Chemical cross-linking and high-performance Fourier transform ion cyclotron resonance mass spectrometry for protein interaction analysis: Application to a calmodulin/target peptide complex

Article abstract of DOI:10.1021/ac0487294

Chemical cross-linking has proved successful in combination with mass spectrometry as a tool for low-resolution structure determination of proteins. The integration of chemical cross-linking with Fourier transform ion cyclotron resonance (FTICR) mass spectrometry to determine protein interfaces was tested on the calcium-dependent complex between calmodulin (CaM) and a 26-amino acid peptide derived from the skeletal muscle myosin light chain kinase (M13). Different amine-reactive, homobifunctional cross-linkers and a "zero-length" cross-linker were employed. The covalently attached complexes were separated from nonreacted proteins by one-dimensional gel electrophoresis, and the bands of interest were excised and in-gel digested with trypsin. Digestion of the cross-linked complexes resulted in complicated peptide mixtures, which were analyzed by nano-HPLC/nano-ESI-FTICR mass spectrometry. The distance constraints obtained by chemical cross-linking were in agreement with the published NMR structure of the CaM/M13 complex, pointing to residues Lys-18 and Lys-19 of M13 being cross-linked with the central α-helix of CaM. Thus, the integrated approach described herein has proven to be an efficient tool for mapping the topology of the CaM/M13 complex. As such it is applicable as a general strategy for the investigation of the spatial organization of protein complexes and complements existing techniques, such as X-ray crystallography and NMR spectroscopy.

Full text of DOI:10.1021/ac0487294

Anal. Chem. 2005, 77, 495-503
Chemical Cross-Linking and High-Performance
Fourier Transform Ion Cyclotron Resonance Mass
Spectrometry for Protein Interaction Analysis:
Application to a Calmodulin/Target Peptide
Complex
†
†
‡
,†
Stefan Kalkhof, Christian Ihling, Karl Mechtler, and Andrea Sinz*
Biotechnological-Biomedical Center, Faculty of Chemistry and Mineralogy, University of Leipzig, D-04103 Leipzig, Germany,
and Institute of Molecular Pathology, Dr. Bohrgasse 7, A-1030 Vienna, Austria
Chemical cross-linking has proved successful in combina-
tion with mass spectrometry as a tool for low-resolution
structure determination of proteins. The integration of
chemical cross-linking with Fourier transform ion cyclo-
tron resonance (FTICR) mass spectrometry to determine
protein interfaces was tested on the calcium-dependent
complex between calmodulin (CaM) and a 26-amino acid
peptide derived from the skeletal muscle myosin light
chain kinase (M13). Different amine-reactive, homobi-
functional cross-linkers and a “zero-length” cross-linker
were employed. The covalently attached complexes were
separated from nonreacted proteins by one-dimensional
gel electrophoresis, and the bands of interest were excised
and in-gel digested with trypsin. Digestion of the cross-
linked complexes resulted in complicated peptide mix-
tures, which were analyzed by nano-HPLC/nano-ESI-
FTICR mass spectrometry. The distance constraints
obtained by chemical cross-linking were in agreement with
the published NMR structure of the CaM/M13 complex,
pointing to residues Lys-18 and Lys-19 of M13 being
cross-linked with the central r-helix of CaM. Thus, the
integrated approach described herein has proven to be
an efficient tool for mapping the topology of the CaM/M13
complex. As such it is applicable as a general strategy for
the investigation of the spatial organization of protein
complexes and complements existing techniques, such
as X-ray crystallography and NMR spectroscopy.
High-resolution structural analysis of proteins is currently
accomplished by NMR spectroscopy, X-ray crystallography, ands
to a certain extentselectron microscopy. These techniques,
however, are time- and material-consuming methods, and more-
over, they are not applicable to all proteins or protein complexes.
A promising strategy with the potential to obtain low-resolution
structural information on minute amounts of protein complexes
within a few days is based on a combination of chemical cross-
1
linking and mass spectrometry. Using chemical cross-linkingsa
well-established technique in protein chemistry, where covalent
bonds between different molecules (intermolecular) or between
parts of a molecule (intramolecular) are formedsit is feasible to
determine distance constraints between functional groups. The
created cross-linking products are subsequently identified by mass
spectrometric techniques.2 The mass of the protein complex to
be investigated is theoretically unlimited, since it is the proteolytic
peptides that are analyzed. Analysis is generally fast and requires
only minimal (down to femtomole or even attomole) amounts of
,3
4
analyte. Furthermore, the broad range of specificities available
for cross-linking reagents toward certain functional groups, such
as primary amines, sulfhydryls, or carboxylic acids, and the wide
range of distances different cross-linking reagents can bridge, offer
1
various options for experimental design. However, despite the
straightforwardness of the cross-linking approach, identification
of the cross-linking products can be hampered by the complexity
of the created reaction mixtures. Several strategies have been
employed to facilitate identification of the cross-linking products,
such as using isotope-labeled cross-linkers or proteins, fluorogenic
cross-linkers, or cross-linkers creating characteristic fragment ions
With recent thrusts in genome sequencing projects, the
number of identified proteins has dramatically increased during
the past few years. The physiological function of many of the newly
discovered proteins, however, remains unclear. Closely related
to studying the function of a protein is the analysis of its binding
partners. After identification of a binding partner, the elucidation
of the three-dimensional structure of a protein complex can yield
important information about the function of a protein.
1
during mass spectrometric detection. An alternative is presented
by using high-performance Fourier transform ion cyclotron
resonance (FTICR) mass spectrometry, which offers unique
analytical possibilities due to its excellent mass accuracy and high
5
resolution. Thus, a cross-linking product can be unambiguously
(
1) Sinz, A. J. Mass Spectrom. 2003, 38, 1225-1237.
