Efficient Enrichment of Glycopeptides by Supramolecular Nanoassemblies that Use Proximity-Assisted Covalent Binding

Abstract

Mass spectrometry (MS)-based analysis of glycoproteins and glycopeptides requires selective separation strategies to eliminate interferences from more abundant non-glycosylated biomolecules. In this work, we describe a two-phase liquid-liquid extraction method using supramolecular polymeric nanoassemblies that can selectively and efficiently enrich glycopeptides for enhanced MS detection. The polymeric nanoassemblies are made selective for glycopeptides via the incorporation of hydrazide functional groups that covalently bind to glycans. The enrichment efficiency is further enhanced via the incorporation of acidic functional groups that provide a proximity-assisted catalysis of the hydrazide-glycan conjugation reaction. Our results further demonstrate the value of designer supramolecular nanomaterials for the selective enrichment of modified peptides from complicated mixtures.

Graphical Abstract

Introduction

Protein glycosylation plays an important role in biological process, such as cell recognition,¹ signalling pathways² and development of genetic disorders.³ Also, it has recently gained attention in the pharmaceutical industry as part of monoclonal antibody characterization.^4–6 The methods used for glycosylation analysis are most often based on mass spectrometry (MS), but because of the low abundance and heterogeneity of glycosylated sites, separation/purification prior to MS analysis is typically required.^7–9

Efforts have been made to improve the separation of glycosylated proteins or peptides through noncovalent and covalent interactions.^8,10 Noncovalent interactions include hydrophilic, lectin and chelation interactions that are relatively easy to incorporate into online analysis before MS detection.^10–13 Covalent binding using boronic acid or hydrazine/ hydrazide functional groups are also widely used because of their higher specificity, which is essential for complex mixtures.^14–21 For better selectivity and efficiency in extraction, covalent and noncovalent interactions have also been used complementarily. For example, Wang et al. used boronic acid nanoparticles for specific binding with glycopeptides and poly(methyl methacrylate) nanobeads for non-specific interactions with non-glycopeptides to achieve better selectivity.²² Also, glycoprotein enrichment based on lectin affinity chromatography and on hydrazide chemistry have been used separately by Song, et al. to determine and confirm glycoprotein/glycopeptide biomarkers in human blood serum more efficiently.²³ Based on the continued need for both high selectivity and efficiency in glycopeptide enrichment approaches, we designed a new strategy that is inspired by the use of covalent and non-covalent interactions. Instead of using both covalent and non-covalent interactions to target peptides, we used covalent binding via hydrazide groups to achieve high selectivity for glycopeptide and a secondary non-covalent interaction to catalyze this normally slow covalent binding reaction, thereby improving extraction efficiency. We employ this strategy in a two-phase liquid-liquid extraction platform because of its advantageous ability to extract, purify and concentrate in one step.^24,25 To achieve this two-phase extraction, we have designed a supramolecular nanoassembly that can be readily tuned for optimal extraction efficiency. The supramolecular nanoassembly described here is based on an amphiphilic polymer design that we recently developed to selectively enrich phosphopeptides and other peptides according to their isoelectric point.^26–31 We find that in the confined environment of the nanoassembly the reaction of hydrazide with glycans is faster than in existing approaches, which in most cases typically take up to 24 h to complete.^{17,23,32–34} Moreover, the extraction efficiency can be further increased by placing functional groups in the assemblies that catalyze the reaction of hydrazide moieties with glycans.

Experimental

Materials and Reagents

Horseradish peroxidase (HRP), bovine serum albumin (BSA), DL-dithiothreitol (DTT), iodoacetamide (IAM), trifluoracetic acid (TFA) and potassium acetate were obtained from Sigma-Aldrich. The monoclonal antibody (IgG1) was purchased from Waters. PNGase F was obtained from New England Biolabs. Trypsin was purchased from Promega. Sodium periodate, 2, 5-dihydroxybenzoic acid (DHB), MOPS, acetic acid, toluene, tetrahydrofuran (THF) and acetonitrile (ACN) were purchased from Fisher Scientific. Urea was purchased from MP Biomedicals. Ammonium bicarbonate (NH₄HCO₃) was obtained from Fluka. THF was distilled before use. Water was purified using a Milli-Q water purification system. All chemicals were used as received from commercial sources.

