This is an open access article published under an ACS AuthorChoice License, which permits

copying and redistribution of the article or any adaptations for non-commercial purposes.

Article

Cite This: Anal. Chem. 2019, 91, 3156−3162 pubs.acs.org/ac

Simple Approach for Improved LC−MS Analysis of Protein

Biopharmaceuticals via Modification of Desolvation Gas
Shunhai Wang,*,† Tao Xing,† Anita P. Liu, Zehong He, Yuetian Yan, Thomas J. Daly, and Ning Li
Analytical Chemistry Group, Regeneron Pharmaceuticals Inc., 777 Old Saw Mill River Road, Tarrytown, New York 10591-6707,
United States
*
S Supporting Information

ABSTRACT: LC−MS based analysis of protein biopharmaceut-

See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

icals could benefit from improved data quality, which can

subsequently lead to improved drug characterization with higher
confidence and less ambiguity. In this study, we created a simple
device to modify the desolvation gas on a Q-Exactive mass
Downloaded via YONSEI UNIV on November 30, 2022 at 02:51:33 (UTC).

spectrometer and to demonstrate the utility in improving both

peptide mapping analysis and intact mass analysis, the two most
routinely and widely applied LC−MS techniques in protein
biopharmaceutical characterization. By modifying the desolvation
gas with acid vapor from propionic acid (PA) and isopropanol
(IPA), the ion suppression effects from trifluoroacetic acid (TFA) in
a typical peptide mapping method can be effectively mitigated, thus
leading to improved MS sensitivity. By modifying the desolvation
gas with base vapor from triethylamine (TEA), the charge reduction effect can be achieved and utilized to improve the spectral
quality from intact mass analysis of protein biopharmaceuticals. The approach and device described in this work suggests a low-
cost and practical solution to improve the LC−MS characterization of protein biopharmaceuticals, which has the potential to be
widely implemented in biopharmaceutical analytical laboratories.

T he liquid chromatography coupled to mass spectrometry

(LC−MS) technique has already become an indispen-
sable tool for in-depth characterization of protein biopharma-
ceuticals in analytical laboratories to support their develop-
ment and license applications.1 Based on a recent retrospective
evaluation of the use of mass spectrometry (MS) in the Food
and Drug Administration (FDA) biological license applications
(BLAs), an overwhelming majority of those applications have
implemented MS or LC−MS based workflows.2 To provide
characterization of different protein attributes, a wide variety of
LC−MS based assays can be performed, in which both peptide
mapping analysis and intact mass analysis are most routinely
and widely applied. To improve the confidence of the analysis
and reduce the ambiguity associated with data interpretation,
continual efforts have been made to improve the data quality
from the LC−MS analysis, including using optimized
experimental procedures, fine-tuned instrument parameters,
as well as more advanced mass spectrometers. Here, we
present a novel approach to improve the LC−MS data quality
from both peptide mapping analysis and intact mass analysis of
protein biopharmaceuticals by modifying the desolvation gas
using a simple device.
LC−MS based peptide mapping analysis is routinely applied
to confirm the primary structure of protein biopharmaceuticals,
in which a protein molecule is first hydrolyzed into small
peptide fragments using a protease with known specificity
(although a nonspecific protease can also be applied), then the
amino acid sequence of each peptide fragment is determined
by LC−MS/MS analysis, taking into consideration the cDNA
predicted sequence and the protease specificity.3−5 Data from
peptide mapping analysis can also be utilized to identify and
quantify post-translational modifications, confirm the disulfide
bond linkages, and even detect amino acid substitution events
present at very low levels (<0.1%).6 During peptide mapping
analysis of protein biopharmaceuticals, LC−MS is often
performed in combination with ultraviolet (UV) detection to
generate so-called UV fingerprints, which alone can be used as
an identification assay during quality control (QC) and drug
release. To effectively separate peptides on a reversed-phase
column with good peak shape, trifluoroacetic acid (TFA) is
commonly used as a mobile phase modifier due to its excellent
ion pairing ability.7 However, TFA is also known for its ion
suppression effects during the electrospray (ESI) process due
to the increased surface tension as well as its ability to form ion
pairs with analytes in the gas phase, thus leading to significantly
decreased MS sensitivity.8 Over the past 2 decades, different
strategies have been investigated and implemented to alleviate
the decrease in MS sensitivity due to TFA. For example,
modifying the TFA-containing mobile phases with acetic acid
or propionic acid has demonstrated a significant MS signal
enhancement without compromising chromatographic integ-
Received: December 19, 2018
Accepted: January 25, 2019
Published: January 25, 2019

