Chiral acidic amino acids induce chiral hierarchical structure in calcium carbonate

Chirality is ubiquitous in biology, including in biomineralization, where it is found in many hardened structures of invertebrate marine and terrestrial organisms (for example, spiralling gastropod shells). Here we show that chiral, hierarchically organized architectures for calcium carbonate (vaterite) can be controlled simply by adding chiral acidic amino acids (Asp and Glu). Chiral, vaterite toroidal suprastructure having a ‘right-handed' (counterclockwise) spiralling morphology is induced by L-enantiomers of Asp and Glu, whereas ‘left-handed' (clockwise) morphology is induced by D-enantiomers, and sequentially switching between amino-acid enantiomers causes a switch in chirality. Nanoparticle tilting after binding of chiral amino acids is proposed as a chiral growth mechanism, where a ‘mother' subunit nanoparticle spawns a slightly tilted, consequential ‘daughter' nanoparticle, which by amplification over various length scales creates oriented mineral platelets and chiral vaterite suprastructures. These findings suggest a molecular mechanism for how biomineralization-related enantiomers might exert hierarchical control to form extended chiral suprastructures.

the modified Kamhi model with vertical carbonates is also inconsistent with our study's chirality observations.
With one orthogonal mirror plane, the (100) surface with "angled" carbonate groups presents the necessary symmetry. There are several possibilities for the termination of the (100) surface. Since the amino acids are negatively charged, and local electrostatic complementarity is needed for favorable binding energies, we discarded the possibility of a net-negatively charged surface termination. We next explored a net-positively charged surface termination and found that it both maintained surface symmetry and strongly adsorbed negatively charged acidic amino acids. However, because of the even spacing of calcium ions on the net-positively charged surface, rotation of any acidic amino acid by 180° creates similar electrostatic contacts between both carboxyl groups and surface calcium atoms. Because these energetically favorable contacts can be made regardless of amino acid orientation, adsorption on net-positively charged surfaces cannot explain the different chiral morphologies resulting from chiral amino acid adsorption.
Finally, we investigated net-neutrally terminated surfaces because they often have low surface energies and are stable in solution ( Supplementary Fig. 10). We explored several net-neutrally terminated surfaces by deleting alternating Ca atoms as well as alternating rows of Ca atoms. We found that deleting alternating rows of Ca atoms perpendicular to the mirror plane maintained symmetry and provided an environment where directionally dependent interactions could be made, yielding an orientational preference. Given that the net-neutrally terminated modified Kamhi (100) surface is expected to be stable in solution, maintains symmetry, and has the potential for orientational preference, we anticipated that docking simulations performed on this surface might generate models capable of explaining the different chiral morphologies resulting from chiral amino acid adsorption. Indeed, as described in the main text, on the modified Kamhi (100) surface, L-Asp and D-Asp have preferred orientations of adsorption, and adsorption in the orientation of the opposite enantiomer changes the binding energy by 0.64 kcal mol -1 . In an ensemble of molecules, this energy difference would result in one orientation being approximately three times as likely as the other, which could create the 4° tilt between nanoparticles observed in the vaterite suprastructure.
In addition to considering how the bulk crystal should be terminated, we also considered the possibility of hydrated species existing at the vaterite surface as a function of pH, as has been found to be important in other systems 3 . Our experiments confirm the formation of chiral morphology from pH 6.46 to 12.40 ( Fig. 3  between pH range 4-9, suggesting the absence of pH-dependent hydration species. The invariance of our observations across a wide range of pH suggests two possibilities. First, a pristine termination may persist for the entire pH range. This is the case represented by our structure prediction models. As shown in Fig. 6a of the main text, a strong hydrogen bond between the amino terminus of Asp and a surface carbonate ion is needed for chiral selectivity. In the second case, at lower pH, the hydrated binding of D-Asp onto the P3 2 21 is favored by 0.33 kcal mol -1 over L-Asp). The difference in adsorption energy arises from the electrostatic complementarity between the P3 1 21 surface and L-and D-Asp (Fig. 9 in main text). Both L-and D-Asp are anchored to the surface via hydrogen bonding between the amino terminus and the surface carbonates as well as by electrostatic contacts between the carboxyl terminus and one surface calcium. With these two functional groups locked in identical positions for both L-and D-Asp, the adsorption energy difference is attributable to the interaction of the charged side chain with the mineral. In the case of L-Asp, the side chain carboxyl group is able to interact favorably with an exposed calcium ion. However, in the case of D-Asp, the side chain carboxyl is forced to interact unfavorably with a highly coordinated calcium ion.

Comparison of proposed mechanisms
Both our "A" and "B" models provide a potential mechanism for the experimentally observed