Effects of Nanocellulose on the Structure of Collagen: Insights from Molecular Dynamics Simulation and Umbrella Sampling

: Collagen-nanocellulose composites have been widely used in biomedicine and tissue 10 engineering. However, the detailed mechanism underlying the effects of nanocellulose on the structure 11 of collagen hasn’t been elucidated. As the main component of skin tissue , the conformational disturbance 12 of collagen triggered by nanocellulose may shed light on the biocompatibility of nanocellulose. Therefore, 13 molecular dynamics simulations were carried out to gain insights into the interactions between 14 nanocellulose and collagen. Four different crystal planes of cellulose ((110), (100), (1-10), (010)) have 15 been constructed and the adsorption of collagen onto the four faces has been investigated respectively. It 16 has been found that the structure of collagen remained intact during the binding without chain separation. 17 The intactness of collagen supported the point that the nanocellulose has good biocompatibility. The 18 results derived from umbrella sampling showed that (110) and (1-10) faces exhibit the strongest affinity 19 with collagen, which may be attributed to its hydrophilicity and rather flat surfaces. The hydrophobicity 20 of (100) facet and roughness of (010) facet diminished the affinity with collagen. The occupancy of hydrogen bonds was low and hydrogen bonding interactions fail to make significant contributions to the binding of nanocellulose and collagen. These findings provided insights into the interactions between cellulose and collagen at an atomic level, which may guide the design and fabrication of collagen- nanocellulose composites. Furthermore, the biocompatibility of nanocellulose validated in the study may 25 help promote the biological application of nanocellulose involved composites.


36
2016), and other 2D nanomaterials, nanocellulose as a carrier material has a lower nanotoxicity, which 37 makes it more widely available. Although nanocellulose and its derivatives have been widely concerned 38 in medicine and biological tissue, the long-term retention of nanocellulose in the human body makes its 39 toxicity study very important due to the lack of cellulose-degrading enzymes in the human body.

40
Collagen can also be a candidate for biomaterials such as tissue-engineered scaffolds and wound 41 dressings (Lee et al. 2019;Ge et al. 2018;Sorushanova et al. 2019). However, the application of pure 42 collagen materials is limited due to their low water resistance, fast biodegradation perishability, and poor 43 thermal stability (Ge et al. 2018). While cellulose and collagen nanocomposite materials overcome the 44 weaknesses of pure collagen materials. Cellulose and its derivatives can be widely used to strengthen 45 various polymer matrix materials due to their high specific surface area, high crystallinity, low density, 46 and high elastic modulus (Manhas et al. 2015;Salimi et al. 2019;Liu et al. 2018;Li et al. 2017).

47
Researchers have proved that collagen/nanocellulose composite has good properties and stability better 48 than pure collagen. Animal experimental studies (Liu et al. 2020b;Liu et al. 2020a) (Collagen/cellulose 49 nanofiber hydrogel scaffold: physical, mechanical and cell biocompatibility properties; A 3D porous 50 microsphere with multistage structure and component based on bacterial cellulose and collagen for bone 51 tissue engineering (Zhang et al. 2020); demonstrated that collagen and nanocellulose composite is a 52 promising material for wound dressings and tissue engineering scaffolds.

53
The advantages and disadvantages of nanocellulose and collagen can effectively complement each 54 other to form a more potential nanocomposite material (Cudjoe et al. 2017), which makes their 55 3 composites have a broader application prospect. However, the interactions between nanocellulose and 56 collagen, which are significantly related to the strength of composites, are still obscure. Furthermore, in 57 vivo and in vitro experiments have shown that nanocellulose and its derivatives have adverse effects on 58 intestinal microorganisms (DeLoid et al. 2019), liver cells (Otuechere et al. 2020), and lung cells (Sai and 59 Fujita 2020). In addition, nanocellulose biological dressings and tissue-engineered materials, as the main 60 components of human tissues, will directly interact with collagen when they contact the human body.

61
Therefore, it is necessary to study the toxicity of nanocellulose.

62
In this study, molecular dynamics simulations were carried out to study the interactions between

83
In this work, Ιβ-cellulose, which is one of the main components of higher plants, was selected to 4 model nanocellulose. Collagen type I was employed as the model collagen, which is the most abundant 85 and widely distributed natural structural protein in the human body (Lin and Liu 2006

92
The initial structure of collagen is obtained by extracting three chains from the crystal structure 93 (Berisio et al. 2009) (PDB code 1K6F) as shown in Fig. 1

97
Dalke and Schulten 1996) to form the initial coordinates of the simulation system and the minimum 98 distances between cellulose faces and collagens were ranging from 0.8 nm to 1 nm ( Fig. 2A). As shown 99 in Fig. 2(A), three parallel simulations of each system were carried out for 500 ns. The composite system 100 was solvated in a cubic box with a TIP3P water model (Mark and Nilsson 2001;Jorgensen et al. 1983) 101 and modeled by a CHARMM36 force field (Lee et al. 2014; Boonstra, Onck and van der Giessen 2016).

