Development of a chemical probe for identifying protein targets of α-oxoaldehydes

Article abstract of DOI:10.1039/c3cc41099d

The development of a chemical probe for identifying the protein targets of reactive electrophilic α-oxoaldehydes such as methylglyoxal is presented. The probe is evaluated against methylglyoxal using human serum albumin as well as using living cells and lysates.

Full text of DOI:10.1039/c3cc41099d

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
Development of a chemical probe for identifying
protein targets of a-oxoaldehydes†
Christian Sibbersen,ab Johan Palmfeldt,c Jakob Hansen,b Niels Gregersen,c
Karl Anker Jørgensena and Mogens Johannsen*b
Cite this: Chem. Commun., 2013,
49, 4012
Received 8th February 2013,
Accepted 24th March 2013
DOI: 10.1039/c3cc41099d
www.rsc.org/chemcomm
The development of a chemical probe for identifying the protein
targets of reactive electrophilic a-oxoaldehydes such as methyl-
glyoxal is presented. The probe is evaluated against methylglyoxal
using human serum albumin as well as using living cells and lysates.
a-Oxoaldehydes (Fig. 1) are toxic metabolites that are formed by
the non-enzymatic degradation of glycolytic intermediates. They
can react with nucleophilic amino acids, primarily arginine, to
form advanced glycation end-products (AGEs). These modifica-
tions are known to disrupt native protein function and have been
implicated in several aging-related disease states such as dia-
betes mellitus, neurodegenerative diseases and cardiovascular
Fig. 2 Methylglyoxal-derived AGEs.
Depending on the local environment and accessibility, some
disease.1 Methylglyoxal is the most important and well-studied amino acid residues are more sensitive to modification by
of the a-oxoaldehydes, though glyoxal and 3-deoxyglucosone are a-oxoaldehydes than others. Collectively, the subset of modified
also physiologically relevant.2 Methylglyoxal has been shown proteins was coined the ‘‘dicarbonyl proteome’’.5 Several proteins
to be a major source of AGEs in endothelial cells and this have been shown to be modified by methylglyoxal in biological
modification of cellular proteins has been implicated in the systems, e.g. histone H1, proteasome subunit 20S and HIF-1a
development of vascular complications in diabetes mellitus.3
co-activator protein p300.6 AGE formation is a non-enzymatic
The arginine-derived hydroimidazolone species MG–H1 is the process, so methylglyoxal will likely have a broad range of
most prevalent AGE formed by methylglyoxal and typically biological targets. Profiling of these targets followed by biochem-
accounts for more than 90% of the modifications.4 Minor ical studies will be essential to understand the impact of methyl-
adducts of methylglyoxal include the fluorescent arginine- glyoxal modification of proteins in vivo.7 In an attempt to obtain a
derived argpyrimidine, the lysine-derived Nd-carboxyethyl lysine broader coverage of proteins prone to modification in vivo, Ahmed
(CEL) and several arginine and lysine-derived crosslinks (Fig. 2). et al. have recently identified 15 different methylglyoxal modified
proteins in senescent fibroblasts using immunodetection with
polyclonal antibodies in combination with two-dimensional gel
electrophoresis and mass spectrometry.8 However, there is still a
lack of methods and tools to study this subproteome in detail,
which led us to develop a chemical probe for mapping cellular
targets of a-oxoaldehyde modification. We envisioned that by
employing a compound with reactivity similar to that of endo-
Fig. 1 Endogenous a-oxoaldehydes.
genous a-oxoaldehydes, the proteins most susceptible to modifi-
cations might be selectively labelled with a chemical tag for
subsequent identification. Chemical probes have been used to
identify targets of both enzymatic and non-enzymatic post-
translational modifications in cells e.g. for targets of lipidation
and lipid-derived electrophilic stress.9 However, to the best of our
knowledge no probes have been developed to elucidate the
a Department of Chemistry, Aarhus University, Langelandsgade 140, 8000 Aarhus C,
Denmark. E-mail: kaj@chem.au.dk
b Department of Forensic Medicine, Bioanalytical Unit, Aarhus University,
Brendstrupgaardsvej 100, 8200 Aarhus N, Denmark. E-mail: mj@forensic.au.dk;
Tel: +45 87168332
c Research Unit for Molecular Medicine, Institute of Clinical Medicine,
Aarhus University Hospital, Brendstrupgaardsvej 100, 8200 Aarhus N, Denmark.