(2) Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M. Science
989, 246, 64-71.
1
*
Corresponding author. Phone: +49-341-9736078, Fax: +49-341-9736115,
(3) Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 60, 2299-2301.
(4) McLafferty, F. W.; Fridriksson, E. K.; Horn, D. M.; Lewis, M. A.; Zubarev,
R. A. Science 1999, 284, 1289-1290.
E-mail: sinz@chemie.uni-leipzig.de.
†
University of Leipzig.
Institute of Molecular Pathology.
‡
(5) Comisarov M. B.; Marshall A. G. Chem. Phys. Lett. 1974, 25, 282-283.
1
0.1021/ac0487294 CCC: $30.25 © 2005 American Chemical Society
Analytical Chemistry, Vol. 77, No. 2, January 15, 2005 495
Published on Web 12/07/2004

identified solely based on determining its exact mass. We recently
presented the elucidation of the interacting regions between
calmodulin (CaM) and its target peptide melittin by integrating
EXPERIMENTAL SECTION
Materials. Bovine brain CaM was obtained from Calbiochem
(Schwalbach am Taunus, Germany) and used without further
purification. The skMLCK peptide (M13) was a generous gift from
Dr. Ad Bax, National Institutes of Health. The purity was checked
by matrix-assisted laser desorption/ionization time-of-flight (MALDI-
TOF) mass spectrometry and SDS-PAGE. The cross-linking
reagents 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydro-
chloride (EDC), N-hydroxysulfosuccinimide (sulfo-NHS), bis-
6
chemical cross-linking and FTICR mass spectrometry. In this
study, we present the combination of chemical cross-linking and
high-performance FTICR mass spectrometry as a generally
applicable, rapid approach for identifying amino acids in protein
complexes. We apply this technique to the complex between CaM
and a target peptide.
CaM is a small (148 amino acids), acidic protein belonging to
the class of EF-hand proteins, which is found ubiquitously in
3
(
sulfosuccinimidyl)suberate (BS ), and disulfosuccinimidyl tartrate
(sulfo-EGS) were purchased from Pierce (Rockford, IL). N-
7
animals, plants, fungi, and protozoa. CaM serves as a calcium-
Hydroxysuccinimide, dicyclohexylcarbodiimide, 1,4-dioxane, and
adipic acid were purchased from Sigma-Aldrich (Vienna, Austria).
Trypsin (sequencing grade) was obtained from Roche Diagnostics
8
dependent regulator in many metabolic pathways. Upon calcium
9
,10
binding, CaM adopts a dumbbell structure
lobes connected by a flexible central helix.
consisting of two
CaM is known to
1
1,12
(
(
Mannheim, Germany). luteinizing hormone releasing hormone
LHRH), sinapinic acid (3,5-dimethoxy-4-hydroxycinnamic acid),
bind with high affinity to various target proteins and peptides,
-
7
-11
7
with dissociation constants ranging between 10 and 10 M.
and proteins and peptides for MALDI-TOFMS calibration were
purchased from Sigma (Taufkirchen, Germany). Buffer reagents
were obtained from Sigma at the highest available purity. Nano-
HPLC solvents were spectroscopic grade (Uvasol, VWR, Darm-
stadt, Germany). Water was purified with a Direct-Q5 water
purification system (Millipore, Eschborn, Germany).
Synthesis of Disuccinimidyl Adipate (DSA). For synthesis
0 8 0 8
of DSA-d /d , adipic acid-d /d (0.727 mmol, 106 mg) and N-
hydroxysuccinimide (1.46 mmol, 168 mg) were dissolved in 10
mL of anhydrous 1,4-dioxane. The flask was continuously flushed
with dry argon. Dicyclohexylcarbodiimide (1.5 mL of a 1 M
The myosin light chain kinase of the smooth muscle (sm-
MLCK) is currently the functionally best-characterized representa-
tive of a class of CaM-dependent enzymes, which catalyze the
phosphorylation of a serine residue at the N-terminus of the
1
3
myosin light chain. Another structurally similar member of this
class of enzymes is the tissue-specific myosin light chain kinase
of the skeletal muscle (skMLCK). The skMLCK consists of ∼600
amino acids depending on the species (for example, the human
isoform comprises 595, the rabbit isoform 607 amino acids).
Binding studies used chymotryptic fragments of the rabbit
skMLCK and synthetic peptides identified a C-terminal peptide
2 2
solution in CH Cl , 1.5 mmol) was added via a syringe under
1
4
of the skMLCK as the CaM-binding region. This 26-amino acid
peptide has been designated as M13, comprising amino acids
continuous stirring, and the reaction was allowed to proceed at
room temperature for 20 h. The precipitated dicyclohexylurea was
removed by vacuum filtration. Dioxane was removed by rotary
evaporation, and the products were dried in high vacuum. The
identity of the products (mp 165-167 °C) was confirmed by ESI
mass spectrometry.
Cross-Linking Reactions. For cross-linking reactions with
EDC/sulfo-NHS, an aqueous CaM stock solution (1 mg/mL,
corresponding to 59.5 µM) was diluted with 20 mM MES buffer
5
77-602 of the rabbit skMLCK (or amino acids 565-590 of the
human skMLCK, respectively). Accordingly, M13 represents the
complete CaM-interacting sequence of skMLCK and thus can be
employed as a substitute for the total kinase in order to analyze
the structures of CaM and the interacting MLCK sequence within
1
5
the complex. The three-dimensional structure of the CaM/M13
complex was solved several years ago by multidimensional NMR
1
5
spectroscopy, making this complex an ideal system to test our
method.
