Article

pubs.acs.org/ac

Ion Mobility, Hydrogen/Deuterium Exchange, and Isotope

Scrambling: Tools to Aid Compound Identification in ‘Omics
Mixtures
Hossein Maleki,† Megan M. Maurer,† Nima Ronaghi, and Stephen J. Valentine*
Department of Chemistry, West Virginia University, Morgantown, West Virginia 26506, United States
*
S Supporting Information

ABSTRACT: Liquid chromatography tandem mass spectrometry (LC-MS/

MS), a widely used method for comparative ‘omics analysis, experiences
challenges with compound identification due to matrix effects, difficulty in
separating isomer and isobaric ions, and long analysis times. Ion mobility
spectrometry (IMS) has proven to be useful in separating isomer and isobar
ions; however, IMS-MS suffers from decreased peak capacity due to the
correlation in ion size and mass. In proof-of-principle experiments, the use of
gas-phase hydrogen/deuterium exchange (HDX) combined with IMS-MS/
MS techniques is demonstrated to offer advantages for compound
identification. Measurements providing unique information for ions include m/z value, drift time in He buffer gas, drift time
in He and D2O buffer gases, deuterium incorporation pattern (isotopic distribution), deuterium incorporation pattern after
collisional activation, and fragment ion deuterium incorporation pattern upon collision-induced dissociation (CID). These
techniques are here shown to be highly reproducible (drift time coefficients of variation < 1.0% and isotopic pattern root-mean-
square deviations of 0.5−1.5%) while demonstrating an increased ability to distinguish individual molecules from diverse classes
of compounds (peptides, catecholamines, nucleosides, amino acids, etc.). The concept of using such (and similar) information
for rapid, high-throughput molecular identification via database searching of standard libraries is briefly discussed, and an example
of such usage is presented for a bonafide metabolite extract sample.

L iquid chromatography tandem mass spectrometry (LC-

MS/MS) is a technique often used for untargeted
metabolite analysis.1,2 It allows for concurrent separations of
many different classes of compounds, whereas techniques such
as GC-MS are better suited for volatile compounds and often
rely on a derivatization step prior to analysis.1 LC-MS/MS is
also more sensitive compared to other approaches such as
NMR, which may not be able to detect low abundance species
in complex mixtures.1,2 However, one of the challenges of LC-
MS/MS is its ability to distinguish between isobars and isomers
that are not well separated by LC.2,3 Ion fragmentation is one
technique used to identify differences in isomeric and isobaric
species,4 but in the case of coeluting compounds, it may be
difficult to discern fragments of one species from the other.
Additionally, fragmentation of isomeric species often produces
ions of the same mass, requiring multistage tandem mass
spectrometry (MSn) for identification.4−6 Another approach
gaining in usage for ion identification is the determination of
molecular formula from accurate mass matches.7−10 That said,
accurate mass matching often requires separation steps to
remove interfering isomeric and isobaric species.10,11 It is also
noted that this approach requires costly, high-end mass
spectrometers.
Ion mobility spectrometry (IMS) has proven to be useful in
separating isomeric and isobaric species.12−18 IMS is a gas-
phase separation strategy in which ion differentiation is
achieved through differences in their ion-neutral collisional
cross sections (CCS) and overall charge. Ions that are more
compact and more highly charged traverse the mobility region
more rapidly than ions that are more elongated and have lower
charge states.19,20 Mobility separation of compounds occurs on
the millisecond time scale, whereas LC separations occur on the
minutes to hours time scale. Thus, it is possible to separate ions
in complex mixtures on much shorter time scales. That said,
because of the correlation between ion CCS and mass, the
combined IMS-MS peak capacity can be less than that of LC-
MS.21 Overall, the efficiency of IMS-MS is relatively high with
regard to the peak generation rate compared with that of LC-
MS.22
Gas-phase hydrogen/deuterium exchange (HDX) can be
coupled to IMS to further differentiate ions based on deuterium
incorporation (uptake). Gas-phase HDX is often used as a tool
to examine protein ion structure and conformational
changes,23−31 but it has also been used to study small
molecules.32−36 The exchange of hydrogens for deuterium
atoms with D2O reagent gas at labile sites (charge sites and
other heteroatoms) occurs via a relay mechanism.37 It has
previously been shown that the overall rate of deuterium uptake
varies in proteins, peptides, and small molecules,37−45 and in
the case of peptides, the uptake is dependent upon the amino
Received: January 7, 2017
Accepted: May 15, 2017
Published: May 15, 2017

© 2017 American Chemical Society 6399 DOI: 10.1021/acs.analchem.7b00075

Anal. Chem. 2017, 89, 6399−6407
Analytical Chemistry
acid sequence.46 The resulting mass spectra show differences in
the isotopic distribution that could be advantageous in pattern
recognition for small molecule identification. Such uptake
patterns are dependent on the partial pressure of D2O gas and
the overall HDX reaction time.43 Because at set partial
pressures of D2O and electric fields ions experience the same
number of collisions with the reagent gas from sample to
sample, the deuterium uptake behavior of individual ions is
expected to be reproducible. It therefore provides a means of
compound identification across many samples.
Hydrogen/deuterium (HD) scrambling, or rearrangement of
labile hydrogen and deuterium atoms, occurs with the addition
of energy which statistically reapportions the incorporated
deuteriums. This can be troublesome when examining protein
structures.46−48 The phenomenon is described by the “mobile”
proton theory in which collisional activation results in
significant proton mobilization.49−52 This process can be
minimized by other fragmentation techniques such as electron
capture dissociation (ECD),53,54 electron transfer dissociation
(ETD),55−57 and matrix assisted laser desorption/ionization in-
source decay (MALDI ISD). 58,59 Several studies have
determined that HD scrambling on peptide ions should be
dependent upon the amino acid sequence46,49,50 and the type of
charge carrier (H+ compared to Na+)46,60,61 and is independent
of the type of mass spectrometer used (Q-TOF vs ion trap46).
Because HD scrambling has been shown to be dependent on
the analyte itself, this characteristic may be exploited as an
additional parameter to further differentiate metabolites
without the use of LC separation. In the study described
here, HD scrambling is used to repopulate some exchange sites
with hydrogens which can then undergo gas-phase HDX,
resulting in an overall increase in the amount of incorporated
deuteriums.
The reproducibility of IMS, HDX, and HD scrambling with
HDX measurements is particularly important when using such
parameters for compound identification. High-reproducibility
provides the opportunity to “dial in” these parameters across
many samples using select measurement conditions. For set
separation and activation electric fields in the drift tube, the
overall reproducibility of these measurements is governed by
the control of the partial pressures of the buffer gases both
within an experimental run and from run to run.
