B American Society for Mass Spectrometry, 2019 J. Am. Soc. Mass Spectrom.

(2019)
DOI: 10.1007/s13361-019-02296-2

RESEARCH ARTICLE

A Case Study to Identify the Drug Conjugation Site

of a Site-Specific Antibody-Drug-Conjugate Using
Middle-Down Mass Spectrometry
Oscar Hernandez-Alba,1 Stéphane Houel,2 Steve Hessmann,1 Stéphane Erb,1
David Rabuka,3 Romain Huguet,2 Jonathan Josephs,2 Alain Beck,4 Penelope M. Drake,3
Sarah Cianférani1
1
Laboratoire de Spectrométrie de Masse BioOrganique, CNRS IPHC UMR 7178, Université de Strasbourg, ECPM R5-0 – 25 Rue
Becquerel, Cedex 2, 67087, Strasbourg, France
2
Thermo Fisher Scientific, 355 River Oaks Pkwy, San Jose, CA 95134, USA
3
Catalent Biologics West, 5703 Hollis Street, Emeryville, CA, 94530, USA
4
IRPF, Centre d’Immunologie Pierre-Fabre (CIPF), Saint-Julien-en-Genevois, France

Abstract. Middle-down mass spectrometry (MD

MS) has emerged as a promising alternative to
classical bottom-up approaches for protein char-
acterization. Middle-level experiments after enzy-
matic digestion are routinely used for subunit
analysis of monoclonal antibody (mAb)-related
compounds, providing information on drug load
distribution and average drug-to-antibody ratio
(DAR). However, peptide mapping is still the gold
standard for primary amino acid sequence as-
sessment, post-translational modifications (PTM), and drug conjugation identification and localization. However,
peptide mapping strategies can be challenging when dealing with more complex and heterogeneous mAb
formats, like antibody-drug conjugates (ADCs). We report here, for the first time, MD MS analysis of a third-
generation site-specific DAR4 ADC using different fragmentation techniques, including higher-energy collisional-
(HCD), electron-transfer (ETD) dissociation and 213 nm ultraviolet photodissociation (UVPD). UVPD used as a
standalone technique for ADC subunit analysis afforded, within the same liquid chromatography-MS/MS run,
enhanced performance in terms of primary sequence coverage compared to HCD- or ETD-based MD ap-
proaches, and generated substantially more MS/MS fragments containing either drug conjugation or glycosyla-
tion site information, leading to confident drug/glycosylation site identification. In addition, our results highlight the
complementarity of ETD and UVPD for both primary sequence validation and drug conjugation/glycosylation site
assessment. Altogether, our results highlight the potential of UVPD for ADC MD MS analysis for drug
conjugation/glycosylation site assessment, and indicate that MD MS strategies can improve structural charac-
terization of empowered next-generation mAb-based formats, especially for PTMs and drug conjugation sites
validation.
Keywords: Middle-down mass spectrometry (MD MS), UVPD fragmentation, ETD, HCD, Site-specific
bioconjugation, Antibody-drug conjugate (ADC)Abbreviations ADC, Antibody-drug conjugate; CDR,

Electronic supplementary material The online version of this article (https://

doi.org/10.1007/s13361-019-02296-2) contains supplementary material, which
is available to authorized users.

Correspondence to: Sarah Cianférani; e-mail: [email protected]

 O. Hernandez-Alba et al.: MD MS of a Site-Specific ADC

Complementarity-determining region; CQA, Critical quality attribute; Cys, , Cysteine; DAR, Drug-to-antibody ratio;
DARav, Average drug-to-antibody ratio; DLD, Drug load distribution; DTT, Dithiothreitol; ETD, Electron transfer
dissociation; Fc, Fragment crystallizable; fGly, Formylglycine; HCD, Higher-energy collisional dissociation; IdeS,
Immunoglobulin-degrading enzyme from Streptococcus pyogenes; IgG, Immunoglobulin G; LC, Light chain; LC-
MS/MS, Liquid chromatography-tandem mass spectrometry; mAb, Monoclonal antibody; MD, Middle-down; MS,
Mass spectrometry; PTM, Post-translational modification; SMD, Small molecule drug; TD, Top-down; UVPD,
Ultraviolet photodissociation

Received: 21 May 2019/Revised: 15 July 2019/Accepted: 18 July 2019

Introduction detection of the payload at the peptide level. So, even if well
established, peptide mapping might be time consuming, labor-

A ntibody-drug conjugates (ADCs) are a variety of mono-

clonal antibody (mAb)-based formats that show un-
matched efficacy for treatment of many diseases, including
intensive, and incomplete in terms of characterization for het-
erogeneous ADC formats. Thus, there is still a need for method
improvement in order to provide unbiased characterization of
cancers and autoimmune diseases [1]. ADCs are comprised ADC sites of conjugation and PTMs.
of a mAb scaffold onto which a highly cytotoxic drug is To overcome the limitations of MS-based bottom-up ap-
covalently bound via a linker [2]. Among critical quality attri- proaches, emergent alternative strategies aiming at direct se-
butes (CQAs) requested by regulatory agencies such as the US quencing of the mAb without any prior sample digestion,
Food and Drug administration or European Medicine Agency, called top-down (TD) or its variation middle-down (MD), are
ADC characterization requires drug load distribution (DLD) appealing [11–14]. In theory, TD MS approaches combining
assessment, along with average drug-to-antibody ratio intact mAb measurements and direct fragmentation of 150–
(DARav) determination, the amount of unconjugated mAb 180 kDa protein ions should provide complete protein se-
(D0), and residual small molecule drugs (SMDs) [3]. Locali- quence coverage, including precise determination of PTM type
zation of the conjugation sites and post-translational modifica- and position, C- or N-terminus truncations, and mutations.
tions (PTMs) is also of utmost importance. The analytical However, in practice, TD MS is still challenging and has
characterization of ADCs relies on a combination of state-of- achieved only limited success [15]. In the best cases, electron
the-art mass spectrometry (MS), chromatography, and electro- transfer dissociation (ETD)-based TD allowed for 30–35%
phoresis techniques. Therapeutic mAbs and their sequence coverage attained of a commercial IgG [12, 16, 17],
immunoconjugates are usually analyzed at different levels due to the rigidity coming from the disulfide bridges across
(intact and subunit mass analyses and peptide mapping) to each IgG domain. One way to circumvent TD MS analysis
achieve the best analytical characterization. shortcomings is to reduce the size of the protein. In the case of
Primary sequence validation and localization of drug con- mAb-related products, this can be achieved by MD MS ap-
jugation or PTMs are usually performed by classical peptide proaches using specific enzymes to produce 25 kDa subunit
mapping strategies that consist of reduction, alkylation, and fragments for which appropriate resolution and fragmentation
enzymatic (trypsin, pepsin or a combination of proteases) di- techniques are available on current MS platforms (Q-TOF and
gestions followed by liquid chromatography-tandem mass Orbitrap) usually available in biopharmaceutical companies.
spectrometry (LC-MS/MS) analysis of the generated peptides. The production of mAb subunits is often performed using IdeS
Even if routinely performed, this approach has a series of (immunoglobulin-degrading enzyme of Streptococcus
limitations. First, digestion often occurs at basic pH and can pyogenes) [18] digestion followed by reduction with dithio-
lead to artifactual modifications of the mAb (e.g., deamidation, threitol (DTT) to produce 25 kDa subunits (Fc/2, LC, and Fd
oxidation [4–6]). In some cases, peptide mapping approaches fragments). Optimized IdeS-based MD ETD MS/MS
can also lead to incomplete sequence coverage, and thus an workflow allowed up to ∼ 50% sequence coverage to be
absence of information on the uncovered regions. Glycosyla- reached for selected IgG fragments in a single 25-min LC
tion site identification is accomplished through detection of run, and up to ∼ 70% when data obtained by distinct LC−MS
specific glycan fragmentation patterns of the glycosylated pep- runs are averaged [19]. Higher sequence coverages for Fc/2 and
tide along with the presence of signature glycan oxonium ions Fd fragments were even reported using IdeS-based MD ETD
at low mass range (204 m/z, and 366 m/z) [7, 8]. In addition, for MS/MS on a 21 Tesla FT-ICR instrument [11].
ADCs, drug conjugation adds a substantial mass increase on In the recent years, to improve TD and MD MS workflows,
mAb peptides that is combined often with a significant increase several groups have evaluated the possibilities of newly devel-
in hydrophobicity of the conjugated peptides. As a conse- oped alternative ion activation techniques. Recently, Brodbelt
quence, conjugated peptides are more difficult to separate and coworkers have demonstrated the benefits of 193 nm ul-
during LC and are challenging for MS/MS analysis [9, 10]. traviolet photodissociation (UVPD) and hybrid activation
In addition, ADC payloads are often prone to dissociation methods for intact protein analysis, including histone
during MS/MS in the mass spectrometer, avoiding the proteoforms [20], HeLa whole-cell lysate [21], and mAb
O. Hernandez-Alba et al.: MD MS of a Site-Specific ADC

