Abstract
Background: Epitope mapping is an increasingly important aspect
of biotherapeutic and vaccine development. Recent advances in
therapeutic antibody design and production has enabled candidate mAbs to
be identified at a rapidly increasing rate resulting in a significant
bottleneck in the characterization of ‘structural’ epitopes, that are
challenging to determine using existing high throughput epitope mapping
tools. Here, Hydrogen/Deuterium Exchange Mass Spectrometry (HDX-MS)
epitope screening workflow was introduced that is well suited for
accelerated characterization of epitopes with a common antigen.
Main methods and major results: The method is demonstrated on
set of 6 candidate mAbs targeting Pertactin (PRN). Using this approach,
five of the six epitopes was unambiguously determined using two HDX
mixing timepoints in 24 hours total run time, corresponding to
substantial decrease in the instrument time required to map a single
epitope using conventional HDX workflows.
Conclusion: An accelerated HDX-MS epitope screening workflow
was developed. The two-timepoint ‘screening’ workflow mapped all six
mAbs and generated high confidence epitopes for five of the six mAbs
assayed. The substantial improvement in the rate of data collection can
advance HDX-MS for higher throughput investigations supporting the
ability to evaluate a broader number of mAb candidates at an earlier
stage of vaccine development.
Introduction
The use of antibody-based therapeutics and reagents has seen
unprecedented growth in recent years, with market projections estimated
at a combined $212 billion USD annually by
2022.[1] Developments in rapid isolation and
enrichment of monoclonal antibodies (mAbs) from challenged serum has
simplified the identification of broadly neutralizing and immunogenic
antibodies, resulting in a glut of new therapeutic (or vaccine relevant)
mAb candiates.[2,3] For vaccines in particular,
mAbs selection, with desirable binding attributes, is essential for
rational design of in vitro potency assays. The challenging
aspect now lies in epitope mapping, where the antibody-antigen complex
is characterized by the interacting amino acids on the antigen. These
amino acids may correspond to a continuous primary sequence (linear
epitope) or a ‘surface’ that is non-contiguous in the primary sequence
and results from the 3D structure of the antigen (conformational
epitope).[4] Methodologies that can identify both
the epitope and binding mechanism are not only desirable but necessary,
as pharmaceutical companies must now identify and disclose their
epitopes in regulatory filings.[5] Furthermore,
pinpointing structurally where the mAb binds to the antigen can provide
a direct link to the functional biology associated with neutralization,
allowing for an evidence-based conclusion to identify the most valuable
mAbs for lead selection.[6],[7]
Currently there are a number of available techniques that can detect
epitopes either with high throughput or high resolution. Frequently,
more than one method is implemented to confirm
identification.[8–10] Synthetic peptide arrays,
which use segments of the antigen sequence to bind a mAb target, have
made remarkable strides towards high-throughput, whole-proteome, peptide
binding assays. Phage display can be used in a similar manner to display
whole libraries of peptides on micro-arrays.[11]Neither of these technologies are ideal to identify conformational
epitopes and provide little direct information on the binding mechanism.
Although, often a time consuming and costly technique,
co-crystallization of the mAb-antigen complex followed by X-ray
crystallography provides the highest resolution snapshot of an
antibody/antigen binding surface and is widely accepted as the ’gold
standard’ method for epitope/paratope mapping. Residues within 4
angstroms are considered as contacts, but no information on the bond
strength of contributing residues or conformational dynamics can be
obtained without screening a panel of point
mutants.[12] Cryo-electron microscopy is becoming
a widely adopted alternative to X-ray crystallography due to its
increasingly comparable resolution to X-ray diffraction while offering
an inherent increase in throughput.[13] Despite
steadily improving high-resolution techniques, it is still desirable
that these methods be complimented by approaches that can assess both
the structural and dynamic aspects of protein complexes and epitopes in
solution. Nuclear magnetic resonance (NMR) spectroscopy is in principle
capable of providing dynamic information through a combination of
classical structural and relaxation experiments, however this technique
requires isotopically labeled material and larger protein complexes
(>100 KDa) are difficult to
analyse.[14,15]
Hydrogen/Deuterium Exchange Mass Spectrometry (HDX-MS) is an alternative
that can characterize both the structural and dynamic aspects of
antibody-antigen complexes in-solution.[16–23] It
also provides a moderate increase in throughput, while substantially
reducing the requirements for purified biological material. Apart from
the structural aspect, HDX-MS has also been used to report binding
affinities for protein-protein interactions, highlighting the
versatility of HDX technology.[24] However,
conventional HDX-MS epitope mapping workflows are typically exceedingly
lengthy, requiring up to 24 hours of experiment time to generate a
single putative epitope. In the current work, we report several
refinements to the conventional HDX-MS workflow that substantially
increase throughput, while maintaining a high degree of accuracy and
precision with epitope assignments using a set of six antibodies
targeting pertactin.