(pH 6.8) to give a CaM concentration of 10.6 µM (volume 946
µL). A 10-µL aliquot of a 100 mM CaCl solution was added, and
2
In the present study, we aimed to demonstrate that a combina-
tion of chemical cross-linking and FTICR mass spectrometry can
rapidly yield structural information on protein complexes. We
show that the resulting data are in close agreement with data
obtained from high-resolution methods, such as NMR spectro-
scopy.
the mixture was incubated for 10 min at room temperature in
order to ensure that CaM was fully loaded with calcium. A 34-µL
aliquot of an aqueous M13 stock solution (1 mg/mL, correspond-
ing to 337.6 µM) was added to give a final M13 concentration of
1
0 µM. After incubation for 10 min, 10 µL of a freshly prepared
aqueous cross-linker solution, containing either 0.5, 1.0, or 2.0 M
EDC in addition to 0.5 M sulfo-NHS, was added. Thus, molar
excesses of 500, 1000, and 2000 of EDC over the protein/peptide
concentration were obtained. For a control, 10 µL of water was
added, instead of the cross-linker, to one CaM/M13 solution. The
reaction mixtures were incubated at room temperature and 200-
µL aliquots were taken after 5, 15, 30, 60, and 120 min. The
reactions were quenched by adding 20 µL of an 220 mM aqueous
DTT solution to each aliquot (final concentration 20 mM). Before
SDS-PAGE and MALDI-TOFMS, the solutions were desalted
using Microcon-YM-3 filters (Millipore, Eschborn, Germany).
(
6) Schulz, D. M.; Ihling, C.; Clore, G. M.; Sinz, A. Biochemistry 2004, 43, 4703-
715.
7) Crivici, A.; Ikura, M. Annu. Rev. Biophys. Biomol. Struct. 1995, 24, 85-
16.
4
(
1
(
(
8) Carafoli, E. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 1115-1122.
9) Babu, Y. S.; Sack, J. S.; Greenhough, T. J.; Bugg, C. E.; Means, A. R.; Cook,
W. J. Nature 1985, 315, 37-40.
(
(
(
10) Babu, Y. S.; Bugg, C. E.; Cook, W. J. J. Mol. Biol. 1988, 204, 191-204.
11) Persechini, A.; Kretsinger, R. H. J. Biol. Chem. 1988, 263, 12175-12178.
12) Barbato, G.; Ikura, M.; Kay, L. E.; Pastor, R. W.; Bax, A. Biochemistry 1992,
1, 5269-5278.
13) Gallagher, P. J.; Herring, B. P.; Stull, J. T. J. Muscle Res. Cell Motil. 1997,
8, 1-16.
14) Herring, B. P.; Stull, J. T.; Gallagher, P. J. J. Biol. Chem. 1990, 265, 1724-
730.
3
(
(
(
1
For cross-linking experiments with the homobifunctional cross-
1
3
linking reagents sulfo-EGS, BS or DSA-d
0
/d
8
, an aqueous CaM
15) Ikura, M.; Clore, G. M.; Gronenborn, A. M.; Zhu, G.; Klee, C. B.; Bax A.
Science 1992, 256, 632-638.
stock solution (1 mg/mL) was diluted with 20 mM Hepes buffer
496 Analytical Chemistry, Vol. 77, No. 2, January 15, 2005

(
pH 7.4) containing 1 mM CaCl
2
to give solutions containing CaM
concentrated on a trapping column (PepMap, C18, 300 µm × 5
mm, 3 µm, 100 Å, Dionex) with water containing 0.1% formic acid
at flow rates of 20 µL/min. After 3 min for desalting, the peptides
were eluted onto the separation column (PepMap, C18, 75 µm ×
150 mm, 3 µm, 100 Å, Dionex), which had been equilibrated with
95% A (solvent A: water containing 0.1% formic acid). Peptides
were separated using the following gradient: 0-30 min, 5-50%
B; 30-31 min, 50-95% B; 31-45 min, 95% B (solvent B:
acetonitrile containing 0.1% formic acid) at flow rates of 200 nL/
min and detected based on their UV absorptions at 214 and 280
nm.
at a concentration of 10.44 µM (volume 957 µL). After an
incubation time of 10 min, 33 µL of an M13 solution (final
concentration 10 µM) was added. The cross-linking reactions were
started by adding 10 µL of solutions containing either sulfo-EGS,
BS (each concentrations of 0.1, 0.05, or 0.01 M in DMSO), or
0 8
DSA-d /d (at concentrations of 0.1, 0.05, or 0.02 M in DMSO),
thus yielding relative molar excesses of cross-linker over protein/
peptide concentration of 100, 50, 10 (for sulfo-EGS and BS ), and
2
cross-linker solution, but 10 µL of DMSO was added instead. The
reaction mixtures were incubated at room temperature, and 200-
µL aliquots were taken after 5, 15, 30, 60, and 120 min. The
reactions were terminated by adding 20 µL of a 220 mM
4 3
NH HCO solution (final concentration 20 µM) to each aliquot.
Before SDS-PAGE and MALDI-TOFMS, the solutions were
desalted using Microcon-YM-3 filters (Millipore).
SDS-PAGE and Enzymatic Proteolysis. Following separa-
tion of the reaction mixtures by one-dimensional SDS-PAGE
(
and in-gel digested as described previously. Trypsin (50 ng/
µL) was used as digestion enzyme for all samples. Depending on
the volume of the gel pieces, between 2 and 5 µL of enzyme
solution was added, and the digests were incubated at 37 °C for
3
3
0 8
0 (for DSA-d /d ). One solution was prepared without adding
The nano-HPLC system was coupled on-line to an APEX II
FTICR mass spectrometer equipped with a 7-T superconducting
magnet (Bruker Daltonics, Billerica, MA) and a nanoelectrospray
ionization source (Agilent Technologies, Waldbronn, Germany).
