Hydrolysis of Flexible Linkers
Digestion of Fusion Proteins with Flexible Linkers
Fusion proteins are created by combining the functionalities of two or more proteins or peptides into a single, tailor-made, molecule to specifically address a therapeutic challenge. The most common way of linking such protein or peptide domains is by inclusion of a flexible linker. These linkers are commonly glycine-rich, for example solely glycine residues (G), or glycine residues interspersed with serine residues (GS) to form repeating sequences. This provides the linker with enough flexibility as not to interfere with the function of the connected protein domains. However, these new modalities pose new analytical challenges and require novel tools to facilitate characterization and product quality monitoring.
Middle-level analysis has become a standard approach for the analysis of monoclonal antibody therapeutics and involves the molecule being digested into a few defined subunits. These are small enough to acquire high quality mass spectra and provide domain-specific information for product quality attributes. Since G and GS linkers are resistant to degradation by common proteases, the separation of the domains for in-depth analytical purposes is difficult. Therefore we developed GlySERIAS, an enzyme that digests flexible glycine-rich linkers. GlySERIAS allows for separation of the different functional domains and enables middle-level approaches for detailed characterization of fusion proteins, multi-specific antibodies and various mAb fragments containing glycine-rich linkers.
Reproducible Digestion of a GS-linked Fusion Protein
The fusion protein dulaglutide consists of two glucagon-like peptide-1 (GLP-1) molecules linked to the Fc region of human IgG4 via flexible GS linkers. To study the peptides and Fc region separately, and thereby identify domain-specific PTMs, dulaglutide was digested with GlySERIAS Lyophilized at 37°C for 1 hour. To reduce the sample complexity, the Fc glycans were removed using the endoglycosidase GlycINATOR Lyophilized and the interchain disulfide bridges were reduced with DTT (Fig. 1a). Analysis of the sample by reversed-phase LC-MS showed that the peptides were completely removed from the Fc region upon linker digestion using GlySERIAS. The multitude of glycine residues in the linker offers many different potential digestion sites for GlySERIAS, which is why both the Fc/2 and the GLP-1 peptide were detected as several variants with different numbers of glycine and serine residues still attached (Fig. 1b, c). In addition, an oxidation of the GLP-1 peptide was identified. Triplicate digests showed reproducible results in the relative amount of the different Fc/2 variants obtained (Fig. 1d), despite GlySERIAS digesting at several sites simultaneously (Fig. 1e).
- Highly reproducible digestion pattern of flexible linkers using GlySERIAS
Figure 1. Replicate digestion of dulaglutide using GlySERIAS Lyophilized. The flexible GS linkers of dulaglutide were digested with GlySERIAS Lyophilized for 1 hour at 37°C under non-denaturing conditions, and the Fc glycans were concurrently hydrolyzed using GlycINATOR Lyophilized to reduce sample complexity. To stop the GlySERIAS reaction, 1 mM ZnCl2 was added. The interchain disulfide bonds were reduced with 20 mM DTT for 30 minutes at 37°C. The digestion was performed in triplicate. a) Illustration of the sample preparation workflow. The samples were analyzed by reversed-phase LC-MS. b) Deconvoluted mass spectra of the Fc/2 subunit. c) Deconvoluted mass spectra of the GLP-1 peptide. d) Relative amount of the identified Fc/2 variants, displaying the mean value between the triplicate digests and error bars representing the standard deviation. The digestion products were separated by reversed-phase chromatography (BioResolve™ RP mAb Polyphenyl, 450 Å, 2.7 µm 2.1 x 100 mm, Waters™) and analyzed with ESI-QTOF MS (Bruker Impact II). e) Schematic image of the flexible linker, connecting the GLP-1 peptide to the Fc/2 subunit, and the identified digestion sites.
In-depth Analysis of Fusion Proteins with Flexible Linkers
Detailed analysis and identification of quality attributes are of high importance during quality control of protein therapeutics. For fusion protein therapeutics with flexible linkers, this includes confirming the quality of the linked proteins as well as that of the protein linkers. Linker digestion using GlySERIAS may aid in this quality control by separating the linked protein domains. However, the multitude of glycine residues in the linker offers many potential digestion sites of the enzyme, leading to simultaneous hydrolysis at several sites within the linker. The digested product often consists of several variants of the linked proteins with different degrees of the linker remaining. While the observed heterogeneity generally does not negatively impact detailed characterization of the protein subunits, but rather facilitate direct characterization of the linker, a more homogenous digestion product is sometimes required for detailed analysis of the individual protein domains.
