Peptide Retention in Hydrophilic Strong Anion Exchange Chromatography Is Driven by Charged and Aromatic Residues

Abstract

Hydrophilic strong anion exchange chromatography (hSAX) is becoming a popular method for the prefractionation of proteomic samples. However, the use and further development of this approach is affected by the limited understanding of its retention mechanism and the absence of elution time prediction. Using a set of 59 297 confidentially identified peptides, we performed an explorative analysis and built a predictive deep learning model. As expected, charged residues are the major contributors to the retention time through electrostatic interactions. Aspartic acid and glutamic acid have a strong retaining effect and lysine and arginine have a strong repulsion effect. In addition, we also find the involvement of aromatic amino acids. This suggests a substantial contribution of cation-π interactions to the retention mechanism. The deep learning approach was validated using 5-fold cross-validation (CV) yielding a mean prediction accuracy of 70% during CV and 68% on a hold-out validation set. The results of this study emphasize that not only electrostatic interactions but rather diverse types of interactions must be integrated to build a reliable hSAX retention time predictor.

Figure 1

Effect of the charged residues on peptide retention in hSAX. (a) Mean residue count per peptide for D/E (red) and K/R (blue) over fraction. Error bars denote the standard deviation. Peptide count per fraction is shown in orange (total 59 297 unique peptides). (b) Effect of D/E count (range 0–5) on peptide retention. (c) Extracted chromatogram of peptides with three and four D/E (red). Subpopulations were defined according to the number of K/R residues (one to three, blue tones for peptides with three D/E residues and green tones for peptides with four D/E residues). Crosses mark the mode of the respective distributions.

Figure 2

Detailed comparison of relative contributions of positively (K/R) and negatively (D/E) charged residues on peptide retention in hSAX. (a) Effect size of K/R residues. Peptides with four D/E residues were divided according to their K and R count (K, green tones; R, blue tones). (b) Effect size of E/D residues. Peptides with either two E or two D residues are shown, split according to their number of K residues (1 or 2).

Figure 3

The effect of neutral amino acids on peptide retention in hSAX. Amino acids were grouped according to their influence on peptide retention in hSAX by linear regression (Supporting Information). (a) Elution behavior of peptides with different numbers of F/Y/W and two D/E, one K/R residues. (b–e) Elution behavior of peptides with different numbers of the indicated amino acids (b, P/A/S; c, Q/T/V; d, I/G/L; e, C/M/N/H) and two D/E, one K/R, zero F/Y/W. Crosses mark the mode of the subpopulations.

Figure 4

Peptide retention time prediction for hSAX using machine learning. (a) Residue retention coefficients from a linear model with length correction parameter. (b) Fraction of correct predictions (accuracy) of different prediction methods, estimated by 5-fold cross-validation based on 35 578 (train) and 8894 (test) peptides in each split. (c) Elution time prediction for the hold-out validation set, FNN classifier (left) and LM (right); ρ indicates the Pearson correlation. Linear Model (LM), Logistic Regression (LR), Feedforward Neural Network (FNN).