analyses [13]. Under optimal conditions, UVPD resulted in ~ voltage was set to 3.6 kV and the temperature of the ion transfer
60% overall coverage of the IgG sequence, in addition to tube was 275 °C. MS/MS spectra were recorded using a mass
unambiguous glycosylation site localization and extensive cov- range of 350–2000 m/z and resolving power of 120000 at
erage of the antigen-binding complementarity-determining re- 200 m/z. The multiplexing method was used and 5 precursor
gions (CDRs) in a single LC-MS/MS experiment. Combining ions (MSX5) were selected over each specific elution window
UVPD and ETD data provided deeper sequencing and greater and subsequently fragmented using the three fragmentation
overall characterization of IgG subunits, highlighting modes. (The 5 most intense charge states were selected for
the potential of MD UVPD MS/MS strategy for the compre- HCD and UVPD. To increase the fragmentation efficiency of
hensive characterization of mAb-based therapeutics [22]. ETD, the highest charge states were targeted.) An isolation
Here, we report the first MD analysis to identify the sites of window of 0.8 m/z was used for each multiplexed ions. For
conjugation and glycosylation of a site-specific ADC with HCD fragmentation, ions were accelerated with 12 eV under a
DARav of 4 (CBW-03-106) using different fragmentation constant N2 pressure of 10−9 mbar. In the case of ETD, the
techniques, including HCD, ETD and 213 nm UVPD. Alto- automatic gain control (AGC) for the precursor ions was set to
gether, our results highlight the potential of UVPD and the 1 × 106 and all the precursor ions were allowed to react with the
complementarity of MS/MS activation techniques for MD anionic fluoranthene reagent in the linear trap for 8 ms. In order
approaches to allow for improved primary sequence coverage to enhance the radical-driven fragmentation, the reagent AGC
and conjugation site identification of the DAR4 site-specific was set to 7 × 105 with a maximum reagent injection time of
ADC. 200 ms. For UVPD MS/MS analysis, the same AGC value was
used as previously specified for ETD fragmentation. Ions were
resonantly activated with 213 nm laser during 20 or 30 ms,
Material and Methods delivering a total energy of 100–150 μJ (2 μJ/pulse).
Site-Specific DAR4 ADC
MS/MS Data Analysis
CBW-03-106 was produced and modified in-house [23]. Brief-
MS/MS spectra were analyzed with BioPharma Finder 3.0 that
ly, the minimal FGE consensus sequence (CXPXR) was cloned
into two specific sites of the mAb heavy chain (HC) using includes Xtract and Prosight algorithms to perform the
deconvolution and match the deconvoluted masses to the se-
molecular biology technologies. The antibody was then pro-
quence, respectively. MS/MS spectra were averaged through
duced in a cell line with an overexpression of human FGE that
different subunit elution windows and then deconvoluted using
oxidizes the cysteine into a formyl-glycine (fGly).
a S/N of 7. The deconvoluted masses were matched to the
sequence with a 5-ppm ion tolerance to reduce the number of
Middle-Level IdeS Digestion
false positives. Different fragment ions were considered as a
Site-specific DAR4 ADC was enzymatically digested using function of the fragmentation technique. Thus, b/y in the case
IdeS enzyme (immunoglobulin-degrading enzyme of of HCD, c/z for ETD, and a/x, b/y, and c/z fragments were
Streptococcus pyogenes) as previously described [24]. Disul- obtained in the case of the UVPD. Since the addition of the
fide bonds were subsequently reduced during 60 min at 37 °C RED-106 drug molecule was performed on cysteine residues
in strong denaturing conditions (6 M guanidine hydrochloride,) that were oxidized into formyl-glycine (fGly), all the cysteine
using DTT (Sigma) as a reducing agent (100 mM final DTT residues contained in the Fd, and Fc/2 subunits were consid-
concentration). For MD MS analysis, the sample was infused ered as putative bioconjugation sites, giving rise to eight
without any previous cleaning step. proteoforms for the Fd, and five proteoforms for the Fc/2
(Figure S2). In the case of the Fc/2 subunit, ions corresponding
Middle-Level LC-MS and LC-MS/MS Analysis to the G0F glycoform were targeted so G0F at the Asp61 was
defined as a fixed modification to search the data.
Separation of the Fc/2, Fd, and LC subunits was performed
with a Thermo Fisher MAbPac™ RP LC column (100 mm,
Bottom-Up Peptide Mapping Analysis
1 mm i.d., 4 μm particle size, and 1500 Å pore size, San Jose,
CA) preheated at 60 °C, using a Thermo Fisher Vanquish Sample Preparation Fifteen micrograms of CBW-03-106
Horizon UHPLC system (San Jose, CA). Mobile phase A ADC was solubilized in 100 mM ammonium acetate, 0.1%
consisted of 0.1% formic acid in water, and mobile phase B RapiGest™ (Waters, Milford, USA) at pH 7.4. Disulfide re-
was 0.1% formic acid in acetonitrile. For LC-MS/MS analysis, duction was performed by incubating the ADC solution with
1 to 1.5 μg of digested ADC was loaded onto the column with 5 mM DTT for 30 min at 60 °C. Alkylation was performed
20% of B at a constant flow rate of 0.1 mL/min. More details with 15 mM iodoacetamide for 30 min in the dark. After these
about the chromatographic method can be found in the legend steps, the samples were split in two for enzymatic digestion
of Figure S5. The LC system was hyphenated to a Thermo using trypsin or pepsin.
Fisher Orbitrap Fusion Lumos Tribrid mass spectrometer (San Digestion was performed by adding trypsin (Promega, Mad-
Jose, CA) equipped with three different fragmentation modes: ison, USA) to a 1:50 enzyme:substrate ratio. Samples were
HCD, ETD, and 213 nm UVPD. For all experiments, the spray incubated overnight at 37 °C. The reaction was quenched by
 O. Hernandez-Alba et al.: MD MS of a Site-Specific ADC

