Low-molecular-weight peptidic and cyclic antagonists of the receptor for the complement factor C5a

Article abstract of DOI:10.1021/jm9806594

Activation of the human complement system of plasma proteins during immunological host defense can result in overproduction of potent proinflammatory peptides such as the anaphylatoxin C5a. Excessive levels of C5a are associated with numerous immunoinflammatory diseases, but there is as yet no clinically available antagonist to regulate the effects of C5a. We now describe a series of small molecules derived from the C-terminus of C5a, some of which are the most potent low-molecular-weight C5a receptor antagonists reported to date for the human polymorphonuclear leukocyte (PMN) C5a receptor. ¹H NMR spectroscopy was used to determine solution structures for two cyclic antagonists and to indicate that antagonism is related to a turn conformation, which can be stabilized in cyclic molecules that are preorganized for receptor binding. While several cyclic derivatives were of similar antagonistic potency, the most potent antagonist was a hexapeptide- derived macrocycle AcF[OPdChaWR] with an IC₅₀ = 20 nM against a maximal concentration of C5a (100 nM) on intact human PMNs. Such potent C5a antagonists may be useful probes to investigate the role of C5a in host defenses and to develop therapeutic agents for the treatment of many currently intractable inflammatory conditions.

Full text of DOI:10.1021/jm9806594

J . Med. Chem. 1999, 42, 1965-1974
1965
Low -Molecu la r -Weigh t P ep tid ic a n d Cyclic An ta gon ists of th e Recep tor for th e
Com p lem en t F a ctor C5a
Angela M. Finch,† Allan K. Wong,‡ Natalii J . Paczkowski,† S. Khe´mar Wadi,† David J . Craik,‡
David P. Fairlie,‡ and Stephen M. Taylor*,†
Department of Physiology and Pharmacology and Centre for Drug Design and Development, University of Queensland,
Brisbane, Queensland 4072, Australia
Received November 23, 1998
Activation of the human complement system of plasma proteins during immunological host
defense can result in overproduction of potent proinflammatory peptides such as the anaphy-
latoxin C5a. Excessive levels of C5a are associated with numerous immunoinflammatory
diseases, but there is as yet no clinically available antagonist to regulate the effects of C5a.
We now describe a series of small molecules derived from the C-terminus of C5a, some of which
are the most potent low-molecular-weight C5a receptor antagonists reported to date for the
1
human polymorphonuclear leukocyte (PMN) C5a receptor. H NMR spectroscopy was used to
determine solution structures for two cyclic antagonists and to indicate that antagonism is
related to a turn conformation, which can be stabilized in cyclic molecules that are preorganized
for receptor binding. While several cyclic derivatives were of similar antagonistic potency, the
most potent antagonist was a hexapeptide-derived macrocycle AcF[OPdChaWR] with an IC50
) 20 nM against a maximal concentration of C5a (100 nM) on intact human PMNs. Such potent
C5a antagonists may be useful probes to investigate the role of C5a in host defenses and to
develop therapeutic agents for the treatment of many currently intractable inflammatory
conditions.
In tr od u ction
recognition site, while the receptor activation domain
was localized to the C-terminal region.9 This ligand-
receptor interaction has been recently termed the “two-
site” binding hypothesis and may be a general charac-
teristic of G protein-coupled receptors and their endoge-
nous high-molecular-weight ligands.10
During host defense the complement system of plasma
proteins initiates inflammatory and cellular responses
to stimuli such as infectious microorganisms (bacteria,
viruses, parasites), chemical and physical injury, radia-
tion, or neoplasia. Prolonged activation of the human
complement system results in elevated C5a levels that
are associated with numerous diseases including rheu-
matoid arthritis,1 Alzheimer’s disease,2 ischemic heart
disease,3 and adult respiratory distress syndrome
(ARDS).4 At present there is no clinically available
antagonist to combat the effects of C5a in disease
development.
C5a receptor (C5aR) antagonists can in principle be
developed through modifying the structure of C5a itself
or that of small components of C5a. Site-directed mu-
tagenesis experiments5-8 on C5a have demonstrated
that multiple residues of this 74-amino acid 4-helix
bundle protein interact with the receptor. The interhe-
In an effort to develop C5aR antagonists, several
groups have synthesized small-peptide analogues (6-
15 amino acids) derived from the C-terminus of C5a
which have increased affinities for the C5a receptor and
are more potent C5a agonists than the corresponding
C-terminal residues of the native C5a sequence.12-18
Some small-peptide agonists derived from the C-
terminal decapeptide sequence of C5a were found to
have partial agonist properties in PMNs14,19 and have
been modified to produce the first complete antagonist,
MeFKPdChaWr.20 On PMN membranes the affinity of
this hexapeptide for C5a receptors was only 0.04% that
of C5a.20 Apart from a recent study from our laborato-
ries,21 there have been no succeeding reports of low-
molecular-weight C5a antagonists with increased po-
tency compared to MeFKPdChaWr. Although high-
affinity protein C5a antagonists have been recently
described by site-directed mutagenesis of C5a,22,23 such
large molecules are expensive to manufacture, have low
bioavailabilities and poor pharmacokinetic properties for
use as drugs, and would need to be administered
parenterally. Indeed, proteins and peptides generally
have low bioavailability and poor metabolic stability
making them unsuitable as oral drug candidates. To
overcome these problems, an orally bioavailable C5a
antagonist will need to be a small molecule (MW ) 1000
Da) with reduced peptide character and greater potency
than MeFKPdChaWr.
lical loops of C5a (loop 1, C5a13-17; loop 3, C5a40-45
)
containing positively charged amino acids are thought
to bind to a negatively charged extracellular region of
the receptor and together constitute a major ligand-
receptor binding domain.10 The receptor binding “effec-
tor” domain of C5a is located at its C-terminus and is
entirely responsible for receptor activation leading to
signal transduction.11 Early studies on C5a and frag-
ments of C5a established that the bulk of the molecule
away from the C-terminal end contained the receptor
* Address correspondence to S. M. Taylor (fax, +61-7-3365-1766;
e-mail, s.taylor@mailbox.uq.edu.au).
