Sample Preparation Matters for Peptide Mapping to Evaluate Deamidation of Adeno-Associated Virus Capsid Proteins Using Liquid Chromatography

Abstract

Adeno-associated viral capsid proteins (AAV VP) are the major components that determine the tissue specificity and immunogenicity, and in vivo transduction performance of the vector. It was reported that asparagine deamidation of AAV capsid proteins leads to charge variants/heterogeneity and altered vector function, reduction of stability, and potency of AAV gene therapy products. Deamidation of asparagine residue is a common post-translational modification of proteins and is mostly detected and quantified by liquid chromatography–tandem mass spectrometry (LC-MS/MS)-based peptide mapping. However, deamidation can be spontaneously introduced during sample preparation before LC-MS/MS analysis. So far, no optimal sample preparations, instead, traditional sample preparation has been used for AAV VP peptide mapping, resulting in exaggerating the original deamidation levels. It is important to accurately monitor and provide true value of asparagine deamidation for development of AAV gene therapy products. In this study, we evaluated denaturation temperatures, digestion durations, and digestion temperatures using three different sample preparation formats for LC-MS/MS-based assessment of deamidation of AAV9 capsid proteins. The results demonstrated that the optimal sample preparation method for AAV9 VP peptide mapping minimized asparagine deamidation artifacts. Although AAV9 was used for method optimization, this study may also provide a guidance on how to control deamidation artifacts for other AAV serotypes.

INTRODUCTION

In recent years, recombinant adenoviral and adeno-associated viral (AAV) vectors with nonpathogenic nature and ability to provide long-term gene expression have taken center stage as a gene delivery vehicle for gene therapy to treat a number of human diseases. AAV is a small nonenveloped virus with a single-stranded DNA genome that is encapsidated in an icosahedral protein capsid shell. The AAV icosahedral capsid is formed by three viral proteins, VP1, VP2, and VP3, at an ∼1:1:10 ratio.¹ These three capsid proteins are splice products that share the common C-terminus, with VP1 being the full-length protein, VP2, being the protein with size between VP1 and VP3, and VP3, the shortest.²

There are 13 AAV serotypes that show 51–99% identity in capsid amino acid sequence.³ AAV capsid proteins are the major components that determine the tissue specificity and immunogenicity, and play important roles in receptor binding, escape of the virus from the endosome, and transport of the viral DNA to the nucleus. The unique N-terminal region of VP1 is necessary for virus infectivity. In addition to a catalytic phospholipase A2 domain, motifs for protein interaction, endosomal sorting, and signal transduction in eukaryotic cells are identified on the VP1/2 N terminus.⁴ It was reported that an amino acid mutation from asparagine (N) at 57 to aspartic acid (D) in AAV8 and AAV9 decreased transduction efficiency by more than 50%.⁵

Deamidation is a common post-translational modification (PTM) of proteins, which may affect protein structural stability and lead to protein denaturation and aggregation and may therefore have significant negative biological consequences.^6

–10

Deamidation is a chemical reaction in which an amide functional group in the side chain of asparagine (N) or glutamine (Q) residue is removed or converted to a hydroxyl group, resulting in + 1 (0.98402) Da mass increase. However, the deamidation rate of asparagine (N) residue is much faster compared with glutamine (Q) residue, therefore, more asparagine deamidation are studied. Typically, asparagine is converted to aspartic acid (D) and/or isoaspartic acid (isoD). Several biophysical factors contribute to the kinetics of spontaneous deamidation. One of the most important factors is the ability of the backbone amide of the N + 1 amino acid (amino acid residue following asparagine) to act as a nucleophile and attack the side chain of the asparagine residue to form succinimide, as well as the flexibility of the peptide backbone.

The N + 1 residue is critical for the local flexibility of the peptide backbone, therefore the deamidation rate of the asparagine.^11

–14 Glycine (G) in the N + 1 position forming an NG motif has the greatest effect on deamidation, followed by histidine (H) and serine (S).^12
–14 Other factors that influence deamidation include the primary and higher-order structures of the proteins, as well as external conditions such as temperature, pH, and components in the solutions.^6,15

Liquid chromatography–tandem mass spectrometry (LC-MS/MS)-based peptide mapping is the most widely used technique to monitor asparagine deamidation. Due to the fact that deamidation can be greatly introduced during sample preparation before LC-MS/MS analysis, optimization of sample preparation conditions is necessary to minimize this artifact. There are two commonly used sample preparation methods for LC-MS/MS, in-solution digestion and in-gel digestion.¹⁶ In addition to these two methods, on-filter digestion,¹⁷ filter-aided sample preparation (FASP),¹⁸ and recently developed on-suspension trapping (S-Trap) digestion^19,20 are also used in the sample preparation for peptide mapping. Regardless of the sample preparation methods used, to date, there is no report on AAV sample preparation optimization to minimize the exaggerated level of original deamidation.

