Inhibition of peptidylglycine α-amidating monooxygenase by exploitation of factors affecting the stability and ease of formation of glycyl radicals

Article abstract of DOI:10.1021/ja046204n

Peptidylglycine α-amidating monooxygenase catalyzes the biosynthesis of peptide hormones through radical cleavage of the C-terminal glycine residues of the corresponding prohormones. We have correlated ab initio calculations of radical stabilization energies and studies of free radical brominations with the extent of catalysis displayed by peptidylglycine α-amidating monooxygenase, to identify classes of inhibitors of the enzyme. In particular we find that, in closely related systems, the substitution of glycolate for glycine reduces the calculated radical stabilization energy by 34.7 kJ mol ^-1, decreases the rate of bromination with N-bromosuccinimide at reflux in carbon tetrachloride by a factor of at least 2000, and stops catalysis by the monooxygenase, while maintaining binding to the enzyme.

Full text of DOI:10.1021/ja046204n

Published on Web 09/23/2004
Inhibition of Peptidylglycine r-Amidating Monooxygenase by
Exploitation of Factors Affecting the Stability and Ease of
Formation of Glycyl Radicals
Brendon J. W. Barratt,† Christopher J. Easton,*,† David J. Henry,† Iris H. W. Li,†
Leo Radom,†,‡ and Jamie S. Simpson†
Contribution from the Research School of Chemistry, Australian National UniVersity,
Canberra, ACT 0200, Australia, and School of Chemistry, UniVersity of Sydney,
Sydney, NSW 2006, Australia
Received June 27, 2004; E-mail: easton@rsc.anu.edu.au
Abstract: Peptidylglycine R-amidating monooxygenase catalyzes the biosynthesis of peptide hormones
through radical cleavage of the C-terminal glycine residues of the corresponding prohormones. We have
correlated ab initio calculations of radical stabilization energies and studies of free radical brominations
with the extent of catalysis displayed by peptidylglycine R-amidating monooxygenase, to identify classes
of inhibitors of the enzyme. In particular we find that, in closely related systems, the substitution of glycolate
for glycine reduces the calculated radical stabilization energy by 34.7 kJ mol-1, decreases the rate of
bromination with N-bromosuccinimide at reflux in carbon tetrachloride by a factor of at least 2000, and
stops catalysis by the monooxygenase, while maintaining binding to the enzyme.
Scheme 1
Introduction
Peptidylglycine R-amidating monooxygenase (PAM) cata-
lyzes the biosynthesis of a wide variety of C-terminal peptide
amides through oxidative cleavage of the corresponding glycine-
extended precursors (Scheme 1). The products include mam-
malian peptide hormones,1 as well as physiologically active
peptides of other organisms such as insects2,3 and cnidarians,4,5
and the C-terminal amide moiety is vital to the activity of many
of these compounds. Amidated peptide hormones are important
in cellular communication, in particular as neuropeptides,1 and
are implicated in a broad range of pathological conditions,
including asthma,6 inflammation,7 and cancers.8-10 The bio-
synthesis of amides from glycine-extended precursors other than
peptides, such as fatty acids,11 bile acids,12 nicotinic acid,13 and
aspirin,14 has also been attributed to PAM. Fatty acid amides
are known to affect a number of neurochemical communication
pathways, including sleep regulation,15,16 while the amidation
of aspirin may be important in its metabolic processing.
The importance of PAM in pathological conditions has led
to interest in its regulation and the development of a number of
inhibitors. One of the first was trans-4-phenylbut-3-enoic acid,17
which is effective in vivo in reducing serum PAM activity18 as
well as showing anti-inflammatory and analgesic effects.19
† Australian National University.
‡ University of Sydney.
(1) Merkler, D. J. Enzyme Microb. Technol. 1994, 16, 450-456.
(2) Konopinska, D.; Rosinski, G.; Sobotka, W. Int. J. Pept. Protein Res. 1992,
39, 1-11.
(3) Schoofs, L.; Veelaert, D.; VandenBroeck, J.; DeLoof, A. Peptides 1997,
18, 145-156.
(4) Hauser, F.; Williamson, M.; Grimmelikhuijzen, C. J. P. Biochem. Biophys.
Res. Commun. 1997, 241, 509-512.
(12) King, L.; Barnes, S.; Glufke, U.; Henz, M. E.; Kirk, M.; Merkler, K. A.;
Vederas, J. C.; Wilcox, B. J.; Merkler, D. J. Arch. Biochem. Biophys. 2000,
374, 107-117.
(5) Williamson, M.; Hauser, F.; Grimmelikhuijzen, C. J. P. Biochem. Biophys.
Res. Commun. 2000, 277, 7-12.
(6) Groneberg, D. A.; Springer, J.; Fischer, A. Pulm. Pharmacol. Ther. 2001,
14, 391-401.
