Rapid Highly-Efficient Digestion and Peptide Mapping of Adeno-Associated Viruses

Abstract

Adeno-associated viruses (AAVs) comprise an area of rapidly growing interest due to their ability to act as a gene delivery vehicle in novel gene therapy strategies and vaccine development. Peptide mapping is a common technique in the biopharmaceutical industry to confirm the correct sequence, product purity, post-translational modifications (PTMs), and stability. However, conventional peptide mapping is time-consuming and has proven difficult to reproduce with viral capsids because of their high structural stability and the suboptimal localization of trypsin cleavage sites in the AAV protein sequences. In this study, we present an optimized peptide mapping-based workflow that provides thorough characterization within 1 day. This workflow is also highly reproducible due to its simplicity having very few steps and is easy to perform proteolytic digestion utilizing thermally stable pepsin, which is active at 70 °C in acidic conditions. The acidic conditions of the peptic digestions drive viral capsid denaturation and improve cleavage site accessibility. We characterized the efficiency and ease of digestion through peptide mapping of the AAV2 viral capsid protein. Using nanoflow liquid chromatography coupled with tandem mass spectrometry, we achieved 100% sequence coverage of the low-abundance VP1 capsid protein with a digestion process taking only 10 min to prepare and 45 min to complete the digestion.

Graphical Abstract

A deno-associated viruses (AAVs) comprise an area of rapidly growing interest due to their ability to act as a gene delivery vehicle, or vector, in novel gene therapy strategies. AAV therapy has been shown to modulate long-term gene expression and disease correction with low toxicity in animal models, and the AAV treatment-based approaches also demonstrated impressive performance in human clinical trials.¹ AAVs and other viral vectors have also been used for vaccine production.² The nonpathogenic viral vector can deliver the genes required for an immune response to targeted cells with good infectivity. They can also induce high immunogenicity without an adjuvant.² Many types of viruses have now been developed as vaccine vectors, including recent COVID-19 vaccines.^3,4 These viral capsid proteins assemble into an icosahedral protein capsid shell (i.e., with 20 triangular equilateral faces) with a molecular mass of ~4 MDa encapsulating a single-stranded DNA genome. The primary sequence of VP1 contains the entire sequence of VP2 and VP3 within it.² The assembled AAV structure is very stable, which adds to the viability as a drug delivery vector. The AAV surface structure mediates its genome trafficking, cell binding, and subsequent internalization. There are 11 known serotypes of AAVs, and the AAV serotype greatly influences the virus target cell and tissue specificity.^5–7

AAV therapy has been used to modulate long-term gene expression, and AAV vaccines have proven to induce high immunogenicity. The AAV surface structure mediates its genome trafficking, cell binding, and subsequent internalization. Changes in the viral capsid protein sequence or post-translational modifications (PTMs) could impact viral targeting and infectivity, so it is essential to have robust technologies to structurally characterize the final vector product thoroughly.

Approaches to target-specific cellular receptors have utilized the structural motifs of the different AAV serotypes along with genetically engineered modifications typed of original vectors, resulting in chimeric and mosaic AAV capsids.⁸ These capsid modifications are being incorporated into vector production and purification methods that provide the ability to scale up the manufacturing process to support human clinical trials. The structures of multiple AAV serotypes have been resolved through X-ray crystallography, and changes in their structures with the response to environmental factors have been studied by cryo-electron microscopy and circular dichroism.^9,10 Changes in the viral capsid protein sequence or PTMs could impact viral targeting and infectivity.¹¹ Thus, it is essential to have facile and robust technologies to structurally characterize the final vector product thoroughly.

