Ion-mobility mass spectrometry is utilized to examine the metacluster formation of serine， asparagine， isoleucine， and tryptophan. These amino acids are representative of different classes of noncharged amino acids. We show that they can form relatively large metaclusters in solution that are difficult or impossible to observe by traditional solution techniques. We further demonstrate， as an example， that the formation of Ser metaclusters is not an ESI artifact because large metaclusters can be detected in negative polarity and low concentration with similar cross sections to those measured in positive polarity and higher concentration. The growth trends of tryptophan and isoleucine metaclusters， al... More
Ion-mobility mass spectrometry is utilized to examine the metacluster formation of serine， asparagine， isoleucine， and tryptophan. These amino acids are representative of different classes of noncharged amino acids. We show that they can form relatively large metaclusters in solution that are difficult or impossible to observe by traditional solution techniques. We further demonstrate， as an example， that the formation of Ser metaclusters is not an ESI artifact because large metaclusters can be detected in negative polarity and low concentration with similar cross sections to those measured in positive polarity and higher concentration. The growth trends of tryptophan and isoleucine metaclusters， along with serine， asparagine， and the previously studied phenylalanine， are balanced among various intrinsic properties of individual amino acids (e.g.， hydrophobicity， size， and shape). The metacluster cross sections of hydrophilic residues (Ser， Asn， Trp) tend to stay on or fall below the isotropic model trend lines whereas those of hydrophobic amino acids (Ile， Phe) deviate positively from the isotropic trend lines. The growth trends correlate well to the predicted aggregation propensity of individual amino acids. From the metacluster data， we introduce a novel approach to score and predict aggregation propensity of peptides， which can offer a significant improvement over the existing methods in terms of accuracy. Using a set of hexapeptides， we show that the strong negative deviations of Ser metaclusters from the isotropic model leads a prediction of microcrystalline formation for the SFSFSF peptide， whereas the strong positive deviation of Ile leads to prediction or fibril formation for the NININI peptide. Both predictions are confirmed experimentally using ion mobility and TEM measurements. The peptide SISISI is predicted to only weakly aggregate， a prediction confirmed by TEM.