© 2019 American Chemical Society 3156 DOI: 10.1021/acs.analchem.8b05846

Anal. Chem. 2019, 91, 3156−3162
Analytical Chemistry
rity during bioanalysis of some basic compounds.9 Postcolumn
addition of a mixture of propionic acid and isopropanol is
another commonly used strategy that does not require the
modification of the LC method. However, the experimental
setup requires additional pumps, consumes large quantities of
reagents, and is not suitable for continuous analysis of large
sample sets.10 Acid vapor−assisted ESI within an enclosed
spray chamber is another elegant solution to counteract the
signal suppression effects of TFA.11 However, its utility in
protein biopharmaceutical characterization has been limited so
far, presumably due to the requirement of a special ESI source.
Finally, advances in reversed-phase column chemistry,
particularly in the development of charged-surface C18
stationary phases, significantly reduced the dependence of
using TFA to achieve good peak shape during peptide mapping
analysis.12 Replacing TFA with a MS-friendly mobile phase
modifier (e.g., formic acid), however, will inevitably reduce the
retention of most peptides, rendering some short and
hydrophilic peptides undetectable due to coelution with the
solvent front, resulting in decreased sequence coverage.
Nevertheless, until those new columns have been routinely
and widely adopted in protein biopharmaceutical character-
ization, TFA-based LC−MS methods are still the standard in
peptide mapping analysis. In this study, we describe a novel
approach to counteract TFA ion suppression during LC−MS
analysis by modifying the desolvation gas with acid vapor from
propionic acid and isopropanol. The modification was
conveniently implemented using a device assembled from all
commercially available parts. Unlike previous reports in which
the acid vapor-assisted ESI setups were often realized using
nano- or capillary-flow,11,13−15 the developed device over-
comes these limitations and can be easily integrated into a
standard Ion Max ion source (Thermo Scientific) for routine
peptide mapping analysis using analytical flow.
We have also demonstrated that the same device can be
utilized to modify the desolvation gas with weak base, in order
to improve the LC−MS based intact mass analysis of protein
biopharmaceuticals via charge reduction. Charge reduction
strategies have already been widely exploited in many intact
mass applications by either simplifying the MS spectra of
complex protein samples or by better preserving noncovalent
interactions during native MS experiments.16−19 During
evaluation of the developed device, we found that charge
reduction strategies can also be utilized to stabilize highly
charged protein analytes during the MS analysis and prevent
them from undesired fragmentation and decay, thus leading to
overall improved spectral quality. We also demonstrated the
utility of this approach by successfully deciphering a highly
heterogeneous protein molecule that cannot be resolved by the
standard method. Again, unlike many previous reports where
charge reduction was achieved by either modifying the mobile
phases or postcolumn addition of the weak bases,16,19 the new
device enables the charge reduction via the desolvation gas,
thus eliminating concerns over chromatographic performance
as well as retaining the practicality for routine analysis.
■ EXPERIMENTAL SECTION
Materials. Deionized water was provided by a Milli-Q
integral water purification system installed with a MilliPak
Express 20 filter (Millipore Sigma, Burlington, MA). NIST
Monoclonal Antibody Reference Material 8671 (NISTmAb,
humanized IgG1κ monoclonal antibody) was purchased from
the National Institute of Standards and Technology (Gaithers-
3157
Article
burg, MD). Protein-X is a highly glycosylated protein
recombinantly produced at Regeneron (Tarrytown, NY).
Rapid Peptide N-glycosidase F (Rapid PNGase F) with 5×
Rapid PNGase F buffer was purchased from New England
Biolabs Inc. (Ipswich, MA). TCEP-HCl (tris(2-carboxyethyl)
phosphine hydrochloride), Tris-HCl pH 7.5 (UltraPure),
trifluoroacetic acid (LC−MS grade), formic acid (LC−MS
