Electrospray Ionization Fourier Transform Ion
Cyclotron Resonance Mass Spectrometric
Analysis of the Recombinant Human
Macrophage Colony Stimulating Factor ?
Claudia S. Maiera, Xuguang Yana, Mark E. Harderb,
Michael I. Schimerlikb, and Max L. Deinzera
Departments ofaChemistry andbBiochemistry & Biophysics, Oregon State University, Corvallis, Oregon,
Ljiljana Pas ˇa-Tolic ´ and Richard D. Smith
Environmental Molecular Sciences Laboratory, Pacific Northwest National Laboratory, Richland,
The potential of electrospray ionization (ESI) Fourier transform ion cyclotron mass spectrom-
etry (FTICR-MS) to assist in the structural characterization of monomeric and dimeric
derivatives of the macrophage colony stimulating factor ? (rhM-CSF?) was assessed. Mass
spectrometric analysis of the 49 kDa protein required the use of sustained off-resonance
irradiation (SORI) in-trap cleanup to reduce adduction. High resolution mass spectra were
acquired for a fully reduced and a fully S-cyanylated monomeric derivative (?25 kDa). Mass
accuracy for monomeric derivatives was better than 5 ppm, after applying a new calibration
method (i.e., DeCAL) which eliminates space charge effects upon high accuracy mass
measurements. This high mass accuracy allowed the direct determination of the exact number
of incorporated cyanyl groups. Collisionally induced dissociation using SORI yielded b- and
y-fragment ions within the N- and C-terminal regions for the monomeric derivatives, but
obtaining information on other regions required proteolytic digestion, or potentially the use of
alternative dissociation methods. (J Am Soc Mass Spectrom 2000, 11, 237–243)
American Society for Mass Spectrometry
bonds [1–4]. Mass spectrometric approaches provide a
fast and sensitive method to monitor thiol modification
reactions because the number of modified thiol groups
can be directly deduced from the observed mass in-
crease. However, if large proteins are studied, the
resolving power of conventional mass spectrometers
using electrospray ionization (ESI) is often insufficient
to resolve the mass differences caused by the incorpo-
ration of small thiol blocking groups, i.e., cyanyl
groups. ESI  coupled to Fourier transform ion cyclo-
tron mass spectrometry (FTICR-MS) [6, 7] often pro-
vides the resolving power, mass accuracy, and sensitiv-
ity necessary to detect major and minor constituents in
complex protein mixtures. Furthermore, its nondestruc-
n vitro folding studies involving disulfide-contain-
ing proteins rely on the use of thiol trapping reac-
tions to study the progressive formation of disulfide
tive detection method can enable the use of multistage
tandem mass spectrometric experiments to obtain se-
quence information . The most frequently used tan-
dem mass spectrometric techniques in combination
with FTICR-MS to induce fragmentation of large pro-
teins are sustained off-resonance irradiation collision-
ally induced dissociation (SORI-CID)  and infrared
multiphoton dissociation (IRMPD) . Both methods
access the lowest dissociation pathways, and are the
methods of choice for optimum sensitivity and protein
identification in combination with existing databases
(i.e., where only limited sequence related information is
Macrophage colony stimulating factor (M-CSF) is a
disulfide-linked dimeric glycoprotein with a molecular
weight ranging from 45 to 100 kDa depending on the
glycosylation pattern. M-CSF acts as a hematopoietic
growth factor which regulates proliferation, differenti-
ation, and survival of monocytes, macrophages, and
their early bone-marrow progenitor lineage . M-CSF
can be recombinantly expressed in mammalian cells [12,
Address reprint requests to Professor Max L. Deinzer, Department of
Chemistry,Oregon State University,
Corvallis,OR 97331. E-mail:
© 2000 American Society for Mass Spectrometry. Published by Elsevier Science Inc.
