Skip to main content
NCBI home page
Bioactive Materials logoLink to Bioactive Materials
. 2023 Mar 29;27:113–124. doi: 10.1016/j.bioactmat.2023.03.020

Zwitterionic peptides: Tunable next-generation stealth nanoparticle modifications

Clyde Overby a,b, Soomin Park c, Austin Summers d, Danielle SW Benoit a,b,e,f,g,∗
PMCID: PMC10063383  PMID: 37006823

Abstract

Adsorption of proteins to nanoparticles (NPs), a complex process that results in a protein corona, is controlled by NP surface properties that define NP interactions in vivo. Efforts to control adsorbed protein quantity through surface modification have led to improvements in circulation time or biodistribution. Still, current approaches have yet to be identified to control adsorbed protein identities within the corona. Here, we report the development and characterization of diverse zwitterionic peptides (ZIPs) for NP anti-fouling surface functionalization with specific and controllable affinity for protein adsorption profiles defined by ZIP sequence. Through serum exposure of ZIP-conjugated NPs and proteomics analysis of the resulting corona, we determined that protein adsorption profiles depend not on the exact composition of the ZIPs but on the sequence and order of charges along the sequence (charge motif). These findings pave the way for developing tunable ZIPs to orchestrate specific ZIP-NP protein adsorption profiles as a function of ZIP charge motif to better control cell and tissue specificity and pharmacokinetics and provide new tools for investigating relationships between protein corona and biological function. Furthermore, overall ZIP diversity enabled by the diversity of amino acids may ameliorate adaptive immune responses.

Keywords: Nanoparticles, Protein corona, Surface functionalization, Peptide design, Cellular uptake

Graphical abstract

Image 1

Open in a new tab

Highlights

  • •

    Many zwitterionic peptides can act as anti-fouling surfaces for nanoparticles.

  • •

    Nanoparticle protein corona can be tuned by zwitterionic peptide sequence.

  • •

    Many unique peptides can be generated with similar protein interactions.

  • •

    Tuned protein corona via zwitterionic peptide affects uptake in macrophages.

  • •

    Library generation of diverse peptides generates data for future models.

1. Introduction

Nanoparticles (NP) are revolutionary drug delivery systems (DDSs) used in several clinical applications to deliver small molecule and nucleic acid drugs. In addition to enabling successful delivery of therapeutics with rapid degradation, poor bioavailability, and poor circulation times, NP DDSs can alter drug bioactivity via modified biodistribution and cellular interactions. However, the interactions between biological fluids and NPs and the subsequent adsorption of proteins to form a protein corona create many challenges for NP DDS design that remain primarily unresolved [1]. Systemic NP administration has generally proven to have unexpectedly poor tissue targeting mainly due to the mononuclear phagocytic system (MPS) and, particularly, liver macrophage uptake [2,3]. The protein corona, rather than the as-synthesized nanoparticle surface, is the determinant of biological activity of many NP DDS [[4], [5], [6], [7]], and the relationship between protein corona and NP surface is an active area of investigation [[8], [9], [10]].

Systemic delivery of NP DDS is particularly challenging. Upon exposure to blood, a wide range of proteins adsorbs rapidly, including complement proteins, fibronectin, immunoglobulins, albumin, and apolipoproteins. Many of these proteins act as opsonins, resulting in MPS uptake and clearance [2,11]. In parallel, some proteins are dysopsonins, such as apolipoprotein E, clusterin, and serotransferrin, to prolong circulation or alter biodistribution and cellular interactions [12,13]. While efforts have been made to map protein corona to in vivo DDS behavior, designing NP to achieve controllable protein adsorption profiles is non-trivial. NP DDSs are often highly constrained by requirements for colloidal stability, drug loading compatibility and capacity, delivery mechanisms, biocompatibility, and synthesis, leaving limited design space for accommodating protein corona-modifying features except through surface functionalization.

Existing surface functionalization approaches, such as poly(ethylene glycol) (PEG) and zwitterionic polymers, are successful in reducing the quantity of adsorbed protein for many NPs and enhancing systemic circulation time [[14], [15], [16], [17]]. However, these modifications often fail to improve the non-MPS target tissue accumulation [[18], [19], [20]]. While the benefits of these “stealth” or anti-fouling coatings are often attributed to reduced NP-protein interactions, there is evidence that the identity of adsorbed proteins critically affects the in vivo properties of PEGylated NP DDS and, in the absence of specific protein adsorption, stealth effects are mitigated [21]. Established anti-fouling materials, such as PEG and polybetaines, have limited chemical diversity for biasing protein identity within the corona. Additionally, immunological reactions to PEG through consumer product exposure may hinder its long-term use [[22], [23], [24]], and the repetitive chemical structures of alternative materials may pose similar immunological challenges [25]. Purified protein pre-adsorption strategies to form a defined NP corona before systemic administration has shown promise [1], but these strategies add multiple additional manufacturing and purification steps that increase complexity and expense. In sum, new methods are needed to control adsorbed protein corona quantity and identity and avoid adaptive immunity.

Simple zwitterionic peptides composed of lysine and glutamic acid have proven efficacious in reducing protein adsorption to gold surfaces [26] and enhancing protein drug circulation time [27]. Still, the tremendous diversity of possible zwitterionic sequences has yet to be explored. Using only lysine, glutamic acid, and aspartic acid, 1.4 × 107 15-mer peptides can be synthesized. Including serine increases this to over 109 15-mer peptides. This diversity may enable the selective adsorption of proteins and avoid adaptive immune responses to drug delivery systems. Standard and well-established peptide synthesis techniques allow for the precise chemical definition of peptides, allowing both significant and subtle variations to be explored. Zwitterionic anti-fouling properties are driven by charge. The many possible ZIP compositions of cationic and anionic amino acids allow for many chemically unique surfaces with potentially tunable protein adsorption behaviors. However, the relative contributions of charge motif and amino acid composition to ZIP-protein interactions have yet to be examined.

To this end, diverse zwitterionic peptides (ZIPs) were designed to explore anti-fouling behavior using gold NPs as a model nanoparticle system as an alternative functionalization strategy for NPs instead of PEG. We hypothesize that ZIPs will act as tunable anti-fouling NP surface modifications, controlling protein adsorption quantity and identity as a function of sequence and charge motif and not bulk amino acid composition. This approach could provide tremendous potential utility in tunable surface functionalization for NP DDSs, which are currently limited by material choice. Specifically, this study investigated a related set of 32 zwitterionic peptide-conjugated 30 nm gold nanoparticles (NPs) with PEG functionalization as a control. NPs were evaluated for colloidal stability using dynamic light scattering and zeta potential analyses, human serum protein adsorption through the bicinchoninic acid assay, adsorbed protein identity via SDS-PAGE and mass-spectrometry, and phagocytic uptake in macrophages.

2. Materials and methods

Materials. All materials were used as received without further purification. Ultrapure water was used for all synthesis and washes. All chemicals were acquired from Sigma-Aldrich unless otherwise specified. All solutions use ultrapure water as a solvent unless otherwise specified.

Gold Nanoparticle Synthesis. Gold nanoparticles were synthesized using a two-step reduction procedure described in [63]. Briefly, in the first step, 15 nm gold NP was synthesized by adding 1 mL 3% w/v trisodium citrate to a rapidly stirred boiling solution of 100 mL 0.25 mM gold chloride solution in ultrapure water for 10 min. The reaction was then chilled on ice and quenched with 100 mL of ice-cold ultrapure water. The resulting nanoparticles were confirmed to be 15 nm via a 518 nm absorbance peak and dynamic light scattering. To synthesize 30 nm gold NP, 98 equivalents of ultrapure water was cooled to 4 °C, and a 1:1:12.7 ratio of 15 mM trisodium citrate, 25 mM gold chloride, and 2.4 mM 15 nm gold nanoparticles were added. While rapidly mixing, one equivalent of 25 mM hydroquinone was quickly added to the flask. The reaction was allowed to complete overnight while stirring at 4 °C. Size was confirmed with absorbance (526 nm peak) and dynamic light scattering. Next, 30 nm gold NP stock solutions were concentrated via centrifugation at 3200g in 50 mL conical tubes for 30 min. An equivalent of 5 mL of OD1 30 nm gold NP stock was added to each tube. After removing the supernatant, four tubes of gold NP concentrated were combined and washed with 750 μL of 0.25 mM trisodium citrate and 0.05 v/v Tween20 wash buffer. Three more rounds of centrifugation and two more 750 μL washes were performed at 3500 RCF for 30 min in 1.7 mL microcentrifuge tubes, and the final resuspension after centrifugation was with 100 μL ultrapure water to yield washed gold NP concentrate.

Peptide Computational Design. An algorithm was developed to predict peptide sequences with defined β-strand interaction energies derived from Trovato et al. [30]. Peptides were limited to 15-mers and included positive, negative, or neutral amino acids, each with numerous possible specific amino acid probabilities. Sequences are generated in silico randomly to achieve overall charges within ±1 with free acid and free base termini. A peptide β-strand interaction prediction algorithm derived from [30] was used to score the average interaction energies of each charge sequence with itself, and only peptides with minimum (attractive) and maximum (repulsive) ΔG per batch were synthesized (typically −6 kJ/mol and +2 kJ/mol). All peptides are 15-mers and approximately 1.7–2.0 kDa. All sequences and compositions are listed in Table 1. 2 kDa PEG was used as a control of similar molecular weight as the ZIPs, and simple lysine/glutamic acid and lysine/aspartic acid repeats were used for comparison.