grade), and acetonitrile (Optima LC/MS grade) were
purchased from Thermo Fisher Scientific (Waltham, MA),
and trypsin (sequencing grade) was purchased from Promega
(Madison, MI). Triethylamine, iodoacetamide, propionic acid,
and acetic acid were purchased from Sigma-Aldrich, Co. (St.
Louis, MO). 2-Propanol (IPA) (HPLC grade) was purchased
from VWR International, LLC (Radnor, PA).
Sample Preparation. To prepare the tryptic digests for
peptide mapping analysis, NISTmAb stock sample (100 μg)
was diluted into 5 mM acetic acid and reduced with 5 mM
TCEP-HCl at 80 °C for 10 min. After adjusting the pH to 7.5
using 1 M Tris-HCl (pH 7.5), 10 mM iodoacetamide, and 5 μg
of trypsin (E/S ratio at 1:20) were added, and the sample was
incubated at 37 °C for 3 h. Finally, the solution was acidified
by 1% TFA to quench the digestion. To reduce NISTmAb for
intact mass analysis, 10 mM TCEP-HCl was added to diluted
NISTmAb solution (1 μg/μL), and the sample was incubated
at 50 °C for 30 min. To remove the N-glycans present on
reagent Protein-X, the protein sample was first diluted to 0.25
μg/μL using 5× Rapid PNGase F buffer (containing reducing
agent), and the solution was then incubated at 80 °C for 10
min. Subsequently, Rapid PNGase F was added to the solution,
and the sample was incubated at 50 °C for 30 min.
Modification of the Desolvation Gas. The sheath gas
flow from a Q-Exactive mass spectrometer was redirected to a
Duran pressure plus bottle (SCHOTT North America, Inc.,
Elmsford, NY) through a Canary-Safe Cap (Analytical Sales
and Services, Inc., Flanders, NJ) using 1/8 in. Teflon tubing
(Figure 1). The outgoing tubing from the bottle was then
connected back to the HESI-II probe in a Thermo Scientific
Ion Max ion source. For peptide mapping analysis, 150 mL of
propionic acid (PA) and 50 mL of isopropanol (IPA) were
mixed and transferred into the bottle. For intact mass analysis,
1% triethylamine (TEA) (v/v) in 200 mL of acetonitrile was
Figure 1. Representation of the device for desolvation gas
modification on a Q-Exactive mass spectrometer.
DOI: 10.1021/acs.analchem.8b05846
Anal. Chem. 2019, 91, 3156−3162
Analytical Chemistry
transferred into the bottle. The bottle, containing concentrated
acid or base, was then placed into a polyethylene secondary
container (BEL-ART acid/solvent bottle carrier, Wayne, NJ)
with a 16 mm opening in the top for insertion of tubing. To
disable the modification, the sheath gas tubing was directly
connected to the HESI-II probe without passing through the
device. Photographs of the actual device and the device in
connection to a Q-Exactive mass spectrometer are shown in
Figures S1 and S2 (Supporting Information).
Safety Considerations. For desolvation gas modified
experiments, the sheath gas was set to 15 arbitrary units. A
higher setting might be possible, but the pressure within the
bottle should be measured and made sure not to exceed the
pressure rating of the bottle. A pressure resistant bottle (e.g.,
Duran pressure plus bottle, pressure rating, +1.5 bar) is highly
recommended for this application. A secondary container was
also required for this setup to prevent possible acid or base
spills.
LC−MS Analysis. For peptide mapping analysis of
NISTmAb, aliquots (2 μg) of digests were separated using
an ACQUITY UPLC Peptide BEH C18 column (130 Å, 1.7
μm, 2.1 mm × 150 mm) (Waters, Milford, MA) for online
LC−MS/MS analysis on a Q-Exactive mass spectrometer. For
formic acid (FA)-based analysis, mobile phase A was 0.1% FA
(v/v) in water, and mobile phase B was 0.1% FA in acetonitrile
(ACN). For TFA-based analysis, mobile phase A was 0.05%
TFA (v/v) in water, and mobile phase B was 0.045% TFA (v/
v) in ACN. Detailed LC gradient and MS parameters are
included in the Supporting Information. For intact mass
analysis of reduced NISTmAb and reagent Protein-X, aliquots
(4 μg) of each sample were separated using a BioResolve RP
mAb Polyphenyl column (50 mm × 2.1 mm, 2.7 μm) (Waters,
Milford, MA) for online LC−MS analysis on a Q-Exactive
mass spectrometer. Detailed LC gradient and MS parameters
are included in the Supporting Information.
■ RESULTS AND DISCUSSION
