Trifluoracetic acid-assisted crystallization of vitamin B₁₂ results in protonation of the phosphate group of the nucleotide loop: Insight into the influence of crystal packing forces on vitamin B₁₂ structures

In the course of experiments concerning our ongoing project on the synthesis of vitamin B₁₂ (cyanocobalamin, CNCbl) bioconjugates for drug-delivery purposes, we observed the formation of well-shaped red parallelepipeds from a concentrated aqueous solution of the HPLC-purified vitamin. The X-ray structural investigation (MoK_α) at 98 K on these crystals revealed a CNCbl-TFA salt of formula [CNCbl(H)](TFAc) ·14H₂O (1, where TFA = trifluoracetic acid; TFAc^- = trifluoracetate anion), in which a proton transfer from the trifluoracetic acid to the phosphate-O4P oxygen atoms is observed. 1 crystallizes in the standard orthorhombic P2₁2₁2₁ space group, a = 16.069(2) Å, b = 20.818(2) Å, c = 24.081(2) Å, Z = 4. The final full-matrix least-squares refinements on F² converged with R ₁ = 4.1% for the 18957 significant reflections, a very low crystallographic residual for cobalamins, which facilitated the analysis of the extensive network of hydrogen bonds within the lattice. To the best of our knowledge, this is the first cobalamin structure to show protonation of the phosphate group of the cobalamin nucleotide loop. In this work, the crystal structure of 1 is analyzed and compared to other CNCbls reported in the literature, namely, CNCbl·3PrOH·12H₂O (2, PrOH = propyl alcohol), CNCbl·acetone·20H₂O (3), CNCbl·2LiCl·10.2H₂O (4), and CNCbl·2KCl· 10.6H₂O (5). The analysis confirmed that protonation of the phosphate leaves the major CNCbl structural parameters unaffected, so that 1 can be considered an "unmodified" Cbl solvate. However, comparison between 1-5 led to interesting findings. In fact, although the cobalt(III) coordination sphere in 1-5 is similar, significant differences could be noted in the upward fold angle of the corrin macrocycle, a parameter commonly related to the steric hindrance of the axial lower "α" nucleotide-base and the electronic trans influence of the upper "β" substituent. This suggests that crystal-packing forces may influence the corrin deformation as well. Herein we explore, on the basis of the newly acquired structure and reported crystallographic data, whether the incongruities among 1-5 have to be attributed to random crystal packing effects or if it is possible to associate them with specific crystal packing (clusters).