Received July 21, 1999
Revised November 8, 1999
Accepted November 8, 1999
13] and Escherichia coli . In the latter expression
system it has been expressed from several clones that
provide recombinant proteins that differ in the number
of amino acid residues. Recombinantly expressed M-
CSF with residues 4-221 of the human protein is termed
rhM-CSF? . In vitro refolding of the inactive protein
isolated from inclusion bodies yields biologically active
protein in high yields . This was not completely
expected because the mature protein possesses nine
disulfide bridges; three intersubunit disulfides (Cys
31-Cys? 31, Cys 157/159-Cys? 157/159) and in each
subunit three intramolecular disulfide bonds (Cys 48-
139, 7-90, 102-146) [15, 16]. We are currently exploring
the possibility of using cyanylation to block sulfhydryl
groups during oxidative in vitro refolding of rhM-
CSF?. This report evaluates the potentials and pitfalls of
using SORI approaches with FTICR-MS/MS to gain
structural information for rhM-CSF? and two deriva-
tives, the fully reduced and the S-cyanylated monomer.
Purified biologically active rhM-CSF? expressed in E.
coli and formulated in 20 mM citrate containing 0.5%
sodium chloride and 1% mannitol was obtained as a
lyophilized powder from Dr. C. Cowgill (Chiron Corp.
Emeryville, CA). The freshly dissolved protein was
extensively dialysed against 1 mM ammonium acetate
or 5% acetic acid prior to mass spectrometric analysis.
Fully reduced monomeric M-CSF? was prepared by
incubating rhM-CSF? in 8 M urea (Fluka) or 6 M
guanidinium chloride (Calbiochem) prepared with 100
mM tris/5 mM EDTA buffer (pH 8.5) containing 10 mM
dithiothreitol (Sigma Chemicals). Reaction periods var-
ied from 5 to 8 h at room temperature. The cyanylation
protocol used was adapted from Wu and Watson .
Fully reduced monomeric M-CSF? was transferred into
0.1 M tris buffer (pH 3.5) containing 6 M guanidinium
chloride (GuHCl) by use of a Bio-Rad 1 mL-spin column
packed with Bio-Gel P-6. Cyanylation of free sulfhydryl
groups was initiated by adding an approximately 25
molar excess of 1-cyano-4-dimethylaminopyridinium
fluoroborate (CDAP, Sigma Chemicals) over the total
cysteine content. After a short reaction period (10–15
min), cyanylation was terminated by a second buffer
exchange into 0.1 M citrate buffer containing 6 M
GuHCl utilizing the above described spin column meth-
odology and stored at 5°C. Reduced as well as cyany-
lated monomeric rhM-CSF? was purified by reversed-
phase high performance
(HPLC) employing a binary linear gradient elution with
0.05% trifluoroacetic acid as solvent A and acetonitrile
containing 0.05% trifluoroacetic acid as solvent B. A
capillary column (0.3 mm ? 100 mm) self-packed with
Vydac C18bulk material (10 ?m particle size, 300 Å pore
size) was utilized. An ABI model 140B solvent delivery
system equipped with a precolumn splitter was used.
The flow rate was adjusted so that a solvent flow of
5–10 ?L/min through the capillary column was
achieved. The UV trace at 217 nm was monitored with
an ABI Spectroflow 783 absorbance detector equipped
with a microflow cell. Protein fractions were collected
ESI-FT ICR-MS and MS2
The 7 tesla FTICR-MS equipped with a custom built ESI
source that incorporates an rf quadrupole for collisional
focusing of ions is described elsewhere . Sample
delivery was performed with a Harvard Apparatus
syringe pump (South Natick, MA). The flow rate was
set to 0.5 ?L/min. Mass spectra were obtained utilizing
standard experimental sequences that employ in-trap
ion accumulation (i.e., ion injection, pump-down, and
excitation/detection). Background pressure in the ICR
trap was maintained at ?10?9torr by a custom
cryopumping assembly that provides pumping speeds
in excess of 105L/s, which allows rapid transition
between in-trap ion accumulation (i.e., 10?5, torr) and
high performance ion excitation/detection (i.e., 10?9
torr) events. A piezoelectric pulse valve (Lasertech-
niques, Albuquerque, NM) was used to inject N2(to
?10?5torr) for accumulated trapping of ions and for
ion activation in in-trap cleanup and CID experiments.