† Department of Physiology and Pharmacology.
‡ Centre for Drug Design and Development.
10.1021/jm9806594 CCC: $18.00 © 1999 American Chemical Society
Published on Web 05/07/1999

1966 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 11
Ta ble 1. Structure-Activity Relationships of C5a Antagonist Peptides
Finch et al.
receptor affinity
antagonist potency
-log IC50 ((SE)b IC50 (µM)b
no.
peptide
MeFKPdChaWr
MeFKPdChaWR
F KPdChaWr
AKPdChaWr
AcFKPdChaWr
-log IC50 ((SE)a
IC50 (µM)a
n
n
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16a
16b
5.70 ( 0.06
4.95 ( 0.06*
5.71 ( 0.08
2.0
11
2.0
29
5
2
2
2
3
3
5
3
3
3
3
3
9
3
3
3
6.97 (0.09
5.71 ( 0.35*
5.17 ( 0.22*
ND
ND
6.57
0.1
2
6.8
17
6
3
>3
>1000
5.32 ( 0.10
5.58 ( 0.26
5.34 ( 0.08
5.91 ( 0.11
6.76 (0.08*
5.72 ( 0.04
6.10 ( 0.26
5.52 ( 0.14
5.45 ( 0.14
5.04 ( 0.42
5.23 ( 0.04
4.29 ( 0.10*
3.94 ( 0.10*
4.8
2.6
4.6
1.2
0.2
1.9
0.8
3.0
3.6
9.0
5.9
51
KFKPdChaWr
0.3
0.5
0.2
0.07
1
1
4
4
F KFKPdChaWr
C5a 12-20-a h x-FKPdChaWr
C5a 12-20-a h x-GGGG-FKPdChaWr
C5a 40-68-FKPdChaWr
AcFOPdChaWr
6.30
6.61 ( 0.12
7.14 ( 0.16
ND
6.72 ( 0.19
6.06 ( 0.07*
5.85 ( 0.38*
5.44 ( 0.11*
6.50 ( 0.17
ND
0.2
0.9
1.4
3.7
0.3
3
3
3
4
3
AcFxPdChaWrc
AcFNlePdChaWr
FAPdChaWr
FWPdChaWr
AcFKPdChaZrd
AcFKPdChaZre
116
ND
a
b
-log IC50 and IC50, concentration of peptide resulting in 50% inhibition in the binding of [125I]C5a to intact PMNs. -log IC50 and
IC50, concentration of peptide resulting in 50% inhibition of C5a (100 nM) causing the release of MPO from PMNs. c x, â-aminoalanine
or 2,3-diaminopropanoic acid (Dap). Z, fluorenylalanine isomer I (see Experimental Section). e Z, fluorenylalanine isomer II (see
d
Experimental Section); O, ornithine; n, number of experiments performed. Substituted amino acids are highlighted in boldface. *Significant
change in affinity/potency compared to MeFKPdChaWr (p < 0.05). ND, not determined.
We have recently shown that the hexapeptide C5a
antagonist MeFKPdChaWr (1) has significant structure
in solution,21 and we now report attempts to stabilize
its bioactive turn conformation by specific side chain to
backbone cyclizations of the peptide. This approach can
increase receptor affinity and antagonist potency while
reducing the peptide character of the antagonist. Struc-
ture-function relationships are explored for a series of
acyclic and cyclic antagonists derived from the C-
terminus of C5a which illustrate the merit of structur-
ally preorganizing the antagonist for receptor binding.
(1) on receptor affinity and antagonist potency, as
measured respectively by competition with [125I]C5a on
whole PMNs and inhibition of myeloperoxidase (MPO)
release from PMNs induced by a maximal concentration
(100 nM) of C5a. The antagonist MeFKPdChaWr (1) has
N-methyl and D-Arg residues to protect both ends of the
molecule from peptide cleavage by endopeptidases in
serum, but the effects of other substitutions at the N-
and C-termini were not known and were investigated
first.
Substitution at the C-terminus of L-Arg for D-Arg (2
vs 1) led to a 5-20-fold decrease in both receptor affinity
and antagonist potency (Figure 1). We therefore re-
tained D-Arg (designated r) at the C-terminus in all
other peptides shown in Table 1. At the N-terminus, the
phenylalanine is necessary for both receptor binding and
antagonism since replacement by alanine dramatically
reduces activity to undetectable levels (4 vs 3). Neither
methylation nor acetylation of this N-terminal phenyl-
alanine significantly affects receptor affinity (1 and 5
vs 3), suggesting the possibility of chain extension to
increase activity.