Experimental Section
HDX-MS
The HDX experiments were performed as previously described with few
modifications.[16,25]For antigen-only HDX experiments,
15 µM of PRN was used and for the complex, equimolar of 15 µM of PRN and
mAb were incubated in buffer E (10 mM potassium phosphate buffer pH
7.5). 7.5 µL of the sample was mixed with 32.5 µL of deuterated buffer,
buffer L (10 mM potassium phosphate pD 7.5) at 25°C. In the mapping
workflow, HDX was conducted at 5 mixing timepoints: 20 sec, 2 min, 10
min, 60 min, and 240 min whereas in the screening workflow, 2 mixing
timepoints: 10 min and 30 min were used. All the mapping experiments
were conducted in triplicates whereas the screening experiments were
performed in duplicates. Mixing in this ratio will attain maximum
deuteration of ~80%. 40 µL of the reaction was quenched
in 40 µL of chilled quenching buffer QH (100 mM potassium phosphate,
7.5M GdnHCl, 0.5M TCEP, pH 2.5, 0°C). The quenched samples were then
mixed with 80 µL of 0.1% formic acid with subsequent injection of 100
µL of the sample into the nanoACQUITY UPLC HDX module containing
protease column for digestion, along with a Waters BEH C18 guard column
and ACQUITY CSH C18 analytical column for desalting and separation
respectively. The samples were allowed to digest/desalt for 3 min in
0.1% formic acid at a flow rate of 100 µL/min with subsequent
separation in the analytical column at a flow rate of 40 µL/min using a
7 min gradient starting from 1% to 60% of water/acetonitrile in 0.1%
formic acid. The eluted peptides are detected using Waters Synapt G2-Si
mass spectrometer (MA, USA) with acquisition of mass/charge (m/z )
range between 300-1700 GluFib (785.8426 m/z) was used as a lock mass
solution to maintain mass calibration of <10 ppm. Blank
injections (0.1% FA in LC-MS water) were incorporated in between HDX
runs to prevent potential carry-over from previous runs.
Peptide identification
and HDX analysis
Non-deuterated (0 sec time control) digested peptides were identified by
mass accuracy and MS/MS fragmentation (Waters HDMSe function) with the
use of software, ProteinLynx Global Server (Waters Corp., MA, USA).
Digestion was carried out as same manner as the HDX runs except
non-deuterated buffer was used. Fragmentation (HDMSe) was activated in
the transfer cell by ramping up the collision energy from 15-65 V.
The quantitative analysis of the HDX data was carried out as previously
described with a few modifications.[26,27]The differences for identical
peptides between the two states at all the HDX timepoints were
calculated and summed. A positive or negative value indicates a decrease
or increase in deuterium uptake upon complexation respectively whereas a
neutral value indicates no difference. For the mapping workflow, if the
differences of absolute sum of the 5 HDX timepoints differences exceeds
1.1 Da and surpass 3 times the standard deviation (σ), then it is
considered statistically significant.[27] For
screening workflow, if the differences of the summed HDX timepoints
exceed 0.5 Da and surpass 3σ, then it is considered significant as well.
σ is defined as the propagated standard deviation of the replicates at
mixing timepoints for both states (PRN and PRN/mAb complex). The HDX
processing parameters and raw data is attached as supplementary
information adhering to the HDX community
standards.[28]
The assigned epitopes were mapped onto a three-dimensional structural
model generated through homology modelling (Phyre2) using a pre-existing
crystal structure, PDB code 1DAB, as a
template.[29] Homology model was used as the
existing crystal structure lacked the C-terminal linker region. A
structural comparison is shown in Figure S1.
Biolayer Interferometry (BLI) peptide/mAb binding experiments.
PRN linear epitope validation experiments were conducted using Biolayer
Interferometry (BLI) on a ForteBio Octet RED384 system. All experiments
were conducted using ForteBio kinetics buffer (10x kinetic buffer
diluted down to 1x using DPBS). Overlapping peptides spanning the
predicted epitope site of the mAb were synthesized and biotinylated at
its NH2 terminal (Thermo Fisher Scientific).
Furthermore, 2 additional N-terminal biotinylated peptides were
synthesized for testing each mAb as negative controls, one that was 100
amino acids upstream and one 100 amino acids downstream from the
predicted epitope binding site. The peptides (20 µg/mL) were captured on
streptavidin coated biosensors. Sensograms were recorded when
peptide-loaded biosensors were incubated with the mAb (20 µg/mL in
ForteBio kinetics buffer) and allowed to associate.
Results and discussion
Protease digestion column selection.
Prior to the HDX experiments, a reference peptide library must be
generated. This is achieved by subjecting the antigen to proteolysis in
the acid-protease digestion column. To maximize relevant information for
HDX experiments, the digestion method has to be optimized for sequence
coverage and sequence overlap.[30] Digest-peptides
should be 5 – 10 amino acids in length with sequence redundancy, where
overlapping peptides can improve the spatial resolution of the data and
increase confidence in the measurements.[30] There are several
commercially available acid protease options including: pepsin, protease
type XIII from Aspergiluus saitoi, protease type XVIII fromRhizhopus and nepenthesin from Nepenthes genus of
carnivorous plants among others.[31–33] The
digestion efficiency of each protease is dependent on the amino acid
properties. For example, pepsin preferentially cleaves bulky hydrophobic
residues whereas protease XIII tends to hydrolyze peptide bond with
basic amino acids on the C-terminal side.[34,35]For this study, pepsin and protease XIII/pepsin columns were evaluated
for digestion efficiency on the model protein pertactin (PRN).