Table 1.

List of synthesized zwitterionic peptides. Charges in each charge motif sequence are represented with + for positively charged amino acids, 0 for neutral amino acids, and - for negatively charged amino acids. All peptides have an N-terminal cysteine and a C-terminal glycine. Each charge sequence can be synthesized with various amino acids without changing the net charge or order of charges. Amino acid composition indicates which amino acids were used for each sequence. For example, peptide 7.KES is CKKKEESEEKESESKKG, while 7.KEN is CKKKEENEEKENENKKG.

Sequence Number Charge Motif Amino Acid Composition Net Charge Predicted Self-Assembly Energy (kJ/mol)
1 0–0+-- + -0+-0++ KES −1 −2.9
2 ++- + -00+0----++ KDN 0 0.33
3 ++- + --0+00- + --+ KEN 0 1.0
4 +00+-- + - + ---+0+ KES 0 −1.0
5 +++-0-- + --0-++0 KES 0 −0.43
6 +0+--+0+--+0+-- KES 0 −0.06
7 +++--0-- + -0-0++ KES KDS KEN 0 1.5
8 0--+0-++--+0-++ KES KDS KEN 0 −1.0
9 +0+0-- + -- + --0++ KES KDS KDN 0 1.3
10 ++-0+--0++-0+-- KES KDS KDN 0 −0.94
11 ++-0-++0–0++-- KEN KET KDT 0 0.98
12 000+--+++---++- KEN KET KDT 0 −4.2
13 +-++-0-0-++-0-+ KES KDS KDT 0 0.36
14 0-+0----++++0-+ KES KDS KDT 0 −4.2
15 +- + - + - + - + - + - + -+ KE KD 0 −5.6
Open in a new tab

Peptide Similarity Analysis. To compare the similarity of different ZIPs, a MATLAB script was used. Briefly, the method was to find the number of positions that differed between two aligned sequences (either with itself or with another peptide) and repeat the measurement when shifting the alignment of the peptides by increasing values until all possible alignments were considered. The maximum value of every peptide's possible alignment was recorded. For cluster analysis, individual maximums of every unique peptide comparison within the clusters being compared were averaged by the number of unique comparisons. Standard deviations were calculated from all unique maximum values. The same method was used to compare charge motifs by first converting the peptides from single-letter amino acid sequences to −1, 0, or +1 values based on the charge of either negative, neutral, or positive values, respectively.

Peptide Synthesis. Peptides (ZIPs) were synthesized using solid phase peptide synthesis at 0.1 mmol scale. All peptide sequences were designed with a C-terminus cysteine for conjugation to gold. 0.4 M Fmoc-protected amino acid solutions were made in dimethylformamide within 1 week of synthesis. Fmoc-Gly-Wang resin was used (Sigma-Aldrich). Deprotection cycles used 10% w/v piperazine in 90/10% dimethylformamide (DMF)/ethanol solution. Coupling reactions were performed at 4-fold molar excess using 1 M diisopropyl carbodiimide and 1 M OxymaPure solutions in DMF. After synthesis, peptides were cleaved using 4 mL 90%/2.5%/2.5%/2.5%/2.5% trifluoroacetic acid/3,6-dioxane-1,8-octane-dithiol/triispropysilane/thioanisole/water for 4 h at room temperature and then precipitated into 50 mL diethyl ether. Precipitates were centrifuged at 3200 RCF for 10 min, the supernatant decanted, and then the pellet was washed in 25 mL diethyl ether and centrifuged twice more before drying overnight under vacuum. Synthesis was confirmed using MALDI-ToF (Shimadzu Axima Performance). All peptide reagents purchased from Aapptec unless otherwise specified.

Gold Nanoparticle Conjugation. Dried, purified peptides were added to ultrapure water to yield 2 mM solutions. 4.0 μL 1 mM peptide solutions were added to 50 μL gold NP concentrate solution (estimated total surface area ∼51 cm2) to give a target 1 peptide/nm2 coverage on the gold NP. Peptide-AuNP solution was heated to 60 °C for 1 h to complete the reaction. Peptide-NP was then centrifuged at 3500 g for 30 min. The resulting supernatant was tested using Ellman's reagent to detect free thiol compared to the pre-conjugation peptide thiol concentration. The difference was used to determine the conjugation efficiency (∼100% for one peptide/nm2). The remaining supernatant was removed, and NP conjugate was washed in 750 μL of 0.25 mM trisodium citrate and 0.05% v/v Tween20, centrifuged at 3500 RCF, the supernatant discarded, and the pellet was resuspended in 50 μL water.

Protein Adsorption. ZIP-NPs 51 cm2 in 50 μL were added to 450 μL pooled male AB human serum (Sigma-Aldrich H4522) in 1.7 mL microcentrifuge tubes. Samples were incubated at 37 °C for 2 h while gently agitated on a shaker plate. Samples were then spun down at 3500 g for 30 min. The supernatant was discarded, and 750 μL of 1x phosphate buffered solution with 0.25 mM trisodium citrate and 5 mM EDTA were used to wash the pellet before another round of centrifugation and washing. After the second wash, the sample was centrifuged at 3500 RCF for 30 min, the supernatant discarded, and the pellet resuspended in 100 μL 1% SDS in PBS overnight.

Adsorbed Protein Quantification. ZIP-NP in 1% SDS solution was centrifuged at 3500 g for 30 min, and the supernatant was collected for quantification. 20 μL of supernatant per replicate was added to a flat-bottomed, transparent 96-well plate, with 3 replicates per condition and three technical replicates. The Pierce BCA assay kit (ThermoFisher 23225) was used to quantify protein abundance using bovine serum albumin as a standard, with 200 μL of BCA solution added to each sample. Plates were incubated at 37 °C for 30 min before sample absorbance at 562 nm was read on a plate reader (Biotek Citation 5). Technical replicates were averaged, and samples were compared using 1-way and 2-way ANOVAs with Tukey's posthoc test.

Adsorbed Protein Gel Electrophoresis. 20 μL adsorbed protein in the supernatant of the washed NPs was mixed with 5 μL of 5x lane marker reducing loading buffer (Thermo). Protein solutions were then incubated at 95 °C for 10 min and cooled. 20 μL of the reduced protein were loaded per lane in a 4–20% miniPROTEAN gel along with BioRad Precision Plus Unstained Protein Ladders. Gels were run at 100 V for ∼75 min in SDS/Tris/Glycine buffer. Gels were then fixed with 50% methanol, 7% acetic acid, and 43% ultrapure water and stained with SyproRuby overnight. Gels were washed in 10% methanol, 7% acetic acid, and 83% ultrapure before rinsing in water. Gels were then imaged using a BioRad GelDoc at 0.4s exposures.

SDS-PAGE Image Analysis. Imaged gels were despeckled using ImageJ before processing in GelAnalyzer. All peaks were chosen manually. Relative mobility (Rf) was calibrated to the BioRad Precision Plus Unstained Protein Ladders. Band area and Rf for each lane were then binned by molecular weight. The summed intensity of each band within each molecular weight bin was then used for histogram plots and clustering analysis.

SDS-PAGE Protein Analysis. Histograms of quantified band intensities were plotted as histograms and analyzed using hierarchical clustering. Hierarchical clustering is an unsupervised machine learning technique for unbiased grouping of data by similarity (JMP SAS), and the resulting output is a grouping of data sets known as clusters. Clusters with fewer than 3 members were excluded from the analysis, leaving 4 clusters. Each cluster, consisting of 4–12 members, was then evaluated for sequence and compositional similarity with each other cluster. Similarity was determined by the maximum length of identical sequence or charge motif without registering exact positional alignment. Averages and variances were taken for each category of comparison before 1-way ANOVA analysis with Tukey's post-hoc testing.

Proteomics Mass Spectrometry Sample Preparation. As above, 20 μL of washed NP supernatant was mixed with 5 μL of 5x lane marking reducing loading buffer (Thermo), incubated at 95 °C for 10 min, and cooled. Next, 20 μL of the reduced protein was loaded per lane in a 4–20% miniPROTEAN gel, which was run at 100 V briefly (∼5–10 min) until the lane marker moved approximately 1 cm. Gels were removed and stained with Bio-Safe™ Coomassie stain for 30 min before overnight destaining in ultrapure water. Gels were then submitted to the University of Rochester Mass Spectrometry Resource Laboratory for proteomics analysis.

Proteomics Sample Digestion. Sections of protein gels containing stained protein were cut into 1 mm cubes, de-stained, reduced with dithiothreitol, and alkylated with iodoacetamide before dehydration with acetonitrile. Next, 10 ng/μL of trypsin in 50 mM ammonium bicarbonate was added to the dehydrated gel until the gels were covered. The solutions were incubated at room temperature for 30 min before additional ammonium bicarbonate solution was added to completely submerge the gel, followed by incubation at 37 °C overnight. Peptides were extracted with 50% acetonitrile, 0.1% trifluoroacetic acid (TFA), and water before drying in a CentriVap concentrator (Labconco). Peptides were desalted using homemade C18 spin columns before drying and final reconstitution in 0.1% TFA.

