Case Study: 

The KBI Biopharma Approach

KBI Biopharma developed a top-down, stepwise LC-MS strategy beginning with intact protein analysis to confirm the signal, followed by subunit characterization to localize and isolate the modification, and concluding with de novo peptide sequencing to determine the identity of the unknown species.

• Step 1: Confirm the Signal Is Real

  Before proceeding with a customized characterization strategy, the existing data were validated. While SDS-PAGE had flagged the variant (Fig. 1), SDS-PAGE is susceptible to imaging artifacts and lacks the molecular resolution needed for definitive characterization. Intact mass analysis using TOF mass spectrometry confirmed the 2,518-Da mass shift between the intended species and the variant, establishing a reproducible baseline for the investigation.

  At this stage, multiple explanations remained viable. The mass shift did not align cleanly with N-glycan species typically observed in CHO-derived products, though a sequence variant combined with glycosylation could not be ruled out. In such a scenario, no observable shift would be expected using cIEF or SEC, which is consistent with the negative results from those assays. O-glycosylation was considered but deemed less likely given the magnitude of the shift. Incomplete signal peptide cleavage was also evaluated: signal peptide sequences provided by the customer were screened, and while some masses were similar, none aligned precisely with the observed species. A frameshift mutation was considered as well, though mass data alone was insufficient to confirm or exclude it.

• Step 2: Localize and Isolate the Modification

  Disulfide reduction was applied to determine whether the modification resided on the light chain or heavy chain. As shown in Figure 2, the light chain showed no mass shift, while the heavy chain showed the full 2,520-Da shift following reduction, indicating that the unknown species contained cysteine residues engaged in disulfide bond formation. This behavior is inconsistent with glycan modification and is more consistent with an additional peptide sequence.

  Subunit analysis was then performed using IdeS proteolysis followed by reverse phase HPLC. IdeS cleaves the heavy chain below the hinge into Fc and above the hinge into Fd, allowing variant species to be resolved chromatographically. As shown in Figure 3, a new species, designated FcΔ, separated from the standard Fc fraction. Deconvoluted mass spectra of the FcΔ peak revealed two discrete species: one with the expected 2,520-Da shift and a second at 2,448 Da. The 71-Da mass difference between them corresponds to the residue mass of alanine, consistent with a short peptide extension.

  This analysis localized the modification to the Fc region, enabling isolation of the material at tens to hundreds of micrograms. Following isolation, both the Fc and FcΔ fractions were re-analyzed using RP-HPLC to confirm enrichment. The FcΔ fraction was highly enriched in the variant species, making peptide mapping analysis feasible. Prior peptide mapping attempts on bulk, unenriched material had not detected the variant, as its relative abundance was insufficient for detection under those conditions.

  

  Fig. 2 — Disulfide reduction localizes the mass shift to the heavy chain, consistent with an additional peptide containing cysteine residues involved in disulfide bond formation.

  

  Fig. 3 — IdeS subunit analysis resolves the FcΔ species chromatographically from the standard Fc fraction, enabling isolation of material at tens to hundreds of micrograms for downstream characterization.)

• Step 3: Sequence the Unknown Species

  Peptide mapping was performed on the isolated FcΔ fraction using sequential enzymatic digestion with trypsin and Asp-N. Each digest was analyzed by reverse phase HPLC on a C18 column and characterized by high-resolution Orbitrap LC-MS. Because the unknown sequence was absent from the reference database, de novo sequencing was applied to derive a partial peptide sequence directly from the fragmentation spectra.

  A UniProt search against the derived partial sequence returned no hits, excluding a contaminating protein or naturally occurring peptide variant and pointing toward either a frameshift mutation or an aberrant splicing event, both of which would generate non-canonical sequences absent from standard reference databases.

• Step 4: Identify the Root Cause

  A targeted literature review on C-terminal extensions in recombinant antibodies identified published reports describing cases in which splice donor-acceptor sites at the C-terminus of an antibody construct can recruit noncoding downstream promoter sequences, which are then incorporated into the expressed protein.

  This mechanism was consistent with the analytical data. In collaboration with the customer, two candidate splicing variants were identified, designated Q35 and Q36 in Figure 4, and differentiated by the presence or absence of a single alanine residue. This distinction corresponded directly to the 71-Da mass difference identified in Step 2. Reprocessing the original tryptic digest data with the confirmed sequence verified the full C-terminal extension, confirming that the variant originated from an aberrant splicing event introduced during cell line development.

  The extension included several hydrophobic residues containing cysteines, consistent with the disulfide behavior observed in Figure 2, and three histidine residues near the C-terminus. While the molecule's overall isoelectric point was unchanged under standard conditions, those histidines confer additional positive charge at low and mildly acidic pH, making the variant amenable to separation by cation exchange chromatography. A quantitative cation exchange method was subsequently developed and implemented in the customer's QC laboratory as an ongoing control strategy.

  

  Fig. 4 — Aberrant splicing at the heavy chain C-terminus incorporates downstream promoter sequence into the expressed protein, generating two variant forms, designated Q35 and Q36, differentiated by a single alanine residue.