Sequence coverage resulting from protease XIII/pepsin digestion and
pepsin digestion are shown in Figure S2. Peptides generated by pepsin
cleavage resulted in higher sequence coverage (~88% of
high-quality peptides on average) compared to protease XIII/pepsin
column (~64% on average). While, the protease
XIII/pepsin column resulted in slightly higher number overlapping
peptides as shown by higher redundancy (2.57 versus 2.25), it also left
several gaps in the coverage map. The missing sequences resulting from
both protease columns were mapped onto three-dimensional structure is
shown in Figure S2. The higher sequence coverage from pepsin can be
explained by prevalence of hydrophobic residues in
PRN.[36] Pepsin was chosen for the subsequent HDX
experiments since it resulted in higher sequence coverage with
sufficient sequence redundancy compared to protease XIII/pepsin. Note
that fully optimized protease and digestion protocols in epitope mapping
are antigen specific, however, there are approaches that will work well
for a wide range of antigens and so complete optimization is not always
necessary.
Epitope Mapping Workflow.
Six mAbs were evaluated for ‘mapping’ and ‘screening’ workflows: mAb
3-4, 3-16, 3-3, 3-21, 3-5, and 1-16. In the ‘mapping’ workflow, five HDX
mixing timepoints, and three replicates, were collected for each
antibody. To identify potential binding sites, HDX uptake differences
between two states (free PRN versus PRN/mAb complex) were required to
exceed the criteria described in the materials and methods section
(i.e., difference > 1.1 Da and 3σ).
Differential analysis of the free antigen and the PRN/mAb 1-16 complex
revealed one peptide, residues 557-568 surpassing the significance
threshold (~1.4 Da difference) as shown in Figure 1. The
representative kinetic plot of peptide 557-568 is provided in Figure S3.
The two partially overlapping peptides (555-561) and (558-578) also
exhibited decreases in deuterium uptake upon complex formation, though
they did not pass the significance threshold (~0.3 Da
and ~0.3 Da respectively) making residues 562-568 the
representative epitope as shown in the heatmap (Figure S4). The heatmap
analysis is performed by averaging the relative deuterium uptake of the
overlapping peptides, hence providing better resolution of the epitope
at the peptide level. This determination provides an example of how
overlapping sequences can further resolve the HDX at shorter peptic
fragments and potentially to individual amides. The results indicate
that the epitope is linear, which was also confirmed using Biolayer
Interferometry (BLI) binding studies, where representative peptides
containing the linear epitope sequence were evaluated with mAb 1-16. An
association curve was generated confirming the binding of the peptide
containing the epitope with mAb 1-16 (Figure S5). The assigned epitope
was mapped onto the three-dimensional structure of PRN, as shown in
Figure 1. This C-terminal region is known to play a critical role in the
secretion of PRN through the bacterial membrane as well as providing a
template for proper folding of the beta-helix structure of
pertactin.[37]
Analysis of the PRN/mAb 3-3 complex also revealed a binding site at the
C-terminus. Two peptides (575-591 and 579-591) localized on the
C-terminal linker region exhibited the most protection from deuterium
update (~5 Da on average) as shown in Figure 1. The
representative kinetic plot of peptide 575-591 is shown in Figure S3.
The adjacent peptides, 575-587 and 579-587, however, displayed minimal
changes (below 0.25 Da) enabling identification of 588-591 as the
epitope. This epitope was also confirmed by BLI binding studies as shown
in Figure S6. Interestingly, four other peptides across the sequence
exhibited a significant increase in deuterium uptake upon complexation,
including residues 35-46, 194-212, 231-245, and 304-313. Although it is
unclear why this mAb induced such structural changes, allosteric effects
upon antibody/antigen complexation are not a rare
phenomenon.[38]
For the PRN/mAb 3-5 complex, only one peptide, 234-244, presented an
observable decrease in deuterium uptake during antigen complexation that
passed the significance criteria (~1.5 Da) as shown in
Figure 1. The representative kinetic plot of peptide 234-244 is provided
in Figure S3. The rest of the protein exhibited negligible differences.
A BLI binding study was conducted on mAb 3-5 and peptide 234-244.
However, no binding was observed. When PRN was heat denatured at 70°C
for various incubation times, there was also reduction in binding as
function of time, shown in Figure S7. This result suggest that the
epitope is conformational where only one of the binding sites (peptide
234-244) was detectable by HDX (since a linear epitope would not be
affected by heat denaturation). Upon analyzing the missing sequences on
the 3D structure shown in Figure S2, the loop coverage close to peptide
234-244 was missing. Most likely, this loop was also involved in the
binding, but was excluded due to non-detection in the protein digest.
Peptide 234-244 is part of the polymorphic region of PRN (residues
~230-260) where variation has been linked to immune
evasion.[39,40] Hence binding to this region may
impact immune escape functionality.
