Spatially resolved protein hydrogen exchange measured by subzero-cooled chip-based nanoelectrospray ionization tandem mass spectrometry

Article abstract of DOI:10.1021/ac300268r

Mass spectrometry has become a valuable method for studying structural dynamics of proteins in solution by measuring their backbone amide hydrogen/deuterium exchange (HDX) kinetics. In a typical exchange experiment one or more proteins are incubated in deuterated buffer at physiological conditions. After a given period of deuteration, the exchange reaction is quenched by acidification (pH 2.5) and cooling (0 °C) and the deuterated protein (or a digest thereof) is analyzed by mass spectrometry. The unavoidable loss of deuterium (back-exchange) that occurs under quench conditions is undesired as it leads to loss of information. Here we describe the successful application of a chip-based nanoelectrospray ionization mass spectrometry top-down fragmentation approach based on cooling to subzero temperature (-15 °C) which reduces the back-exchange at quench conditions to very low levels. For example, only 4% and 6% deuterium loss for fully deuterated ubiquitin and β₂-microglobulin were observed after 10 min of back-exchange. The practical value of our subzero-cooled setup for top-down fragmentation HDX analyses is demonstrated by electron-transfer dissociation of ubiquitin ions under carefully optimized mass spectrometric conditions where gas-phase hydrogen scrambling is negligible. Our results show that the known dynamic behavior of ubiquitin in solution is accurately reflected in the deuterium contents of the fragment ions.

Full text of DOI:10.1021/ac300268r

Article
pubs.acs.org/ac
Spatially Resolved Protein Hydrogen Exchange Measured by
Subzero-Cooled Chip-Based Nanoelectrospray Ionization Tandem
Mass Spectrometry
Sabine Amon, Morten B. Trelle, Ole N. Jensen, and Thomas J. D. Jørgensen*
Department of Biochemistry and Molecular Biology, University of Southern Denmark, Campusvej 55, DK-5230 Odense M, Denmark
S
* Supporting Information
ABSTRACT: Mass spectrometry has become a valuable method for studying structural
dynamics of proteins in solution by measuring their backbone amide hydrogen/
deuterium exchange (HDX) kinetics. In a typical exchange experiment one or more
proteins are incubated in deuterated buffer at physiological conditions. After a given
period of deuteration, the exchange reaction is quenched by acidification (pH 2.5) and
cooling (0 °C) and the deuterated protein (or a digest thereof) is analyzed by mass
spectrometry. The unavoidable loss of deuterium (back-exchange) that occurs under
quench conditions is undesired as it leads to loss of information. Here we describe the
successful application of a chip-based nanoelectrospray ionization mass spectrometry
top-down fragmentation approach based on cooling to subzero temperature (−15 °C)
which reduces the back-exchange at quench conditions to very low levels. For example, only 4% and 6% deuterium loss for fully
deuterated ubiquitin and β2-microglobulin were observed after 10 min of back-exchange. The practical value of our subzero-
cooled setup for top-down fragmentation HDX analyses is demonstrated by electron-transfer dissociation of ubiquitin ions under
carefully optimized mass spectrometric conditions where gas-phase hydrogen scrambling is negligible. Our results show that the
known dynamic behavior of ubiquitin in solution is accurately reflected in the deuterium contents of the fragment ions.
he conformational dynamics of proteins are essential for
graphy−mass spectrometry (LC−MS).6,10 Importantly, the
Ttheir function. For example, molecular recognition chromatographic separation of the labeled peptides must be
carried out at quench conditions and as fast as possible to
minimize the unavoidable loss of deuterium caused by back-
exchange with the HPLC solvents. The occurrence of back-
exchange has been described as “The single biggest problem
with solution-phase H/D exchange” by Emmett et al.11 Direct
fragmentation of the labeled protein in the mass spectrometer
by electrons12−21 or matrix-assisted laser desorption ionization
in source decay (MALDI ISD)22 is an attractive alternative
approach to the traditional protease-based bottom-up method.
