Download Analyzing glycerol-mediated protein oligomerization by electrospray

RAPID COMMUNICATIONS IN MASS SPECTROMETRY
Rapid Commun. Mass Spectrom. 2005; 19: 2636–2642
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/rcm.2113
Analyzing glycerol-mediated protein oligomerization
by electrospray ionization mass spectrometry
Maria Anita Mendes1,2, Bibiana Monson de Souza1,2, Lucilene Delazari dos Santos1,2,
Keity Souza Santos2 and Mario Sergio Palma1,2*
1
Department of Biology, CEIS/IBRC, UNESP, Rio Claro, SP, CEP: 13506-900, Brazil
Institute of Immunological Investigations/MCT-CNPq, CEIS/IBRC, UNESP, Rio Claro, SP, CEP: 13506-900, Brazil
2
Received 12 April 2005; Revised 20 July 2005; Accepted 21 July 2005
Glycerol is widely used as protein stabilizer, in both local and commercial preparations, so it has
become necessary to develop methods for mass spectrometric analysis of protein preparations in
the presence of glycerol. However, this stabilizing agent may cause signal suppression when
present in high concentrations, and is also known to induce protein supercharging even at low
concentrations. This work reports the use of electrospray ionization (ESI) mass spectrometry to
characterize glycerol-mediated protein oligomerization. This phenomenon seems to involve the
formation of strong non-covalent interactions between protein and glycerol involving close contact
between the monomers, leading to formation of protein oligomers adducted with glycerol molecules under the characteristic analytical conditions of the ESI interface. At high orders of oligomerization a lower number of glycerol molecules is required to maintain the high oligomeric states
than for the dimers and trimers, and it is possible that for the higher oligomers the monomers
become so close to one another that non-covalent bonds between the side chains of the amino
acid residues in the proteins may be established. Copyright # 2005 John Wiley & Sons, Ltd.
Biomolecular mass spectrometry received an enormous
impetus with the introduction of new soft ionization methods such as matrix-assisted laser desorption/ionization
(MALDI),1,2 and electrospray ionization (ESI).3 –5 The ESI
method has emerged as a powerful technique for producing
intact ions in vacuo from large and complex species in solution, with many applications for biomolecules.6,7 Nowadays,
the ESI technique is routinely used to produce intact gasphase ions of proteins, nucleic acids, and specific non-covalent biomolecular complexes, for analysis by mass spectrometry (MS).8 ESI has the advantage that it can produce
multiply charged ions, especially for large molecules; this
multiple charging phenomenon results in a distribution of
molecular ions with m/z values typically in the range between
500 and 3000, which permits the use of conventional mass
spectrometers for the accurate mass measurement of large
molecules. The mass accuracy for proteins provided by
ESI-MS is generally within 0.01–0.05% of the calculated
masses.9,10 The technique is also used in the structural characterization of smaller biomolecules by tandem mass spectrometry (MS/MS) or multi-stage (MSn) experiments, and thus
is used for peptide sequencing,9–11 for location of sites of
post-translational modifications of proteins and peptides,12
or even to determine gas-phase protein conformations.7
Despite the wide range of different analytical applications
of ESI-MS to proteins, several factors have been shown to
*Correspondence to: M. S. Palma, Department of Biology, CEIS/
IBRC, UNESP, Rio Claro, SP, CEP: 13506-900, Brazil.
E-mail: mspalma@rc.unesp.br
Contract/grant sponsor: São Paulo State Research Foundation
(FAPESP).
