The First Zero-Length Mass Spectrometry-Cleavable Cross-Linker for Protein Structure Analysis

Article abstract of DOI:10.1002/anie.201708273

Combining the properties of a zero-length cross-linker with cleavability by tandem mass spectrometry (MS/MS) poses great advantages for protein structure analysis using the cross-linking/MS approach. These include a reliable, automated data analysis and the possibility to obtain short-distance information of protein 3D-structures. We introduce 1,1′-carbonyldiimidazole (CDI) as an easy-to-use and commercially available, low-cost reagent that ideally fulfils these features. CDI bridges primary amines and hydroxy groups in proteins with the lowest possible spacer length of one carbonyl unit (ca. 2.6 ?). The cross-linking reaction can be conducted under physiological conditions in the pH range between 7.2 and 8. Urea and carbamate cross-linked products are cleaved upon collisional activation during MS/MS experiments generating characteristic product ions, greatly improving the unambiguous identification of cross-links. Our innovative analytical concept is exemplified and applied for bovine serum albumin (BSA), wild-type tumor suppressor p53, an intrinsically disordered protein, and retinal guanylyl cyclase activating protein-2 (GCAP-2).

Full text of DOI:10.1002/anie.201708273

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: The First "Zero-Length" Mass Spectrometry-Cleavable Cross-
Linker for Protein Structure Analysis
Authors: Christoph Hage, Claudio Iacobucci, Anne Rehkamp, Christian
Arlt, and Andrea Sinz
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Angew. Chem. Int. Ed. 10.1002/anie.201708273
Angew. Chem. 10.1002/ange.201708273
Link to VoR: http://dx.doi.org/10.1002/anie.201708273
http://dx.doi.org/10.1002/ange.201708273

10.1002/anie.201708273
Angewandte Chemie International Edition
COMMUNICATION
The First “Zero-Length” Mass Spectrometry-Cleavable Cross-
Linker for Protein Structure Analysis
Christoph Hage#, Claudio Iacobucci#, Anne Rehkamp, Christian Arlt, and Andrea Sinz*
Abstract: Combining the properties of a “zero-length” cross-linker
with cleavability by tandem mass spectrometry (MS/MS) poses great
advantages for protein structure analysis using the cross-linking/MS
approach. These include a reliable, automated data analysis and the
possibility to obtain short-distance information of protein 3D-
structures. We introduce 1,1’-carbonyldiimidazole (CDI) as an easy-
to-use and commercially available, low-cost reagent that ideally
fulfils these features. CDI bridges primary amines and hydroxyl
groups in proteins with the lowest possible spacer length of one
carbonyl unit (~2.6 Å). The cross-linking reaction can be conducted
at physiological conditions in the pH range between 7.2 and 8. Urea
and carbamate cross-linked products are cleaved upon collisional
activation during MS/MS experiments generating characteristic
product ions, greatly improving the unambiguous identification of
cross-links. Our innovative analytical concept is exemplified and
applied for bovine serum albumin (BSA), wild-type tumor suppressor
p53, an intrinsically disordered protein, and retinal guanylyl cyclase
activating protein-2 (GCAP-2).
based MS-cleavability with virtually “zero-length” (2.6 Å) spacer
length.
Phosgene and its derivatives[3] are well-known reagents for the
preparation of urea compounds under anhydrous conditions.
Interestingly, Padiya et al.[4] described imidazole carbonylation of
amines in water for preparing urea and carbamate compounds.
Inspired by this finding, we envisioned the use of phosgene
derivatives as amine- and hydroxyl-reactive, ultra-short cross-
linkers. We screened the reactivities of four reagents (Scheme
1) towards three different proteins at various temperatures and
pH values. This systematic study allowed us to identify CDI as
effective protein cross-linker (Table S1, SI). The inefficacy of the
three other reagents can be explained by either their low
reactivity or fast hydrolysis rate. This is in agreement with the
hydrolysis kinetics reported by Staab[5] who suggested the
following reactivity order of CDP < CDI < CD-1,2,4-T.
Chemical cross-linking/mass spectrometry (MS) has evolved as
a powerful tool for 3D-structural analysis of proteins and protein
complexes.[1] The cross-linker has a defined length and acts as
a “molecular ruler”, imposing distance constraints between
connected amino acids. After the cross-linking reaction,
enzymatic digestion of the covalently connected protein(s) yields
complex peptide mixtures that are notoriously difficult to handle.