The top-down approach omits the need for in solution
digestion as well as chromatographic separation. Therefore, it
opens for the possibility to attain very low levels of back-
exchange. One analytical strategy is based on minimizing the
back-exchange time by having a continuous flow of the protein
in the labeling buffer where the quench solution is mixed into
the flow just prior to entering the ion source.15,16 The
acidification upon quenching causes the protein ions to attain
higher charge states in the electrospray process, and this is
highly beneficial for the electron-based fragmentation as higher
charge states fragment more efficiently than lower charge
states.23,24 This continuous flow strategy is, however, only
compatible with volatile buffers as the flow of the exchange
buffer mixed with quench buffer enters the ion source without a
depends on conformational plasticity in the binding interface
when a protein adapts its structure upon binding to different
binding partners.1 Allosteric regulation of protein activity also
relies on structural flexibility.2−4 Hydrogen/deuterium ex-
change monitored by mass spectrometry has recently become
an important technique for characterizing protein structural
dynamics.5,6 The basis of this method is that backbone amide
hydrogens that are engaged in stable hydrogen bonds in the
natively folded conformation are protected against exchange
with the solvent.7 According to the thermodynamic principle
that proteins unfold and refold continuously, the protecting
hydrogen bonds will break when the protein cycles through
locally (and globally) unfolded states thereby allowing exchange
to occur. The exchange kinetics of amide backbone hydrogens
is therefore a direct probe for the structural dynamics of the
protein backbone. When a natively folded protein is incubated
in D2O, it will gradually exchange its amide hydrogens with
deuterium. The protein mass increases by 1 Da for each
exchange event, and the global deuterium uptake is thus readily
measured by mass spectrometry.8 The traditional approach9 to
localize the incorporated deuterium is based on proteolytic
cleavage of the deuterium-labeled protein at cold acidic
conditions where the isotopic exchange rate is slowed down
several orders of magnitude (i.e., quench conditions). Pepsin or
other proteases with a broad specificity are used to digest the
labeled protein yielding a relatively complex mixture of labeled
peptides which are typically analyzed by liquid chromato-
Received: January 26, 2012
Accepted: March 30, 2012
Published: April 26, 2012
© 2012 American Chemical Society
4467
dx.doi.org/10.1021/ac300268r | Anal. Chem. 2012, 84, 4467−4473

Analytical Chemistry
Article
possibility for desalting. Another strategy25 is to avoid
quenching in solution altogether by electrospraying the
exchange solution and thereby monitoring the protein
hydrogen exchange kinetics in real time20 (the isotopic
exchange reaction in solution is quenched when the protein
ions are desolvated in the ion source). To ensure efficient
fragmentation, the charge states are increased by having a
supercharging reagent (m-nitrobenzyl alcohol, m-NBA)26 in the
exchange buffer. The presence of this reagent at a relatively
high concentration (1%) may, however, affect the protein
stability and hence its dynamics. The real-time strategy is also
limited to volatile buffers. In the continuous flow strategy, the
protein should be exposed to quench conditions as short time
as possible to minimize back-exchange, whereas the real-time
strategy completely omits the quench step. Here we propose an
alternative analytical strategy for minimizing back-exchange in
top-down fragmentation studies by quenching at subzero
temperature. We utilize a chip-based nanoelectrospray
ionization setup (Advion TriVersa NanoMate) equipped with
a device that has been especially adapted to minimize the level
of back-exchange by cooling the quenched sample to −15 °C.
Under these conditions, the level of back-exchange is reduced
to a very low level even after the protein has been stored in the
quench solution for a prolonged period (e.g., only 4%
deuterium loss for fully deuterated ubiquitin after 10 min of
back-exchange). This allows for relatively long acquisition times
to accumulate signals from low-abundant deuterium-labeled
fragment ions. Our setup offers the same advantages as
nanoelectrospray needles over conventional microflow electro-
spray, such as lower sample consumption, higher sensitivity,
and no risk of carry over. Moreover, the chip-based setup offers
some additional advantages over nanoelectrospray needles such
as high spray reproducibility and spray stability and the
possibility for automation.27,28 Furthermore, a desalting step is
easy to implement in the hydrogen/deuterium exchange mass
spectrometry (HDX-MS) workflow if nonvolatile salts are
present in the exchange buffers. The protein sample could be
rapidly desalted by reversed-phase chromatography at normal
quench conditions and eluted by an acidic methanolic solution
directly into the 96-well plate of the NanoMate.
(Advion BioSciences, Inc., Ithaca, NY, U.S.A.) for subsequent
analysis by mass spectrometry (see below for a further
description of the system). To measure the initial deuterium
content of ubiquitin (i.e., at tex = 0 min), an aliquot of fully
deuterated ubiquitin was transferred by 100-fold dilution
directly to the ice-cold quench solution (1:1 (v/v) water/
methanol mixture with 0.20 M formic acid) and analyzed
immediately. As a reference, equilibrium D-to-H exchanged
ubiquitin was prepared by 50-fold dilution of fully deuterated
ubiquitin into 1H2O at 55 °C for 3 days. This sample represents
a homogeneously labeled protein with 2% deuterium at all
labile sites. To measure the level of back-exchange, an aliquot of
fully deuterated ubiquitin was transferred by 100-fold dilution
directly to the ice-cold quench solution (1:1 (v/v) water/
methanol mixture with 0.20 M formic acid) and immediately
analyzed with or without cooling of the front end of the
NanoMate pipet tip (see below). The dead time in these
measurements with the NanoMate system was 30 s (i.e., the
time from loading the sample solution into the 96-well plate
until the sample is spraying). Without cooling, the sample
solution will reach room temperature (22 °C). Ubiquitin
contains 76 residues and 72 backbone amide hydrogens (76
residues − 1 N-terminal residue − 3 Pro residues = 72
backbone NH). Fully deuterated β2-microglobulin was
prepared by dissolving lyophilized β2-microglobulin (1.0 mM)
in deuterated buffer (100 mM ammonium-d4 acetate-d3) and
incubating for at least 6 h at 37 °C as previously described.29
Global back-exchange kinetics was measured by a 100-fold
dilution of fully deuterated β2-microglobulin into ice-cold
quench solution (1:1 (v/v) water/methanol mixture with 0.20
M formic acid) and immediately analyzed with or without
cooling of the front end of the NanoMate pipet tip. β2-
Microglobulin has a total of 93 exchangeable backbone amide
hydrogens (99 residues − 1 N-terminal residue − 5 Pro
residues = 93 backbone NH). The global back-exchange
measurements measure predominantly the deuterium content
of the backbone amides. However, the ε-NH in the side chain
of arginine contributes to the deuterium content at short
exchange times, as its exchange rate is only ∼10-fold higher
than that of the poly-D/L-alanine backbone amides at quench
conditions.30
Mass Spectrometry. Samples were introduced into a
Thermo LTQ Orbitrap XL equipped with electron-transfer
dissociation (ETD) (Thermo Fisher Scientific, Waltham, MA,
U.S.A.) via a TriVersa NanoMate system operated in the
infusion mode. An amount of 5 μL of sample solution was
picked up from the 96-well plate by a disposable conductive
pipet tip which subsequently docked onto the electrospray
ionization (ESI) chip (A-chip with a nozzle of i.d. 5.5 μm). For
spraying a voltage of 1.25 kV and a nitrogen pressure of 0.3 psi
were applied. During the spraying process the pipet tip of the
NanoMate was cooled in order to minimize back-exchange.
