Company

Isoleucine and Leucine determination in de novo antibody sequencing

digital proteomics
July 1, 2019

Isoleucine and leucine determination

De novo antibody sequencing by tandem mass spectrometry(MS/MS) uses the differences in mass of different amino acids to accurately determine the correct sequence. Distinguishing between isoleucine (Ile) and leucine (Leu) residues is troublesome in MS/MS-based de novo sequencing as the side chains are isomers and have the same mass, 113.08406 Da. Two peptides differing in amino acid composition by a substitution of Ile with Leu have identical masses and commonly used fragmentation modes, such as CID or HCD yield the same peptide fragment masses.

Isomeric leucine and isoleucine structures
Due to the lack of discriminatory signal, many de novo sequencing tools report isoleucines or leucines in a peptide as an ambiguous Xle residue. However, Ile and Leu in antibodies can be distinguished using alternate proteomics strategies, specifically:

  1. Enzyme-cleavage specificity to leucine
  2. Diagnostic w-ions in MS2 and MS3 spectra
  3. V/J gene homology

Isoleucine and leucine determination methods

Enzyme-cleavage specificity

MS/MS proteomics workflows typically use enzymes to cleave proteins into smaller peptides since MS/MS instruments produce more interpretable fragmentation data from peptides than proteins. Each digestion enzyme has a propensity to cleave at specific residues.

As shown in the figure to the right, chymotrypsin tends to cleave after tyrosine (Tyr), tryptophan (Trp), leucine (Leu), and phenylalanine (Phe), while pepsin is less specific and cleaves before or after glutamic acid (Glu), tyrosine (Tyr), leucine (Leu), phenylalanine (Phe), and alanine (Ala). Both enzymes are more prone to cleave around leucine than isoleucine.

Cleavage patterns of chymotrypsin and pepsin

EThcD fragmentation of Ile and Leu side chains

While fragmentation typically occurs along the peptide backbone, fragmentation of residue side chains is also possible. Ile and Leu residues share the same elemental composition, but have different bonding patterns (see figure at top of post).

Side-chain fragmentation can be produced using ETD fragmentation for MS2 fragmentation of peptides and HCD fragmentation to generate MS3 for targeted z-ions. The z-ions where the unknown Xle is at N − Cα bond breakage results in Ile and Leu characteristic w-ions. As shown in the spectrum to the right, Leu side chain fragmentation results in an ejection of an isopropyl group (loss of C3H7 from the z-ion), whereas the Ile fragmentation results in an ejection of an ethyl group (loss of C2H5 from the z-ion).

W ion losses in EThcD spectra

V/J gene homology

A third method for determining Ile and Leu is to use antibody sequence homology. Antibody sequences can be traced back to their germline gene origins and predicted Ile and Leu can be confirmed by checking the encoding codon for antibody labeled V and J gene segments. Of note, antibodies contain complementarity-determining regions (CDRs) that are highly prone to mutation. Ile and Leu validation by checking the translated nucleotides from genes for CDR1 and CDR2 are suspect. CDR3’s of antibodies are uniquely derived with no corresponding gene segments to validate against.

Case study in de novo antibody sequencing

To illustrate Confirmation of Leu and Ile Presence (CLIPTM), MS/MS data was generated from two monoclonal antibodies, anti-ERBB2 (Absolute Antibody Ab00103) and anti-β Gal (Digital Proteomics) antibody. Each antibody heavy and light chain was digested by four enzymes: chymotrypsin, trypsin, pepsin, and elastase. Digested peptides were analyzed by nano-LC MS/MS in doublet HCD and EThcD mode. Unlike, previous methods that used MS3 fragmentation, both the HCD and EThcD spectra were at the MS2 level. De novo protein sequencing was performed using Digital Proteomics’ ValensTM algorithm, and CLIPTM was used to call Ile and Leu residues.

CLIPTM applies Bayesian inference that incorporates enzyme specificity evidence and characteristic z- and w-ion evidence to produce a log-likelihood score for an unknown Xle site being a leucine or isoleucine. Of the 62 sites on the anti-ERBB2 antibody, the accuracy of correctly calling isoleucine or leucine at a log-likelihood threshold of 1.0 was 89%. Only 9 Xle sites did not have a log-likelihood score surpassing the threshold. The anti-β Gal antibody had 67 sites, with similar 83% accuracy and 88% call rate at the same threshold. Prediction errors were typically the result of Ile and Leu appearing close together in sequence. Peptides containing both Ile and Leu can produce decoy w-ions through radical site migration and cause erroneous calls (Zhokhov et al. 2017).

This case study demonstrates that CLIPTM can confidently distinguish the majority of Ile and Leu residues. When ambiguity persists at the MS2 level, a targeted strategy employing MS3 can be used to provide additional support.

Resources

A recent blog post gives an overview of the de novo antibody sequencing process.

You can read more about our ValensTM service. We’ve been de novo antibody sequencing for over a decade!

Papers that were helpful in the development of CLIPTM include:

Distinguishing between Leucine and Isoleucine by Integrated LC-MS Analysis Using an Orbitrap Fusion Mass Spectrometer

An EThcD-Based Method for Discrimination ofLeucine and Isoleucine Residues in Tryptic Peptides