adding 1% of trifluoroacetic acid. RapiGest™ was eliminated and the peak intensity observed on the MS/MS spectra,
by centrifugation at 10,000g for 5 min. using Proline 1.5 software [25].
For pepsin digestion, pH was decreased to 2.0 prior to the
addition of pepsin (Promega, Madison, USA). Digestion was
performed by adding pepsin at a 1:50 enzyme:substrate ratio.
Samples were incubated at 37 °C for 3 h. The reaction was Results
stopped by heating at 95 °C for 10 min. RapiGest™ was
eliminated by a centrifugation at 10,000g for 5 min. To keep Middle-Level LC-MS Analysis of IdeS Digested
the peptides in solution, 10% of isopropanol was added after CBW-03-106
digestion. The antibody CBW-03-106 used in this study is a site-specific
DAR4 ADC generated through aldehyde-specific
LC-MS/MS Analysis NanoLC-MS/MS analysis was per- bioconjugation [26–32]. In this case, four cysteine residues
formed using a nanoAcquity Ultra-Performance-LC (Waters, found within a specific pentapeptide consensus sequence were
Milford, USA) coupled to the Q-Exactive Plus Orbitrap mass oxidized to formyl-glycine (fGly) with a formyl-glycine-
spectrometer (Thermo Scientific, Bremen, Germany) with a generating enzyme [27, 29]. The fGly residues were further
nanoSpray source. The peptides were trapped on a modified using aldehyde-specific chemistries in order to selec-
nanoACQUITY UPLC precolumn (C18, 180 μm × 20 mm, tively conjugate four RED-106 drug molecules to the mAb
5 μm particle size), and then separated on a nanoACQUITY structure, located on the Fc/2 (C213) and Fd (C163) subunits
UPLC column (C18, 75 μm × 250 mm with 1.7 μm particle [32]. We first performed a middle-level characterization of
size, Waters, Milford, USA) maintained at 60 °C. Mobile phase intact CBW-03–106 using IdeS enzymatic digestion followed
A was 0.1% (v/v) formic acid in water and mobile phase B was by DTT reduction. An improved LC method compared to the
0.1% (v/v) formic acid in acetonitrile. A gradient (1–8% B for one described in Botzanowski et al. [24] was developed and
2 min, 8–35% B for 58 min, 35–90% B for 1 min, 90% B for allowed separation of Fc/2, LC, and Fd ADC subunits within a
5 min, 90–1% B for 1 min, and maintained 1% B for 20 min) 10-min run (see “Material and Methods”). This LC-MS anal-
was used at a flow rate of 450 nL/min. The Q-Exactive Plus ysis was used as a high-resolution LC-MS survey run to
Orbitrap source temperature was set to 250 °C and spray provide information on subunit retention times, the charge state
voltage to 1.8 kV. Full scan MS spectra (300–1800 m/z) were distribution (isotopic resolution of each peak), and subsequent
acquired in positive mode at a resolution of 140 000, a maxi- sub 3 ppm mass accuracies within a single LC-MS run. Three
mum injection time of 50 ms, and an AGC target value of 3 × peaks were observed corresponding to the Fc/2, the light chain
10 6 charges, with lock-mass option being enabled (LC), and Fd fragments, respectively (Figure 1). Peak 1
(polysiloxane ion from ambient air at 445.12 m/z). The 10 most (27307.51 ± 0.01 Da, 2 ppm) corresponds to the theoretical
intense multiply charged peptides per full scan were isolated mass of the G0F glycoform of the Fc/2 with one molecule of
using a 2 m/z window and fragmented using higher-energy RED-106; peak 2 could be unambiguously attributed to the LC
collisional dissociation (normalized collision energy of 27). of the mAb (23389.31 ± 0.01 Da, 2 ppm), and peak 3 was
MS/MS spectra were acquired with a resolution of 17 500, a assessed to the conjugated Fd subunit with one RED-106
maximum injection time of 100 ms, and an AGC target value molecule (26770.34 ± 0.01 Da, 3 ppm). No signals correspond-
of 1 × 105, and dynamic exclusion was set to 60 s. The system ing to the unconjugated Fd and Fc/2 subunits could be detected
was fully controlled by XCalibur software v3.0.63, 2013 (Ther- from extracted ion chromatograms. High-resolution LC-MS
mo Scientific) and NanoAcquity UPLC console v1.51.3347 analysis also revealed the presence of two common modifica-
(Waters). tions on the Fc/2 subunit, namely C-terminal K-clipping and N-
glycosylation.
Altogether, middle-level LC-MS analysis and accurate mass
Bottom-up Data Interpretation Raw data collected were measurements enabled us to make conclusions about the drug
processed and converted in .mgf format. The mgf files from localization at the subunit level: one drug was linked to the Fc/2
trypsin and pepsin digestions were merged using Mass Spec- subunit and the other bound to the Fd fragment, as expected
trometry Data Analysis 2.7.3 (MSDA). from the position of the aldehyde tags in the heavy chain mAb
The MS/MS data were interpreted using a local Mascot sequence. No signal corresponding to unbound Fd or Fc/2 was
server with MASCOT 2.5.0 algorithm (Matrix Science, detected, leading to a DARav of 4.0 ± 0.0, in good agreement
London, UK). Spectra were searched with a mass tolerance with previous data obtained from intact and middle-level char-
of 5 ppm for MS and 0.07 Da for MS/MS data, using none acterization [24].
as enzyme. G0F glycosylation (+1445.45 Da), the linker
modification (+652.37 Da), and the whole RED-106 pay-
Peptide Mapping for Identification of Drug Conju-
load (+1198.59 Da) were specified as variable modifica-
gation Sites of the Site-Specific DAR4 ADC
tions. Protein identifications were validated with Mascot
ion score above 25. Each conjugation site was manually Peptide mapping is the gold standard for mAb and ADC
validated based on the presence of y-ion and b-ion series analysis to obtain optimal primary sequence coverage and to
O. Hernandez-Alba et al.: MD MS of a Site-Specific ADC

