In situ structural studies during denaturation of natural and synthetically crosslinked collagen using synchrotron SAXS
aLeather and Shoe Research Association of New Zealand, PO Box 8094, Palmerston North 4446, New Zealand, and bChemical Engineering Building, National Tsing Hua University, No. 101, Section 2, Guangfu Road, East District, Hsinchu City 300, Taiwan
*Correspondence e-mail: email@example.com
Collagen is an important biomacromolecule, making up the majority of the extracellular matrix in animal tissues. Naturally occurring crosslinks in collagen stabilize its intermolecular structure in vivo, whereas chemical treatments for introducing synthetic crosslinks are often carried out ex vivo to improve the physical properties or heat stability of the collagen fibres for applications in biomaterials or leather production. Effective protection of intrinsic natural crosslinks as well as allowing them to contribute to collagen stability together with synthetic crosslinks can reduce the need for chemical treatments. However, the contribution of these natural crosslinks to the heat stability of collagen fibres, especially in the presence of synthetic crosslinks, is as yet unknown. Using synchrotron small-angle X-ray scattering, the in situ role of natural and synthetic crosslinks on the stabilization of the intermolecular structure of collagen in skins was studied. The results showed that, although natural crosslinks affected the denaturation temperature of collagen, they were largely weakened when crosslinked using chromium sulfate. The development of synergistic crosslinking chemistries could help retain the intrinsic chemical and physical properties of collagen-based biological materials.
Keywords: in situ SAXS; collagen structure; crosslinks; denaturation.
During the assembly of collagen molecules to fibrils in vivo, enzymatic crosslinking is known to affect the mechanical properties of biological tissues by preventing the slippage of molecules under load (Lucero & Kagan, 2006). The crosslinking is induced by lysyl oxidase which deaminates the ∊-amino groups of telopeptidyl lysine and hydroxylysine residues to form aldehydes (Knott & Bailey, 1998). These aldehydes then form Schiff base linkages by condensing with lysine or hydroxylysine residues in the helical region of adjacent collagen molecules (Knott & Bailey, 1998). To date, five types of crosslinks have been identified in bovine skins (Avery & Bailey, 2008; Naffa et al., 2016), i.e. divalent dehydro-hydroxylysinonorleucine (deH-HLNL) and dehydro-dihydroxylysinonorleucine (deH-DHLNL), trivalent crosslinked pyridinoline (Pyr) and histidinohydroxylysinonorleucine (HHL), and tetravalent crosslinked dehydro-histidinohydroxymerodesmosine (deH-HHMD), although its presence in vivo is disputed (Naffa et al., 2019; Avery & Bailey, 2008). However, amongst these five crosslinks, deH-HLNL, deH-DHLNL and deH-HHMD contain a C=N bond and are labile to acidic conditions (Avery & Bailey, 2008). The chemical modification of hides and skins during leather processing or even collagen extraction can lead to a loss of labile crosslinks, lessening their intrinsic mechanical and heat stability (Avery & Bailey, 2008).
Synthetic crosslinkers are often introduced after the removal of intrinsic crosslinks to stabilize collagen and modify its structure as required (Ashwin & McDonnell, 2010). The mechanism by which synthetic crosslinkers cause an increment in the heat stability of collagen can be explained by the dehydration of the collagen fibres (Miles et al., 2005). Chromium sulfate can covalently bind to the carboxylate groups in collagen and, potentially, form linkages in the collagen matrix (Covington, 2009). It is commonly used to convert animal skins to leather by imparting the material with a higher denaturation temperature and characteristic organoleptic properties (Covington, 2009). However, stringent environmental legislations and increasing pressure by regulatory agencies on the disposal of chromium wastes have led researchers to use less chrome or develop chrome-free alternatives (Lyu et al., 2018; Beghetto et al., 2019; Zhou et al., 2018; Ding et al., 2019). Similarly, biomaterials stabilized by synthetic crosslinkers such as glutaraldehyde can improve the denaturation temperature and protect against enzymatic degradation (Damink et al., 1995; Yang et al., 2017; Skopinska-Wisniewska et al., 2016), although there is the potential to liberate toxic aldehydes during the degradation process (Eybl et al., 1989). By retaining the natural crosslinks that are formed in vivo to contribute to the collagen heat stability, the need for additional crosslinking reagents ex vivo can be reduced.
