4-Diazocyclohexa-2,5-dienone derivatives as photoaffinity labels for proteins: Scope and limitations

Article abstract of DOI:10.1055/s-1999-3102

4-Diazocyclohexa-2,5-dienone derivatives are presented as photosensitive reagents generating reactive carbene species which allow efficient photo- coupling reactions with proteins. However, during such coupling reactions, the photogenerated hydroquinones

Full text of DOI:10.1055/s-1999-3102

LETTER
889
4-Diazocyclohexa-2,5-dienone Derivatives as Photoaffinity Labels for
Proteins: Scope and Limitations
Christophe Colas, Thomas Grutter, Maurice Goeldner
Lab. de Chimie Bioorganique, UMR 7514 CNRS, Faculté de Pharmacie, Université L. Pasteur Strasbourg, BP 24, 67401 Illkirch Cedex
France
Fax +33(3)88678891; E-mail: goelder@bioorga.u-strasbg.fr
Received 1 February 1999
chemical conditions (RT solution photochemistry). Nu-
Abstract: 4-Diazocyclohexa-2,5-dienone derivatives are presented
merous publications refer to the photochemical properties
of diazocyclohexadienones, assessing the high reactivity
as photosensitive reagents generating reactive carbene species
which allow efficient photo-coupling reactions with proteins. How-
of the photogenerated carbenes i.e. insertion in C-H bonds
from benzene or cyclohexane4-6 or addition reactions to
unsaturated carbon carbon bonds7,8. Also, the lifetimes of
these carbenes have been determined, reaching
subnanoseconds9 and illustrative of the high reactivity of
these species. However most of these experiments were
rarely adapted to biological systems. Independently, we
tested the photochemical reactivity of diazocyclohexadi-
enones in isopropyl alcohol, a possible substitute for wa-
ter in a hydrophobic proteic environment and we noticed
that the singlet / triplet reactivity10 ( mainly O-H insertion
/ hydrodediazoniation) is directly dependent on the substi-
tution on the ring. Finally, the spectral properties of diaz-
ocyclohexadienones are excellent for photoaffinity
labeling investigations, having their lmax around 350 nm asso-
ciated to high extinction coefficients. In addition these
spectral properties allow singlet-singlet energy transfer
induced photoactivation using tryptophanes as energy do-
nors, leading to efficient and specific photoaffinity label-
ing experiments11.
ever, during such coupling reactions, the photogenerated hydro-
quinones can initiate undesired redox side reactions.
Key words: photochemistry, carbenes, hydroquinones, redox
reaction
Photoaffinity labeling is a classical method used either to
identify unknown functional proteins or to characterize, at
a molecular level, a protein binding site by identifying the
peptides and the amino acids constituting the site. Photo-
affinity labeling requires the synthesis of a photosensitive
ligand analog which when incubated with the biological
receptor, forms, after irradiation, a covalent bond within
the target binding site. This labeling reaction induces a
concomitant loss of the biological function which should
be prevented when the irradiation is achieved in the pres-
ence of an excess of a specific ligand (specificity of label-
ing). Finally, the use of radiolabeled probes allows to
undertake the analytical identification by characterizing
first the radiolabeled polypeptide(s), further the radiola-
beled peptides after proteolysis and finally the labeled
amino acids after microsequencing. Since the discovery of
this method by F. Westheimer and colleagues1, numerous
review articles2 have described in detail this methodology
including a monograph3 containing an extensive study on
the used chemical functions. We describe here, 4-diazo-
cyclohexa-2,5-dienone derivatives as photosensitive re-
agents of potential use in photoaffinity labeling.
The criteria for a satisfactory photoaffinity reagent are the
following: the probes should be chemically stable in the
incubation medium (neutral buffer), while photochemi-
cally, they should generate very reactive species, without
undergoing rearrangements to less reactive entities, and
which are able to react with any chemical function in the
binding site surrounding. Finally the irradiation condi-
tions used (wavelength, time) should not be harmful for
the proteic molecule. 4-Diazocyclohexa-2,5-dienones
possess all these properties: they are very stable in neutral
buffered medium (several days) they generate photo-
chemically the corresponding carbene which, by opposi-
tion to the 6-diazocyclohexa-2,4-dienonyl carbene, does
not undergo a Wolff rearrangement under usual photo-
Figure 1
Photoaffinity labels
Molecules (A – E) (Figure 1) represent different potential
photolabels containing the diazocyclohexadienonyl moi-
ety. Molecule A was designed to label the membrane-
associate domains of proteins,12 compound B was synthe-
sized as a channel blocker associate to the GABA
receptor13 compound C was described for a potential use
Synlett 1999, S1, 889–892 ISSN 0936-5214 © Thieme Stuttgart · New York