(13) Merkler, D. J.; Glufke, U.; Ritenour-Rodgers, K. J.; Baumgart, L. E.;
DeBlassio, J. L.; Merkler, K. A.; Vederas, J. C. J. Am. Chem. Soc. 1999,
121, 4904-4905.
(7) Wiesenfeld-Hallin, Z.; Xu, X. J. Eur. J. Pharmacol. 2001, 429, 49-59.
(8) Schally, A. V.; Comaru-Schally, A. M.; Nagy, A.; Kovacs, M.; Szepeshazi,
K.; Plonowski, A.; Varga, J. L.; Halmos, G. Front. Neuroendocrinol. 2001,
22, 248-291.
(14) DeBlassio, J. L.; deLong, M. A.; Glufke, U.; Kulathila, R.; Merkler, K.
A.; Vederas, J. C.; Merkler, D. J. Arch. Biochem. Biophys. 2000, 383, 46-
55.
(9) Du, J. L.; Keegan, B. P.; North, W. G. Cancer Lett. 2001, 165, 211-218.
(10) Jimenez, N.; Jongsma, J.; Calvo, A.; van der Kwast, T. H.; Treston, A.
M.; Cuttitta, F.; Schroder, F. H.; Montuenga, L. M.; van Steenbrugge, G.
J. Int. J. Cancer 2001, 94, 28-34.
(15) Cravatt, B. F.; Lerner, R. A.; Boger, D. L. J. Am. Chem. Soc. 1996, 118,
580-590.
(16) Cravatt, B. F.; Prosperogarcia, O.; Siuzdak, G.; Gilula, N. B.; Henriksen,
S. J.; Boger, D. L.; Lerner, R. A. Science 1995, 268, 1506-1509.
(17) Bradbury, A. F.; Mistry, J.; Roos, B. A.; Smyth, D. G. Eur. J. Biochem.
1990, 189, 363-368.
(11) Wilcox, B. J.; Ritenour-Rodgers, K. J.; Asser, A. S.; Baumgart, L. E.;
Baumgart, M. A.; Boger, D. L.; DeBlassio, J. L.; deLong, M. A.; Glufke,
U.; Henz, M. E.; King, L.; Merkler, K. A.; Patterson, J. E.; Robleski, J. J.;
Vederas, J. C.; Merkler, D. J. Biochemistry 1999, 38, 3235-3245.
(18) Mueller, G. P.; Driscoll, W. J.; Eipper, B. A. J. Pharmacol. Exp. Ther.
1999, 290, 1331-1336.
9
13306
J. AM. CHEM. SOC. 2004, 126, 13306-13311
10.1021/ja046204n CCC: $27.50 © 2004 American Chemical Society

Inhibition of Peptidylglycine
R
-Amidating Monooxygenase
A R T I C L E S
Others include R,â-unsaturated acids,20-22 a peptide terminating
in R-vinylglycine,23 and diastereomers of a peptide terminating
in an R-styrylglycine.24 These are all mechanism-based inhibitors
in that they show turnover-dependent inactivation of the enzyme.
Inhibitors of other types have also been reported, such as
inorganic sulfite,25 benzyl hydrazine,26 and N-formylamides,27
as well as derivatives of â-mercaptostyrene28 and homocys-
teine.29
to identify and characterize the precatalytic complex of PHM
with copper, oxygen, and the substrate. That study was the first
to delineate the role of copper in the activation of dioxygen in
this or any other enzyme system. The authors also drew a
correlation between the reactivity of PHM substrates and the
stability of the corresponding radical intermediates. They
reported that our earlier theoretical studies37 showed a peptide
R-carbon-centered alanyl radical to be 9.1 kJ mol-1 less stable
than a corresponding glycyl radical and noted that, even so,
N-acetyl-(S)-tryptophanyl-(R)-alanine is still processed by the
enzyme during X-ray diffraction. They also predicted, on the
basis of our calculations with related amino acids, that an
R-carbon-centered threonyl radical would be even less stable
than an alanyl radical, and accordingly found that a peptide
containing (R)-threonine instead of (R)-alanine at the C-terminus
was less effectively turned over by the enzyme. Our calculations
of glycyl and alanyl radical stability37 were based on the radicals
3c and 5d (Chart 1). Contrary to the above discussion, their
relative radical stabilization energies (RSEs), which correspond
to the negative of the relative bond dissociation energies of the
corresponding closed-shell molecules, actually showed the alanyl
radical 5d to be less stable than the glycyl radical 3c by only
1.6 kJ mol-1. We have not studied threonine derivatives, but
Rauk et al.39 used two different methods to calculate that an
R-carbon-centered threonyl radical is destabilized relative to a
glycyl radical by only 7-14 kJ mol-1. Consequently, the
destabilization of alanyl and threonyl radicals appears to have
been somewhat over-estimated by Prigge et al.,32 with the result
that the relationship between the stabilization energies of radicals
and the ease of their formation through PAM catalysis warrants
further investigation. We also include in the present study a
further comparison with the relative rates of formation of
radicals in conventional brominations.