Using sodium borohydride, the C=N bonds in acid-labile natural crosslinks can be reduced and stabilized (Eyre et al., 1984). However, it is essential to understand the interplay between natural and synthetic crosslinks and their contribution to the heat stability of collagen structure, which still remains unclear. Using synchrotron-based small-angle X-ray scattering (SAXS), changes in the collagen structure during crosslinking and denaturation can be monitored in situ (Buchanan et al., 2019). Also, liquid chromatography–mass spectrometry (LC–MS) can be used to quantify the natural crosslinks present in skin collagen (Naffa et al., 2016). In this work, we studied the effect of natural and synthetic crosslinks on the stability of collagen using a combination of in situ synchrotron SAXS and differential scanning calorimetry (DSC).
2. Materials and methods
2.1. Materials and sample preparation
Raw bovine skins were obtained from a local tannery in New Zealand. Commercial reagents including sodium borohydride, sodium hydrosulfide, sodium sulfide, calcium hydroxide, ammonium chloride, hydrogen peroxide, sulfuric acid, sodium chloride, magnesium oxide and Chromosal B (25% Cr2O3, 33% basicity) were used without further modification.
Raw bovine skins were treated with 3 wt% (wt% of the skin, same below) sodium borohydride at 25°C for 24 h, termed `reduced control (R-CT)', whereas the negative controls without borohydride treatment were termed `non-reduced control (N-CT)'. The skins were then treated overnight at 25°C in a solution of 2.5 wt% sodium sulfide, 1.5 wt% sodium hydrosulfide and 1.5 wt% calcium hydroxide and rinsed with water to remove hair. Then, the skins were neutralized with a solution of 2.5 wt% ammonium chloride and 0.5 wt% hydrogen peroxide for 2 h, followed by rinsing with water. Samples collected at this stage are termed `dehaired' (N-DH and R-DH). The skins were then acidified using 1.2 wt% sulfuric acid with 8 wt% sodium chloride to pH = 3.0, followed by sample collection as `acidified' (N-AC and R-AC). Finally, the skins were treated using 8 wt% chromium sulfate overnight at 25°C, followed by the addition of 0.4 wt% magnesium oxide to adjust the pH to 4.0, termed `crosslinked' (N-CR and R-CR). All samples were kept at 4°C prior to further analysis.
2.2. In situ SAXS analysis
In situ isotropic SAXS measurements were performed on beamline 23A1 at the National Synchrotron Radiation Research Centre (NSRRC) in Hsinchu, Taiwan. Experiments were carried out in the standard beamline temperature-controlled cell with the X-ray beam perpendicular to the skin surface. Samples 0.5 cm × 0.5 cm × 1 mm (L × W × H) in size were cut out and loaded onto the cell and sealed in polyimide film to prevent sample dehydration. Measurements of 30 s with an energy of 15 keV were performed. Scattered radiation was collected using a Pilatus 1M detector located at a distance of 3.233 m from the sample.
The scattered intensity I(q) is presented as a function of scattering vector q, with q = 4π sin(θ/2)/λ, where θ is the angle between the incident and scattered radiation. The acquired SAXS scattering patterns were radially integrated into 1D plots, and the diffraction peaks were fitted to calculate the d-spacing following a previously reported approach (Zhang et al., 2018; Ingham et al., 2009). The diffraction peaks are modelled as Gaussians,
where An is the area of peak n, d is the D-period (in Å), and w(q) is the width, expressed as w(q) = a + bq where a and b are fitted parameters. During the temperature increase, the onset temperatures of D-period (Tonset) are calculated to indicate the beginning of structural events.