This was achieved by means of a flow of cooled nitrogen (see
below for a further description). The ion source parameters of
the mass spectrometer were kept at gentle declustering
conditions in order to avoid hydrogen scrambling during the
ionization process (capillary voltage, 30 V; capillary temper-
ature, 100 °C; tube lens potential, 100 V). MS and MS/MS
spectra were acquired with a resolution of 100 000 and an
automated gain control of 1 × 106 for MS and 2 × 105 for MS/
MS spectra. The maximum allowed fill time for MS spectra was
500 ms, and for MS/MS spectra it was 1000 ms. The scan
range was set to m/z 500−2000 for full scan spectra and m/z
EXPERIMENTAL SECTION
■
Materials and Reagents. Bovine ubiquitin (BioUltra,
purity ≥98%), methanol (LC−MS grade), formic acid (p.A.
grade), ammonium-d4 acetate-d3 (99 atom % D), acetic-d3 acid-
d (99.5 atom % D), and methyl alcohol-d (99.5 atom % D)
were obtained from Sigma-Aldrich (St. Louis, MO, U.S.A.).
D2O (99.9% atom% D) was purchased from Cambridge
Isotope Laboratories, Inc. (Andover, MA, U.S.A.). Water was
purified by a Millipore Milli-Q system (Bedford, MA, U.S.A.).
Human β2-microglobulin was obtained from Innovative
Research (Novi, MI, U.S.A.).
Hydrogen/Deuterium Exchange. Bovine ubiquitin (500
μM) was fully deuterated in 3% acetic-d3 acid-d in a 1:1 (v/v)
mixture of D2O and methyl alcohol-d for 3 days at 37 °C,
followed by lyophilization, dissolution in 50 mM ammonium-d4
acetate-d3, and subsequent incubation for 3 days at 37 °C.
Native state D-to-H exchange was initiated by 50-fold dilution
into 1H2O at 25.0 °C, pH 6.6. After the desired exchange time,
the exchange reaction was quenched by mixing 1:1 (v/v) with
ice-cold 0.40 M formic acid in methanol. An aliquot of the
quenched sample solution (20 μL) was directly pipetted into a
96-well plate cooled to 4 °C by the TriVersa NanoMate system
4468
dx.doi.org/10.1021/ac300268r | Anal. Chem. 2012, 84, 4467−4473

Analytical Chemistry
Article
200−1800 for MS/MS spectra. MS/MS experiments of
ubiquitin were carried out on charge state 12+, where the
precursor isolation window was set to 20 m/z in order to avoid
side band activation resulting in hydrogen scrambling. ETD was
performed at a reaction time of 10 ms using a target population
of fluoranthene anions of 2 × 105 at a maximum fill time of 100
ms. The acquisition time was 0.5 min for MS spectra and 1 min
for MS/MS spectra (35 scans), respectively. Fragmentation of
deuterated ubiquitin by ETD yielded 44 c and 32 z fragment
ions with sufficiently high signal-to-noise ratios (S/N) (>5) to
allow a reliable determination of their deuterium content
(Figure S1, Supporting Information). The signal intensity of
the base peak (z173+) in the ETD spectra of [ubiquitin +
12H]12+ was approximately 105. The signal intensity of the
fragment ions that were used for the calculation of deuterium
content was 104 or higher.
tip. A flexible dial indicator holder (from a lathe) mounted on
the NanoMate allowed easy positioning of the Teflon tubing.
The nitrogen flow rate was 3 L/min, and the temperature of the
gas was approximately −15 °C measured at the location of the
tip. This experimental setup ensures efficient cooling of a
sample volume of approximately 2 μL, sufficient for
approximately 15−20 min of measuring time.
Data Analysis. Mass spectra and peak lists were processed
by Xcalibur version 2.1 (Thermo Fisher Scientific). Isotopic
patterns of the fragment ions were extracted and the intensity-
weighted average mass for each fragment ion was calculated by
an in-house written spreadsheet macro (Microsoft Excel,
Redmond, WA). The deuterium content of the fragment ions
was calculated by the mass difference between the fragment ion
obtained from deuterium-labeled ubiquitin and the correspond-
ing fragment ion obtained from equilibrium D-to-H exchanged
ubiquitin (i.e., homogeneously labeled ubiquitin with 2%
deuterium). The values in the plots represent the average of
at least three independent measurements. For reasons of clarity,
error bars are only displayed if the length of the error bar is
larger than the data marker symbol.