Structural characterization of therapeutic proteins via peptide mapping is a common technique in the biopharmaceutical industry to confirm the correct sequence, product purity, PTMs, and stability.^12–17 Conventional sample preparation for thorough peptide mapping involves several time-consuming steps, including (1) denaturation with chaotropic agents (urea or guanidine), (2) reduction of the disulfide bridges with dithiothreitol or tris(2-carboxyethyl)-phosphine (TCEP), (3) alkylation of the free thiol groups with iodoacetamide or other alkylating reagents, and (4) buffer exchange or dilution to remove the chaotropes, which would otherwise denature the protease required for enzymatic digestion. These steps are followed by a slow digestion step of several hours, often overnight, due to the low enzyme-to-protein ratio essential to avoid high levels of autolysis products from the protease.^18,19 Several recent studies attempted to optimize peptide mapping of biotherapeutics to increase reproducibility and preserve the integrity of PTMs through decreasing the number of sample preparation steps and processing time as well as by optimizing buffer composition and other conditions in sample preparation.^20,21

In the majority of peptide mapping workflows, protein denaturation is an essential first step because unfolded proteins yield greater protease accessibility to the target cleavage sites and thereby result in higher digestion efficiency, leading to more informative sequence maps. In addition to chaotrope-driven, protein denaturation can also be achieved thermally, where heat is applied to samples to reach the temperature just above the melting point (TM) of the majority of proteins. At this temperature, the protein unfolds and loses its higher-order structure, resulting in denaturation. The effect of heat as a denaturant on the AAV viral capsid has been demonstrated through electron microscopy and circular dichroism.^5,6 The thermal stability of viral capsids has been studied both as a means of serotype identification and to understand the mechanism of recognition and infection. In the thermally induced denaturation, the temperature was found to be serotype-specific and can be influenced by solvation factors such as pH and buffering salt.⁵ Cryo-EM studies showed a loss of structural integrity at high temperatures and confirmed that this thermal shift was correlated to the denaturation of the protein.^5,6 AAV serotypes 1–10 were found to have TMs ranging from 66 to 90 °C, with AAV2 being least stable and AAV5 being most stable. All AAV serotypes had unique TMs except for AAV7 and AAV9; however, by altering the buffering ion, there was a shift in the TM of AAV7 by 3 °C to allow differentiation for these serotypes.⁵ All serotypes have a reduced melting point at lower pH and denature promptly at pH 4.0.

There is a known difficulty in digesting AAV capsid proteins with the commonly used in-solution trypsin digestion protocols.¹⁹ For example, there are stretches of AAV2 viral capsid protein sequence, which give rise to very short (≤3 amino acid residues) and very long (≥60 amino acid residues) tryptic peptides, resulting in suboptimal elution and recovery properties on reversed-phase stationary phases and poor fragmentation efficiency in LC-MS/MS analysis.¹⁸ These deficiencies lead to incomplete sequence coverage and structural as well as quantitative characterization of the protein. In a recent study, three separate protease digestions using enzymes with different cleavage specificities followed by LC-MS of each digest were required to achieve full sequence coverage of an AAV sample.²² Also, the viral particles are very stable at neutral pH, which is necessary for tryptic digestion. This stability reduces significantly as the pH is lowered.⁵

In this study, we examined the enzyme-immobilized beads with different protease activities for their ability to fully digest the AAV capsids. The beads utilize thermally stable enzymes, which are optimally active at 70 °C and are proposed as a solution to simplify and improve the digestion process. Digestion was performed under protein denaturing conditions to minimize the number of steps in the sample processing. We characterized the efficiency and ease of digestion through peptide mapping of the AAV2 viral capsid protein. Using nanoflow liquid chromatography coupled with tandem mass spectrometry (nanoLC-MS/MS), we explored the ability of thermally active trypsin and pepsin to generate proteolysis products and the resulting sequence coverage of the AAV2 viral capsid protein. As an alternative to the commonly used trypsin, in this work, we chose to evaluate pepsin for AAV2 peptide mapping because of its strong activity at very low pH and the orthogonal target site-specificity.²³

EXPERIMENTAL SECTION

Chemicals and Reagents.

Water with 0.1% formic acid (Optima, LC-MS grade) and acetonitrile with 0.1% formic acid were obtained from Fisher Scientific. All reagents and materials required for SMART digestion kits were obtained from Thermo Fisher Scientific (Sunnyvale, CA).