Gentle in-trap removal of unspecific noncovalent ad-
ducts from the protein was achieved using broadband
cleanup with low energy chirp (i.e., 10 to 110 kHz at 10
?s?1sweep rate, ?100 Vpp) as previously described
In MS2experiments, stored-waveform inverse Fou-
rier transform (SWIFT)  radial ejection was used to
remove ions of all but selected m/z. The m/z-selected
parent ions are then translationally excited to dissociate
by means of collisional activation employing low am-
plitude (typically ?25 Vpp) SORI ?1.5 kHz of resonance
from the precursor ion for 100 ms. MS2experiments
were run in signal averaging mode: ion injection, isola-
tion, dissociation, and excitation/detection events were
repeated 30 times and the resulting spectra were accu-
mulated before Fourier transformation of the transient
(acquired with 256K data points at 290 kHz), followed
by magnitude calculation and frequency-to-m/z conver-
An Odyssey data station (Finnigan, Madison, WI)
provided ICR trap control, data acquisition, and stor-
age. Typically, coherent ICR motion was excited by
dipolar frequency sweep excitation (?125 Vppampli-
tude, excitation over a 125 kHz bandwidth with a 35
Hz/?s sweep rate).
Data analysis was performed using the ICR2LS soft-
ware package developed at PNNL . A new method
of calibration was used to remove the effects of space
charge on high accuracy mass measurements without
the requirement for internal calibrants . The pro-
gram IsoPro 3.0 (http://members.aol.com/msmssoft) was
utilized to aid in MS2data interpretation.
MAIER ET AL.J Am Soc Mass Spectrom 2000, 11, 237–243
rhM-CSF? is a 49 kDa homodimer. Each monomeric
subunits consists of 218 amino acid residues including
nine cysteines of which six form intramolecular disul-
fides and the remaining three link the two subunits via
disulfide bridges . Initial ESI mass spectrometric
experiments were performed with rhM-CSF? dialyzed
against 5% acetic acid. The broadband spectrum
showed heavily adducted ions making the assignment
of charge states impossible (Figure 1A). Recently, sev-
eral FTICR-MS specific techniques were reported which
allow gentle removal of noncovalently adducted salt
ions from proteins and DNAs directly in the ICR trap:
infrared multiphoton dissociation (IRMPD)  and
low energy SORI-CID . In-trap SORI cleanup was
successful in removing nonspecific adducts and yielded
a broadband low resolution mass spectrum that al-
lowed assignment of charge states. Attempts to remove
additional adduction using increased SORI activation of
the molecular ions by increasing the amplitude induced
dissociation of the covalent bonds. Hence, rhM-CSF?
was further purified by reversed-phase HPLC prior to
ESI-FTICR mass spectrometric analysis. The acquired
broadband mass spectrum showed less adducted mo-
lecular ions (Figure 1B) and gentle SORI-CID in-trap
cleanup provided a high quality (but not isotopically
resolved) mass spectrum (Figure 1C) that showed a
charge state distribution encompassing the molecular
ions (M ? 15H)15?to (M ? 20H)20?. The theoretical
average molecular mass of rhM-CSF? is 49,029.7 u (i.e.,
based on the empirical formula C2146H3342N572O686S28).
The experimental mass obtained from the charge state
distribution using DeCAL  was 49,029.1 u corre-
sponding to a difference of only 12.2 ppm from that
predicted for this cytokine.
Analysis of the reduced rhM-CSF? monomer by
ESI-FTICR-MS yielded a high resolution mass spectrum
showing a charge state distribution consisting of isoto-
pically resolved 13- to 22-fold protonated molecular
ions after minimal SORI in-trap cleanup (Figure 2A).