For nano-ESIMS, distal coated fused-silica PicoTips (tip i.d. 8 µm,
New Objective, Woburn, MA) were applied. The capillary voltage
was set to -1400 V. Mass spectral data were acquired in the
broadband mode over an m/z range of 400-2000 with 256K data
points, four scans were accumulated per spectrum, and 400 spectra
were recorded for each LC/MS run. MS data acquisition was
initialized with a trigger signal from the HPLC system 7 min after
initiation of the LC gradient. Data were acquired over 34.5 min.
Calibration of the instrument was performed with CID fragments
(capillary exit voltage 200 V) of the LHRH peptide (y5 (m/z
499.2987), LHRH ([M + H]2 m/z 592.2358), y7 (m/z 749.3941),
1
6
15%, Coomassie staining ), the bands of interest were excised
1
7
1
6 h.
MALDI-TOF Mass Spectrometry. MALDI-TOF mass spec-
+
+
trometry was performed on a Voyager DE RP Biospectrometry
Workstation (Applied Biosystems, Foster City, CA) equipped with
a nitrogen laser (337 nm). The instrument was run in positive
ionization mode, and measurements were performed in the linear
mode (detection range m/z 16 000-24 000) as well as in the
reflector mode (detection range m/z 2500-4000) using sinapinic
acid and R-cyano-4-hydroxycinnamic acid as matrixes. A saturated
matrix solution was prepared in 30% (v/v) acetonitrile, 69.9% (v/
v) water, and 0.1% (v/v) TFA. Samples were prepared using the
dried droplet method by spotting 0.5 µL of matrix solution and
and y8 (m/z 935.4734), LHRH ([M + H] m/z 1183.5643)). Data
acquisition was performed using the XMASS software (version
6.1.2); data processing was performed with the XMASS software
(version 5.0.10) (Bruker Daltonics). Prior to Fourier transforma-
tion, the time-domain signals were doubly zero filled to enhance
the digitization of the spectra and to improve their visualization
followed by apodization with a sine function. A total of 400 single
spectra were projected into one final mass spectrum using the
Projection tool in the XMASS software.18
Identification of Cross-Linking Products. Cross-linking
products were identified using the GPMAW (General Protein
Mass Analysis for Windows) software, version 6.01 (Lighthouse
Data, Odense, Denmark) (available at: http://welcome.to/gp-
maw), the ExPASy Proteomics tools in the Swiss-Prot Database
(available at: http://www.expasy.org), and the ASAP (Automatic
Spectrum Assignment Program) software (available at: http://
roswell.ca.sandia.gov/∼mmyoung /asap.html). Proteolytic cleav-
ages at modified amino acids, such as the trimethylated K115 in
CaM as well as amino acids modified by cross-linking reagents
were excluded. The N-terminus of CaM was excluded from
possible cross-linking since it is acetylated.
Determination of Distances between Atoms in the CaM/
M13 Complex. The NMR structure of the CaM/M13 complex
is deposited in the RCSB Protein Data Bank (http://www.
rcsb.org/pdb/) under the entry 2BBM.15 Based on the atom
coordinates of this structure, the distances between the nitrogen
atoms of the ꢀ-amine groups of lysine residues and the N-terminus
of M13, respectively, were determined using the VMD-XPLOR
visualization package19 (available at: http://vmd-xplor.cit.nih.gov).
The VMD-XPLOR software also calculated distances between
0
.5 µL of sample solution onto the target. Spectra from 100 laser
shots were accumulated to one spectrum when operating the
instrument in the reflector mode, and 300 laser shots were added
when spectra were acquired in the linear mode. In the linear mode,
the instrument was calibrated using cytochrome c ([M + H] average
at m/z 12 361) and myoglobin ([M + H]+average at m/z 16 952).
For calibration in the reflector mode, signals of angiotensin I ([M
+
1
1
were performed using the Voyager software version 5.1 and the
Data Explorer software version 4.0 (Applied Biosystems, Foster
City, CA).
Nano-HPLC/Nano-ESI FTICR Mass Spectrometry. Peptide
mixtures from the enzymatic digests were separated by nano-
HPLC. Nano-HPLC was carried out on an Ultimate Nano-LC
system (Dionex, Idstein, Germany) equipped with a Switchos II
column switching module and a Famos microautosampler with a
+
+
+
H] mono at m/z 1296.69), substance P ([M + H] mono at m/z
+
347.74), and somatostatin (reduced form) ([M + H] mono at m/z
637.72) were employed. Data acquisition and data processing
2
0-µL sample loop. Samples were injected by the autosampler and
(
16) Laemmli, U. K. Nature 1970, 227, 680-685.
(
17) Jensen, O. N.; Shevchenko, A.; Mann, M. In Protein Structure, A Practical
Approach; Creighton, T, E., Ed.; Oxford University Press: Oxford, U.K.,
(18) Bruker Daltonics BioAPEX User’s Manual v 1.1; Billerica, MA, 1996.
(19) Schwieters, C. D.; Clore, G. M. J. Magn. Reson. 2001, 149, 239-244.
1
997; pp 48.
