Environment influences on the aromatic character of nucleobases and amino acids

Abstract

Geometric (HOMA) and magnetic (NICS) indices of aromaticity were estimated for aromatic rings of amino acids and nucleobases. Cartesian coordinates were taken directly either from PDB files deposited in public databases at the finest resolution available (≤ 1.5 Å), or from structures resulting from full gradient geometry optimization in a hybrid QM/MM approach. Significant environmental effects imposing alterations of HOMA values were noted for all aromatic rings analysed. Furthermore, even extra fine resolution (≤ 1.0 Å) is not sufficient for direct estimation of HOMA values based on Cartesian coordinates provided by PDB files. The values of mean bond errors seem to be much higher than the 0.05 Å often reported for PDB files. The use of quantum chemistry geometry optimization is strongly advised; even a simple QM/MM model comprising only the aromatic substructure within the QM region and the rest of biomolecule treated classically within the MM framework proved to be a promising means of describing aromaticity inside native environments. According to the results presented, three consequences of the interaction with the environment can be observed that induce changes in structural and magnetic indices of aromaticity. First, broad ranges of HOMA or NICS values are usually obtained for different conformations of nearest neighborhood. Next, these values and their means can differ significantly from those characterising isolated monomers. The most significant increase in aromaticities is expected for the six-membered rings of guanine, thymine and cytosine. The same trend was also noticed for all amino acids inside proteins but this effect was much smaller, reaching the highest value for the five-membered ring of tryptophan. Explicit water solutions impose similar changes on HOMA and NICS distributions. Thus, environment effects of protein, DNA and even explicit water molecules are non-negligible sources of aromaticity changes appearing in the rings of nucleobases and aromatic amino acids residues.

Fig. 1

Heterogeneity of the structural index of aromaticity (HOMA) calculated for four aromatic amino acids residues. Two sets of plots correspond to the systematic analysis of mean bond length errors (gray lines) and to protein environment (black lines with open circles or triangles). Only high quality structures (resolution better than 1.5 Å) from the Protein Database [30] were used. For detailed statistics, see Table S1. Standard deviations of the analysed HOMA values distribution are given in brackets. The amount of mean bond length errors is indicated in the legends. HOMA(x) Aromaticity of the x-membered ring

Fig. 2

The relationship between skewness and kurtosis for HOMA distributions of amino acid (a) and nucleobase (b) aromatic rings affected by various bond length inaccuracies. Larger symbols represent skewness and kurtosis of experimental values estimated from Cartesian coordinates taken directly from PDB files [30, 31]

Fig. 3

Correlation between HOMA values of aromatic amino acids estimated using Cartesian coordinates taken directly from PDB files (HOMA^PDB) and those obtained after QM/MM optimization (HOMA^QM/MM) at ONIOM(QM:B3LYP/6-311 + G**//MM:AMBER) level. Vertical lines HOMA values estimated for geometries of isolated monomers optimized in the gas phase using B3LYP/6-311 + G** method

Fig. 4

Sensitivity of histidine aromaticity to interactions with the iron cation in the model heme system. Solid lines HOMA values, dotted lines Fe-HIS distances. The distance of both histidine molecules from Fe²⁺ after full optimization is 2.041(2.026) Å, while HOMA values are 0.891(0.895). Values in brackets correspond to water solution (PCM model)

Fig. 5

Heterogeneity of HOMA values characterising aromaticity of the pyrimidine rings of nucleobases. The two sets of plots correspond to systematic bond length error analysis (gray lines) and native DNA environment (coordinates directly taken from Nucleic Acid Database [31]) (black lines with symbols). Only high quality structures (resolution better than 1.5 Å) were taken into consideration. Standard deviations of the analysed HOMA values distribution are given in brackets

Fig. 6

Heterogeneity of HOMA values characterising aromaticity of the purine [HOMA(9)] and imidazole (HOMA5) rings of adenine and guanine. The two sets of plots correspond to systematic bond length error analysis (gray lines without symbols) and native DNA environment (taken from Nucleic Acid Database [31]) (black lines with symbols). Only high quality structures (resolution better than 1.5 Å) were taken into consideration. Standard deviations of the analysed HOMA values distribution are given in brackets

Fig. 7

Correlation between HOMA values of (a) pyrimidine rings and (b) imidazole/purine rings of nucleobases estimated using Cartesian coordinates taken directly from PDB files (HOMA^PDB) and those obtained after QM/MM optimizations (HOMA^QM/MM) at ONIOM(QM:B3LYP/6-311 + G**//MM:AMBER) level. Vertical lines HOMA values estimated for geometries of isolated monomers optimized in the gas phase (using B3LYP/6-311 + G** level)

Fig. 8

Heterogeneity of the magnetic measure of aromaticity [NICS(1)] of amino acids and nucleobases inside protein (PDB [30]) or DNA (NDB [31]) interiors, respectively. The lines represent reference levels of isolated monomers optimized in the gas phase

Fig. 9

Heterogeneity of the structural index of aromaticity (HOMA) corresponding to nucleobases or amino acids inside DNA (NDB [31]), or protein (PDB [30]) interiors. Gray symbols Distributions of HOMA values estimated for rings geometries optimized in explicit water solutions using QM/MM approach, lines reference levels of isolated monomers optimized in the gas phase