Cooling Device. We modified the NanoMate by
introducing a flow of cold nitrogen gas to the front of the
pipet tip that is in direct contact with the ESI chip (as shown in
Figure 1). The flow of cold nitrogen gas was generated by
RESULTS AND DISCUSSION
■
Back-Exchange for Intact Proteins. Top-down fragmen-
tation of deuterium-labeled proteins from solution-phase HDX
experiments holds great promise as an analytical tool to
investigate protein structural dynamics with high spatial
resolution. One of the benefits of this approach is that the
labeled residues are localized by fragmentation in the solvent-
free gas phase where back-exchange does not occur. However,
it is not possible to avoid the undesired loss or gain of
deuterium due to back-exchange with the solvent in the
quenched sample solution (i.e., before the sample is transferred
to the gas phase). Our approach to limit the undesired back-
exchange in the quenched sample solution is to cool it to
subzero temperature (−15 °C) while it is electrospraying. We
cooled the NanoMate pipet tip containing the quenched
sample solution by directing a flow of cold nitrogen gas to the
front end of the pipet tip. The efficiency of the cooled
NanoMate setup in regard to minimizing the deuterium loss
was investigated by monitoring the global back-exchange
kinetics of fully deuterated ubiquitin in a quench solution
with and without cooling. Figure 2 shows that cooling is highly
efficient in reducing the back-exchange of ubiquitin to a very
low level (4% deuterium loss after 10 min of back-exchange). A
close inspection of the data points reveal that ∼71 deuterons
(out of a total number of 72) are retained in the first half
minute. Even after 10 and 15 min of back-exchange, ∼69 and
∼68 deuterons are retained, respectively. Without cooling, the
level of back-exchange is dramatically increased, and after just
half a minute significantly less deuterons are retained (∼63)
than after 15 min with cooling (∼68). Further deuterium
depletion occurs rapidly without cooling and after 10 and 15
min, only ∼27 and ∼22 deuterons are retained, respectively.
Clearly, our cooling device efficiently reduces the level of back-
exchange and allows ample time for accumulating ETD spectra
of the labeled protein. Although ubiquitin rapidly looses
deuterons without cooling at quench conditions, it is known to
attain a partially folded conformation (called the A state) in
water/alcohol mixtures at acidic conditions.31−34 This partially
folded state will slow down the global back-exchange kinetics
relative to a random coil structure. Therefore, we also
investigated the back-exchange kinetics of β2-microglobulin
Figure 1. Schematic drawing of the cooling device adapted to a chip-
based nanoflow electrospray ionization (ESI) setup. A flow of nitrogen
gas cooled to subzero temperature is generated by passing it through a
coil of copper tubing immersed in dry ice. The inset shows how the
cold gas flow is directed to the front of the pipet tip (containing the
sample solution) that is in direct contact with the ESI chip. The
labeled parts are (a) entrance cone of Orbitrap mass spectrometer, (b)
holder for ESI chip, (c) ESI chip, (d) holder for conductive pipet tip,
(e) conductive pipet tip, (f) outlet tubing for cold nitrogen gas, (g)
copper tubing, (h) dry ice, and (i) thermoinsulated box. Only the front
end parts of the NanoMate are shown.
passing nitrogen through a coil of cooled copper tubing (length
20 m, i.d. 2 mm). The tubing was cooled by dry ice in an
insulated container (inner dimensions approximately 30 × 25 ×
20 cm3). The copper tubing outlet was mantled with thermal
insulation tubing (EPDM rubber tube, wall thickness 2 cm). A
short noninsulated Teflon tubing (10 cm, i.d. 2 mm) directed
the cold nitrogen gas flow to the NanoMate pipet tip
containing the sample solution for electrospraying. The tubing
was carefully positioned approximately 2 mm above the pipet
4469
dx.doi.org/10.1021/ac300268r | Anal. Chem. 2012, 84, 4467−4473

Analytical Chemistry
Article
crosses that of ubiquitin already after 1 min clearly reflecting a
higher degree of unfolding of β2-microglobulin at quench
conditions (Figure 2). This underscores the efficiency of
cooling to subzero degrees with respect to limiting back-
exchange at quench conditions in top-down fragmentation
experiments.
Importantly, the subzero temperature cooling does not have
any adverse effect on the stability of the electrospray (Figure
S2, Supporting Information). The cooled front end of the pipet
tip contains approximately ∼2 μL which allows for approx-
imately 15 min of measurement time for mass spectrometric
experiments. After ∼15 min of spraying, the remainder of
sample solution which was initially not cooled enters the
electrospray nozzle and this is readily detected by the
appearance of a bimodal peak pattern (Figure S3, Supporting
Information). The high-mass peak in the bimodal distribution
represents the labeled protein molecules which were initially
present in the cooled volume, whereas the low-mass peak
represents protein molecules that were initially outside the
cooled volume and hence have undergone a higher degree of
back-exchange. Note that such a bimodal peak pattern was also
observed for peptides (data not shown).
Figure 2. Time course of back-exchange. The deuterium content of
ubiquitin (circles) and β2-microglobulin (triangles) is shown as a
function of exchange time in quench solution with (closed symbols) or
without (open symbols) cooling to subzero temperature (−15 °C).
Error bars represent average
SD for measurements made in
quadruplicate. Note: the back-exchange time includes the instrumental
dead time (0.5 min) and the acquisition time (0.5 min).
Spatially Resolved Hydrogen/Deuterium Exchange
Measurements of Ubiquitin. Ubiquitin is a well-charac-
terized protein, and its backbone amide hydrogen exchange rate
constants under various conditions are known from NMR
studies.32,38,39 Therefore, we selected ubiquitin as a model
protein for our present HDX ETD top-down fragmentation
study. The subzero cooling did not affect the charge state
which is more highly unfolded at low pH as well as upon
addition of organic solvent.35−37 With cooling, the deuterium
loss from β2-microglobulin is only slightly higher than that of
ubiquitin (2% point higher loss after 10 min of back-exchange).