Figure 1. Middle-level LC-MS analysis of site-specific DAR4 CBW-03-106 ADC after Ides digestion and DTT reduction. Chro-
matogram of the IdeS-digested ADC (a). Mass spectra obtained for each subunit peak (b) and corresponding measured masses (c)

gain information on glycosylation and drug conjugation sites. bond (+ 652.37 Da on the conjugated amino acid). However,
In our routine mAb/ADC peptide mapping workflow, we no peptide containing the fragmented RED-106 product could
usually combine results from two independent digestions using be identified. Two explanations can account for that (1) the
two complementary enzymes (trypsin and pepsin, see “Mate- poor solubility of more hydrophobic drug-containing peptides,
rial and Methods”). As expected, 100% and 99% sequence and (2) the internal dissociation of the drug-linker interaction in
coverages were obtained for CBW-03-106 light and heavy HCD. Further manual data interpretation using extracted ion
chains (see Figure S1), respectively, which meets regulatory chromatograms (XIC) of tryptic peptides ions bearing the
agencies’ requirements (FDA, EMA) in terms of primary ami- CBW-03-106 payload (1175.57 m/z and 868.44 m/z) allowed
no acid sequence verification. Peptide mapping data were next detection of peptides. However, the most intense signal in MS/
interpreted for glycosylation and drug conjugation site identi- MS spectra consists of the dissociation of the drug from the
fications. The first analysis using automatic workflows (includ- conjugated peptide, resulting in an intense overwhelming sig-
ing glycosylation as variable modification, see “Material and nal at 547.22 and 485.22 m/z that hinders peptide backbone
Methods”) for bottom-up data interpretation did not provide fragmentation (Figure S1b). As a consequence, the conjugation
any information about glycosylation position. After further site is deduced from indirect XIC data interpretation of diag-
manual analysis, the masses measured from two precursor ions nostic ions from the payload fragmentation, without specific
(EEQYNSTYR and TKPREEQYNSTYR) could be potential- ions of the peptide bearing the payload.
ly assigned to glycopeptides bearing the G0F glycoform, but
their fragmentation spectra (Figure S1c) revealed that the main
MD MS Analysis with UVPD Activation Affords
fragments were obtained from the glycan fragmentation, hin-
Drug Conjugation Site Identification of the Site--
dering glycan peptide sequencing, and thus peptide validation
Specific DAR4 ADC
by automated search tools.
For drug conjugation, C163 (Fd subunit) and C213 (Fc/2 Since MD MS might be well suited to circumvent bottom-up
subunit) are expected to be modified from bioconjugation limitations, we thus used UVPD in a MD MS workflow to
reaction. No peptide bearing the intact RED-106 payload (+ evaluate the capabilities of this activation technique to provide
1198.59 Da) was detected. As the internal fragmentation of the fragment ions characteristic of the position of the two modifi-
ADC payload (linker-cytotoxic molecule, especially the ester cations of interest (N-glycosylation and payload). This strategy
bond between the linker and the drug [33]) can be induced at avoids solubility issues related to hydrophobic drug-
increasing ion internal energies, MS data were analyzed taking conjugated peptides and allows the use of UVPD, less prone
into account the disruption of the linker-cytotoxic covalent to PTM fragmentation as compared to HCD [13, 34].
 O. Hernandez-Alba et al.: MD MS of a Site-Specific ADC

Altogether, UVPD MS/MS MD experiments of the Ides- UVPD fragmentation to provide efficient fragmentation of the
digested CBW-03-106 allowed for global amino acid sequence Fc/2 backbone while preserving relative labile modifications
coverage of ~ 50% (44% Fc/2, 50% Fd, 62% LC), as already such us glycoforms (Figure 2a).
reported for unconjugated mAbs (Figure 2) [13, 22]. As the The identification of the conjugation site on the Fd subunit is
conjugation strategy targeted two cysteine residues located in more complex owing to both the position of the conjugated
the Fc/2 (C213) and Fd (C163) subunits and the N-glycoforms cysteine in the interior regions of the Fd sequence (C163) and
are also located on N61 of the heavy chain of the antibody, the the UV fragmentation pattern (Figure 2b). In spite of the signifi-
discussion will mainly focus on the Fc/2 and Fd subunits. To cant fragmentation yield of the Fd subunit upon UV irradiation
take into account possible RED-106 internal fragmentation, the (50% of sequence coverage), no clear-cut evidences about the
presence of the linker (without the cytotoxic drug) was includ- conjugation site can be afforded mostly due to the size of the
ed as a search parameter (see “Material and Methods”). fragment ions (59 amino acids average length) but also to the
After one LC-MS/MS (20 ms UVPD) run, the sequence absence of N-terminal fragments bearing the RED-106 payload.
coverage of the Fc subunit was 44% among which 34 frag- In this case, 22 C-terminal fragment ions (x, y, and z ions) contain
ments contained the RED-106 molecule (Figure 2a). Among the RED-106 payload. However, all these fragments include
the 34 payload-containing fragments, 8 C-terminal fragment several other cysteines that can be considered as putative conju-
ions containing the C213 conjugation site can be distinguished, gation sites. So the identification of the conjugation site of the Fd
allowing the unambiguous identification of the Fc/2 conjuga- subunit cannot be delimited to one specific cysteine.
tion site (Figure 2a). Similar to the identification of the conju- For these reasons, the variation of additional parameters, such
gation site, UVPD activation allows for the characterization of as the sequence coverage and the number of payload-containing
the N-glycosylation of the CBW-03-106. The fragmentation fragments, as a function of the hypothetical position of the payload
map of the Fc/2 subunit exhibits 31 fragment ions that contain conjugation site was taken into account to guide the determination
the intact G0F scaffold, thus pointing out the suitability of the of the specific conjugated cysteine. In the case of the Fd subunit,
the sequence coverage increases from 37% when C22 was as-
sumed to bear the drug to 50% for the C163-conjugated
proteoform (Table S1a). In addition, 8 MS/MS fragments that
contained the RED-106 molecule were observed when conjuga-
tion was supposed to be on C22 compared to 22 MS/MS frag-
ments when C163 was hypothesized to be conjugated
(Table S1a). This latter parameter (number of MS/MS fragments)
remains constant when the payload is assumed on cysteine resi-
dues closer to the C-terminus (C202, C222, C228, and C231)
while the global sequence coverage decreases (Table S1), which
allowed us to reinforce the confidence that the RED-106 is more
likely conjugated on the C163 of the Fd subunit.