Polymer synthesis, self-assembly and co-assembly in toluene

The synthesis and characterization of polymer P1 (Scheme 1) are described in the Supporting Information. Polymers P2-P6 were synthesized as previously described.^26,29,30 Nanoassembly solutions of individual polymers were prepared by dissolving solid polymers in toluene. Nanoassemblies of co-assembled polymers were prepared by mixing their already prepared nanoassembly solutions. The concentration of P1 was 0.3 mg/mL and of P2-P6 were 0.6 mg/mL. Dynamic light scattering on the resulting polymer solutions after self-assembly and co-assembly were conducted and can be found in Supporting Information (Figure S2).

Scheme 1.

Polymers P1-P6 that were used in this work.

Protein oxidation and digest

To facilitate enrichment of the glycosylated peptides by the hydrazide-containing nanoassemblies, samples of 0.5 mg of IgG1 or 1 mg of HRP were mildly oxidized with 10 mM sodium periodate in 1 mL of 0.1 M sodium acetate buffer at pH 5.5 to produce free aldehydes capable of reacting with hydrazide. Samples were incubated in the dark at room temperature for 30 min and then were washed 3 times and desalted by a 10 kDa molecular weight cut-off filter to remove excess periodate before digestion. To prepare the samples for proteolytic digestion, they were dissolved in 500 μL denaturing buffer containing 8 M urea, 50 mM NH₄HCO₃ and 5 mM DTT and incubated for 1 h at 37 °C with gentle agitation to reduce the disulfide bonds in the proteins. The protein solutions were brought to room temperature, and 10 mM IAM was added into the solutions before incubating in the dark at room temperature for 30 min to alkylate the reduced disulfide bonds. 5 mM DTT was added again, and the samples were incubated in the dark at room temperature for another 30 min to stop overalkylation. The solutions then were diluted with 50 mM NH₄HCO₃ to reduce the urea concentration to 1.2 M. Trypsin was added at an enzyme-to-protein ratio of 1:50 and incubated for 12 h at 37 °C. BSA was digested in the same manner as the other proteins.

Liquid-liquid extraction and release

Before extraction with the nanoassemblies, the protein digests were diluted to the desired pH and concentration. 200 μL of the nanoassembly solution was added to 500 μL of the protein digest and vortex mixed vigorously for 30 min. The sample was incubated in a sand bath at 50 °C for 1 h, vortexed again for 30 min, and then incubated again at 50 °C for 1 h. After extraction, centrifugation at 14000 rpm for 30 min was used to separate the two phases. The aqueous phase was removed, and 500 μL of fresh acetate buffer was added to wash the organic phase by mixing the two phases for 30 min and removing the aqueous phase after centrifugation. The organic phase was dried by vacuum centrifuge. 300 μL of THF was added to break the supramolecular assembly and 200 μL of MOPS buffer (pH 7) was added to wash the polymer thoroughly. Then, the THF was evaporated and aqueous phase was removed. To release the glycopeptides from the polymer, the dry residue was re-dissolved in 300 μL of fresh THF and 200 μL of 0.1 M HCl was added to hydrolyze the hydrazone group between the polymer and the glycopeptides. The mixture was incubated for 1 h at 60 °C to complete peptide release. The aqueous phase with released glycopeptides was collected after centrifugation. Before adding PNGase F, vacuum centrifugation was used to remove the remaining THF, and the pH was adjusted to 8.4 in 50 mM NH₄HCO₃. 1 μL of PNGase F was then added to 20 μL of the peptide solution and incubated at 37 °C with gentle agitation for 12 h.

Mass spectrometry analysis

The peptide solutions resulting from nanoassembly enrichment before and after deglycosylation were analysed by MALDI-MS. 0.5 μL of peptide solution was mixed with 0.5 μL of DHB matrix solution (25 mg/mL in 70% ACN, 29% H2O and 1% TFA) and spotted on MALDI target for analysis. MALDI analysis was performed on a Bruker UltrafleXtreme MALDI-TOF/TOF. Spectra were obtained in positive ionization mode using reflectron detection with a repetition rate of 2 kHz and an acceleration voltage of 20 kV. Peptides were identified using MS/MS using the LIFT mode³⁵ at a 1 kHz laser repetition rate, applying 7.5 kV for initial acceleration of ions and 19 kV for reacceleration of the product ions in the LIFT cell. 5000 laser shots were accumulated per spectrum.