It is important to accurately monitor and provide true value of asparagine deamidation for development of AAV gene therapy products. In this study, we evaluated denaturation temperatures, concentration of denaturants, digestion durations, and digestion temperatures using three different sample preparation formats, in-solution, on-S-Trap, and in-gel digestions, for LC-MS/MS-based assessment of deamidation of AAV9 capsid proteins. The results demonstrated that the optimal sample preparation method for AAV9 peptide mapping reached much lower levels of asparagine deamidation artifacts. This is the first report on optimization of sample preparation for peptide mapping for characterization of AAV capsid proteins. This study provides a guidance on how to control deamidation artifacts for all serotypes of AAV peptide mapping to support AAV gene therapy development.

EXPERIMENTAL

Materials

AAV9 samples used for deamidation assessments were produced at Novartis Gene Therapies. All AAV9 samples were kept frozen at ≤−60°C before sample preparation for peptide mapping. The sample preparation of AAV9 samples for deamidation comparison was carried out at the same time. Ammonium bicarbonate, 1 M triethylammonium bicarbonate (TEAB) solution, 20% sodium dodecyl sulfate (SDS) solution in water, 2,2,2-trifluoroethanol (TFE), and iodoacetamide (IAM) were purchased from Sigma (St. Louis, MO). Water, acetonitrile, methanol, 0.1% trifluoroacetic acid (TFA) in acetonitrile, 0.1% TFA in water, 0.1% formic acid in acetonitrile, and 0.1% formic acid in water were LC-MS grade and purchased from Honeywell Burdick & Jackson (Muskegon, MI). LC-MS grade TFA, LC-MS grade formic acid, 0.5 M tris (2-carboxyethyl) phosphine (TCEP), and phosphoric acid (85%) were purchased from Thermo Fisher Scientific (Rockford, lL). Trypsin/LysC Mix Mass Spec Grade was purchased from Promega (Madison, WI). S-Trap was purchased from ProtiFi (Huntington, NY).

All of the following items for running SDS–polyacrylamide gel electrophoresis were purchased from Thermo Fisher Scientific (Carlsbad, CA): NuPage™ Novex™ 4–12% Bis-2-carboxyethyl phosphine midi gel, 12 + 2-well (1.0 mm), NuPage^® lithium dodecyl sulfate (LDS) Sample Buffer (4 × ), 20 × MOPS Running Buffer, SeeBluePlus2 Prestained Standard, NuPage 10 × Sample Reducing Agent, NuPAGE^® Antioxidant, and SimplyBlue™ SafeStain.

Methods

Sample preparation with In-solution digestion format

AAV9 sample was denatured with 50% TFE and reduced with 10 mM of TCEP at 70°C for 30 min, then alkylated with 10 mM of IAM at room temperature (RT) for 30 min in the dark. After the samples was diluted with 25 mM of Ammonium Bicarbonate to final TFE concentration lower than 10%, the sample was digested with trypsin at 1:20 w/w of enzyme to protein ratio with the concentration of trypsin greater than 4 ng/mL at 37°C for 1, 2, 4 h, and overnight. The trypsin digestion was quenched by adding TFA to the final concentration at 0.4% v/v. The samples were concentrated using a SpeedVac, the final volume of concentrated samples was adjusted with 25 mM ammonium bicarbonate to double the initial volume. The samples were then transferred into HPLC sample vials for LC-MS/MS analysis or kept at ≤−60°C before LC-MS/MS analysis.

Sample preparation with on-S-Trap digestion format

S-Trap is a new powerful FASP method that consists in trapping acid-aggregated proteins in a quartz filter before enzymatic proteolysis. The S-Trap protocol provided by ProtiFi was followed with some modifications.

