The structural role of osteocalcin in bone biomechanics and its alteration in Type-2 Diabetes

This study presents an investigation into the role of Osteocalcin (OC) on bone biomechanics, with the results demonstrating that the protein’s α-helix structures play a critical role in energy dissipation behavior in healthy conditions. In the first instance, α-helix structures have high affinity with the Hydroxyapatite (HAp) mineral surface and provide favorable conditions for adsorption of OC proteins onto the mineral surface. Using steered molecular dynamics simulation, several key energy dissipation mechanisms associated with α-helix structures were observed, which included stick–slip behavior, a sacrificial bond mechanism and a favorable binding feature provided by the Ca2+ motif on the OC protein. In the case of Type-2 Diabetes, this study demonstrated that possible glycation of the OC protein can occur through covalent crosslinking between Arginine and N-terminus regions, causing disruption of α-helices leading to a lower protein affinity to the HAp surface. Furthermore, the loss of α-helix structures allowed protein deformation to occur more easily during pulling and key energy dissipation mechanisms observed in the healthy configuration were no longer present. This study has significant implications for our understanding of bone biomechanics, revealing several novel mechanisms in OC’s involvement in energy dissipation. Furthermore, these mechanisms can be disrupted following the onset of Type-2 Diabetes, implying that glycation of OC could have a substantial contribution to the increased bone fragility observed during this disease state.


Supplementary Figures
. The dissipated energy and number of hydrogen bonds. These curves show a direct relation between the energy dissipation rate and the number of hydrogen bonds for different simulations of R1-4A-#2 for parallel pulling. a) 1 Å⁄ns simulation #4, b) 1 Å⁄ns simulation #5, c) 10 Å⁄ns simulation #1 and d) 10 Å⁄ns simulation #3. Figure SI2. The simulation results for the parallel pulling of OC in healthy Case 3 (R54-3A-#1). These results show the stick-slip mechanism occurrence. Dissipated energy and number of hydrogen bonds for a) simulation #1, b) simulation #2 and c) simulation #3. TYR1 displacement for d) simulation #1, e) simulation #2 and d) simulation #3. As long as the TYR1 is stuck on the surface (or its parallel pulling velocity is lower than the pulling velocity) the dissipated energy increases. After a jump in the TYR1 displacement there is no more energy dissipation (fixed total dissipation amount. e.g. horizontal curve) Figure SI3. The simulation results of the SMD simulation of R1-4A-#2 for perpendicular pulling. a) the maximum force -velocity shows the maximum force convergence for the pulling speeds of lower than 1 Å⁄ns. b) Dissipated energy (blue curve) and hydrogen bond number (red curve) -displacement curves for the pulling speed of 1 Å⁄ns averaged over three simulations with different random seed numbers illustrating a direct relation between the hydrogen bonds number and the energy dissipation rate.

Energy Dissipation
Final snapshot of the simulation R1-4A-#2 (Table SI1) is utilized to run SMD simulations with pulling velocities between 1 − 100 Å⁄ to obtain a competent velocity value. The maximum force -velocity curve reaches a plateau before the velocity of 10 Å⁄ ( Figure SI4-a), with simulations carried out at the lowest velocities of 1 Å⁄ and 10 Å⁄ showing similar behavior, even though each simulation peaks at different displacement value ( Figure SI4-b). This behavior can be considered as a reminiscent for a mechanism with a non-deterministic nature.
In physiologically relevant pulling velocities, the covalent bonds in the OC protein do not bear external loading. Instead, load is transmitted through non-bonded interactions. For the pulling speed of 1 Å⁄ the maximum force is equal to 256.09 ± 28.31 . If the whole protein backbone is under this force, according to Thompson et al, the carbon-carbon bond energy will be equal to 0.7 / which is a little higher than a and is around 2 percent of the dissipated energy 1 . In addition, the strength of 2+ mediated ion networks in osteopontin which are also present in the OC-HAp interface is estimated to be between 100 and 1100 2 . The value of 256.09 ± 28.31 measured here is in this range.
Thus, the selected pulling rate is small enough for the purpose of the current study.

Adsorption for 5 th Orientation
As mentioned before, the fifth configuration was itself subdivided into four initial orientations. For this case just the orientations in which the random coil segment (residues 1-12) faces toward the HAp surface are considered. For this case the most important contact residues are TYR1, ASP28, PRO27, GLU31 and ARG44 with the percentage of 37.74%, 28.82%, 24.24%, 19.86% and 16.18% among the contact residues, respectively. The final simulation snapshots for R51-3A-#2 and R54-3A-#1 are almost the same as each other ( Figure SI5). The protein adsorption results showed that for the simulation R51-3A-#2 the protein rotates and switches to an orientation like the simulation R54-3A-#1. As a result, it seems that the former had an opportunity to opt to other orientations and it chooses the latter because of its favorability. Thus, the final snapshot of R51-3A-#2 is selected for the SMD simulation and surprisingly enough its main contact residues are TYR1, PRO27, ASP28 and GLU31 similar to the most prominent contact residues among all the simulations for the complete OC adsorption on HA.

Energy Dissipation for Parallel pulling
The contact residues during the SMD pulling for three simulations done at 1 Å⁄ are summarized in Table SI2. At the beginning of the simulation, the main contact residues are GLU40 and ARG43 like table SI1. As the simulation proceeds, the helix in contact with the HAp surface switches from α3 to α2. After this change, the contact residues of protein are ASP28, GLU31, ASP34 and HSD35. It seems that the protein rolls over on the surface during the pulling and it finds itself a preferable contact mode with the HAp surface. It is interesting that the main contact part of the protein switches to α2 helix which showed more importance in the adsorption simulations (Table SI1).
To shed more light on the OC mediated energy dissipation, SMD simulations with lower spring constants of 2 and 0.2 Å 2 ⁄ ⁄ were carried out as lower SMD spring constant slows down the process. Changes in the spring constant does not affect the amount of dissipated energy for the higher spring constants ( Figure SI6). For the lower spring constant of 0.2 Å 2 ⁄ ⁄ , at the end of the simulation (40 ) the protein is still attached to the surface. The dissipated energy will be higher than the value reported here for this spring constant. As a result, in the case of a very flexible bond between the OC and the rest of the bone hierarchal structure the dissipation energy will be higher. Figure SI4. The dissipated energy of the simulation R1-4A-#2 for parallel pulling with different SMD spring constants. The energy dissipation is the same for different spring constants