Modification of the desolvation gas was demonstrated on a Q-
Exactive mass spectrometer equipped with a standard Ion Max
ion source. The accessibility of both sheath gas and auxiliary
gas from this instrument made implementation of the
modification straightforward. The device can be conveniently
enabled or disabled within 1 min without any need for system
equilibration. Although it should be feasible to implement
similar modifications to other mass spectrometers from
different vendors, more extensive modifications might be
required.
To improve the peptide mapping analysis by counteracting
the TFA ion suppression, a mixture of propionic acid (PA) and
isopropanol (IPA) at a 3 to 1 ratio was transferred into the
device to deliver the acid vapor into the sheath gas. This recipe
was directly adopted from a previous study, in which Apffel et
al. used a factorial experimental approach and systematically
evaluated the effect of different weak acids and organic solvents
for the “TFA Fix”.10 Although PA and IPA were delivered into
the LC flow via a postcolumn addition in that study, the
proposed mechanism of action and conclusion should be
generally applicable. Briefly, in the presence of PA, the
deprotonated TFA anions, which are the main culprits for ion
suppression by forming ion pairs with analytes, can be
reprotonated and removed from the ESI process through
evaporation due to the relatively high volatility. As a weak acid,
PA does not exhibit a strong ion pairing effect with analytes,
3158
Article
thus facilitating the ionization process. IPA mainly serves as a
carrier for the acid vapor but might also contribute to the
reduced surface tension of the ESI droplets, thus leading to
better desolvation and ionization.20
To evaluate the MS sensitivity improvement as a result of
desolvation gas modification, LC−MS based peptide mapping
analysis was performed using the tryptic digests of NISTmAb
as a testing standard. Both FA- and TFA-containing mobile
phases were tested with the desolvation gas modification either
enabled or disabled for the TFA-based analysis. Figure 2 shows
Figure 2. BPCs from the LC−MS analysis of NISTmAb digests using
the FA-based method (top panel), the TFA-based method (middle
panel), and the TFA-based method with PA/IPA modified desolva-
tion gas (bottom panel). The MS signal from the TFA control
experiment was amplified two times for better visualization. P1−P6
were six representative tryptic peptides from NISTmAb.
the base peak chromatograms (BPCs) from the LC−MS
analysis of the tryptic digests of NISTmAb using a BEH C18
column. It is evident that the separation and retention of
tryptic peptides on the C18 column were significantly different
when FA or TFA were used as the mobile phase modifier. In
general, TFA outperforms FA from the chromatographic
performance perspective, exhibiting better retention of hydro-
philic peptides, better peak shapes and overall higher peak
capacity. For example, a small tryptic peptide, EYK (P1, Figure
2), was not retained on the C18 column when FA was used,
whereas this peak was retained on the same column when TFA
was applied. This feature might improve the sequence coverage
of protein biopharmaceuticals in the peptide mapping analysis.
On the other hand, TFA-based analysis exhibited a significant
loss in MS sensitivity compared to the FA-based analysis.
Subsequently, modification of the desolvation gas with acid
vapor from PA and IPA led to a significant increase in MS
sensitivity for TFA-based analysis.
As demonstrated by six representative tryptic peptides (P1−
P6) of different sizes and retention times (Table S1,
Supporting Information), a 4.9−5.8-fold increase in MS
sensitivity was readily achieved by this simple approach,
regardless of the peptide size and mobile phase composition.
Comparable background noise was also observed before and
after the desolvation gas modification, which subsequently led
to the overall improved signal-to-noise ratio (Figure S3,
Supporting Information). In addition, a subtle increase in
DOI: 10.1021/acs.analchem.8b05846
Anal. Chem. 2019, 91, 3156−3162
Analytical Chemistry Article