890
C. Colas et al.
LETTER
as a peptidic probe14, while molecules D and E containing of a competitive inhibitor (tacrine) used as a protector. Fi-
either a tertiary amine or a quaternary ammonium salt nally, the analysis of the variation of the inactivation rate
were synthesized to label cholinergic proteins, of the enzyme as a function of irradiation wavelength was
acetylcholinesterase15 and the nicotinic cholinergic in agreement with an inactivation process depending on
receptor16 respectively.
the photodecomposition of the diazocyclohexadienone.
All the above mentioned experiments possessed the nec-
essary criteria for a photosuicide inactivation.
Alternatively, compound F (Figure 2) which is a p-diazo-
nium ester has been synthesized as a potential photosui-
cide inactivator17 of cholinesterases, this molecule
generating after enzymatic ester hydrolysis 3,5-dimethyl-
4-diazocyclohexa-2,5-dienone which can be selectively
photoactivated at its absorption wavelength (~350 nm).
Figure 3
Compound D, a reversible inhibitor of acetylcholinest-
erase (Ki = 8 x 10-7M) gave satisfactory photoaffinity la-
beling results15, leading to an efficient inactivation of the
enzyme after photoirradiation at 350nm of an incubation
mixture of the protein with this inhibitor. This inactivation
was demonstrated to be specific since it could be prevent-
ed by using a protector such as edrophonium, a selective
ligand for the active site. However, besides this successful
example(s), it appeared that most labeling experiments
gave results which were disappointing. Indeed they lead
either to limited specificity in the labeling reaction (com-
pounds B and E). Even for molecule F, the inactivation ki-
netics showed the contribution of factors which are
different from the expected photosuicide process. Despite
their almost ideal photochemical properties for photoaf-
finity labeling reactions, diazocyclohexadienones gave
often poor results in photochemical coupling reactions
with proteins and it appeared very rapidly that the photo-
generated hydroquinones (reaction of the photogenerated
carbenes with water molecules, Figures 3 and 4) were re-
sponsible for this fact.
Photolabeling results
We selected the p-anisoylaryldiazonium ester F as a po-
tential photosuicide substrate for butyrylcholinesterase
(BuChE). This probe which is stable in buffer (half life
~24h) was a slow substrate of BuChE : Km =32mM and
Vm = 1.1 mmol/min/mg generating 3,5-dimethyl-4-diazo-
cyclohexa-2,5-dienone. The principle of photosuicide
substrates is to induce a selective photodecomposition of
the enzyme product, as soon it is formed in the active site.
This can be achieved by choosing an appropriate irradia-
tion wavelength. Irradiation of an incubation mixture of
the enzyme containing the substrate F, at a wavelength
~350 nm (the lmax of the diazoquinone), led to an effi-
cient enzyme inactivation (Figure 3). This inactivation
process was time- and substrate concentration-dependent.
Also, the enzyme inactivation could be prevented when
the irradiation was performed in the presence of an excess
Figure 4
Synlett 1999, S1, 889–892 ISSN 0936-5214 © Thieme Stuttgart · New York