ETD and HCD Fragmentations for More Confi-

dence in Glycosylation and Conjugation Sites
Characterization
UVPD fragmentation capabilities for CBW-03-106 characteriza-
tion were benchmarked against more conventional fragmentation
techniques, like HCD [35] and ETD [36, 37], more frequently
used in biopharmaceutical R&D labs, to evaluate the limitations/
benefits associated with each individual technique and to comple-
ment our UVPD dataset.
HCD led in fact to the lowest primary amino acid sequence
coverage and the lowest number of cleavage sites that contain the
RED-106 conjugation or the N-glycosylation compared to UVPD
Figure 2. Fragmentation maps of Fc/2 (a), Fd (b), and LC (c) or ETD (Figure S3). In this case, HCD fragmentation yielded very
subunits after irradiation with 213 nm UVPD when RED-106
low number of fragment ions characteristic of both conjugation
payload is assumed on C213 for the Fc subunit and C163 for
the Fd. The N-glycosylation (G0F) and the payload conjugation
sites (1 fragment ion characteristic of the C213 Fc/2 and 0
sites are highlighted with green squares. Blue and red circles fragment ion corresponding to the C163 on the Fd subunit). In-
show the identified cleavage sites that contain the G0F depth analysis of MS/MS spectra showed that internal fragmen-
glycoform and the RED-106 payload, respectively. Specific tation of the RED-106 payload was favored under HCD condi-
fragment ions of C213 conjugation site are depicted with black tions compared to UVPD (Figure S4). Intensities of diagnostic
arrows ions for payload internal fragmentation (547.22 m/z) were higher
O. Hernandez-Alba et al.: MD MS of a Site-Specific ADC

with HCD than with UVPD MS/MS spectra (Figure S4). This obtained by ETD, showing its suitability to characterize the posi-
higher payload internal fragmentation might also explain the tion of the glycoforms in mAb structures. In the case of the Fd
results obtained by peptide mapping upon HCD MS/MS frag- subunit, despite the detection of 16 ETD fragments bearing the
mentation (no RED-106 containing peptide detected). Addition- RED-106 payload, no specific fragment ions of the C163 conju-
ally, the site of N-glycosylation with HCD is difficult to identify, gation position that could led to the univocal identification of the
due to the low number of identified cleavage sites containing the conjugation site on C163 was observed, as previously observed
intact G0F glycoform (Figure S3d), which corroborates peptide upon UVPD activation (Figure 3b). While conjugation sites in the
mapping observations. central region of the subunit sequences require the presence of N-
ETD fragmentation was then carried out to induce the back- terminal and C-terminal fragments specific of one single conju-
bone cleavage of the site-specific DAR4 ADC subunits gation site to unambiguously determine the site of conjugation,
(Figure 3). Electron-driven fragmentation techniques have already other signature fragment ions can be used to discard the co-
proven their utility for mAbs sequencing at top- [12, 16, 17] and existence of different proteoforms. In our case, four ETD fragment
middle-level [13, 22, 38] characterization. In our case, after 8 ms ions (c147, c148, c154, and c157) point out that the C143 is not
of reaction with the anionic reagent, the sequence coverage of the conjugated, and at least 29 fragment ions support that the conju-
Fc/2, Fd, and LC subunits was 37%, 47%, and 52%, respectively gation site is not located on cysteine amino acids at the C-terminal
(Figure 3), leading to an overall mAb sequence coverage of 45%. side (C202, C222, C228, and C231) of the Fd subunit (Figure 3b).
Albeit slightly lower sequence coverage was obtained compared In addition, the best sequence coverage is obtained when the
to UVPD, overall, the fragmentation maps corresponding to the RED-106 position is assumed on the C163 as previously observed
ETD fragmentation exhibit fragment ions that contain the RED- using UVPD fragmentation (Table S1b). Altogether, these results
106 modification. ETD generates 19 cleavage sites that contain lead to the indirect conclusion that the conjugation site of the Fd
the payload in the Fc/2 structure (Figure 3a) and four of these subunit is located on the C163.
cleavage sites are specific of the C213 conjugation site, strength-
ening the results obtained with UVPD activation and thus provid-
ing more confident drug conjugation site identification. Similarly,
Complementarity Between UVPD, HCD, and ETD
17 cleavage sites that contain the G0F moiety along with unam-
in MD MS of CBW-03-106
biguous localization of G0F site owing to the identification of the Even though UVPD provides the best outcome concerning the
c60 and c61 consecutive fragment ions (Figure 3d) were also three principal parameters for ADC characterization (number

Figure 3. Fragmentation maps of Fc/2 (a), Fd (b), and LC (c) subunits after ETD fragmentation. The N-glycosylation (G0F) and the
payload conjugation sites are highlighted with green squares. Blue and red circles show the identified cleavage sites that contain the
G0F glycoform and the RED-106 payload, respectively. Specific fragment ions of C213 conjugation site are depicted with black
arrows. The deconvoluted ETD fragment spectrum of the Fc/2 subunit showing a mass difference between the two consecutives c60
and c61 ions that corresponds to the mass of the intact G0F structure (d)
 O. Hernandez-Alba et al.: MD MS of a Site-Specific ADC

of RED-106 cleavage sites, number of G0F cleavage sites, and

global amino acid sequence coverage), substantial benefits are
obtained when MS/MS data from the three activation tech-
niques are combined, highlighting the complementarity be-
tween HCD, ETD, and UVPD (Figure 4). After combining
the results from the three activation techniques, the primary
sequence coverage of the three ADC subunits rose to 64%,
74% and 83% for Fc/2, Fd and LC, respectively (Figure 4a),
which approach the highest sequence coverages reported in
literature using MD approaches [13, 22]. In addition, only a
limited number of MS/MS fragments are common to the three
activation modes (18 for both Fc/2 and Fd subunits, and 37 for
LC), highlighting the complementarity of UVPD, HCD, and
ETD for overall sequence coverage (Figure S5). For CBW-03-
106 primary sequence coverage using MD approaches, our
results demonstrate a limited contribution of HCD data, but a
strong complementarity between ETD and UVPD activation
techniques (Figure S5), as already reported for unconjugated
mAbs [13, 22]. Figure 5. Venn diagram of RED-106 bearing cleavage sites
When focusing on identification of the RED-106 drug con- obtained throughout the Fc/2 (a), and Fd (b) subunits upon
jugation (on C163 and C213) or N-glycosylation sites (N61) UVPD, ETD, and HCD fragmentations. Number of cleavage
(Figures 4 and 5), the complementarity between UVPD, ETD, sites containing the G0F glycoform (N61) in the Fc/2 domain
and HCD is even more obvious. Indeed, no RED-106 contain- after fragmentation with the three different techniques (c)
ing fragments are common to all three activation modes
(Figure 5). Again, the number of RED-106 cleavage sites higher number of drug-containing MS/MS fragment ions along
considerably increases when results of the three fragmentation with the number of ions characteristic of a single conjugation
techniques were combined: 58 RED-106 containing fragments site obtained by UVPD compared to ETD or HCD reveals the
can be identified within the Fc/2 sequence and 34 in the case of potential of UVPD for more confident drug-conjugation site
the Fd subunit (Figure 4b), which are significantly higher than determination. However, specific MS/MS fragments allowed
those obtained after using only UVPD (34 and 22 respectively). excluding the positioning of RED-106 on cysteine’s upstream
Furthermore, among the total number of RED-106 containing from C163. Moreover, pairwise association of two fragmenta-
MS/MS fragments (92), 47% result specifically from UVPD tion modes increased the confidence in drug conjugation bind-
compared to 27% and 10% specific fragments generated by ing site identification (Figure 5), through validation of cleavage
ETD and HCD, respectively. It is important to notice that sites by at least two different fragmentation techniques (3, 5,
UVPD afforded also the highest number of payload- and 2 RED-106 containing MS/MS spectra of Fc/2 shared
containing fragments specific of the C213 on the Fc/2 subunit between UVPD/ETD, UVPD/HCD, and ETD/HCD,
(8 fragments) compared to ETD (4 fragments) and HCD (1 respectively).
fragment). Conversely, all fragmentation techniques tested in The benefits of combining HCD, ETD, and UVPD can also
the present work failed in direct identification of a fragment be observed for the identification of the G0F glycosylation site.
bearing RED-106 on Cys163. Altogether, the significantly Thereby, 44 G0F-containing cleavage sites were identified

