Catalytic inactivation of human carbonic anhydrase I by a metallopeptide-sulfonamide conjugate is mediated by oxidation of active site residues

Article abstract of DOI:10.1021/ja0778038

Oxidation of active site residues (His and Trp), following catalytic inactivation of human carbonic anhydrase I by a copper-ATCUN conjugate of sulfanilimide, is evidenced by mass spectrometric analysis of tryptic and chymotryptic digests of the modified C

Full text of DOI:10.1021/ja0778038

Published on Web 02/02/2008
Catalytic Inactivation of Human Carbonic Anhydrase I by a
Metallopeptide-Sulfonamide Conjugate is Mediated by Oxidation of Active
Site Residues
Nikhil H. Gokhale, Seth Bradford, and J. A. Cowan*
EVans Laboratory of Chemistry, Ohio State UniVersity, 100 West 18th AVenue, Columbus, Ohio 43210
Received October 11, 2007; E-mail: cowan@chemistry.ohio-state.edu
Scheme 1. Synthetic Route to a Metallopeptide-Sulfonamide
Conjugate as a Catalytic Inactivator of Human Carbonic
Anhydrase Ia
Zinc-dependent metalloproteases are a well-recognized drug
target for treatment of a variety of disease states.1 Inhibitors are
often substrate-based, involving recognition of the active site
coupled with coordination to the zinc cofactor.2 Potent drug inhib-
itors typically show dissociation constants in the nanomolar or
submicromolar range.2 Recently, we have reported an alternative
strategy to the classical design of enzyme inhibitors,3-5 exploring
the development of catalytic drugs that are selective for a therapeutic
target and mediate the irreversible inactivation of that target. Such
an approach is complementary to traditional concepts underlying
drug development but has the potential to lower both dosage and
dose frequency.
Our prior reports of catalytic inactivation of ACE and ECE-I
described metallopeptide-mediated time-dependent inactivation in
the presence of a physiologically relevant co-reactant such as
ascorbate.3,4 The mechanism of catalytic inactivation presumably
involves metal-associated reactive oxygen species that promote
either protein cleavage3 and/or modification of active site residues.
Herein we report the results of experiments designed to elucidate
the mechanism of enzyme inactivation by metallopeptide-promoted
formation of reactive oxygen species. This work also establishes a
non-peptide moiety as a targeting agent when coupled to a copper-
ATCUN motif,8 using a metallopeptide-drug conjugate to catalyti-
cally inactivate the zinc-dependent enzyme, human carbonic
anhydrase I (CA-I).
a The coupling of sulfanilamide (SLN) to Z-GGH (Z-Gly-Gly-His,
where Z is carboxybenzyl) in the presence of EDC ((1-ethyl-3-[3-dimethyl
aminopropyl] carbodiimide hydrochloride) resulted in the N-terminally
protected ligand that upon hydrogenation yielded the zero-linker-length
bifunctional ligand GGHSLN. THE Cu2+ complex of GGHSLN demon-
strated the characteristic absorbance signature for GGH-bound Cu2+ (λmax
) 525 nm)6 as well as a Cu3+/2+ redox couple of 1.23 V (vs NHE).7
at the concentration used. Under oxidative conditions and subsatu-
rating concentrations of Cu-GGHSLN (5 µM), time-dependent
inactivation of CA-I activity was observed (Figure 1), yielding kobs
∼ 0.038 min-1 (k2 ∼ 7600 M-1 min-1). The initial activity of ∼76%
reflects the inhibitory influence of the complex prior to initiation
of irreversible chemistry in the active site. Within 3 h, CA-I activity
was eliminated. The residual 20% activity reflects the intrinsic
hydrolytic activity of the Cu-GGHSLN complex toward the
substrate. Inactivation by Cu2+(aq) under oxidative conditions was
considerably slower (kobs ∼ 0.003 min-1). No evidence of protein
cleavage was found by SDS-PAGE gel electrophoresis, but several
residues were found to be susceptible to oxidation following a
detailed mass spectrometric analysis (Supporting Information).
Following incubation of CA-I with Cu-GGHSLN for 3 h under
oxidative conditions, the Q-TOF2 mass spectrum yielded clear
evidence for oxidation of 1 or 2 residues per protein molecule. More
extensive oxidation (at least 12 residues) was observed with
Cu2+(aq), although the relatively slow inactivation chemistry
suggests that surface residues are the most probable target sites for
free copper ion. Control reactions with Cu-GGH and ascorbate
also did not show any evidence of protein oxidation, consistent
with a lack of binding by the latter (as noted earlier). Characteriza-
tion of oxidized residues was achieved by enzymatic digestion using
both chymotrypsin and trypsin, and the modified residues were
mapped by mass spectrometric analysis (nano LC/MS/MS) (Sup-
porting Information). Clear evidence for specific amino acid residue
modification in the proximity of the enzyme active site was obtained
(Supporting Information). The most prominent oxidation was
observed for select histidine residues (H40, H64, H67, H94, H96,
H103, H200, and H243). Prior reports indicate that histidine residues
Carbonic anhydrases (CAs, EC 4.2.1.1) carry out a number of
biosynthetic reactions including the reversible hydration of CO2 to
form bicarbonate.9 Possessing a (His)3-coordinated active site Zn2+
,
these enzymes are inhibited by sulfonamides,9-12 and numerous
enzyme-sulfonamide complexes have been crystallographically
characterized.11,12 The simplest of these sulfonamides, sulfanilamide
(Scheme 1), represents the active pharmacophore and the building
block for several generations of carbonic anhydrase inhibitors
(CA-Is).9 We have utilized a sulfanilamide derivative as a targeting
motif for CA-I in a Gly-Gly-His (GGH)-tagged conjugate
(GGHSLN) (Scheme 1) that directs a Cu2+-bound GGH complex
to mediate oxidative damage within the active site of the enzyme.13
The amino terminal copper/nickel binding motif (ATCUN) is
naturally found in serum albumins, demonstrating high affinity
toward Cu2+ and Ni2+ metal ions under physiological conditions.8
The metal is redox active in the Cu3+/2+ states, with high affinity
to the ATCUN ligand in both states. Accordingly, and as evidenced
elsewhere,4,6,14 the metal ion is not released during catalytic
turnover. The most likely active species is a Cu2+-associated
hydroxyl radical, or Cu3+dO (copper oxone species).3,4,14-17
Under hydrolytic conditions (no reducing agent), Cu-GGHSLN
inhibited CA-I (KI ) 5.00 µM) in a dose-dependent manner (relative
to KI ) 8.20 µM for the free sulfanilamide motif) and comparable
to the Cu2+-free ligand (GGHSLN, KI ) 5.20 µM). Neither
Cu2+(aq) nor Cu-GGH alone demonstrated significant inhibition
9
2388
J. AM. CHEM. SOC. 2008, 130, 2388-2389
10.1021/ja0778038 CCC: $40.75 © 2008 American Chemical Society