E-mail: nig@ki.au.dk
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c3cc41099d equally important carbohydrate-derived dicarbonyl stress.10
c
4012 Chem. Commun., 2013, 49, 4012--4014
This journal is The Royal Society of Chemistry 2013

View Article Online
Communication
ChemComm
methylglyoxal, we performed an experiment under similar condi-
tions. This led to the formation of the MG–H1 (Fig. 2) analogue
AlkMG–H1 5 (Fig. 3), which could be characterized.
Human serum albumin (HSA) is a well-known target of
methylglyoxal and its modifications have been described exten-
sively in the literature,15 and therefore this protein was chosen
for the initial investigation. Methylglyoxal and AlkMG 4 were
both reacted with isolated HSA for 24 h at 37 1C to form AGEs.
The protein samples were then digested with trypsin and the
resulting peptides were analysed by Orbitrap LC-MS for high
mass accuracy detection of peptides with an additional mass
corresponding to one of the aforementioned modifications.
Several arginine residues were found to be extensively labelled
by both methylglyoxal and AlkMG 4 at levels of 100 mM and with
both precursors only the MG–H1-type AGE species was detected.
For methylglyoxal, 13 peptides were found to be modified on an
Fig. 3 Synthesis of the chemical probe and its primary arginine adduct.
We have now devised a probe that includes an a-oxoaldehyde moiety arginine residue. These were also found to be modified in the
to react with nucleophilic amino acid residues, as well as an experiment with AlkMG 4, along with two additional peptides
alkyne tag for subsequent fluorescent labelling or affinity chro- (Fig. 4). These findings confirm that the probe reacts with
matography.11 In this communication we describe the synthesis, arginine residues under physiological conditions and with a
validation and biological evaluation of this chemical probe.
The probe AlkMG 4 was synthesized starting from commercially making it useful for detecting targets of dicarbonyl stress.
available precursors as shown in Fig. 3. First, the Weinreb amide 1, Next we set out to explore a-oxoaldehyde modifications in
reactivity and selectivity that resembles that of methylglyoxal,
formed from ethyl diethoxyacetate, undergoes a Grignard reaction cell lysates. We synthesized a fluorescent rhodamine-azide
with the magnesium bromide alkyne species, leading to 2. TMS- (Fig. 5) tag using the Ugi reaction as described by Brauch
deprotection of 2 provides the acetal-protected probe 3. Since et al.16 This tag can be attached to proteins labelled by AlkMG
a-oxoaldehydes are unstable the removal of the acetal protecting 4 using the Sharpless–Meldal copper-catalyzed alkyne azide
group in the final step must be carried out immediately before cycloaddition,17 and when conjugated to these proteins it
applying the probe. AlkMG 4 cannot be characterized unambigu- allows for direct visualization of the labelled proteome by
ously using NMR spectroscopy due to acetal formation and it in-gel fluorescence. Cultured HEK293 cells were lysed and
decomposes when subjected to mass spectrometry.12 In order to treated with AlkMG 4 at 37 1C with varying probe concentra-
demonstrate its formation we initially estimated the yield of the tions and incubation time. The proteins were then precipitated
deprotection by oxidizing 4 to organic acids and titrating with a
base as described by Friedemann.13 This experiment indicated
quantitative conversion of 3 to 4. Additionally, we performed the
deprotection in D2O and subsequent NMR analysis showed that
AlkMG 4 exists in solution as a 1 : 0.3 mixture of the mono- and
dihydrate. Methylglyoxal has been shown to react primarily with
arginine residues in vivo, and Hellwig et al. have demonstrated that
this reaction also takes place with free arginine on a mmol scale.14
Fig. 5 Fluorescent rhodamine-azide tag.
To test if
4 would form adducts analogous to those of
Fig. 4 Mass spectrometric identification of adducts formed from the reaction of AlkMG or methylglyoxal with human serum albumin (HSA). Arginine residues
modified by both AlkMG and methylglyoxal are highlighted in green, whereas blue highlighting shows residues modified only by the AlkMG probe.
c
This journal is The Royal Society of Chemistry 2013
Chem. Commun., 2013, 49, 4012--4014 4013

View Article Online
ChemComm
Communication
most prone to modification by a-oxoaldehydes and thereby shed
light on some of the mechanisms by which these reactive meta-
bolites affect the disease states they are implicated in. We are
currently pursuing this direction of research.