interfere with the ESI charge-state distributions, including
acid-base chemistry both in solution and in the gas phase,7
solvent composition, the pH of the solvent, the presence of
non-volatile salts, contaminants, volatile buffers and percentage of organic modifier,13 as well as some instrumental
factors.8
The effects of many salts, detergents and chaotropic agents
are relatively well documented in the deterioration of
performance of ESI-MS protocols.14,15 There are some studies
in the literature characterizing the effect of protein stabilization by glycerol; however, in some circumstances, unexpected analytical behavior is observed in the presence of
glycerol when ESI-MS is used, such as the suppression of
protein signal.9,16 It was also demonstrated that the addition
of glycerol to electrosprayed solutions might cause supercharging of proteins and peptides, as well as glycerol
adduction.8,17,18 At glycerol concentrations higher than 1%,
some dramatic changes were reported in both the average
charge and in the shape of charge-state distributions of
protein spectra. The higher average charge seems to be
related to the broadening of charge-state distributions, which
generally is associated with the unfolding of proteins.19 It was
demonstrated that minor conformational changes occurred
upon glycerol addition to the native proteins, suggesting that
the protein structure in the presence of the additive becomes
slightly compressed compared with its state in water.20
Many commercial recombinant proteins, used as molecular biology tools, and even some academically made
preparations, are maintained in the presence of high glycerol
concentrations after purification to stabilize the biological
activity. Thus, there is a need for methodology for MS
Copyright # 2005 John Wiley & Sons, Ltd.
Glycerol-mediated protein oligomerization by ESI-MS
characterization of proteins in the presence of high concentrations of glycerol. Since ESI-MS is commonly used for many
of these analyses, the effect of this compound on MS
performance must be properly understood. Here we show
that, in addition to the effects described above, glycerol also
may mediate protein oligomerization involving formation of
strong non-covalent interactions between protein and
glycerol, which promote close contacts between the protein
monomers leading to the formation of different orders of
oligomers with adduction of glycerol.
EXPERIMENTAL
Experiments were performed using a Quattro II triple-quadrupole mass spectrometer (Micromass, Altrincham, UK),
equipped with a standard ESI source. During all experiments
the source temperature was maintained at 808C and the needle voltage at 2.8 kV; a drying gas flow (nitrogen) of 200 L/h
and a nebulizer gas flow of 15 L/h were used. The mass spectrometer was calibrated with intact horse heart myoglobin
and its characteristic cone-voltage-induced fragments; the
cone-to-skimmer lens voltage controlling the ion transfer to
the mass analyzer was manually scanned from 30–50 V.
The samples were injected into a flow of transport solvent
using a micro-syringe (250 mL) coupled to a micro-infusion
pump (KD Scientific) at a flow rate of 4 mL/min. The ESI
spectra were acquired in the continuum acquisition mode,
scanning from m/z 200–2500, with a scan time of 7 s. The
mass spectrometer data acquisition and processing system
was equipped with MassLynx, Transform and MaxEnt
software for handling and deconvoluting spectra. The
abundances of the charge states are reported relative to that
of the most abundant charge state in each mass spectrum; the
average charge state was computed as described elsewhere.8
Typical conditions were used to perform the MS/MS
experiments: a capillary voltage of 2.8 kV, a cone voltage of
2637
4
50 V, and collision gas pressure of 4 10 mbar and a
desolvation gas temperature of 808C. In these experiments Q1
was used to select the precursor ion staying in radio-frequency
(rf)-only mode. The ion of interest was mass-selected in Q1 and
structurally characterized by collision-induced dissociation
(CID). It was subjected to about 35 eV collision energy and
4 104 mbar collision gas pressure (argon) in q2. The CID
fragments were analyzed by scanning Q3.
Chicken egg ovalbumin (>95%) and horse heart myoglobin (95–100%) were purchased from Sigma Chemical Co.
(St. Louis, MO, USA), and used as standard proteins in all MS
experiments without further purification. Stock solutions
of these proteins were prepared with analyte concentrations
of 106 M. All the solutions used in the present study
contained 5% (v/v) acetonitrile containing 0.1% (v/v)
formic acid.
RESULTS AND DISCUSSION
The positive-mode ESI mass spectrum of myoglobin in the
absence of glycerol, obtained using a cone voltage of 30 V,
revealed a single (unimodal) envelope of peaks from m/z
678–1541 (Fig. 1a), corresponding to protein molecules containing from 25 to 11 charges (Fig. 1(b)), with the most abundant charge state at 20 (m/z 848.6). The deconvolution of this
spectrum resulted in a molecular mass value of 16 948 Da.