Briefly, an AAV9 sample was denatured with 1% or 5% SDS solubilization/denaturation buffer and reduced with 10 mM of TCEP at 60°C or 95°C for 10 min, then alkylated with 10 mM of IAM at RT for 30 min in the dark after cooling the protein solution to RT. The sample was further denatured by acidification to ∼pH 1 by adding 12% aqueous phosphoric acid at 1:10 v/v for a final concentration of ∼1.2%. The acidified SDS sample was diluted six- to nine-fold using S-Trap-binding buffer, and the mixture of the acidified SDS sample and S-Trap-binding buffer was loaded on a S-Trap column. After washing the column with S-Trap-binding buffer four times, digestion buffer (50 mM TEAB in water, pH8.5) containing trypsin at 1:20 w/w of enzyme to protein ratio with the concentration of trypsin >4 ng/mL was added onto the top of the column. Samples on the Trap column were incubated for 1 h at 37°C or 47°C for trypsin digestion, or overnight at 37°C. Then the tryptic peptides were eluted with 0.2% aqueous formic acid first and then 50% acetonitrile containing 0.2% formic acid to recover hydrophobic peptides.

Sample preparation with In-gel digestion format

In-gel digestion of proteins isolated by gel electrophoresis is widely used for identification of impurities in protein samples by using MS-based proteomics. Briefly, AAV9 sample at ∼E13 vector genomes (vg)/mL was denatured and reduced in a sample reducing agent by incubating at 70°C for 15 min. The sample, along with Mark 12 protein standard, was run in a NuPAGE 4–12% Bis 2-carboxyethyl phosphine gel in 1 × MOPS buffer for 60 min at a constant voltage of 200 V. The gel was then stained using Simple Blue Gel Stain following the manufacturer's instructions. After destaining the gel, the gel bands of VP1, VP2, and VP3 were excised and separated from the SDS-PAGE with care of minimizing common laboratory contaminations, such as keratin, for mass spectrometry. The protocol²¹ for in-gel digestion was followed and different digestion period times with 3 h and overnight at 37°C, and 30 min at 55°C were tested.

Liquid chromatography–tandem mass spectrometry

The LC-MS/MS was performed for this study using a Thermo Scientific™ ULtiMate™ 3000 RSLC nano system (Thermo Fisher Scientific) coupled with Thermo Scientific Orbitrap Eclipse™ Tribrid™ Mass Spectrometer (Thermo Fisher Scientific). AAV9 tryptic peptides were separated on an EASY-Spray™ column, PepMap RSLC C18, 3 μm particle size, 75 μm × 150 mm (Thermo Fisher Scientific) at a column temperature of 35°C, using mobile phase A (0.1% formic acid in water) and mobile phase B (0.1% formic acid in acetonitrile) at a flow rate of 300 nL/min. The peptides were eluted with a gradient from 2% to 10% mobile phase B over 2 min, then 10–20% mobile phase B over 40 min, 20–40% mobile phase B over 12 min, 40–75% mobile phase B over 8 min, then held at 75% mobile phase B for 5 min before changing mobile phase B back to 2%.

Mass spectrometry data were acquired using a top 10 data-dependent acquisition method on an Orbitrap Eclipse™ Tribrid Mass Spectrometer fitted with a nanospray ionization source (EASY-Spray Source) (Thermo Fisher Scientific).

The mass spectrometer parameters used for peptide mapping are described below: Ion source properties include static spray voltage at 3,000 V for positive ion polarity mode, and ion transfer tube temperature at 300°C. MS full survey scan (m/z = 200–2,000) was acquired in the Orbitrap mass analyzer at a resolution of 120,000 at 200 m/z, the S-lens radio frequency level was set at 30%, Automatic Gain Control (AGC) target: standard, maximum injection time mode: auto. Peptides of charge states 2–7 were selected with a signal intensity threshold of 1 × 10⁴. Precursor ions with single, unassigned, or seven and higher charge states will be excluded. The method also used dynamic exclusion with 20 s exclusion duration and mass tolerance of low 10 ppm and high 10 ppm. MS² scans were acquired in the quadrupole mass analyzer using higher energy collision dissociation (HCD) at a collision energy of 30, with the isolation window of 0.7 m/z. The resolution for HCD spectra was set to 60,000 at 200 m/z.

