under the band peaks at 1655 cm−1 and 1648 cm−1 represents the ␣-helix contribution (46.8%) to the elastin amide I profile. The area under the band peak at 1639 cm−1 indicates the -sheet participation (20%) in this profile, and the areas corresponding to band peaks at 1628 cm−1 and 1620 cm−1 are assigned to the contribution of unordered structures to this profile, presumably correlated with the poor birefringence phenomenon in the elastin fragments. Consulting the FT-IR profile library provided by the Grams/AI microspectroscopic software, elastin keeps similarity to Conex–Toray aramid fibers within the 1360–1189 cm−1 IR spectral range.
Acknowledgments The author is indebted to Dr. Maria Luiza S. Mello for discussing and reviewing the manuscript. The support of Fundac¸ão de Amparo à Pesquisa do Estado de São Paulo (FAPESP), Brazil (grants no. 2003/04597-0, 2007/058251-8 and 2013/11078-0) is gratefully acknowledged.
References Aldrovani M, Guaraldo AMA, Vidal BC. Fluorescence, birefringence and confocal microscopy of the abdominal aorta from nonobese diabetic (NOD) mice. Acta Histochem 2007;109:248–54.
1366
B.d.C. Vidal / Acta Histochemica 116 (2014) 1359–1366
Ayer JP. Elastic tissue. In: Hall DA, editor. International review of connective tissue research, vol. 2. New York and London: Academic Press; 1964. p. 33–100. Braga-Vilela AS, Vidal BC. Identification of elastic fibers and lamellae in porcine pericardium and aorta by confocal, fluorescence and polarized light microscopy. Acta Histochem 2006;108:125–32. D’Souza AJ, Hart DS, Middaugh CR, Gehrke SH. Characterization of the changes in secondary structure and architecture of elastin–mimetic triblock polypeptides during thermal gelation. Macromolecules 2006;39:7084–91. Eglinton G. Application of infrared spectroscopy to organic chemistry. In: Scheinmann F, editor. An introduction to spectroscopy methods for the identification of organic compounds, vol. 1. Oxford: Pergamon Press; 1970. p. 129–41. Eyre DR, Paz MA, Gallop PM. Cross-linking in collagen and elastin. Ann Rev Biochem 1984;53:717–48. Fabian H, Mäntele W. Infrared spectroscopy of proteins. In: Chalmers JM, Griffiths PR, editors. Handbook of vibrational spectroscopy. Chichester: Wiley; 2002. p. 3399–425. Fabian H, Naumann D. Methods to study protein folding by stoppedflow FT-IR. Methods 2004;34:28–40. Gibson MA, Kumaratilake JS, Cleary EG. Immunohistochemical and ultrastructural localization of MP78/70 (ig-h3) in extracellular matrix of developing and mature bovine tissues. J Histochem Cytochem 1997;45:1683–97. Goissis G, Suzigan S, Parreira DR, Raymundo SRO, Chaves H, Hussain KMK. Malhas de polipropileno recobertas por colágeno polianiônico ou como dupla camada com poli(cloreto de vinila) para reconstruc¸ão da parede abdominal. Rev Bras Eng Biomed 2001;17:67–78. Gordon PL, Huang R, Lord RC, Yannas IV. The far-infrared spectrum of collagen. Macromolecules 1974;7:954–6. Goormaghtigh E, Ruyssharet J-M, Raussens V. Evaluation of the information content in infrared spectra for protein secondary structure determination. Biophys J 2006;90:2946–57. Kielty CM, Sherratt MJ, Shuttleworth CA. Elastic fibers. J Cell Sci 2002;115:2817–28. Kumashiro KK, Ho JP, Niemczura WP, Keeley FW. Cooperativity between the hydrophobic and cross-linking domains of elastin. J Biol Chem 2006;281:23757–65. Mecham RP. Methods in elastic tissue biology: elastin isolation and purifications. Methods 2008;45:32–41. Mello MLS, Vidal BC. Anisotropic properties of toluidine bluestained collagen. Ann Histochim 1973;18:103–22. Mello MLS, Vidal BC. Analysis of the DNA Fourier transform-infrared microspectroscopic signature using an all-reflecting objective. Micron 2014;61:49–52. Popescu M-C, Vasile C, Craciunesco O. Structural analysis of some soluble elastins by means of FT-IR 2D IR correlation spectroscopy. Biopolymers 2010;93:1072–83.
Ribeiro JF, dos Anjos EHM, Mello MLS, Vidal BC. Skin collagen fiber molecular order: a pattern of distributional fiber orientation as assessed by optical anisotropy and image analysis. PLoS One 2013;8:e54724. Romhányi G. Specific topo-optical reactions of connective tissue elements and their ultraestructural interpretation. Connect Tissue Res 1986;15:13–6. Romhányi G, Deák GY, Fischer J. Aldhyde-bisulfite-toluidine blue (PTB) staining as a topo-optical reaction for demonstration of linear order of vicinal OH groups in biological structures. Histochemistry 1975;43:333–48. Silva DFT, Gomes ASL, Vidal BC, Ribeiro MS. Birefringence and second harmonic generation on tendon collagen following red linearly polarized laser irradiation. Ann Biomed Eng 2013;41:752–62. Uitto J, Paul JL, Brockley K, Pearce RH, Clark JG. Elastic fibers in human skin: quantification of elastic fibers by computerized digital image analyses and determination of elastin by radioimmunoassay of desmosine. Lab Invest 1983;49: 499–505. Vidal BC. The use of 8-anilino-naphtalene sulfate (ANS) for collagen and elastin histochemistry. J Histochem Cytochem 1978;26:196–201. Vidal BC. Aorta and tendon collagen reactivity to 8-anilinonaphtalene sulfate (ANS) and dansylchloride. Cell Mol Biol 1980;26:583–8. Vidal BC. From collagen type I solution to fibers with a helical pattern: a self-assembly phenomenon. CR Acad Sci Paris, Sci de la Vie 1995;318:831–6. Vidal BC. Image analysis of tendon helical superstructure using interference and polarized light microscopy. Micron 2003;34:423–32. Vidal BC. Using the FT-IR linear dichroism method for molecular order determination of tendon collagen bundles and nylon 6. Acta Histochem 2013;115:686–91. Vidal BC, Mello MLS. Collagen type I amide I infrared spectroscopy. Micron 2011;42:283–9. Vidal BC, Ghiraldini FG, Mello MLS. Changes in liver cell methylation status in diabetic mice affect its FT-IR characteristics. PLoS One 2014;9:e102295. Wilson EB, Wells AJ. The experimental determination of the intensities of the infra-red absorption bands. 1. Theory of the methods. J Chem Phys 1946;14:578–80. Yugami H, Shibayama Y, Matsuo S, Ishigame M, Shin S. Proton sites and defect-interactions in SrZrO3 single crystals studied by infrared absorption spectroscopy. Solid State Ionics 1996;85:319–22. Zoumi A, Lu X, Kassab GS, Tromberg BJ. Imaging coronary artery microstructure using second-harmonic and two-photon fluorescence microscopy. Biophys J 2004;87:2778–86.