To acquire a well-resolved spectrum of myoglobin in the
presence of 0.5% (v/v) glycerol, it was necessary to increase
the sample cone voltage to 50 V, and, under this condition,
several overlapped envelopes of peaks were observed in the
ESI mass spectrum (Fig. 2(a)). Two major envelopes were
observed, and are represented in centroid form (Figs. 2(b)
and 2(c)) to more easily distinguish them from one another.
The most intense envelope extends over charge states 6 to 29,
with the most abundant charge state 14 (m/z 1211.6); this
distribution corresponds to the monomer of myoglobin
Figure 1. (a) ESI mass spectrum of myoglobin in the absence of glycerol, acquired in the
continuum mode and using a sampling cone voltage of 30 V and (b) centroid mass spectrum
representation of the ESI mass spectrum in (a).
Copyright # 2005 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2005; 19: 2636–2642
2638
M. A. Mendes et al.
Figure 2. (a) ESI mass spectrum of myoglobin in the presence of 0.5% of glycerol (v/v), acquired in the
continuum mode, with the cone voltage adjusted to 50 V; (b) centroid mass spectrum representation of
the ESI mass spectrum of myoglobin shown in (a); (c) centroid mass spectrum representation of the
peaks corresponding to dimers of myoglobin; and (d) deconvoluted continuum mass spectrum showing
the polymer molecules formed by myoglobin in the presence of 0.5% of glycerol (v/v).
(Fig. 2(b)). The second most abundant envelope observed in
this spectrum extends over charge states 24 to 54, with
the most abundant charge state 35 (m/z 981.6), as shown
in Fig. 2(c); deconvolution of this distribution resulted in
accurate determination of the molecular mass of the dimer of
myoglobin that had formed an adduct with five glycerol
molecules (34 327 Da). In addition to these peak envelopes,
Fig. 2 also shows at least three other overlapped peak
envelopes corresponding to higher order myoglobin oligomers adducted with glycerol. The complete deconvolution of
the ESI mass spectrum (Fig. 2(a)) revealed a series of
oligomers of this protein from monomer to tetramer,
as shown in Fig. 2(d), i.e., 16 948 Da (myoglobin), 34 327 Da
(2 myoglobin þ 5 glycerol), 51 141 Da (3 myoglobin þ 3
glycerol), and 67 892 Da (4 myoglobin þ 1 glycerol). A
magnified region from m/z 1050–1350 of the myoglobin
spectrum obtained in continuum mode in the presence of
0.5% (v/v) glycerol is shown in Fig. 3, where it is possible to
Copyright # 2005 John Wiley & Sons, Ltd.
observe four clusters of peaks characteristic of the ions (in
different charge states) from monomers, dimers and trimers
of myoglobin; the ions from monomers range from charge
state 13 (m/z 1303) to 16 (m/z 1059); the ions from the dimers
extend from charge state 26 (m/z 1319) to 32 (m/z 1072); and
ions from the trimers range from charge state 38 (m/z 1345) to
48 (m/z 1065).
Figure 4(a) shows the ESI mass spectrum of ovalbumin
obtained in the absence of glycerol and using a cone voltage of
30 V; this spectrum consists of an envelope of peaks from m/z
602–2800. The deconvolution of this ESI mass spectrum
results in a molecular mass of 44 553 Da; the centroid
representation of this spectrum is shown in Fig. 4(b), and
makes it clear that the peaks in this envelope correspond to an
ionized ovalbumin population containing from 15 to 74
charges, with the most abundant charge state 32 (m/z 1392).