For example, for the ion peak cluster at m/z 1291.85 a
resolving power of approximately 64,000 was achieved
(inset in Figure 2A). The “zero-charge” spectrum (Fig-
ure 2B) generated from the charge states 18?to 23?
yielded an experimental mass of 24,523.04 u (Mris
given as the mass of the highest abundant isotope)
which compared well with the theoretical mass of
reduced rhM-CSF? monomer of 24,522.93 u (4.4 ppm
error). A minor constituent with an experimental mass
of 24,565.83 u was tentatively assigned as monocar-
bamylated rhM-CSF? monomer (deviation 4.0 ppm
from the theoretical mass of 24,565.93 u). Carbamyla-
tion causes a nominal mass increase of 43 u and
probably occurred during reductive denaturation in the
presence of urea. Variants with methionine sulfoxide or
possible misincorporation of norleucine for methionine
were not detected. (The latter would result in a nominal
mass decrease of ?18 u in the variant protein.) Incor-
poration of norleucine instead of methionine residues
was recently reported for a truncated form of M-CSF
. Methionine sulfoxide formation as a common
aging-related degradation reaction in proteins would
cause a nominal mass increase of ?16 u.
One long term goal of our present research efforts is
the characterization of in vitro folding pathway(s) of
rhM-CSF?. An important prerequisite to investigate if
Broadband FTICR mass spectrum (m/z 1000–4000) after electros-
praying from 5% acetic acid showing heavily adducted ions. (B)
Gentle in-trap cleanup with low energy chirp (i.e., 10 to 110 kHz
at 10 ?s?1sweep rate, ?100 Vpp) in the presence of ?10?5torr of
N2resulted in a low resolution mass spectrum showing molecular
ions (M ? 18H)18?to (M ? 21H)21. (C) Low resolution broadband
mass spectrum of HPLC-purified rhM-CSF? after gentle in-trap
cleanup [conditions as in (B)]. The experimental mass obtained for
rhM-CSF? was 49,029.1 u. The protein (?0.3–0.5 mg/mL) was
dissolved in approximately 80% acetonitrile containing 0.1% trif-
Positive ion ESI-FTICR mass spectra of rhM-CSF?. (A)
J Am Soc Mass Spectrom 2000, 11, 237–243 FT ICR-MS OF RHM-CSF? AND DERIVATIVES
the reassociation step of the monomers during renatur-
ation depends on the formation of disulfide linkages is
the availability of a monomer that cannot form any
disulfide bonds. Therefore, rhM-CSF? monomer with
fully S-protected sulfhydryl group was prepared by
cyanylation with CDAP under acidic condition .
First attempts to achieve complete cyanylation in the
reduced rhM-CSF? monomer with CDAP (5–35 molar
excess over the total cysteine content) in citrate buffer
(pH 3) were not successful even after prolonged reac-
tion periods (1 h). The reaction was eventually carried
out in 0.1 M tris/6 M GuHCl buffer adjusted to pH 3.5
in the presence of a 25 molar excess of CDAP relative to
the total cysteine content. The HPLC-purified protein
was subjected to ESI-FTICR mass spectrometric analy-
sis. Under the conditions used, the broadband ESI mass
spectrum showed a charge state distribution ranging
from 11?to 22?with the 16-fold protonated molecular
ion predominating. The mass resolution of ?55,000
allowed direct determination of the distribution of
cyanyl groups. The mass for the dominant species was
24,747.85 u indicating the incorporation of nine cyanyl
groups (1.3 ppm error using DeCAL ).