Results and Discussion

Glycopeptide enrichment using hydrazide-containing nanoassemblies

The immiscible two-phase liquid-liquid extraction procedure for enriching glycopeptides from an aqueous phase by the hydrazide-containing nanoassemblies in toluene is illustrated in Scheme 2. IgG1, with its four well characterized glycans on Asn292,³⁶ was used as a model protein to evaluate the extraction performance of nanoassemblies of polymer P1. Before enrichment, the two most abundant glycopeptides from the IgG1 digest are detected (Figure 1a), while the other two glycopeptides are not detected. After mild oxidation, enrichment with the nanoassemblies and release, all four glycopeptides are exclusively detected (Figure 1b). Some heterogeneity in the detected glycopeptides is observed due to the oxidation reaction, but most of the glycan information is retained after enrichment (Figure S3). The extraction selectivity is further confirmed by PNGase F treatment of the enriched/released glycopeptides. Only a single peptide containing the glycosylated residue Asn292 is measured after deglycosylation (Figure 1c), confirming that only glycopeptides are extracted.

Scheme 2.

Workflow for glycopeptide enrichment using supramolecular nanoassemblies and analysis by mass spectrometry.

Figure1.

MALDI-TOF mass spectra of the oxidized IgG1 tryptic digests before and after enrichment. (a) Mass spectrum before enrichment; (b) mass spectrum after enrichment using nanoassemblies of the hydrazide polymer P1 at pH 4; and (c) mass spectrum after enrichment and deglycosylation by PNGase F.

Improved enrichment efficiency via proximity-assisted reactivity

For more heterogeneous glycoproteins, enrichment using polymer P1 can sometimes lead to relatively inefficient detection of some low level glycopeptides. As an example, enrichment of HRP using nanoassemblies of P1 can enable the selective detection of all 7 known glycosylation sites (Figure 2a), even though most are undetectable before enrichment (Figure S4), but some of the glycopeptides have low relative abundance. To improve the enrichment efficiency, we hypothesized that a copolymer with both hydrazide and weak acid functional groups would facilitate the binding in the nanoassemblies as the hydrazide-aldehyde conjugation reaction is known to be catalysed by an acid and conjugate base.³⁷ We reasoned that a nearby acid and conjugate base in the local microenvironment would accelerate the reaction via a proximity-based effect. We tested this idea using mixed nanoassemblies with 10 nm diameters (Figure S2) of polymer P1 and P2 (Scheme 1). A polymer containing phosphonate groups (i.e. polymer P2) was chosen because it is a polyprotic acid with pK_a values of 2.5 and 7.5 that could provide both acidic and conjugate base functionality at the enrichment pH of 5.0. Upon extraction of the HRP digest by the mixed nanoassembly, we find a significant increase in the total ion intensity of the glycosylated peptides (Figure 2b), and even detect new glycopeptides that are due to enzymatic mis-cleavages that were not detected upon enrichment with nanoassemblies of P1 (Figure 2a).

Figure 2.

Enhanced glycopeptide detection after proximity-assisted enrichment. (a) MALDI-TOF mass spectrum of an oxidized HRP tryptic digest after enrichment using nanoassemblies of the hydrazide polymer P1; the labels above the peaks indicate the Asn residue containing the glycan; (b) MALDI-TOF mass spectrum after enrichment using nanoassembly mixtures of hydrazide polymer P1 and phosphonate polymer P2; (c) total intensity of glycopeptides extracted by nanoassemblies of P1 and of P1+P2, P1+P3 and P1+P4; and (d) proposed proximity effects that improve the extraction efficiency of glycopeptides.

Two sets of control experiments were employed to further test our hypothesis that a nearby acid and conjugate base improve enrichment efficiency. Mixed nanoassemblies of P1 and (i) a quaternary amine containing polymer, P3, that does not yield a proton or conjugate base or (ii) an ethyleneglycol-containing polymer, P4, without acidic or basic functionality were also studied (Scheme 1). Using total glycopeptide intensity as an indicator of the enrichment efficiency, we find that mixed assemblies with the phosphonate polymer P2 increase glycopeptide signal by more than a factor of 2 compared to P1 alone, while mixed assemblies with polymers P3 and P4 decrease glycopeptide signal by factors of 3 and 2.3, respectively (Figure 2c). The increased glycopeptide signal caused by the presence of P2 in the nanoassemblies together with the decreased signal in the presence of the other polymers supports our hypothesis that nearby proton donors and acceptors are responsible for the improved enrichment efficiency. We postulate that the improved efficiency is due to proximity effects that accelerate the hydrazide-glycan conjugation reaction (Figure 2d).