Peptide Fragment Mass Spectrometry. Peptides from each fraction were injected onto a homemade 30 cm C18 column with 1.8 μm beads (Sepax), with an Easy nLC-1200 HPLC (Thermo Fisher), connected to a Fusion Lumos Tribrid mass spectrometer (Thermo Fisher). Solvent A was 0.1% formic acid in water, while solvent B was 0.1% formic acid in 80% acetonitrile. Ions were introduced to the mass spectrometer using a Nanospray Flex source operating at 2 kV. The gradient began at 3% B and held for 2 min, increased to 10% B over 6 min, increased to 38% B over 35 min, then ramped up to 90% B in 5 min and was held for 3 min, before returning to starting conditions in 2 min and re-equilibrating for 7 min, for a total run time of 60 min. The Fusion Lumos was operated in a data-independent mode. The full MS1 scan was done over a range of 390–1010 m/z, with a resolution of 60,000 at m/z of 200, an AGC target of 4e5, and a maximum injection time of 50 ms. Precursor ions were measured using a staggered windowing scheme of 20 m/z with 10 m/z overlaps. For example, the first cycle fragmented precursor ions between 400 and 420 m/z, 420–440 m/z, etc., with the final window fragmenting ions between 980 and 1000 m/z. The next cycle fragmented precursor ions between 390 and 410 m/z, then 410–430 m/z, etc., with the final fragmentation of the cycle being ions between 990 and 1010 m/z. Each cycle consisted of 30 MS2 scans, with fragment ions collected between 200 and 2000 m/z. Precursor ions were fragmented by higher energy C-trap dissociation (HCD) using a collision energy of 33%. MS2 scans were collected in the orbitrap with a resolution of 15,000, an AGC target of 4e5, and a maximum injection time of 23 ms.

Mass Spectrometry Data Analysis. The raw data were processed with DIA-NN version 1.8.1 (https://github.com/vdemichev/DIA-NN) [64]. Data analysis was carried out for all experiments using library-free analysis mode in DIA-NN. To annotate the library, the human UniProt ‘one protein sequence per gene’ database (UP000005640, downloaded 4/7/2021) was used with ‘deep learning-based spectra and RT prediction’ enabled. For precursor ion generation, the maximum number of missed cleavages was set to 1, a maximum number of variable modifications to 1 for Ox(M), peptide length range to 7–30, precursor charge range to 2–3, precursor m/z range to 400–1000, and fragment m/z range to 200–2000. In addition, the quantification was set to ‘Robust LC (high precision)’ mode with normalization turned off, MBR enabled, protein inferences set to ‘Genes,’ and ‘Heuristic protein inference’ turned off. MS1 and MS2 mass tolerances, along with the scan window size, were automatically set by the software. Precursors were subsequently filtered at library precursor q-value (1%), library protein group q-value (1%), and posterior error probability (50%). Protein quantification was carried out using the MaxLFQ algorithm as implemented in the DIA-NN R package (https://github.com/vdemichev/diann-rpackage), and the number of peptides quantified in each protein group was counted as implemented in the DiannReportGenerator Package (https://github.com/kswovick/DIANN-Report-Generator) [65].

Proteomics Data Processing. Adsorbed protein identities and abundances were averaged over multiple runs per sample and then ranked by abundance. Proteins with fewer than 10 peptide fragments detected on average over all samples were eliminated from the list, as was the trypsin used for digestion, leaving 45 proteins. The mean abundance per protein over all samples was calculated and used to normalize each protein's relative abundance to the mean. A log2 transformation was applied to the relative abundance to regularize the data for analysis. For values where a protein was not detected in a sample, it was set to −6.

Proteomics Data Analysis. The trimmed and log-transformed regularized relative protein abundance was analyzed using k-means clustering and hierarchical clustering. Using JMP 16.0 software, charge motifs were clustered using hierarchical clustering with cluster numbers identified using the maximum cubic clustering criterion. Means and standard deviations were calculated within each cluster and compared.

Phagocytic Cell Uptake. RAW 264.7 cells were cultured in high glucose (4.5 g/L) Dulbecco's modified Eagle's medium with 10% fetal bovine serum supplement at 37 °C in 5% CO2. For uptake experiments, cells were plated at 16,000 cells/cm2 on Grenier Bio-One Cellstar 24 well plates 24 h before dosing with ZIP-NPs, PEG-NPs, or NPs. NP conjugates were pre-incubated in 10% human serum (Sigma-Aldrich H4522) in PBS for 2 h at 37 °C before washing. To wash, samples were spun down at 3500 g for 30 min. The supernatant was discarded, and 750 μL of 1x phosphate buffered solution with 0.25 mM trisodium citrate and 5 mM EDTA were used to wash the pellet before another round of centrifugation. Supernatant was removed and serum-adsorbed NP conjugates were resuspended in 100 μL 1x PBS. 60 μg of NP (30 μL) was added to 1.2 mL of AlphaMEM serum-free media. Cells were washed with warm 1x PBS twice (approximately 400 μL) before 400 μL of NP in AlphaMEM was added to each well (20 μg NP/well dose). Cells were incubated at 37 °C in 5% CO2 for 2 h before being washed with 1x PBS thrice. Cells were kept in 500 μL 1x PBS for imaging. Images were taken using a Biotek Cytation 5 with a 20x objective in color brightfield. Image data was redacted, and the samples were analyzed visually for the presence of red or black cellular inclusions indicative of gold NP uptake.

Statistical analysis. All statistical analyses, unless otherwise specified, were conducted using GraphPad Prism 7.04. Programming was conducted in either Python 3.8 or Matlab 2017a.

3. Results and discussion

Protein adsorption to NPs occurs immediately upon exposure to biological fluids. The quantity and identity of adsorbed proteins depend on the fluid's composition and NP surface properties. Serum is a particularly challenging but very relevant biological fluid for analyzing NP protein corona formation. NPs must maintain colloidal stability in the presence of serum to maintain circulation and avoid activating the complement system [1]. Specific serum proteins within the NP corona enhance circulation time, reduce phagocytic uptake, and reach specific tissues. Still, there is limited ability to tune NP surface chemistry to control the adsorption of specific proteins. For zwitterionic peptides, it is unclear whether charge motif or average amino acid composition affects these different aspects of NP-serum interactions. To explore these relationships, myriad ZIP charge motifs with various amino acid compositions were conjugated to gold nanoparticles (Scheme 1).

Scheme 1.

Scheme 1

Open in a new tab

Zwitterionic peptide (ZIP) and ZIP-NP library generation. Gold nanoparticles are functionalized with different sequences of zwitterionic peptides that a common charge motif or a common composition can group.

3.1. Computational design enables generation of a diverse zwitterionic peptide library

The peptides investigated herein were designed to attain diversity with minimal constraints and patterning bias. Peptides were first developed as amino-acid agnostic charge patterns, termed charge motifs (CM). Charge motifs were created randomly by incorporating positive, negative, and neutral residues using limits on net charge (±1) and the number of neutral residues arbitrarily limited to 3 or 4. These parameters are arbitrary but rationalized for peptide diversity; even 3 neutral residues tremendously increase the possible number of both charge motifs and individual peptides without changing synthetic difficulty, as does allowing a slight variation in charge.

Using this approach, thirty 15-mer peptides with zwitterionic characteristics were designed with C-terminal cysteines to enable conjugation to 30 nm gold NP through thiol-gold reactions. Sequence length was chosen to approximate 2 kDa PEG, which has been previously reported as effective for reducing protein adsorption to gold NP [28]. Amino acids were chosen based on charge, hydrophilic properties, and ease of peptide synthesis. To limit complexity, sequences were limited to natural amino acids. Only lysine was included in sequences of positively charged amino acids due to the synthetic intractability of arginine (lower reactivity, longer reaction cycles, and more complex deprotection chemistries) and the neutral charge of histidine at physiological pH. Both negatively charged amino acids, glutamic acid and aspartic acid, were used. Polar uncharged amino acids serine, threonine, and asparagine were used to increase peptide sequence diversity while maintaining similar hydrophilicity to the charged amino acids.

To guide the initial selection of peptides, charge motifs were computationally evaluated for self-assembly potential through β-strand interactions, which have been previously reported as a driver of peptide-NP conjugate stability [29], using an adaptation of an existing aggregation prediction algorithms [30] (see Methods). These β-strand interactions account for amino acid backbone and side-chain interactions without considering secondary or tertiary structure. However, short peptides typically lack defined structure, and NP surfaces can partially denature proteins [31], leading to β-strand interactions dominating other types of interactions [32]. This method was also chosen for computational simplicity, allowing for unbiased screening of 100,000 potential peptides on a personal computer on hour timescales. In contrast, more sophisticated peptide-peptide interaction models are too computationally intensive to perform in this way. Charge motifs were selected with a range of self-interaction potentials of 5000 randomly assembled and evaluated charge motifs in-silico, which were then used for composition assignment.

The resulting sequences and their associated parameters are listed in Table 1. This constraint produced only net-neutral zwitterionic sequences except for charge motif 1. Sequences 1–6 enabled testing of the influence of charge motifs with similar compositions, while sequences 7–14 enabled analyses of the impact of composition within the same charge motifs. Amino acid compositions were designed to include all combinations of the previously selected positive, negative, and neutral amino acids with a minimum coverage of at least three peptides of each composition. Sequence 15 (Table 1), a 15-mer lysine/glutamic acid repeat, was included for comparison with a previous report [33]. Peptides were synthesized using solid-phase peptide synthesis for conjugation to NPs (see Methods).