cyanylated rhM-CSF? monomer. (A) Broadband high resolution
mass spectrum of HPLC-purified cyanylated rhM-CSF? mono-
mer. The protein (?0.3–0.5 mg/mL) was electrosprayed from
approximately 80% acetonitrile containing 0.05% trifluoroacetic
acid. (B) SWIFT-isolated (M ? 18H)18?ions. (C) SORI-CID spec-
trum of the 18?-charge state. The inset shows the resolved isotopic
distribution of the doubly charged fragment ion b21(2?) and its
adjacent isotopic envelope that probably arose from neutral loss of
Series of spectra obtained from SORI-CID of the
reduced rhM-CSF? monomer. (A) Broadband high resolution
FTICR mass spectrum (m/z 950–4500, four transients were accu-
mulated before Fourier transformation) of rhM-CSF? monomer
electrosprayed from a solution of acetonitrile/0.05% trifluoracetic
acid (4/1) containing 5% acetic acid. Protein concentration was
approximately 0.3–0.5 mg/mL. The inset shows the enlarged
region of the (M ? 19H)19?ion demonstrating a resolving power
of approximately 64,000 at m/z ?1291.6. (B) Comparison of the
reconstructed “zero-charge” mass spectrum generated from
charge states 18?through 23?(upper trace) with the calculated
isotopic distributions for the reduced rhM-CSF? monomer
(C1073H1680N286O343S14) and the monocarbamylated rhM-CSF?
monomer (C1074H1681N287O344S14) (lower trace).
Positive ion ESI-FTICR mass spectra of HPLC-purified
MAIER ET AL. J Am Soc Mass Spectrom 2000, 11, 237–243
Additionally, the utility of SORI-CID for obtaining
sequence data for the monomeric rhM-CSF? derivatives
(25 kDa) was evaluated. The 18?-charge state of the
reduced and the cyanylated rhM-CSF? monomer was
isolated by SWIFT  and subjected to SORI-CID 
using identical experimental conditions. ESI-FTICR-
MS/MS spectra obtained for cyanylated rhM-CSF?
monomer are presented in Figure 3. The SORI-CID
spectra of the reduced and the cyanylated rhM-CSF?
monomer showed a large number of fragment ions
(?75). High resolution obtained for fragment peaks
allowed unambiguous peak assignment (?10 ppm av-
erage error). This information was then used in a
computer program that matched the observed values
with a list of possible fragment ions (a, b, c and x, y, z)
calculated from the protein sequence. In spite of the
high resolving power, uncertainty in the assignment of
some fragment ions remained and, therefore, a probable
assignment was made only if other ions of the same
fragment ion series were observed (Table 1). In the case
of overlapping fragment ion peak clusters the isotopic
distributions were simulated (Figure 4). Fragment ions
that resulted from neutral loss of small molecules (i.e.,
H2O and/or NH3) were also observed. The loss of small
neutral molecules is a common low energy fragmenta-
tion process in SORI-CID . Most of the fragment
assignments (to ?10 ppm) belonged to b- and y-frag-
ment ion type and corresponded to cleavages of the N-
and C-terminal regions of the protein (Figure 5). Other
fragment ions detected (but not assigned in this work)
likely originated from internal cleavages. The SORI-CID
spectrum of the 18?-charge state of cyanylated rhM-
CSF? monomer showed many of the same fragment
species as for the reduced rhM-CSF? monomer. As
expected, none of the observed y-fragment species were
shifted in mass because of the presence of cyanyl
group(s). On the other hand, the triply charged frag-
ment ions b21and b23to b26in the SORI-CID spectrum
of cyanylated rhM-CSF? monomer were shifted to
higher masses. The mass difference of ?25.0 u sug-
gested the incorporation of one cyanyl group near the
N-terminus, most likely at the cysteine residue in posi-
tion 7. Additionally, a weak doubly charged ion 29.49 u
lower than the highly abundant b21(2?) ion at m/z
1191.545 was present which implied neutral loss of
HSCN (Figure 3C). Furthermore, S-cyanylation of the
second cysteine residue at position 31 was marked by
the m/z shift of fragment ion b30(4?) to m/z 854.605
Assigned product ions observed in SORI-CID spectra of the 18?-charge state of the reduced M-CSF? monomer
aCharge states of the isotopic distribution.
bObserved mass-to-charge ratio of the most abundant isotope after applying DeCAL .
cExperimental monoisotopic mass.
dExperimental mass of the most abundant isotope.
eTheoretical monoisotopic mass.
fError (in ppm) based on the mass difference observed between the measured most abundant and simulated most abundant isotope.
gFragment ion assignment according to Biemann .
hTheoretical mass of the most abundant isotope.