PAM consists of two functional subunits, peptidylglycine
R-hydroxylating monooxygenase (PHM, E.C. 1.14.17.3) and
peptidylamidoglycolate lyase (PAL, E.C. 4.3.2.5.) (Scheme 1).
PHM catalyzes the copper-, molecular oxygen-, and ascorbate-
dependent hydroxylation of a C-terminal glycine residue of a
peptide substrate. The product hydroxyglycine is then hydro-
lyzed to the corresponding amide and glyoxylate, a process that
is catalyzed by PAL at physiological pH. The determination of
the crystal structure of PHM in both reduced and oxidized
forms,30-32 and kinetic33-35 and mutagenesis studies,36 have
resulted in a detailed picture of the mechanism of action of this
enzyme. In particular, it has been concluded that, in the first
step, a copper-bound superoxide radical abstracts the pro-S
hydrogen from the glycine residue, to give a glycyl radical.
In the present work we have sought to exploit factors affecting
the formation of such radicals in order to design analogues of
the substrates of PAM that competitively bind to, but are not
processed by, the enzyme and therefore inhibit reaction of the
substrates. To this end, we have compared the results of ab initio
calculations and studies of relative reaction rates in free radical
brominations, which identify factors affecting the stability and
ease of formation of glycyl and related radicals,37,38 with the
kinetic parameters defining the interactions of analogous
compounds with PAM.
Publication of our results is very timely in the light of a quite
recent paper by Prigge et al.,32 in which crystallography of
frozen protein soaked with a slowly reacting substrate was used
Results
The natural substrates of PAM all have in common an
N-acylated glycine, and they are therefore comprised of acyl,
amido, methylene, and carboxyl groups. Since the carboxyl
group of a substrate is known to be important for binding to
PAM,31 we have not examined alternatives to this moiety in
the present study. However, the effect of modifying each of
the other groups was explored. A range of N-acyl substituents
was investigated, since these are known to affect the stability
and ease of formation of glycyl radicals38,40 and to be tolerated
by PAM. Although derivatives of R-substituted amino acids tend
not to bind to PAM, small R-alkyl substituents, such as the
methyl group of (R)-alanine,32,41 the vinyl moiety of (R)-
vinylglycine,42 and the hydroxyethyl group of (R)-threonine,32
are accommodated. The incorporation of a trifluoromethyl group
was therefore studied since â,â,â-trifluoroalanine derivatives
are known to be resistant to R-carbon-centered radical forma-
tion.37 The effect of replacing the amido group with an ester or
ketone using derivatives of glycolate or γ-keto acids instead of
N-acylglycines was also explored. Glycolate inhibitors of PAM
have been previously reported22,42 in studies of broad ranges of
(19) Ogonowski, A. A.; May, S. W.; Moore, A. B.; Barrett, L. T.; O’Bryant, C.
L.; Pollock, S. H. J. Pharmacol. Exp. Ther. 1997, 280, 846-853.
(20) Feng, J.; Shi, J.; Sirimanne, S. R.; Mounier-Lee, C. E.; May, S. W. Biochem.
J. 2000, 350, 521-530.
(21) Moore, A. B.; May, S. W. Biochem. J. 1999, 341, 33-40.
(22) Katopodis, A. G.; May, S. W. Biochemistry 1990, 29, 4541-4548.
(23) Zabriskie, T. M.; Cheng, H. M.; Vederas, J. C. J. Am. Chem. Soc. 1992,
114, 2270-2272.
(24) Andrews, M. D.; O’Callaghan, K. A.; Vederas, J. C. Tetrahedron 1997,
53, 8295-8306.
(25) Merkler, D. J.; Kulathila, R.; Francisco, W. A.; Ash, D. E.; Bell, J. FEBS
Lett. 1995, 366, 165-169.
(26) Merkler, D. J.; Kulathila, R.; Ash, D. E. Arch. Biochem. Biophys. 1995,
317, 93-102.
(27) Klinge, M.; Cheng, H. M.; Zabriskie, T. M.; Vederas, J. C. J. Chem. Soc.,
Chem. Commun. 1994, 1379-1380.
(28) Casara, P.; Ganzhorn, A.; Philippo, C.; Chanal, M. C.; Danzin, C. Bioorg.
Med. Chem. Lett. 1996, 6, 393-396.
(29) Erion, M. D.; Tan, J.; Wong, M.; Jeng, A. Y. J. Med. Chem. 1994, 37,
4430-4437.
(30) Prigge, S. T.; Kolhekar, A. S.; Eipper, B. A.; Mains, R. E.; Amzel, L. M.