The analysis of ovalbumin in the presence of 0.5% (v/v)
glycerol again required the use of a cone potential of 50 V in
Rapid Commun. Mass Spectrom. 2005; 19: 2636–2642
Glycerol-mediated protein oligomerization by ESI-MS
2639
Figure 3. Expanded region of the continuum ESI-MS spectrum of a solution of 106 M
myoglobin in presence of 0.5% (v/v) glycerol (Fig. 2), obtained by using cone voltage
adjusted to 50 V, showing the ions characteristic of the monomer, dimer and trimer.
order to overcome the ESI signal suppression caused by
glycerol; under this condition the ESI mass spectrum shows a
series of overlapped peak envelopes (Fig. 5(a)), corresponding to different orders of protein oligomerization. Figures 5(b)
and 5(c) show the centroid representations of the two most
abundant peak envelopes. The most intense distribution
ranges from charge state 13 to 27 with the most abundant
charge state 21 (m/z 2122) corresponding to the ovalbumin
monomer (44 553 Da), as shown in Fig. 4(b). The second most
abundant envelope (Fig. 4(c)) covers charge states 25 to 60,
with the most abundant charge state 46 (m/z 1954) corresponding to the ovalbumin dimer adducted with glycerol (2
ovalbumin þ 9 glycerol, 89 909 Da). The ESI mass spectrum
(Fig. 5(a)) also contains other peak envelopes (not assigned in
the figure) that were deconvoluted, resulting in determination of molecular masses of a series of ovalbumin oligomers
adducted with glycerol, as shown in Fig. 5(d): 44 553 Da
(ovalbumin), 89 908 Da (2 ovalbumin þ 9 glycerol), 133 918 (3
ovalbumin þ 3 glycerol), and 178 352 Da (4 ovalbumin þ 1
glycerol).
A detailed comparison between the centroid forms of the
ESI mass spectra of myoglobin, in particular the distributions
corresponding to the non-adducted monomer, in the absence
and presence of glycerol (Figs. 1(b) and 2(b), respectively),
reveals the effect of glycerol in increasing the maximum
charge state from 25 to 29, a shift of the most abundant charge
state from 20 to 14, and a shift of the average charge state from
18.0 to 17.5. These results appear to indicate an overcharging
of myoglobin promoted by the presence of glycerol, as
previously demonstrated by Lavarone et al.8 for the ESI-MS
analysis of cytochrome C in the presence of this compound.
The ions of dimeric protein complexes usually carry less than
twice the amount of charge as the monomeric protein,7 which
complicates the interpretation of the experimental data on the
charge distributions of the envelopes of peaks obtained for
the dimers and other oligomers.
Figure 4. (a) ESI mass spectrum of ovalbumin in the absence of glycerol acquired in the
continuum mode and using a cone sample voltage of 30 V and (b) centroid mass spectrum
representation of the ESI mass spectrum shown in (a).
Copyright # 2005 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2005; 19: 2636–2642
2640
M. A. Mendes et al.
Figure 5. ESI mass spectrum of ovalbumin in the presence of 0.5% (v/v) glycerol acquired in the continuum
mode, with the cone voltage adjusted to 50 V; (b) centroid mass spectrum representation of the ESI mass
spectrum of ovalbumin shown in (a); (c) centroid mass spectrum representation of the peaks corresponding to
dimers of myoglobin; and (d) deconvoluted continuum mass spectrum showing the polymer molecules formed
by ovalbumin in the presence of 0.5% (v/v) glycerol.
When the same comparison is performed for the ESI mass
spectra of ovalbumin, it is observed that glycerol caused a
large decrease in the maximum charge state of ovalbumin
from 74 to 27, a shift in the most abundant charge state from
32 to 20, and a decrease of the average charge state from 44 to
20, apparently indicating that the glycerol partially suppressed the ionization of ovalbumin molecules. Note that the
supercharging effect of glycerol was previously reported
only for small proteins, while large proteins such as
ovalbumin were not previously investigated.8 The downward shift in the charge-state distribution observed for
the ovalbumin monomer in the presence of glycerol may be
due to proton transfer to glycerol, based on the gas-phase
basicities of multiply protonated ions.21
Taking into account the observations of Grandori et al.20 for
the relationship between some changes in the envelope peaks
generated by a protein during ESI-MS analysis and the
Copyright # 2005 John Wiley & Sons, Ltd.
folding/unfolding process undergone by this protein,20 the
decrease in the average charge state from 44 to 20 suggests
that monomeric ovalbumin suffered a strong compression by
interaction with glycerol.