Accurate and automated identification of cross-links is currently
best achieved by using MS-cleavable cross-linkers.[2] These
novel linkers, such as disuccinimidyl sulfoxide (DSSO)[2c] and
Scheme 1. Structure of reagents (disuccinimidyl carbonate (DSC), 1,1’-
carbonyldiimidazole (CDI), 1,1’-carbonyl-di-1,2,4-triazole (CD-1,2,4-T), 1,1’-
carbonyldipyrazole (CDP)) that were tested as urea-based, ultra-short protein
cross-linkers. CDI has been identified herein as the most effective reagent.
Reactivity. CDI belongs to an extraordinary class of electrophilic
reactive amides, termed azolides, where the amide nitrogen is
part of a quasi-aromatic five-membered heterocycle containing
at least two nitrogens, i.e. azoles.[5] The amide nitrogen of
azolides is 'pyrrole-like' and its lone electron pair contributes to
the aromatic sextet of the azole. This significantly reduces the
partial double bond character of the amide bond as well as the
electron density of the nitrogen atom. This polarizes the
exocyclic bond even further, increasing the electrophilicity of the
central carbonyl. The reactivity of azolides towards nucleophiles
is comparable to those of acyl halides and anhydrides, but they
benefit from a tunable reactivity and hydrolysis rate, depending
on the number and position of nitrogen atoms in the ring.[5]
Among the azolides evaluated in this study (CDP, CDI, and CD-
1,2,4-T; Scheme 1; Table S1, SI), CDI emerged as efficient
cross-linker that is applicable in an aqueous environment at
physiological pH conditions. CDI is inexpensive (~ 0.30 €/g) and
cross-links proteins in the pH range between 7.2 and 8.0 at
temperatures higher than 10°C.
disuccinimidyl dibutyric urea (DSBU or BuUrBu)[2f]
,
are
commercially available and drastically diminish potential false-
positive identification rates of cross-links.
The spacer lengths of the MS-cleavable cross-linkers are 10.1 Å
for DSSO and 12.5 Å for DSBU, allowing bridging side chains of
amino acids that are spatially apart from each other. DSSO and
DSBU are highly valuable for thoroughly mapping large protein
assemblies, however, they do not ensure to derive geometrical
constraints for higher resolution computational modeling due to
their relatively long spacers.
Along the lines of extending our analytical concept of designing
urea-based MS-cleavable cross-linkers, we aimed at shortening
the spacer of DSBU to obtain more valuable geometrical
constraints for protein modeling. In this work, we describe 1,1’-
carbonyldiimidazole (CDI) as the first cross-linker merging urea-
*
Department of Pharmaceutical Chemistry and Bioanalytics, Institute
of Pharmacy, Martin Luther University Halle-Wittenberg Wolfgang-
Langenbeck-Str. 4, D-06120 Halle/Saale (Germany), E-mail:
andrea.sinz@pharmazie.uni-halle.de
CDI has been found to react with primary amines in lysine side
chains yielding urea products. Interestingly, CDI has also been
#
Both authors contributed equally to this paper.
SI for this article is given via a link at the end of the document.
This article is protected by copyright. All rights reserved.

10.1002/anie.201708273
Angewandte Chemie International Edition
COMMUNICATION
the most common sources of cross-links misassignments.[9] In
fact, cross-links composed of consecutive amino acid
sequences are isobaric to peptides involving identical amino
acid sequences that are modified by a partially hydrolyzed
cross-linker (Figure S1, SI).[9]
MS-cleavability. Upon collisional activation, CDI cross-linked
products follow the well-known fragmentation pathway of the
urea-based DSBU linker[2f]. In detail, two characteristic 26-u
doublets are visible in the product ion mass spectrum, which
allows distinguishing between different cross-link types. In case
of the CDI cross-linker, one of the two diagnostic product ion
signals of each doublet corresponds to the unmodified peptide
(Scheme 2). The second one depends on the nature of cross-
linked amino acid. It can either correspond to the peptide
decorated with i) an isocyanate group, in case of reaction with
an amine group of lysine or the N-terminus; ii) an oxonium ion, in
case of reaction with a hydroxyl group of serine, threonine, and
tyrosine (Scheme 2a). Exemplary product ion spectra of CDI-
cross-links are presented in Scheme 2 and in the Supporting
Information (SI).