Extension of the hexapeptide at its N-terminus by one
or two additional residues (6, 7), Lys and Phe being
chosen because of their reported importance in longer
agonist peptides,18 failed to increase affinity for the
receptor. Addition of a longer peptide sequence, corre-
Resu lts a n d Discu ssion
Our previous NMR structure study of MeFKPdChaWr
(1) led us to believe that a turn conformation centered
around the K-P-dCha-W residues would engender
antagonist activity.21 That study revealed that while an
inverse γ-turn involving K-P-dCha residues persisted
in solution, there was only a small conformational popu-
lation with a larger â-turn involving K-P-dCha-W
residues, and the N- and C-termini were quite flexible.
We considered that cyclization might stabilize such a
turn conformation, reducing the conformational freedom
of the N- and C-termini thereby increasing antagonist
potency, but needed to first discover which residues in
1 could be modified without losing antagonist potency.
P ep tid ic An ta gon ists. Table 1 summarizes the
effects of amino acid substitutions in MeFKPdChaWr
sponding to one of the interhelical loops (C5a12-20
,
KYKHSVVKK, loop 1) of the 4-helix bundle of C5a, to
the N-terminus of the peptide agonist YSFKPMPLaR
has been shown to increase receptor binding affinity and
activation of the C5a receptors on human PMNs.18 To
investigate if the binding affinity and biological potency
of the antagonist could be similarly enhanced, loop 1
(C5a12-20) was connected to the N-terminus of FKPd-
ChaWr via an ꢀ-aminohexanoic acid linker (8), and
peptide 9 has an additional four glycines in the linker
to further increase the separation between the two
motifs. This longer spacer led to significantly increased
receptor binding for 9, but this was not accompanied
by an increase in antagonist potency (Table 1). Addition

Antagonists of the C5a Receptor
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 11 1967
The tryptophan of 1 has previously been reported to
be important for antagonist activity,20 but we wondered
whether other bulky substituents might be tolerated at
this position. In particular we were interested in
incorporating fluorescent substituents as labels to per-
mit monitoring of ligand-receptor interactions. A fluo-
renylalanine side chain (16) significantly decreased
affinity for the receptor, consistent with the previously
observed requirement for a side chain of specific bulk
(e.g., Trp) at this position to confer antagonism.20
Clearly, while Phe and Leu are smaller than required
to cause antagonism,20 the fluorenylalanine is sterically
too bulky and agonism was observed instead (EC50 ) 1
µM, 16a ; 8 µM, 16b). We conclude that for complete
antagonism the hydrophobic space accommodating this
substituent must be optimally occupied and that smaller
or larger substituents than the indole ring of tryptophan
may lead to agonist activity.
Cyclic An ta gon ists. Next we aimed to limit the
conformational flexibility of the peptide MeFKPdChaWr
(1) by producing cyclic derivatives which could poten-
tially stabilize a turn motif. The peptide can in principle
be cyclized through backbone-to-backbone connection of
N- and C-termini (e.g., A) or through various backbone-
to-side chain connections (e.g., B). The effects of cycliza-
tion on receptor affinity and antagonist potency are
summarized in Table 2 for some molecules representing
examples A and B.
F igu r e 1. Effects of L-Arg and D-Arg on receptor affinity and
antagonist potency: differences between linear and cyclic C5a
antagonists. Panel A: Receptor binding as percent inhibition
of binding of [125I]C5a to human PMNs. Panel B: C5a antago-
nist potency as inhibition of MPO release from human PMNs:
MeFKPdChaWr, 1 (9); MeFKPdChaWR, 2 (0); F[OPdChaWR],
22 (b); and F[OPdChaWr], 21 (O). Data are means
SEM.
(
of a different sequence corresponding to interhelical loop
3 of C5a (C5a37-46, RAARISLGPR) did not increase
receptor affinity of MeFKPdChaWr (10; Table 1), as
noted for modified agonist peptides.18
Returning to the hexapeptide 1, substitution of lysine
by ornithine (11), â-aminoalanine (12), norleucine (13),
alanine (14), or tryptophan (15) had a relatively small
effect on receptor binding (Table 1). Since the side chain
length, charge, and hydrophobicity at this position have
only a small effect on affinity for the C5a receptor, we
considered this site to be promising for linking to the
C-terminus of the antagonist to form cyclic analogues
(see ahead).
The role of proline has been previously investigated
for the C-terminal decapeptide agonist YSFKDMQLGR,
in which a proline scan showed optimum affinity for
YSFKDMP LaR.18 Since proline is in a corresponding
position in the turn-containing antagonist 1, and be-
cause proline plays a unique role in restricting confor-
mational flexibility in peptides due to its restricted φ,
ψ angles, we did not attempt to substitute it with other
amino acids that could compromise the constraint at
this position.
Similarly, the role of hydrophobic residues has previ-
ously been investigated in the agonist octapeptide
AcHKDCh a Ch a LaR and in the antagonists MeFKPd -
Ch a Wr and MeFKPLWr, in which dCha confers at least
a 10-fold increase in receptor affinity.12,13,20 Since cy-
clohexylalanine is one of the larger hydrophobic amino
acids, we did not attempt to further increase steric bulk
or hydrophobicity at this position.
The simplest approach to cyclize the peptide is to
covalently link the N- and C-termini through backbone-
to-backbone cyclization (A). Thus [FWPdChaWr] (17)
and its methylated derivative Me[FWPdChaWr] (18)
were synthesized and assayed (Table 2). For synthetic
convenience these cycles contained Trp in place of the
Lys of 1 to enable the use of an unprotected peptide in
the cyclization reaction, this replacement being shown
to maintain affinity in peptide 15 (Table 1). Cycle 18
(Figure 4) had an equivalent affinity to that of the
acyclic 1 for the PMN C5a receptor, whereas the
nonmethylated cycle 17 had considerably reduced af-
finity. Thus, peptide 1 may be cyclized without apparent
loss of affinity for the receptor, but backbone-to-
backbone cyclization did not increase receptor affinity
or antagonist potency.