We acknowledge The Danish Council for Independent
Research, Medical Sciences and The John and Birthe Meyer
Foundation for financial support.
Notes and references
1 (a) P. J. Thornalley, Amino Acids, 1994, 6, 15–23; (b) M. Brownlee,
Nature, 2001, 414, 813–820; (c) J. Li, D. Liu, L. Sun, Y. Lu and
Z. Zhang, J. Neurol. Sci., 2012, 317, 1–5; (d) Z. Hegab, World J.
Cardiol., 2012, 4, 90–102; (e) T. H. Fleming, P. M. Humpert,
P. P. Nawroth and A. Bierhaus, Gerontology, 2011, 57, 435–443.
2 P. J. Thornalley, A. Langborg and H. S. Minhas, Biochem. J., 1999,
344, 109–116.
3 (a) M. Shinohara, P. J. Thornalley, I. Giardino, P. Beisswenger,
S. R. Thorpe, J. Onorato and M. Brownlee, J. Clin. Invest., 1998,
101, 1142–1147; (b) See also ref. 1b.
Fig. 6 HEK293 lysates were incubated with varying levels of AlkMG 4 for 1 h,
then reacted with rhodamine-azide 6 and analysed by SDS-PAGE and in-gel
fluorescence followed by Coomassie staining.
4 N. Rabbani and P. J. Thornalley, Amino Acids, 2012, 42, 1133–1142.
5 N. Rabbani and P. J. Thornalley, Ann. N. Y. Acad. Sci., 2008, 1126,
124–127.
and reacted with rhodamine-azide 6. The lysates, now contain-
ing fluorescently labelled proteins, were loaded onto an SDS-
PAGE gel and subsequently analysed for fluorescence. The gels
were then stained with Coomassie brilliant blue to verify that
protein levels were comparable in each lane (Fig. 6).
6 (a) S. Pashikanti, G. A. Boissonneault and D. Cervantes-Laurean, Free
Radical Biol. Med., 2011, 50, 1410–1416; (b) M. A. Queisser, D. Yao,
S. Geisler, H.-P. Hammes, G. Lochnit, E. D. Schleicher, M. Brownlee
and K. T. Preissner, Diabetes, 2010, 59, 670–678; (c) H. Thangarajah,
D. Yao, E. I. Chang, Y. Shi, L. Jazayeri, I. N. Vial, R. D. Galiano,
X.-L. Du, R. Grogan, M. G. Galvez, M. Januszyk, M. Brownlee and
G. C. Gurtner, Proc. Natl. Acad. Sci. U. S. A., 2009, 106, 13505–13510.
7 See e.g.: (a) A. Bierhaus, T. Fleming, S. Stoyanov, A. Leffler, A. Babes,
It was observed that labelling of proteins by AlkMG 4
increased with probe concentration. Furthermore, the degree
of labelling increased over several hours (ESI,† Fig. S2). This was
initially surprising because methylglyoxal is known to be very
reactive. However, it has been demonstrated that a large fraction
of the methylglyoxal in biological systems exists in a reversibly
bound pool that is in equilibrium with e.g. free cysteine.18 The
reversibly bound methylglyoxal is slowly released over time,
which would explain our observations. We investigated whether
some of the binding sites could be blocked by pre-incubating the
cell lysates with methylglyoxal (ESI,† Fig. S3). This was success-
ful, but required high levels of methylglyoxal, indicating that
there are numerous potential binding sites in the proteome.