Nat. Struct. Biol. 1999, 6, 976-983.
(31) Prigge, S. T.; Kolhekar, A. S.; Eipper, B. A.; Mains, R. E.; Amzel, L. M.
Science 1997, 278, 1300-1305.
(32) Prigge, S. T.; Eipper, B. A.; Mains, R. E.; Amzel, L. M. Science 2004,
304, 864-867.
(33) Francisco, W. A.; Merkler, D. J.; Blackburn, N. J.; Klinman, J. P.
Biochemistry 1998, 37, 8244-8252.
(34) Francisco, W. A.; Knapp, M. J.; Blackburn, N. J.; Klinman, J. P. J. Am.
Chem. Soc. 2002, 124, 8194-8195.
(35) Francisco, W. A.; Blackburn, N. J.; Klinman, J. P. Biochemistry 2003, 42,
1813-1819.
(39) Rauk, A.; Yu, D.; Taylor, J.; Shustov, G. V.; Block, D. A.; Armstrong, D.
A. Biochemistry 1999, 38, 9089-9096.
(36) Bell, J.; El Meskini, R.; D’Amato, D.; Mains, R. E.; Eipper, B. A.
Biochemistry 2003, 42, 7133-7142.
(40) Easton, C. J. Chem. ReV. 1997, 97, 53-82.
(41) Landymore-Lim, A. E. N.; Bradbury, A. F.; Smyth, D. G. Biochem. Biophys.
Res. Commun. 1983, 117, 289-293.
(37) Croft, A. K.; Easton, C. J.; Radom, L. J. Am. Chem. Soc. 2003, 125, 4119-
4124.
(38) Croft, A. K.; Easton, C. J.; Kociuba, K.; Radom, L. Tetrahedron:
Asymmetry 2003, 14, 2919-2926.
(42) Ping, D.; Mounier, C. E.; May, S. W. J. Biol. Chem. 1995, 270, 29250-
29255.
9
J. AM. CHEM. SOC. VOL. 126, NO. 41, 2004 13307

A R T I C L E S
Chart 1
Barratt et al.
Table 1. Correlation of Studies of Free Radical Brominations and ab Initio Calculations of Radical Stabilization Energies with the Extent of
Catalysis Displayed by Peptidylglycine R-Amidating Monooxygenase
PART C
kinetic parameters for interactions of various substrates and inhibitors with
peptidylglycine
R
-amidating monooxygenase
PART A
PART B
relative rates of reaction of 1a
−
8a
radical stabilization energies
[0 K, RMP2/G3large]
kinetic parameters
with N-bromosuccinimide to give the
corresponding radicals 1c
−8c
of derivatives of glycyl and related radicals
c
glycine derivative
or substitute
VM,app
mol min-1 mg-
KM,app
(mM)
KI
IC50
1
compd
relative rate of reactiona
radical
RSEb
compd
(µ
)
(mM)
(mM)
N-benzoyl
N-pentafluorobenzoyl
N-acetyl
N-trifluoroacetyl
other N-acyl
1a
2a
3a
4a
1.0d
0.25f
1.2f
1b
6.5e
1.3e
3c
4c
79.1
69.9
3b
4b
6.4e
1.4e
9.3e
0.05f
4.1e
10a
11a
12a
13a
14a
12.6e
0.1e
0.03
3.3g
8.2e
5.6g
0.0012g
0.096e
0.0079g
R-methyl
R-trifluoromethyl
5a
6a
0.33h
5d
6d
78.8
39.9
<0.0005i,j
9a
9b
7b
10b
11b
12b
13b
14b
15
inhibitor
inhibitor
inhibitor
inhibitor
inhibitor
inhibitor
inhibitor
inhibitor
inhibitor
inhibitor
5
5
glycolate
7a
<0.0005j
7d
44.4
0.25
0.04
0.5
0.05
0.0598g
0.0452g
γ-keto acid
8a
<0.0005j
8d
34.9
6
3
16
a From mixtures of substrates, in carbon tetrachloride at reflux. b Radical stabilization energies (RSEs) were calculated as the energy change in the isodesmic
reaction R• + CH4 f RH + •CH3. The RSEs correspond to the differences between the bond dissociation energies (BDEs) of methane and RH,45 and reflect
the stability of R• compared with •CH3, relative to the corresponding closed-shell systems. c Corresponds to loss of 50% of the catalytic activity of PAM in
processing the substrate (R)-tyrosyl-(S)-valylglycine at a concentration of 0.1 mM, under conditions where KM,app for the substrate is 0.2 mM. Although
K
M,app, KI, and IC50 values are not directly comparable as measures of enzyme binding affinity, they are adequate as used in the present work, mainly to
establish that compounds interact with PAM. d Assigned as unity. e Data from ref 11. f Data from ref 38. g Data from ref 42. h Data from ref 44. i Data from
ref 37. j No detectable bromination.
compounds, but there has been no analysis of, nor explanation
for, their behavior. One γ-keto acid has been previously
investigated, but it was found not to interact with PAM, either
as a substrate or an inhibitor.43
Compiled in Table 1 (part A) are the relative rates of reaction
of the acylated glycine derivatives 1a-4a, as well as the
derivatives of alanine 5a, trifluoroalanine 6a, glycolate 7a, and
γ-ketopropionate 8a, with N-bromosuccinimide at reflux in
carbon tetrachloride. They were derived by measuring the
(43) Tamburini, P. P.; Jones, B. N.; Consalvo, A. P.; Young, S. D.; Lovato, S.