Mass spectra data analysis

Chromatographic separation of peptides, sequence coverage, missed cleavage (1), and PTMs were used for the evaluation of peptide mapping methods. The Byos™ software (v3.8, Protein Metrics Inc.) was used for MS/MS data analysis to determine AAV9 VP primary structure, sequence coverage, and PTMs. AAV9 VP1 sequence (Uniprot Q6JC40) was used for performing sequence matching. For peptide identification, peptide MS/MS spectra match ≥2, and for protein identification, identified peptides ≥2 were set as acceptance criteria to assure the accuracy and confidence of peptide/protein identification and assignment.

Parameters applied for sequence matching include carbamidomethylation (+57.021464 Da) set as a fixed modification, and oxidation (+15.994915 Da), deamidation (+0.984016 Da), and N-terminal initial start methionine loss followed by acetylation on alanine, the second amino acid residue of AAV9 VP1 (−89.02992 Da) set as variable modifications, precursor mass tolerance set 10 ppm, and fragment mass tolerance set 15 ppm. These data were validated at a 1% false discovery rate using standard reverse-decoy techniques. The peptide spectra and all PTM mass spectra were also manually validated to confirm identification.

The deamidation percentage of an asparagine residue in a peptide was calculated by dividing the peak area of the deamidated peptide by the sum of the peak areas from the deamidated and unmodified peptides as shown in the following equation; the peak areas were based on extracted ion chromatograms (XIC) from the software. $% D e a m i d a t i o n = \frac{P e a k A r e a o f D e a m i d a t e d P e p t i d e}{P e a k A r e a o f D e a m i d a t e d P e p t i d e + P e a k A r e a o f U n m o d i f i e d P e p t i d e} \times 100 %$

Since asparagine deamidates into two products, aspartic acid (D) and isoD, the peak area of deamidated peptide includes peak areas of peptides in which asparagine converts to both aspartic acid and isoD.

RESULTS

There are 58 asparagine residues in AAV9 capsid protein VP1, including four NG motif sites—N57G, N329G, N452G, and N512G. As discussed in the Introduction section, many factors, including backbone flexibility and solvent-exposure environment, contribute to deamidation rates. Literature search on the crystal structure of AAV9 viral protein with starting amino acid D219 revealed that three of the NG motifs (N329G, N452G, and N512G) are located in or near surface-exposed region.^2,5 As N57 was not included in the crystal structure, the solvent-exposure environment is unknown. Based on our data, it was observed that the asparagine residues not with NG motif showed no or slight deamidation changes, therefore, deamidation assessment of asparagine residues with NG motif are reported in this study.

In-solution digestion

In-solution digestion samples were prepared with different digestion times for 1, 2, 4 h, and overnight at 37°C. Based on the results, the sequence coverages of the above samples were similar, >96% (data not shown), regardless of digestion times. Although the sequence coverages of all samples were similar, the deamidation of asparagine residues at four NG sites, N57G, N329G, N452G, and N512G, increased significantly in the sample with overnight digestion as shown in Fig. 1, indicating that overnight digestion at 37°C exaggerates deamidation and should not be used for deamidation analysis for AAV9 capsid proteins.

Figure 1.

AAV9 NG motif deamidation comparison with different in-solution digestion durations. The AAV9 DS samples were digested with trypsin in the format of in-solution digestion. The digestion was conducted at 37°C for 1 h (), 2 h (), 4 h (), and overnight (O/N) (). The deamidation levels of the four NG sites (N57, N329, N452, and N512) were plotted. AAV9, adeno-associated virus serotype 9.

Notably, the deamidation of asparagine residues at four NG sites, N57G, N329G, N452G, and N512G, in the samples digested for 1, 2, 4 h, and overnight showed similar increasing pattern. Because N57 deamidation in VP1 N-terminus impacts AAV9 transduction efficiency,⁵ N57 deamidation was mainly used for deamidation assessment in the following experimental studies.

On-S-Trap digestion

Solubilization/denaturation with 5%SDS and a digestion temperature of 47°C are recommended by the S-Trap supplier (ProtiFi). Two sets of experiments were carried out. The first experiment was set to compare solubilization/denaturation with 1%SDS and 5%SDS, denaturation temperature at 60°C and 95°C, and digestion temperature at 37°C and 47°C for 1 h. The second experiment was set to compare digestion time for 1 h and overnight at 37°C with 1%SDS for solubilization/denaturation and denaturation temperature at 60°C.