charge states was also consistently observed for many tryptic antibodies can be greatly improved by adding a trace amount
peptides after the desolvation gas modification (Table S1, of base additive into mobile phases.16 To evaluate if similar
Supporting Information). This feature might lead to improved improvements can be achieved via desolvation gas modifica-
spectral quality of the tandem MS due to more efficient tion, reduced NISTmAb was used as a test article. Triethyl-
fragmentation of higher charged species, resulting in increased amine (TEA), a frequently used charge reducing reagent, was
identification confidence. Overall, by modifying the desolva- diluted to 1% (v/v) using acetonitrile (ACN) and used to
tion gas with acid vapor using the developed device, a deliver the base vapor into the desolvation gas. After
significantly more sensitive peptide mapping method can be chromatographic separation on a reversed-phase (RP)
readily achieved. The increase in MS sensitivity is of great polyphenol column, the MS spectra of the reduced heavy
value to characterize low-abundance attributes present in chain of the NISTmAb acquired from both a control method
protein biopharmaceuticals, such as PTMs and sequence and from the desolvation gas modified method are shown in
variants. Figure 4. Following desolvation gas modification, a significant
As previously discussed, although postcolumn addition of a
weak acid is an effective and straightforward approach to
counteract TFA ion suppression, it is not suitable for
continuous analysis of large sample sets and consumes a
large quantity of reagents. To evaluate if the new approach can
be applied to continuous analysis and if the improved
sensitivity can be maintained over a long period of time,
peptide mapping analysis of NISTmAb digests were repeatedly
performed over two consecutive days. Again, after enabling the
desolvation gas modification with acid vapor, the increase in
MS sensitivity, compared to the control method, was
immediately achieved and maintained for at least 37
consecutive runs (∼3800 min) as tested (Figure 3). The