J.; Gilligan, J. P.; Wennogle, L. P.; Erion, M.; Jeng, A. Y. Arch. Biochem.
Biophys. 1988, 267, 623-631.
9
13308 J. AM. CHEM. SOC. VOL. 126, NO. 41, 2004

Inhibition of Peptidylglycine
R
-Amidating Monooxygenase
A R T I C L E S
relative rates of consumption of 1a-8a from mixtures.37,38,44
Since the reactions involve radical bromination, with hydrogen
atom transfer from the carbon adjacent to the ester group of
each substrate determining the relative rate at which that
compound is brominated, the relative rates of reaction cor-
respond to the relative ease of formation of the radicals 1c-8c.
For comparison with the relative rates of bromination of 1a-
8a, and the relative ease of formation of the radicals 1c-8c in
those reactions, RSEs for the radicals 3c, 4c, and 5d-8d were
determined (Table 1, part B). The RSEs relate the stability of
the radicals in question to that of methyl radical (relative to the
corresponding closed-shell species), with a more stable radical
having a more positive RSE.45 The geometries and zero-point
vibrational energies were determined at the B3-LYP/6-31G(d)
level, while improved relative energies were obtained by
carrying out single-point calculations on these optimized
structures at the RMP2 level with the 6-311+G(2df,p) and
G3large basis sets.46 The results quoted in Table 1, part B,
correspond to RMP2/G3large//B3-LYP/6-31G(d) RSEs at 0 K.
The radicals 1c, 2c, and 5c-8c were not studied because their
aryl groups would substantially increase the complexity of the
calculations. The acetamides 5d-8d were examined instead of
the benzamides 5c-8c. The effect of different acyl groups was
examined with only 3c and 4c, and not 1c and 2c. The level of
theory used in these calculations is significantly higher than
that used in our earlier studies.37,38 Nevertheless, the results are
in good agreement where comparisons are possible. In particular,
they show the RSEs of the radicals 3c and 5d to differ only
slightly (by 0.3 kJ mol-1), consistent with our previous work
(which gave a difference of 1.6 kJ mol-1).
literature procedures.19,22,28 There was no evidence of enzyme-
catalyzed reaction for 7b, 9a,b, 10b-12b, 15, or 16. A double-
reciprocal plot of 1/V vs 1/[S] was used to determine the KM,app
value for 11a, and Dixon plots of 1/V vs [I] were used to
determine the IC50 values for 7b, 9a,b, 10b-12b, 15, and 16.
Also shown in Table 1 (part C) are kinetic parameters for
the interactions of the glycine derivatives 1b, 3b, 4b, and 10a-
14a, the trifluoroalanine derivatives 9a and 9b, the glycolates
7b and 10b-14b, and the γ-keto acids 15 and 16 with PAM.
The data for the glycine derivatives 1b, 3b, and 4b, and the
glycolate 7b, allow for a direct comparison with the rates of
bromination of the analogous esters 1a, 3a, 4a, and 7a, and the
stability and ease of formation of the corresponding radicals
1c, 3c, 4c, and 7c. The properties of the glycine derivatives
10a-14a and the glycolates 10b-14b allow for further analysis
of the relationships between these two classes of compounds.