Based on the first set of experiment results, the sequence coverages of all samples were similar. The samples solubilized/denatured with 1%SDS or 5%SDS, denaturation at 60°C, and digestion at 37°C showed similar N57 deamidation level (Table 1A). As shown in Table 1B and C, denaturation temperature at 95°C or digestion temperature at 47°C resulted in higher deamidation than denaturation temperature at 60°C and digestion temperature at 37°C.

Table 1.

On suspension trapping digestion—adeno-associated virus serotype 9 N57 deamidation

Sample Preparation Conditions					Deamidation at N57
% SDS for Denaturation	Denaturation Temperature	Denaturation Duration	Digestion Temperature	Digestion Duration	Deamidation at N57
A: Comparison of denaturation with different concentrations of SDS
1% SDS	60°C	10 min	37°C	1 h	16.9%
5% SDS	60°C	10 min	37°C	1 h	14.6%
B: Comparison of denaturation temperatures
1% SDS	60°C	10 min	37°C	1 h	14.6%
1% SDS	95°C	10 min	37°C	1 h	35.7%
C: Comparison of digestion temperatures
1% SDS	60°C	10 min	37°C	1 h	14.6%
1% SDS	60°C	10 min	47°C	1 h	24.4%
D: Comparison of digestion durations
1% SDS	60°C	10 min	37°C	1 h	22.8%
1% SDS	60°C	10 min	37°C	Overnight	76.2%

The comparing conditions were in bold.

SDS, sodium dodecyl sulfate.

From the second set of experiment, the samples digested for 1 h or overnight at 37°C showed similar sequence coverages, however, the deamidation at N57 in the sample digested overnight was significantly higher than that in the sample digested for 1 h as shown in Table 1D. Therefore, for peptide mapping method using S-Trap, solubilization/denaturation with 1%SDS, denaturation temperature at 60°C, and digestion temperature at 37°C for 1 h are optimal sample preparation conditions for deamidation analysis for AAV9 capsid proteins.

In-gel digestion

The protocol²¹ for in-gel digestion for mass spectrometric characterization of proteins and proteomes was followed. In addition to the two digestion conditions, 37°C overnight and 55°C for 30 min recommended in the protocol, in-gel digestion at 37°C for 3 h was also carried out. Sequence coverages of in-gel digestion of AAV9 VP1 band under these three digestion conditions were similar, all greater than 96%. As shown in Table 2, N57 deamidation of in-gel digested AAV9 VP1 band at 55°C for 30 min was 20%, lower than that at 37°C for 3 h (26%) and overnight (57%). Overnight in-gel digestion resulted in significantly higher N57 deamidation of AAV9. Therefore, for deamidation analysis, in-gel digestion should not be carried out overnight.

Table 2.

In-gel digestion—adeno-associated virus serotype 9 N57 deamidation

In-Gel Digestion Temperature and Duration		% Deamidation at N57 (Average from Duplicate Injections)
Temperature	Duration	% Deamidation at N57 (Average from Duplicate Injections)
55°C	30 min	20.1% (±0.9%)
37°C	3 h	26.2% (±0.4%)
37°C	Overnight	56.8% (±0.4%)

Because separation of AAV VP1, VP2, and VP3 by SDS-PAGE is done before proteolytic enzyme in-gel digestion, the benefit of using in-gel digestion is that sequence coverages of VP1, VP2, and VP3 can be obtained individually. In contrast, it is difficult to get individual VP1, VP2, and VP3 sequence coverages using in-solution digestion or on-S-Trap digestion since VP1, VP2, and VP3 share a common C-terminal of over 530 amino acids.

DISCUSSION

In the biopharmaceutical industry, LC-MS-based peptide mapping has become a routinely used method to confirm protein identity, and qualitative (nontargeted) or quantitative (targeted) analysis of PTMs such as deamidation.

The peptide mapping involves a digestion step in which proteins are cleaved at amide bonds between specific amino acid residues by proteolytic enzymes to generate a predictable set of peptides. Most proteins exist as large complex structures with unique three-dimensional shapes/higher-order structures for their functions; the higher-order structures of proteins may hinder full access of proteolytic enzymes to cleavage sites and result in low sequence coverage.