1968 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 11
Finch et al.
Ta ble 2. Structure-Activity Relationships of Cyclic C5a Antagonists
receptor affinity
antagonist potency
no.
peptide
-log IC50 ( SEa
IC50 (µM)a
n
-log IC50 ( SEb
IC50 (µM)b
n
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
[FWPdChaWr]
Me[FWPdChaWr]
F[KPdChaWr]
F[KPdChaWR]
F[OPdChaWr]
4.37 ( 0.36*
5.53 ( 0.11
5.09 ( 0.08
6.48 ( 0.16*
5.51 (0.07
7.25 ( 0.03*
5.42 ( 0.05
6.51 ( 0.02*
4.39 ( 0.10*
4.98 ( 0.05*
6.57 ( 0.05*
5.49 ( 0.22
4.60 ( 0.06*
4.77 ( 0.14*
5.02 ( 0.07
43
3
3
3
3
3
3
3
5
3
3
3
4
4
3
3
ND
3.0
8.1
0.3
3.1
0.06
3.8
0.3
41
5.82 ( 0.18*
5.55 ( 0.57*
6.69 ( 0.04
5.79 ( 0.34*
7.46 (0.09*
6.70 ( 0.04
7.36 ( 0.13
ND
5.63 ( 0.13*
7.69 (0.13*
7.07 ( 0.29
6.41 ( 0.10
6.09 ( 0.08*
4.71 ( 0.23*
2.1
2.8
0.2
1.6
0.03
0.2
0.04
3
3
3
3
6
3
3
F[OPdChaWR]
F[Da bPdChaWr]
F[Da bPdChaWR]
F[Da p PdChaWr]
F[Da p PdChaWR]
AcF[OPdChaWR]
AcF[KPdChaWr]
AcF[OPdChaWr]
AcF[Da bPdChaWr]
AcF[Da p PdChaWr]
10
2.4
0.02
0.1
0.4
0.8
3
6
5
4
4
3
0.3
3.2
16
17
9.5
20
a
b
-log IC50 and IC50, concentration of peptide resulting in 50% inhibition in the binding of [125I]C5a to intact PMNs. -log IC50 and
IC50, concentration of peptide resulting in 50% inhibition of C5a (100 nM) causing the release of MPO from PMNs. [ ], cyclization between
the ends of the bracketed residues; Dap, L-2,3-diaminopropanoic acid; Dab, L-2,4-diaminobutanoic acid; O, L-ornithine; n, number of
experiments performed. Substituted amino acids are highlighted in boldface. *Significant change in affinity/potency compared to
MeFKPdChaWr (p < 0.05). ND, not determined.
On the other hand, backbone-to-side chain cyclization
(B), specifically by connecting the C-terminal carboxy-
late to the lysine side chain which had not affected
receptor binding (Table 1), was more promising. The
resulting cycle 19 (n ) 4; R ) H; D-Arg) did not affect
affinity for the C5a receptor on human PMNs versus 1,
but its diastereomer 20 (n ) 4; R ) H, L-Arg) had almost
an order of magnitude higher affinity than 1. This
arginine residue is directly involved in cyclization so its
chirality could be expected to influence the conformation
F igu r e 2. Temperature dependence (20-70 °C) of chemical
of the cycle. By varying the length of the linker in 19 (n
) 1-4), the size of the cycle could also be easily altered,
shifts for amide NH protons of 18 (top) and 27 (bottom) in
DMSO-d6: D-Arg6, 0; L-Arg6, 9; Trp5, 2; dCha4, O; Orn2, 4;
and because of the activity difference between 19 and
20, we compared the activities for each pair of arginine
isomers for each size cycle.
Phe1, [; Trp2, b.
40 nM (Table 2). Thus, the size of the cycle had a
significant effect on the binding affinity of these an-
tagonists with n ) 2 or 3 preferred, corresponding to
macrocycles with 17 or 18 ring atoms. The acetylated
analogue of 22 (namely 27) was even more potent with
antagonist activity IC50 ) 20 nM (Table 2). This is the
most active compound reported to date as a C5a
antagonist.
Compounds 19 and 20 were thus compared with
analogues in which lysine was replaced with progres-
sively smaller amines such L-ornithine (Orn, n ) 3; 21,
22), L-2,4-diaminobutanoic acid (Dab, n ) 2; 23, 24), and
L-2,3-diaminopropanoic acid (Dap, n ) 1; 25, 26). In each
case the antagonist activity was greatest when the
cycles contained L-arginine instead of D-arginine (Table
2, Figure 1), which is the opposite trend to that observed
for the acyclic compounds (Table 1). The larger cycles
with Lys, Orn, and Dab were also more potent com-
pounds in terms of both receptor affinity and antagonist
potency. Compared to the smaller cycles containing Dap
(25, 26) which had 5-20-fold reduced affinity versus 1,
cycles containing L-arginine and ornithine (22) or Dab
(24) bound the receptor with 7-30-fold higher affinity
than 1 and displayed antagonist potencies of about 30-
In our initial study on C5a antagonists,21 we reported
the activity of one of the cyclic compounds tested as a
mixture of Arg diastereoisomers, AcF[OPdChaWr] (29)
and AcF[OPdChaWR] (27). Since the D-Arg conferred
higher affinity and antagonist potency in the corre-
sponding linear peptides, we stereoselectively synthe-
sized 29 and determined its NMR solution structure in
that study. We have now isolated and tested both dia-
stereomers and find that, contrary to the results for the

Antagonists of the C5a Receptor
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 11 1969
Ta ble 3. 1H NMR Assignmentsa for 18 in DMSO-d6
residue
HN
HR
Hâ
2.92
2.87, 3.10
2.02, 1.89
1.28
Hγ
others
MePhe1
Trp2
Pro3
dCha4
Trp5
D-Arg6
3.79
4.93
4.25
4.30
4.58
1.82b
8.80
NH 10.80, 7.10c
3.42, 3.69d
0.74, 1.22
1.74
7.82
8.07
8.48
2.97, 3.10
NH 10.80, 7.10b
3.62d
4.46e 1.67, 1.89e 1.92
4.57f
1.80, 2.02f
a
b
Referenced to residual DMSO-d5 at 2.50 ppm. N-Me. c Aro-
matics. Hδ. e Isomer I. f Isomer II.