It was not feasible to examine the interaction of either the
trifluoroalanine derivative 6b or the γ-keto acid 8b with PAM,
for comparison with the glycine derivative 1b, due to their poor
enzyme-binding affinities. Instead, the effect of these substitu-
tions was explored using the more tightly binding trifluoroala-
nine derivatives 9a and 9b, and the γ-keto acids 15 and 16,
and relating their behavior to that of the analogous glycinated
dipeptide derivatives 13a and 14a. The data for 1b, 3b, 4b,
10a, 12a-14a, 13b, and 14b have been previously re-
ported.11,22,42 The trifluoroalanine derivatives 9a,b, the glycolates
7b and 10b-12b, the dipeptide 11a, and the γ-keto acids 15
and 16 were prepared as described in the Supporting Informa-
tion. Their interactions with PAM were assayed using modified
Discussion
N-Benzoyl-, pentafluorobenzoyl-, acetyl-, and trifluoroacetyl-
glycine methyl esters 1a-4a differ only in their N-acyl
substituents. In their reactions with N-bromosuccinimide, the
benzamide 1a is 4 times more reactive than the pentafluoro-
benzamide 2a, and the acetamide 3a is 24 times more reactive
than the trifluoroacetamide 4a (Table 1, part A). These relative
reactivities can be attributed to differences between the π-electron-
donating abilities of the various amido groups, to stabilize and
therefore facilitate formation of the corresponding radicals 1c-
4c. The inductively electron-withdrawing fluorines of the
pentafluorobenzamide 2a and the trifluoroacetamide 4a decrease
the extent of resonance stabilization and rates of formation of
the radicals 2c and 4c, relative to those radicals 1c and 3c
derived from the non-halogenated precursors 1a and 3a,
respectively. This effect of the fluorines is reflected in the
calculated RSEs in the cases of the radicals 3c and 4c, where
the trifluoroacetamide 4c is less stable than the acetamide 3c
by 9.2 kJ mol-1 (Table 1, part B). The fluorines also exert an
inductive effect that decreases the pKas of the carboxylic acid
analogues of the amido groups, due to stabilization of the
corresponding carboxylate anions (the pKas of benzoic acid,
pentafluorobenzoic acid, acetic acid, and trifluoroacetic acid are
4.2, 1.5, 4.8, and 0.6, respectively47). As a result, there is a strong
correlation between the effect of the fluorines on the relative
(44) Burgess, V. A.; Easton, C. J.; Hay, M. P. J. Am. Chem. Soc. 1989, 111,
1047-1052.
(45) See, for example: (a) Parkinson, C. J.; Mayer, P. M.; Radom, L. Theor.
Chem. Acc. 1999, 102, 92-96. (b) Parkinson, C. J.; Mayer, P. M.; Radom,
L. J. Chem. Soc., Perkin Trans. 2 1999, 2305-2313. (c) Henry, D. J.;
Parkinson, C. J.; Mayer, P. M.; Radom, L. J. Phys. Chem. 2001, 105, 6750-
6756.
(46) These levels of theory have been found to be suitable for the calculation
of radical stabilization energies: see ref 45c.
(47) Krygowski, T. M.; Guilleme, J. J. Chem. Soc., Perkin Trans. 2 1982, 531-
534.
9
J. AM. CHEM. SOC. VOL. 126, NO. 41, 2004 13309

A R T I C L E S
Barratt et al.
reactivity of the glycine derivatives (1a-4a), the relative acidity
of the acids, and the RSEs of the radicals 3c and 4c. The effect
of the fluorines on the reactivity of 1a-4a and the acidity of
the corresponding carboxylic acids is greatest with the trifluo-
roacetamide 4a and trifluoroacetic acid, relative to the acetamide
3a and acetic acid, respectively.
tight binding to the enzyme. This is seen from the IC50 value
of 5 mM for each of the trifluoroalanine derivatives 9a and 9b.
As an indirect comparison, the KM,app value of the corresponding
glycine derivative 14a is 7.9 µM.
In a manner similar to the effect of introducing the triflu-
oromethyl group, replacing the glycine moiety with glycolate,
which is a simple substitution of the glycine NH by O, reduces
the RSE of the glycolyl radical 7d compared with that of the
corresponding glycyl radical 3c by 34.7 kJ mol-1 (Table 1, part
B). This can be attributed to the decreased π-electron-donating
and increased σ-electron-withdrawing ability of the acetoxy
group of 7d relative to the acetamido substituent of 3c, as
In analogous systems, the fluorines have a related impact on
the interactions of N-acetylglycine 3b and N-trifluoroacetyl-
glycine 4b with PAM. They decrease the overall rate of turnover
(VM,app) of 4b relative to 3b, via the corresponding glycyl
radicals, by a factor of 4.5 (Table 1, part C). Consequently, the
correlation of the differences between the RSEs of the N-acetyl-
and trifluoroacetyl-glycyl radicals 3c and 4c, the relative rates
of bromination of N-acetyl- and trifluoroacetyl-glycine methyl
esters 3a and 4a, and the pKas of acetic and trifluoroacetic acids
extends to the relative rates of the enzyme-catalyzed reactions
of N-acetyl- and trifluoroacetyl-glycine 3b and 4b. The fluorines
reduce the RSE of the radical 4c (by 9.2 kJ mol-1), the rate of
bromination of the ester 4a (by a factor of 24), the pKa of
trifluoroacetic acid (by 4.2 units), and the rate of turnover of
the acid 4b by the enzyme (by a factor of 4.5). By way of
comparison, replacing the acetyl substituent of N-acetylglycine
methyl ester 3a, N-acetylglycine 3b, and acetic acid with the
benzoyl group of N-benzoylglycine methyl ester 1a, N-ben-
zoylglycine 1b, and benzoic acid has relatively little effect on
any of these properties. It decreases the rate of bromination of
the ester 1a by a factor of only 1.2, decreases the pKa of benzoic
acid by only 0.6, and changes the rate of the enzyme-catalyzed
reaction of the acid 1b by less than 2%.