Without cooling, the back-exchange curve of β2-microglobulin
Figure 3. Electrospray ionization mass spectra of ubiquitin sprayed from a solution at subzero temperature (−15 °C). The charge state distribution
of ubiquitin is shown in panel a. Electron-transfer dissociation (ETD) mass spectrum of the +12 charge state of ubiquitin is shown in panel b; the
asterisked peaks correspond to charge-reduced protein ions. Panel c shows the +12 charge state after D-to-H exchange for various periods of time,
while panel d shows the corresponding ETD spectra displaying the isotopic envelopes of fragment ions z4+ and c3+ originating from a fast and slow
exchanging region in ubiquitin, respectively.
4470
dx.doi.org/10.1021/ac300268r | Anal. Chem. 2012, 84, 4467−4473

Analytical Chemistry
Article
distribution of ubiquitin or the signal intensity (Figure S2,
Supporting Information). The ESI spectrum of ubiquitin in
Figure 3a shows a charge state distribution dominated by highly
charged ions which are characteristic for proteins electro-
sprayed at denaturing conditions. Fragmentation of the four
highest charge states (+13 to +10) of ubiquitin by ETD yielded
similar fragmentation patterns (data not shown). This was also
observed by Sterling and Williams.20 Figure 3b shows the ETD
spectrum of the +12 charge state which yielded 59 backbone
bond cleavages representing 82% of the cleavable peptide
bonds. This allows for a detailed analysis of the deuterium
levels of backbone amides in the protein. Native-state D-to-H
exchange was carried out by incubating fully deuterated
1
ubiquitin in H2O at neutral pH. Under these conditions, the
most stable regions in ubiquitin are protected against isotopic
1
exchange with H2O and they will retain their deuterium,
whereas the more dynamic and flexible regions undergo
exchange. The initial deuterium content of ubiquitin was 71
D (i.e., at tex = 0 min) (Figure 3c, upper spectrum). After 60
min of D-to-H exchange, ubiquitin retained 26 deuterium
atoms (Figure 3c, middle spectrum). The retained deuterium
can be assigned to individual residues by the mass shift of the
fragment ions obtained from ETD fragmentation. To illustrate
this, Figure 3d displays the isotopic envelope of the c3 and z4
ion obtained from ubiquitin after D-to-H exchange for 60 min.
Also shown are the isotopic envelopes for tex = 0 min displaying
the initial deuterium content (2.5 and 3.4 D for c3 and z4) and
for tex = 72 h at 55 °C displaying the exchange end point (i.e.,
homogeneously labeled ubiquitin with ∼2% deuterium). The c3
ion encompasses the first three N-terminal backbone amides
(i.e., Q2, I3, and F4) of which the latter two are highly
protected against exchange as they are part of a stable β-sheet in
the natively folded protein (protection factors are larger than
105 as determined by NMR).38,39 In contrast, the three C-
terminal backbone amides of the z4 ion are rapidly exchanging
as they belong to the highly dynamic C-terminal region of
ubiquitin.38,40,41 Accordingly, Figure 3d shows that the c3 ion
has retained ∼1.9 D after 60 min of exchange, whereas the z4
ion is devoid of deuterium. These values are in excellent
agreement with those calculated from known NMR exchange
rates, i.e., 2.0 and 0.0 D (values are not corrected for back-
exchange). The deuterium content of the c ion series up to and
including c40 is plotted in Figure 4. A nearly perfect correlation
is observed between the deuterium content of the c ion series
and the graph showing the cumulative deuterium content
obtained from known NMR exchange rate constants39
(compare filled circles with red graph in Figure 4). The
cumulative deuterium graph exhibits a distinctive profile, where
the horizontal region of the graph from c6 to c11 corresponds to
a flexible region in ubiquitin that is devoid of deuterium (a turn
between β-strands 1 and 2), whereas the neighboring protected
regions (β-strands 1 and 2) retain deuterium and exhibit a
nearly linear increase. The known solution behavior of
ubiquitin is thus readily apparent by the top-down
fragmentation data. A nearly horizontal region from c30 to c40
corresponds to the loop between the central α-helix α1 and β-
strand 3. Interestingly, the first horizontal region (i.e., turn
β1−β2) has recently been reported to undergo a large
amplitude pincer-like movement toward the loop α1-β3 in
solution.41 This motional mode is essential for molecular
recognition by enabling ubiquitin to adapt to different protein
binding partners.41
Figure 4. Dissecting the solution structural dynamics of the N-
terminal half of ubiquitin. Deuterium levels of the c fragment ions of
ubiquitin after native state D-to-H exchange for 60 min measured by
electron-transfer dissociation mass spectrometry (filled circles) and
NMR (red line). Error bars represent average SD for measurements
made in triplicate. NMR deuterium levels were obtained from
exchange rate constants determined by Johnson et al. (ref 39). The
theoretical deuterium content in the case of 100% gas-phase H/D
scrambling is indicated (dotted gray line). The inset displays the
crystal structure of ubiquitin (PDB 1UBQ). The turn between β-
strands 1 and 2 and the loop between the central α-helix α1 and β-
strand are highlighted in red (see text for further information).
Regions of secondary structure of the N-terminal half are illustrated
above the figure for reference.
Hydrogen scrambling is a major concern when using mass
spectrometric fragmentation as an experimental tool to
determine deuterium levels of individual residues in proteins
and peptides.22,29,42−49 The level of hydrogen scrambling in our
measurements is clearly negligible, as our experimental data
closely resembles the known solution deuteration profile and
not the theoretical graph for 100% scrambling (Figure 4). The
deuterium content of the z ions correlates well with that of the
c ions (which is in excellent agreement with the known solution
behavior) corroborating that hydrogen scrambling is negligible
(Figure S4, Supporting Information).