The efficiency of digestion is dependent on the degree of denaturation of a protein before its digestion. Generally, the first step of peptide mapping is to denature proteins, a critical step for efficient proteolytic enzyme digestion. There are three common ways to denature proteins, denaturation with chaotropic agents and surfactants,²² volatile solvent denaturation,^22,23 and heat denaturation.²⁴

Guanidinium chloride or urea at high concentration is the commonly used chaotropic agent for in-solution digestion. However, the high concentration of chaotropic agents can reduce enzyme activity and are not compatible with mass spectrometry. Usually an additional step, buffer exchange using filtration or dialysis, is carried out to remove chaotropic agents from the sample solutions before digestion. Offline buffer exchange is time consuming and difficult to automate. On the other hand, volatile solvents for nonsalt-based denaturation of proteins can be diluted to a proper concentration suitable for retention of enzyme, such as tryptic activity, and easily removed from samples using a SpeedVac before MS analysis. Thermal denaturation is preferred over chemical denaturation because it does not require buffer exchange or concentration before MS analysis. However, longer thermal denaturation at higher temperature may be needed to completely denature some proteins, resulting in asparagine deamidation artifacts.

In general, a combination of denaturation methods is used for efficient protein denaturation and reduced asparagine deamidation artifacts. In this study, the combination of solvent denaturation using 50% TFE and heat at 70°C for 30 min for in-solution digestion, the combination of denaturation with surfactant 1%SDS and heat at 60°C for 10 min for on-S-Trap digestion, and the combination of denaturation with surfactant LDS and heat at 70°C for 15 min for in-gel digestion were used for optimal denaturation of AAV9 capsid proteins. As shown in Table 1B, a denaturation temperature of 95°C resulted in more deamidation at N57 of AAV9 than denaturation at 60°C. Optimization of denaturation conditions is needed for peptide mapping for AAV9 capsid proteins, such as deamidation monitoring.

The optimal digestion temperature of most proteolytic enzymes is 37°C. As shown in Fig. 1, Tables 1 and 2, the longer digestion duration at 37°C, the higher AAV9 N57 deamidation level was observed, indicating that the conversion of asparagine at 57 to aspartic acid and/or isoD in AAV9 is enhanced with elongated digestion time at the optimal digestion temperature (37°C). Decreasing the digestion time for tryptic digestion could dramatically reduce deamidation artifacts of AAV9 capsid proteins. Therefore, optimization of digestion conditions is also needed for peptide mapping for deamidation assessments of AAV9 capsid proteins.

Sequence alignment of VP1 N-termini of 13 AAV serotypes is shown in Fig. 2. All 13 AAV serotypes capsid proteins have the same deamidation motif, NG, asparagine at 57 or 56 (AAV4, AAV5, and AAV13), and followed by glycine. The underlined amino acids in Fig. 2 indicate that the theoretical tryptic peptides containing N57 or N56 (AAV4, AAV5, and AAV13) of each serotype can be generated by trypsin. Therefore, deamidation at N57 or N56 is not AAV serotype specific and the optimized denaturation and digestion conditions used in this study for AAV9 capsid protein peptide mapping will be suitable for peptide mapping for other AAV serotype capsid proteins.

Figure 2.

Sequence alignment of VP1 N-termini of 13 AAV serotypes. Asparagine residue at 57 or 56 (AAV4, AAV5, and AAV13) are highlighted in red, and the N + 1 glycine residues are in bold in the NG motif. The underlined amino acids indicate the theoretical tryptic peptides containing N57 or N56 of each serotype, which can be generated by trypsin.

Since higher pH (>7) can also enhance the conversion of asparagine to aspartic acid and/or isoD, we compared N57 deamidation in AAV9 samples digested at different pHs with pepsin versus trypsin using on-S-Trap digestion format. The digestion buffer for pepsin was LC-MS grade water, pH 3–5 adjusted using hydrochloric acid. The digestion buffer for trypsin was 50 mM aqueous TEAB, pH 8.5, same as described in the section of sample preparation with on-S-Trap digestion format. Similar N57 deamidation was observed in the AAV9 samples digested with pepsin (17.5 ± 7.4%) and trypsin (16.3 ± 0.3%), indicating that N57 deamidation was not enhanced by the sample preparation during the digestion using the digestion buffer at pH 8.5.