•
reflected in the RSEs of the radicals MeCONHCH2 and MeCO2-
•
CH2 of 41.3 and 17.1 kJ mol-1, respectively.48 The effect is
magnified in the captodatively stabilized49,50 glycyl and glycolyl
radicals 3c and 7d, where the extent of the synergy displayed
by the acetamido and acetoxy substituents in combination with
the carboxyl group corresponds to 17.6 and 7.1 kJ mol-1. These
values are based on the differences between the RSEs of
MeCONHCH2• (41.3 kJ mol-1), •CH2CO2Me (20.2 kJ mol-1),
and 3c (79.1 kJ mol-1) and MeCO2CH2• (17.1 kJ mol-1), •CH2-
CO2Me, and 7d (44.4 kJ mol-1), respectively. The substitution
of glycine by glycolate prevents bromination of methyl OR-
benzoylglycolate 7a (Table 1, part A). It also blocks catalysis
of the reactions of the glycolates 7b and 10b-14b by PAM
(Table 1, part C). The glycine derivatives 1b and 10a-14a are
all turned over by the enzyme, but there is no evidence of
reaction of any of the corresponding glycolates 7b and 10b-
14b. In this case, the substitution does not severely disrupt
binding to the enzyme. The IC50 and KI values of the glycolates
7b and 10b-14b are all in the 0.04-0.5 mM range. The
corresponding glycine derivatives 1b and 10a-14a have KM,app
values between 1 µM and 1 mM. Therefore, the glycolates
constitute a general class of inhibitors of the enzyme because
they do not readily undergo hydrogen atom transfer, yet they
bind effectively to PAM.
The effect of the fluorines, although reducing the rate of
turnover, is clearly insufficient to prevent the PAM-catalyzed
reaction of N-trifluoroacetylglycine 4a. On the basis of the
calculations with the N-acetyl- and trifluoroacetyl-glycyl radicals
3c and 4c, this indicates that a reduction in the RSE of 9.2 kJ
mol-1 is not sufficient to stop glycyl radical formation by the
enzyme. This is consistent with reductions in the RSEs of alanyl
and threonyl radicals relative to those of the corresponding
glycyl radicals, of 0.3 and 7-14 kJ mol-1, respectively,37,39
being too small to block the enzyme-catalyzed reactions of
alanine and threonine derivatives.32 The effect of the trifluoro-
methyl substituent on the stability of the trifluoroalanyl radical
6d is much larger, reducing the calculated RSE relative to that
of the corresponding glycyl radical 3c by 39.2 kJ mol-1 (Table
1, part B). This substituent also prevents bromination of
N-benzoyl-â,â,â-trifluoroalanine methyl ester 6a (Table 1, part
A) and the processing of the trifluoroalanine-containing dipep-
tides 9a,b by PAM (Table 1, part C). N-Benzoyl-â,â,â-
trifluoroalanine methyl ester 6a was inert on treatment with
N-bromosuccinimide, either alone or as a mixture with N-
benzoylglycine methyl ester 1a, that nevertheless reacted
smoothly to give the corresponding R-bromoglycine derivative.
There was no evidence of reaction of either of the trifluoroala-
nine-containing dipeptides 9a or 9b being catalyzed by PAM,
even though competitive experiments with (R)-tyrosyl-(S)-
valylglycine established that both 9a and 9b bind to the enzyme,
albeit with IC50 values of around 5 mM (Table 1, part C). It is
therefore apparent that the decrease in the RSE brought about
by introducing the trifluoromethyl group is sufficient to prevent
radical formation by PAM. However, this does not give effective
enzyme inhibitors, as the trifluoromethyl substituent prevents
Replacing the acylglycine with a γ-keto acid or ester, through
substitution of the glycine NH by CH2, has effects similar to
the swapping of glycine for glycolate. The RSE of the keto ester
radical 8d is less than that of the analogous glycyl radical 3c
by 44.2 kJ mol-1 (Table 1, part B). The γ-keto ester 8a is inert
to bromination (Table 1, part A), and the γ-keto acids 15 and
16 are not processed by PAM (Table 1, part C), even though,
unlike the example reported previously,43 these keto acids 15
and 16 do bind to some extent to the enzyme, with IC50 values
of 6 and 3 mM, respectively. Again it is apparent that a reduction
in the RSE of around 35-45 kJ mol-1 is sufficient to stop both
the bromination with N-bromosuccinimide and the catalysis by
the monooxygenase.