Herein, a strategy that utilizes IMS, HDX, and HD
scrambling with HDX is proposed to rapidly distinguish
compounds in a complex mixture. The approach is evaluated
for its ability to provide unique parameters for identification of
a diverse set of compounds as well as for its overall
reproducibility. The former evaluation addresses issues with
regard to limited peak capacity, while the latter indicates
feasibility with regard to automated pattern matching. Although
the demonstration reported here was performed using relatively
rudimentary control of reagent gas partial pressure and model
systems, the ion distinguishing and reproducibility character-
istics of the approach suggest the possibility of incorporation
into routine ‘omics workflows in the future.
■
EXPERIMENTAL SECTION
Materials. Dopamine hydrochloride (99%, AK Scientific),
guanosine (99%, Acros Organics), acetaminophen (98%,
Sigma-Aldrich, St. Louis, MO), and alanine (98%, Alfa Aesar)
were purchased and used without further purification. LC-MS
grade water, acetonitrile, formic acid, and acetic acid were
purchased from Sigma-Aldrich. D2O (99.9% D) was purchased
6400
Article
from Sigma-Aldrich. Ammonium bicarbonate (≥99.5%), urea
(ACS reagent grade), ubiquitin (bovine, 98.0%), and pepsin
(porcine, ≥2500 units/mg protein) were also purchased from
Sigma-Aldrich. Bradykinin (≥95.0%), substance P (≥95.0%),
and synthetic peptide KKDDDDDIIKKII (KKD peptide,
94.1%) were all purchased from GenScript (Piscataway, NJ).
These materials were used without subsequent purification.
Bovine heart extract was purchased from Lipid Avanti, Inc.
Sample Preparation for Reproducibility Studies.
Ubiquitin was solubilized in ultrapure (18 MΩ) water to a
concentration of 1.0 mg/mL. The protein solution was then
acidified with glacial acetic acid (0.1 M) to a pH of ∼2.8. Pepsin
was added to the protein solution to a final concentration of
0.33 mg/mL. Protein digestion was performed at 37 °C for 60
min. Acetonitrile and acetic acid were then added to the protein
solution such that the final ESI solution composition was 1:1
water:acetonitrile with 1% acetic acid. The sample was loaded
into a gastight syringe; the syringe was placed into a syringe
pump, and the sample was infused at 300 nL/min through the
electrospray needle which was biased at 2000 V relative to the
desolvation region entrance.
A mixture of model peptides containing 45 μg/mL
bradykinin, 45 μg/mL substance P, and 450 μg/mL KKD
peptide was generated in 1:1 water:acetonitrile with 0.1% (v/v)
acetic acid. Samples were loaded into a gastight syringe for
direct infusion into the ion source.
The dopamine, guanosine, acetaminophen, and alanine
standard samples were generated by dissolving 10 mg of each
metabolite in methanol and ACN (1:1 v/v) to a final volume of
100 mL (0.1 mg/mL). Acetic acid (1%) was used to enhance
the ionization of these compounds.
A lipid working solution was prepared by mixing 250 mL of
methylene chloride, 250 mL of methanol, and 5 mL of
ammonium acetate buffer solution (1 M solution prepared by
dissolving 1.54 g of ammonium acetate in 20 mL of methanol).
The bovine heart extract was prepared by diluting 400 μL of
bovine heart extract with 3600 μL of lipid working solution.
Data Collection and Analysis. Data were collected on a
home-built, dual-gate IMS device coupled to a Thermo LTQ
Velos mass spectrometer (Thermo Fisher, San Jose, CA) which
has been described in detail elsewhere.62 The front gate was
open for a duration of 150 μs, and the back gate was opened at
a delay with respect to the first gate for 200 μs. Ions were
allowed to traverse the drift tube and were then collected in the
ion trap for 100 ms prior to mass analysis for each delay time
setting. Data were collected for 30 s at each drift time, resulting
in an averaged mass spectrum.
For sequential and random drift time data collection, a total
of 3 sample sets were collected. For sequential drift time data
collection, the delay times sequentially stepped through (5.0,
5.2, 5.4 ms, etc.) the drift time range. For random drift time
data collection, the delay times were randomized (7.8, 6.2, 10.4
ms, etc.). The order of the type of acquisition was randomly
assigned. The run order was: random, sequential, sequential,
random, sequential, random. The random sequence of drift
time selection was achieved with the pseudorandom number
generator in the Excel software. The mobility selection using
the second ion gate followed an increasing order of the
associated random numbers. The helium buffer gas pressure
within the drift tube was maintained at ∼2.60 Torr for
reproducibility experiments. This pressure was maintained in
the drift tube for the duration of the data acquisition for the six
trials.
DOI: 10.1021/acs.analchem.7b00075
Anal. Chem. 2017, 89, 6399−6407
Analytical Chemistry
For the HDX and scrambling experiments, the partial
pressures of He and D2O were maintained at ∼2.50 and
∼0.02 Torr, respectively. The reservoir containing the D2O was
maintained at a temperature of 35 °C, and the tubing from the
reservoir to the drift tube was heated to prevent D2O
condensation which could lead to partial pressure fluctuations.
Notably, this partial pressure of D2O results in relatively little
reagent consumption (100 g in tens of hours). Drift time
reproducibility using ubiquitin digest samples subjected to
HDX was obtained by randomly stepping through the delay
times with three trials for each sample type. HDX
reproducibility studies using isotopic distributions were
performed on the model peptide mixture without drift
selection. The reproducibility studies from experiments
employing HD scrambling with HDX were also performed
with the model peptide mixture without drift time selection. In
all, six replicates were performed for each HDX and HD
scrambling with HDX study using this sample. Between each
trial, the drift tube voltage was removed, and the drift tube
gases were evacuated. Care was taken to return the gas partial
pressures to the same approximate values by using the same
leak valve settings for each run (Granville Phillips, MKS
Instruments, Andover, MA). The buffer gas temperature was
determined using a thermocouple attached to the stainless steel
plate through which the buffer gas was introduced; notably,
gentle heating of D 2 O introduction lines to prevent
condensation did not change the temperature significantly.
The buffer gas pressure was measured using a capacitance
manometer.
Data from each IMS-MS data set were compiled into a three-
column text file [drift time, mass-to-charge (m/z), and
intensity]. Two-dimensional (2D) contour plots of m/z versus
drift time of the data were constructed using DPlot (v. 2.3.5.3,
HydeSoft Computing, LLC; Vicksburg, MS, United States).