Analytical Chemistry, Vol. 77, No. 2, January 15, 2005 497

depends on a detailed description of their respective primary
structures. ESI-FTICRMS of CaM yielded the most abundant mass
at m/z 16 790.921 (most abundant mass from simulation at m/z
1
6 790.884, ∆m ) 0.037 u, 2.1 ppm). Peptide mass fingerprinting
of CaM using trypsin yielded full sequence coverage. CaM was
confirmed to be N-terminally acetylated and to contain a trim-
ethylated lysine at position 115.6 The monoisotopic mass of M13
was determined by ESI-FTICR mass spectrometry as 2962.702 u
,20
(calculated 2962.717 u, ∆m ) 0.015 u, 5.0 ppm). The fragments
created by tryptic cleavage did not cover the complete sequence
of M13; however, when additional data from the cross-linked
CaM/M13 complex were also considered, a total of 100% sequence
coverage was achieved. The amino acid sequences of CaM and
M13 are shown in Figure 1.
Figure 1. Amino acid sequences of (A) CaM and (B) M13. CaM
was found to be N-acetylated (Ac-Ala) and to contain a trimethylated
lysine in position 115 (the modified amino acids Ala-1 and Lys-115
are underlined).
Cross-Linking Reactions. Cross-linking experiments between
CaM and M13 were performed as described in the Experimental
Section. Figure 2 provides a schematic overview of the analytical
strategy employed. The zero-length cross-linker EDC in combina-
C-atoms of carboxylic acids of glutamic or aspartic acid residues
and the N-atoms of amine groups of lysine residues or the
N-terminus of M13.
tion with sulfo-NHS, as well as the homobifunctional, amine-
RESULTS AND DISCUSSION
3
Characterization of CaM and M13. The investigation of the
reactive cross-linking reagents sulfo-EGS, BS , and DSA-d
0
/d
8
interaction sites between CaM and M13 in the CaM/M13 complex
(Table 1), were used for conducting the cross-linking experiments.
Figure 2. General analytical strategy for mapping the interacting sequences between CaM and M13.
98 Analytical Chemistry, Vol. 77, No. 2, January 15, 2005
4

Table 1. Structures and Spacer Lengths of the
Cross-Linking Reagents Used for This Study
Figure 3. SDS-PAGE of cross-linking reaction mixtures of CaM
10 µM) and M13 (10 µM) with EDC (20 mM)/sulfo-NHS (500 µM).
The samples are labeled as follows: Protein standard (S), CaM/M13
1:1) mixture without cross-linker (R), reaction mixtures after 5 (5),
5 (15), 30 (30), 60 (60), and 120 min (120).
(
(
1
between ∼17 000 and 20 000 were observed on the gel, corre-
sponding to CaM and the CaM/M13 (1:1) complex (Figure 3).
Additionally, a band in the low molecular weight region represent-
ing M13 indicated that cross-linking was not complete. The
appearance of a band in the low molecular weight region in
addition to the M13 band might be caused by intramolecular cross-
linking of M13 causing a compact structure of the peptide with a
higher mobility than M13 itself. Aggregation of proteins caused
by excessive cross-linking was not observed for any of the cross-
linking reagents as was evidenced by the absence of any gel bands
in the higher mass range.
MALDI-TOF Mass Spectrometry. MALDI-TOFMS was em-
ployed to estimate the extent of chemical cross-linking over the
course of the cross-linking reaction. MALDI-TOF mass spectrom-
etry yields information on the number of intra- and intermolecular
cross-linking products as well as on the number of modifications
caused by hydrolyzed cross-linker. Figure 4 shows MALDI-TOF
mass spectra of the nondigested reaction mixtures of CaM and
M13 after incubation times of 5, 15, 30, 60, and 120 min with sulfo-
EGS at 50 molar excess over the protein/peptide concentration.
Based on the cross-linking experiments conducted with sulfo-EGS,
complex formation was found to be complete after ∼60 min when
sulfo-EGS was applied at 50- or 100-fold molar excess. The number
of incorporated cross-linker molecules increased from three to
four during the second hour of the reaction (Figure 4). With
increasing reaction time, the formation of a small amount of CaM/
M13 (1:2) complex was also observed. Applying a low cross-linker
concentration resulted in reduced complex formation as well as
reduced modification with hydrolyzed cross-linker. For the reac-
tion employing a 10-fold molar excess of cross-linking reagent over
the protein/peptide concentration, we observed a highly reduced
yield of cross-linking product after 120 min (data not shown). The
cross-linker EDC/sulfo-NHS had to be used at a significantly
higher molar excess in comparison to the homobifunctional amine-
reactive cross-linkers employed in this study. Optimization of
cross-linking reaction conditions resulted in the formation of a
EDC reacts with the carboxylic acid group of an aspartic or
glutamic acid or the C-terminus of a protein to form a short-lived
active O-acylurea intermediate. Sulfo-NHS reacts with the EDC
active-ester complex and extends the half-life of the activated
carboxylate. In the presence of an ꢀ-amino group of lysine or
the free N-terminus of a protein, an amide bond is formed, leading
2
1
to the loss of an H
8.011 u per cross-link.
Homobifunctional cross-linking reagents contain two identical
2
O molecule and yielding a mass decrease of
1
functional groups on either side of the molecule that are separated
by a spacer bridging a defined distance. Homobifunctional sulfo-
3
NHS-esters, such as sulfo-EGS, BS , and DSA, are highly reactive
toward primary amine groups; however, they are susceptible to
3
hydrolysis. Upon cross-linking, sulfo-EGS, BS , and DSA produce
amide bond cross-linked molecules causing a mass shift of
1
13.995, 138.068, and 110.0368 u (DSA-d
partially hydrolyzed cross-linkers exhibit a mass increase of
32.006, 156.079, and 128.0472 u (DSA-d ), respectively.