2.3. LC–MS analysis
Natural crosslinks of bovine skins were analysed based on a previously reported method (Naffa et al., 2016), using a Cogent Diamond Hydride HPLC column (150 mm × 2.1 mm, particle size 2.2 µm, pore size 100 Å; Microsolv Technology, Leland, NC, USA) and Q-Exactive hybrid quadrupole-orbitrap high-resolution accurate-mass spectrometry (Thermo Fisher Scientific, San Jose, USA). In brief, lyophilized skins were weighed and rehydrated in a phosphate buffered saline then reduced with sodium borohydride at 25°C for 24 h. The reduction was quenched by adjusting the pH to 3.0 using acetic acid, and then the reduced samples were washed three times with water and lyophilized. The lyophilized reduced samples were hydrolysed in 3 ml of 6 M HCl containing 3% phenol at 105°C for 24 h and the resulting hydrolysates were then dried and rehydrated in 1 ml of water. A volume of 10 µl was removed from each sample for the analysis of hydroxyproline content in order to estimate the collagen content in skin based on previous work (Kliment et al., 2011). The remaining samples were subjected to a CF-11 column and the water eluent containing the crosslinks was then lyophilized and dissolved in 1 ml of water. The total concentration of natural crosslinks was determined and normalized with respect to the collagen content of the sample.
2.4. DSC analysis
DSC measurements were carried out using a Q2000 calorimeter (TA Instruments). Lyophilized samples were rehydrated overnight with deionized water of the same weight and then encapsulated in hermetically sealed aluminium pans (10 µl). All measurements were carried out over the temperature range from 30°C to 120°C under N2 purge and at a heating rate of 5°C min−1. The onset temperatures of the heat flow curve (Td) were marked to highlight the starting point of the thermal events.
3. Results and discussion
3.1. Quantitative analysis of the natural crosslinks: LC–MS
The reduction of C=N to CH—NH [Fig. 1(a)] is known to stabilize natural crosslinks from acidic degradation (Eyre et al., 1984). The quantity of natural crosslinks was analysed using LC–MS showing the variation across the eight groups of samples [Fig. 1(b)] (see Section 2.1 for differences in sample preparation). From control (CT) to dehaired (DH), natural crosslinks remain unchanged for both reduced and non-reduced skin samples (N-CT, R-CT, N-DH and R-DH) owing to the mild pH conditions (pH = 8). When skin samples were acidified (AC) to pH = 3, the number of total crosslinks decreased in the non-reduced samples (N-AC), whereas no change was observed for the reduced samples (R-AC). This is direct evidence of the completed reduction of labile crosslinks in the reduced skin, which makes the labile crosslinks stable under acidic conditions. Interestingly, after synthetic crosslinking using chromium sulfate, the detected amounts of natural crosslinks decrease for both reduced and non-reduced samples (N-CR and R-CR). A detrimental effect of chromium on natural crosslinks was previously reported in ovine skins (Zhang et al., 2018), which can also explain our observation here in bovine skins.
3.2. Effect of natural and synthetic crosslinks on collagen structure during denaturation: in situ SAXS
In situ monitoring of the structural changes during collagen denaturation was conducted using SAXS. Typical SAXS 2D patterns of collagen were observed [Fig. 2(a) shows a pattern of an N-CR sample] and a 1D profile is displayed for each type of sample [Fig. 2(b)]. The increase in overall peak intensity confirmed the synthetic crosslinking effect by chromium sulfate, in agreement with previous reports (Buchanan et al., 2019). The intermolecular packing of collagen follows a quarter-staggered arrangement, resulting in a gap and overlap banding structure with a characteristic periodicity in the axial direction (D-period) (Petruska & Hodge, 1964). When collagen denatures under moist heat, the triple helix uncoils and the intermolecular structure in the collagen fibril loses the long-range order as well as collapsing in the axial direction, as depicted in Fig. 2(a). This leads to a diminished intensity of the diffraction rings on the SAXS pattern [Figs. 2(a) and S3] and a decreased D-period [Figs. 2(f)–2(m)] (Zhang et al., 2018). However, in this study, R-CT and N-CT [Figs. 2(f) and 2(j)] showed unchanged D-periods during denaturation, whereas the other samples [Figs. 2(g)–2(i) and 2(k)–2(m)] showed clear decreasing trends. The unchanged D-period in CT samples during denaturation could be attributed to the non-covalent binding of inter-fibrillary non-collagenous proteins such as dermatan sulfate which constrains the collagen structure (Stinson & Sweeny, 1980). These proteins are removed during dehairing (Covington, 2009), thereby enabling the changes in D-period when denaturation occurs. Interestingly, the propensity of collagen fibrils to shrink in the axial direction due to the uncoiling of collagen molecules during denaturation is not suppressed by the presence of natural crosslinks. When the collagen triple helix uncoils during denaturation, steric hindrance due to the binding of collagen with non-collagenous proteins in the helical region could play a more important role than the constraining force by natural crosslinks in the telopeptidyl region.