Data features corresponding to ubiquitin digest peptide ions
were compared across replicate data sets to determine the
reproducibility in drift time. Coefficients of variation (CVs)
were calculated for multiple digest peptide ions.
Post-HDX isotopic envelope reproducibility for ubiquitin
digest peptide ion conformers (mobility selected) was
performed using an algorithm to extract drift times and m/z
values surrounding each peptide ion of interest. The
dimensions of the extraction box varied based on conformer
proximity, the total number of isotopes, and the charge state of
the peptide ion. Typically, a drift time range of <1 ms wide was
used. Ion intensities within the region were summed for each
m/z bin (0.0833 units wide) in the extracted region and then
normalized by the total intensity of the region. The root-mean-
square deviation (RMSD) for each experimental run was
compared to the group mean.
Isotopic distribution reproducibility for HDX and HD
scrambling with HDX measurements for the model peptide
mixture (bradykinin, substance P, and KKD peptide) was
examined without drift selection. HD scrambling was induced
by collisionally activating ions in the last portion of the drift
tube (∼0.9 m from the ion source). Continued HDX was then
allowed to proceed in the remaining portion of the drift tube
(∼0.1 m). Six replicates of both the HDX and HD scrambling
with HDX analyses were collected for the model peptide
mixture. Between each trial, the mass spectrometer was set to
stand by; the applied voltage was removed, and the buffer gases
were evacuated from the drift tube. Subsequently, the gases
were reintroduced manually by dialing the leak valves to their
6401
Article
previous setting values, and the drift voltages were reapplied.
Thus, each replicate essentially mimics a new “cold start”
measurement. Isotopic distribution reproducibility was deter-
mined as described above for select digest peptides and small-
molecule metabolites. A separate demonstration examined the
fragment ions produced by peptide ions undergoing HD
scrambling with HDX; the fragment ions were produced by
collision-induced dissociation (CID) in the linear ion trap mass
spectrometer. Isotopic distribution comparisons were per-
formed (RMSD values) for some of the resulting fragment
ions. Isotopic distributions for alanine and an isobaric
metabolite ion were compared before and after ion activation.
■ RESULTS AND DISCUSSION
Deuterium Uptake for Distinguishing Molecules. As
mentioned above, the ability to utilize a distinguishing
measurement relies on its probing of a separate physicochem-
ical property of the ion. With regard to gas-phase HDX using
D2O reagent gas, the property is labile site accessibility, which
has been interpreted as accessibility to charge site and the ion’s
surface (collisions with D2O).30,37,43−45,63,64 With different
arrangements of charge sites and heteroatoms, a large number
of polar compounds should exhibit different HDX character-
istics. To demonstrate the nature of gas-phase HDX as a
compound-specific measurement, a model peptide mixture was
used to provide examples of different HDX behavior. Notably,
the use of HDX to increase the overall peak capacity of the
measurement in this manner is similar to seminal studies that
employed parallel dissociation65,66 and drift shift reagents67,68
to disrupt the correlation between IMS and MS measurements.
Unique deuterium uptake patterns of the different peptide
ions were observed for the model peptide mixture (Figure 1).
For example, for the [M+2H]2+ ions of substance P, little to no
deuterium (<1) uptake is observed, as shown in Figure 1A. In
contrast, for [M+2H]2+ bradykinin ions from the same
experiment, Figure 1B shows that, on average, the [M+2H]2+
bradykinin ions incorporate ∼3 deuteriums. For the [M+3H]3+
KKD peptide ions, the average deuterium uptake is ∼12
deuteriums (Figure 1C).
For these experiments, the partial pressure of D2O is
sufficient to exchange the rapidly exchanging sites (readily
accessible).30,44,45,64 To examine the origin of the differences in
deuterium incorporation (Figure 1), it is instructive to consider
the numbers of exchangeable hydrogens on these peptide ions.
There are five residues on substance P (RPKPQQFFGLM)
containing side-chain exchange sites, namely: one arginine, one
lysine, two glutamines, and one methionine. Therefore,
including these sites and those at the backbone amide,
amino- and carboxy-termini, and the two extra protons, the
total number of exchangeable hydrogens is 24 for [M+2H]2+
substance P ions. In comparison, there are 19 and 31
exchangeable hydrogens for the [M+2H]2+ bradykinin and
[M+3H]3+ KKD peptide ions, respectively. In comparing the
two doubly charged ions, the percent deuterium incorporation
for substance P and bradykinin is 4.2 and 15.8%, respectively.
This translates into nearly a fourfold efficiency increase for the
bradykinin ions. Therefore, although the ions have the same
number of charges and nearly the same number of residues and
exchangeable hydrogens, they exhibit significantly different
deuterium incorporation capabilities. Presumably this is a
reflection of differences in exchange site accessibility to charge
sites and surface collisions (i.e., the ion structures).30,43,44,69
Overall, the triply charged KKD peptide ions exhibit the
DOI: 10.1021/acs.analchem.7b00075
Anal. Chem. 2017, 89, 6399−6407
Analytical Chemistry
Figure 1. Model peptides showing different deuterium uptake patterns
with D2O added to the drift tube. (A) [M+2H]2+ substance P ions. (B)
[M+2H]2+ bradykinin ions. (C) [M+3H]3+ KKD peptide ions. The
buffer gas composition is labeled for conditions in which He (blue)
and He + D2O (orange) are used.
greatest efficiency (38.7%) in gas-phase HDX (nearly double
that of the bradykinin ion).
Reproducibility of Drift Time in HDX Experiments. To
take advantage of spectral matching for peptide ion assignment,
it is requisite that drift time distributions resulting from
experiments using the buffer gas mixture are highly
reproducible. Indeed, such spectral reproducibility is a strength
of metabolite assignment using GC-MS data where the
fragmentation spectra from electron ionization (EI) are
extremely consistent. The instrument used in these experiments
(described in Donohoe et al.62) uses a longer gate time at the
back of the drift tube compared to the front (200 μs vs 150 μs).
This is required because of the relatively lengthy second gate
6402
Article
(∼3 mm). That is, a longer time is required to allow the ions to
fully traverse the distance of the second gate and any extra
region caused by impinging fields during the off-transmit
setting. Thus, to determine if gating differences affect the
measurement reproducibility, data were collected by sequen-
tially or randomly stepping through the drift times using a
buffer gas of He only. Three replicates of each type of
acquisition were performed in a random order for a peptic
digest of ubiquitin. The average CVs for 16 peptide ion peaks
were determined to be 0.52 ± 0.11 and 0.33 ± 0.23% for
random and sequential data collection, respectively. 2D drift
time and m/z contour plots for the 16 peptide ions are shown
in Figures S1 (random data collection) and S2 (sequential data
collection), respectively, in the Supporting Information. The
identities of the peptide ions and their respective drift times are
also provided in Tables S1 and S2. This analysis serves as the
benchmark against which the drift time reproducibility of
peptide ions in the buffer gas mixture (He + D2O) is compared.