C O M M U N I C A T I O N S
His and (formyl)kynurenine formation] were calculated. These
distances are listed in Figure 2 under each residue label. Taken
together, the results showed two nonspecific long-range residue
oxidations (H40 and H103) that could be attributed to the diffusion
of metal-associated radicals or nonspecific surface reactions. Neither
residue lies close to a secondary inhibitor binding site that has been
structurally characterized.23 Overall, the oxidation of the amino acid
residues was restricted to the proximity of the active site (5-20
Å) where the inhibitor tail with the Cu-GGH motif would be
oriented. The pattern of oxidation also reveals a certain degree of
freedom to the bound inhibitor that can transfer metal-bound oxygen
species to neighboring residues. Localization of modified residues
around the active site supports the nondiffusive nature of the oxygen
species and association with the catalytic copper center. Three of
the modified residues sites (His64, His67, and His200) have been
implicated in rate-limiting proton transfer within the active site.24,25
Enzyme inactivation through targeted catalytic modification of
active site residues represents a novel mechanism for enzyme
inhibition. Such an approach is not only limited to the design of
catalytic inhibitors that, to the best of our knowledge, have not
been explored but also would give deeper insights on protein-
substrate interactions.
Figure 1. Human carbonic anhydrase I (CA-I) activity was measured at
the specified time intervals. Reactions were set up as individual reactions
(0.3 mL final volume) containing 1.1 µM CA-I, with zero time started at
the simultaneous incubation of all reactions with Cu-GGHSLN (5 µM)
and ascorbate (1 mM) in Tris buffer (12.5 mM, 75 mM NaCl, pH 8.0).
The inhibitor, Cu-GGHSLN, concentration was employed at the KI
concentration (5 µM). The activity at each time interval was determined
under initial velocity conditions with 1 mM substrate (4-nitrophenyl acetate).
Control reactions containing no inhibitor but only ascorbate were run
simultaneously. The initial velocity data were converted to % CA-I activity
with respect to control and plotted as a function of time. The figure shows
(O) control reaction with ascorbate only and (b) the time-dependent
inactivation in the presence of Cu-GGHSLN and ascorbate.
Acknowledgment. This work was supported by the National
Institutes of Health (GM063740). Mass spectrometry was carried
out by the OSU MS and Proteomics Facility.
Supporting Information Available: Synthetic procedures, enzyme
assays, UV-vis and electrochemical characterization of copper com-
plexes, mass spectrometric data, and control assays for Cu2+(aq). This
material is available free of charge via the Internet at http://pubs.acs.org.
References
(1) Hooper, N. M. Zinc Metalloprot. Health Dise. 1996, 1-21.
(2) Babine, R. E.; Bender, S. L. Chem. ReV. 1997, 97, 1359-1472.
(3) Gokhale, N. H.; Cowan, J. A. J. Biol. Inorg. Chem. 2006, 11, 937.
(4) Gokhale, N. H.; Cowan, J. A. Chem. Commun. 2005, 5916-5918.
(5) Jin, Y.; Cowan, J. A. J. Am. Chem. Soc. 2006, 128, 410-411.
(6) Jin, Y.; Cowan, J. A. J. Am. Chem. Soc. 2005, 127, 8408-8415.