To measure the site-specific level of back-exchange, fully
deuterated ubiquitin was analyzed by ETD top-down
fragmentation under quench conditions at subzero degrees.
The measured deuterium levels exhibit a nearly linear increase
with residue number (Figure 5), as expected for a fully
deuterated uniform labeling at all backbone amides. The black
graph in Figure 5 shows the theoretical cumulative deuterium
content of all backbone amides assuming an average deuterium
loss of 3% per residue (except for the N-terminal amide at
residue Q2 where the loss was 45% due to the accelerating
effect of the positively charged N-terminus).30 The very low
level of back-exchange at individual residues (3%) corroborates
the efficiency of our chip-based setup with subzero cooling. The
small horizontal segments of the graph correspond to proline
residues which lack amide hydrogens.
The spatial resolution (the ability to obtain site-specific
deuterium levels) in top-down fragmentation HDX experi-
ments of small proteins approaches single-residue resolution.
For electron-based fragmentation of proteins, however, the
proportion of observed cleaved interresidue bonds decreases
4471
dx.doi.org/10.1021/ac300268r | Anal. Chem. 2012, 84, 4467−4473

Analytical Chemistry
Article
method;20 however, fast exchanging amides are amenable for
analysis by our approach as the exchange reaction is efficiently
quenched during ETD spectrum acquisition. If higher charge
states are desired for ETD fragmentation then a supercharging
reagent26 could be added to the quench solution. Importantly,
any denaturing effect of the supercharging reagent54 would not
affect the native state kinetic measurements as the reagent is
only present in the quench buffer. Although it is not important
for our method whether a protein partially retains its structure
or completely unfolds at the present quench conditions, it is
nevertheless interesting to speculate how these conditions affect
protein structure. The acidic pH and the presence of organic
solvent are in general destabilizing conditions for the native
structure. Furthermore, cold denaturation can also occur at low
temperature. For example, cold denaturation of natively folded
ubiquitin (at pH 5.0) is expected to occur around −50 °C55
and the presence of denaturant (guanidinium hydrochloride)
increases the cold denaturation temperature.56 On the other
hand, subzero temperature has been reported to preserve
enzyme activity in methanolic solutions57 and several organic
solvents have been used as cryosolvents to study slowed-down
enzyme-catalyzed reactions at temperatures below 0 °C.58 In
our setup, the presence of methanol serves several important
purposes. It is an antifreeze agent, and it enhances the
electrospray ionization process by reducing surface tension.
Methanol also reduces the back-exchange rate as it lowers
concentration of the exchange catalyst ions of water, i.e., H+
and OH−.59 In this regard, the back-exchange rates at quench
conditions could be even further reduced by decreasing the
temperature and increasing the concentration of methanol. For
example, a 90% methanol solution with 0.2 M formic acid (v/v)
can be cooled to the temperature of dry ice (−78 °C) without
freezing. As back-exchange is very efficiently reduced by
subzero cooling, we envision that this will be used as a general
method for acquiring top-down fragmentation MS spectra of
deuterium-labeled proteins. In this regard, we note that it
should be possible to mount our cooling device on other types
of nanoESI sources.
Figure 5. Spatially resolved deuterium content of fully deuterated
ubiquitin at subzero temperature. Deuterium levels of the c fragment
ions (black circles) and z fragment ions (red triangles) obtained from
fully deuterated ubiquitin are plotted with the graph indicating the
cumulative deuterium content of backbone amides assuming an
average deuterium loss of 3% per residue. Error bars represent average
SD for measurements made in triplicate.
with protein size.50 Consequently, the spatial resolution in top-
down fragmentation HDX experiments decreases with protein
size. For larger proteins, supplemental activation is required in
order to observe the fragment ions generated by ETD or
electron capture dissociation (ECD).51 Importantly, supple-
mental activation of proteins should not be expected to induce
scrambling as supplemental activation of peptides (after having
received electrons) appears to be sufficiently gentle to avoid the
occurrence of scrambling.52
CONCLUSION
■
The deuterium loss caused by back-exchange at quench
conditions is of major concern in HDX-MS analyses as it
leads to loss of information. Here we demonstrate that very low
levels of back-exchange can be obtained by cooling the
quenched sample to subzero temperature (−15 °C) while it
is analyzed with a chip-based nanoelectrospray ionization setup.
It ensures a nearly constant deuterium content of the protein
throughout the measurement, and this allows for prolonged
acquisition time of ETD spectra at quench conditions so that
good fragment ion statistics can be obtained. This is essential
for an accurate determination of the deuterium content of the
fragment ions. Furthermore, with a prolonged acquisition time
the effect of various instrumental settings on the level of
scrambling is readily explored. Our approach is particularly
useful for electron-based top-down fragmentation of deute-
rium-labeled proteins. For the traditional bottom-up approach
(based on pepsin digestion) subzero cooling is not feasible as
the acidic aqueous mobile phase has a freezing point close to
zero degrees. We note, however, that transient supercooling can
occur in refrigerated HPLC systems.53 As the exchange reaction
is efficiently quenched at subzero temperature in our setup, it is
possible to acquire and average scans for up to 15 min for a
sample corresponding to a single exchange time point. In the
real-time kinetic data measuring method,20 native state
exchange proceeds while ETD spectra are acquired and
averaged for 2 min for each exchange time point (the
experimental dead time is also ∼2 min). This compromises
kinetic analysis of fast exchanging amides by the real-time
ASSOCIATED CONTENT
■
S
* Supporting Information
Additional information as noted in text. This material is
available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: tjdj@bmb.sdu.dk.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors thank Torben Christensen, Klaus Lehn Petersen,
and Dr. Christer Ejsing (University of Southern Denmark) for
excellent technical assistance. We gratefully acknowledge
financial support by the Danish Council for Independent
Research in Natural Sciences (FNU Grant No. 09-070686,
T.J.D.J. and S.A.). This work was also supported by a Grant
from the Danish National Research Foundation to the Center
for Epigenetics (O.N.J.).