In case of cross-linking a primary amine and an alcohol, the
resulting product ion spectrum does not always show the
complete doublet-of-doublet pattern. The protonation of the
carbamate moiety, which triggers its cleavage in the gas-phase,
is influenced by the lower gas-phase basicity of the ester oxygen
compared to that of the amide nitrogen. Moreover, the unstable
oxonium ion, created by cleavage of the amide bond of the
carbamate, might suffer further CO2 loss. Nevertheless, the
partial lack of characteristic product ions does not hamper a safe
and potentially automated cross-link identification. Recently, Liu
et al. reported an optimized strategy for the automated
identification of DSSO cross-linked products using only one of
the four diagnostic fragments.[2b]
Scheme 2. a) Reactivity of CDI with amines (lysines, N-termini) and hydroxyls
(serines, threonines, tyrosines) in proteins. b) Diagnostic fragmentation pattern
of the urea moiety upon collisional activation in tandem MS experiments. The
two characteristic doublets are highlighted in the product ion mass spectrum of
a doubly charged ion at m/z 843.4611 (shown in green). The tandem mass
spectrum has been automatically assigned using MeroX[6] and identified as a
CDI-cross-linked product between Lys-370 and Lys-382 (364AHSSHLKSK372
–
382KLMFK386) in p53. The characteristic doublet signals caused by cleavage of
the CDI cross-linker are shown in yellow, b-and y-type fragment ions of the
peptide backbone are shown in blue and red.
BSA cross-linking. We employed BSA as model protein to
evaluate the reactivity of DSC, CDI, CD-1,2,4-T, and CDP. All
four compounds were reacted for 30 minutes at two pH values
(7.5 and 8.0), three temperatures (0, 10, 20°C), and at equimolar
concentration as well as at 10-, 20-, and 50-fold molar excess
compared to BSA (experimental details are given in the SI). CDI
emerged as a powerful reagent for protein cross-linking at
temperatures above 10°C and at both pH values (Figures S2
and S3, SI). Applying our established beam-type collision
induced dissociation (HCD)-tandem MS strategy for analyzing
cross-linked products in combination with the MeroX software[6]
for data analysis, we identified 26 unique cross-links for BSA.
The number of cross-link identifications slightly increased with
temperature and pH (Figure S2, SI). Remarkably, the high
reproducibility was striking for all reaction conditions examined
in this study (Figure S3, SI). To verify the validity of our data,
cross-links were mapped onto the published structure of BSA
(PDB ID: 3v03, resolution: 2.7 Å) to determine the Cα-Cα
distances of these residues that were connected within the BSA
monomer (Figure S4, SI). The Cα-Cα distance for elongated
lysine side chains is ~13.4 Å plus ~2.6 Å for the carbonyl group
of CDI, which results in a distance of ~16 Å to be bridged by CDI.
The average distance measured for BSA was 12.5 Å. A
maximum distance of 17.8 Å was found in our cross-linking
found to connect primary amines in lysines with hydroxyl groups
of serines, threonines, and tyrosine sides giving carbamates
(Scheme 2a). Overall, we identified ~ 43% Lys-Lys and ~ 57%
Lys-Ser/Thr/Tyr cross-links. The large percentage of cross-links
involving hydroxyl groups can be explained by a higher reactivity
of CDI. Also, the carbamates created with CDI are more stable
in aqueous solution compared to the esters formed with hydroxyl
groups using N-hydroxysuccinimide (NHS)-based cross-
linkers.[7]
Remarkably, peptides modified with a partially hydrolyzed cross-
linker (type-0 or so-called “dead-end” cross-links)[8] were not
identified as the created carbamic acid and carbonate
compounds are unstable. Carbamic acids decarboxylate and
release the free amine or alcohol, which can further react with
another CDI molecule. This presents a distinct advantage of CDI
compared to conventional amine-reactive NHS-based cross-
linkers, which undergo irreversible hydrolysis to form stable
carboxylic acid derivatives.[9] The absence of peptides that are
modified by a partially hydrolyzed cross-linker eliminates one of
This article is protected by copyright. All rights reserved.

10.1002/anie.201708273
Angewandte Chemie International Edition
COMMUNICATION
experiments, which can be explained by the flexibility of BSA
and the X-ray structure’s resolution. The distribution of Cα-Cα
distances is provided in Table S2, SI. These results suggest that
the cross-linking conditions do not disturb the conformation of
BSA. The low molar excess required for CDI cross-linking has a
crucial role in minimizing eventually detrimental effects on the
protein structure by excessive cross-linking. These results
encouraged us to further apply this reagent to structural
investigations of the tumor suppressor protein p53, an
intrinsically disordered protein (IDP), and GCAP-2.
