Predicted Adsorption Affinity for Enteric Microbial Metabolites to Metal and Carbon Nanomaterials

Ingested nanomaterials are exposed to many metabolites
that are
produced, modified, or regulated by members of the enteric microbiota.
The adsorption of these metabolites potentially affects the identity,
fate, and biodistribution of nanomaterials passing the gastrointestinal
tract. Here, we explore these interactions using in silico methods,
focusing on a concise overview of 170 unique enteric microbial metabolites
which we compiled from the literature. First, we construct quantitative
structure–activity relationship (QSAR) models to predict their
adsorption affinity to 13 metal nanomaterials, 5 carbon nanotubes,
and 1 fullerene. The models could be applied to predict log k values for 60 metabolites and were particularly applicable
to ‘phenolic, benzoyl and phenyl derivatives’, ‘tryptophan
precursors and metabolites’, ‘short-chain fatty acids’,
and ‘choline metabolites’. The correlations of these
predictions to biological surface adsorption index descriptors indicated
that hydrophobicity-driven interactions contribute most to the overall
adsorption affinity, while hydrogen-bond interactions and polarity/polarizability-driven
interactions differentiate the affinity to metal and carbon nanomaterials.
Next, we use molecular dynamics (MD) simulations to obtain direct
molecular information for a selection of vitamins that could not be
assessed quantitatively using QSAR models. This showed how large and
flexible metabolites can gain stability on the nanomaterial surface
via conformational changes. Additionally, unconstrained MD simulations
provided excellent support for the main interaction types identified
by QSAR analysis. Combined, these results enable assessing the adsorption
affinity for many enteric microbial metabolites quantitatively and
support the qualitative assessment of an even larger set of complex
and biologically relevant microbial metabolites to carbon and metal
nanomaterials.