Summary

This article outlines 

• This application note describes the comparison of six strategies for label-free peptide quantitation, based on protease and charge state selection.
• The preferred strategy was implemented in Protein Metrics’ Byos® software for two scenarios: Sequence
  Variant Analysis and PTM Quantitation.
• Byos peptide workflows include report templates for relative quantification, which can be designed around specific requirements.
• Reports include a “dynamic columns” feature which allows the creation of custom columns and calculations.

Introduction

The accurate quantitation of post-translational modifications (PTMs) in therapeutic proteins is a significant challenge in mass spectrometry (MS) analysis. Various factors, such as the presence of multiple charge states, missed cleavages and low signal intensities of modified species complicate the reliable measurement of PTMs.
A multitude of variables exist. The question is, which ones are best predicators of real-world efficacy? For example, some groups perform quantitation based on the most intense charge state of a single species, ignoring whether the selected peptide contains missed cleavages.
Label-free quantitation in MS analysis provides relative measurements, rather than absolute quantification. As such, it is only suitable for trend analysis. Despite this limitation, many research groups define a quantitation strategy and adhere to it throughout the life cycle of a potential new drug candidate to maintain consistency. Stakeholders still require accurate (relative) measurements, as they play a crucial role in assessing the risk associated with PTMs and their potential impact on the success of drug development projects.

This application note summarizes the work presented by Roche (see this link), which compares the accuracy of different strategies for PTM analysis. We also demonstrate how Protein Metrics software can be used to implement their recommended strategy by creating a customized report template, utilizing the dynamic columns feature.

Experimental

Assessment of Quantitation Strategies for Sequence Variant Analysis in Therapeutic Proteins 

In this study, six different quantitation strategies were evaluated for sequence variant analysis. Due to a disadvantageous distribution of Trypsin cleavage sites, nonspecific enzymes were employed to achieve complete sequence coverage. Nine different enzymes were used in combination with 18 different molecules, giving a total of 162 samples. Each sample was spiked with a molecule of similar sequence (differing by a single amino acid) to simulate sequence variants at a known amount of 1%. μ-Flow-LC-MSMS was used to acquire the data (Orbitrap Fusion or Q-Exactive), which was subsequently processed in Protein Metrics Byos® (vendor neutral software), using the PTM workflow. The table below gives an overview of the assessed quantitation strategies.

Table 1. Overview of the different quantitation strategies..

All strategies are described in detail in the presentation recording (see the link). Here we focus on two strategies to enhance clarity. Strategy 4 selects only wildtype peptides with an absolute XIC area greater than 1e7 (instrument specific). All charge states having a wildtype with a corresponding modified species are chosen, regardless of their digestion origin. The XIC ratios (XIC AUC[modified]/XIC AUC[wildtype]) are averaged and the resulting value serves as the quantitation estimate of the sequence variant's abundance.

Figure 1. Selection of peptides for quantitation strategy number 4 (Mean1e7). Quantitation is based on all enzymes and calculated by averaging the XIC ratios.

Quantification strategy number 6 on the other hand considers all wildtypes at all charge states with corresponding modified species, irrespective of their digestion origin. The XIC AUC intensities of the modified species are summed and divided by the total XIC AUC intensities of the wildtypes, yielding an estimate that assigns higher weight to higher intensity values.

Figure 2. Selection of peptides for quantitation strategy number 6 (Weighted All Enzymes). Quantitation is based on all enzymes and calculated by a weighted average.

Results and Discussion

Results of the Assessment

The table below presents the outcomes of the six quantitation strategies based on four factors. The first factor shows the percentage of captured sequence variants, with approaches solely considering trypsin exhibiting lower values due to the challenging tryptic cleavage site distribution. Expanding to other specific enzymes, as in Strategy 5, improves the capture but still falls short in capturing all variants, highlighting the need for unspecific digests alongside standard sample preparation.

Table 2. Overview of the results of the assessment. SVs (%): Percentage of identified sequence variants.5-fold lover and upper quant: Number of sequence variants that either have been 5 times over- or underestimated. Mean Deviation to 1%: Mean deviation to known sequence variant concentration.

The next two factors reveal the frequency of fivefold over- or underestimation of sequence variant abundance. Underestimation is particularly critical as it may mask crucial attributes as non-relevant. Lastly, the mean deviation from the known abundance of 1% is shown.

Conclusion of the Assessment

Based on the results, two strategies stand out as superior. Strategies 5 and 6 both resulted in low deviation, with minimal overestimated values. While Strategy 5 captures all sequence variants, Strategy 6 shows a higher number of overestimated values. The study concludes that a hybrid approach, incorporating elements from both Strategy 5 and 6, is most effective for accurate quantitation. For digestion, the recommended strategy is to begin with specific enzymes and fill potential sequence gaps with non-specific digests.