Three mAbs described thus far bind to PRN on continuous sections of the
sequence, mainly comprised of loops or disordered regions localized on
the surface of the protein, as shown in Figure 1. These results align
with other studies where epitopes were most commonly found to be in
surface exposed, less-structured regions of
antigens.[41,42] Binding to linear, continuous
epitope does not depend on the tertiary fold of the antigen but is
linked only to the binding sequence.[4,43] Hence
antibodies binding to a linear epitope can be tested using a denatured
antigen or an unstructured peptide containing the epitope
sequence.[4,43] Conformational epitopes, however,
are dependent on the fold of the antigen. As such, binding is highly
sensitive to the correct tertiary structure of the antigen,
corresponding to at least one configuration of the ‘native’
conformational ensemble.[4,43] It has been
estimated that ~90% of the antibodies recognize
conformational epitopes.[44]
Differential analysis of the PRN/mAb 3-16 complex detected two peptides,
36-46 and 56-67, in the N-terminus with deuterium uptake differences of
4.5 Da and ~3 Da respectively (Figure 1). Structurally,
these two peptides are in close proximity to one another, indicating
characteristics of a conformational epitope. The kinetic plots for these
two peptides are provided in Figure S3. All other peptides across the
sequence showed negligible differences except for increase in deuterium
uptake in peptide 231-245 which is part of the polymorphic regions of
PRN responsible for immune evasion, similar to mAb
3-3.[39,40] The N-terminus is immunogenic and some
of the known immunogenic epitopes in the N-terminus are known to be
masked by the polymorphic region.[45,46] Increased
flexibility of the polymorphic region could indicate “unmasking” of
the epitope in the N-terminus for efficient binding of mAb
3-16.[46]
In the PRN/mAb 3-21 complex, five peptides from the N-terminal region,
35-46, 47-54, 55-67, 68-78 and 109-116, exhibited decreases in deuterium
that surpassed the significance criteria suggesting a conformational
epitope (Figure 1). This epitope included the sequences detected for mAb
3-16 (peptides 36-46 and 56-67) which suggests that these two mAbs share
a common epitope. However, when an epitope binning experiment (i.e., a
competitive binding assay) was performed using biolayer interferometry,
both mAbs were able to bind to PRN simultaneously, suggesting different
epitope location. The remaining sequences: 47-54, 68-78 and 109-116 were
then mapped onto the protein structure. Residues 47-54 and 109-116 are
in close proximity whereas 68-78 is localized on the opposite face of
the antigen, same side as residues 36-46 and 56-67, as shown in Figure
S8. Hence a tentative assignment for a conformational epitope
corresponding to peptides 47-54 and 109-116 was made. The kinetic plots
for two representative peptides, 47-54 and 108-116 are shown in Figure
S3.
Finally, epitope mapping analysis of PRN/mAb 3-4 revealed a decrease in
deuterium uptake localized on the C-terminal end of PRN specifically
peptides 420-436, 455-464, 477-492, and 505-517 as shown in Figure 1.
The kinetic plots for these four peptides are shown in Figure S3. These
sites are all in close proximity to each other, with a majority of
residues in loop regions. It is plausible that some of these residues
took up less deuterium due to steric hindrance effects, since a large
surface area was impacted.[47] Nevertheless, the
data indicates that this is a conformational epitope. This epitope is
localized near another PRN polymorphic variation region (residues
545-566).[39,40] This PRN polymorphic region
contains tandemly repeated sequences (Pro-Gln-Pro-) that have been
implicated in immune evasion by either insertion or deletion of repeat
units. Hence mAb binding to this epitope may directly impact the immune
escape functionality of PRN by affecting the flexibility of the
polymorphic region such that hidden epitopes downstream of C-terminus
are unmasked.[46] Similar to the linear epitopes,
all the identified conformational epitopes were localized on
surface-exposed loops and/or sheet-loop
structures.[42,47]
Screening workflow.
In developing the screening workflow, two HDX timepoints (10 min or 30
min) were acquired. Single timepoints were also assessed for feasibility
of epitope detection. HDX results for individual timepoints are shown in
Figure S9. For all mAbs, results from the 30 min timepoint were more
ambiguous than for the 10 min timepoint, as demonstrated by weaker or no
deuterium-uptake signals. For example, PRN/mAb 1-16 and PRN/mAb 3-5
complexes did not show clear epitope detection at 30 minutes labeling
time, although one peptide in the mAb 1-16 epitope (575-587) and one in
the mAb 3-5 epitope (234-244) showed the highest uptake difference
(qualitatively) while surpassing the 3σ. Nonetheless, these uptake
differences were comparable to those observed in other peptides across
the antigen sequence. In the PRN/mAb3-3 complex, there were two
peptides, 229-236 and 579-591, that exhibited HDX differences at 30
minutes of labelling. However, these two peptides were spatially distant
and defining an epitope against the protein structure was difficult.
Analysis of 10 minutes HDX timepoint showed clear HDX differences
allowing for slightly improved epitope assignment. However, the results
were often only marginally substantial and in some cases were
indistinguishable from differences occurring across the sequence
(i.e ., ‘noise’ in HDX difference data). Ultimately, the HDX
differences for both timepoints (10 and 30 min) were summed in the
screening workflow, which amplified non-random differences to the
minimum extent required to enable confident epitope assignments, as
shown in Figure 2. As with the ‘mapping’ workflow, a significant change
was defined as uptake differences exceeding 0.5 Da in both HDX
timepoints and surpassing 3σ from at least two technical replicates.
The HDX differences observed using the in the screening workflow were
lower than those acquired from the full time-course, mapping workflow
(an unsurprising consequence of adding together fewer timepoint
measurements). These changes can be seen in deuterium uptake (y-axis) of
the Difference plots comparing mapping results in Figure 1 to screening
results in Figure 2. The primary benefit to the screening approach is
that it provides substantial enhancements to throughput; all six mAb/PRN
complexes were screened in the same amount of time required to ‘map’ a
single antibody/antigen interaction using the full time-course workflow.
The acquisition and analysis for all states of PRN complexed with the
different mAbs, was performed simultaneously in one dataset, as opposed
to data collection in separate experiment/days. Thus, this type of
analysis can also more easily identify artifact peptides/signals.