To further understand how acidic functional groups in the polymer nanoassemblies influence the enrichment efficiency, we synthesized polymers with carboxylate (P5) and sulfonate (P6) groups (Scheme 1) and tested their enrichment efficiencies as a function of pH. These functional groups, along with the phosphonate functional groups, provide pK_a values that range from about −1 (−SO₃H) to 7.5 (−PO₃H⁻), providing a test of the importance of both the acid and conjugate base functionality in enhancing the enrichment reaction. We find that the enrichment efficiency is improved by forming mixed assemblies with all the acidic polymers when the extraction is done at low pH (Figure 3). At higher pH values, however, the enrichment efficiencies become more similar to the hydrazide assembly alone, suggesting that deprotonation of the acidic group leads to less enhancement. Interestingly, the drop in the enhancement as the pH is increased is more extensive for the sulfonate polymer (P6), which has the lowest pK_a and is the least extensive for the phosphonate polymer (P2), which has the highest pK_a. Indeed, the phosphonate group has two acidic protons causing its enhancement to exist over a wider range of pH values. As a whole, these results provide additional evidence for a proximity-based effect in the nanoassemblies that improve glycopeptide enrichment. These improvements, together with previous work from our group using similar materials, highlights the advantages of using supramolecular nanomaterials for peptide enrichment in general.

Figure 3.

Total intensity of glycopeptides extracted from tryptic digests of oxidized HRP at pH values ranging from 3 to 7 using nanoassemblies of P1, P1+P2, P1+P6 and P1+P5.

Extraction of glycopeptides from complicated samples using co-assembled nanoassemblies

With a better understanding of the polymer features that influence enrichment efficiency, we then tested the ability of mixed nanoassemblies to extract trace-level glycopeptides in a mixture. For this purpose, IgG1 and HRP were digested in the presence of 100-fold molar excesses of BSA. Before enrichment of the protein digest mixture, no glycopeptides for IgG1 or HRP can be detected (Figure 4a and d); only non-glycosylated peptides are detected. After using nanoassemblies of P1 (Figure 4b) or mixed nanoassemblies of P1 and P6 (Figure 4c), the glycopeptide from IgG1 is selectively detected after release and deglycosylation, with higher signal being observed for the mixed nanoassemblies. A similar effect is observed for the glycopeptides from HRP. The mixed nanoassemblies containing P1 and P6 allow the selective detection of more glycopeptides (Figure 4f) than the nanoassembly with hydrazide polymer (P1) alone (Figure 4e). These results highlight the high degree of selectivity possible with these supramolecular nanoassemblies.

Figure 4.

Enrichment and sensitive detection of trace-level glycopeptides in protein digest mixtures. (a) MALDI-TOF mass spectra of a tryptic digest of IgG1 (10 nM) and BSA (1 μM) before enrichment; (b) after enrichment using nanoassemblies of polymer P1 at pH 4 followed by release and deglycosylation; and (c) after extraction using mixed nanoassemblies of polymers P1 and P6 at pH 4 followed by release and deglycosylation. (d) MALDI mass spectra of a tryptic digest mixture of HRP (50 nM) and BSA (5 μM) before enrichment; (e) after enrichment using nanoassemblies of polymer P1 at pH 4 followed by release and deglycosylation; and (f) after extraction using mixed nanoassemblies of polymers P1 and P6 at pH 4 followed by release and deglycosylation.

Conclusions

We have developed a simple method to improve glycopeptide enrichment efficiency and MS detection using supramolecular nanoassemblies based on amphiphilic polymers with hydrazide functional groups. The hydrazide-containing polymer itself forms nanoassemblies that are highly selective for oxidized glycopeptides. Co-assembly of the hydrazide polymer together with acidic polymers further improves enrichment efficiency via a proximity effect that catalyses the hydrazide-glycan conjugation reaction, allowing us to efficiently extract glycopeptides in about 3 h. Using the nanoassembly mixtures allow the selective and efficient enrichment and detection of glycopeptides that are present at low levels in mixtures. This study further demonstrates the value of designer supramolecular materials based on amphiphilic polymers as a platform for enriching and detecting biomolecules of interest. Future work will apply these nanoassemblies for the detection of glycosylated peptides or proteins in cell lysate, which would make them valuable materials for glycoproteomics studies.