The AAV2 samples were purchased from Vector Biolabs (Malvern, PA) at a concentration of 1e13 GC in 500 μL of phosphate buffer. This quantity is calculated to be roughly 5 μg of material with a concentration of 1 μg/25 μL.

Trypsin High-Temperature Proteolysis.

SMART Digest kit protocol was followed with minor adjustments. A volume of 25 μL of water was added to 25 μL of the AAV sample and then diluted 1:4 (v/v) with the SMART Digest buffer (50 mM Tris pH 7.2) provided with the kit. A volume of 5 μL of 0.5 M TCEP was added to the digestion mixture. To initiate digestion, 15 μL of the Trypsin Magnetic SMART digest resin was added to the sample (corresponding to 14 μg of heat-stable immobilized trypsin), and the reaction vessels were placed on an Eppendorf thermomixer equilibrated at 70 °C.

In all samples, enzymatic digestion was allowed to proceed at 70 °C for 45 min while shaking at a rate of 1400 rpm. After the digestion, the digestion vessel was placed in a side magnetized plate, and the supernatant was transferred to a fresh tube. To ensure that no residual resin was present in the sample, the side magnetic attraction of enzymatic beads and transfer of supernatant was repeated. All samples were diluted with 0.1% formic acid (FA) in water to a final protein concentration of 1 μg in 200 μL, and 5 μL of this digest mixture was loaded on the column for each chromatographic analysis.

Pepsin High-Temperature Proteolysis.

A volume of 25 μL of water was added to the 25 μL AAV sample and then diluted 1:4 (v/v) with the Pepsin SMART Digest buffer (pH 2) provided with the kit. A volume of 5 μL of 0.5 M TCEP was added to the digestion mixture. To initiate digestion, 15 μL of the Magnetic SMART digest resin was then added to the sample (corresponding to 14 μg of heat-stable immobilized pepsin), and the reaction vessels were placed on an Eppendorf thermomixer equilibrated at 70 °C.

In all samples, the peptic digestion was allowed to proceed at 70 °C for 45 min at 1400 rpm. After the digestion, the digestion vessel was placed in a side magnetized plate, and the supernatant was transferred to a fresh tube. To ensure no residual resin was present in the sample, the side magnetic attraction of enzymatic beads and transfer of supernatant was repeated. All samples were diluted with 0.1% formic acid (FA) in water to a final protein concentration of 1 μg in 200 μL, and 5 μL of this digest mixture was loaded on the column for each chromatographic analysis.

LC-MS of AAV Digests.

Nanoflow liquid chromatography with tandem high-resolution accurate mass spectrometry (nano-LC-MS/MS) peptide mapping analysis for AAV2 was performed on an EASY nLC-1200 nano-HPLC online with an Orbitrap QE Plus hybrid mass spectrometer (Thermo Fisher Scientific). The separation was performed using an EASY Spray column (ES803A), packed with Acclaim PepMap C18 stationary phase which has an i.d. of 75 μm, an o.d. of 500 mm, and a particle size of 2 μm with 100 Å pore size. A flow rate of 300 nL/min and a column temperature of 40 °C were the conditions used throughout the separation. The peptides were eluted using a linear gradient from 2% to 60% B (0.1% formic acid in acetonitrile) over 100 min. The source parameters for the mass spectrometer were as follows: source voltage, 1.8 kV; capillary temperature; 320 °C; S-lens RF level of 70%. Data were acquired using a top 10 data-dependent method with a full scan resolution of 70 000 (300–2000 m/z), and MS/MS was done using higher energy collisional excitation (HCD) at a normalized collision energy of 30 at a resolution of 17 500. The full scan data accuracy was improved by mass locking on the 445.1200 m/z ambient air contaminant.

Mass spectrometer parameters were set for an AGC target of 5 × 10⁵ with a maximum IT of 55 ms. The minimum intensity required for MS2 was set to 2 × 10⁴ with a dynamic exclusion time of 60 s. Singly charged ions were excluded for tryptic digests. However, as pepsin digests produce an abundance of singly charged ions, these were not excluded.