However, the caveat for simultaneous analysis is that it requires
consistent, high quality peptide signals for all the PRN states. Even if
a subset of states shows poor peptide signal response, that peptide must
be excluded from analysis, which may impact sequence coverage, or
sequence redundancy. Nonetheless in the screening dataset presented
here, the sequence coverage is ~88%, which is identical
to the ‘mapping’ workflow. The summary of epitopes identified from the
‘screening’ workflow is shown in Table S1.
Analysis of the PRN/mAb 3-4 complex found residues 427-436, 455-464, and
477-492 exhibiting the greatest uptake differences (Figure 2). This
represents a substantial overlap with regions identified in the mapping
workflow (Figure 1), however; the screening data lacked sequence
coverage between 504 and 541, preventing a direct comparison in the 505
– 517 region, which was assigned to the epitope in the mapping
workflow. For the PRN/mAb 3-16 complex, regions of 35-46 and 56-67
exhibited the most difference (Figure 2), perfectly superimposing the
differences from the mapping workflow (Figure 1). PRN/mAb 3-5 also
exhibited the greatest HDX difference in the 237-244 region (Figure 2),
in alignment with the epitope identified in the mapping workflow (region
234-244, Figure 1). In the PRN/mAb 3-3 complex, only one peptide near
the C-terminus, 579-591, showed a significant difference, which was also
consistent with the mapping workflow (Figure 1 and Figure 2). The
PRN/mAb 1-16 complex did not exhibit a large difference in uptake upon
complexation in either workflow, which made the data interpretation
challenging. Nevertheless, one peptide, 558-574, surpassed the 3σ
significance criterion in the screening data (Figure 2). While too large
to provide good localization of an epitope, this sequence still contains
the segment identified in the mapping workflow (562-568, Figure 1). For
the PRN/mAb 3-21complex, the screening experiment identified residues
35-67 and 108-117 as the most affected regions (Figure 2), while peptide
68-78 showed negligible difference (refer to Figure S8 for peptide 68-78
localization on the structure). While not identical, this is similar to
the mapping workflow, which identified segments 47 – 54 and 109-116 to
be the epitope (Figure 1).
Overall, the epitopes identified through the screening workflow were in
good agreement with the mapping workflow for both conformational and
linear epitopes. The mapping workflow utilizes many time points across
wide range (seconds – hours). The advantage of using such a wide range
is that it can improve confidence in the assignments of epitopes,
especially for cases where summing differences across multiple
timepoints is required to achieve statistical certainty based on strict
significance criteria. Furthermore, the kinetics data acquired in the
mapping workflow can also provide additional information on the kinetics
of the protein complex dissociation relative to the rate of the
deuterium uptake.[48] The screening workflow, on
the other hand, takes two snapshots, which offers at best highly limited
rate information. However, when the aim is specifically to determine the
epitope and not to characterize the antibody/antigen interaction in
detail, the screening workflow balances confidence in the epitope
assignment with substantially improved throughput.
Concluding remarks
In this work, we developed an accelerated workflow for epitope mapping
using hydrogen deuterium exchange mass spectrometry and demonstrated it
using six mAbs targeting the pertussis antigen PRN. The classical five
timepoint ‘mapping’ workflow required 144 hours of instrument time to
map all six epitopes and generated high confidence epitopes for five of
the six mAbs assayed. The two-timepoint ‘screening’ workflow required 24
hours of instrument time to map all six mAbs and generated high
confidence epitopes for five of the six mAbs assayed. The remaining
epitope was correctly identified, but not fully mapped (mAb 3-5) in both
workflows. The high degree of correlation between these two methods
suggests that, at least for the strict purpose of identifying epitopes,
acquiring more than two time points may be of little value. Given the
increasingly rapid rate at which new therapeutic and test mAbs are being
generated, we are confident this approach, which allows a seven-fold
improvement in the rate of data collection, will be an attractive
alternative, opening the door to medium throughput epitope screening by
HDX-MS.
The accelerated workflow introduced here will provide a means to advance
HDX-MS for higher throughput investigations in vaccine and drug
development, supporting the ability to evaluate a broader number of mAb
candidates at an earlier stage of development.
ACKNOWLEDGMENT
The authors would like to thank Marin Ming, Jin Su, and Beata Gajewska
from the Immunology Platform, Analytical Sciences of Sanofi Pasteur,
Toronto, for their invaluable support and helpful discussions for this
project.
CONFLICTS OF INTEREST
Shaolong Zhu (SZ) and Peter Liuni (PL) were industrial postdoctoral
fellows when experiments in this study were performed. Camille Houy (CH)
was part of the Volunteer for International Experience (VIE) program at
Sanofi Pasteur Lmt Canada at the time this study was performed. SZ and
CH are now employees of Eurofins PSS. PL is now an employee of Sciex.
Tricia Chen and D. Andrew James are employees of Sanofi Pasteur Lmt
Canada. Derek J. Wilson declares no competing interests.
REFERENCES
[1] Carter, P. J., Lazar, G. A., Next generation antibody drugs:
pursuit of the ‘high-hanging fruit’. Nature Reviews Drug
Discovery 2017 , 17:197–223.
[2] Wine, Y., Boutz, D. R., Lavinder, J. J., Miklos, A. E., Hughes,
R. A., Hoi, K. H., Jung, S. T., Horton, A. P., Murrin, E. M., Ellington,
A. D., Marcotte, E. M., Georgiou, G., Molecular deconvolution of the
monoclonal antibodies that comprise the polyclonal serum response.PNAS 2013 , 110:2993–2998.