Variability in buffer gas pressure is a greater concern with the
buffer gas mixture considering that only a small fraction of the
composition is D2O. Therefore, a small change in partial
pressure could result in large changes in drift times. This could
be a significant source of error especially for a system that
utilizes an evaporative introduction system employing a leak
valve such as is used here. Another challenge could be the
undersampling of drift time peaks as required by the wide gate
times. That is, a peak apex partitioning between two large bins
can skew centroid determination and erode overall reproduci-
bility. The purpose of the following experiments was to
determine how much the drift time reproducibility deviates
from the benchmark values upon using the buffer gas mixture.
Figure 2 provides an indication of the run-to-run variability in
drift time for ubiquitin digest peptide ions that undergo HDX
in the drift tube. Data set features for nine [M+H]+, [M+2H]2+,
and [M+3H]3+ ions show very little difference in drift time
(Figure 2A). Consider the CV for peak 9 (Figure 2B) which has
been assigned to the peptide ion [VKTLTGKTITL+3H]3+.
Across the 3 replicate analyses, the CV is 0.36% (Figure 2B).
Separate data set features examined in the drift time
distributions of the ubiquitin sample were found to provide
an average CV of 0.55 ± 0.20% (Figures 2 and S3). The
reproducibility is not significantly different when separately
dialing in the buffer gas mixture components with the leak
valves. This is still within the high reproducibility range of IMS
experiments20 despite the usage of relatively large drift bin sizes
(see above). Table S3 provides the drift times from replicate
runs for the assigned digest peptide ions.
Reproducibility of Deuterium Uptake in HDX Experi-
ments. As with the drift times, comparisons of HDX reactivity
reproducibility can be achieved with spectral comparisons.
Here, due to the presence of multiple isotopologue ions, an
RMSD comparison was employed. To best illustrate the
reproducibility, results for [M+2H]2+ bradykinin ions from the
peptide mixture are presented in Figure S4 in the Supporting
Information. In these experiments, the cold start procedure was
employed (see Experimental Section). Figure S4 shows the
isotopic distribution of the [M+2H]2+ bradykinin ions across
six trials. The RMSDs of the trials were less than 0.2%. The
RMSDs of the other peptides within the model peptide mix
were also determined for the [M+2H]2+ KKD peptide and
substance P ions; [M+3H]3+ bradykinin, KKD peptide, and
substance P peptide ions; and the [M+4H]4+ KKD peptide
ions. The RMSDs were all found to be less than 1% (see
DOI: 10.1021/acs.analchem.7b00075
Anal. Chem. 2017, 89, 6399−6407
Analytical Chemistry
Figure 2. IMS-MS profile of two ubiquitin digest trials (A) demonstrating the run-to-run reproducibility of compounds in a complex mixture after
addition of D2O to the drift tube. Panel B shows the CV for select peptide ions. m/z and average drift time values are provided for the peptide ions.
CV values for data set features tentatively identified as (1) QRLIFAGKQ+3H]3+; (2) [NIQ+H]+; (3) [HLVLRL+2H]2+; (4) [GIPPDQQ+2H]2+;
(5) [KIQDKE+2H]2+; (6) [VLRLRGG+2H]2+; (7) [IFAGKQL+2H]2+; (8) [DKE+H]+; and (9) [VKTLTGKTITL+3H]3+ ions as labeled in (B).
respective panels in Figures S5 and S6 in the Supporting
Information).
A question arises as to whether or not comparisons of such
isotopic distributions could be of utility for complex mixture
analysis where significant abundance differences are encoun-
tered. Here, it is noted that the highly reproducible drift time
separation (Figure 2) is useful. That is, in mobility separation
experiments, it is possible to select for data set features and
obtain the high reproducibility of the isotopic distributions
demonstrated for the bradykinin ions. For example, mobility
selection of the features associated with [DKE+H] + ,
[VLRLRGG+2H]2+, and [QRLIFAGKQ+3H]3+ peptide ions
provides isotopic distribution RMSD values that are all below
1.5% (Figure S7 in the Supporting Information). Notably, the
RMSD increases for lower intensity species such as the
[HLVLRL+2H]2+ peptide ions (Figure S7), yielding a value of
∼2.5%; these ions have an abundance that is 20-fold less than
that of the [VLRLRGG+2H]2+ ions. Finally, the degradation of
RMSD values with drift window location is reported. For the
[VLRLRGG+2H]2+ ions, at 10% height of the drift time peak
apex, the RMSD value increases by a factor of 3 (Figure S7).
Overall, even for lower abundance ions, the mobility and
reactivity information together can be used for compound
matching provided that the isotopic distributions of other
species in complex mixtures do not shift and interfere with that
of the desired analyte.
Isotopic Distribution Reproducibility after HDX and
HD Scrambling with HDX. One issue in determining ion
identities in the manner proposed here is whether or not there
is a sufficient number of identifiers to obtain high-confidence
assignments. Ideally, increased parametrization (separate
measurements) would be required to address mixtures of
greater complexity. A question arises as to whether or not
additional measurements can be made with the instrumentation
employed here that would provide unique ion characterization.
One approach could involve the use of HD scrambling followed
by HDX. From the “mobile proton” model for peptide ion
6403
Article
dissociation,49,50,52 collisional activation causes intramolecular
proton transfer events where a severing of a chemical bond can
occur at the site of proton migration. Because deuterons can be
mobilized as well as protons,46,70 HD scrambling can follow
collisional activation. HD scrambling offers an opportunity in
that hydrogens can be scrambled to accessible sites, and
deuteriums can be scrambled to sites less accessible by HDX.
Therefore, more extensive deuterium uptake can occur as the
newly populated sites with hydrogen undergo subsequent
HDX.
To test the reproducibility of a process in which ions are
subjected to HDX and then HD scrambling with subsequent
HDX, experiments were performed for the peptide mixture
sample as outlined in the Experimental Section. Again, cold
start conditions were employed. Figure S8 in the Supporting
Information shows the effect of HD scrambling with further
HDX on the same peptide ions from the model peptide mixture
shown in Figure 1. Here, it is demonstrated that even with ion
activation there is little to no further deuterium incorporation
for the [M+2H]2+ substance P and bradykinin ions. In
comparison, [M+3H]3+ KKD peptide ions incorporate
significantly more (∼3−4) deuteriums after scrambling. The
isotopic distribution RMSDs for the 6 HD scrambling trials of
all model peptide ions were less than 1%, also demonstrating
suitable reproducibility for routine use in ion identification
analyses. These distributions are provided in Figures S9 and
S10 in the Supporting Information.