3.2. ZIP-functionalized NP are colloidally stable

ZIPs were conjugated to gold NP via N-terminal cysteine thiols. Dynamic light scattering (DLS) and zeta potential measurements were used to interrogate the effects of ZIP surface modification on NP size in PBS and after serum incubation compared to PEG-functionalized and unfunctionalized controls (Fig. 1). In general, ZIP functionalization increased NP diameter to a greater extent than PEG despite similar molecular weights, with many ZIPs doubling NP size, suggesting some NP aggregation. ZIPs increased NP hydrodynamic diameter from 32 nm to 50–100 nm. (Fig. 1A), yielding colloidally stable NPs with zeta potentials of −20 to −40 mV (Fig. S1). All modifications (ZIPs and PEG) were >90% conjugated as measured by a thiol depletion assay [28] (see Fig. S2). PEG functionalization resulted in a small (∼15 nm) increase in size consistent with PEG monolayer functionalization. ZIP sequence (Fig. 1B) and composition (Fig. 1C) had no significant effect on size (p = 0.55, 2-way ANOVA). While not statistically significant due to large variability, a trend was observed for NP functionalization, with charge motifs 7, 8, and 14 having the smallest change in size. At the same time, 9, 10, and 12 more than doubled the original NP hydrodynamic diameter. This suggests some NP aggregation was occurring in ultrapure water but not to the extent of colloidal instability. Suspending the NP conjugates in pH 7.4 phosphate-buffered saline (PBS), which is similar in pH and ionic concentration to blood, for 24 h led to colloidal instability of unfunctionalized NP, as expected [34], while PEG-NP and the majority of ZIP-NP conjugates were stable (Fig. 1D). Citrate-capped gold NPs are stabilized by electrostatic repulsion, which is inhibited in the presence of salts. In contrast, PEG-NPs are stabilized by hydration-mediated steric hindrance [35], independent of ionic conditions. The stability of CMs 2, 3, 7, 8, 12, and 14 in 1x PBS (150 mM NaCl) suggests that the stability of these ZIP-NPs is not driven by long-range electrostatic repulsion, which is expected from their net neutral charge, and is likely to be a similar steric hindrance from strong hydration effects of local charges [14,36,37].

Fig. 1.

Fig. 1

Open in a new tab

Zwitterionic peptide-conjugated NPs are colloidally stable in water regardless of composition or charge motif as measured by DLS (A). DLS size data organized by sequence (B) and amino acid composition (C) reveals no significant differences between charge motifs or compositions. Size increases for all NP conjugates over NP, consistent with surface conjugation. For B, C, E, and F data markers represent mean values of technical triplicates for independent experiments. When suspended in physiological pH 7.4 phosphate-buffered saline for 24 h at 4 °C, NP conjugates exhibit variable colloidal stability (D). NP conjugates with a Z-avg size greater than 200 nm were marked as unstable. After incubation in human serum for 2 h, washed NP conjugates were analyzed in pH 7.4 PBS for size as a function of charge motif (E) and amino acid composition (F). * indicates p < 0.05 for comparison. Statistical comparisons were made by 1-way ANOVA with Tukey's posthoc testing. Error bars indicate standard deviation.

ZIP-NPs were incubated with serum for 2 h, and sizes were measured via DLS (Fig. 1E and F). Serum protein adsorption is used as a model for blood exposure, as would be experienced for systemic administration of NPs. Serum incubation reduced hydrodynamic diameters for all ZIP-NPs compared to water, suggesting the adsorption of serum proteins may further stabilize and reduce aggregation of the ZIP-NPs. All NP conjugates were stable in serum and 1x PBS after protein adsorption. The formation of stabilizing protein coronas has been described for gold NPs previously and has been attributed to the replacement of citrate moieties with a hard protein corona of serum proteins [34]. This effect in ZIP-NPs may be due to a similar mechanism of hard corona formation and suggests that adsorbed proteins are not significantly denatured, which would lead to irreversible aggregation [31].

Additionally, all serum-incubated ZIP-NPs had similar sizes to serum-incubated NP. In contrast, PEG-functionalized NP was significantly smaller than NP after serum incubation, consistent with the anti-fouling properties of PEG [28].

3.3. ZIP conjugation reduces NP protein adsorption as a function of sequence and composition

To investigate ZIP anti-fouling properties, the adsorption of serum proteins onto functionalized NPs was analyzed. The total mass of protein adsorbed by ZIP-NP in serum was reduced up to 60% compared to NP, varying with ZIP charge motif and composition (Fig. 2). No statistically significant differences in serum adsorption were observed due to ZIP composition (Fig. 2B). Comparing bulk compositions (Fig. 2C) reveals that ZIPs comprised of KDS and KET have significantly reduced protein adsorption compared to other ZIPs, averaging a ∼50% versus ∼25% reduction of adsorbed protein compared to NP controls. KES-functionalized NP had highly variable protein adsorption between individual peptides. When analyzed individually, peptides with the KES composition and varying charge motifs, protein adsorption ranged from 40 to 100% of unfunctionalized NP controls (∼0.23–0.58 μg/cm2) (Fig. 2D). PEGylation significantly lowered protein adsorption by 80% versus NP controls (from 0.58 to 0.12 μg/cm2) compared with all ZIP-NPs except 14.KES.

Fig. 2.

Fig. 2

Open in a new tab

Serum protein adsorption on NPs is reduced by ZIP conjugation with a strong dependence on peptide sequence as measured by bicinchoninic acid assay normalized to citrate-stabilized gold NP. (A) Summary of all protein adsorption data. Data is organized by charge motif (B), the bulk composition of various sequences (C), and peptides of the same design with differing sequences (D). (B) Charge motif is not predictive of serum protein adsorption abundance. (C) Composition analysis reveals KDS and KET to have reduced protein adsorption compared to other ZIP compositions. (D) Sequence within KES strongly influences protein adsorption, varying from 100% to 40% of unfunctionalized NP control protein adsorption. (E) Protein adsorption is not an effect of NP size before incubation. All samples absent not significant (n.s.) are significantly different (p < 0.05) from other groups. # indicates statistical significance versus all other samples not marked by n. s. Statistical comparisons were made using 1-way ANOVA and Tukey's posthoc test. All data points are averages of quadruplicate measurements of independent experiments, n = 3–30. All error bars indicate standard deviation.

A significant reduction in protein adsorption via PEGylation is expected based on previous data [28]. Still, the variation in protein adsorption by ZIP-functionalization and the impact of sequence and composition has yet to be reported. The observed variation in adsorption is not due to ZIP-NP size prior to serum exposure (Fig. 2E) and is, therefore, not an effect of NP surface area. The variation may be due to the identity of proteins adsorbing to the various ZIP-NP surfaces. However, the abundance and identity of the adsorbed proteins are not necessarily correlated. The specificity of protein adsorption may be due to underlying electrostatic interactions between the proteins and peptides similar to those observed in protein-protein interactions [38]. Still, a linear regression analysis suggests that differences in protein adsorption are not due to the ZIP terminal charge or the number of positive, negative, or neutral residues within the eight peptides of the ZIP N-terminus or any other simple correlations such as NP size or ZIP-protein β-strand interaction energy (see Fig. S3). Predicting peptide-protein interactions is only possible with an experimentally determined starting point [39]. Therefore, in the small library of zwitterionic peptides used here for screening, it is possible ZIP sequence-specific binding is occurring, thus motivating the experimental investigation of protein corona identity.

3.4. ZIP charge motif correlates with corona protein profiles

To interrogate protein adsorption patterns of various ZIP-NPs, the molecular weight of adsorbed serum proteins was evaluated using SDS-PAGE [40]. Using a methodology adapted from [28], serum proteins within NP coronas were electrophoresed in 4–20% polyacrylamide gels under reducing conditions and stained with fluorescent protein-dye (SYPRO Ruby™) for imaging. This technique resolved 15–30 bands for each sample, but similar molecular weights of many serum proteins prevented direct protein band comparisons between samples. Instead, bands were separated into molecular weight ranges, and relative intensities were compared (Fig. 3A). Hierarchical clustering analysis of the resulting histograms was performed to enable unbiased data grouping by similarity using the maximal cubic clustering criterion [41], revealing 4 clusters with more than three members (Fig. 3B).

Fig. 3.

Fig. 3

Open in a new tab

Cluster analysis of adsorbed human serum proteins to various ZIP-NP conjugates shows adsorption profiles correlate strongly with motif charge and not composition. (A) Histograms of SYPRO Ruby stained protein fluorescent intensities binned by molecular weight determined via SDS-PAGE. (B) Unsupervised hierarchical clustering analysis of SDS-PAGE histograms finds high sequence diversity within the clusters and minimal correlation between cluster and ZIP amino acid composition. Clusters with fewer than 3 NP members were excluded from the analysis. (C) Peptide sequences within clusters were compared to other members of the same cluster and all other peptides based on sequence and charge identity. Charge motif differences within groups were lower than sequence differences, p = 0.038. (D) Each cluster's average amino acid composition was calculated and compared; No cluster had significant compositional differences from any other group. (E) Cluster sequences compared directly with sequences in other clusters reveals that peptide members of each cluster are more similar to their cluster than other clusters. * indicates p < 0.05 vs. combined sequence similarity. 2-way ANOVA with Tukey's posthoc testing was used for statistical comparisons, n = 3–30. Error bars indicate standard deviation.

Comparing ZIPs within each cluster revealed a remarkable similarity of charge motifs versus sequence similarity (p < 0.05) by 2-way ANOVA, indicating that charge motifs with different amino acid substitutions preferentially grouped within the same cluster (Fig. 3C). No cluster of sequences had significantly different amino acid composition compared to other clusters, indicating that clustering was not a result of composition (Fig. 3D). Together, the charge motif similarity and lack of significance of compositional variation suggest that cluster groupings are driven by charge motif and sequence. Indeed, when comparing the sequences of each cluster, clusters 1 and 2 share approximately 34% sequence similarity, while clusters 3 and 4 bear less resemblance than other clusters (Fig. 3E). Cluster 4 shares less than 14% the sequence similarity of clusters 1, 2, and 3, and clusters 2 and 3 have only 20% sequence similarity. All clusters had at least 44% internal similarity, further suggesting that sequence similarities dominate clustering. The significant difference in cluster sequences combined with no considerable composition differences indicates that the ZIP charge motifs contribute to protein corona profiles while the underlying ZIP amino acid composition does not.