d
there were some exceptions, notably compounds 9 and
20 (Tables 1 and 2). The reasons for this are not yet
clear, but such results reinforce the notion that mea-
surements of both affinity and antagonist potency
should be carried out to fully characterize the in vitro
activity of new antagonists.
F igu r e 3. Backbone and proline heavy atoms of 29 low-
energy-minimized NMR structures of backbone cycle 18 in
DMSO-d6 (24 °C). Two sets of conformations 18a (left, 18
structures) and 18b (right, 11 structures) are shown along with
the putative H-bond length (dotted lines) for Trp CO‚‚‚HN
dCha (18a , 2.77 Å; 18b, 2.76 Å).
Recep tor Selectivity. To determine the selectivity
of the cyclic antagonists for the C5aR, compounds 22
and 27 were tested for their ability to block PMN
enzyme secretion to other G protein ligands, such as
formylmethionyl-leucyl-phenylalanine (fMLP), leuko-
triene B4 (LTB4), and platelet-activating factor (PAF).
Maximally effective agonist concentrations of fMLP (1
µM), LTB4 (0.1 µM), or PAF (1 µM) were used to cause
PMN MPO secretion. In concentrations up to 100 µM,
neither 22 nor 27 affected fMLP MPO release, and 22
at 100 µM did not affect either LTB4 or PAF MPO
release. Some inhibition (∼50%) of PAF and LTB4 MPO
secretion was seen with 100 µM 27. Since both C5a
antagonists at concentrations below 1 µM completely
inhibited MPO secretion to a maximally effective con-
centration of C5a (100 nM), the results demonstrate the
highly selective nature of the inhibition for the C5aR.
NMR Solu tion Str u ctu r e of Cycle 18. An NMR
investigation of the solution structure of cycle 18 was
carried out to determine the conformational preference
for the backbone of this cycle. DQF-COSY and TOCSY
two-dimensional NMR spectra were used to identify
residue types from their side chain connectivities.
NOESY spectra were then used to obtain sequential
assignments and chemical shifts listed in Table 3.
H/D exchange and temperature dependence studies
were carried out to provide information about intramo-
1
lecular hydrogen bonds. The 1D H NMR spectrum of
peptide 18 in DMSO-d6 at 25 °C shows four distinct
resonances in the amide region. A 10-fold excess of D2O
was added to this solution to promote exchange of amide
NH’s with deuterium, and 1D spectra were recorded at
regular intervals. Due to fast H/D exchange, the amide
NH of Trp2 disappeared rapidly compared to the other
three amide NH protons for D-Arg, Trp5, and dCha
which remained after 120 min. These peaks persisted
for at least another hour after addition of a further 10-
fold amount of D2O. This behavior is characteristic of
protection from solvent and suggests the possibility of
intramolecular hydrogen bonding.
The temperature dependence (27-67 °C) of amide NH
chemical shifts for cycle 18 is shown in Figure 2. The
temperature coefficients (∆δ/T) were 3.0 ppb/deg (dCha
NH); 4.8 ppb/deg (Trp2 NH); 5.8 ppb/deg (Trp5 NH); 6.4
ppb/deg (D-Arg NH). Values of ∆δ/Τ ∼ 3 ppb/deg or less
usually indicate solvent protection,23 so the amide NH
of dCha is likely hydrogen-bonded in a majority of
F igu r e 4. Structures of cycles 18 (top) and 27 (bottom)
showing components and connectivity.
linear peptide diastereomers, 27 with L-Arg is actually
much more active than 29 which contains D-Arg. Figure
1 summarizes this activity difference. Panel A compares
the receptor binding of L- and D-Arg isomers for a linear
and a cyclic antagonist, while panel B shows the
importance of L- versus D-arginine for antagonist po-
tency. For this reason we report in this study the
solution structure of the active diastereoisomer 27 for
comparison with the previously determined structure
of the less active antagonist 29.
In our earlier preliminary study on C5a antagonists,
we reported that there was a good correlation between
binding affinity and antagonist potency.21 Although
increases in affinity for the PMN C5a receptor were
usually accompanied by increased antagonist potency,

1970 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 11
Ta ble 4. Backbone φ, ψ Angles (deg) for Peptides 18 and 27
Finch et al.