Figure 3. MS intensities of six representative tryptic peptides from

NISTmAb digests using control method (runs 1 and 2) and
desolvation gas modified method (runs 3−39). P1−P6 were six
representative tryptic peptides from NISTmAb. The relative standard
deviation (RSD) was calculated based on the MS intensities of each
peptide from 37 modified runs.
consumption of the PA and IPA mixture, under the applied
experimental conditions, was estimated to be 1 mL per hour.
As a result, the developed approach is considered highly
suitable for routine peptide mapping analysis of protein
biopharmaceuticals, in which continuous analysis of large
sample sets might be required.
Besides peptide mapping analysis, the same device can also
be utilized to improve the intact mass analysis of protein
biopharmaceuticals via charge reduction, which is achieved by
modifying the desolvation gas with a weak base. By shifting the
charge state envelope toward a high m/z region, the MS
spectrum can be greatly simplified due to the increased spacing
between two adjacent charge states. In a recent study, Ding et
al. have found the intact mass analysis of reduced monoclonal
3159
Figure 4. Mass spectra of reduced heavy chain of NISTmAb from (a)
the control method and (b) charge reduction method. F, fucose; G,
galactose; Gc: N-glycolylneuraminic acid; M, mannose; Hex, hexose;
A1, biantennary with only one GlcNAc; and A2, biantennary with
both GlcNAcs.
charge reduction on the NISTmAb heavy chain, from charge
states 23+ to 55+ to charge states 14+ to 28+ was achieved.
Close examination of charge state 39+ from the control
method revealed a high level of background noise, which was
likely attributed to the decay of the highly charged heavy chain
species during the MS analysis, such as neutral losses of water
(−18 Da), ammonia (−17 Da), and carboxylic acid (−44 Da).
In contrast, the zoom-in view of charge state 21+ from the
modified method exhibited a greatly improved signal-to-noise
ratio, thus allowing the identification of several low-abundance
glycoforms that were not detected in the control method. The
DOI: 10.1021/acs.analchem.8b05846
Anal. Chem. 2019, 91, 3156−3162
Analytical Chemistry
Figure 5. LC−MS analysis of a highly heterogeneous protein with multiple O-glycans: (a) raw mass spectra acquired using control method (top)
and charge reduction method (bottom) and (b) deconvoluted mass spectrum from the charge reduction method.
deconvoluted mass spectrum (Figure S4, Supporting Informa-
tion) further demonstrated that the improved method could
confidently identify glycoforms present at levels as low as 0.3%
on the heavy chain of NISTmAb.
The dramatic improvement in spectral quality is not only
attributed to the highly simplified spectrum resulting from less
crowded charge state envelopes but more importantly is a
result of stabilized protein analytes. This is consistent with the
well-established knowledge that the collision energy associated
with a protein ion during MS analysis is proportional to its
charge state. Specifically, lower charge states decay slower than
the higher charge states of the same analyte during the analysis
within the Orbitrap. Notably, efforts to mitigate the extensive
decay of highly charged heavy chain species from the control
method were not successful, even after completely removing
the source-induced dissociation (SID) energy. On the
contrary, high levels of SID energy (75 eV used in this
experiment) can be readily tolerated from the charge reduction
method without noticeable decay or fragmentation of the
heavy chain, which further improves the spectral quality via
more efficient desolvation and adduct removal. Similar
improvement in spectral quality was also achieved for the
reduced light chain of the NISTmAb (Figure S5, Supporting
3160
Article
Information). It is worth noting that the application of a charge
reduction strategy to improve MS spectral quality is most
helpful for fully denatured and reduced proteins, as they are
often highly charged (due to increased protein surface areas21)
during ESI-MS analysis and are most susceptible to undesired
fragmentation and decay.
The developed charge reduction method described here,
achieved by modifying the desolvation gas with TEA, can also
be utilized to tackle the high mass heterogeneity present in
complex protein samples. These complex samples might
include various protein reagents that are critically important
to support different assays during the development of protein
biopharmaceuticals. Using intact mass analysis to confirm the
identities of those protein reagents, no matter if they are
obtained from commercial sources or produced in-house, is
frequently required to support the downstream studies
supporting clinical development activities. Some protein
reagents are highly heterogeneous in molecular weight due
to extensive glycosylation, which can present significant
challenges for the standard intact mass method. To
demonstrate the utility of the developed charge reduction
method, a complex reagent protein (Protein-X) with high mass
heterogeneity was used as a test article. Prior to the analysis,
DOI: 10.1021/acs.analchem.8b05846
Anal. Chem. 2019, 91, 3156−3162
Analytical Chemistry
Protein-X was first treated with PNGase F under both
denaturing and reducing conditions, in order to remove the
mass heterogeneity introduced by the presence of N-glycans.
As shown in Figure 5, even after the PNGase F treatment,
Protein-X exhibited a highly convoluted MS spectrum that
cannot be deciphered using a regular intact mass method. This
was likely attributed to the overlapping charge envelopes from
different coeluting mass forms of this molecule at low m/z
region. In contrast, after enabling the desolvation gas
modification with TEA, the MS signal of the same sample