(7) Jin, Y.; Lewis, M. A.; Gokhale, N. H.; Long, E. C.; Cowan, J. A. J. Am.
Chem. Soc. 2007, 129, 8353-8361.
(8) Harford, C.; Sarkar, B. Acc. Chem. Res. 1997, 30, 123-130.
(9) Supuran, C. T.; Scozzafava, A.; Casini, A. Med. Res. ReV. 2003, 23, 146-
189.
(10) Chakravarty, S.; Kannan, K. K. J. Mol. Biol. 1994, 243, 298-309.
(11) Di Fiore, A.; De Simone, G.; Menchise, V.; Pedone, C.; Casini, A.;
Scozzafava, A.; Supuran, C. T. Bioorg. Med. Chem. Lett. 2005, 15, 1937-
1942.
(12) Khadikar, P. V.; Sharma, V.; Karmarkar, S.; Supuran, C. T. Bioorg. Med.
Chem. Lett. 2005, 15, 923-930.
(13) Stadtman, E. R. Am. J. Clin. Nutr. 1991, 54, 1125S-1128S.
(14) Gokhale, N. H.; Bradford, S.; Cowan, J. A. J. Biol. Inorg. Chem. 2007,
12, 981-987.
Figure 2. Summary of residues that are susceptible to oxidative modifica-
tion (H40, H64, H67, H103, H200, H243, W97, W123) in human CA-I
(PDB: 1CZM) following interaction with Cu-GGHSLN in the presence
of ascorbate. A maximum of two residues are modified per protein. The
catalytic zinc ion is coordinated by three histidines that are not oxidized.
Overall, the figure shows modified amino acids to be localized in the
proximity of the enzyme active site (except H40 and H103). Distances from
the oxidized residues to the catalysts are listed under each label as detailed
in the main text.
(15) Bossu, F. P.; Margerum, D. W. J. Am. Chem. Soc. 1976, 98, 4003-4004.
(16) Margerum, D. W.; Owens, G. D. Met. Ions Biol. Syst. 1981, 12, 75-132.
(17) McDonald, M. R.; Fredericks, F. C.; Margerum, D. W. Inorg. Chem. 1997,
36, 3119-3124.
(18) Schoneich, C. J. Pharm. Biomed. Anal. 2000, 21, 1093-1097.
(19) Uchida, K.; Kawakishi, S. Arch. Biochem. Biophys. 1990, 283, 20-26.
(20) Hovorka, S. W.; Biesiada, H.; Williams, T. D.; Huhmer, A.; Schoneich,
C. Pharm. Res. 2002, 19, 530-537.
(21) Uchida, K.; Enomoto, N.; Itakura, K.; Kawakishi, S. Arch. Biochem.
Biophys. 1990, 279, 14-20.
(22) Griffiths, H. R. Free Radical Res. 2000, 33, S47-S58.
(23) Jude, K. M.; Banerjee, A. L.; Haldar, M. K.; Manokaran, S.; Roy, B.;
Mallik, S.; Srivastava, D. K.; Christianson, D. W. J. Am. Chem. Soc. 2006,
128, 3011-3018.
(24) Engstrand, C.; Jonsson, B.-H.; Lindskod, S. Eur. Biochem, J. 1995, 229,
696-702.
(25) Mohanty, A. K.; Kannan, K. K.; Mahajan, S. K. J. Biosci. 1998, 23,
235-246.
are effectively converted to the 2-oxo-histidine upon exposure to
Cu-ATCUN-derived oxygen-based radicals.18-20 Trp was also
found to be selectively oxidized (W97, W123).21,22 Significantly,
none of the zinc-bound histidine residues were oxidized (Supporting
Information), and so the time-dependent inhibition of CA-I by
Cu-GGHSLN does not involve a random oxidation pathway but
seems to be specific to residues that lie close to the Cu-GGH
domain of the metallopeptide-drug conjugate.
Figure 2 shows the location of oxidized residues with respect to
the active site and a crystallographically determined bound sul-
fonamide inhibitor.10 Using this structural data (PDB 1CZM),
distances to the oxidized residues from the terminal para-amino
nitrogen of the sulfanilamide inhibitor to the presumed site of
reaction on the His and Trp rings [C2, respectively, assuming 2-oxo-
JA0778038
9
J. AM. CHEM. SOC. VOL. 130, NO. 8, 2008 2389