The peptide identification was performed using BioPharma Finder for mass fingerprinting against the AAV2 VP1 viral capsid protein sequences with MS and MS/MS tolerances of 10 ppm. The sequence of VP1 was used for the peptide mapping analysis as the sequence of VP2 and VP3 are contained in VP1. N-terminal acetylation, methionine oxidation, glycation, and asparagine deamidation were included as variable modifications. Cleavage specificity for trypsin was defined as highly specific for cleavage at KR residues. Cleavage specificity for pepsin was defined as low specificity for cleavage at residues CDEFLMTWY. The cleavage positions can be automatically displayed with the determined abundance values in the BioPharma Finder software. The specificity of the immobilized enzyme was found to be comparable with previously reported specificities using the in-solution pepsin. Furthermore, duplicated experiments show the specificity of the immobilized enzyme to be stable and the qualitative and quantitative results of peptide mapping experiments were reproducible, which indirectly confirmed the high level of control over digestion conditions.

RESULTS AND DISCUSSION

Experimental Workflow.

The aim of this study was to develop an efficient and easy-to-use sample processing method for the viral capsids that would be amenable to automation prior to LC-MS/MS characterization. The full workflow is schematically represented in Figure 1A. In short, heat-induced capsid protein denaturation was used with heat-stable proteases to remove several sample processing steps, which are typically necessary with chemical denaturants. In the peptic digestion, the digestion buffer, with a pH of 3.5, provided an additional layer of pH-induced denaturation. TCEP was added to all digest mixtures to reduce disulfide bonds and unfold both the capsid and viral proteins to allow better access to the specific cleavage sites. The chemical stability of TCEP prevents disulfide bond scrambling for peptides containing free cysteine residues after digestion. The digestion is initiated by increasing the temperature. The reaction is stopped by removing the magnetic beads from the samples at the end of the set digestion time. Pipetting errors typically introduced during dispensing of reagents in conventional protocols, using a set trypsin-to-substrate ratio, are eliminated in the developed approach that relies on magnetic beads due to the excess of the immobilized protease. This is particularly important with AAV samples which are typically dilute and variable in protein concentration. The magnetic bead-based platform results in a highly reproducible sample processing that is very easy to use. These are the significant improvements over traditional in-solution digestion methods that require more steps in the experimental protocol and a substantially longer time.

Figure 1.

Experimental workflow and the main outcomes of the AAV characterization using immobilized pepsin. (A) Description of experimental workflow yielding (B) 81% and 100% sequence coverage of the VP1 component of AAV2 when digesting using trypsin and pepsin, respectively. The sequence gaps that cannot be characterized using the trypsin digestion are highlighted in red, with the corresponding amino acid residue positions of these gaps indicated in the blue boxes (on the right). Deamidation and oxidation sites are highlighted in green and purple, respectively. (C) Comparative detection of process-induced deamidation and oxidation reveals that peptic proteolysis conditions practically eliminate deamidation artifacts.

Following AAV viral capsid protein proteolysis, digest mixtures were analyzed by nano LC-MS/MS. RAW files were parsed and analyzed using BioPharma Finder software searching against the AAV2 VP1 viral capsid protein sequence. Sequence coverage for tryptic digestion and peptic digestion were compared as well as the observation of process-induced modifications.