In order to investigate whether the a-oxoaldehyde moiety is
specifically responsible for the high reactivity of the probe, we
performed a comparative control experiment with AlkMG 4,
compound 3, hex-5-ynal and hex-5-yn-2-one in cell lysates. This
was confirmed as only low levels of labelling were observed for
the other compounds (ESI,† Fig. S4). Finally we investigated the
ability of the probe to cross the cell membrane. Living HEK-293
cells were incubated with various concentrations of AlkMG 4
¨
C. Neacsu, S. K. Sauer, M. Eberhardt, M. Schnolzer, F. Lasitschka,
F. Lasischka, W. L. Neuhuber, T. I. Kichko, I. Konrade, R. Elvert,
W. Mier, V. Pirags, I. K. Lukic, M. Morcos, T. Dehmer, N. Rabbani,
P. J. Thornalley, D. Edelstein, C. Nau, J. Forbes, P. M. Humpert,
M. Schwaninger, D. Ziegler, D. M. Stern, M. E. Cooper,
U. Haberkorn, M. Brownlee, P. W. Reeh and P. P. Nawroth, Nat.
Med., 2012, 18, 926–933; (b) M. G. Distler, L. D. Plant, G. Sokoloff,
A. J. Hawk, I. Aneas, G. E. Wuenschell, J. Termini, S. C. Meredith,
M. A. Nobrega and A. A. Palmer, J. Clin. Invest., 2012, 122,
2306–2315; (c) C. F. Bento, R. Fernandes, J. Ramalho, C. Marques,
F. Shang, A. Taylor and P. Pereira, PLoS One, 2010, 5, e15062;
(d) D. Dobler, N. Ahmed, L. Song, K. E. Eboigbodin and
P. J. Thornalley, Diabetes, 2006, 55, 1961–1969.
8 E. K. Ahmed, A. Rogowska-Wrzesinska, P. Roepstorff, A.-L. Bulteau
and B. Friguet, Aging Cell, 2010, 9, 252–272.
9 See e.g.: (a) R. N. Hannoush and J. Sun, Nat. Chem. Biol., 2010, 6, 498–506;
(b) A. T. Jacobs and L. J. Marnett, Acc. Chem. Res., 2010, 43, 673–683; For
other carbon electrophile probes see e.g. also: (c) E. Weerapana,
G. M. Simon and B. F. Cravatt, Nat. Chem. Biol., 2008, 4, 405–407.
10 Recently a phenylglyoxal-based probe was developed for visualiza-
tion of protein citrullination: K. L. Bicker, V. Subramanian,
A. A. Chumanevich, L. J. Hofseth and P. R. Thompson, J. Am. Chem.
Soc., 2012, 134, 17015–17018.
11 A. E. Speers, G. C. Adam and B. F. Cravatt, J. Am. Chem. Soc., 2003,
125, 4686–4687.
´
´
´
and washed several times to ensure that extracellular proteins 12 I. Nemet, D. Vikic-Topic and L. Varga-Defterdarovic, Bioorg. Chem.,
and excess probe were removed before cell lysis and reaction
with rhodamine-azide 6. This experiment showed a low level of
2004, 32, 560–570.
13 T. E. Friedemann, J. Biol. Chem., 1927, 73, 331–334.
14 M. Hellwig, S. Geissler, R. Matthes, A. Peto, C. Silow, M. Brandsch
labelling from 0.1 to 1 mM; however, at 5 mM extensive
labelling was observed (ESI,† Fig. S5). This verifies that the
probe is able to permeate cell membranes and label intracel-
lular proteins, and is thus useful for experiments in living cells.
The present work demonstrates that the chemical probe AlkMG
reacts with arginine residues with a reactivity and selectivity that
and T. Henle, ChemBioChem, 2011, 12, 1270–1279.
15 See for example: (a) N. Ahmed and P. J. Thornalley, Biochem. J., 2002,
364, 15–24; (b) N. Ahmed, D. Dobler, M. Dean and P. J. Thornalley,
J. Biol. Chem., 2005, 280, 5724–5732; (c) M. J. Kimzey, H. N. Yassine,
B. M. Riepel, G. Tsaprailis, T. J. Monks and S. S. Lau, Chem.–Biol.
Interact., 2011, 192, 122–128.
16 S. Brauch, M. Henze, B. Osswald, K. Naumann, L. A. Wessjohann, S. S.
van Berkel and B. Westermann, Org. Biomol. Chem., 2012, 10, 958–965.
resembles that of methylglyoxal. The probe has been applied 17 (a) V. V. Rostovtsev, L. G. Green, V. V. Fokin and K. B. Sharpless,
Angew. Chem., Int. Ed., 2002, 41, 2596–2599; (b) C. W. Tornøe,
C. Christensen and M. Meldal, J. Org. Chem., 2002, 67, 3057–3064.
18 A. Dhar, K. Desai, J. Liu and L. Wu, J. Chromatogr., B: Anal. Technol.
in vitro to modify proteins sensitive to a-oxoaldehydes and visua-
lize them using rhodamine-azide and in-gel fluorescence. We
believe that this new chemical probe can help identify the proteins
Biomed. Life Sci., 2009, 877, 1093–1100.
c
4014 Chem. Commun., 2013, 49, 4012--4014
This journal is The Royal Society of Chemistry 2013