The error bar from trypsin digestion was small, yet large for pepsin digestion. This is because trypsin cuts peptide chains specifically at the carboxyl side of lysine or arginine residues, and only one unmodified peptide (YLGPG N ⁵⁷GLDK) peak area and two deamidated peptides (YLGPG D ⁵⁷GLDK and YLGPG isoD ⁵⁷GLDK) peak areas were used for % deamidation calculation. In contrast, pepsin cuts peptide chains nonspecifically, % deamidation of AAV9 N57 was calculated from the average of peak areas of three unmodified peptides (KYLGPG N ⁵⁷GLD, LVLPGYKYLGPG N ⁵⁷GLD, PGYKYLGPG N ⁵⁷GLD) and six deamidated peptides (KYLGPG D ⁵⁷GLD, KYLGPG isoD ⁵⁷GLD, LVLPGYKYLGPG D ⁵⁷GLD, LVLPGYKYLGPG isoD ⁵⁷GLD, PGYKYLGPG D ⁵⁷GLD, PGYKYLGPG isoD ⁵⁷GLD).

The identification of deamidated peptides by mass spectrometry is straightforward, as deamidation adds +0.984 Da to the mass of an intact peptide, yet it is challenging for mass spectrometry to differentiate the two isomers (D and isoD) through collision-induced dissociation (CID) or high energy CID (HCD), due to the fact that they have the same elemental composition, identical mass, same formal charge, and indistinguishable fragmentation patterns. Fortunately, the structural changes induced by isomerization alter the retention time of the peptide in reversed-phase LC, resulting in isomer separation, peaks are typically assigned based on the elution order,²⁵ which may vary depending on the chromatographic conditions.¹⁴

Our mass spectrometry data showed that AAV9 asparagine 57 deamidates to both aspartic acid and isoD since there is one peak from the undeamidated tryptic peptide (YLGPG N ⁵⁷GLDK), yet two peaks from the corresponding deamidated tryptic peptide in the plot of the XIC. In the current work, CID/HCD was used and the two isomers cannot be distinguished, therefore, the peak areas from both peptides are summed to represent the total deamidated form.

It was reported that electron capture dissociation or electron transfer dissociation mass spectrometry can generate a single pair of reporter ions (c + 57 and z − 57) that are unique to isoD, which may be helpful to distinguish these two isomers.^14,25

CONCLUSION

The results of this study indicate that sample preparation matters for assessment of deamidation of AAV capsid proteins. Sample preparation using overnight digestion introduced extreme deamidation artifacts. However, overnight digestion was previously reported for LC-MS/MS-based characterization of AAV capsid proteins.^3,5,26 This is the first report on optimization of sample preparation for peptide mapping for characterization of AAV capsid proteins. This study will provide a guidance on how to control deamidation artifacts using optimal denaturation and digestion conditions for all serotypes of AAV peptide mapping to support AAV gene therapy development.

Footnotes

AUTHOR DISCLOSURE

No competing financial interests exist.

FUNDING INFORMATION

No funding was received for this article.

References

Rose

, Maizel

Jr , Inman

, et al. Structural proteins of adenovirus-associated viruses. J Virol, 1971; 8:766–770.

DiMattia

, Nam

, Vliet

, et al. Structural insight into the unique properties of adeno-associated virus serotype 9. J Virol, 2012; 86:6947–6958.

Jin

, Liu

, Nass

, et al. Direct liquid chromatography/mass spectrometry analysis for complete characterization of recombinant adeno-associated virus capsid proteins. Hum Gene Ther, 2017; 28:255–267.

Popa-Wagner

, Porwal

, Kann

, et al. Impact of VP1-specific protein sequence motifs on adenoassociated virus type 2 intracellular trafficking and nuclear entry. J Virol, 2012; 86:9163–9174.

Giles

, Sims

, Turner

, et al. Deamidation of amino acids on the surface of adeno-associated virus capsids leads to charge heterogeneity and altered vector function. Mol Therapy, 2018; 26:2848–2862.

Robinson

, Robinson

. Molecular clocks. Proc Natl Acad Sci USA, 2001; 98:944–949.

Solstad

, Flatmark

. Microheterogeneity of recombinant human phenylalanine hydroxylase as a result of nonenzymatic deamidations of labile amide containing amino acids. Effects on catalytic and stability properties. Eur J Biochem, 2000; 267:6302–6310.