was immediately shifted to the high m/z region and exhibited
much better resolved charge states. The deconvoluted
spectrum (Figure 5b) indicated that this protein was
extensively modified by O-glycans with possibly 10 different
glycosylation sites and 4 different O-glycan forms. Overall,
nearly 35 different mass species from this protein were
confidently identified and assigned. It is worth noting that the
improved spectral quality might also be partly attributed to
stabilized protein ions after charge reduction, as the
polysaccharide moiety on this molecule could be labile under
regular ESI-MS conditions. It is also worth pointing out that
the developed approach might be particularly applicable for the
characterization of PEGylated protein products, as they
typically exhibit a high degree of mass heterogeneity.
Finally, for some other LC−MS based intact mass methods,
where the use of TFA is inevitable to ensure chromatographic
performance, the developed method could also be tested to
counteract TFA ion suppression using PA/IPA modified
desolvation gas. For example, hyphenation of size exclusion
chromatography (SEC) to MS using mobile phases containing
ACN, TFA, and formic acid has been used for reduced mAb
analysis.22 Hydrophilic interaction chromatography coupled to
MS using TFA-containing mobile phases has been used to
study the low molecular weight impurities in mAb samples.23
Application of PA/IPA modified desolvation gas in both
methods led to significant improvement in MS sensitivity of
mAb fragments (e.g., heavy chain, light chain, and smaller
fragments) but not the intact mAb (∼150 kDa) (data not
shown). This observation is consistent with a previously
published study.10 It is thought that a large protein might
accommodate a large number of TFA anions, which cannot be
effectively replaced by PA during the ESI process.
■ CONCLUSIONS
Improved LC−MS based characterization of protein bio-
pharmaceuticals is of great value to better support the drug
development process by providing data with higher confidence
and less ambiguity. We demonstrated a novel approach to
improve the data quality from both peptide mapping analysis
and intact mass analysis via desolvation gas modification using
a simple but effective device. By using PA/IPA modified
desolvation gas, the TFA ion suppression from a typical
peptide mapping method can be effectively mitigated, leading
to improved MS sensitivity. We also showed that the
developed approach can be easily implemented without
changing the LC method and is capable of continuous analysis
of large sample sets, making it particularly suitable for routine
characterization of protein biopharmaceuticals. By using TEA
modified desolvation gas, the new approach could also be
utilized to improve the intact mass analysis of proteins via
charge reduction. Significant improvement in spectral quality
not only allows the detection of minor mass forms otherwise
buried in noise but also enables the mass measurement of
3161
Article
highly heterogeneous proteins. Notably, the modification was
easily implemented via a simple device that was put together
with all commercially available parts at minimal cost and could
be enabled or disabled with little effort. Although we have only
demonstrated the approach on Q-Exactive series instruments,
similar modifications might also be feasible on other
instruments. In addition, the versatility of this device might
be further exploited for other MS-based applications by
changing the ESI conditions. Finally, with the ever-increasing
role played by LC−MS methods in protein biopharmaceutical
characterization, the developed approach can make a deeper
and broader contribution by serving as a low-cost and practical
solution to improve the analytical capability and better support
drug development.
■
*
ASSOCIATED CONTENT
S Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.anal-
chem.8b05846.
LC gradient and MS parameters for peptide mapping
analysis and intact mass analysis; Table S1, comparison
of the MS intensities and averaged charge states of six
representative tryptic peptides from NIST-mAb analyzed
under the three conditions; Figure S1, device in
connection to a Q-Exactive mass spectrometer; Figure
S2, detailed view of the solvent bottle inside the safety
container; Figure S3, comparison of the S/N of six
representative peptides from the LC−MS analysis of
NISTmAb digests using the FA-based, TFA-based, and
TFA-based methods with PA/IPA modified desolvation
gas; Figure S4, deconvoluted mass spectra of reduced
heavy chain of NISTmAb using the control and charge
reduction methods; and Figure S5, deconvoluted mass
spectra of reduced light chain of NISTmAb using the
control and charge reduction methods (PDF)
■
AUTHOR INFORMATION
Corresponding Author
*E-mail: shunhai.wang@regeneron.com.
ORCID
Shunhai Wang: 0000-0003-4055-2187
Author Contributions
†
S.W. and T.X. are co-first authors.
Notes
The authors declare the following competing financial
interest(s): S.W., T.X., A.P.L., Z.H., Y.Y., T.J.D., and N.L.
are full-time employees and shareholders of Regeneron
Pharmaceuticals Inc.
■
ACKNOWLEDGMENTS
This study was sponsored by Regeneron Pharmaceuticals Inc.
The authors would like to thank Ashley Roberts of Scientific
Writing Group for her assistance in editing the manuscript.
■
REFERENCES
(1) Kaltashov, I. A.; Bobst, C. E.; Abzalimov, R. R.; Wang, G.;
Baykal, B.; Wang, S. Biotechnol. Adv. 2012, 30, 210−222.
(2) Rogstad, S.; Faustino, A.; Ruth, A.; Keire, D.; Boyne, M.; Park, J.
J. Am. Soc. Mass Spectrom. 2017, 28, 786−794.
(3) Dick, L. W., Jr.; Mahon, D.; Qiu, D.; Cheng, K. C. J. Chromatogr.
B: Anal. Technol. Biomed. Life Sci. 2009, 877, 230−236.
DOI: 10.1021/acs.analchem.8b05846
Anal. Chem. 2019, 91, 3156−3162
Analytical Chemistry Article