Demonstration of HDX and HD Scrambling for Other
Small Molecules. The application of HDX and scrambling
experiments was extended to small-molecule metabolite
samples to demonstrate the universality of these techniques
for complex mixture analysis. Figures 3A, C, E, and G present
the isotopic distributions for alanine, guanosine, dopamine, and
acetaminophen, respectively, after being subjected to gas-phase
HDX in the drift tube. Figures 3B, D, F, and H show the
isotopic distributions after collisional activation of the same
ions in the drift tube. The molecular structures of these ions
DOI: 10.1021/acs.analchem.7b00075
Anal. Chem. 2017, 89, 6399−6407
Analytical Chemistry Article

Figure 3. Isotopic distributions of precursor ions of (A) alanine, (C) guanosine, (E) dopamine, and (G) acetaminophen. Distributions for mobility
selected ions in the presence of He only and He and D2O are shown as blue and orange traces, respectively. Panels B, D, F, and H show the isotopic
distributions (green traces) for the respective molecular ions obtained after collisional activation in the drift tube.

and recorded reduced mobilities are presented in Table S4 in
the Supporting Information.
[M+H]+ alanine ions have a nominal m/z value of 90. As
demonstrated in Figure 3A, these ions incorporate ∼1.3
deuteriums on average. This is evident from the fact that the
ions with a nominal m/z of 91 are the most abundant ions.
Upon collisional activation of the ions in the drift tube,
increased deuterium incorporation is evident by the fact that
ions of nominal m/z 92 are in equal abundance to those of m/z
91. The average deuterium incorporation increases to 1.7
deuteriums. As with the peptide ions, the isotopic patterns are
extremely reproducible using the evaporative introduction
system; average RMSD values are 1.1 and 1.6% for the
precursor and activated ions, respectively (see Figure S11 in the
Supporting Information).
The behavior of alanine ions can be contrasted with that of
[M+H]+ guanosine ions. Guanosine has seven exchangeable
hydrogens. Protonated guanosine ions have a nominal m/z
6404
value of 284. Upon introduction of D2O in the drift tube, the
isotopic distribution of these ions shows a high intensity peak at
m/z 285, as shown in Figure 3C. Figure 3D shows that, after
activation of the ions, the peak at m/z 285 remains the most
abundant peak. Indeed, the isotopic distributions for the
mobility-selected ions are nearly identical. On average, the
deuterium incorporation is ∼1.0 for the precursor and the
collisionally activated ions. As with the other ions presented in
this study, the reproducibility of the HDX measurements is
high; average RMSD values for the isotopic distribution
comparisons of the precursor and activated ions are 0.7 and
1.5%, respectively (see Figure S12 in the Supporting
Information).
[M+H]+ dopamine ions have a nominal m/z value of 154. As
demonstrated in Figure 3E, these ions do not undergo
appreciable exchange upon introduction of D2O reagent gas
to the drift tube. Additionally, after collisional activation in the
drift tube, the dopamine ions do not significantly increase the
DOI: 10.1021/acs.analchem.7b00075
Anal. Chem. 2017, 89, 6399−6407
Analytical Chemistry Article

level of deuterium incorporation (Figure 3F). Indeed, using the instrumentation geometry. That said, all of the measurements
average mass difference computed from the isotopic distribu- shown here can be performed on an instrument that combines
tions, no deuteriums are incorporated for the precursor ions IMS with TOF MS where the MS measurement is nested in the
and the collisionally activated ions. The average RMSD values IMS experiment.72 Additionally, parallel CID can be conducted
for the isotopic distribution comparisons are 1.1 and 0.5% for for such an instrument at no cost to experimental run time.65,66
the precursor and activated ions, respectively (see Figure S13 in Thus, it would be possible to make these measurements on the
the Supporting Information). seconds time scale. Consider experiments in which drift times
Finally, Figures 3G and H present the HDX behavior of [M and m/z values could be recorded in 5 s for which He is
+H]+ acetaminophen ions. The ions have a nominal m/z value employed as a drift buffer gas. Then, these values could be
of 152 and contain up to 3 exchangeable hydrogens. As shown recorded for 5 s experiments in which a He and D2O buffer gas
in Figure 3G, with the addition of D2O to the drift tube, the mixture is used. Next the m/z shift could be measured in a 5 s
dominant feature has a nominal m/z value of 153 with an HD scrambling with HDX experiment. Another 5 s experiment
increase in features of m/z 154 and 155. After activation of the could be used to perform parallel CID. Overall, 20 s of
ions, the higher m/z peaks slightly increase in relative intensity instrumentation time could produce these 6 pieces of ion
within the isotopic distribution. For isotopic distribution identification information. This is a significant time savings
comparisons, average RMSD values of 1.2 and 1.0% are compared with condensed-phase separations.
obtained for the precursor and activated ions, respectively (see Gas-Phase HDX Behavior to Aid Compound Identi-
Figure S14 in the Supporting Information). The data indicate fication. As a final proof-of-principle demonstration of the use
that acetaminophen ions uptake ∼2.3 deuteriums with the of HDX behavior in compound identification efforts, the
introduction of D2O and that the uptake increases with ion isotopic distributions of [M+H]+ alanine ions and bovine heart
activation. Overall, Figure 3 shows different, distinguishing metabolite extract ions of the same nominal m/z value (90) and
HDX behavior for the various ions. similar mobility (9.1 × 10−4 m2 V−1 s−1, see also Table S4) are
Reproducibility in Experiments Using HDX and HD compared. Figure 4 shows the isotopic distributions obtained
Scrambling with HDX Followed by CID. As a final ion
descriptor, it is useful to consider ion fragmentation in
conjunction with the HDX and HD scrambling with HDX
processes. Such an approach derives conceptually from parallel
dissociation methods performed using ion mobility devices.65,66
Here, although CID results in HD scrambling, the
fragmentation process would provide reproducible ion
fragmentation spectra as well as isotopic distributions for
fragment ions. Furthermore, given that HDX provides a
sufficient shift in m/z value, ion fragmentation could allow
for further disambiguation of sample peaks within the same
drift time window.