The optimum cross-linking reaction results in the high-yield
0
), respectively, whereas
1
0
formation of a cross-linked complex without creating oligomers
and without distorting the tertiary structure of the complex by
excessive cross-linking. Therefore, the right choice of both the
cross-linker concentration and the reaction times is crucial for
successfully conducting cross-linking experiments.
SDS-PAGE. After the cross-linking reaction, the reaction
mixtures were separated by one-dimensional SDS-PAGE and the
gels were stained with Coomassie Brilliant Blue. Figure 3 shows
a one-dimensional gel of CaM, M13, and the cross-linking reaction
product of an equimolar mixture of both proteins employing EDC/
sulfo-NHS (20 mM/500 µM) after incubation times of 5, 15, 30,
6
0, and 120 min. CaM and M13 possess molecular weights of
about 16 800 and 29 00, respectively. Thus, the cross-linked
complex between CaM and M13 is expected to migrate on the
gel at ∼20 000 depending on the extent of chemical cross-linking.
After incubation times of 5, 15, 30, and 60 min, two distinct bands
(
(
20) Watterson, D. M.; Sharief, F.; Vanaman, T. C. J. Biol. Chem. 1980, 255,
62-975.
21) Hermanson, G. T. Bioconjugate Techniques, 1st ed.; Academic Press: San
9
Diego, 1996.
Analytical Chemistry, Vol. 77, No. 2, January 15, 2005 499

Table 2. Intermolecular Cross-Linking Products between CaM and M13 with EDC/Sulfo-NHSa
distance
∆m
(ppm)
identified in
sample
[
M+H]+calc
[M+H]+exp
CaM
M13
(Å)
2
224.113
2224.136
75-86
D 78
D 80
E 82
E 83
E 84
78-86
E 82
E 83
E 84
D 78
D 80
19-26
K 19
K 19
K 19
K 19
K 19
19-26
K 19
K 19
K 19
K 19
K 19
11
7
EDC_30min_2000x
EDC_120min_1000x
2
224.129
13.76
12.47
11.62
3.66
6.90
1836.882
1836.891
5
EDC_60min_2000x
11.62
3.66
6.90
13.76
12.47
a
The reaction times and the excess of EDC in the samples, in which cross-links were identified, are given.
(amino acids 19-26) of M13 with amino acid sequences 75-86
and 78-86 of CaM (Table 2). The sequences of CaM contain as
possible reactive sites two aspartic acid residues at positions 78
and 80 as well as three glutamic acid residues at positions 82, 83,
and 84. Which amino acids were actually cross-linked by reaction
with Lys-19 of M13, however, remains unclear without conducting
MS/MS experiments.
Calculating the atomic distances between the nitrogen atoms
of the amine groups of lysines and the C-atoms of carboxylic acid
groups of aspartic acids and glutamic acids revealed that only a
single distance is under 5 Å. This distance is between Glu-83 (in
CaM) and Lys-19 (in M13). The possible cross-linking product
between these two amino acids is in agreement with the fact that
the zero-length cross-linker EDC is able to connect amine and
carboxylic acid groups, which are within distances of under 5 Å.
Cross-Linking Products Obtained with Sulfo-EGS. In
addition to the zero-length cross-linker EDC, three different amine-
reactive homobifunctional cross-linking reagents with spacer
Figure 4. MALDI-TOF mass spectra (detection range m/z 16 t000-
24t000) of the reaction mixture between CaM and M13 with 50-fold
3
lengths of 8.9 (DSA), 11.4 (BS ), and 16.4 Å (sulfo-EGS) were
molar excess of sulfo-EGS at different reaction times (from bottom
employed. For all three reagents, the formation of a cross-linked
CaM/M13 (1:1) complex was confirmed using MALDI-TOFMS
and SDS-PAGE. Using sulfo-EGS, two cross-linking products
were identified by nano-HPLC/nano-ESI-FTICRMS analysis of the
proteolytic peptide mixtures (Table 3). A signal at m/z 1393.724
was assigned as cross-linking product between amino acids 75-
to top: 5-120 min).
sufficient amount of cross-linked complex after 15 min, if a 500-
2
000-fold molar excess of EDC in combination with a 500-fold
molar excess of sulfo-NHS was employed (data not shown).
Analysis of Cross-Linking Products. Following SDS-PAGE
separation of the cross-linking reaction mixtures, the cross-linked
CaM/M13 (1:1) complexes were excised from the gel and
subjected to enzymatic in-gel digestion with trypsin (Figure 2).
The created peptide mixtures were separated by nano-HPLC and
analyzed by nano-ESI-FTICR mass spectrometry (Figure 2). These
mixtures were rather complex, as they contained peptides derived
from CaM and M13 in addition to inter- and intramolecular cross-
linking products and peptides modified by hydrolyzed cross-linker.
The experimentally obtained monoisotopic masses were compared
to calculated masses of peptides and cross-linking products
employing the GPMAW and ASAP software packages.
7
7 of CaM and amino acids 19-26 of M13. As cross-linked amino
acids, the residues Lys-19 (in M13) and Lys-75 (in CaM) were
identified. As trypsin is expected not to cleave after a modified
lysine residue, Lys-77 of CaM could be excluded as a cross-linked
amino acid. Another cross-linking product composed of amino
acids 17-19 (M13) and amino acids 76-86 (CaM) was identified
based on the appearance of a signal at m/z 1999.916. In addition
to the intermolecular cross-linking products between CaM and
M13, an intramolecular cross-linking product within M13 consist-
ing of the amino acid sequence 17-26 of M13 (Table 4) was
identified in analyzing the CaM/M13 (1:1) complex. This cross-
linking product could only have been created by cross-linking
between M13 lysines at positions 18 and 19. Moreover, several
modified peptides of M13 were identified as lysines at positions
5, 6, 18, and 19 in M13 derivatized with hydrolyzed sulfo-EGS
(data not shown).