106
The energy minimization process was carried out with 1000 cycles of steepest descent and 1,000 cycles 5 of conjugate gradient minimization. Then, equilibration runs were performed for 5 ns in the NVT 108 ensemble and 5 ns in the NPT ensemble with the heavy atoms of protein and cellulose fixed. Finally, 500 109 ns production runs were simulated in the NPT ensemble with the restriction of the protein released. The 110 long-range electrostatic interactions were treated by the particle mesh Ewald (PME) method (Petersen 111 1995), while the short-range van der Waals interactions were calculated with a cutoff distance of 1.0 nm.

112
All covalent bonds containing hydrogen atoms were constrained by the LINCS algorithm (Hess et al. 113 2008). The V-rescale thermostatic (Berendsen et al. 1984) was used to heat the system to 300 K and the

117
The potential of mean force (Roux 1995) (PMF) obtained by pulling simulation and umbrella 118 sampling (Hub 2015) was used to calculate the binding free energy of the system. The cellulose surface 119 was used as a reference point and a harmonic potential was applied to the collagen as a pulling point.

120
The last frame of the MD simulations was selected as the initial conformation, 300 ps umbrella traction 121 was provided for collagen along the z-axis to increase the center of mass (COM) distance between 122 collagen and cellulose. The spring constant used was 2000 kJ mol -1 nm 2 and the pull rate was 0.01 nm/ps.

123
More than 13 umbrella sampling windows were selected according to the interval size of COM values.

124
1ns of simulations in NPT was performed on each sample, then 10 ns of MD process was carried out.

126
The relevant modules in GROMACS were used to calculate the backbone root mean square  In the initial simulation system, the distance between protein and cellulose crystal faces was 133 controlled between 0.8-0.1nm as shown in Fig. 2A

136
The protein slowly contacted the cellulose surface. As shown in Fig. 2(B-E), no significant structural 137 changes were observed except that the overall structure of collagen was slightly bent at the end of the

214
The chain of collagen is composed of repeating tripeptide sequence Gly-Pro-Pro, which are all polar 215 amino acids. Polar interactions between hydrophilic cellulose faces and collagen enhance their affinity.

216
Though (010) face is hydrophilic, the interaction between these faces and collagen is the weakest among 217 the four systems, which seems quite counterintuitive. The origin of this behavior is attributed to the 218 topography of (010) faces, which hinders the binding of collagen with half of the hydroxyl grouping

226
To further validate the deduction, radial distributions of water molecules were calculated to evaluate 227 the hydrophilicity of the four faces. As shown in Fig. 6(B-E), their first peak positions occur at the same 228 position 3.8 Å, respectively. The height of the first g (r) peak belonging to (100) face is about 0.4, which 229 is significantly lower than that of (110), (1-10) and (010) faces with the heights of the first peak all about 230 0.6. Thus, the heights of the first peaks describe a distinguishable difference in hydrophilicity among the 231 four faces and the surface hydration of (100) face is weaker than the other three faces. To evaluate the 232 effects of surface morphology on the interaction between collagen and cellulose, the contact number of 233 11 heavy atoms was calculated with 0.5 nm as the threshold. In general, the loading of collagen on the 234 cellulose surface was fast with molecules of collagen adsorbed on cellulose within 200 ns. Based on the 235 heavy atom contact numbers between collagen and cellulose, it has been deduced that there is an obvious 236 correlation between surface roughness and contact numbers. As shown in Figure 7, (100) face displays 237 the largest contact number with the smoothest surface, while, (010) face exhibits the least contact number 238 with the greatest surface roughness. (110) and (1-10) faces are in between with the contact numbers larger 239 than that of (100) face but less than that of (010) face. As the fundamental part of molecular interactions,

247
Cellulose molecules contain a large number of free hydroxyl groups, which might be involved in 248 hydrogen bonding interactions. Therefore, the average occupancy of hydrogen bonds between the 249 collagen and cellulose in the four systems was calculated respectively. As shown in Fig. 8

266
This study provides theoretical guidance for the design and fabrication of collagen-nanocellulose 267 composites. Furthermore, the intactness of collagen structure supported the viewpoint that nanocellulose 268 is quite biocompatible.