3.5. Proteomic analysis of ZIP-NP protein adsorption profiles reveals dependence on ZIP charge motif

Proteomic analysis was performed on charge motifs 1, 3, 7, and 8, as these charge motifs had low aggregation, significant differences in protein adsorption, and were present in 3 of the 4 clusters identified via protein gel analyses. 111 unique proteins were identified within the corona (see SI), of which 45 were quantitatively comparable (Fig. 4A). This aligns with previous experiments in which NPs of ∼30 nm maximally adsorb 20–60 serum proteins as part of the “hard corona,” which is the surface layer of the adsorbed proteins with the most significant interaction directly with the NP [13,42]. This hard corona drives the specificity of the soft corona that forms secondarily on the NPs [43]. Corona proteins are listed in order of average integrated mass-spectrometry signal. Molecular weight distributions (Fig. 4B) and isoelectric point distributions (Fig. 4C) reveal that the majority of proteins adsorbed are 40–80 kDa with a pI < 6, consistent with serum proteins [44]. Hierarchical clustering analysis of the NP conjugates revealed a similar clustering pattern to the SDS-PAGE clustering analysis, with 3.KEN-NP, PEG-NP, and NP have the greatest separation from the more similar 1.KES-NP, CM-7-NPs, and CM-8-NPs (Fig. 4D). Proteins were then categorized by function (see SI Table 1) and compared between charge motifs and the bulk average of all ZIPs (Fig. 4E). CM-1 had significantly higher transport protein adsorption than other ZIP-NPs, PEG-NP, and NP, while CM-7 had significantly lower immunoglobulin adsorption than unfunctionalized NP. Complement protein adsorption was significantly lower in all ZIP-NPs compared to NPs as was the case for apolipoprotein adsorption except for CM-3. Notably, individual protein adsorption varied more than the average functional groupings. Certain proteins, such as apolipoprotein A1 (APOA1), apolipoprotein E (APOE), and clusterin (CLU), have been reported to have significant effects on nanoparticle behavior in vivo through either enhanced circulation time or changes in biodistribution [13]. Comparing the adsorption of select proteins from the functional categories between the individual charge motifs and ZIP-NPs revealed significant differences between ZIP-NP adsorption profiles for Apolipoprotein A4 (APOA4), APOE, apolipoprotein H (APOH), albumin (ALB), and serotransferrin (TF) (Fig. 4F).

Fig. 4.

Fig. 4

Open in a new tab

Proteomics analysis of proteins adsorbed to nanoparticles indicates substantial differences in adsorbed protein identity as a function of charge motif. (A) Heatmap of relative abundance of the top 45 most abundant proteins identified within NP coronas compared to mean values of all samples. (B) The histogram of identified protein masses has a median value of 70 kDa. (C) The histogram of identified protein isoelectric points has a median value of 6.0. (D) Hierarchical clustering of the NP conjugates, with colors indicating clusters as identified using SDS-PAGE analysis previously. The relative hierarchy of the clustering is similar to that identified in SDS-PAGE. (E) Comparison of categorized protein adsorbed of ZIP-NPs batched by charge motif (see SI Table 1 for categorization criteria). CM-1-NP showed a significant increase in transport protein adsorption compared to all other NPs and NP-conjugates. Both PEG-NPs and ZIP-NPs significantly reduced average apolipoprotein and complement protein adsorption. CM-7-NP significantly reduced immunoglobulin binding compared to other ZIP-NPs. (F) Comparison of select proteins of interest adsorbed to ZIPs batched by charge motif. ZIP-NPs had reduced APOA1 and APOA4 adsorption compared to NP, but no effect was observed with APOB. CM-3-NP showed selective enhancement of APOE adsorption. CM-7-NP had significantly enhanced ALB and TF adsorption compared to all other NPs and NP-conjugates. # indicates p < 0.05 vs mean of all ZIP-NP, $ indicates p < 0.05 vs PEG-NP, and % indicates p < 0.05 vs NP. 1-way ANOVA with Tukey's posthoc analysis on raw abundance data per category/protein, n = 2–8.

The most significant variance was found for TF and APOE adsorption, which has been shown to enhance non-MPS target tissue biodistribution [7,12] and circulation time in NP drug delivery systems [9,13], respectively. ZIP 1.KES-NP serotransferrin abundance was 5-, 11-, and 30-fold higher than all other ZIP-NPs, unfunctionalized NP, and PEG-NP, respectively. APOE had a 9-fold greater abundance on CM-3-functionalized NPs than other ZIP-NPs and was similar to unfunctionalized NP and PEG-NP.

Proteomics analysis shows ZIP functionalization significantly reduces the overall abundance of complement proteins versus NP (Fig. 4E). Indeed, complement proteins are nearly universally observed within NP corona in vivo due to antigen-specific binding to existing proteins in the corona [45] or through non-specific adsorption [46]. Interestingly, proteins in all three complement pathways are observed in the corona. While the alternative pathway can be activated via non-specific C3 adsorption, activation of the classic or lectin-mediated complement is unlikely, as the pathways rely on antigen-specific conformational changes to initiate enzymatic amplification [47,48]. Specific antibody binding against ZIPs is not expected from the pooled human serum used, as the ZIPs used are both unique to this study and do not match existing protein domains. Nevertheless, proteins in the classic and lectin-mediated pathways were observed in the corona, including C1, C4, Ficolin-3, and Mannose-Associated Serine Protease 1 (MASP-1). C2, a critical component for classic and lectin-based activation, is absent, as expected, due to a lack of epitope-specific antibody production [46] and suggests complete pathway activation. Conversely, in the alternative pathway, C3 spontaneously hydrolyzes to C3b and attracts Complement Factor B (CFB), and no antigen or pathogen-associated molecular pattern is necessary to initiate enzymatic amplification. C3 and CFB were detected in all ZIP-NPs, which could lead to the formation of C3Bb (a C3 convertase) and C3Bb3b (a C5 convertase), and result in membrane attack complex (MAC) proteins C5, C6, C7, C8, and C9 being recruited. However, no stabilizing Complement Factor P was detected, suggesting that any C3Bb complexes were unstable and may be short-lived [48]. The abundance of Complement Factors H and I, both of which are C3 inhibitors, and the C4 inhibitor, C4 Binding Protein Alpha, suggests that corona contain inactivated C3 and C4. The abundance of alternate complement activation proteins, classic complement activation proteins C1 and C4, and complement protein inhibitors are significantly greater in NP than all ZIP-NP and PEG-NP conjugates. In parallel, there is no variation in lectin complement activation pathway abundance and MAC component abundance (see Fig. S4). No statistical differences in complement system abundances between ZIP-NPs and PEG-NP were observed. These results indicate that ZIPs effectively reduce complement protein adsorption to NPs, similar to PEG. However, the alternative pathway may still be activated, as has been observed in other drug delivery systems and biomaterials [15,45,46,49,50].

3.6. ZIP charge motif modulates macrophage uptake of serum-exposed ZIP-NPs

NP protein corona has been correlated strongly with cellular and tissue interactions, especially cellular uptake [10]. Notably, phagocytosis of NPs in blood circulation by the mononuclear phagocytic system (MPS) often reduces circulation time and increases the accumulation of NP in the liver, lungs, and spleen [2]. Adsorbed proteins tend to be strong opsonins for nanoparticles without anti-fouling modifications due to denaturation [2,6]. Therefore, cellular uptake of ZIP charge motifs 1, 3, 7, and 8 functionalized NPs was tested in RAW 264.7 cells, a well-established murine macrophage model. ZIP-NPs were incubated in serum and washed before addition to RAW 264.7 cells in serum-free media to analyze NP uptake (Fig. 5). Uptake was analyzed as a ratio of NP-positive cells to total cells. Unmodified NP had ∼80% cell uptake in this system, while PEG-NP had ∼15% uptake on average. ZIP-NPs varied greatly, with 1. KES having similar uptake to NP, while 3.KEN had similar uptake to PEG-NP (Fig. 5A). Other ZIP-NPs fell in between these extremes, and together ZIP-NPs had an average macrophage uptake of ∼50%. Organized by composition, only KEN had a significant reduction in macrophage uptake versus NP (Fig. 5B) while organization by charge motif indicates charge motifs 3, 7, 8, and 15 were significantly different in uptake versus NP (Fig. 5C).

Fig. 5.

Fig. 5

Open in a new tab

Uptake of ZIP-NP in RAW264.7 macrophage cells is reduced as a function of both composition and charge motif. Cellular uptake of NP is reduced by PEG conjugation and is highly dependent on ZIP sequence, with large variation between sequences (A). Analyzing by composition (B) reveals that KEN has significantly lower cellular uptake than NP, while KES and KDS are trending lower but are not significant. When analyzing by charge motif (C), all charge motifs have reduced cellular uptake vs NP except for CM-1. % indicates p < 0.05 vs NP. 1-way ANOVA with Tukey's posthoc analysis, n = 3–6 biological replicates per ZIP-NP.