The onset temperature of D-period (Tonset) for N-DH (52°C) was similar to that of R-DH (54°C) [Figs. 2(g) and 2(k)]. After acidification, the Tonset increased to 54°C in N-AC and 60°C in R-AC [Figs. 2(h) and 2(l)]. To prevent damage to the fibril structure caused by swelling at pH = 3, sodium chloride was added along with the sulfuric acid, which can result in reduced intermolecular hydration and a slight increase in Tonset. Also, Tonset was higher by 6°C in the reduced sample R-AC (60°C) than in the non-reduced sample N-AC (54°C). The difference in Tonset matches with the greater concentration of natural crosslinks in R-AC than N-AC. We speculate that retained natural crosslinks result in a dehydration effect in the collagen fibres, leading to denser packing of collagen molecules, reducing their entropy thereby increasing the heat stability (Miles & Ghelashvili, 1999; Miles et al., 2005).
When acidified skins are synthetically crosslinked using an excess of chromium sulfate, both N-CR and R-CR show similar Tonset at 78°C [Figs. 2(i) and 2(m)]. When both natural and synthetic crosslinks are presented in the collagen matrix, the denaturation temperature of collagen is determined by the labile domains in its triple helix with relatively less steric hindrance and a lower activation energy required for the polypeptide chains to uncoil (Miles & Bailey, 2001). Our results suggest that the contribution to the activation energy required for the uncoiling of the collagen triple helix from synthetic crosslinks is much higher than the retained natural crosslinks. Therefore, in our case, the Tonset of the collagen molecule is dominated by the contribution of the synthetic crosslinks. We also observe a second peak around 105°C in both N-CR and R-CR [Figs. 2(i) and 2(m)]. In our previous work, we suggested two mechanisms during the uncoiling of the collagen triple helix when crosslinked using chromium sulfate (Zhang et al., 2018) and these are further confirmed by the results in this study. The first Tonset at 78°C can be assigned to the conformational changes in the thermally labile domains in collagen molecules due to the covalent crosslinking with chromium sulfate. The second peak can be attributed to the complete uncoiling of collagen triple helices to random coils and the disordering of the intermolecular structure of collagen in the fibrils due to non-covalent interactions with chromium sulfate when an excess is used to crosslink the collagen.