J Am Soc Mass Spectrom 2000, 11, 237–243FT ICR-MS OF RHM-CSF? AND DERIVATIVES
The purpose of this study was to explore the potential
of ESI and low energy SORI-CID with FTICR to struc-
turally characterize rhM-CSF?, a 49 kDa protein, and its
monomeric derivatives (?25 kDa). Of particular interest
was a cyanylated derivative that is used as the simplest
protein folding intermediate in presently ongoing in
vitro protein folding studies. One question was whether
a “top down” approach , i.e., direct gas phase
sequencing of the whole protein to produce a comple-
mentary set of fragments that eventually add up to the
whole protein, is a realistic alternative to “bottom up”
characterization which involves extensive proteolysis of
the protein and identification of the smaller peptides
generated (peptide mapping) [27, 28].
A problematic issue in this work was the tendency of
rhM-CSF? to form nonspecific noncovalent adduction
which prevent acquiring isotopically resolved mass
spectrometric data using 7 tesla instrumentation for the
intact protein even after in-trap SORI cleanup. The
observed noncovalent adduct formation was in part
attributed to the strong ion pairing properties of triflu-
oroacetic acid which was used as modifier in the HPLC
solvents. Thus, future efforts will have to include the
replacement of trifluoroacetic acid with an ion pairing
reagent more suitable for ESI such as formic or acetic
acid. Despite these issues, it should be noted that highly
accurate mass measurements could still be obtained
using SORI in-trap sample cleanup and a new calibra-
tion method (i.e., DeCAL ) which eliminates space
charge effects on high accuracy mass measurements.
High quality isotopically resolved mass spectra were
obtained for both monomeric derivatives of rhM-CSF?,
the reduced and the cyanylated one. Noteworthy, loss
of HSCN during gentle SORI in-trap cleanup was not
observed. The high mass accuracy achieved for the
monomeric derivatives (?5 ppm error) allowed the
precise determination of the number of cyanyl groups
incorporated. CDAP proved to be a trapping reagent that
provides high thiol selectivity and trapping efficiency.
Derivatization of monomeric rhM-CSF? was successful
after optimization of the reaction conditions. ESI-FTICR
results confirmed complete S-cyanylation of all nine
cysteine residues of the reduced rhM-CSF? monomer.
Not surprisingly, SORI-CID of the intact monomeric
derivatives as an alternative approach to obtain se-
quence-specific information was only partly successful.
Sequence data obtained were largely limited to rela-
tively short partial sequences near N- and C-termini.
The high resolving power of FTICR-MS2allowed assign-
ment of fragment ions with more confidence and was
beneficial in the assignment of fragment ions with similar
m/z values (Figure 4). However, like IRMPD, SORI-CID is
most useful in cases where only sparse fragmentation is
required along with greater sensitivity. Higher energy
CID methods can provide greater fragmentation but at the
cost of significant reduction in sensitivity (i.e., greater
sample consumption). Another alternative that exploits
the power of FTICR is the electron capture method re-
cently developed by McLafferty and co-workers .
However, this again requires greater sample consumption
and poses much greater challenges for data analysis.
In this work charge-remote fragmentation gave rise
to a series of y-fragment ions in the C-terminal proline-
rich region of the monomeric rhM-CSF? derivatives. It
is noteworthy that many of the more abundant disso-
SORI-CID spectrum of reduced (A) and cyanylated (B) rhM-CSF?
monomer. The high resolution capability of FT-ICR allowed the
assignment of overlapping isotopically resolved fragment ion
peak clusters. The insets show the simulated isotope pattern of the
observed fragment ions in the corresponding region. Successful
cyanylation of cysteine residues in position 7 and 31 is reflected in
the observed m/z shift of the fragment ion b30(4?).
Comparison of the 838 ? m/z ? 856 region of the
Residues are numbered according to the human sequence. Cys-
teine residues are shaded. The four ?-helical regions (double line)
are marked. Vertical lines mark the fragmentation sites that were
observed in SORI CID experiments in this study.