(4) Bongers, J.; Cummings, J. J.; Ebert, M. B.; Federici, M. M.;

Gledhill, L.; Gulati, D.; Hilliard, G. M.; Jones, B. H.; Lee, K. R.;
Mozdzanowski, J.; Naimoli, M.; Burman, S. J. Pharm. Biomed. Anal.
2000, 21, 1099−1128.
(5) Mouchahoir, T.; Schiel, J. E. Anal. Bioanal. Chem. 2018, 410,
2111−2126.
(6) Zeck, A.; Regula, J. T.; Larraillet, V.; Mautz, B.; Popp, O.;
Gopfert, U.; Wiegeshoff, F.; Vollertsen, U. E.; Gorr, I. H.; Koll, H.;
Papadimitriou, A. PLoS One 2012, 7, No. e40328.
(7) Shibue, M.; Mant, C. T.; Hodges, R. S. J. Chromatogr A 2005,
1080, 68−75.
(8) Annesley, T. M. Clin. Chem. 2003, 49, 1041−1044.
(9) Shou, W. Z.; Naidong, W. J. Chromatogr. B: Anal. Technol.
Biomed. Life Sci. 2005, 825, 186−192.
(10) Apffel, A.; Fischer, S.; Goldberg, G.; Goodley, P. C.; Kuhlmann,
F. E. J. Chromatogr A 1995, 712, 177−190.
(11) Chen, J.; Liu, Z.; Wang, F.; Mao, J.; Zhou, Y.; Liu, J.; Zou, H.;
Zhang, Y. Chem. Commun. (Cambridge, U. K.) 2015, 51, 14758−
14760.
(12) Lauber, M. A.; Koza, S. M.; McCall, S. A.; Alden, B. A.; Iraneta,
P. C.; Fountain, K. J. Anal. Chem. 2013, 85, 6936−6944.
(13) Kharlamova, A.; Prentice, B. M.; Huang, T. Y.; McLuckey, S. A.
Anal. Chem. 2010, 82, 7422−7429.
(14) Kharlamova, A.; McLuckey, S. A. Anal. Chem. 2011, 83, 431−
437.
(15) Zhu, Y.; Nugent, K. D. Method and apparatus for nano-capillary/
micro electrospray for use in liquid chromatography-mass spectrometry.
U.S. Patent 8,227,750, July 24, 2012.
(16) Ding, W.; Qiu, D.; Bolgar, M. S.; Miller, S. A. Anal. Chem. 2018,
90, 1560−1565.
(17) Pacholarz, K. J.; Barran, P. E. EuPa Open Proteomics 2016, 11,
23−27.
(18) Hopper, J. T.; Sokratous, K.; Oldham, N. J. Anal. Biochem.
2012, 421, 788−790.
(19) Huang, L.; Gough, P. C.; Defelippis, M. R. Anal. Chem. 2009,
81, 567−577.
(20) Iavarone, A. T.; Jurchen, J. C.; Williams, E. R. J. Am. Soc. Mass
Spectrom. 2000, 11, 976−985.
(21) Kaltashov, I. A.; Mohimen, A. Anal. Chem. 2005, 77, 5370−
5379.
(22) Liu, H.; Gaza-Bulseco, G.; Chumsae, C. J. Am. Soc. Mass
Spectrom. 2009, 20, 2258−2264.
(23) Wang, S.; Liu, A. P.; Yan, Y.; Daly, T. J.; Li, N. J. Pharm.
Biomed. Anal. 2018, 154, 468−475.

3162 DOI: 10.1021/acs.analchem.8b05846

Anal. Chem. 2019, 91, 3156−3162