To corroborate the formation of these oligomers, CID
experiments were performed. Using tandem mass spectrometry, by selecting some peaks from the envelope produced
from myoglobin during ESI-MS analysis and submitting
them to CID conditions, a careful search was performed for
observation of the peaks associated with either lower and
higher charge states, varying collision energy at a collision
gas pressure of 4 104 mbar. Figure 6 shows the CID
spectrum of the 21 charge state (m/z 808) from the monomer
myoglobin obtained at 35 V collision potential. Figure 6(a)
represents the precursor ion selected at m/z 808, while Fig. 6(b)
shows the resulting CID spectrum; as the collision energy is
increased, the fragment ions of myoglobin itself fill the
Rapid Commun. Mass Spectrom. 2005; 19: 2636–2642
Glycerol-mediated protein oligomerization by ESI-MS
Figure 6. CID mass spectrum of the precursor ion of m/z
808 (the 21 charge state from the envelope of the monomer
myoglobin): (a) selected precursor ion and (b) spectrum
obtained under 35 eV collision energy.
spectrum, and thus 35 V was the optimal value of collision
potential energy used in all CID experiments. In this case
there is clear evidence that only peaks associated with charge
states lower than the value initially selected were observed in
the CID spectrum; thus, the 21 charge state was dissociated to
form the 20 charge state. A similar CID experiment was
performed by selecting the ion with 40 charge state (precursor
ion at m/z 858) from the envelope of the myoglobin dimer.
Figure 7(a) represents the precursor ion selected at m/z 858,
while Fig. 7(b) shows that this CID experiment results in a
spectrum in which the precursor ion is almost fully
dissociated under 35 V collision potential, forming a symmetrical distribution of peaks corresponding to charge states
from 14 to 27. The observation of a symmetrical charge
distribution around the precursor peak cannot be explained
simply by proton loss due to collisions; these results suggest
2641
that, under CID conditions, the dimer dissociated to the
monomer due to collisions, providing additional evidence for
the formation of protein oligomers in the presence of glycerol.
The mixture of oligomer charge states produces some peaks
at exactly the same mass-to-charge ratio values of the
monomer charge states and also between the monomer
charge states peaks, as can be easily seen in Fig. 3.
Generally, glycerol is added to protein solutions due to its
stabilizing effect on protein molecules. However, high levels
of glycerol influence strongly the mass spectrometric
performance due to the ESI suppression effect caused by
glycerol, which requires the use of skimmer voltages higher
than those typically used in routine analysis of proteins with
an ESI interface, to disrupt the strong protein-glycerol
interactions.9
Generally, in the absence of glycerol, the basic and/or
acidic side chains of the amino acid residues are ionized in
water, depending on the pH of the solution. However,
glycerol displaces the water molecules from the protein
surface and establishes strong interactions between itself and
the side-chain groups of the amino acid residues.16,17 These
interactions may be due to hydrogen bonds between glycerol
molecules and the amino/imino groups in the basic amino
acid residues; the carboxyl hydrogens from the acidic amino
acid residues may also be involved in such interactions. The
interactions with glycerol may be sufficiently strong that
glycerol-adducted protein monomers may interact with one
another to form oligomers maintained by non-covalent
interactions under analytical conditions characteristic of the
ESI interface.
CONCLUSIONS
Proteins may interact in vivo, or even under mass spectrometric analysis conditions, with alike or different proteins
forming homo/heterooligomers. The investigation of protein
Figure 7. CID mass spectrum of the precursor ion of m/z 858 (the 40 charge state from the
envelope of the dimer myoglobin): (a) selected precursor ion and (b) spectrum obtained
under 35 eV collision energy.
Copyright # 2005 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2005; 19: 2636–2642
2642
M. A. Mendes et al.
oligomerization is of interest to improve our understanding
of protein stability and of the types of interaction involved
in this process. Glycerol is widely used as a protein stabilizer,
both in academic and commercial protein preparations.
However, unusual effects caused by this compound are
known to occur in ESI-MS analysis of proteins. For example,
it may cause signal suppression when it is present in high
concentrations, requiring the use of high voltage at the
sampling (skimmer) cone to overcome this problem;9,16 it is
also known to induce supercharging in some proteins and a
partial suppression of degree of ionization for other proteins,
even at low concentrations.