Interestingly, none of the intermolecular cross-links identified
with DSBU[14] have been observed for CDI. We assume that the
relevant residues are sufficiently far apart to prevent CDI cross-
linking. This again testifies the valuable complementarity of the
MS-cleavable cross-linkers CDI and DSBU for structure analysis
of a large variety of proteins.
p53 cross-linking. The short spacer length of CDI may be
particularly beneficial for studying 3D-structures of IDPs. These
proteins, which comprise ~25% of mammalian proteomes, are
mainly involved in cell signaling and are characterized by a lack
of defined 3D-structures.[10] The gold standard for the structural
analysis of IDPs is currently NMR spectroscopy.[11] We
employed CDI to investigate the elusive structure of the full-
length, wild-type human tumor suppressor p53. Known as
“guardian of the genome”, human p53 consists of 393 amino
acids that are organized as biologically active homotetramer, for
which different structural concepts have been proposed. P53
cross-linking has been performed at pH 7.2 at 10°C in the
absence of DNA yielding 23 unique cross-links (see SI).
Interestingly, the cross-links were located not only in the C-
terminal region, harboring the majority of p53’s lysines, but also
in the tetramerization domain and the DNA-binding domain
(Figure S5, SI). The latter p53 domain has never been
Figure 1. Cα–Cα distances bridged by the CDI cross-linker are mapped into
p53 tetrameric model structure by Okorokov et al.[13]. Cα–Cα distances of
cross-linked residues are indicated as colored lines; green: distances <17 Å;
yellow: distances <27 Å; red: distances >27 Å.
connected by any of the cross-linkers employed previously[14]
.
Guanylyl cyclase-activating protein-2 (GCAP-2) cross-linking.
GCAP-2 is a Ca2+-binding protein that belongs to the neuronal
calcium sensor protein family.[15] GCAPs regulate guanylyl
cyclases in the retina cells and play a key role for light
adaptation.[16] N-terminally myristoylated, Ca2+-bound bovine
GCAP-2 was reacted with CDI yielding 19 unique cross-links
(see SI). They were compared with previously obtained BS2G
cross-links of GCAP-2[17], showing 8 identical cross-links (six
intramolecular and two intermolecular ones). Among the cross-
links of GCAP-2 monomer, three of them could be mapped into
the NMR structure of Ca2+-bound GCAP-2 (PDB ID: 1JBA,
resolution: 3.5 - 5.2 Å)[18]. The Cα-Cα distances measured were
in the range between 14.6 to 17.4 Å, which are in perfect
agreement with the distance CDI can bridge. The remaining
three cross-links, identified with both CDI and BS2G, involve K-
200 (or S-201) located in the C-terminal domain of GCAP-2 (S-
191 to F-204), which is structurally not resolved in the NMR
structure. The four amino acids involved, K-30, K-50, K-96, K-
126 (or K-128), cover a triangular surface on GCAP-2 (Figure 2).
Interestingly, this surface is located on the opposite side of the
three Ca2+-binding EF-hand motifs. The most C-terminal amino
acid resolved in the NMR structure is P-190 that is located
above this triangular surface. The cross-links between S-37 (or
T-39) and K-200 as well as between K-50 and K-200 (or S-201)
can be explained by a high flexibility of GCAP-2’s C-terminus.
However, the C-terminus of GCAP-2 was not found to be cross-
linked with amino acids in the EF-hand motifs of GCAP-2 that
are located on the opposite site of the molecule.
The identification of complementary cross-links in p53 is
probably mainly attributed to an improved reactivity of CDI with
hydroxyl groups compared to NHS ester cross-linkers[14]. This
underlines the great potential of CDI to shed light onto the
spatial arrangement of p53 and IDPs in general.
Unique CDI-cross-links were mapped onto two modeled
structures of the p53 tetramer, as proposed by Tidow et al.[12]
and by Okorokov et al.[13], to assess whether they are consistent
with one of p53 models. As mentioned above, the maximum Cα-
Cα distance to be bridged by CDI is ~16 Å, which however might
be considered as too stringent in this case.
The great majority of cross-links was found to be consistent with
the intramolecular Cα-Cα distances measured in the model
proposed by Okorokov et al.[13] These cross-links fall within the
range of 6.6 to 30 Å with an average of 19.3 Å and those
exceeding the maximum bridging distance of ~16 Å are located
in flexible loop regions[13]. In contrast, the SAXS-based model by
Tidow et al.[12] does not agree with our cross-linking data. Here,
average of Cα-Cα distances is 42.4 Å for intramolecular 65.5 Å
for intermolecular cross-links (Figure S6, SI).