The model peptide mixture was subjected to HDX combined
with HD scrambling with HDX. Then, a peak at a drift time of
6.7 ms and having m/z 526 corresponding to extensively
deuterated [M+3H]3+ KKD peptide ions was selected (mobility
and m/z), and the precursor ions were subjected to CID in the
linear ion trap at a normalized collision energy of 40. Three
replicates were collected with the cold start approach. Figure
S15A in the Supporting Information presents the MS/MS
spectra for the [M+3H]3+ KKD peptide ions. Notably, the ion
fragmentation spectra are very similar, as may be expected from
the highly reproducible process of CID.71 For example, the
relative intensities of the identified y-ion series are very similar
across the fragmentation spectra. Figure S15B shows an
expanded region of the fragmentation spectra encompassing Figure 4. Isotopic distributions of precursor ions having nominal m/z
90 obtained from a bovine heart extract. Panel A shows distributions
the isotopic distribution of the y122+ fragment ion. The isotopic for ions in the presence of He only and He and D2O as blue and
distribution was found to be highly reproducible (RMSDs < orange traces, respectively. Panel B shows the isotopic distribution
1%) for this fragment ion (Figure S15). Comparisons of other (green trace) for these ions obtained after collisional activation in the
identified fragment ions exhibited similar RMSD values (data drift tube.
not shown).
Although the experiments described above provide proof-of-
principle demonstrations of combined IMS-HDX-MS/MS for the ions having nominal m/z 90 from the bovine heart
approaches for obtaining information that could serve as extract. Upon introduction of D2O to the drift tube, similar
unique identifiers for ions, the discussion has thus far not isotopic distributions for these ions and the alanine ions (Figure
addressed the time scale of these measurements. Admittedly, 3A) are observed. A possible exception is the presence of a
the combination of IMS with the scanning MS instrumentation feature at m/z 90 (ions that do not undergo HDX) for the
results in a relatively long time scale measurement (e.g., MS are unknown compound in the metabolite extract (Figure 4A). The
collected at separate delay time settings). The IMS-MS data most noticeable difference in HDX behavior is observed for the
reported here required time scales of tens of minutes for this activated ions for which alanine shows an increase in deuterium
6405 DOI: 10.1021/acs.analchem.7b00075
Anal. Chem. 2017, 89, 6399−6407
Analytical Chemistry
incorporation (Figure 3B), while no similar increase is observed
for the heart extract ions (Figure 4B). Considering the
reproducibility of the measurements as described above, this
would be an indication that such ions from the metabolite
extract could not be assigned to the amino acid alanine. This
assessment is supported by ion fragmentation experiments
showing different fragmentation patterns for the mobility-
selected ions (see Figure S16 in the Supporting Information).
Although the above example shows an approach in which a
false positive assignment is identified, with a database of
sufficient size (i.e., one containing the metabolite), assignment
confirmation would be achieved by matching all of the new
measurement parameters.
■ CONCLUSIONS
The experiments described here provide proof-of-principle
examples of how IMS combined with gas-phase HDX and MS
analysis techniques can be used to provide information-rich
data sets. Such data sets could have a number of unique ion
descriptors to be used in identification efforts for compounds in
mixtures such as those encountered in ‘omics analyses. The
studies show that, even with rudimentary buffer gas
introduction systems, sufficient reproducibility can be obtained
for these descriptors such that ion identification could be
achieved through comparisons with values recorded for
molecular standards. Although such an approach would require
a Herculean effort to populate comparison databases, the
rapidity of the measurements would allow for significant
improvements in sample throughput, which is a tremendous
challenge for biomarker discovery efforts. In addition to the
database challenge, scoring schemes would need to be devised
to provide an assessment of confidence in ion matches.
Although important, such strategies could benefit from
algorithms developed previously. For example, scoring isotopic
distribution matches could be achieved using an approach that
is similar to the cross correlation approach employed by the
SEQUEST software suite used in proteomics experiments.73
As a final consideration, it is instructive to briefly examine a
potentially significant advantage of the IMS-HDX-MS/MS
approach. Over the years, significant effort has been exerted to
develop analytical region prediction tools based on biomo-
lecular ion structure. These include methods for predicting ion-
neutral collision cross sections74−76 and the gas-phase HDX
behavior of specific ions.30,43,44,57,77 In the context of
comparative ‘omics, the ability to predict ion-neutral collision
cross sections for small molecules could be beneficial for
species for which no measurements are available76 such as
newly emerging synthetic drugs and their metabolites or newly
discovered neuropeptides. Adding the ability to predict isotopic
distributions upon HDX may significantly enhance the ion
distinguishing capabilities of such an approach.77 Notably,
significant work to build out and utilize ion-neutral collision
cross section and HDX databases as well as to construct
predictive algorithms would be required to avail such an
approach, and this is a focus of future research efforts. Because
of the extremely high reproducibility of the approach, it may be
possible to incorporate HDX database searching into ‘omics
workflows in much the same way that GC-MS employs small-
molecule databases.
6406
Article
■
*
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.7b00075.
Tables S1−S3, drift time reproducibility data; Table S4,
reduced mobilities and structures of small molecules used
in the study; Figure S1−S3, mobility reproducibility;
Figures S4−S7 and S9−S15, isotopic distribution
reproducibility after HDX; Figure S8, different HDX
behavior by activated ions; Figure S16, fragmentation
patterns of isobaric ions (PDF)
■
AUTHOR INFORMATION
Corresponding Author
*E-mail: stephen.valentine@mail.wvu.edu; Phone: (304) 293-
4937; Fax: (304) 293-4904.
ORCID
Stephen J. Valentine: 0000-0003-4360-2928
Author Contributions
†
H.M. and M.M.M. contributed equally to this work.
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS
We are grateful for the financial support of this work by the
National Science Foundation (Grant CHE-1553021).
■
REFERENCES
(1) Zhang, A.; Sun, H.; Wang, P.; Han, Y.; Wang, X. Analyst 2012,
137, 293−300.
(2) Alonso, A.; Marsal, S.; Julià, A. Front. Bioeng. Biotechnol. 2015, 3,
1−20.
(3) Schrimpe-Rutledge, A. C.; Codreanu, S. G.; Sherrod, S. D.;
McLean, J. A. J. Am. Soc. Mass Spectrom. 2016, 27, 1897−1905.
(4) Vaniya, A.; Fiehn, O. TrAC, Trends Anal. Chem. 2015, 69, 52−61.
(5) Cao, M.; Fraser, K.; Rasmussen, S. Metabolites 2013, 3, 1036−
1050.
(6) Wang, X.; Peng, Q.; Li, P.; Zhang, Q.; Ding, X.; Zhang, W.;
Zhang, L. Anal. Chim. Acta 2016, 940, 84−91.