4472
dx.doi.org/10.1021/ac300268r | Anal. Chem. 2012, 84, 4467−4473

Analytical Chemistry
Article
(35) Katou, H.; Kanno, T.; Hoshino, M.; Hagihara, Y.; Tanaka, H.;
Kawai, T.; Hasegawa, K.; Naiki, H.; Goto, Y. Protein Sci. 2002, 11,
2218−2229.
(36) Platt, G. W.; McParland, V. J.; Kalverda, A. P.; Homans, S. W.;
Radford, S. E. J. Mol. Biol. 2005, 346, 279−294.
(37) Jorgensen, T. J. D.; Cheng, L.; Heegaard, N. H. H. Int. J. Mass
Spectrom. 2007, 268, 207−216.
(38) Bougault, C.; Feng, L. M.; Glushka, J.; Kupce, E.; Prestegard, J.
H. J. Biomol. NMR 2004, 28, 385−390.
(39) Johnson, E. C.; Lazar, G. A.; Desjarlais, J. R.; Handel, T. M.
Struct. Folding Des. 1999, 7, 967−976.
(40) Brand, T.; Cabrita, E. J.; Morris, G. A.; Gunther, R.; Hofmann,
H. J.; Berger, S. J. Magn. Reson. 2007, 187, 97−104.
(41) Lange, O. F.; Lakomek, N. A.; Fares, C.; Schroder, G. F.; Walter,
K. F. A.; Becker, S.; Meiler, J.; Grubmuller, H.; Griesinger, C.; de
Groot, B. L. Science 2008, 320, 1471−1475.
(42) Bache, N.; Rand, K. D.; Roepstorff, P.; Jorgensen, T. J. D. Anal.
Chem. 2008, 80, 6431−6435.
REFERENCES
■
(1) James, L. C.; Roversi, P.; Tawfik, D. S. Science 2003, 299, 1362−
1367.
(2) Hauge, C.; Antal, T. L.; Hirschberg, D.; Doehn, U.; Thorup, K.;
Idrissova, L.; Hansen, K.; Jensen, O. N.; Jorgensen, T. J.; Biondi, R.
M.; Frodin, M. EMBO J. 2007, 26, 2251−2261.
(3) Hindie, V.; Stroba, A.; Zhang, H.; Lopez-Garcia, L. A.; Idrissova,
L.; Zeuzem, S.; Hirschberg, D.; Schaeffer, F.; Jorgensen, T. J. D.; Engel,
M.; Alzari, P. M.; Biondi, R. M. Nat. Chem. Biol. 2009, 5, 758−764.
(4) Rand, K. D.; Jorgensen, T. J. D.; Olsen, O. H.; Persson, E.;
Jensen, O. N.; Stennicke, H. R.; Andersen, M. D. J. Biol. Chem. 2006,
281, 23018−23024.
(5) Konermann, L.; Pan, J. X.; Liu, Y. H. Chem. Soc. Rev. 2011, 40,
1224−1234.
(6) Wales, T. E.; Engen, J. R. Mass Spectrom. Rev. 2006, 25, 158−170.
(7) Milne, J. S.; Mayne, L.; Roder, H.; Wand, A. J.; Englander, S. W.
Protein Sci. 1998, 7, 739−745.
(8) Katta, V.; Chait, B. T. Rapid Commun. Mass Spectrom. 1991, 5,
214−217.
(43) Bache, N.; Rand, K. D.; Roepstorff, P.; Ploug, M.; Jorgensen, T.
J. D. J. Am. Soc. Mass Spectrom. 2008, 19, 1719−1725.
(44) Jorgensen, T. J. D.; Bache, N.; Roepstorff, P.; Gardsvoll, H.;
Ploug, M. Mol. Cell. Proteomics 2005, 4, 1910−1919.
(45) Rand, K. D.; Adams, C. M.; Zubarev, R. A.; Jorgensen, T. J. D. J.
Am. Chem. Soc. 2008, 130, 1341−1349.
(9) Zhang, Z.; Smith, D. L. Protein Sci. 1993, 2, 522−531.
(10) Johnson, R. S.; Walsh, K. A. Protein Sci. 1994, 3, 2411−2418.
(11) Emmett, M. R.; Kazazic, S.; Marshall, A. G.; Chen, W.; Shi, S. D.
H.; Bolanos, B.; Greig, M. J. Anal. Chem. 2006, 78, 7058−7060.
(12) Abzalimov, R. R.; Kaplan, D. A.; Easterling, M. L.; Kaltashov, I.
A. J. Am. Soc. Mass Spectrom. 2009, 20, 1514−1517.
(13) Charlebois, J. P.; Patrie, S. M.; Kelleher, N. L. Anal. Chem. 2003,
75, 3263−3266.
(46) Rand, K. D.; Jorgensen, T. J. D. Anal. Chem. 2007, 79, 8686−
8693.