Amino acid sequence of the rhM-CSF? monomer.
MAIER ET AL. J Am Soc Mass Spectrom 2000, 11, 237–243
ciations occurred at amide bonds adjacent to proline Download full-text
residues which is in agreement with the concept that
the 5-membered ring of the proline side chain con-
strains the N–C?bond rotation hindering intramolecu-
lar energy transfer . Yu et al.  observed the facile
gas-phase cleavage of Asp–Xxx peptide bonds and
proposed a mechanism that is based on cyclization of
the aspartic acid side chain upon cleavage of the Asp–Xxx
bond. Formation of the abundant fragment ion b21(2?) is
the basicity of the nitrogen in the pyrrolidine ring of
proline residues greatly facilitates cleavage of Asp–Pro
bonds. This might be reflected in the dominance of frag-
ment ion y8(1?) and the tentatively assigned singly
charged internal fragment ion (b210y18)10at m/z 1074.52.
Interestingly, S-cyanyl groups were apparently suf-
ficiently stable in SORI-CID MS2experiments and neu-
tral loss of HSCN represented only a minor decompo-
sition pathway. This is somewhat surprising because
loss of small molecules (H2O and/or NH3) is frequently
reported as low energy fragmentation reactions in
SORI-CID experiments [10, 25]. Notably, in MS2exper-
iments employing an ion trap mass spectrometer loss of
HSCN through gas-phase ?-elimination from S-cyano-
cysteine containing peptides has been reported as a
major fragmentation pattern .
In the present work, SORI-CID provided no assign-
able fragment ions from regions that form (in the
mature protein) the rigid four-helical-bundle fold (Fig-
ure 5). The strong hydrophobic interactions and the
extended hydrogen bonding network between these
residues may make peptide bonds less susceptible to
cleavage . Therefore, in the case of rhM-CSF? and its
derivatives, traditional peptide mapping using pepsin
or trypsin for proteolysis and identification of gener-
ated peptides by MS2(“bottom up” approach) proved
to be more successful in identifying sites of modifica-
tion. However, it is likely that much greater coverage of
these regions could be obtained using harsher energy
CID or electron capture dissociation  with FTICR.
The modification of solution conditions prior to analy-
sis may also serve to improve the dissociation efficiency
for protein regions having strong noncovalent associa-
tions (i.e., the four-helical bundle). Thus, we conclude
that the appropriate methodology depends upon the
details of the questions being addressed. Pragmatic issues,
most importantly available sample size, often dictate the
best approaches. Clearly, the combination of available and
developing dissociation methods combined with accu-
racy, resolution, and sensitivity of FTICR is enabling a
new approach to protein characterization.
LPT and RDS thank the Office of Biological and Environmental
Research for support of the work at Pacific Northwest National
Laboratory (PNNL) operated by Battele Memorial Institute for the
U.S. Department of Energy under Contract DC-AC06-76RLO 1830.
We thank Dr. Cynthia Cowgill of Chiron Inc. for rhM-CSF? and
Gordon A. Anderson of PNNL for help with data analyses. This
work was supported in part by NIEHS grant no. 00040 to MLD
1. Weissman, J. S.; Kim, P. S. Science 1991, 253, 1386–1393.
2. Jensen, P. K.; Harrata, A. K.; Lee, C. S. Anal. Chem. 1998, 70,
3. Rothwarf, D. M.; Scheraga, H. A. Biochemistry 1993, 32, 2671–
4. Wu, J.; Yang, Y.; Watson, J. T. Protein Sci. 1998, 7, 1017–28.
5. Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse,
C. M. Mass Spectrom. Rev. 1990, 9, 37–70.
6. Marshall, A. G.; Grosshans, P. B. Anal. Chem. 1991, 63, 215A–
7. Kelleher, N. L.; Senko, M. W.; Siegel, M. W.; McLafferty, F. W.
J. Am. Soc. Mass Spectrom. 1997, 8, 380–383.
8. McLafferty, F. W. Acc. Chem. Res. 1994, 27, 379–386.
9. Gauthier, J. W.; Trautman, T. R.; Jacobson, D. B. Analyt. Chim.
Acta 1991, 246, 211–225.