Implementation in Byos

The rationale behind the selection of sequence variants for assessing label-free quantitation strategies was driven by the fact that amino acid substitutions are not impacted by the sample preparation as oxidation and deamidation are. Nevertheless, the suggested quantitation strategy is not limited to sequence variants and smoothly transitions to the all peptide workflows in Byos (PTM, disulfide bond,…). It is also not a requirement to run the analysis with multiple enzymes.
Since the quantitation strategy is just a way to group the peptides and normalize their quantitation values across the group, only the Byos report needed to be customized for this application. Byos provides report templates for the most common applications that can be customized, or new templates can be created from scratch. With the functionality to add dynamic columns that are coded in JavaScript the versatility of the report is tremendously high. For this application a dynamic column was created to group all peptides containing the same modified residue independent of the utilized enzyme, modification status, charge state or the number of missed cleavages. Those peptides that can be assigned to multiple groups are duplicated. By assigning all those peptides with the common residue to the same category, each of the variants can be quantified within the group. 

General Instructions

Following project creation and inspection, a custom tab can be added to an existing report (File --> Load From File, available upon request at support@proteinmetrics.com). This tab incorporates two new columns with names Quantitation Key^T and Quantitation Key Type^T. The superscripted ‘T’ indicates that these columns are only available in this tab.

Figure 3. Example of a quantitation key with its assigned peptides.

The Quantitation Key Type^T field contains values that in the following text are referenced to as keys. These keys are generated in three steps. In the first step, all modified residues are listed and a temporary key is created based on the protein name, the residue character and the protein position. Each peptide that fulfills the criteria of originating from the same protein and containing the residue in the key, gets assigned to it (multiple assignments are possible). After the second step each key has a peptide list and those keys with an identical peptide list get merged in the third step into the final key with following pattern:

Figure 4. Pattern of the quantitation key.

The list of peptides for each key may need some further curation, since it may contain wildtype peptides with no matching modified species, or the charge states are not the same. The second column “Quantitation Key Type^T” categorizes the peptides for each key in four different groups:

Figure 5. Overview of the different Quantitation Key Types.

By filtering those wildtypes and modified species with no matching species, each key will only contain a list of peptides with matching wildtypes or modified species.

Case Scenario 1: Sequence variants analysis with multiple enzymes

The NISTmab was digested with four specific enzymes (Trypsin, Chymotrypsin, AspN, LysC+GluC) and five non-specific enzymes (Pronase, ThermoLysin, Elastase, Pepsin and ProAlanase). After acquisition, the resulting data set was split based on the specificity of the utilized enzyme in two data sets that ran with the SVA workflow. For this note, only one sequence variant is shown that already has been identified by two independent laboratories. ¹

Figure 6. Example report for sequence variant analysis.

The data values in columns 1-5 are summed normalized Total XIC AUC Averagine values (Level 1). One special feature in Byos is to add column totals, giving a weighted average based on user-defined levels:

Figure 7. Current tab settings after enabling column totals.

Case Scenario 2: PTM analysis with Trypsin

An IgG molecule was subjected to stressed conditions for seven days, before being digested with trypsin and ran on a Thermo LTQ Orbitrap XL. Raw data files were processed using the Byos PTM workflow. The default PTM report was amended to include the additional tab as described above. Additionally, the subtotals of each sample can be added with these steps: File --> Presets --> Post pivot columns --> TotalGroup.postpvtjs. This will granulate the “Totals” column into subtotal columns for each sample.

Figure 8. Example report for PTM analysis.

Conclusions

In conclusion, Protein Metrics Byos® (vendor neutral software) stands as a robust and versatile solution for protein characterization, in part due to its extensive reporting capabilities. The platform offers users a wide array of tools, including flexible reporting templates and dynamic columns, which greatly enhance the flexibility and depth of data analysis.
The PTM and SVA workflows within the software incorporate a default report template, comprising multiple tabs that offer critical insights such as peptide coverage information and label-free quantitation. Additionally, the dynamic column feature empowers users to tailor their data analysis by creating custom columns and performing specific calculations.
This application note has effectively summarized the outcomes of Roche's assessment of different quantitation strategies and how a hybrid of two of the assessed methods outperforms the others. The integration into Byos demonstrated the practical application of dynamic columns highlighting their values in addressing specific analytical challenges.
For researchers seeking further information on the range of report templates and dynamic columns available, including the template discussed in this application note, we encourage contacting support@proteinmetrics.com.
 

References

[1] Zhang, A.; Chen, Z.; Li, M.; Qiu, H.; Lawrence, S.; Bak, H.; Li, N. A General Evidence-Based Sequence Variant Control Limit for Recombinant Therapeutic Protein Development. mAbs 2020, 12 (1), 1791399. https://doi.org/10.1080/19420862.2020.1791399.