(47) Rand, K. D.; Lund, F. W.; Amon, S.; Jorgensen, T. J. D. Int. J.
Mass Spectrom. 2011, 302, 110−115.
(14) Konermann, L.; Tong, X.; Pan, Y. J. Mass Spectrom. 2008, 43,
1021−1036.
(48) Rand, K. D.; Zehl, M.; Jensen, O. N.; Jorgensen, T. J. D. Anal.
Chem. 2010, 82, 9755−9762.
(49) Zehl, M.; Rand, K. D.; Jensen, O. N.; Jorgensen, T. J. D. J. Am.
Chem. Soc. 2008, 130, 17453−17459.
(15) Pan, J.; Han, J.; Borchers, C. H.; Konermann, L. J. Am. Chem.
Soc. 2008, 130, 11574−11575.
(50) Zubarev, R. A.; Horn, D. M.; Fridriksson, E. K.; Kelleher, N. L.;
Kruger, N. A.; Lewis, M. A.; Carpenter, B. K.; McLafferty, F. W. Anal.
Chem. 2000, 72, 563−573.
(16) Pan, J. X.; Han, J.; Borchers, C. H.; Konermann, L. J. Am. Chem.
Soc. 2009, 131, 12801−12808.
(17) Pan, J. X.; Han, J.; Borchers, C. H.; Konermann, L. Anal. Chem.
2010, 82, 8591−8597.
(51) Horn, D. M.; Ge, Y.; McLafferty, F. W. Anal. Chem. 2000, 72,
4778−4784.
(18) Pan, J. X.; Han, J.; Borchers, C. H.; Konermann, L. Anal. Chem.
2011, 83, 5386−5393.
(52) Rand, K. D.; Pringle, S. D.; Morris, M.; Engen, J. R.; Brown, J.
M. J. Am. Soc. Mass Spectrom. 2011, 22, 1784−1793.
(53) Keppel, T. R.; Jacques, M. E.; Young, R. W.; Ratzlaff, K. L.;
Weis, D. D. J. Am. Soc. Mass Spectrom. 2011, 22, 1472−1476.
(54) Sterling, H. J.; Daly, M. P.; Feld, G. K.; Thoren, K. L.; Kintzer,
A. F.; Krantz, B. A.; Williams, E. R. J. Am. Soc. Mass Spectrom. 2010, 21,
1762−1774.
(19) Stefanowicz, P.; Petry-Podgorska, I.; Kowalewska, K.; Jaremko,
L.; Jaremko, M.; Szewczuk, Z. Biosci. Rep. 2010, 30, 91−99.
(20) Sterling, H. J.; Williams, E. R. Anal. Chem. 2010, 82, 9050−
9057.
(21) Kaltashov, I. A.; Bobst, C. E.; Abzalimov, R. R. Anal. Chem.
2009, 81, 7892−7899.
(22) Rand, K. D.; Bache, N.; Nedertoft, M. M.; Jorgensen, T. J. D.
Anal. Chem. 2011, 83, 8859−8862.
(23) Syka, J. E. P.; Coon, J. J.; Schroeder, M. J.; Shabanowitz, J.;
Hunt, D. F. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 9528−9533.
(24) Zubarev, R. A.; Kelleher, N. L.; McLafferty, F. W. J. Am. Chem.
Soc. 1998, 120, 3265−3266.
(55) Davidovic, M.; Mattea, C.; Qvist, J.; Halle, B. J. Am. Chem. Soc.
2009, 131, 1025−1036.
(56) Ibarra-Molero, B.; Makhatadze, G. I.; Sanchez-Ruiz, J. M.
Biochim. Biophys. Acta 1999, 1429, 384−390.
(57) Bielski, B. H. J.; Freed, S. Biochim. Biophys. Acta 1964, 89, 314
ff.
(58) Dunn, B. M.; Uversky, V. N. Curr. Protein Pept. Sci. 2009, 10,
408−415.
(59) Dempsey, C. E. J. Am. Chem. Soc. 1995, 117, 7526−7534.
(25) Eyles, S. J.; Speir, J. P.; Kruppa, G. H.; Gierasch, L. M.;
Kaltashov, I. A. J. Am. Chem. Soc. 2000, 122, 495−500.
(26) Iavarone, A. T.; Jurchen, J. C.; Williams, E. R. Anal. Chem. 2001,
73, 1455−1460.
(27) Zhang, S.; Van Pelt, C. K.; Henion, J. D. Electrophoresis 2003,
24, 3620−3632.
(28) Van Pelt, C. K.; Zhang, S.; Henion, J. D. J. Biomol. Tech. 2002,
13, 72−84.
(29) Rand, K. D.; Zehl, M.; Jensen, O. N.; Jorgensen, T. J. D. Anal.
Chem. 2009, 81, 5577−5584.
(30) Bai, Y.; Milne, J. S.; Mayne, L.; Englander, S. W. Proteins 1993,
17, 75−86.
(31) Brutscher, B.; Bruschweiler, R.; Ernst, R. R. Biochemistry 1997,
36, 13043−13053.
(32) Pan, Y. Q.; Briggs, M. S. Biochemistry 1992, 31, 11405−11412.
(33) Wilkinson, K. D.; Mayer, A. N. Arch. Biochem. Biophys. 1986,
250, 390−399.
(34) Hoerner, J. K.; Xiao, H.; Kaltashov, I. A. Biochemistry 2005, 44,
11286−11294.
4473
dx.doi.org/10.1021/ac300268r | Anal. Chem. 2012, 84, 4467−4473