Executive Summary
Deamidation is common in protein therapeutics and is often the focus of stress and forced degradation studies conducted during drug development. Many analysts in the biotherapeutic industry quantitatively compare stress studies for early lead candidate profiling or even to show comparability of lots and batches in later development. Consequently, accurately quantifying these modifications is critical for both internal development and for reporting to regulatory agencies. As shown in Figure 1, asparagine residues can deamidate either by hydrolysis to form aspartic acid or through a cyclic succinimide intermediate which then can form both isomers of aspartic acid.
|
Figure 1 - Deamidation of asparagine (N) to aspartic acid (D) by hydrolysis or through a cyclic, succinimide intermediate to aspartic acid (D) or isoaspartic acid (iso-D). Work by Robinson and Robinson referenced in the text has proposed that deamidation route depends on the neighboring C-terminal residue. |
Because there is only a 0.984 Da shift, deamidation can be difficult to quantitatively measure when there is a moderate or large degree of co-elution with the unmodified Asn form. Here we present a novel method for deconvolving, quantifying, displaying, and reporting deamidation as well as isomerization (Asn to Asp and iso-Asp) using Protein Metrics’ Byos software.
Method
In this study, a series of purified IgG1 antibody samples was exposed to stress conditions designed to induce deamidation (pH = 8.5 for 0, 1, 3, and 7 days). The samples were digested using trypsin and LC-MS/MS datasets were collected using an LTQ-Orbitrap XL mass spectrometer (Thermo Fisher).
A project was created using Byos PTM workflow, which identifies peptides using MS/MS (legacy Byonic search), and quantifies the peptides based on their extracted ion chromatogram (XIC) areas.
Deamidated species often co-elute with their unmodified wildtype peptides. It can be helpful to initially view the extracted ion chromatograms (XICs) to get a high-level sense of potential co-elution.
|
Figure 2 –Total extracted ion chromatogram (XIC) of an …NGQPENNY peptide before [left] and after [right] deconvolution of deamidation within Protein Metrics Byos software. |
In Figure 2, the XIC on the top-left shows the ion current of the monoisotopic peak of deamidated peptides as well as the first 13C isotope of the unmodified peptide. The XIC on the top-right shows the Byos deconvolution of XICs for the deamidated form based only on the non-deamidated XIC (bottom-left and bottom-right) along with the theoretical isotopic distribution of the peptide in question.
In this Application Note, as an example, we report with some detail on the deamidation of one common IgG1 tryptic peptide for a stress test consisting of exposure at alkaline conditions: GFYPSDIAVEWESNGQPENNYK. This peptide has three asparagine residues shown in ORANGE below, each one with a different susceptibility to deamidation under these conditions:
GFYPSDIAVEWESN387GQPEN392N393YK
Deamidated N392 and N393
To examine the deamidation of N392, the user clicks on the GFYPSDIAVEWESNGQPEnNYK peptide in the Peptides table and Byos shows the annotated MS2 spectra of the deamidated form (top, Figure 3) juxtaposed with the corresponding unmodified form of the same peptide (bottom, Figure 3). As shown in the Figure, Byos allows a user to zoom in and with a cursor inspect that the y4-ion has a 0.984 Da shift relative to the unmodified form. The fragmentation indicates that the deamidation occurs on the N392 of the NN pair.
|
Figure 3 - Annotated MS2 spectra of a deamidated N392N moiety and its corresponding unmodified form. As shown in the Figure, Byos provides the b- and y-ion fragmentation annotation and enables a user to zoom in and precisely inspect identified shifts in mass. Here, the y4-ion shows a 0.984 Da shift relative to its unmodified form. |
In addition to the annotated MS2 spectra, Byos also displays the associated XICs for the modified peptide with a red dot indicating the retention time of the selected scan for MS2 display, and pink dots indicating the retention time from all scans that have an exact match in terms of sequence and modification type and location. Gray dotsare associated with other modification forms of the peptide having the same mass (isomers).
|
Figure 4 - Deconvolved XIC of the deamidated form of the N392N moiety. |
The deconvolved XIC for the GFYPSDIAVEWESNGQPEnNYK peptide is shown in Figure 4. Note that the cluster of pink and red dots indicate that the deamidated N392N form occurs in the first trailing peak of the XIC, at a retention time centered at about 92.5 minutes. Adjustable, vertical bars within the XIC window can be used to define the integration limits for quantification.