Figure 4. Contribution of HCD (blue), ETD (red), UVPD (green), ETD + UVPD (pink) and the combination of the three activation
techniques (gray) to overall sequence coverage (a), number of RED-106 containing fragments (b), and number of G0F containing
payloads (c)
O. Hernandez-Alba et al.: MD MS of a Site-Specific ADC

after the combination of MS/MS data coming from the three dissociated in MS/MS, resulting in hardly detectable intact
activation techniques, while only 31 G0F containing fragments drug-conjugated peptides. Apart from primary acid verifica-
were detected using UVPD (Figures 4 and 5). In spite of the tion, DLD, DARav, and glycoprofiles are not attainable by
significantly lower number of G0F-fragments obtained upon peptide mapping.
ETD fragmentation [17] compared to UVPD [31], this activa- For all the above-mentioned reasons, MD approaches have
tion technique allowed the identification of the glycosylation gained interest and may even complement conventional and
site of the CBW-03-106 at the amino acid level, in agreement routine bottom-up analyses when they reach their limits. Espe-
with UVPD. Similarly to drug conjugation, no N61 G0F- cially for ADC characterization, we demonstrate here using the
contaning MS/MS fragment was common to the three activa- CBW-03-106 site-specific DAR4 ADC that state of the art MD
tion techniques, illustrating the complementarity of the frag- approaches provide rapid and accurate drug conjugation site
mentation techniques for G0F site identification. assessment, information that was not easily obtained by peptide
Altogether, our MD results illustrate a clear complementar- mapping. Advantages of middle-level approaches include
ity between UVPD, ETD, and HCD for global amino acid faster analysis time (10 min/run) combined with limited sample
sequence coverage and both drug conjugation and glycosyla- preparation. At the MS2 level, MD MS/MS analysis can pro-
tion site identification. vide additional specific information, like PTM (glycosylation)
Taking into account that these results imply four indepen- or drug conjugation site determination at the amino acid level.
dent LC-MS/MS runs (one LC-MS, and three LC-MS/MS) and Even if peptide mapping is qualified as a critical quality test by
that HCD has only minor contribution to the MS/MS data, we regulatory agencies for routine amino acid sequence validation,
evaluated the benefits of combining one LC-UVPD MS/MS innovative complementary approaches must be strengthened
and one LC-ETD MS/MS analysis. Interestingly, only minor and pushed forward to improve the portfolio of analytical
differences can be observed in terms of sequence coverage and workflows for the more complex mAb formats being
the number of identified payload/G0F-containing cleavage developed.
sites after combining MS/MS data from ETD and UVPD In MD approaches, UVPD used as a standalone technique
(pink bars Figure 4). For CBW-03-106 MD MS characteriza- showed better performance compared to HCD and ETD not
tion, the combination of ETD and UVPD is the best option to only for overall sequence coverage, but also for drug or glyco-
provide the highest sequence coverage (~ 65%), along with sylation binding site identification. Our results are in line with
unambiguous drug conjugation and glycosylation site identifi- those obtained by the groups of Brodbelt and Kelleher on
cation from two independent fragmentation techniques. therapeutic unconjugated mAbs [13, 22], suggesting that the
presence of hydrophobic drugs on the amino acid scaffold does
not affect UVPD performances.
A clear complementarity between UVPD and ETD in MD
Conclusions approaches was also demonstrated for CBW-03-106, providing
We report here for the first time the use of MD approaches for complementary MS/MS information for improved sequence
the characterization of a third-generation site-specific DAR4 coverages and more accurate drug conjugation and PTM site
ADC (CBW-03-106). Three fragmentation techniques (HCD, assessment. The combination of one MS1 and two MS2
ETD, and 213 nm UVPD) have been used to induce the UVPD/ETD MS/MS run appears to be a good compromise
fragmentation of the Fc/2, Fd, and LC subunits of CBW-03- for drug-conjugation site assessment. Indeed, a relatively low
106. Our work highlights the great potential of UVPD, a new amount of redundant information was obtained from MS/MS
commercially available fragmentation technique on Orbitrap data from ETD and UVPD, while the overall sequence cover-
platforms, for drug conjugation and glycosylation site assess- age and the confidence in identification of PTM/conjugation
ment of ADCs. sites of the three subunits rose significantly, pinpointing a
ADC characterization requires assessment of a series of synergic effect between both activation techniques. However,
CQAs [39], including primary amino acid sequence verifica- identification of conjugation/PTM sites in the interior region of
tion, glycoprofile and glycosylation site determination, DLD, subunit sequences can be challenging for MD approaches since
DARav assessment, and identification of drug conjugation specific N-terminal and C-terminal fragments are needed to
sites. Currently, no unique analytical workflow tackles all these assign these modifications to one specific amino acid.
issues, and multilevel analyses (intact and subunit mass analy- We believe that the routine implementation of MD ap-
sis and peptide mapping) are advised for extensive primary proaches in R&D laboratories will strongly depend on the
structure determination [39]. Peptide mapping approaches development of tailored software modules based on statistical
consisting of enzymatic digestion followed by routine LC- studies to correlate a specific score to each putative
MS/MS analysis are the gold standard for primary amino acid conjugation/PTM site. Altogether, our results highlight the
sequence assessment, including PTM identification and local- benefits of combining MD MS analysis with multiple ion
ization. Bottom-up limitations are clear in ADC drug- activation techniques to provide an extensive primary structure
conjugation site identification, as many drugs increase the characterization of ADCs. With the ongoing progress in mass
hydrophobicity of the resulting peptides, often leading to less spectrometry instrumentation, activation techniques, and de-
soluble entities. In addition, drug-linker interactions might be velopment of adapted user-friendly software for automated
 O. Hernandez-Alba et al.: MD MS of a Site-Specific ADC