Therefore, our CDI cross-linking results strongly support the
structural organization of the p53 tetramer as proposed by
Okorokov et al.[13], which presents a more compact arrangement
than the elongated cross-shaped structure introduced by Tidow
et al.[12]. This finding is also in good agreement with recent p53
cross-linking experiments, in which DSBU was employed.[14]
Compared to DSBU, CDI yields ~33% shorter geometrical
constraints and suggests a more compact structure of p53 as
Among the remaining
9 cross-links that were exclusively
perceived in the p53 model proposed by Tidow et al.[12]
.
identified with CDI only two are in agreement with the structure
This article is protected by copyright. All rights reserved.

10.1002/anie.201708273
Angewandte Chemie International Edition
COMMUNICATION
of Ca2+-bound GCAP-2. The remaining seven cross-links exceed
the CDI distance limit within the GCAP-2 monomer, but they can
be explained by the GCAP2 dimer.
Keywords: 1,1’-Carbonyldiimidazole (CDI) • cross-linking/mass
spectrometry (MS) • MS-cleavable cross-linker • p53 • GCAP- •
protein structure.
[1]
a) A. Sinz, Mass Spectrom. Rev. 2006, 25, 663-682; b) J. Rappsilber, J
Struct Biol 2011, 173, 530-540; c) A. Leitner, T. Walzthoeni, A.
Kahraman, F. Herzog, O. Rinner, M. Beck, R. Aebersold, Mol. Cell.
Proteomics 2010, 9, 1634-1649; d) C. Iacobucci, S. Reale, F. De Angelis,
Chembiochem 2013, 14, 181-183.
[2]
a) F. Liu, D. T. S. Rijkers, H. Post, A. J. R. Heck, Nat. Methods 2015, 12,
1179-1184; b) F. Liu, P. Lössl, R. Scheltema, R. Viner, A. J. R. Heck,
Nat. Commun 2017, 8; c) A. H. Kao, C. L. Chiu, D. Vellucci, Y. Y. Yang,
V. R. Patel, S. H. Guan, A. Randall, P. Baldi, S. D. Rychnovsky, L.
Huang, Mol. Cell. Proteomics 2011, 10; d) C. Arlt, M. Gotze, C. H. Ihling,
C. Hage, M. Schafer, A. Sinz, Anal. Chem. 2016, 88, 7930-7937; e) C.
Iacobucci, C. Hage, M. Schafer, A. Sinz, J. Am. Soc. Mass Spectr. 2017;
f) M. Q. Müller, F. Dreiocker, C. H. Ihling, M. Schäfer, A. Sinz, Anal.
Chem. 2010, 82, 6958-6968.
[3]
[4]
F. Bigi, R. Maggi, G. Sartori, Green Chem. 2000, 2, 140-148.
K. J. Padiya, S. Gavade, B. Kardile, M. Tiwari, S. Bajare, M. Mane, V.
Gaware, S. Varghese, D. Harel, S. Kurhade, Org. Lett. 2012, 14, 2814-
2817.
[5]
[6]
[7]
[8]
[9]
H. A. Staab, Angew. Chem. Int. Ed. 1962, 1, 351, Angew. Chem. 1962,
74, 407.
M. Götze, J. Pettelkau, R. Fritzsche, C. H. Ihling, M. Schäfer, A. Sinz, J.
Am. Soc. Mass Spectr. 2015, 26, 83-97.
C. L. Swaim, J. B. Smith, D. L. Smith, J. Am. Soc. Mass Spectr. 2004,
15, 736-749.
B. Schilling, R. H. Row, B. W. Gibson, X. Guo, M. M. Young, J. Am. Soc.
Mass Spectr. 2003, 14, 834-850.
C. Iacobucci, A. Sinz, Anal. Chem. 2017, 89, 7832-7835.
Figure 2. Front and side views of GCAP-2. The flexible C-terminal region of
GCAP-2 (S-191 to F-204) is not resolved in the NMR structure (PDB ID:
1JBA)[18]. The most C-terminal residue resolved in the NMR structure, P-190
(red), resides above a triangular surface (green dashed lines) that is formed by
the lysines K-30, K-96, K-126/128 (green) cross-linked with K-200 (or S201).
K50, S-37 (or T-39) (blue), were cross-linked to K-200 (or S-201) only by CDI.
The opposite site of GCAP-2 is defined by three Ca2+ ions (yellow spheres),
shown as black dashed line.