[3] Corti, D., Langedijk, J. P. M., Hinz, A., Seaman, M. S.,
Vanzetta, F., Fernandez-Rodriguez, B. M., Silacci, C., Pinna, D.,
Jarrossay, D., Balla-Jhagjhoorsingh, S., Willems, B., Zekveld, M. J.,
Dreja, H., O’Sullivan, E., Pade, C., Orkin, C., Jeffs, S. A.,
Montefiori, D. C., Davis, D., Weissenhorn, W., McKnight, Á., Heeney, J.
L., Sallusto, F., Sattentau, Q. J., Weiss, R. A., Lanzavecchia, A.,
Analysis of Memory B Cell Responses and Isolation of Novel Monoclonal
Antibodies with Neutralizing Breadth from HIV-1-Infected Individuals.PLOS ONE 2010 , 5:e8805.
[4] Van Regenmortel, M. H. V., What is a B-cell epitope?Methods Mol. Biol. 2009 , 524:3–20.
[5] Davidson, E., Doranz, B. J., A high-throughput shotgun
mutagenesis approach to mapping B-cell antibody epitopes.Immunology 2014 , 143:13–20.
[6] Howell, K. A., Qiu, X., Brannan, J. M., Bryan, C., Davidson, E.,
Holtsberg, F. W., Wec, A. Z., Shulenin, S., Biggins, J. E., Douglas, R.,
Enterlein, S. G., Turner, H. L., Pallesen, J., Murin, C. D., He, S.,
Kroeker, A., Vu, H., Herbert, A. S., Fusco, M. L., Nyakatura, E. K.,
Lai, J. R., Keck, Z.-Y., Foung, S. K. H., Saphire, E. O., Zeitlin, L.,
Ward, A. B., Chandran, K., Doranz, B. J., Kobinger, G. P., Dye, J. M.,
Aman, M. J., Antibody Treatment of Ebola and Sudan Virus Infection via a
Uniquely Exposed Epitope within the Glycoprotein Receptor-Binding Site.Cell Rep 2016 , 15:1514–1526.
[7] Saphire, E. O., Schendel, S. L., Fusco, M. L., Gangavarapu, K.,
Gunn, B. M., Wec, A. Z., Halfmann, P. J., Brannan, J. M., Herbert, A.
S., Qiu, X., Wagh, K., He, S., Giorgi, E. E., Theiler, J., Pommert, K.
B. J., Krause, T. B., Turner, H. L., Murin, C. D., Pallesen, J.,
Davidson, E., Ahmed, R., Aman, M. J., Bukreyev, A., Burton, D. R.,
Crowe, J. E., Davis, C. W., Georgiou, G., Krammer, F., Kyratsous, C. A.,
Lai, J. R., Nykiforuk, C., Pauly, M. H., Rijal, P., Takada, A.,
Townsend, A. R., Volchkov, V., Walker, L. M., Wang, C.-I., Zeitlin, L.,
Doranz, B. J., Ward, A. B., Korber, B., Kobinger, G. P., Andersen, K.
G., Kawaoka, Y., Alter, G., Chandran, K., Dye, J. M., Systematic
Analysis of Monoclonal Antibodies against Ebola Virus GP Defines
Features that Contribute to Protection. Cell 2018 ,
174:938-952.e13.
[8] Abbott, W. M., Damschroder, M. M., Lowe, D. C., Current
approaches to fine mapping of antigen–antibody interactions.Immunology 2014 , 142:526–535.
[9] Buus, S., Rockberg, J., Forsström, B., Nilsson, P., Uhlen, M.,
Schafer-Nielsen, C., High-resolution Mapping of Linear Antibody Epitopes
Using Ultrahigh-density Peptide Microarrays. Mol Cell Proteomics2012 , 11:1790–1800.
[10] Reineke, U., Schutkowski, M., Epitope mapping protocols2009 Humana Press, New York.
[11] Rojas, G., Tundidor, Y., Infante, Y. C., High throughput
functional epitope mapping: Revisiting phage display platform to scan
target antigen surface. MAbs 2014 , 6:1368–1376.
[12] Chakraborti, S., Prabakaran, P., Xiao, X., Dimitrov, D. S., The
SARS Coronavirus S Glycoprotein Receptor Binding Domain: Fine Mapping
and Functional Characterization. Virology Journal 2005 ,
2:73.
[13] Merk, A., Bartesaghi, A., Banerjee, S., Falconieri, V., Rao,
P., Davis, M. I., Pragani, R., Boxer, M. B., Earl, L. A., Milne, J. L.
S., Subramaniam, S., Breaking Cryo-EM Resolution Barriers to Facilitate
Drug Discovery. Cell 2016 , 165:1698–1707.
[14] Blech, M., Peter, D., Fischer, P., Bauer, M. M. T., Hafner, M.,
Zeeb, M., Nar, H., One target-two different binding modes: structural
insights into gevokizumab and canakinumab interactions to
interleukin-1β. J. Mol. Biol. 2013 , 425:94–111.
[15] Zuiderweg, E. R. P., Mapping Protein−Protein Interactions in
Solution by NMR Spectroscopy. Biochemistry 2002 ,
41:1–7.