A common feature was also observed in DH, AC and CR samples where the D-period showed an initial decrease followed by an increase after the first Tonset peak [Figs. 2(g)–2(i) and 2(k)–2(m)]. This could be caused by the presence of two populations of fibrils during heating, one remains intact and the other becomes denatured (Fig. 3). When the temperature reaches Tonset of a skin sample, the thermally unstable collagen molecules in the fibrils start to uncoil (Fig. 3). During uncoiling, the axial packing of the collagen molecules becomes denser (decrease in D-period), accompanied by swelling in the lateral direction of the fibrils as previously reported (Zhang et al., 2018). However, at Tonset the other population of collagen fibrils remains intact with the original D-period. As the temperature increases the overall peak intensity is reduced due to the intensity decrease from the denaturing population and the unchanged intensity from the intact population (Fig. S2 of the supporting information). Therefore, it is observed as the combination of diffraction signal from (i) intact collagen fibrils, with unchanged D-period and intensity, and (ii) denaturing collagen fibrils, showing a decrease in both D-period and intensity. When the intermolecular structure of the denaturing population is partially disordered, the peak intensities from both populations are comparable. When the diffraction peaks combine, we observe a shift in the peak towards lower D-period. When the denaturing fibrils are completely denatured, the observed intensity arises only from the intact fibrils, albeit weakly (Fig. S4). The presence of the two populations (intact and denaturing) can be explained on the basis of structural constraints. Miles & Ghelashvili (1999) previously reported an increase in the heat stability of the collagen triple helix from structural constraints due to the denser packing of the collagen molecules in the fibrils. Structural constraints may be caused by high crosslink density in the intact population of fibrils, or from swelling of the denaturing fibrils adjacent to the intact ones during temperature ramp (Fig. 3).
3.3. Effect of natural and synthetic crosslinks on the denaturation temperature of collagen: DSC
DSC results (Fig. 4) from the denaturation of collagen were generally in agreement with the in situ SAXS analyses. The denaturation temperatures (Td) of N-CT and R-CT were both 64°C but decrease to 54°C and 56°C for N-DH and R-DH, possibly due to the removal of non-collagenous proteins. When acidified, a Td difference of 4°C was observed for N-AC (55°C) and R-AC (59°C), confirming the contribution of natural crosslinks to the heat stability of collagen. However, in the crosslinked samples (N-CR and R-CR), DSC results show a single peak instead of two as shown in Figs. 2(i) and 2(m). The Td of both crosslinked samples (106°C and 107°C) match the second peak in the SAXS plots [∼105°C and ∼107°C, Figs. 2(i) and 2(m)], whereas no changes in heat flow were observed at temperatures around the first Tonset found in SAXS (78°C). Based on these observations, we suggest that: (i) the changes in collagen D-period at 78°C are driven by the entropy changes during the conformational changes in the collagen molecules, which involve only minor enthalpy changes, (ii) the complete uncoiling of collagen triple helices to random coils above 100°C involves both significant enthalpy changes and structural changes. Nevertheless, the effects of retained natural crosslinks on the heat stability of collagen were overwhelmed by the strong chemical binding from chromium sulfate.
In summary, natural crosslinks that are formed in vivo during collagen assembly can improve the heat stability of collagen molecules, but do not constrain the intermolecular structure to prevent the changes in D-period when denatured. However, the quantity of natural crosslinks decreases when chromium sulfate is involved, and their contributions towards the heat stability of the collagen structure are largely overwhelmed by synthetic crosslinks. Further research on the interplay between natural crosslinks and a wide range of synthetic crosslinkers could help retain the intrinsic properties of collagen in biological materials.
Figures S1 to S5 and Table S1. DOI: https://doi.org//10.1107/S1600577520009479/ju5017sup1.pdf
SP, YZ and BM thank NSRRC for SAXS beam time on the TLS BL23A1 beamline. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
The following funding is acknowledged: Ministry of Business, Innovation and Employment, New Zealand (grant No. LSRX-1801).