(7) Brown, M.; Dunn, W. B.; Dobson, P.; Patel, Y.; Winder, C. L.;
Francis-McIntyre, S.; Begley, P.; Carroll, K.; Broadhurst, D.; Tseng, A.;
Swainston, N.; Spasic, I.; Goodacre, R.; Kell, D. B. Analyst 2009, 134,
1322−1332.
(8) Junot, C.; Madalinski, G.; Tabet, J. C.; Ezan, E. Analyst 2010, 135,
2203−2219.
(9) Weber, R. J. M.; Viant, M. R. Chemom. Intell. Lab. Syst. 2010, 104,
75−82.
(10) Rathahao-Paris, E.; Alves, S.; Junot, C.; Tabet, J.-C.
Metabolomics 2016, 12, 10.
(11) Dunn, W. B.; Erban, A.; Weber, R. J. M.; Creek, D. J.; Brown,
M.; Breitling, R.; Hankemeier, T.; Goodacre, R.; Neumann, S.; Kopka,
J.; Viant, M. R. Metabolomics 2013, 9, S44−S66.
(12) Bylda, C.; Thiele, R.; Kobold, U.; Bujotzek, A.; Volmer, D. A.
Anal. Chem. 2015, 87, 2121−2128.
(13) Regueiro, J.; Negreira, N.; Berntssen, M. H. G. Anal. Chem.
2016, 88, 11169−11177.
(14) Kyle, J. E.; Zhang, X.; Weitz, K. K.; Monroe, M. E.; Ibrahim, Y.
M.; Moore, R. J.; Cha, J.; Sun, X. F.; Lovelace, E. S.; Wagoner, J.;
Polyak, S. J.; Metz, T. O.; Dey, S. K.; Smith, R. D.; Burnum-Johnson,
K. E.; Baker, E. S. Analyst 2016, 141, 1649−1659.
(15) Reading, E.; Munoz-Muriedas, J.; Roberts, A. D.; Dear, G. J.;
Robinson, C. V.; Beaumont, C. Anal. Chem. 2016, 88, 2273−2280.
DOI: 10.1021/acs.analchem.7b00075
Anal. Chem. 2017, 89, 6399−6407
Analytical Chemistry
(16) Lalli, P. M.; Corilo, Y. E.; Rowland, S. M.; Marshall, A. G.;
Rodgers, R. P. Energy Fuels 2015, 29, 3626−3633.
(17) Pu, Y.; Ridgeway, M. E.; Glaskin, R. S.; Park, M. A.; Costello, C.
E.; Lin, C. Anal. Chem. 2016, 88, 3440−3443.
(18) Maleki, H.; Ghassabi Kondalaji, S.; Khakinejad, M.; Valentine, S.
J. Energy Fuels 2016, 30, 9150−9161.
(19) Kanu, A. B.; Dwivedi, P.; Tam, M.; Matz, L.; Hill, H. H. J. Mass
Spectrom. 2008, 43, 1−22.
(20) Clemmer, D. E.; Jarrold, M. F. J. Mass Spectrom. 1997, 32, 577−
592.
(21) Bohrer, B. C.; Merenbloom, S. I.; Koeniger, S. L.; Hilderbrand,
A. E.; Clemmer, D. E. Annu. Rev. Anal. Chem. 2008, 1, 293−327.
(22) Deng, L. L.; Ibrahim, Y. M.; Hamid, A. M.; Garimella, S. V. B.;
Webb, I. K.; Zheng, X. Y.; Prost, S. A.; Sandoval, J. A.; Norheim, R. V.;
Anderson, G. A.; Tolmachev, A. V.; Baker, E. S.; Smith, R. D. Anal.
Chem. 2016, 88, 8957−8964.
(23) Katta, V.; Chait, B. T.; Carr, S. Rapid Commun. Mass Spectrom.
1991, 5, 214−217.
(24) Hemling, M. E.; Conboy, J. J.; Bean, M. F.; Mentzer, M.; Carr, S.
A. J. Am. Soc. Mass Spectrom. 1994, 5, 434−442.
(25) Kaltashov, I. A.; Doroshenko, V. M.; Cotter, R. J. Proteins:
Struct., Funct., Genet. 1997, 28, 53−58.
(26) Hoofnagle, A. N.; Resing, K. A.; Ahn, N. G. Annu. Rev. Biophys.
Biomol. Struct. 2003, 32, 1−25.
(27) Wales, T. E.; Engen, J. R. Mass Spectrom. Rev. 2006, 25, 158−
170.
(28) Konermann, L.; Pan, J. X.; Liu, Y. H. Chem. Soc. Rev. 2011, 40,
1224−1234.
(29) Khakinejad, M.; Ghassabi Kondalaji, S.; Donohoe, G. C.;
Valentine, S. J. J. Am. Soc. Mass Spectrom. 2016, 27, 451−461.
(30) Khakinejad, M.; Kondalaji, S. G.; Tafreshian, A.; Valentine, S. J.
J. Am. Soc. Mass Spectrom. 2015, 26, 1115.
(31) Mistarz, U. H.; Brown, J. M.; Haselmann, K. F.; Rand, K. D.
Anal. Chem. 2014, 86, 11868−11876.
(32) Green, M. K.; Lebrilla, C. B. Mass Spectrom. Rev. 1997, 16, 53−
71.
(33) Kostyukevich, Y.; Kononikhin, A.; Popov, I.; Kharybin, O.;
Perminova, I.; Konstantinov, A.; Nikolaev, E. Anal. Chem. 2013, 85,
11007−11013.
(34) Tittebrandt, S.; Edelson-Averbukh, M.; Spengler, B.; Lehmann,
W. D. Angew. Chem., Int. Ed. 2013, 52, 8973−8975.
(35) Chipuk, J. E.; Brodbelt, J. S. Int. J. Mass Spectrom. 2007, 267,
98−108.
(36) Zhang, J. M.; Brodbelt, J. S. J. Am. Chem. Soc. 2004, 126, 5906−
5919.
(37) Campbell, S.; Rodgers, M. T.; Marzluff, E. M.; Beauchamp, J. L.
J. Am. Chem. Soc. 1995, 117, 12840−12854.
(38) Freitas, M. A.; Hendrickson, C. L.; Emmett, M. R.; Marshall, A.
G. Int. J. Mass Spectrom. 1999, 185, 565−575.
(39) Gard, E.; Green, M. K.; Bregar, J.; Lebrilla, C. B. J. Am. Soc. Mass
Spectrom. 1994, 5, 623−631.
(40) Green, M. K.; Lebrilla, C. B. Int. J. Mass Spectrom. Ion Processes
1998, 175, 15−26.