Comparing the extremes of ZIP-NP uptake reveals unexpected patterns. 1.KES-NP has the greatest macrophage uptake and 3.KEN-NP the least. 1.KES- and 3.KEN-NPs have similar sizes as well as similar protein adsorption quantities (∼90–100% of NP), although 1.KES-NP is not colloidally stable in PBS while 3.KEN-NP is. Importantly, these two ZIP-NPs have vastly different corona identities. Cluster analysis placed 1.KES and 3.KEN nearly as far apart as possible with the greatest differences found in low-molecular weight proteins (Fig. 3). Similarly, proteomic analysis of the coronas of ZIP-NPs showed 1.KES and 3.KEN having the most different coronas of all ZIP-NPs, with 1.KES having greater albumin and transferrin abundance and 3.KEN having higher APOE abundance and a trend for higher APOA1 abundance. APOE is a known dysopsonin and may play a role in the observed differences [12]. These results suggest that the protein corona identity is responsible for the differences in observed macrophage uptake.

Reductions in phagocytic uptake in vitro have only been weakly correlated with changes in in vivo MPS uptake [51]. However, existing materials are fundamentally more limited in diversity than the zwitterionic peptides developed here. The subtle perturbations possible using ZIPs may enable finer interrogation of corona-biofunctional relationships in the future.

The studies here suggest that rapidly synthesized, compositionally simple, highly tunable peptides and subsequent conjugation to NPs yield NPs with favorable anti-fouling properties. Specifically, we found many designs capable of stabilizing gold NPs in the 30–35 nm range in various physiologically-relevant salt concentrations, a feature that had not previously been reported [29]. Furthermore, ZIP functionalization reduced serum protein adsorption by up to 60% compared to citrate-capped gold NPs and controlled protein corona as a function of peptide sequence similarity, reducing phagocytic uptake by macrophages. The protein corona of serum-exposed ZIP-NPs varied with sequence, and some charge motifs increased the abundance of the dysopsonins APOE and TF compared with citrate-capped NPs, PEG-NP, and other ZIP-NPs. As for opsonins, all ZIP-NPs reduce complement protein abundance compared with citrate-capped NPs and some reduced immunoglobulin abundance. These coronal dysopsonin and opsonin abundance changes have been correlated with improved pharmacokinetics in vivo and altered biodistribution compared tounfunctionalized NPs. Except for ZIP 15.KE, all peptides reported herein are novel and previously unreported.

The combination of highly specific and controlled synthesis of peptides with the observed protein adsorption selectivity makes for a powerful combination to tune nanoparticle behavior in biological contexts. Understanding protein-peptide interactions in the context of NP surface modification is currently limited and typically characterized in the context of targeting. We did not observe a correlation between a simple peptide-peptide interaction energy and protein adsorption behavior in ZIP-NPs. Nevertheless, further investigation of libraries of peptide-functionalized NPs may allow for the correlation of physical and chemical properties to adsorption behavior. Indeed, leveraging library-mediated adsorption data would enable the development of models beyond our naive peptide-protein energy interaction model to predict the adsorption behaviors of peptide-NPs. This library generation and screening strategy was previously used to identify quantitative structure-function relationships in antimicrobial peptides [52]. Understanding and modeling ZIP-NP protein adsorption profiles as a function of sequence and composition would be a tremendous boon to developing nanoparticle drug delivery systems by enabling more efficient development of peptides of interest than existing screening methods.

Combining these methods with rapid peptide synthesis techniques may allow for fast experimental iterative approaches to develop robust physical models. However, such models would depend on identifying the physical and chemical basis of the peptide-protein interactions observed. Due to the tremendous variety of possible interactions and conformations of both peptides and proteins, such a descriptive model may be more difficult to derive and implement than a machine learning model. Machine learning (ML) requires large data sets and functions as a black box, where the interior model parameters do not necessarily correlate with any physical basis. Nevertheless, ML has successfully solved multiple long-standing peptide and protein folding and interaction problems [53,54].

Furthermore, machine learning has been used to develop antibacterial peptides from large databases of existing peptides [55]. However, implementations of prediction to experimental verification for peptide-based materials in other applications still need to be improved. The diverse anti-fouling ZIPs explored in this work represent a new avenue to generate data for such machine-learning efforts. The resulting models may allow for drastically improved control over NP protein coronas through ZIP functionalization.

The gold nanoparticles used here as a model system are suitable for several therapeutic applications, having been used for medical imaging, cancer treatment, and diagnostics [56,57]. Nevertheless, gold NPs rely on surface adsorption for drug loading or their intrinsic physical properties for therapeutic delivery or diagnostics [58]. Therefore, extending this ZIP system to polymeric NPs or liposomes would provide additional translational benefits. Furthermore, liposomal and polymeric NPs have different physical and chemical properties, particularly stimuli-responsive behaviors that are affected by protein coronas [4,59]. They may give additional avenues to explore the relationship between NP drug delivery function and ZIP functionalization.

Amino acids were readily substituted for other, similarly charged hydrophilic amino acids within a peptide sequence while maintaining a similar protein adsorption profile. If this trend holds with individual amino acid substitutions, generating multiple ZIPs with similar protein adsorption profiles and unique antigenic properties may be possible. Single amino acid substitutions have been linked with immunoevasion in viruses [60], and existing work in establishing peptide-antigenic response relationships in nanoparticle systems has shown minimal variations in repeating zwitterionic motifs can create distinct antigens with no cross-reactive antibody responses [61]. Given that ZIPs can have variable composition without disrupting corona identity, there is a potential path toward mitigating immune responses. Additionally, non-natural and d-amino acids, shown to eliminate antibody cross-reactivity when used to substitute for l-amino acids [62], could be employed to alter the structure and further increase diversity for both interaction tuning and altering antigenicity. In sum, the potential of ZIPs to incorporate amino acid substitutions with minimal variation in protein corona could allow for novel immunoevasive substitution strategies without affecting biological function of ZIP-NPs.

4. Conclusions

Controlling the protein corona of nanoparticle drug delivery systems is critical for therapeutic development. After investigating the use of diverse zwitterionic peptides to modify the surface of gold nanoparticles, results indicate a successful reduction in protein adsorption and alteration in the protein corona composition through control of the ZIP charge motif and, to a lesser degree, composition, which has biological consequences, specifically via macrophage uptake of ZIP-functionalized NPs. These results indicate that zwitterionic peptides are suitable for reducing serum protein adsorption to nanoparticle surfaces and suggest that the charge motif of zwitterionic peptides can be tuned to bias for specific serum protein binding semi-independently of amino acid composition. Detailed adsorption profiles of proteins on nanoparticle coronas have been associated with significantly altered circulation time and biodistribution [9,12]. Still, adsorption profiles were observed from either nanoparticle surfaces with limited tuning flexibility or bulk surface modification. The differences in protein adsorption stem from subtle variations of amino acid positioning has not been demonstrated before. This possible tuning of protein corona by ZIP modification offers an alternative to complex purified protein pre-adsorption methods currently being explored. Furthermore, the chemical diversity of ZIPs likely from a single charge motif may ameliorate long-term immunogenicity concerns prevalent amongst NP surface modifications by providing libraries of ZIPs with different potential immunogenicity but similar protein coronas and biological behavior. The combination of anti-fouling, tunability, ease of synthesis, and diversity of zwitterionic peptides bring a revolutionary new approach to the nanoparticle corona problem.

Informed consent

This study did not use human subjects and is not clinical in nature and as such informed consent is not applicable.

Animal experiments

This study did not use animal subjects.

CRediT authorship contribution statement

Clyde Overby: Conceptualization, Methodology, Software, Investigation, Formal analysis, Investigation, Data curation, Writing – original draft, Visualization. Soomin Park: Validation, Investigation. Austin Summers: Investigation, Visualization. Danielle S.W. Benoit: Resources, Supervision, Project administration, Funding acquisition, Conceptualization, Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known conflicts of interest that could have appeared to influence the work in this paper.

Acknowledgements

Funding for this study was provided by the National Institute of Health (NIH) F31AR076874, R01DE018023, R01AR056696, and the National Science Foundation (NSF) CBET-1450897 and DMR-2103553. The authors thank Professor James L. McGrath (University of Rochester, Biomedical Engineering) for using his laboratory equipment. The authors also thank Kevin Welle and Kyle Swovick (University of Rochester, Mass Spectroscopy Resource Laboratory) for their assistance and Prof. James E. Hutchison (University of Oregon) for valuable discussions. The content is solely the responsibility of the authors and does not necessarily represent the official view of the National Institutes of Health or National Science Foundation.

Footnotes

Peer review under responsibility of KeAi Communications Co., Ltd.

Appendix A

Supplementary data to this article can be found online at https://doi.org/10.1016/j.bioactmat.2023.03.020.

Appendix A. Supplementary data

The following are the Supplementary data to this article.