[10] a) A. C. Joerger, A. R. Fersht, Annu. Rev Biochem. 2016, 85, 375-404;
b) A. C. Joerger, A. R. Fersht, Annu. Rev Biochem. 2008, 77, 557-582.
[11] J. T. Nielsen, F. A. Mulder, Front. Mol. Biosci. 2016, 3, 4.
[12] H. Tidow, R. Melero, E. Mylonas, S. M. Freund, J. G. Grossmann, J. M.
Carazo, D. I. Svergun, M. Valle, A. R. Fersht, P. Natl Acad. Sci. USA
2007, 104, 12324-12329.
[13] a) A. L. Okorokov, M. B. Sherman, C. Plisson, V. Grinkevich, K.
Sigmundsson, G. Selivanova, J. Milner, E. V. Orlova, The EMBO J.
2006, 25, 5191-5200; b) R. Aramayo, M. B. Sherman, K. Brownless, R.
Lurz, A. L. Okorokov, E. V. Orlova, Nucleic Acids Res. 2011, 39, 8960-
8971.
[14] a) C. Arlt, V. Flegler, C. H. Ihling, M. Schäfer, I. Thondorf, A. Sinz,
Angew. Chem. Int. Ed. 2017, 56, 275-279; Angew. Chem. 2017, 129,
281-285; b) C. Arlt, C. H. Ihling, A. Sinz, Proteomics 2015, 15, 2746-
2755.
[15] W. A. Gorczyca, M. P. Graykeller, P. B. Detwiler, K. Palczewski, P. Natl
Acad. Sci. USA 1994, 91, 4014-4018.
Conclusion. Our ongoing search for novel cross-linking
principles led to the identification of the azolide CDI cross-linker.
CDI is the first “zero-length” cross-linker that is cleavable under
MS/MS conditions and delivers characteristic product ions that
facilitate the analysis of cross-linked products in protein
conformational studies. The cross-linking reaction can be
conducted at near-physiological pH conditions between 7.2 and
8. The short spacer length of CDI yields complementary
distance information compared to the so far mainly used NHS-
based cross-linkers. As CDI possesses similar reactivites
towards amine and hydroxyl groups in proteins, it allows
targeting additional regions in proteins that have so far been not
accessible to cross-linking with NHS esters. Also, sample
complexity is reduced due to the absence of “dead-end” (type 0)
cross-links. Other advantages of CDI are its low costs and its
straightforward application that will make the cross-linking/MS
approach available to a large number of laboratories having
access to mass spectrometers with MS/MS capabilities.
[16] K. W. Koch, D. Dell'Orco, ACS Chem. Neurosci. 2013, 4, 909-917.
[17] a) J. Pettelkau, I. Thondorf, S. Theisgen, H. Lilie, T. Schröder, C. Arlt, C.
H. Ihling, A. Sinz, J. Am. Soc. Mass Spectr. 2013, 24, 1969-1979; b) J.
Pettelkau, T. Schröder, C. H. Ihling, B. E. S. Olausson, K. Kölbel, C.
Lange, A. Sinz, Biochemistry 2012, 51, 4932-4949.
[18] J. B. Ames, A. M. Dizhoor, M. Ikura, K. Palczewski, L. Stryer, J. Biol.
Chem. 1999, 274, 19329-19337.
Acknowledgements
The authors would like to thank X. Wang for her contribution in
data analysis of BSA cross-links and D. Tänzler for GCAP-2
expression and purification. Dr. C. Ihling is acknowledged for his
valuable advice regarding LC/MS/MS analyses. A.S.
acknowledges financial support by the DFG (project Si 867/15-2).
C.I. is funded by the Alexander von Humboldt Foundation.
This article is protected by copyright. All rights reserved.

10.1002/anie.201708273
Angewandte Chemie International Edition
COMMUNICATION
COMMUNICATION
1,1’-carbonyldiimidazole (CDI) is the
shortest possible (~2.6 Å) urea-
based, MS cleavable cross-linker. It
reacts with amine and hydroxyl
groups in proteins at physiological
pH. CDI has been successfully
applied to cross-link BSA, the tumor
suppressor p53, and GCAP-2. It
allows deriving valuable short-
distance constraints in proteins,
which proved especially valuable for
Christoph Hage#, Claudio Iacobucci#,
Anne Rehkamp, Christian Arlt, and
Andrea Sinz*
Page No. – Page No.
probing
intrinsically
disordered
proteins (IDPs), such as p53.
This article is protected by copyright. All rights reserved.