Ta ble 5. 1H NMR Assignmentsa for 27 in DMSO-d6
residue HNb HR
Hâ Hγ others
AcPhe 8.10 4.51 2.69, 2.96
1.74 (Me), 7.22c
φ, ψ
φ, ψ
residue
18a
18b
residue
27
Orn
Pro
7.95 4.55 1.44, 1.64 1.22 2.76, 3.38,d 7.04 (NHꢀ)
Phe1
Trp2
Pro3
dCha4
Trp5
D-Arg6
70, -176
65, -174
-75, 43
132, 162
55, 86
51, -74
-96, 178
-68, 78
131, 157
42, 40
Phe1
Orn2
Pro3
dCha4
Trp5
Arg6
-167, 56
-97, 146
-71, 160
65, 176
-91, 28
-151, 53
4.57 1.63, 1.99 1.63 3.41, 3.62d
dCha
Trp
Arg
8.16 4.01 1.15, 1.26 0.93 0.62, 0.69
8.40 4.28 2.96, 3.26
7.15, 7.39,c 10.86 (NH)
7.81 4.11 1.63, 1.87 1.50 3.11,d 7.57 (NHꢀ)
a
b
53, 90
60, -173
Referenced to residual DMSO-d5 at 2.50 ppm. Amide NH’s,
3J NH-CRH (Hz): 8.51 (Phe), 7.14 (Orn), 4.94 (dCha), 7.14 (Trp), 8.23
(Arg). c Aromatic CHs. Hδ.
d
conformers. On the other hand, the slow exchange
behavior but high temperature coefficient of the D-Arg
and Trp5 amide NH’s reflects solvent protection/in-
tramolecular hydrogen bonding in a much smaller
population of solution conformers.
respectively. Hence as found for linear peptide 1,20 an
inverse γ-turn is the major feature of the solution
structure of cycle 18. Potential hydrogen bonding be-
tween Trp2 CdO and Trp5 N-H to form a 10-membered
ring â-turn was eliminated as a possibility, at least for
a majority population of conformers, because of the large
distance between the donor and acceptor atoms.
NOESY experiments afforded spectra with a good
signal-to-noise ratio, and several of the identified NOEs
for this cycle indicated the presence of a folded con-
former.25 These include i, i+2 NOEs from Trp2 to dCha
and a dCha to Trp5 NH-NH NOE, which suggest a turn
in the peptide centered on the Pro-dCha residues.
Other significant NOEs were observed between Trp2 RH
and the two Pro δH protons. These confirm that the
Trp-Pro bond exists in the trans conformation. The
presence of cis-Pro bonds is not uncommon in small
peptides, but in this case there is no evidence of any
minor populations containing a cis form of the Trp-Pro
bond. However, doubling of signals from residues near
the Phe suggested the presence of both cis and trans
isomers of the methyl amide bond between Phe and
D-Arg, with the trans form predominant.
NOEs detected for the trans isomer of 18 supported
the existence of a well-defined conformer and were used
to derive a set of 41 interproton distance restraints.
These restraints were used to generate a set of 50
structures using a simulated annealing algorithm within
the program XPLOR. All of these structures showed a
good convergence to a folded conformation; however
there are two distinct conformational populations as
shown in Figure 3. In the first family of 18 structures a
pairwise rmsd of 0.30 Å was obtained for all of the
backbone CR, N, C atoms, while for the second family
consisting of 11 structures there was a pairwise rmsd
of 0.25 Å. Thus, there was a good superimposition
within each family of structures.
The φ, ψ angles for the average structure generated
from the members of each of the families 18a and 18b
are given in Table 4. These data reveal that the main
differences between the two families reside in MePhe,
Trp2, and D-Arg, whereas the region Pro to Trp5 is
conserved between the two families. Analysis of the
structures revealed the formation of a hydrogen bond
between the CdO of Trp2 and NH of dCha to form a
7-membered ring associated with an inverse γ-turn.
This is consistent with hydrogen bond formation involv-
ing the dCha NH as indicated by the temperature
coefficient data. The φ, ψ angles for Pro in both families
are also consistent with the presence of an inverse
γ-turn,25 and a schematic drawing of the cycle 18 is
shown in Figure 4. The hydrogen bond associated with
the proposed inverse γ-turn is defined by the dashed
line. The inverse γ-turn identified in both structural
families of the cycle 18 had a dCha-NH‚‚‚OC-Trp
heavy atom distance of 2.77 and 2.76 Å and a N-H‚‚‚O
angle of 150° and 135° for families 18a and 18b,
NMR Solu tion Str u ctu r e of Cycle 27. 1H NMR
spectral data for cycle 27 (Table 5) in DMSO-d6 at 24
°C shows six distinct amide resonances which are
attributed to the amide NH’s of residues Phe, Orn,
dCha, Trp, and Arg and the side chain of Orn which is
3
linked to the C-terminus. The J NH-RH coupling con-
stants for the amide NH protons of cycle 27 were
measured from the 1D spectrum and are reported in
Table 5. It is unusual for backbone coupling constants
in small peptides to differ from the conformationally
averaged value of ∼7 Hz;25 therefore, it is significant
that the constants are higher for Phe (8.51 Hz) and
D-Arg (8.23 Hz) but lower for dCha (4.94 Hz) than the
conformational average. These coupling constants
strongly indicate the presence of well-defined local
structure near these residues in a significant proportion
of the solution conformations.