[16] Zhu, S., Liuni, P., Ettorre, L., Chen, T., Szeto, J., Carpick,
B., James, D. A., Wilson, D. J., Hydrogen–Deuterium Exchange Epitope
Mapping Reveals Distinct Neutralizing Mechanisms for Two Monoclonal
Antibodies against Diphtheria Toxin. Biochemistry 2019 ,
58:646–656.
[17] Malito, E., Biancucci, M., Faleri, A., Ferlenghi, I.,
Scarselli, M., Maruggi, G., Lo Surdo, P., Veggi, D., Liguori, A.,
Santini, L., Bertoldi, I., Petracca, R., Marchi, S., Romagnoli, G.,
Cartocci, E., Vercellino, I., Savino, S., Spraggon, G., Norais, N.,
Pizza, M., Rappuoli, R., Masignani, V., Bottomley, M. J., Structure of
the meningococcal vaccine antigen NadA and epitope mapping of a
bactericidal antibody. Proc Natl Acad Sci U S A 2014 ,
111:17128–17133.
[18] Huang, R. Y.-C., Krystek, S. R., Felix, N., Graziano, R. F.,
Srinivasan, M., Pashine, A., Chen, G., Hydrogen/deuterium exchange mass
spectrometry and computational modeling reveal a discontinuous epitope
of an antibody/TL1A Interaction. MAbs 2018 , 10:95–103.
[19] Bertoldi, I., Faleri, A., Galli, B., Lo Surdo, P., Liguori, A.,
Norais, N., Santini, L., Masignani, V., Pizza, M., Giuliani, M. M.,
Exploiting chimeric human antibodies to characterize a protective
epitope of Neisseria adhesin A, one of the Bexsero vaccine components.FASEB J 2016 , 30:93–101.
[20] Li, J., Wei, H., Krystek, S. R., Bond, D., Brender, T. M.,
Cohen, D., Feiner, J., Hamacher, N., Harshman, J., Huang, R. Y.-C.,
Julien, S. H., Lin, Z., Moore, K., Mueller, L., Noriega, C., Sejwal, P.,
Sheppard, P., Stevens, B., Chen, G., Tymiak, A. A., Gross, M. L.,
Schneeweis, L. A., Mapping the Energetic Epitope of an
Antibody/Interleukin-23 Interaction with Hydrogen/Deuterium Exchange,
Fast Photochemical Oxidation of Proteins Mass Spectrometry, and Alanine
Shave Mutagenesis. Anal Chem 2017 , 89:2250–2258.
[21] Batsuli, G., Deng, W., Healey, J. F., Parker, E. T., Baldwin,
W. H., Cox, C., Nguyen, B., Kahle, J., Königs, C., Li, R., Lollar, P.,
Meeks, S. L., High-affinity, noninhibitory pathogenic C1 domain
antibodies are present in patients with hemophilia A and inhibitors.Blood 2016 , 128:2055–2067.
[22] Giuliani, M., Bartolini, E., Galli, B., Santini, L., Lo Surdo,
P., Buricchi, F., Bruttini, M., Benucci, B., Pacchiani, N., Alleri, L.,
Donnarumma, D., Pansegrau, W., Peschiera, I., Ferlenghi, I., Cozzi, R.,
Norais, N., Giuliani, M. M., Maione, D., Pizza, M., Rappuoli, R., Finco,
O., Masignani, V., Human protective response induced by meningococcus B
vaccine is mediated by the synergy of multiple bactericidal epitopes.Sci Rep 2018 , 8:3700.
[23] Sun, H., Ma, L., Wang, L., Xiao, P., Li, H., Zhou, M., Song,
D., Research advances in hydrogen–deuterium exchange mass spectrometry
for protein epitope mapping. Anal Bioanal Chem 2021 ,
413:2345–2359.
[24] Zhu, M. M., Rempel, D. L., Du, Z., Gross, M. L., Quantification
of protein-ligand interactions by mass spectrometry, titration, and H/D
exchange: PLIMSTEX. J. Am. Chem. Soc. 2003 ,
125:5252–5253.
[25] Wales, T. E., Fadgen, K. E., Gerhardt, G. C., Engen, J. R.,
High-Speed and High-Resolution UPLC Separation at Zero Degrees Celsius.Anal. Chem. 2008 , 80:6815–6820.
[26] Houde, D., Engen, J. R., Conformational analysis of recombinant
monoclonal antibodies with hydrogen/deuterium exchange mass
spectrometry. Methods Mol. Biol. 2013 , 988:269–289.
[27] Houde, D., Berkowitz, S. A., Engen, J. R., The utility of
hydrogen/deuterium exchange mass spectrometry in biopharmaceutical
comparability studies. J Pharm Sci 2011 , 100:2071–2086.
[28] Masson, G. R., Burke, J. E., Ahn, N. G., Anand, G. S.,
Borchers, C., Brier, S., Bou-Assaf, G. M., Engen, J. R., Englander, S.
W., Faber, J., Garlish, R., Griffin, P. R., Gross, M. L., Guttman, M.,
Hamuro, Y., Heck, A. J. R., Houde, D., Iacob, R. E., Jørgensen, T. J.
D., Kaltashov, I. A., Klinman, J. P., Konermann, L., Man, P., Mayne, L.,
Pascal, B. D., Reichmann, D., Skehel, M., Snijder, J., Strutzenberg, T.