Deamidation of N392 reveals only one peak and one product, which is consistent with the work of Robinson and Robinson that proposes that asparagine deamidates mainly by hydrolysis forming only aspartic acid unless it is followed by a glycine on the C-terminus side.i,ii Note also that N393 has a large hydrophobic residue on its C-terminus (tyrosine) which generally translates to very slow deamidation; in this study no deamidation on N393 was observed.
Byos also enables an analyst to quantify the extent of the N392 degradation over a series of samples by examining the deconvoluted XICs from the same deamidated peptides across a set of samples within the same project (Figure 5). Byos calculates a field XIC Ratio % as a simple, label-free quantification as a ratio of XICs to the corresponding wildtype form:
An alternate method of quantification that include accounts for multiple forms of deamidation is presented in the attached Appendix. Table 1 summarizes the quantification of the N392 deamidation based on an XIC Ratio % relative to the unmodified form for each day of exposure under alkaline conditions, pH 8.5.
Figure 5 – Series of XICs quantifying N392 deamidation of the …NGQPENNYK peptide for 0, 1,3 and 7 days of stress at pH = 8.5.
Table 1 – Quantification of Deamidated N392 (XIC Ratio) relative to the unmodified form.
|
Day 0 |
Day 1 |
Day 3 |
Day 7 |
|
|
N392 Deamidation |
1.43% |
8.75% |
20.3% |
37.9% |
Deamidated N387G Moiety
To examine the deamidation of N387, a user clicks on the GFYPSDIAVEWESnGQPENNYK peptide in the Peptides table. Byos shows the annotated MS2 identification of the deamidated N387G form, seen here in Figure 6.
|
Figure 6 - Annotated MS2 spectra of a deamidated N387G moiety and its corresponding unmodified form. As shown in the Figure, Byos provides the b- and y-ion fragmentation annotation and enables a user to zoom in and inspect identified shifts in mass. Here, the y9-ion (but not the smaller y-ions) shows a 0.984 shift in m/z relative to its unmodified form. |
As mentioned above, Byos allows a user to zoom in and inspect the data, revealing that the y-9 ion has a 0.984 Da shift relative to the unmodified form.
The associated deconvolved XIC is shown in Figure 8. Note the two clusters of pink and red dots indicating that the N387G form degrades into an early eluting peak (at about 91.10 minutes) and a late eluting peak (at about 93.25 minutes), indicative of two distinct, deamidated forms.
|
Figure 7 - Deconvolved XIC is shown for the N387G moiety showing a bifurcation of elution times. |
The bifurcation shown in Figure 7 means that Byos can distinguish between the aspartic acid and iso-aspartic acid forms shown in Figure 1. To quantify the leading and trailing elution peaks individually, it is convenient to have Byos create a copy of an existing specific peptide as shown in Figure 8 in order to report both quantifications simultaneously. This is accomplished by right clicking on the peptide of interest and selecting Create new peptide from current…. The new copy can be used, for example, to quantify the trailing peak (aspartic acid in this case) and the original can be used to quantify the leading peak (iso-Asp). Quantification of N387G leading and trailing peaks over the time course of the stress study is shown in Table 2.
|
Figure 8 – Creating a copy of an existing peptide. |
Table 2 – Quantification of Deamidated N387G Moiety (XIC Ratio)
|
Day 0 |
Day 1 |
Day 3 |
Day 7 |
|
|
N387G Leading Peak (iso-aspartic acid) |
2.30% |
7.02% |
15.6% |
29.5% |
|
N387G Trailing Peak (aspartic acid) |
0.42% |
1.53% |
3.77% |
7.29% |
Reporting
Table 3 contains the resulting quantification of the various deamidation species of the …NGQPENNYK peptide, with the Label column being used to indicate the form of the asparagine modification.
Table 3 – Byos resulting quantification of the various deamidation species of the …NGQPENNYK peptide
The results can be exported for reporting in several ways. Protein Metrics software includes a powerful reporting feature which outputs a project summary, various types of tabular data, charts, and plot images. These templated options include basic calculations such as normalization, summation, etc. as seen in Figure 9. The automatically formatted report can be exported by clicking on File > Export > Report….