Dudek

, Lampi

, et al. Ubiquitin proteasome pathway-mediated degradation of proteins: effects due to site-specific substrate deamidation. Investig Ophthalmol Vis Sci, 2010; 51:4164–4173.

Deverman

, Cook

, Manson

, et al. Bcl-xL deamidation is a critical switch in the regulation of the response to DNA damage. Cell, 2002; 111:51–62.

10.

Kato

, Nakayoshi

, Kurimoto

, et al. Mechanisms of deamidation of asparagine residues and effects of main-chain conformation on activation energy. Int J Mol Sci, 2020; 21:7035–7048.

11.

Catak

, Monard

, Aviyente

, et al. Deamidation of asparagine residues: direct hydrolysis versus succinimide-mediated deamidation mechanisms. J Phys Chem A, 2009; 113:1111–1120.

12.

Robinson

, Robinson

. Prediction of protein deamidation rates from primary and three-dimensional structure. Proc Natl Acad Sci USA, 2001; 98:4367–4372.

13.

Robinson

. Protein deamidation. Proc Natl Acad Sci USA, 2002; 99:5283–5288.

14.

, Dai

, Karger

, et al. Analysis of isoaspartic acid by selective proteolysis with Asp-N and electron transfer dissociation mass spectrometry. Anal Chem, 2010; 82:7485–7491.

15.

Pace

, Wong

, Zhang

, et al. Asparagine deamidation dependence on buffer type, pH, and temperature. J Pharma Sci, 2013; 102:1712–1723.

16.

Zhang

, Fonslow

, Shan

, et al. Protein analysis by shotgun/bottom-up proteomics. Chem Rev, 2013; 113:2343–2394.

17.

Manza

, Stamer

, Ham

A-JL

, et al. Sample preparation and digestion for proteomic analyses using spin filters. Proteomics, 2005; 5:1742–1745.

18.

Wisniewski

, Zougman

, Nagaraj

, et al. Universal sample preparation method for proteome analysis. Nat Methods, 2009; 6:359–362.

19.

HaileMariam

, Eguez

, Singh

, et al. S-Trap, an ultrafast sample-preparation approach for shotgun proteomics. J Proteome Res, 2018; 17:2917–2924.

20.

Zougman

, Selby

, Banks

. Suspension trapping (STrap) sample preparation method for bottom-up proteomics analysis. Proteomics, 2014; 14:1006–1010.

21.

Shevchenko

, Tomas

, Havlis

, et al. In-gel digestion for mass spectrometric characterization of proteins and proteomes. Nat Protoc, 2006; 1:2856–2860.

22.

Proc

, Kuzyk

, Hardie

, et al. A quantitative study of the effects of chaotropic agents, surfactants, and solvents on the digestion efficiency of human plasma proteins by trypsin. J Proteome Res, 2010; 9:5422–5437.

23.

Adach

, Kumar

, Zhang

, et al. The human urinary proteome contains more than 1500 proteins, including a large proportion of membrane proteins. Genome Biol, 2006; 7:R80.

24.

Park

, Russell

. Thermal Denaturation: a useful technique in peptide mass mapping. Anal Chem, 2000; 72:2667–2670.

25.

Yang

, Zubarev

. Mass spectrometric analysis of asparagine deamidation and aspartate isomerization in polypeptides. Electrophoresis, 2010; 31:1764–1772.

26.

Mary

, Maurya

, Arumugam

, et al. Post-translational modifications in capsid proteins of recombinant adeno-associated virus (AAV) 1-rh10 serotypes. The FEBS J, 2019; 286:4964–4981.

Sample Preparation Matters for Peptide Mapping to Evaluate Deamidation of Adeno-Associated Virus Capsid Proteins Using Liquid Chromatography–Tandem Mass Spectrometry

Abstract

INTRODUCTION

EXPERIMENTAL

Materials

Methods

Sample preparation with In-solution digestion format

Sample preparation with on-S-Trap digestion format

Sample preparation with In-gel digestion format

Liquid chromatography–tandem mass spectrometry

Mass spectra data analysis

RESULTS

In-solution digestion

On-S-Trap digestion

In-gel digestion

DISCUSSION

CONCLUSION

Footnotes

AUTHOR DISCLOSURE

FUNDING INFORMATION

References