To obtain a more homogenous digestion product of the dulaglutide fusion protein (described in Fig. 1 above), the protein was digested with GlySERIAS Immobilized, consisting of the GlySERIAS enzyme covalently coupled to agarose beads. By coupling the enzyme to a resin, the enzyme-to-protein ratio can be increased, and thereby the reaction can be accelerated further to a more complete linker digestion. In addition, the enzyme will not be present in the final sample preparation, potentially interfering with sample analysis. For comparison, dulaglutide was digested with GlySERIAS Immobilized overnight at 37°C, or with GlySERIAS Lyophilized for 1 hour and overnight at 37°C. To reduce the sample complexity, the Fc glycans were removed using the endoglycosidase GlycINATOR Lyophilized, and the interchain disulfide bridges were reduced (Fig. 2a). Analysis of the samples by reversed-phase LC-MS showed that GlySERIAS Immobilized digested the linkers almost completely from the Fc domain, leaving a close to homogenous Fc/2 product containing one serine residue from the linker (Fig. 2b). Digestion with the GlySERIAS Lyophilized for 1 hour or overnight both resulted in several Fc products with different numbers of serine and glycine residues attached. The GLP-1 peptide was present as two variants in all three samples (Fig. 2c). The homogenous Fc/2 from digestion with GlySERIAS Immobilized enabled the detailed study of quality attributes located specifically to the Fc region. Additionally, no enzyme peaks interfering with the analysis were observed in the sample. As a complement, the Fc/2 products with part of the linker remaining from digestion with GlySERIAS Lyophilized, facilitated the study of modifications located to the GS linker.
- GlySERIAS Immobilized improves linker removal
- GlySERIAS Lyophilized enables linker characterization
Figure 2. Digestion of dulaglutide using the two formats of GlySERIAS. The flexible GS linkers of dulaglutide were digested with GlySERIAS Lyophilized for 1 hour and overnight, or with GlySERIAS Immobilized overnight at 37°C under non-denaturing conditions. This was followed by hydrolysis of the Fc glycans using GlycINATOR Lyophilized to reduce the sample complexity, and denaturation and reduction of the interchain disulfide bonds with 4 M guanidine hydrochloride and 100 mM DTT for 1 hour at 37°C. a) Illustration of the sample preparation workflow. The samples were analyzed by reversed-phase LC-MS. b) Deconvoluted mass spectra of the Fc/2 fragment. c) Deconvoluted mass spectra of the GLP-1 peptide. The digestion products were separated by reversed-phase chromatography (ACQUITY Premier Protein BEH C4, 300 Å, 1.7 µm 2.1 x 100 mm, Waters™) and analyzed with ESI-QTOF MS (Bruker Impact II). d) Schematic image of the flexible linker, connecting the GLP-1 peptide to the Fc/2 subunit.
Romiplostim consists of four identical thrombopoietin (TPO) receptor binding peptides and one Fc fragment, linked together by flexible polyglycine sequences. The protein was, similarly to dulaglutide, digested with GlySERIAS Immobilized overnight at 37°C, or with GlySERIAS Lyophilized for 1 hour and overnight at 37°C. The interchain disulfide bridges were reduced to facilitate the following analysis by reversed-phase LC-MS (Fig. 3a). Similar to what was observed with dulaglutide, digestion with the immobilized enzyme yielded a homogenous Fc/2 with solely one linker residue remaining, here, a glycine residue (Fig. 3b). This facilitated the identification of modifications located specifically in the Fc region. Digestion with the lyophilized enzyme for 1 hour yielded Fc/2 with four remaining glycine residues and a smaller amount of Fc/2 still linked to one TPO peptide. This, on the other hand, enabled the study of the linker region between the Fc/2 and the first peptide without the interference from the second peptide. Overnight digestion with GlySERIAS Lyophilized yielded two Fc/2 variants with four and one glycine residue remaining, respectively. The peptides released using the lyophilized product were present as five or seven variants, depending on the digestion time, with different numbers of glycine residues remaining from the linker, while the peptides from using the immobilized product were present as four variants (Fig. 3c).