Ashwin, P. T. & McDonnell, P. J. (2010). Br. J. Ophthalmol. 94, 965–970. CrossRef PubMed CAS Google Scholar
Avery, N. & Bailey, A. (2008). Collagen, edited by P. Fratzi, pp. 81–110. Boston: Springer. Google Scholar
Beghetto, V., Agostinis, L., Gatto, V., Samiolo, R. & Scrivanti, A. (2019). J. Clean. Prod. 220, 864–872. CrossRef CAS Google Scholar
Buchanan, J. K., Zhang, Y., Holmes, G., Covington, A. D. & Prabakar, S. (2019). ChemistrySelect, 4, 14091–14102. CrossRef CAS Google Scholar
Covington, A. D. (2009). Tanning Chemistry: the Science of Leather. Cambridge: Royal Society of Chemistry. Google Scholar
Ding, W., Yi, Y., Wang, Y., Zhou, J. & Shi, B. (2019). Carbohydr. Polym. 224, 115–169. CrossRef Google Scholar
Eybl, E., Griesmacher, A., Grimm, M. & Wolner, E. (1989). J. Biomed. Mater. Res. 23, 1355–1365. CrossRef CAS PubMed Google Scholar
Eyre, D. R., Paz, M. A. & Gallop, P. M. (1984). Annu. Rev. Biochem. 53, 717–748. CrossRef CAS PubMed Google Scholar
Ingham, B., Li, H., Allen, E. L. & Toney, M. F. (2009). arXiv: 0901.4782. Google Scholar
Kliment, C. R., Englert, J. M., Crum, L. P. & Oury, T. D. (2011). Int. J. Clin. Exp. Pathol. 4, 349–355. PubMed Google Scholar
Knott, L. & Bailey, A. J. (1998). Bone, 22, 181–187. Web of Science CrossRef CAS PubMed Google Scholar
Lucero, H. & Kagan, H. (2006). Cell. Mol. Life Sci. 63, 2304–2316. CrossRef PubMed CAS Google Scholar
Lyu, B., Chang, R., Gao, D. & Ma, J. (2018). ACS Sustain. Chem. Eng. 6, 5413–5423. CrossRef CAS Google Scholar
Miles, C. A., Avery, N. C., Rodin, V. V. & Bailey, A. J. (2005). J. Mol. Biol. 346, 551–556. CrossRef PubMed CAS Google Scholar
Miles, C. & Bailey, A. (2001). Micron, 32, 325–332. CrossRef PubMed CAS Google Scholar
Miles, C. A. & Ghelashvili, M. (1999). Biophys. J. 76, 3243–3252. CrossRef PubMed CAS Google Scholar
Naffa, R., Holmes, G., Ahn, M., Harding, D. & Norris, G. (2016). J. Chromatogr. A, 1478, 60–67. CrossRef CAS PubMed Google Scholar
Naffa, R., Maidment, C., Ahn, M., Ingham, B., Hinkley, S. & Norris, G. (2019). Int. J. Biol. Macromol. 128, 509–520. CrossRef CAS PubMed Google Scholar
Olde Damink, L. H. H., Dijkstra, P. J., Van Luyn, M., Van Wachem, P., Nieuwenhuis, P. & Feijen, J. (1995). J. Mater. Sci. Mater. Med. 6, 460–472. CrossRef CAS Google Scholar
Petruska, J. A. & Hodge, A. J. (1964). Proc. Natl Acad. Sci. USA, 51, 871–876. CrossRef PubMed CAS Google Scholar
Skopinska-Wisniewska, J., Wegrzynowska-Drzymalska, K., Bajek, A., Maj, M. & Sionkowska, A. (2016). J. Mater. Sci. Mater. Med. 27, 67. PubMed Google Scholar
Stinson, R. H. & Sweeny, P. R. (1980). Biochim. Biophys. Acta, 621, 158–161. CrossRef CAS PubMed Google Scholar
Yang, J., Ding, C., Huang, L., Zhang, M. & Chen, L. (2017). Int. J. Biol. Macromol. 97, 1–7. CrossRef CAS PubMed Google Scholar
Zhang, Y., Mansel, B. W., Naffa, R., Cheong, S., Yao, Y., Holmes, G., Chen, H.-L. & Prabakar, S. (2018). ACS Sustain. Chem. Eng. 6, 7096–7104. CrossRef CAS Google Scholar
Zhou, Y., Ma, J., Gao, D., Li, W., Shi, J. & Ren, H. (2018). J. Clean. Prod. 201, 668–677. CrossRef CAS Google Scholar
© International Union of Crystallography. Prior permission is not required to reproduce short quotations, tables and figures from this article, provided the original authors and source are cited. For more information, click here.