scoring, validation, and statistical evaluation of MS/MS spec- antibodies by electron transfer dissociation mass spectrometry. Anal.
Chem. 83(23), 8919–8927 (2011)
tra, we believe that MD MS will soon be mature for more 15. Han, X., Jin, M., Breuker, K., McLafferty, F.W.: Extending top-down
routine use in biopharma companies for next-generation mass spectrometry to proteins with masses greater than 200 kilodaltons.
empowered mAb-based formats characterization. Science. 314(5796), 109–112 (2006)
16. Mao, Y., Valeja, S.G., Rouse, J.C., Hendrickson, C.L., Marshall, A.G.:
Top-down structural analysis of an intact monoclonal antibody by elec-
Acknowledgements tron capture dissociation-Fourier transform ion cyclotron resonance-mass
spectrometry. Anal. Chem. 85(9), 4239–4246 (2013)
The authors would like to thank Région Alsace for financial 17. Fornelli, L., Damoc, E., Thomas, P.M., Kelleher, N.L., Aizikov, K.,
Denisov, E., Makarov, A., Tsybin, Y.O.: Analysis of intact monoclonal
support in purchasing an Orbitrap Exactive Plus EMR instru- antibody IgG1 by electron transfer dissociation orbitrap FTMS. Mol. Cell.
ment, the CNRS, the University of Strasbourg, the “Agence Proteomics. 11(12), 1758–1767 (2012)
Nationale de la Recherche” (ANR) and the French Proteomic 18. Chevreux, G., Tilly, N., Bihoreau, N.: Fast analysis of recombinant
monoclonal antibodies using IdeS proteolytic digestion and electrospray
Infrastructure (ProFI; ANR-10-INBS-08-03). O.A-H acknowl- mass spectrometry. Anal. Biochem. 415(2), 212–214 (2011)
edges the IdeX program of the University of Strasbourg for 19. Fornelli, L., Ayoub, D., Aizikov, K., Beck, A., Tsybin, Y.O.: Middle-
funding his postdoctoral fellowship. down analysis of monoclonal antibodies with electron transfer dissocia-
tion orbitrap Fourier transform mass spectrometry. Anal. Chem. 86(6),
3005–3012 (2014)
20. Greer, S.M., Brodbelt, J.S.: Top-down characterization of heavily modi-
References fied histones using 193 nm ultraviolet photodissociation mass spectrom-
etry. J. Proteome Res. 17(3), 1138–1145 (2018)
21. Cleland, T.P., DeHart, C.J., Fellers, R.T., VanNispen, A.J., Greer, J.B.,
1. Carter, P.J., Lazar, G.A.: Next generation antibody drugs: pursuit of the
Leduc, R.D., Parker, W.R., Thomas, P.M., Kelleher, N.L., Brodbelt, J.S.:
‘high-hanging fruit’. Nat. Rev. Drug Discov. 17(3), 197–223 (2018)
High-throughput analysis of intact human proteins using UVPD and
2. Beck, A., Goetsch, L., Dumontet, C., Corvaia, N.: Strategies and chal- HCD on an orbitrap mass spectrometer. J. Proteome Res. 16(5), 2072–
lenges for the next generation of antibody drug conjugates. Nat. Rev. 2079 (2017)
Drug Discov. 16(5), 315–337 (2017)
22. Fornelli, L., Srzentić, K., Huguet, R., Mullen, C., Sharma, S.,
3. Beck, A., Terral, G., Debaene, F., Wagner-Rousset, E., Marcoux, J., Zabrouskov, V., Fellers, R.T., Durbin, K.R., Compton, P.D., Kelleher,
Janin-Bussat, M.C., Colas, O., Van Dorsselaer, A., Cianferani, S.: N.L.: Accurate sequence analysis of a monoclonal antibody by top-down
Cutting-edge mass spectrometry methods for the multi-level structural and middle-down Orbitrap mass spectrometry applying multiple ion
characterization of antibody-drug conjugates. Expert Rev. Proteomics. activation techniques. Anal. Chem. 90(14), 8421–8429 (2018)
13(2), 157–183 (2016) 23. Barfield, R.M., Rabuka, D.: Leveraging formylglycine-generating en-
4. Zhang, Z.Q., Pan, H., Chen, X.Y.: Mass spectrometry for structural zyme for production of site-specifically modified bioconjugates. In:
characterization of therapeutic antibodies. Mass Spectrom. Rev. 28(1), Lemke, E.A. (ed.) Noncanonical amino acids: methods and protocols,
147–176 (2009) pp. 3–16. Springer New York, New York (2018)
5. Ren, D., Pipes, G.D., Liu, D.J., Shih, L.Y., Nichols, A.C., Treuheit, M.J., 24. Botzanowski, T., Erb, S., Hernandez-Alba, O., Ehkirch, A., Colas, O.,
Brems, D.N., Bondarenko, P.V.: An improved trypsin digestion method Wagner-Rousset, E., Rabuka, D., Beck, A., Drake, P.M., Cianférani, S.:
minimizes digestion-induced modifications on proteins. Anal. Biochem. Insights from native mass spectrometry approaches for top- and middle-
392(1), 12–21 (2009) level characterization of site-specific antibody-drug conjugates. mAbs.
6. Krokhin, O.V., Antonovici, M., Ens, W., Wilkins, J.A., Standing, K.G.: 9(5), 801–811 (2017)
Deamidation of -Asn-Gly-sequences during sample preparation for pro- 25. Carapito, C., Lane, L., Benama, M., Opsomer, A., Mouton-Barbosa, E.,
teomics: consequences for MALDI and HPLC-MALDI analysis. Anal. Garrigues, L., de Peredo, A.G., Burel, A., Bruley, C., Gateau, A.,
Chem. 78(18), 6645–6650 (2006) Bouyssie, D., Jaquinod, M., Cianferani, S., Burlet-Schiltz, O., Van
7. Segu, Z.M., Mechref, Y.: Characterizing protein glycosylation sites Dorsselaer, A., Garin, J., Vandenbrouck, Y.: Computational and mass-
through higher-energy C-trap dissociation. Rapid Commun. Mass spectrometry-based workflow for the discovery and validation of missing
Spectrom. 24(9), 1217–1225 (2010) human proteins: application to chromosomes 2 and 14. J. Proteome Res.
8. Wuhrer, M., Catalina, M.I., Deelder, A.M., Hokke, C.H.: 14(9), 3621–3634 (2015)
Glycoproteomics based on tandem mass spectrometry of glycopeptides. 26. Drake, P.M., Carlson, A., McFarland, J.M., Bañas, S., Barfield, R.M.,
J. Chromatogr. B. 849(1), 115–128 (2007) Zmolek, W., Kim, Y.C., Huang, B.C.B., Kudirka, R., Rabuka, D.: CAT-
9. Janin-Bussat, M.C., Dillenbourg, M., Corvaia, N., Beck, A., Klinguer- 02-106, a site-specifically conjugated anti-CD22 antibody bearing an
Hamour, C.: Characterization of antibody drug conjugate positional iso- MDR1-resistant maytansine payload yields excellent efficacy and safety
mers at cysteine residues by peptide mapping LC-MS analysis. J. in preclinical models. Mol. Cancer Ther. 17(1), 161–168 (2018)
Chromatogr. B Analyt. Technol. Biomed. Life Sci. 981, 9–13 (2015) 27. York, D., Baker, J., Holder, P.G., Jones, L.C., Drake, P.M., Barfield,
10. Wagner-Rousset, E., Janin-Bussat, M.C., Colas, O., Excoffier, M., R.M., Bleck, G.T., Rabuka, D.: Generating aldehyde-tagged antibodies
Ayoub, D., Haeuw, J.F., Rilatt, I., Perez, M., Corvaia, N., Beck, A.: with high titers and high formylglycine yields by supplementing culture
Antibody-drug conjugate model fast characterization by LC-MS follow- media with copper(II). BMC Biotechnol. 16(1), 23 (2016)
ing IdeS proteolytic digestion. Mabs. 6(1), 173–184 (2014) 28. Kudirka, R.A., Barfield, R.M., McFarland, J.M., Drake, P.M., Carlson,
11. He, L.D., Anderson, L.C., Barnidge, D.R., Murray, D.L., Hendrickson, A., Bañas, S., Zmolek, W., Garofalo, A.W., Rabuka, D.: Site-specific
C.L., Marshall, A.G.: Analysis of monoclonal antibodies in human serum tandem Knoevenagel condensation–Michael addition to generate
as a model for clinical monoclonal gammopathy by use of 21 Tesla FT- antibody–drug conjugates. ACS Med. Chem. Lett. 7(11), 994–998 (2016)
ICR top-down and middle-down MS/MS (vol 28, pg 827, 2017). J. Am. 29. Kudirka, R., Barfield, R.M., McFarland, J., Albers, A.E., de Hart, G.W.,
Soc. Mass Spectrom. 28(5), 839–839 (2017) Drake, P.M., Holder, P.G., Banas, S., Jones, L.C., Garofalo, A.W.,
12. Fornelli, L., Ayoub, D., Aizikov, K., Liu, X.W., Damoc, E., Pevzner, Rabuka, D.: Generating site-specifically modified proteins via a versatile
P.A., Makarov, A., Beck, A., Tsybin, Y.O.: Top-down analysis of im- and stable nucleophilic carbon ligation. Chem. Biol. 22(2), 293–298
munoglobulin G isotypes 1 and 2 with electron transfer dissociation on a (2015)
high-field orbitrap mass spectrometer. J. Proteome. 159, 67–76 (2017) 30. Holder, P.G., Jones, L.C., Drake, P.M., Barfield, R.M., Bañas, S., de Hart,
13. Cotham, V.C., Brodbelt, J.S.: Characterization of therapeutic monoclonal G.W., Baker, J., Rabuka, D.: Reconstitution of formylglycine-generating
antibodies at the subunit-level using middle-down 193 nm ultraviolet enzyme with copper(II) for aldehyde tag conversion. J. Biol. Chem.
photodissociation. Anal. Chem. 88(7), 4004–4013 (2016) 290(25), 15730–15745 (2015)
14. Tsybin, Y.O., Fornelli, L., Stoermer, C., Luebeck, M., Parra, J., Nallet, S., 31. Agarwal, P., Kudirka, R., Albers, A.E., Barfield, R.M., de Hart, G.W.,
Wurm, F.M., Hartmer, R.: Structural analysis of intact monoclonal Drake, P.M., Jones, L.C., Rabuka, D.: Hydrazino-Pictet-Spengler ligation
O. Hernandez-Alba et al.: MD MS of a Site-Specific ADC