Figure 3. Digestion of romiplostim using the two formats of GlySERIAS. The polyglycine linkers of romiplostim were digested with GlySERIAS Lyophilized for 1 hour and overnight, or with GlySERIAS Immobilized overnight at 37°C under non-denaturing conditions. This was followed by denaturation and reduction of the interchain disulfide bonds with 4 M guanidine hydrochloride and 100 mM DTT for 1 hour at 37°C. a) Illustration of the sample preparation workflow. The samples were analyzed by reversed-phase LC-MS. b) Deconvoluted mass spectra of the Fc/2 subunit. c) Deconvoluted mass spectra of the TPO receptor binding peptide. The glycine residues can be present at both ends of the peptide. The digestion products were separated by reversed-phase chromatography (ACQUITY Premier Protein BEH C4, 300 Å, 1.7µm 2.1 x 100 mm, Waters™) and analyzed with ESI-QTOF MS (Bruker Impact II).
Middle-level Analysis of a BiTE Molecule
Bispecific T-cell Engager (BiTE) molecules are a type of fusion protein therapeutics where two scFvs are fused, forming a bispecifc molecule with one scFv targeting cytotoxic T-cells, and the other one targeting a cancer antigen. The protein consists of four protein domains, the two scFvs both being made up of a VL and a VH region, joined by a total of three flexible GS linkers. The large size of this protein, 54 kDa, may pose a challenge for product quality monitoring by intact mass spectrometry since standard MS instrumentation may not be able to provide monoisotopic resolution of the protein.
A research grade biosimilar of blinatumomab, a BiTE molecule with specificity towards CD19 and the T-cell marker CD3, was digested with GlySERIAS Immobilized overnight at room temperature (Fig. 4a). Direct analysis of the digested sample by LC-MS showed that all three linker regions were digested, with the α-CD3 VL and VH chains eluting as one chromatographic peak, and the α-CD19 VH and VL chains eluting as separate peaks (Fig. 4b). A small peak from the linked α-CD3 VH and α-CD19 VH domains revealed that this short linker was not completely digested, indicating that GlySERIAS has a lower activity on short linkers. The separation of the four domains provided high quality spectra with monoisotopic resolution, as compared to the intact protein analysis (Fig. 4c, d). A few different variants of each domain could be observed, with different numbers of linker residues remaining. The α-CD3 VH domain could not be fully baseline resolved from the α-CD19 VH domain during chromatographic separation, resulting in some co-elution which could be observed in the associated mass spectra. The separation of the four domains yielded information about the location of protein modifications identified on the intact protein level. For example, a protein modification of 37 Da, which was also identified on the intact level, was assigned to the α-CD3 VH chain. Furthermore, two modifications of 240 Da and 320 Da, respectively, were assigned to the α-CD3 VL domain. The latter also contained a His-tag with five or six histidine residues. These data demonstrate the power of middle-level workflows using GlySERIAS Immobilized for quality control of fusion proteins containing several flexible linkers.
- Hydrolysis of the flexible linkers in multi-subunit fusion proteins provides domain-specific information, thereby highly facilitating protein quality monitoring
Figure 4. Middle-level analysis of BiTE using GlySERIAS Immobilized. The GS linkers of a biosimilar of blinatumomab, a BiTE molecule with specificity towards CD19 and CD3, were digested with GlySERIAS Immobilized overnight at room temperature under non-denaturing conditions. a) Illustration of the sample preparation workflow. The samples were analyzed by reversed-phase LC-MS. b) Chromatographic separation of the digest, total ion chromatogram. c) Deconvoluted mass spectrum of the intact BiTE. d) Deconvoluted mass spectra of the four protein domains separated after GlySERIAS digestion. The glycine and serine residues can be present at both ends of the heavy chains. The processed samples were separated by reversed-phase chromatography (ACQUITY Premier Protein BEH C4, 300 Å, 1.7 µm 2.1 x 100 mm, Waters™) and analyzed with ESI-QTOF MS (Bruker Impact II).