The two proteases chosen for evaluation in this study included trypsin, which is commonly used for peptide mapping, and pepsin, which is not a common enzyme in protein characterization. Trypsin is by far the most widely used proteolytic enzyme due to its high specificity for cleavage at the carboxyl side of arginine and lysine residues.²⁴ However, there are proteins, as in the case of the AAV capsids, that are challenging for trypsin due to the peculiarities of the protein primary and higher-order structures. Such structural features include either closely or distantly spaced cleavage sites that result in producing very small or very large peptides or may lead to spatial hindrance making the cleavage sites inaccessible to trypsin. Both very small and very large peptides are suboptimal for reversed-phase C18-based chromatography and tandem MS fragmentation typically used for peptide mapping in protein characterization.²⁵ Pepsin has alternative cleavage specificity to trypsin, cleaving primarily at the hydrophobic aromatic amino acid residues. Pepsin is known to cleave after phenylalanine and leucine with additional cleavage sites of glutamic acid and aspartic with methionine, tyrosine, threonine, asparagine, and glutamine, showing lower rates of digestion.²⁶ Unlike the high specificity of trypsin, the pepsin digestion pattern alters as the digestion time increases due to slower rates of cleavage activity at alternative digestion sites. These properties largely discouraged the use of pepsin. However, the time of digestion can be carefully controlled by using magnetic beads with the enzyme immobilized at high density. The protease-covered beads can be added and removed from the heated sample wells with high precision, and the activity of the immobilized enzyme can be accurately controlled. Unlike trypsin, pepsin is highly active at low pH. We hypothesized that the proteolytic activity at low pH and the acidic environment would synergistically increase the denaturation efficiency for viral particles. We also hypothesized that an additional potential advantage of pepsin for the characterization of AAV particles isolated using an affinity capture column would be the direct compatibility of the digestion conditions with the low pH elution conditions to release the isolated capsids from the immunoaffinity stationary phase.

Following digestion, we employed nanoflow LC coupled to tandem mass spectrometry for peptide mapping (Figure 1A). The goal was to enhance the sensitivity of the capsid protein characterization by using nanoflow LC, which can be critical due to the low concentration and limited availability of protein found in most AAV samples. In the initial experiments, we used trypsin digestion to establish the baseline in our AAV capsid protein characterization and to assess the challenges described in early reports with the use of this protease for the AAV protein characterization. Figure 1B shows the sequence coverage achieved when using the heat-stable trypsin magnetic beads. The total sequence coverage was 81.2% (full sequence map available in Figures S1 and S2, sequence coverage is shown in Table S1 and Supplementary Table 2), which falls short of the desired 100% coverage mainly because of large regions of the protein sequence that do not contain trypsin cleavages sites. The short sequence from Lys₁₆₁ to Arg₁₇₀ contains six hydrophilic amino acid residues with three consecutive trypsin cleavage sites. The peptides produced from this region are unlikely to be retained by the reversed-phase column, and thus, they were never detected. Following this region, there are no other trypsin cleavage sites up to Arg₂₃₈, which would produce a large 67 amino acid residue-long peptide.

Other sequence gaps, 361–381 and 666–688, contain multiple hydrophobic aromatic amino acid residues, which can be challenging to elute off the C18 column and even keep in solution due to their high hydrophobicity. Although the heat-enabled digestion with trypsin showed some success, the positions of the specific trypsin cleavage sites produced several peptides that are not ideal for tandem MS fragmentation, which then resulted in low sequence coverage (Figure 1B).

Next, we explored the use of an alternative digestion enzyme, pepsin. Pepsin is commercially available on magnetic beads. While these enzymes are not as specific as trypsin, they do have orthogonal site specificities. From the trypsin digestion data, we can see a number of hydrophobic amino acid residues in the missing sequence gaps following tryptic cleavage sites. The data obtained from pepsin digestion demonstrated cleavages at such sites inside the hydrophobic peptide regions not covered using trypsin digestion. This improvement resulted in the increased to 100% sequence coverage for the capsid protein, as shown in Figure 1B (full sequence map available in Figures S1 and S2).