In addition to these effects, the present manuscript reports
the occurrence of glycerol-mediated protein oligomerization,
which must be considered during ESI-MS analysis of proteins
in the presence of this stabilizing agent, even when very low
protein concentrations are used (about 106 M). Regardless of
the mechanism(s), this phenomenon seems to involve
formation of strong non-covalent interactions between
protein and glycerol; thus, due to its high polarity and
capacity to form relatively strong hydrogen bonds, a small
number of glycerol molecules could form non-covalent
interactions with several protein monomers, leading to
formation of oligomers. When the oligomers reach higher
orders they possibly become so close to one another that noncovalent bonds between the side chains of the amino acid
residues from the protein molecules may be established,
without participation of glycerol. The decreasing numbers of
adducted glycerol molecules per protein molecule corroborates this as the higher oligomers are formed.
Acknowledgements
This work was supported by a grant from the São Paulo State
Research Foundation (FAPESP). Maria Anita Mendes is a
postdoctoral fellow from FAPESP (Proc. 01/05060-4); Bibiana
Copyright # 2005 John Wiley & Sons, Ltd.
Monson de Souza is a doctoral student fellow from FAPESP
(Proc. 03/00985-5); and Mario Sergio Palma is a researcher for
the Brazilian Council for Scientific and Technological Development (CNPq, 300377/2003-5).
REFERENCES
1. Tanaka K, Waki H, Ido Y, Akita S, Yoshida Y, Yoshida T.
Rapid Commun. Mass Spectrom. 1988; 2: 151.
2. Hillenkamp F, Karas M. Methods Enzymol. 1990; 193: 280.
3. Fenn JB, Mann M, Meng CK, Wong SF, Whitehouse CM.
Science 1989; 246: 64.
4. Chowdhury SK, Katta V, Chait BT. Biochem. Biophys. Res.
Commun. 1990; 167: 686.
5. Mann M, Wilm M. Trends Biochem. Sci. 1995; 20: 219.
6. Chowdhury SK, Katta V, Chait BT. Rapid Commun. Mass
Spectrom. 1990, 4: 81.
7. Heck AJR, Van den Huvel HH. Mass Spectrom. Rev. 2004; 23:
368.
8. Lavarone AT, Jurchen JC, Willians ER. Anal. Chem. 2001; 73:
1455.
9. Mendes MA, de Souza BM, Marques MR, Palma MS.
Spectroscopy 2004; 18: 339.
10. de Souza BM, Marques MR, Tomazela DM, Eberlin MN,
Mendes MA, Palma MS. Rapid. Commun. Mass Spectrom.
2004; 18: 1095.
11. Yates JR. J. Mass Spectrom. 1998; 33: 1.
12. Kelleher NL, Hendrickson CL, Walsh CT. Biochemistry 1999;
38: 15623.
13. Thomson BA. J. Am. Soc. Mass Spectrom. 1997; 8: 1053.
14. Canarelli S, Fisch I, Freitag R. J. Chromatogr. A 2002; 948: 139.
15. Liu CL, Hofstadler SA, Bresson JA, Udseth HR, Tsukuda T,
Smith RD, Snyder AP. Anal. Chem. 1998; 70: 1797.
16. Mendes MA, Chies JM, Dias ACO, Astofi S, Palma MS.
Rapid Commun. Mass Spectrom. 2003; 17: 672.
17. Lavarone AT, Willians ER. Int. J. Mass Spectrom. 2002; 219:
63.
18. Lavarone AT, Willians ER. J. Am. Chem. Soc. 2003; 125: 2319.
19. Takáts A, Wiseman JM, Gologan B, Cooks RG. Anal. Chem.
2004; 76: 4050.
20. Grandori R, Matecko I, Mayr P, Muller N. J. Mass Spectrom.
2001; 36: 918.
21. Schnier PD, Gross DS, Willians ER. J. Am. Soc. Mass
Spectrom. 1995; 6: 1086.
Rapid Commun. Mass Spectrom. 2005; 19: 2636–2642