Cross-Linking Products Obtained with EDC/Sulfo-NHS.
The “zero-length” cross-linker EDC was employed in combination
with sulfo-NHS at a molar excess of 500-, 1000-, and 2000-fold
compared to protein/peptide concentration. Data analysis yielded
two cross-linking products composed of the C-terminal sequence
500 Analytical Chemistry, Vol. 77, No. 2, January 15, 2005

Table 3. Intermolecular Cross-Linking Products between CaM and M13 with the Homobifunctional, Amine-Reactive
3
a
Cross-Linkers Sulfo-EGS, BS , and DSA (d0/d8)
cross-linker
sulfo-EGS
[M+H]+calc
[M+H]+exp
CaM
M13
distance (Å)
∆m (ppm)
identified in sample
1393.724
1393.734
75-77
K 75
76-86
K 77
19-26
K 19
17-19
K 18
7
8
1
2
1
1
4
1
4
0
EGS_120min_100x
EGS_120min_200x
EGS_60min_200x
EGS_15min_200x
EGS_60min_100x
EGS_120min_100x
1
393.735
18.12
29.79
1
999.916
1999.918
1
1
1
999.920
999.918
999.915
BS3
2380.191
2380.181
75-86
K 75
19-26
K 19
BS _30min_100x
3
3
2
2
380.193
380.182
18.12
20.58
BS _60min_100x
3
K 77
K 19
BS _100x_120min
3
2
2
2
396.186 [M76 oxidized]
2396.185
2699.332
2040.032
1883.904
75-86
K 75
19-26
K 19
BS _60min_100x
18.12
20.58
K 77
K 19
3
3
699.360 (BS + linker)
75-90
K 75
17-19
K 18
10
6
BS _60min_100x
26.24
29.79
K 77
K 18
3
040.032
75-86
K 75
17-19
K 18
BS _60min_100x
26.24
29.79
K 77
K 18
DSA-d
0
8
1883.906
1891.956
76-86
K 77
17-19
K 18
1
2
4
3
DSA_120min_100x
DSA_120min_50x
DSA_120min_100x
DSA_120min_50x
1
883.902
1891.948
891.950
29.79
DSA-d
1
a
The reaction times and the excess of cross-linking reagent in the samples, in which cross-links were identified, are given.
Table 4. Intramolecular Cross-Linking Products in CaM and M13 with the Homobifunctional, Amine-Reactive
Cross-Linkers Sulfo-EGS and BS3a
cross-linker
sulfo-EGS
BS3
[M+H]+calc
1265.662
[M+H]+ex
1265.650
1175.672
component
M13
sequence
distance (Å)
10.62
∆m (ppm)
identified in sample
EGS_60min_50x
17-26
9
K18/K19
17-26
3
3
1175.667
M 13
CaM
CaM
10.62
4
0
2
BS _30min_100x
1
175.667
K18/K19
75-90
BS _120min_100x
3
2
2
122.012
138.007
2122.007
5.38
BS _120min_100x
K75/K77
75-90
3
2138.003
5.38
2
BS _120min_100x
[M76 oxidized]
K75/K77
a
The reaction times and the excess of cross-linking reagent in the samples, in which cross-links were identified, are given.
3
3
Cross-Linking Products Obtained with BS . Using BS , four
intermolecular cross-linking products identified between CaM and
M13 pointed conclusively to a cross-linking between the sequence
of amino acids 75-90 of CaM and the amino acid sequence 17-
19 of M13 (Table 3). Figure 5 shows the deconvoluted ESI-FTICR
mass spectrum of a tryptic peptide mixture of the CaM/M13 (1:
3
1
) complex, which had been cross-linked by BS . The signal at
m/z 2380.193 corresponds to the singly charged ion of a cross-
linking product consisting of sequence 75-86 of CaM and
sequence 19-26 of M13. The signal at m/z 2396.185 was identified
as the Met-oxidized (Met-76 in CaM) derivative of the respective
cross-linking product. Additionally, three intramolecular cross-
linking products were detected (Table 4) and identified as amino
acid Lys-18 of M13 cross-linked with Lys-19 and Lys-75 of CaM
cross-linked with Lys-77. Moreover, we identified several peptides
3
of both M13 and CaM that had been modified by hydrolyzed BS .
In M13, the N-terminal amino acid and in CaM, lysine residues
at positions 21, 77, and 94 were found to be modified, indicating
that the functional groups of these amino acids are highly solvent-
exposed.
Cross-Linking Products Obtained with DSA. DSA was
employed for the cross-linking reaction as a 1:1 mixture with its
Figure 5. Deconvoluted ESI-FTICR mass spectrum of the tryptic
peptide mixture from the CaM/M13 (1:1) complex cross-linked with
3
BS (100-fold molar excess over protein/peptide concentration) at an
incubation time of 60 min. The inset shows the magnified signal of a
cross-linking product between CaM residues 75-86 and M13 resi-
dues 19-26 at m/z 2380.193, which was also detected with one
methionine residue oxidized at m/z 2396.185; n.a., signal not as-
signed.
8
fully deuterated derivative (d ). Therefore, an additional criterion
Analytical Chemistry, Vol. 77, No. 2, January 15, 2005 501

Figure 6. NMR structure of the CaM/M13 complex according to ref 15 (pdb entry 2BBM). The side chains of lysines and acidic amino acids
that are potentially involved in cross-linking are indicated.