Multimedia component 1
mmc1.xlsx (29KB, xlsx)
Multimedia component 2
mmc2.docx (253.4KB, docx)

References

  • 1.Mitchell M.J., Billingsley M.M., Haley R.M., Wechsler M.E., Peppas N.A., Langer R. Engineering precision nanoparticles for drug delivery. Nat. Rev. Drug Discov. 2021;20:101–124. doi: 10.1038/s41573-020-0090-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Ngo W., Ahmed S., Blackadar C., Bussin B., Ji Q., Mladjenovic S.M., Sepahi Z., Chan W.C. Why nanoparticles prefer liver macrophage cell uptake in vivo. Adv. Drug Deliv. Rev. 2022 doi: 10.1016/j.addr.2022.114238. [DOI] [PubMed] [Google Scholar]
  • 3.Wilhelm S., Tavares A.J., Dai Q., Ohta S., Audet J., Dvorak H.F., Chan W.C.W. Analysis of nanoparticle delivery to tumours. Nat. Rev. Mater. 2016;1 [Google Scholar]
  • 4.Malcolm D.W., Varghese J.J., Sorrells J.E., Ovitt C.E., Benoit D.S. The effects of biological fluids on colloidal stability and siRNA delivery of a pH-responsive micellar nanoparticle delivery system. ACS Nano. 2017;12:187–197. doi: 10.1021/acsnano.7b05528. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Corbo C., Molinaro R., Parodi A., Toledano Furman N.E., Salvatore F., Tasciotti E. The impact of nanoparticle protein corona on cytotoxicity, immunotoxicity and target drug delivery. Nanomedicine. 2016;11:81–100. doi: 10.2217/nnm.15.188. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Pearson R.M., Juettner V.V., Hong S. Biomolecular corona on nanoparticles: a survey of recent literature and its implications in targeted drug delivery. Front. Chem. 2014;2:108. doi: 10.3389/fchem.2014.00108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Salvati A., Pitek A.S., Monopoli M.P., Prapainop K., Bombelli F.B., Hristov D.R., Kelly P.M., Åberg C., Mahon E., Dawson K.A. Transferrin-functionalized nanoparticles lose their targeting capabilities when a biomolecule corona adsorbs on the surface. Nat. Nanotechnol. 2013;8:137–143. doi: 10.1038/nnano.2012.237. [DOI] [PubMed] [Google Scholar]
  • 8.Ke P.C., Lin S., Parak W.J., Davis T.P., Caruso F. A decade of the protein corona. ACS Nano. 2017;11:11773–11776. doi: 10.1021/acsnano.7b08008. [DOI] [PubMed] [Google Scholar]
  • 9.Bertrand N., Grenier P., Mahmoudi M., Lima E.M., Appel E.A., Dormont F., Lim J.-M., Karnik R., Langer R., Farokhzad O.C. Mechanistic understanding of in vivo protein corona formation on polymeric nanoparticles and impact on pharmacokinetics. Nat. Commun. 2017;8:777. doi: 10.1038/s41467-017-00600-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Kopac T. Protein corona, understanding the nanoparticle–protein interactions and future perspectives: a critical review. Int. J. Biol. Macromol. 2021;169:290–301. doi: 10.1016/j.ijbiomac.2020.12.108. [DOI] [PubMed] [Google Scholar]
  • 11.Blanco E., Shen H., Ferrari M. Principles of nanoparticle design for overcoming biological barriers to drug delivery. Nat. Biotechnol. 2015;33:941–951. doi: 10.1038/nbt.3330. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Ritz S., Schöttler S., Kotman N., Baier G., Musyanovych A., Kuharev J., Landfester K., Schild H., Jahn O., Tenzer S., et al. Protein corona of nanoparticles: distinct proteins regulate the cellular uptake. Biomacromolecules. 2015;16:1311–1321. doi: 10.1021/acs.biomac.5b00108. [DOI] [PubMed] [Google Scholar]
  • 13.Papini E., Tavano R., Mancin F. Opsonins and dysopsonins of nanoparticles: facts, concepts, and methodological guidelines. Front. Immunol. 2020;11 doi: 10.3389/fimmu.2020.567365. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.García K.P., Zarschler K., Barbaro L., Barreto J.A., O'Malley W., Spiccia L., Stephan H., Graham B. Zwitterionic‐coated “stealth” nanoparticles for biomedical applications: recent advances in countering biomolecular corona formation and uptake by the mononuclear phagocyte system. Small. 2014;10:2516–2529. doi: 10.1002/smll.201303540. [DOI] [PubMed] [Google Scholar]
  • 15.Fam S.Y., Chee C.F., Yong C.Y., Ho K.L., Mariatulqabtiah A.R., Tan W.S. Stealth coating of nanoparticles in drug-delivery systems. Nanomaterials. 2020;10:787. doi: 10.3390/nano10040787. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Li Q., Wen C., Yang J., Zhou X., Zhu Y., Zheng J., Cheng G., Bai J., Xu T., Ji J., et al. Zwitterionic biomaterials. Chem. Rev. 2022;122:17073–17154. doi: 10.1021/acs.chemrev.2c00344. [DOI] [PubMed] [Google Scholar]
  • 17.Jackson M.A., Werfel T.A., Curvino E.J., Yu F., Kavanaugh T.E., Sarett S.M., Dockery M.D., Kilchrist K.V., Jackson A.N., Giorgio T.D. Zwitterionic nanocarrier surface chemistry improves siRNA tumor delivery and silencing activity relative to polyethylene glycol. ACS Nano. 2017;11:5680–5696. doi: 10.1021/acsnano.7b01110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Panagi Z., Beletsi A., Evangelatos G., Livaniou E., Ithakissios D. Vol. 221. 2001. pp. 143–152. (Avgoustakis KJIjop: Effect of Dose on the Biodistribution and Pharmacokinetics of PLGA and PLGA–mPEG Nanoparticles). [DOI] [PubMed] [Google Scholar]
  • 19.Perry J.L., Reuter K.G., Kai M.P., Herlihy K.P., Jones S.W., Luft J.C., Napier M., Bear J.E., DeSimone J.M. PEGylated PRINT nanoparticles: the impact of PEG density on protein binding, macrophage association, biodistribution, and pharmacokinetics. Nano Lett. 2012;12:5304–5310. doi: 10.1021/nl302638g. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Hatakeyama H., Akita H., Harashima H. The polyethyleneglycol dilemma: advantage and disadvantage of PEGylation of liposomes for systemic genes and nucleic acids delivery to tumors. Biol. Pharm. Bull. 2013;36:892–899. doi: 10.1248/bpb.b13-00059. [DOI] [PubMed] [Google Scholar]
  • 21.Schöttler S., Becker G., Winzen S., Steinbach T., Mohr K., Landfester K., Mailänder V., Wurm F.R. Protein adsorption is required for stealth effect of poly (ethylene glycol)-and poly (phosphoester)-coated nanocarriers. Nat. Nanotechnol. 2016;11:372–377. doi: 10.1038/nnano.2015.330. [DOI] [PubMed] [Google Scholar]
  • 22.Zhang P., Sun F., Liu S., Jiang S. Anti-PEG antibodies in the clinic: current issues and beyond PEGylation. J. Contr. Release. 2016;244:184–193. doi: 10.1016/j.jconrel.2016.06.040. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Mohamed M., Abu Lila A.S., Shimizu T., Alaaeldin E., Hussein A., Sarhan H.A., Szebeni J., Ishida T. PEGylated liposomes: immunological responses. Sci. Technol. Adv. Mater. 2019;20:710–724. doi: 10.1080/14686996.2019.1627174. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Verhoef J.J.F., Anchordoquy T.J. Questioning the use of PEGylation for drug delivery. Drug Deliv. Transl. Res. 2013;3:499–503. doi: 10.1007/s13346-013-0176-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Ilinskaya A.N., Dobrovolskaia M.A. Understanding the immunogenicity and antigenicity of nanomaterials: past, present and future. Toxicol. Appl. Pharmacol. 2016;299:70–77. doi: 10.1016/j.taap.2016.01.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Zhao J., Qin Z., Wu J., Li L., Jin Q., Ji J. Zwitterionic stealth peptide-protected gold nanoparticles enable long circulation without the accelerated blood clearance phenomenon. Biomater. Sci. 2018;6:200–206. doi: 10.1039/c7bm00747g. [DOI] [PubMed] [Google Scholar]
  • 27.Yuan Z., Li B., Niu L., Tang C., McMullen P., Jain P., He Y., Jiang S. Zwitterionic peptide cloak mimics protein surfaces for protein protection. Angew. Chem. Int. Ed. 2020;59:22378–22381. doi: 10.1002/anie.202004995. [DOI] [PubMed] [Google Scholar]
  • 28.Walkey C.D., Olsen J.B., Guo H., Emili A., Chan W.C. Nanoparticle size and surface chemistry determine serum protein adsorption and macrophage uptake. J. Am. Chem. Soc. 2012;134:2139–2147. doi: 10.1021/ja2084338. [DOI] [PubMed] [Google Scholar]
  • 29.Egorova E.A., van Rijt M.M.J., Sommerdijk N., Gooris G.S., Bouwstra J.A., Boyle A.L., Kros A. One peptide for them all: gold nanoparticles of different sizes are stabilized by a common peptide amphiphile. ACS Nano. 2020;14:5874–5886. doi: 10.1021/acsnano.0c01021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Trovato A., Chiti F., Maritan A., Seno F. Insight into the structure of amyloid fibrils from the analysis of globular proteins. PLoS Comput. Biol. 2006;2 doi: 10.1371/journal.pcbi.0020170. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Barbero F., Russo L., Vitali M., Piella J., Salvo I., Borrajo M.L., Busquets-Fité M., Grandori R., Bastús N.G., Casals E., et al. formation of the protein corona: the interface between nanoparticles and the immune system. Semin. Immunol. 2017;34:52–60. doi: 10.1016/j.smim.2017.10.001. [DOI] [PubMed] [Google Scholar]