A temperature dependence (20-60 °C) study of the
amide NH chemical shifts was also carried out for
peptide 27, and the results are shown graphically in
Figure 2. Only one of the NH protons has chemical shift/
temperature coefficients (∆δ/T) < 3 ppb/deg, and this
is assigned to the Arg (2.72 ppb/deg). This low temper-
ature coefficient is highly significant for such a small
peptide, indicating intramolecular hydrogen bond for-
mation.
The NOE data are consistent with the cyclic 27
adopting a well-defined solution conformation. In com-
bination with the coupling constant data, several highly
characteristic medium range NOEs confirm the pres-
ence of an inverse γ-turn involving the residues dCha-
Trp-DArg. In particular, the presence of small and large
coupling constants for the two residues dCha and Arg,
respectively, together with an RH-NHꢀ NOE between
Trp5 and Orn, an RH-NH i, i+2 NOE between dCha
and Arg, and an NH-NH NOE between Trp5 and Arg
suggest an inverse γ-turn conformation. This agrees well
with the observed low temperature coefficient for the
Arg NH proton. From a set of 66 NOE-based distance
restraints and 3 dihedral angle (φ) constraints, a family
of 50 structures was calculated. The structures con-
verged with high precision, and the 20 lowest-energy
structures are shown in Figure 5. The rmsd over all
backbone atoms of the cycle is 0.35 Å reflecting good
superposition of the structures.
The φ, ψ angles for the average structure generated
from members of family 27 is given in Table 4. The

Antagonists of the C5a Receptor
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 11 1971
of the Trp residue from 4.28 to 4.70 ppm with increasing
water concentration. Given the invariance of the other
shifts and coupling constants, this most likely reflects
minor orientation changes of the Trp side chain (ex-
pected to have a large effect on the Trp RH shift due to
ring-current effects) rather than a change in the global
backbone conformation.
Su m m a r y
We have determined structure-activity relationships
for a series of cyclic and acyclic analogues of the peptide
MeFKPdChaWr, an antagonist of the C5a receptor on
human PMNs. Except for the fluorenylalanine deriva-
tives, none of the peptides tested showed any agonist
activity in causing the release of MPO from human
PMNs in the concentration range of 10-10-10-3 M.
Structure-activity data and NMR solution structures
indicate that a general turn conformation is a key
structural determinant of antagonist activity in the
peptides, a point demonstrated by the improved receptor
affinity and antagonist potency of cyclic analogues. Such
cycles stabilize a turn conformation, and those with a
ring size of 17 or 18 atoms were the most potent
compounds, with AcF[OPdChaWR] showing the best
C5a antagonist activity (IC50 ) 20 nM) for human
PMNs. Our results provide significant progress toward
the development of a pharmacophoric model for small
C5a antagonists derived from the C-terminus of C5a and
may catalyze the development of a new drug candidate
for the treatment of inflammatory diseases. As a step
toward this goal, we have recently found that one of
these potent cyclic C5a antagonists, F[OPdChaWR], is
an effective inhibitor of C5a- and endotoxin- induced
neutropenia when administered intravenously to rats.27
F igu r e 5. Backbone and proline heavy atoms of 20 lowest-
energy-minimized NMR structures of 27 in DMSO-d6 (24 °C).
Putative H-bond length (dotted line) for dCha CO‚‚‚HN Arg
is 3.13 Å.
temperature-dependence study of amide resonances
implicated the Arg NH in hydrogen bonding. Analysis
of potential hydrogen bonding in the converged struc-
tures revealed an inverse γ-turn between the Arg NH
and dCha CdO as the most reasonable donor and
acceptor residues for hydrogen bonding (H-bond shown
in Figure 5). For the inverse γ-turn hydrogen bond in
this family an Arg NH‚‚‚OdC dCha heavy atom distance
of 3.13 Å and N-H‚‚‚O angle of 140° were found, these
values being consistent with inverse γ-turn hydrogen
bonds in proteins.26
The solution structure of cycle 27 differs somewhat
from that determined for either its diastereomer AcF-
[OPdChaWr] or the linear peptide 121 and also from the
backbone cycle 18 (Figure 3) in that the inverse γ-turn
is now situated between residues dCha and Arg, instead
of between residues Orn and dCha. This new position
for the inverse γ-turn in 27 still enforces a similar
overall (turn) conformation upon the cyclic backbone as
in 29 or 1 but may orient the side chains differently.
Such subtle repositioning of the side chains accounts
for the different (increased) binding affinity and an-
tagonist potency for this cycle. While ultimately it would
be valuable to directly determine the receptor-bound
conformation to confirm this suggestion, we are cur-
rently limited to an examination of solution conforma-
tions and to the assumption that the cyclic nature of
the peptides means that receptor-bound conformations
will be similar to that determined in DMSO. The poor
solubility of 27 in water precluded a determination of
its structure in aqueous solution; however in an attempt
to assess possible contributions to side chain conforma-
tions from hydrophobic collapse in aqueous solution, we
measured chemical shifts and main-chain coupling
constants of 27 in a range of DMSO/water mixtures and
found that the backbone RH shifts of Orn, dCha, Arg,
and Phe varied by less than 0.1 ppm for DMSO/water
ratios from 100/0% to 15/85%, and the coupling con-
stants for all backbone residues varied by less than 10%
over the same range of solvent mixtures. This strongly
suggests that the backbone conformation (and by im-
plication the side chain orientations) do not vary
significantly with the solution environment. It can thus
reasonably be assumed the receptor-bound conformation
is similar to that determined above. In the DMSO/water
titrations there was a monotonic shift of the RH peak
Exp er im en ta l Section
Abbr evia tion s: TFA, trifluoroacetic acid; DIEA, diisopro-
pylethylamine; HBTU, O-benzotriazole N′,N′,N′,N′-tetram-
ethyluronium hexafluorophosphate; BOP, benzotriazol-1-
yloxytris(dimethylamino)phosphonium hexafluorophosphate;
DMF, dimethylformamide; O, L-ornithine; dCha, D-cyclohexyl-
alanine; amino acid symbols “r” and “a” represent D-Arg and
D-Ala, respectively; tR, retention time; ES-MS, electrospray
mass spectrometry.