Expectedly, more peptides are produced with pepsin digestion due to the reduced cleavage specificity of the enzyme and a greater number of susceptible cleavage sites. All observed peptic peptides are reported in the Supplementary File 2. We compared the relative quantitation of selected post-translational modifications for the detected peptide populations, shown in Figure 1C. Oxidation of methionine residues, which could be caused by both post-translational modifications and sample processing, was observed at two sites with pepsin and one site with trypsin due to lack of coverage of M235 in the tryptic sequence map. The degree of oxidation at M604 were very similar using both digestion techniques, indicating that these modifications could be reliably monitored using both the pepsin and trypsin digestion conditions. The levels of deamidation showed significant differences between the two digestion conditions (Figure 1C). This phenomenon can be explained by the differences in the pH at which the digestions were performed. Moderate-level deamidation levels between 4 and 10% were observed at five different sites as a result of trypsin digestion that was conducted at pH 7.2 for 45 min at 70 °C. All these sites except for N552 are in close proximity to glycine, serine, or threonine residues, known to increase susceptibility to deamidation.²⁷ High pH and high temperature are also known to increase deamidation rates.²⁸ This is indirectly confirmed by the pepsin digestion conducted under acidic conditions. The analysis of peptic digests showed only minor deamidation at a single site, N382 (Figure 1C). This amino acid residue is in the sequence region that corresponds to the Gap2 361–389 in the sequence coverage that could not be characterized based on the tryptic digest. Therefore, the level of detected deamidation could not be directly compared with the tryptic digestion experiments. The similarity in the levels of oxidation between the two digestion protocols and the known challenges with sample processing-induced deamidation at higher pH and temperature suggest that pepsin digestion may yield more accurate and reproducible quantitative data in the characterization of protein structure and post-translational modifications of AAV capsid proteins.

Figure 2 demonstrates the sequence gaps in sequence coverage of VP1 in the tryptic digestion experiments (shown in red font in Figure 1B) that has the sequence information recovered with peptides released from the pepsin digestion. These sequence regions of VP1 (Gap 1 residues, 162–145; Gap 2 residues, 361–389; Gap 3 residues, 666–688) contain a relatively high level of hydrophobic amino acid residues. Even the short peptide FMVPQYGY resulting from digestion with pepsin is retained strongly and elutes late in the gradient at a retention time of 103 min. The peptic peptides required to fill in each tryptic sequence gap (Figure 2) are easily detected with high signal intensity and show high-quality MS/MS fragmentation data, resulting in high confidence peptide identifications.

Figure 2.

Improved characterization of commonly challenging protein sequence sections. Sequence gaps in the tryptic digestion experiments (162–245, 361–381, and 666–688 shown in Figure 1) contain multiple hydrophobic aromatic amino acid residues, which can be challenging to elute off the C18 column and even keep in solution. These gaps in sequence coverage are recovered with peptides released using the pepsin digestion. Representative examples of peptide sequences, fragmentation patterns, and the corresponding extracted ion chromatograms are shown in panels A-C.

The peptides released as an outcome of trypsin digestion have a charged lysine or arginine residue at the C-terminus. The peptides resulting from tryptic digestion of VP1 typically lead to electrospray mass spectrometry-based detection of ion species with charge states of 2+ and higher. The digestion specificity of pepsin typically produces peptides that have a hydrophobic aromatic amino acid at the C-terminus. This encourages more hydrophobic peptides, where even short peptides of two or three amino acids in length are retained well and do not compromise the peak shape, efficiency, and separation selectivity on reversed-phase columns. The peptides released from digestion with pepsin shown in Figure 2 are more hydrophobic than most of the peptides observed in tryptic digests based on GRAVY scores of the identified peptides. The differences in the cleavage specificity of trypsin and pepsin required slight adjustments to the gradient elution conditions commonly used for tryptic digests of proteins as a significant number of pepsin peptides can retain strongly and elute later. Peptic digestion produces smaller peptides that elute more efficiently over a wider hydrophobicity range from a C18 column, as shown in Figure 3A,B. Additionally, due to pepsin’s lower specificity, a larger number of peptides are generated overall and with a wider range of detectable hydrophobicities (Figure 3D). Linear elution gradients optimized for the pepsin digests run up to 60% acetonitrile instead of the conventionally used 45% for tryptic digests to enable the more hydrophobic peptic peptides be eluted from the column. In addition, most of the pepsin-produced peptides are detected as singly- and doubly-charged ion species (Figure 3C), as opposed to tryptic peptides, which are generally larger and have a higher charge state distribution (Figure 3C), using ESI-MS detection. For example, the small peptide FMVPQYGY is predominantly singly charged yet still produces rich sequence fragmentation information and high signal intensity, as shown in Figure 2B.

