ACS PRF | ACS
All e-Annual Reports
44321-AC4
Solid-State NMR as a Probe of Flavin Interactions, Electronics and Reactivity
Research during the past year benefited from the recruitment and training of one excellent student. However this was a relatively inefficient process, requiring initial recruitment of three. One of the two students recruited initially had to leave the group, and a third student (recruited to replace that person) then decided against completing a Ph. D.. Each of these attempts cost eight months. Thus, identifying quality personnel with a commitment to research has been a challenge that has taken time, and cost continuity. This is the reason that we are behind both in expenditures and in results. However, I plan to try to recruit a postdoc. next year, to join me when I return from my sabbatical. The retained student is about to stand for her qualifying oral exam.
We have published two papers on the work proposed. One appeared in JACS. and the other in Tet. Lett. Both describe essential foundations for more work to occur this fall. The JACS paper reports our success in Objective 3, as it reports 13C principal chemical shift values obtained via 2-dimentional PASS spectra collected on lumiflavin. These augment our 15N-NMR studies of tetraacetyl riboflavin (TARF), presented in the grant application. One interesting finding that emerged from the 15N experiments was the striking difference between the chemical shift principal values of the flavin N5 and N1 sites. Despite the fact that these two are both considered to be ‘pyridine-type' sites, N5 displayed a span more than twice that of N1, in addition to the previously documented isotropic shifts that differ by some 150 ppm. By comparison with a review of 15N isotropic chemical shifts, that of N1 is less shifted than the least shifted previous example and N5 is more shifted than the most strongly shifted previous example. I find this particularly remarkable considering that N5 and N1 are linked to one-another in a conjugated butadiene system. Thus, the flavin ring system is very strongly polarizing. The trends we observe are also consistent with the known distinct reactivities of the flavin N1 and N5 sites. N5 behaves as an electrophile, accepting hydride, whereas N1 acts as a nucleophile and is the site of flavin protonation at pH 2. Thus the flavin incorporates two superficially similar sites with very different, complementary reactivities, into a single cofactor. This provides a chemical rationale for the evolutionary selection of flavins for use in such a wide variety of biological chemistry.
As described under Objective 2, we have begun density functional theory (DFT) calculations of the complexes we proposed to observe experimentally. Unfortunately this was to be the project of the student who left, and it suffered. Nonetheless, two positive indications are in hand. Since an easy sample to make is hydrated TARF, for comparison with dry TARF, we added water molecules around the periphery of the flavin ring, employing molecular dynamics to achieve initial placement and then moving stepwise to increasingly large basis sets using density functional theory. In the case that converged well, chemical shift principal values were calculated as well as isotropic values. In a case with only one water molecule H-bonding to the flavin, our calculated isotropic shifts all constituted moves towards those reported for flavin mononucleotide (FMN) in water, as hoped. In future calculations of this sort, we will try providing diffuse functions immediately after initial molecular dynamics optimization, as these are likely necessary for mediating the weak long-range interactions in which H2O engages. The natural bond order (NBO) population analyses resulting from this calculation revealed increased polarization of the flavin, and striking electron density excess at the C8' and C7' positions. This surprising behaviour (C8' and C7' are both methyls) is consistent with NMR observation that the protons of C8' are subject to exchange with solvent, which is not normal for methyl protons. Thus, incomplete as they are, our results are headed in the right direction based on comparison with experiments.
We have also undertaken calculations of the complex between lumiflavin and diamido pyridine (DAP). Consistent with the tight complexes formed in solution, the calculated complex adopted a single geometry and converged readily at low levels of theory. However at higher levels of theory expected to better describe H-bonding, the complex displayed signatures of strong single-well H-bonds. Thus, we found very short distances between the O and N atoms sharing the H+, and very similar distances from the H+ to each partner, blurring the distinction between donor and acceptor. Experimental evidence does not support short, strong single-well H-bonds (although it does support good ones). We hypothesize that our calculations over-estimate the strength of these H-bonds because they are performed in vacuuo. We propose to resume this work using a dielectric. Nonetheless, it is far better to have an over-strong positive result, than a negative result.