  • 32.Roberts C.J. Therapeutic protein aggregation: mechanisms, design, and control. Trends Biotechnol. 2014;32:372–380. doi: 10.1016/j.tibtech.2014.05.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Nowinski A.K., White A.D., Keefe A.J., Jiang S. Biologically inspired stealth peptide-capped gold nanoparticles. Langmuir. 2014;30:1864–1870. doi: 10.1021/la404980g. [DOI] [PubMed] [Google Scholar]
  • 34.Luby A.O., Breitner E.K., Comfort K.K. Preliminary protein corona formation stabilizes gold nanoparticles and improves deposition efficiency. Appl. Nanosci. 2016;6:827–836. [Google Scholar]
  • 35.Yang Q., Lai S.K. Vol. 7. Wiley Interdiscip Rev Nanomed Nanobiotechnol; 2015. pp. 655–677. (Anti-PEG Immunity: Emergence, Characteristics, and Unaddressed Questions). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Cao Z., Jiang S. Super-hydrophilic zwitterionic poly(carboxybetaine) and amphiphilic non-ionic poly(ethylene glycol) for stealth nanoparticles. Nano Today. 2012;7:404–413. [Google Scholar]
  • 37.Kane R.S., Deschatelets P., Whitesides G.M. Kosmotropes form the basis of protein-resistant surfaces. Langmuir. 2003;19:2388–2391. [Google Scholar]
  • 38.Zhou H.-X., Pang X. Electrostatic interactions in protein structure, folding, binding, and condensation. Chem. Rev. 2018;118:1691–1741. doi: 10.1021/acs.chemrev.7b00305. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Ciemny M., Kurcinski M., Kamel K., Kolinski A., Alam N., Schueler-Furman O., Kmiecik S. Protein–peptide docking: opportunities and challenges. Drug Discov. Today. 2018;23:1530–1537. doi: 10.1016/j.drudis.2018.05.006. [DOI] [PubMed] [Google Scholar]
  • 40.Cedervall T., Lynch I., Lindman S., Berggard T., Thulin E., Nilsson H., Dawson K.A., Linse S. Understanding the nanoparticle-protein corona using methods to quantify exchange rates and affinities of proteins for nanoparticles. Proc. Natl. Acad. Sci. U.S.A. 2007;104:2050–2055. doi: 10.1073/pnas.0608582104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Milligan G.W., Cooper M.C. An examination of procedures for determining the number of clusters in a data set. Psychometrika. 1985;50:159–179. [Google Scholar]
  • 42.Fedeli C., Segat D., Tavano R., Bubacco L., De Franceschi G., de Laureto P.P., Lubian E., Selvestrel F., Mancin F., Papini E. The functional dissection of the plasma corona of SiO₂-NPs spots histidine rich glycoprotein as a major player able to hamper nanoparticle capture by macrophages. Nanoscale. 2015;7:17710–17728. doi: 10.1039/c5nr05290d. [DOI] [PubMed] [Google Scholar]
  • 43.Zhang Y., Wu J.L.Y., Lazarovits J., Chan W.C.W. An analysis of the binding function and structural organization of the protein corona. J. Am. Chem. Soc. 2020;142:8827–8836. doi: 10.1021/jacs.0c01853. [DOI] [PubMed] [Google Scholar]
  • 44.Pirmoradian M., Aarsland D., Zubarev R.A. Isoelectric point region pI≈7.4 as a treasure island of abnormal proteoforms in blood. Discoveries. 2016;4 doi: 10.15190/d.2016.14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Chen F., Wang G., Griffin J.I., Brenneman B., Banda N.K., Holers V.M., Backos D.S., Wu L., Moghimi S.M., Simberg D. Complement proteins bind to nanoparticle protein corona and undergo dynamic exchange in vivo. Nat. Nanotechnol. 2017;12:387. doi: 10.1038/nnano.2016.269. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Quach Q.H., Kah J.C.Y. Non-specific adsorption of complement proteins affects complement activation pathways of gold nanomaterials. Nanotoxicology. 2017;11:382–394. doi: 10.1080/17435390.2017.1306131. [DOI] [PubMed] [Google Scholar]
  • 47.Lu J., Kishore U. C1 complex: an adaptable proteolytic module for complement and non-complement functions. Front. Immunol. 2017;8 doi: 10.3389/fimmu.2017.00592. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Merle N.S., Church S.E., Fremeaux-Bacchi V., Roumenina L.T. Complement system Part I – molecular mechanisms of activation and regulation. Front. Immunol. 2015;6 doi: 10.3389/fimmu.2015.00262. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Mödinger Y., Teixeira G.Q., Neidlinger-Wilke C., Ignatius A. Role of the complement system in the response to orthopedic biomaterials. Int. J. Mol. Sci. 2018;19:3367. doi: 10.3390/ijms19113367. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Nilsson B., Ekdahl K.N., Mollnes T.E., Lambris J.D. The role of complement in biomaterial-induced inflammation. Mol. Immunol. 2007;44:82–94. doi: 10.1016/j.molimm.2006.06.020. [DOI] [PubMed] [Google Scholar]
  • 51.Gustafson H.H., Holt-Casper D., Grainger D.W., Ghandehari H. Nanoparticle uptake: the phagocyte problem. Nano Today. 2015;10:487–510. doi: 10.1016/j.nantod.2015.06.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Haney E.F., Brito-Sánchez Y., Trimble M.J., Mansour S.C., Cherkasov A., Hancock R.E.W. Computer-aided discovery of peptides that specifically attack bacterial biofilms. Sci. Rep. 2018;8:1871. doi: 10.1038/s41598-018-19669-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Jumper J., Evans R., Pritzel A., Green T., Figurnov M., Ronneberger O., Tunyasuvunakool K., Bates R., Žídek A., Potapenko A. Highly accurate protein structure prediction with AlphaFold. Nature. 2021;596:583–589. doi: 10.1038/s41586-021-03819-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Tsaban T., Varga J.K., Avraham O., Ben-Aharon Z., Khramushin A., Schueler-Furman O. Harnessing protein folding neural networks for peptide–protein docking. Nat. Commun. 2022;13:176. doi: 10.1038/s41467-021-27838-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Cherkasov A., Hilpert K., Jenssen H., Fjell C.D., Waldbrook M., Mullaly S.C., Volkmer R., Hancock R.E.W. Use of artificial intelligence in the design of small peptide antibiotics effective against a broad spectrum of highly antibiotic-resistant superbugs. ACS Chem. Biol. 2009;4:65–74. doi: 10.1021/cb800240j. [DOI] [PubMed] [Google Scholar]
  • 56.Elahi N., Kamali M., Baghersad M.H. Recent biomedical applications of gold nanoparticles: a review. Talanta. 2018;184:537–556. doi: 10.1016/j.talanta.2018.02.088. [DOI] [PubMed] [Google Scholar]
  • 57.Zong J., Cobb S.L., Cameron N.R. Peptide-functionalized gold nanoparticles: versatile biomaterials for diagnostic and therapeutic applications. Biomater. Sci. 2017;5:872–886. doi: 10.1039/c7bm00006e. [DOI] [PubMed] [Google Scholar]
  • 58.Ackun-Farmmer M.A., Overby C.T., Haws B.E., Choe R., Benoit D.S. Biomaterials for orthopedic diagnostics and theranostics. Curr. Opin. Biomet. Eng. 2021;19 doi: 10.1016/j.cobme.2021.100308. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 59.Sharifi S., Caracciolo G., Mahmoudi M. Biomolecular corona affects controlled release of drug payloads from nanocarriers. Trends Pharmacol. Sci. 2020;41:641–652. doi: 10.1016/j.tips.2020.06.011. [DOI] [PubMed] [Google Scholar]
  • 60.Rees-Spear C., Muir L., Griffith S.A., Heaney J., Aldon Y., Snitselaar J.L., Thomas P., Graham C., Seow J., Lee N. The effect of spike mutations on SARS-CoV-2 neutralization. Cell Rep. 2021:34. doi: 10.1016/j.celrep.2021.108890. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 61.McMullen P., Qiao Q., Luozhong S., Cai L., Fang L., Shao Q., Jiang S. Motif-based zwitterionic peptides impact their structure and immunogenicity. Chem. Sci. 2022;13:10961–10970. doi: 10.1039/d2sc03519g. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 62.Benkirane N., Friede M., Guichard G., Briand J.-P., Van Regenmortel M., Muller S. Antigenicity and immunogenicity of modified synthetic peptides containing D-amino acid residues. Antibodies to a D-enantiomer do recognize the parent L-hexapeptide and reciprocally. J. Biol. Chem. 1993;268:26279–26285. [PubMed] [Google Scholar]
  • 63.Walkey C.D., Olsen J.B., Guo H., Emili A., Chan W.C. Nanoparticle size and surface chemistry determine serum protein adsorption and macrophage uptake. J. Am. Chem. Soc. 2012;134:2139–2147. doi: 10.1021/ja2084338. [DOI] [PubMed] [Google Scholar]
  • 64.Demichev V., Messner C.B., Vernardis S.I., Lilley K.S., Ralser M. DIA-NN: neural networks and interference correction enable deep proteome coverage in high throughput. Nat. Methods. 2020;17:41–44. doi: 10.1038/s41592-019-0638-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 65.Cox J., Hein M.Y., Luber C.A., Paron I., Nagaraj N., Mann M. Accurate proteome-wide label-free quantification by delayed normalization and maximal peptide ratio extraction, termed MaxLFQ*. Mol. Cell. Proteomics. 2014;13:2513–2526. doi: 10.1074/mcp.M113.031591. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Multimedia component 1
mmc1.xlsx (29KB, xlsx)
Multimedia component 2
mmc2.docx (253.4KB, docx)

Articles from Bioactive Materials are provided here courtesy of KeAi Publishing

RESOURCES