S., Underbakke, E. S., Wagner, C., Wales, T. E., Walters, B. T., Weis,
D. D., Wilson, D. J., Wintrode, P. L., Zhang, Z., Zheng, J., Schriemer,
D. C., Rand, K. D., Recommendations for performing, interpreting and
reporting hydrogen deuterium exchange mass spectrometry (HDX-MS)
experiments. Nat Methods 2019 , 16:595–602.
[29] Kelley, L. A., Mezulis, S., Yates, C. M., Wass, M. N.,
Sternberg, M. J. E., The Phyre2 web portal for protein modeling,
prediction and analysis. Nat Protoc 2015 , 10:845–858.
[30] Nirudodhi, S. N., Sperry, J. B., Rouse, J. C., Carroll, J. A.,
Application of Dual Protease Column for HDX-MS Analysis of Monoclonal
Antibodies. Journal of Pharmaceutical Sciences 2017 ,
106:530–536.
[31] Zhang, Z., Smith, D. L., Determination of amide hydrogen
exchange by mass spectrometry: a new tool for protein structure
elucidation. Protein Sci. 1993 , 2:522–531.
[32] Yang, M., Hoeppner, M., Rey, M., Kadek, A., Man, P., Schriemer,
D. C., Recombinant Nepenthesin II for Hydrogen/Deuterium Exchange Mass
Spectrometry. Anal. Chem. 2015 , 87:6681–6687.
[33] Cravello, L., Lascoux, D., Forest, E., Use of different
proteases working in acidic conditions to improve sequence coverage and
resolution in hydrogen/deuterium exchange of large proteins. Rapid
Commun. Mass Spectrom. 2003 , 17:2387–2393.
[34] Zhang, H.-M., Kazazic, S., Schaub, T. M., Tipton, J. D.,
Emmett, M. R., Marshall, A. G., Enhanced digestion efficiency, peptide
ionization efficiency, and sequence resolution for protein
hydrogen/deuterium exchange monitored by Fourier transform ion cyclotron
resonance mass spectrometry. Anal. Chem. 2008 ,
80:9034–9041.
[35] Sachdev, G. P., Fruton, J. S., Secondary enzyme-substrate
interactions and the specificity of pepsin. Biochemistry1970 , 9:4465–4470.
[36] Emsley, P., Charles, I. G., Fairweather, N. F., Isaacs, N. W.,
Structure of Bordetella pertussis virulence factor P.69 pertactin.Nature 1996 , 381:90–92.
[37] Junker, M., Schuster, C. C., McDonnell, A. V., Sorg, K. A.,
Finn, M. C., Berger, B., Clark, P. L., Pertactin beta-helix folding
mechanism suggests common themes for the secretion and folding of
autotransporter proteins. Proc Natl Acad Sci U S A 2006 ,
103:4918–4923.
[38] Mazza, M. M., Retegui, L. A., Monoclonal antibodies to human
growth hormone induce an allosteric conformational change in the
antigen. Immunology 1989 , 67:148–153.
[39] Gravekamp, C., Rosner, B., Madoff, L. C., Deletion of repeats
in the alpha C protein enhances the pathogenicity of group B
streptococci in immune mice. Infect. Immun. 1998 ,
66:4347–4354.
[40] Hijnen, M., Mooi, F. R., van Gageldonk, P. G. M., Hoogerhout,
P., King, A. J., Berbers, G. A. M., Epitope structure of the Bordetella
pertussis protein P.69 pertactin, a major vaccine component and
protective antigen. Infect. Immun. 2004 , 72:3716–3723.
[41] Rubinstein, N. D., Mayrose, I., Halperin, D., Yekutieli, D.,
Gershoni, J. M., Pupko, T., Computational characterization of B-cell
epitopes. Mol. Immunol. 2008 , 45:3477–3489.
[42] Sollner, J., Grohmann, R., Rapberger, R., Perco, P., Lukas, A.,
Mayer, B., Analysis and prediction of protective continuous B-cell
epitopes on pathogen proteins. Immunome Res 2008 , 4:1.
[43] Barlow, D. J., Edwards, M. S., Thornton, J. M., Continuous and
discontinuous protein antigenic determinants. Nature1986 , 322:747–748.
[44] Haste Andersen, P., Nielsen, M., Lund, O., Prediction of
residues in discontinuous B-cell epitopes using protein 3D structures.Protein Sci 2006 , 15:2558–2567.
[45] Inatsuka, C. S., Xu, Q., Vujkovic-Cvijin, I., Wong, S.,
Stibitz, S., Miller, J. F., Cotter, P. A., Pertactin is required for
Bordetella species to resist neutrophil-mediated clearance.Infect. Immun. 2010 , 78:2901–2909.
[46] Hijnen, M., de Voer, R., Mooi, F. R., Schepp, R., Moret, E. E.,
van Gageldonk, P., Smits, G., Berbers, G. A. M., The role of peptide
loops of the Bordetella pertussis protein P.69 pertactin in antibody
recognition. Vaccine 2007 , 25:5902–5914.
[47] Kringelum, J. V., Nielsen, M., Padkjær, S. B., Lund, O.,
Structural analysis of B-cell epitopes in antibody:protein complexes.Molecular Immunology 2013 , 53:24–34.
[48] Zhang, Z., Vachet, R. W., Kinetics of Protein Complex
Dissociation Studied by Hydrogen/Deuterium Exchange and Mass
Spectrometry. Anal Chem 2015 , 87:11777–11783.