as a biocompatible method for the generation of stable protein conjugates. 36. Coon, J.J., Ueberheide, B., Syka, J.E.P., Dryhurst, D.D., Ausio, J.,
Bioconjug. Chem. 24(6), 846–851 (2013) Shabanowitz, J., Hunt, D.F.: Protein identification using sequential ion/
32. Rabuka, D., Rush, J.S., de Hart, G.W., Wu, P., Bertozzi, C.R.: Site- ion reactions and tandem mass spectrometry. Proc. Natl. Acad. Sci. U. S.
specific chemical protein conjugation using genetically encoded aldehyde A. 102(27), 9463–9468 (2005)
tags. Nat. Protoc. 7, 1052 (2012) 37. Syka, J.E.P., Coon, J.J., Schroeder, M.J., Shabanowitz, J., Hunt, D.F.:
33. Sang, H., Lu, G., Liu, Y., Hu, Q., Xing, W., Cui, D., Zhou, F., Zhang, J., Peptide and protein sequence analysis by electron transfer dissociation
Hao, H., Wang, G., Ye, H.: Conjugation site analysis of antibody-drug- mass spectrometry. Proc. Natl. Acad. Sci. U. S. A. 101(26), 9528–9533
conjugates (ADCs) by signature ion fingerprinting and normalized area (2004)
quantitation approach using nano-liquid chromatography coupled to high 38. Srzentic, K., Nagornov, K.O., Fornelli, L., Lobas, A.A., Ayoub, D.,
resolution mass spectrometry. Anal. Chim. Acta. 955, 67–78 (2017) Kozhinov, A.N., Gasilova, N., Menin, L., Beck, A., Gorshkov, M.V.,
34. Halim, M.A., MacAleese, L., Lemoine, J., Antoine, R., Dugourd, P., Aizikov, K., Tsybin, Y.O.: Multiplexed middle-down mass spectrometry
Girod, M.: Ultraviolet, infrared, and high-low energy photodissociation as a method for revealing light and heavy chain connectivity in a mono-
of post-translationally modified peptides. J. Am. Soc. Mass Spectrom. clonal antibody. Anal. Chem. 90(21), 12527–12535 (2018)
29(2), 270–283 (2018) 39. Beck, A., Terral, G., Debaene, F., Wagner-Rousset, E., Marcoux, J.,
35. Olsen, J.V., Macek, B., Lange, O., Makarov, A., Horning, S., Mann, M.: Janin-Bussat, M.-C., Colas, O., Dorsselaer, A.V., Cianférani, S.:
Higher-energy C-trap dissociation for peptide modification analysis. Nat. Cutting-edge mass spectrometry methods for the multi-level structural
Methods. 4, 709 (2007) characterization of antibody-drug conjugates. Expert Rev. Proteomics.
13(2), 157–183 (2016)