10. Little, D. P.; Speir, J. P.; Senko, M. W.; O’Connor, P. B.;
McLafferty, F. W. Anal. Chem. 1994, 66, 2809–2815.
11. Stanley, E. R. In The Cytokine Handbook; Thomson, A., Ed.;
Academic: London, 1994; pp 387–418.
12. Weaver, J.; McCormick, F.; Manos, M. M. Bio/Technol. 1987, 6,
13. Halenbeck, R.; Shadle, P. J.; Lee, P.-J.; Lee, M.-T.; Koth, K.
J. Biotech. 1988, 8, 45–58.
14. Halenbeck, R.; Kawasaki, E.; Wrin, J.; Koths, K. Bio/Technol.
1989, 7, 710–715.
15. Pandit, J.; Bohm, A.; Jancarik, J.; Halenbeck, R.; Koths, K.; Kim,
S.-H. Science 1992, 258, 1358–1362.
16. Glocker, M. O.; Arbogast, B.; Schreurs, J.; Deinzer, M. L.
Biochemistry 1993, 32, 482–488.
17. Wu, J.; Watson, J. T. Protein Sci. 1997, 6, 391–398.
18. Winger, B. E.; Hofstadler, S. A.; Bruce, J. E.; Udseth, H. R.;
Smith, R. D. J. Am. Soc. Mass Spectrom. 1993, 4, 566–577.
19. Pasa-Tolic, L.; Bruce, J. E.; Lei, Q. P.; Anderson, G. A. Anal.
Chem. 1998, 70, 405–408.
20. Marshall, A. G.; Wang, T.; Ricca, T. L. J. Am. Chem. Soc. 1985,
21. Anderson, G. A.; Bruce, J. E. Pacific Northwest National
Laboratory, Richland, WA, 1995.
22. Bruce, J. E.; Anderson, G. A.; Brands, M. D.; Pasa-Tolic, L.;
Smith, R. D. J. Am. Soc. Mass Spectrom., in press.
23. Randhawa, Z. I.; Witkowska, H. E.; Cone, J.; Wilkens, J. A.;
Hughes, P.; Yamanishi, K.; Yasuda, S.; Masui, Y.; Arthur, P.;
Kletke, C.; Bitsch, F.; Shackleton, C. Biochemistry 1994, 33,
24. Wu, J.; Watson, J. T. Anal. Biochem. 1998, 258, 268–276.
25. Wu, Q.; Van Orden, S.; Cheng, X.; Bakhtiar, R.; Smith, R. D.
Anal. Chem. 1995, 67, 2498–2509.
26. Kelleher, N. L.; Lin, H. L.; Valaskovic, G. A.; Aaserud, D. J.;
Fridriksson, E. K.; McLafferty, F. W. J. Am. Chem. Soc. 1999,
27. Shevchenko, A.; Wilm, M.; Vorm, O.; Mann, M. Anal. Chem.
1996, 68, 850–858.
28. Maier, C. S.; Arbogast, B.; Hahn, U.; Deinzer, M. L.; Kuchu-
mov, A. R.; Vinogradov, S. N.; Walz, D. A. J. Am. Soc. Mass
Spectrom. 1997, 8, 352–364.
29. Zubarev, R. A.; Kruger, N. A.; Fridriksson, E. K.; Lewis, M. A.;
Horn, D. M.; Carpenter, B. K.; McLafferty, F. W. J. Am. Chem.
Soc. 1999, 121, 2857–2862.
30. Yu, W.; Vath, J. E.; Huberty, M. C.; Martin, S. A. Anal. Chem.
1993, 65, 3015–3023.
31. Yan, X., personal communication.
32. Biemann, K. Methods Enzymol. 1990, 193, 455–479.
J Am Soc Mass Spectrom 2000, 11, 237–243FT ICR-MS OF RHM-CSF? AND DERIVATIVES