3
for the identification of cross-linking products was introduced, as
every cross-linking product should exhibit two signals at a distance
of 8 u with approximately the same intensity.22 Two signals
showing this mass difference were detected, which were identified
as deuterated and nondeuterated cross-linking product between
between the amino acid sequence 76-86 in CaM and the amino
acid sequence 17-19 in M13 (Table 3). As trypsin was cleaving
after Lys-19 in M13, lysine in position 18 could be unambiguously
identified as the amino acid cross-linked with Lys-77 of CaM.
Intermolecular Cross-Linking Products between CaM and
M13. Based on the published NMR structure (pdb entry 2BBM),
the distances of N atoms of all lysine residues in CaM and M13
were calculated. The NMR structure exhibits only one intermo-
lecular distance lower than 10 Å, another distance under 15 Å,
and nine intermolecular distances of under 20 Å. For the amine-
reactive homobifunctional cross-linkers, we successfully identified
seven cross-linking products containing sequences from the
central R-helix of CaM (residues 75-90) and amino acids 18 and
linkers used for this study: DSA, BS , and sulfo EGS (spacer
lengths between ∼9 and 16 Å). According to the published NMR
15
structure, the N-atoms of these two lysine residues are ∼29 Å
apart from each other (Figure 6). However, a closer look at the
published structure reveals that the region comprising amino acids
74-82 in CaM is poorly defined and that the nitrogen atoms of
Lys-77 in CaM and Lys-18 in M13 are within ∼20 and 35 Å of
each other. Apparently, the amine groups of Lys-77 in CaM and
3
Lys-18 in M13 are coming close enough to be cross-linked by BS ,
DSA, and sulfo-EGS.
CONCLUSION
We successfully demonstrated the applicability of chemical
15
cross-linking in combination with high-resolution FTICR mass
spectrometry as a rapid method to determine interfaces between
proteins or between a protein and a peptide. One advantage of
the described strategy consists of the possibility of analyzing the
topology of protein complexes within a few dayssgiven that data
analysis can be automated. (For the present study, however,
analysis of the CaM/M13 complex required several weeks, which
was mainly caused by the time-intense assignment of cross-linking
products in the mass spectra.) Other advantages of the presented
strategy include the low material consumption, allowing for
analysis of femtomole amounts of proteins, as well as the
possibility of obtaining solution structures of proteins. When
lysines serve as targets for cross-linking reactions, valuable
structural information of a protein can be obtained despite the
flexibility of the lysine side chains. For a comprehensive structural
analysis of a protein complex, however, one should apply a variety
of cross-linking reagents with different reactivities, targeting both
flexible and rigid amino acid side chains.
1
9 of M13. Figure 6 shows the published NMR structure of the
CaM/M13 complex, indicating the amino acid side chains that
are potentially involved in cross-linking. It can be noted that all
identified cross-linking products are in agreement with published
structural data. However, we did not identify any cross-linking
products containing lysines in positions 13, 21, 94, or 148 with
any of the lysine residues of M13, although they are located less
than 20 Å from each other.
Another interesting aspect is that Lys-18 of M13 was found to
be cross-linked with Lys-77 of CaM with all amine-reactive cross-
(22) Pearson, K. M.; Pannell, L. K.; Fales, H. M. Rapid Commun. Mass Spectrom.
2002, 16, 149-159.
502 Analytical Chemistry, Vol. 77, No. 2, January 15, 2005

In general, the described strategy should not be considered
as competing with high-resolution methods for three-dimensional
structure analysis, such as NMR spectroscopy or X-ray crystal-
lography, as only low-resolution structures are obtained. One
major drawback of this approach is that not all predicted cross-
linking products were found, although the functional groups with
the right specificities were the correct distance for cross-linking.
For example, no cross-linking products were identified between
Lys-21 in CaM and Lys-6 in M13 although the two nitrogen atoms
are within a distance of ∼15 Å. One cannot draw any conclusion
if a defined cross-linking product is not found, since the cross-
linking product may not have been created at all or the created
product may not have been amenable to subsequent mass
spectrometric analysis. There is also a slight possibility that cross-
linked peptides do not elute from the gel after enzymatic in-gel
digestion and thus escape from analysis.
exact identification of cross-linked amino acids, MS/MS experi-
ments need to be conducted. With our instrumentation, however,
it is not feasible to perform on-line MS/MS experiments, for
example, using electron capture dissociation or infrared multipho-
ton dissociation during an LC/MS analysis. For future studies,
we are aiming to employ instruments such as the novel com-
mercially available hybrid FTICR mass spectrometers, which
offer the possibility to conveniently acquire MS/MS data during
an LC/MS run.
2
3
ACKNOWLEDGMENT
The junior research group of A.S. is funded by the Saxon State
Ministry of Higher Education, Research and Culture. Financial
support from the Thermo Electron Corp. (Mattauch-Herzog award
of the German Society for Mass Spectrometry to A.S.) is also
gratefully acknowledged. The authors are indebted to Dr. Ad Bax
for the generous gift of M13, to Dr. Mike Lewis for critical reading
of the manuscript, and to Ms. Daniela Schulz for valuable
discussions.
When studying relatively small complexessas in the present
casesfor which the amino acid sequences of the binding partners
are known, mass measurement accuracies of up to ∼15 ppm are
sufficient in order to unambiguously identify a cross-linking
product. For larger complexes, additional information, e.g., by
using isotope-labeled cross-linkers, is needed. Moreover, for an
Received for review August 26, 2004. Accepted October
22, 2004.
(
23) Gershon, D. Nature 2003, 424, 581-587.
AC0487294
Analytical Chemistry, Vol. 77, No. 2, January 15, 2005 503