(41) Wood, T. D.; Chorush, R. A.; Wampler, F. M.; Little, D. P.;
O'Connor, P. B.; McLafferty, F. W. Proc. Natl. Acad. Sci. U. S. A. 1995,
92, 2451−2454.
(42) Suckau, D.; Shi, Y.; Beu, S. C.; Senko, M. W.; Quinn, J. P.;
Wampler, F. M.; McLafferty, F. W. Proc. Natl. Acad. Sci. U. S. A. 1993,
90, 790−793.
(43) Wyttenbach, T.; Bowers, M. T. J. Am. Soc. Mass Spectrom. 1999,
10, 9−14.
(44) Valentine, S. J.; Clemmer, D. E. J. Am. Soc. Mass Spectrom. 2002,
13, 506−517.
(45) Valentine, S. J.; Clemmer, D. E. J. Am. Chem. Soc. 1997, 119,
3558−3566.
(46) Demmers, J. A. A.; Rijkers, D. T. S.; Haverkamp, J.; Killian, J. A.;
Heck, A. J. R. J. Am. Chem. Soc. 2002, 124, 11191−11198.
(47) Rand, K. D.; Zehl, M.; Jorgensen, T. J. D. Acc. Chem. Res. 2014,
47, 3018−3027.
6407
Article
(48) Deng, Y. Z.; Pan, H.; Smith, D. L. J. Am. Chem. Soc. 1999, 121,
1966−1967.
(49) Dongre, A. R.; Jones, J. L.; Somogyi, A.; Wysocki, V. H. J. Am.
Chem. Soc. 1996, 118, 8365−8374.
(50) Tsaprailis, G.; Nair, H.; Somogyi, Á .; Wysocki, V. H.; Zhong,
W.; Futrell, J. H.; Summerfield, S. G.; Gaskell, S. J. J. Am. Chem. Soc.
1999, 121, 5142−5154.
(51) Harrison, A. G.; Yalcin, T. Int. J. Mass Spectrom. Ion Processes
1997, 165, 339−347.
(52) McCormack, A. L.; Somogyi, A.; Dongre, A. R.; Wysocki, V. H.
Anal. Chem. 1993, 65, 2859−2872.
(53) Pan, J. X.; Han, J.; Borchers, C. H. Int. J. Mass Spectrom. 2012,
325, 130−138.
(54) Kostyukevich, Y.; Kononikhin, A.; Popov, I.; Nikolaev, E. J. Mass
Spectrom. 2014, 49, 989−994.
(55) Abzalimov, R. R.; Kaplan, D. A.; Easterling, M. L.; Kaltashov, I.
A. J. Am. Soc. Mass Spectrom. 2009, 20, 1514−1517.
(56) Zehl, M.; Rand, K. D.; Jensen, O. N.; Jorgensen, T. J. D. J. Am.
Chem. Soc. 2008, 130, 17453−17459.
(57) Khakinejad, M.; Kondalaji, S. G.; Maleki, H.; Arndt, J. R.;
Donohoe, G. C.; Valentine, S. J. J. Am. Soc. Mass Spectrom. 2014, 25,
2103−2115.
(58) Rand, K. D.; Bache, N.; Nedertoft, M. M.; Jorgensen, T. J. D.
Anal. Chem. 2011, 83, 8859−8862.
(59) Lemaire, P.; Debois, D.; Smargiasso, N.; Quinton, L.; Gabelica,
V.; De Pauw, E. A. Rapid Commun. Mass Spectrom. 2013, 27, 1837−
1846.
(60) Solouki, T.; Fort, R. C.; Alomary, A.; Fattahi, A. J. Am. Soc. Mass
Spectrom. 2001, 12, 1272−1285.
(61) Jurchen, J. C.; Cooper, R. E.; Williams, E. R. J. Am. Soc. Mass
Spectrom. 2003, 14, 1477−1487.
(62) Donohoe, G. C.; Maleki, H.; Arndt, J. R.; Khakinejad, M.; Yi, J.
H.; McBride, C.; Nurkiewicz, T. R.; Valentine, S. J. Anal. Chem. 2014,
86, 8121−8128.
(63) Winger, B. E.; Lightwahl, K. J.; Rockwood, A. L.; Smith, R. D. J.
Am. Chem. Soc. 1992, 114, 5897−5898.
(64) Bohrer, B. C.; Atlasevich, N.; Clemmer, D. E. J. Phys. Chem. B
2011, 115, 4509−4515.
(65) Hoaglund-Hyzer, C. S.; Li, J. W.; Clemmer, D. E. Anal. Chem.
2000, 72, 2737−2740.
(66) Sun, W. J.; May, J. C.; Russell, D. H. Int. J. Mass Spectrom. 2007,
259, 79−86.
(67) Hilderbrand, A. E.; Myung, S.; Clemmer, D. E. Anal. Chem.
2006, 78, 6792−6800.
(68) Kerr, T. J.; Gant-Branum, R. L.; McLean, J. A. Int. J. Mass
Spectrom. 2011, 307, 28−32.
(69) Khakinejad, M.; Kondalaji, S. G.; Donohoe, G. C.; Valentine, S.
J. J. Am. Soc. Mass Spectrom. 2016, 27, 462−473.
(70) Engen, J. R. Anal. Chem. 2009, 81, 7870−7875.
(71) Wells, J. M.; McLuckey, S. A. Methods Enzymol. 2005, 402,
148−185.
(72) Hoaglund, C. S.; Valentine, S. J.; Sporleder, C. R.; Reilly, J. P.;
Clemmer, D. E. Anal. Chem. 1998, 70, 2236−2242.
(73) Eng, J. K.; McCormack, A. L.; Yates, J. R. J. Am. Soc. Mass
Spectrom. 1994, 5, 976−989.
(74) Valentine, S. J.; Counterman, A. E.; Hoaglund-Hyzer, C. S.;
Clemmer, D. E. J. Phys. Chem. B 1999, 103, 1203−1207.
(75) Valentine, S. J.; Counterman, A. E.; Clemmer, D. E. J. Am. Soc.
Mass Spectrom. 1999, 10, 1188−1211.
(76) Valentine, S. J.; Ewing, M. A.; Dilger, J. M.; Glover, M. S.;
Geromanos, S.; Hughes, C.; Clemmer, D. E. J. Proteome Res. 2011, 10,
2318−2329.
(77) Khakinejad, M.; Ghassabi Kondalaji, S.; Tafreshian, A.;
Valentine, S. J. J. Am. Soc. Mass Spectrom. 2017, 28, 960.
DOI: 10.1021/acs.analchem.7b00075
Anal. Chem. 2017, 89, 6399−6407