Figure 9 – An automatically formatted report generated by Byos software showing a summary page with sequence coverage of wildtype and modified proteins, two peptide tables with all relevant information about each modified peptide, bar charts, and a plots tab including MS1, MS/MS, and XIC plots.
When calculating XIC Ratio% value of deamidated peptides, Byos compares the monoisotopic ion XICs of both deamidated and corresponding wildtype peptides. The XIC Ratio% column, shown in the inspection view, is calculated per modification and per charge state, as shown in eq 1: XICmod ÷ (XICmod + XICwildtype). The PTM default report’s 2nd tab uses this value, arithmetically averaged across charge states and possibly missed cleavages.
Quantification based on the simple XIC Ratio % is often sufficient and is accurate when the extent of modification is low. When there is significant degradation or modification, as seen in the 7-day samples, it may be desirable to ratio against a sum of all relevant modifications. This will both account for a loss of wildtype peptide and account for the presence of significant amounts of other modified forms. Thus, an alternative calculation would be to use the total area under the curve for each XIC (AUC):
The Byos reports perform this calculation automatically. The report's 4th tab takes into account all other modifications of the same sequence peptides and re-calculates relative abundances via normalization. If one of the modifications is deamidation, the XIC of the deamidated peptide is normalized, so that the isotopes used to generate XICs match between deamidated, wildtype, and other modifications. This table also sums XICs of different charge states before calculating ratios. The calculation can be considered as weighted average. “XIC Area Summed isoX Normalized” is the dynamic column which performs the normalization. It normalizes XIC area summed values for peptides that contain deamidation alongside other modifications, because the XIC area summed may be calculated using different isotopes for the deamidated peptide (monoisotope) and the other modifications (determined by MS extract options table). This column normalizes deamidated XICs by using the formula: XIC Area Wildtype (isotope determined by MS extract options table) * XIC Ratio/ (1- XIC Ratio).
NOTE: When the isotope used for Wildtype XIC Area is > 0 (determined by MS extract options table), it can sometimes be important for the XIC range of the Wildtype peptide to exclude nearby eluting deamidation for the PTM Report to properly calculate percent modification values. Usually this is not a concern, unless all of the following apply (See Figure 10 for an example):
- Deamidation elutes close to the Wildtype.
- The MS extract options table selects an isotope > 0, typically peptides with mass >1800 Da.
- Closely eluting deamidation is abundant enough to significantly skew the Wildtype XIC Area (iso > 0) and subsequently skew PTM Report percent modification calculations
|
Figure 10 – GFYPSDIAVEWESNGQPENNYK after 7 days of stress to induce deamidation. Panes A and B show deconvolved deamidation XIC iso=0 and Wildtype XIC iso=0. The XIC Areas of the top two panes are used to calculate the % Ratio for deamidation signals. Panes C and D show the Wildtype XIC iso=1 (iso determined by MS extract options table). The Wildtype XIC Area iso=1 is used to calculate the % Ratio for all modifications except for deamidation and is used for the PTM 4th report tab percent modification calculation and "XIC Area Summed isoX Normalized" for normalizing deamidation XIC area. The Wildtype XIC Area iso=1 can be skewed too high by the nearby eluting deamidation signals at 90.9 and 92.6 minutes. In Pane C the Wildtype XIC boundaries exclude the nearby eluting deamidation peaks. In Pane D the Wildtype XIC boundaries include nearby eluting deamidation, which can skew the PTM report percent modification calculations. |
Conclusions
The deamidation of the …NGQPENNYK peptide based on XIC Ratio % relative to the unmodified form can be summarized in reporting by the chart in Figure 11. Deamidation of the N392 asparagine into aspartic acid and the N387 asparagine into both aspartic and iso-aspartic acid was observed and quantified. No deamidation of the N393 asparagine was observed. In this study, Byos was able to quantify partially co-eluting deamidation modifications as low as a fraction of one percent relative to the unmodified form.
|
Figure 11 – Summary of the deamidation of the …NGQPENNYK peptide at pH=8.5 over 7 days. |
Byos offers a unique software platform for the rapid and confident analysis of all types of PTMs including co-eluting deamidated species. Analysis of tandem mass spec data from a large number of data sets can be performed with high confidence and in a fraction of the time it would take to manually process the data. Byos also provides a unique means for all analysts to produce comprehensive, consistent reports.