From the KM,app values of the glycine derivatives 13a and
14a, 96 and 7.9 µM, respectively, the KI values of the
corresponding glycolates 13b and 14b, 59.8 and 45.2 µM, and
the IC50 values of the analogous γ-keto acids 15 and 16, 6 and
3 mM (Table 1, part C), it appears that the replacement of an
acylglycine with a γ-keto acid has a more adverse effect on the
(48) RSE values quoted in the text, in addition to those listed in Table 1, were
also calculated using the RMP2/G3large level of theory.
(49) Welle, F. M.; Beckhaus, H.-D.; Ru¨chardt, C. J. Org. Chem. 1997, 62, 552-
558.
(50) Schulze, R.; Beckhaus, H.-D.; Ru¨chardt, C. Chem. Ber. 1993, 126, 1031-
1038.
9
13310 J. AM. CHEM. SOC. VOL. 126, NO. 41, 2004

Inhibition of Peptidylglycine
R
-Amidating Monooxygenase
A R T I C L E S
binding to PAM than does the substitution of glycine by
glycolate. One explanation for this is that both acylglycines and
glycolates can form hydrogen bonds with the enzyme, through
NH and the corresponding O, respectively, to facilitate binding,
but this is not possible with the methylene of an analogous
γ-keto acid. An alternative explanation for the less effective
binding of the γ-keto acids relates to their greater conformational
freedom in free solution. As a result of conjugation of the amide
NH of an acylglycine and the analogous ester O of an
acylglycolate, in each case there is easy rotation only around
the bonds to the R-carbon. By contrast, with a γ-keto acid there
is easy rotation around the bonds to both the R- and â-carbons.
Assuming that the acylglycines, glycolates, and keto acids all
bind to PAM with similar orientations, and that the planarity
of the amide and ester groups of the acylglycine and acylgly-
colate is maintained, there is therefore a greater degree of
conformational constraint and loss of entropy on binding of a
γ-keto acid.
In any event, it is clear that introducing an R-trifluoromethyl
substituent into the glycine residue of a PAM substrate, or
replacing the acylglycine of a substrate with either a glycolate
or a γ-keto-acid, stops catalysis by the enzyme, because it
reduces by around 35-45 kJ mol-1 the RSE of the radical that
would be formed. By contrast, swapping the acylglycine for a
â,γ-unsaturated acid is likely to increase the RSE of the radical
intermediate by around 21.1 kJ mol-1, based on the values
calculated for the acetamido and propenyl radicals 3c and 17,
79.1 and 100.2 kJ mol-1, respectively. Thus, it is not surprising
that, as outlined in the Introduction, â,γ-unsaturated acids are
processed by PAM.17,23,24
glycyl and related radicals in free radical brominations, and the
extent of catalysis of the reactions of analogous compounds by
peptidylglycine R-amidating monooxygenase (PAM). A de-
crease in the RSE of around 10 kJ mol-1 relative to that of a
normal peptide glycyl radical slows but does not stop the rate
of radical formation through either bromination or enzyme
catalysis. However, reducing the RSE by around 35-45 kJ
mol-1, either through introduction of a trifluoromethyl sub-
stituent at the R-position or by replacing the acylglycine NH
with a glycolate O or a γ-keto acid CH2, stops both bromination
and enzyme processing. Of these changes, only the glycolate
substitution does not adversely affect the enzyme binding
affinity. Consequently, glycolates are a general class of PAM
inhibitors because they do not readily undergo hydrogen atom
transfer, yet they bind effectively to the enzyme. We are now
beginning to study the associated in vivo effects of this class
of compounds, and already we have found that both the
glycolates 10b and 12b actively reduce the levels of the peptide
hormone Substance P produced by rat DRG cells over a 6 h
period.51
Acknowledgment. We gratefully acknowledge receipt of
financial support from the Australian Research Council and
generous allocations of computing time from the National
Facility of the Australian Partnership for Advanced Computing
and the Australian National University Supercomputing Facility.
Supporting Information Available: Details of the procedures
and methods used to determine the relative rates of reaction of
1a-8a with N-bromosuccinimide, the methods used for the ab
initio calculations, the Gaussian archive entries for the
UB3-LYP/6-31G(d) optimized structures of derivatives of glycyl
and related radicals, calculated RMP2/G3large and RMP2/
6-311+G(2df,p) total energies and corresponding RSEs; pro-
tocols for the syntheses of 7b, 9a,b, 10b, 11a,b, 12b, 15, and
16 with studies of their interactions with PAM. This material
is available free of charge via the Internet at http://pubs.acs.org.
Conclusion
JA046204N
In summary, there are strong correlations between calculated
radical stabilization energies (RSEs), the ease of formation of
(51) Easton, C. J.; Inoue, A.; Wright, A. Unpublished results.
9
J. AM. CHEM. SOC. VOL. 126, NO. 41, 2004 13311