LETTER
4-Diazocyclohexa-2,5-dienones as Photoaffinity Labels
891
Interference of the photo-products with the labeling Compound F: During the photosuicide inactivation pro-
reaction of proteins
cess described in the previous chapter, the only character-
istic which was puzzling, was an observed lag-time in the
inactivation kinetics and we were able to demonstrate that
the photochemically generated dimethylhydroquinone (F-
HQ) initiated a redox reaction with the diazonium mole-
cule F (Figure 5). A characteristic of this redox process
leading to the dediazoniation of the salt F, was the forma-
tion of several intermediate radical species18 which could
indeed react with the protein leading to its inactivation.
Therefore, despite an apparent well demonstrated photo-
suicide inactivation mechanism, the major part of the in-
activation process was due to a reaction intiated by the
photogenerated side product F-HQ. We could also show
an additional inhibitory effect of the redox generated
quinone F-Q, which emphasizes the complexity of the ob-
served inhibition.
Compound E : Using the tritiated molecule E to label the
acetylcholine binding site on the nicotinic acetylcholine
receptor, we noticed a radioactivity incorporation in the
expected molecular weight band on SDS polyacrylamide
electrophoresis gels but this incorporation could not be
satisfactorily prevented in the presence of excess of a spe-
cific ligand used as a protector. The labeling reaction oc-
curred on the protein outside the binding site (non specific
labeling). To explain this side-reaction, we postulated that
the photogenerated hydroquinone (E-HQ) was oxidized
in the incubation medium to the corresponding quinone
(E-Q) which could in turn react with free cysteines exist-
ing on the protein complex. To demonstrate this hypothe-
sis, we first verified that addition of exogenous tritiated E-
Q, in the absence of a photochemical reaction, induced ra-
dioactivity incorporation on the protein. Secondly, addi- In conclusion, the work presented here showed that, de-
tion of glutathione (GSH), a reactive cysteine-containing spite remarkable photochemical properties, diazocyclo-
tripeptide, to the incubation medium, reduced the non spe- hexadienones can be subjected to side reactions which
cific radioactivity incorporation, presumably by quench- might be deleterious to the desired site-directed photo-
ing the formed quinone E-Q (we were able to check alkylation reaction of a protein. In particular, the reductive
independently this reaction). Finally, by pretreatment of power of the photogenerated hydroquinones can generate
the receptor with N-phenyl maleimide (a thiol specific re- in the medium the corresponding chemically reactive
agent), non specific incorporation was reduced again. quinones (compound E) or induce a redox process with
These different control experiments agree with non spe- the starting compound as illustrated with molecule F.
cific reactions occuring between the formed E-Q mole-
cule with extra-site cysteines (Figure 4). To overcome the
non specific labeling problem, we selected reaction condi-
Ackowledgement
The support from the Centre National de la Recherche Scientifique,
the Ministère de l’Enseigenement Supérieur, The Association Fran-
çaise contre les Myopathies and the Région Alsace are gratefully
acknowledged.
tions which allowed to remove as rapidly as possible all
the photochemical side-products from the incubation me-
dium without altering the biochemical properties of the
protein. The use of a CeIV/GSH mixture allowed a very
efficient sequential oxidation / addition process: oxidation
of E-HQ to E-Q followed by a nucleophilic GSH addition References and Notes
reaction and permitted an efficient and very specific label-
ing reaction of the acetylcholine binding site on the nico-
tinic receptor.
(1) Singh, A.; Thornton, E.R.; Westheimer, F.H. J. Biol. Chem.
1962, 237, 3006.
(2) (a) Kotzyba-Hibert, F.; Kapfer, I.; Goeldner, M. Angew.
Chemie Int. Ed. Engl. 1995, 34, 1296 ; (b) Fleming, S.A.
Tetrahedron 1995, 51, 12479. (a) and (b) and references cited
therein.
(3) Bailey, H. In Photogenerated Reagents in Biochemistry and
Molecular Biology ; Work, T.S.; Burdon, R.H., Eds.; Elsevier:
Amsterdam, 1983.
(4) Dürr, H.; Kober, H. Tetrahedron Lett. 1972, 1259.
(5) Süs, O.; Möller, K.; Heiss, H. Ann., 1956, 598, 123.
(6) Alcaraz, M-L.; Peng, L.; Klotz, P.; Goeldner, M. J. Org.
Chem.. 1995, 61, 192.
(7) Pirkle, W.H.; Koser, G.F. Tetrahedron Lett. 1968, 3959.
(8) Fleming, J.C.; Shechter, H. J. Org. Chem. 1969, 34, 3962.
(9) Arnold, B.R.; Scaiano, J.C.; Bucher, G.F.; Sander, W.W. J.
Org. Chem. 1992, 57, 6469.
(10) Regitz, M.; Maas, G. In Diazo Compounds ; Regitz, M.; Maas,
G., Eds.; Academic Press, Inc., 1986.
(11) Goeldner, M.; Hirth, C. Proc. Natl. Acad. Sci. USA 1980, 77,
6439.
(12) Peng, L.; Alcaraz, M-L.; Klotz, P.; Kotzyba-Hibert, F.;
Goeldner, M. FEBS Lett. 1994, 346, 127.
Figure 5
(13) Dugave, C.; Kessler, P. Tetrahedron Lett. 1994, 35, 9557.
(14) Goeldner, M.; Hawkinson J.E.; Casida, J.E. Tetrahedron Lett.
1989, 30, 823.
Synlett 1999, S1, 889–892 ISSN 0936-5214 © Thieme Stuttgart · New York

892
C. Colas et al.
LETTER
(18) Brown, K.C.; Doyle, M.P. J. Org. Chem. 1988, 53, 3255.
(15) Kessler, P.; Ehret-Sabatier, L.; Goeldner, M.; Hirth, C.
Tetrahedron Lett. 1990, 31, 1275.
(16) Kotzyba-Hibert, F.; Kessler, P.; Zerbib, V.; Grutter, T.;
Bogen, C.; Takeda, K.; Hammadi, A.; Knerr, L.; Goeldner, M.
Bioconjugate Chem. 1997, 8, 472.
Article Identifier:
1437-2096,E;1999,0,S1,0889,0892,ftx,en;W04899ST.pdf
(17) Peng, L.; Colas, C.; Goeldner, M. Pure & Appl. Chem. 1997,
69, 755.
Synlett 1999, S1, 889–892 ISSN 0936-5214 © Thieme Stuttgart · New York