Figure 3.

Peptic digestions produce a greater number of peptides with (a) a wider range of hydrophobicity scores, (b) overall smaller molecular masses, and (c) overall lower charge state distribution. (d) The experimentally determined C-terminal cleavage specificity.

CONCLUSIONS

In this study, we showed that digestion of the capsid proteins of AAV2 with heat-stable immobilized proteases is fast, simple, and easy-to-use. This proteolytic method makes use of high temperature and low pH-induced AAV capsid denaturation combined with a heat stable enzyme to increase sequence coverage of viral capsid proteins by peptide mapping. This protocol dramatically reduces the number of steps required for proteolytic digestion. The magnetic beads with immobilized proteolytic enzymes will also potentially facilitate straightforward automation to improve the ease of use and reproducibility of sample processing. The developed technique demonstrates the simplicity of the protocol with increased sequence coverage of the AAV2 capsid proteins. The preliminary steps of preparing samples for digestion take approximately 10 min and, unlike in-solution digestions, do not require the preparation of fresh buffers and reagents as all buffers are included in the commercially available kit and have a shelf life of 1 year. The optimized pepsin digestion requires another 30 min. Thus, the whole peptide mapping experiment, including digestion and LC-MS analysis, can be completed in approximately 90 min. It is a major improvement over the conventional in-solution digestion protocols that can take several hours. The developed application of pepsin provided 100% sequence coverage of AAV2 capsid proteins and demonstrated the potential for PTM analysis and more in-depth structural and quantitative characterization of AAVs than the conventional trypsin-based approaches showed. The suboptimal performance of peptide mapping using trypsin digestion is mostly caused by the stability of the AAV capsids and the primary structure of the AAV capsid proteins being unfavorable for trypsin digestion. The low pH conditions required to activate pepsin helped drive the disruption of the capsid structure integrity and protein denaturation, thus, decreasing the digestion time and also reduced the level of process-induced protein modifications. The use of pepsin requires only minor changes in the analytical workflow. The elution gradient in LC—MS needs minor adjustments, and the data acquisition and analysis require the inclusion of singly charged peptides. There are significantly more peptides produced with pepsin, but the current approaches in data processing, such as the one used in this study, are capable of efficient handling of the resulting data. In this study, we took advantage of nanoflow LC—MS to increase the sensitivity in the characterization of the low-level sample amounts that are usually available in the characterization of AAV samples. Using the sensitive nanoflow LC—MS system, we readily achieved 100% sequence coverage of the VP1 component, which is only present at a tenth of the concentration of VP3 in AAV2 intact viral capsid particles. Minor optimization in sample amounts and buffer dilutions in the digestion will increase the final peptide concentrations that would help transfer this method to the more conventional 1 or 2 mm i.d. LC columns. This will allow the use of the higher flow rate UHPLC systems for this analysis. Such optimization of the scaled-up analyses may involve increased sample amounts to offset the reduced sensitivity of the higher flow rate LC—MS systems. Furthermore, in our hands, it has been proven difficult to concentrate AAV samples with conventional techniques. In this study, we developed the sample preparation workflow based on the application of an excess of the immobilized pepsin, which did not require any sample manipulation prior to digestion. This sample preparation approach followed by nanoflow LC-MS analysis resulted in structurally more informative and quantitatively more accurate peptide mapping data than the conventional trypsin-based techniques. It remains to be determined if the additional sample concentration steps would be required to allow the use of standard flow HPLC systems. The technique reported here may help accelerate progress toward the development of AAV-based vaccines, viral vectors, and other new modality therapies as well as in the characterization of membrane, viral, and other difficult-to-digest proteins.