Quantitative proteomics analysis of adsorbed plasma proteins classifies nanoparticles with different surface properties and size

Abstract

Nanoparticle biological activity, biocompatibility and fate can be directly affected by layers of readily adsorbed host proteins in biofluids. Here, we report a study on the interactions between human blood plasma proteins and nanoparticles with a controlled systematic variation of properties using (18)O-labeling and LC-MS-based quantitative proteomics. We developed a novel protocol to both simplify isolation of nanoparticle bound proteins and improve reproducibility. LC-MS analysis identified and quantified 88 human plasma proteins associated with polystyrene nanoparticles consisting of three different surface chemistries and two sizes, as well as, for four different exposure times (for a total of 24 different samples). Quantitative comparison of relative protein abundances was achieved by spiking an (18)O-labeled "universal" reference into each individually processed unlabeled sample as an internal standard, enabling simultaneous application of both label-free and isotopic labeling quantification across the entire sample set. Clustering analysis of the quantitative proteomics data resulted in distinctive patterns that classified the nanoparticles based on their surface properties and size. In addition, temporal data indicated that the formation of the stable protein corona was at equilibrium within 5 min. The comprehensive quantitative proteomics results obtained in this study provide rich data for computational modeling and have potential implications towards predicting nanoparticle biocompatibility.

Figure 1

Experimental scheme of protocol optimization using mouse blood plasma and the 50 nm amine-modified nanoparticles.

Figure 2

Evaluation of the performance of the “SDS buffer elution” protocol and the “on-nanoparticle” digestion protocol. Numbers of identified mouse plasma proteins are shown for acetone-precipitated reference (without nanoparticle), and 50 nm amine-modified nanoparticle bound protein using “SDS buffer elution” and “on-nanoparticle digestion” protocol, respectively (a). The overlap of mouse plasma proteins identified from three process replicates using 50 nm carboxylate-modified nanoparticles was shown for “on-nanoparticle digestion” protocol (b).

Figure 3

Unsupervised hierarchical clustering analysis of the relative abundance of human blood plasma proteins bound to nanoparticles with two different sizes (50 nm and 100 nm), three types of surface modifications (amine, carboxylate, and unmodified), and 4 different incubation time periods (5 min, 15 min, 1 h, and 5 h). Color scheme is based on Log2 values of the ¹⁶O/¹⁸O ratios.

Figure 4

Nanoparticle-bound protein abundances are significantly influenced by a) surface modifications (Apolipoprotein E) and b) nanoparticle size (haptoglobin related protein).

Figure 5

Blood plasma proteins that showed differential binding to nanoparticles based on nanoparticle surface properties and size. a) ANOVA analysis generated 42 nanoparticle-bound proteins with significant abundance changes resulting from surface modification factors (p<0.05), which were clustered into three groups with different abundance change patterns. b) Similarly, 41 nanoparticle-bound proteins with significant abundance changes were resulting from nanoparticle size factors (p<0.05), which were clustered into two groups with different abundance change patterns. Color scheme is based on Log2 values of the ¹⁶O/¹⁸O ratios.

Figure 5

Blood plasma proteins that showed differential binding to nanoparticles based on nanoparticle surface properties and size. a) ANOVA analysis generated 42 nanoparticle-bound proteins with significant abundance changes resulting from surface modification factors (p<0.05), which were clustered into three groups with different abundance change patterns. b) Similarly, 41 nanoparticle-bound proteins with significant abundance changes were resulting from nanoparticle size factors (p<0.05), which were clustered into two groups with different abundance change patterns. Color scheme is based on Log2 values of the ¹⁶O/¹⁸O ratios.