Gen er a l Meth od s. Protected amino acids and resins were
obtained from Novabiochem. TFA, DIEA, and DMF (peptide
synthesis grade) were purchased from Auspep. All other
materials were reagent grade unless otherwise stated. Pre-
parative scale reverse-phase HPLC separations were per-
formed on a Vydac C18 reverse-phase column (2.2 × 25 cm);
analytical reverse-phase HPLC on Waters Delta-Pak PrepPak
C18 reverse-phase column (0.8 × 10 cm); using gradient
mixtures of solvent A ) water/0.1% TFA and solvent B ) water
10%/acetonitrile 90%/0.09% TFA. The molecular weight of the
peptides was determined by electrospray mass spectrometry
recorded on a triple quadrupole mass spectrometer (PE SCIEX
API III) as described elsewhere.28 1H NMR spectra were
recorded on a Bruker 500 or 750 NMR spectrometer at 24 °C.
Proton assignments were determined by 2D NMR experiments
(DFCOSY, TOCSY, NOESY). All assay data were analyzed by
nonlinear regression and statistics using one-way ANOVA
followed by a Newman-Keuls multiple comparisons test.
P ep tid e Syn th esis. The linear peptides 1-16 were as-
sembled by manual stepwise solid-phase peptide synthesis
using HBTU activation and DIPEA in situ neutralization.
Couplings were monitored by the quantitative ninhydrin test.
Boc chemistry was employed for temporary NR-protection of
amino acids with two 1-min treatments with TFA for Boc
group removal. Peptides were synthesized on a Novabiochem

1972 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 11
Finch et al.
Ta ble 6. Peptide Retention Times (tR) and Mass Spectral
Ta ble 7. Characterization Data (rp-HPLC Retention Times
and ES-MS) for Cyclic Peptidesa
Dataa
MH+
MH+
exptl
tR
peptide
sequence
MeFKPdChaWr
MeFKPdChaWR
FKPdChaWr
AKPdChaWr
AcFKPdChaWr
KFKPdChaWr
FKFKPdChaWr
C5a12-20-ahx-FKPdChaWr
C5a12-20-ahx-GGGGFKPdChaWr 28.8 2325.4 2326.2
C5a40-68-FKPdChaWr
AcFOPdChaWr
AcF(Dap)PdChaWr
AcFNlePdChaWr
FAPdChaWr
(min) exptl calcd
peptide
sequence
tR (min)
calcd
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16a
16b
32.0 900.5 900.7
33.4 900.7 900.7
33.5 886.7 886.5
28.8 810.5 810.6
31.8 928.5 928.7
28.7 1014.6 1015.0
30.5 1161.7 1162.0
34.4 2097.3 2097.6
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
[FWPdChaWr]
Me[FWPdChaWr]
F[KPdChaWr]
F[KPdChaWR]
F[OPdChaWr]
41.6
42.2
35.3
33.6
36.1
33.8
32.6
33.6
35.4
32.9
39.8
44.0
43.7
42.0
41.2
926.5
940.5
868.5
868.5
854.5
854.5
840.5
840.5
826.5
826.5
896.5
910.5
896.5
882.5
868.5
926.5
940.5
868.5
868.5
854.5
854.5
840.5
840.5
826.5
826.5
896.5
910.5
896.5
882.5
868.5
F[OPdChaWR]
F[DabPdChaWr]
F[DabPdChaWR]
F[DapPdChaWr]
F[DapPdChaWR]
AcF[OPdChaWR]
AcF[KPdChaWr]
AcF[OPdChaWr]
AcF[DabPdChaWr]
AcF[DapPdChaWr]
31.7 4206.1 4205.1
38.4 914.5 914.7
39.8 886.5 886.7
36.2 913.5 913.1
33.2 829.5 829.7
35.4 944.5 944.7
44.7 977.5 977.7
47.2 977.5 977.7
FWPdChaWr
AcFKPdChaZr (I)
AcFKPdChaZr (II)
a
Dap, L-2,3-diaminopropionic acid; Dab, L-2,4-diaminobutanoic
acid; O, L-ornithine; [], cyclization between the ends of the
bracketed residues.
a
Dap, L-2,3-diaminopropanoic acid; Z, fluorenylalanine which
was synthesized as a mixture of enantiomers and was introduced
into the peptide during peptide synthesis as a mixture. The two
peptide diastereomers were subsequently separated using rp-
HPLC and are arbitrarily denoted as isomers I and II.
and the Fmoc protecting group removed by treating with 30%
piperidine/DMF (4 mL) for 1 h. The solvents were removed in
vacuo, and the cyclic peptide was purified by rp-HPLC (4.8
mg, 22%): tR ) 33.6 min; MS [M + H]+ calcd ) 868.5, [M +
H]+ exptl ) 868.5.
Boc-D-Arg(Tos)-PAM or Boc-L-Arg(Tos)-PAM resin with
a
substitution value of approximately 0.5 mmol/g. The peptides
were fully deprotected and cleaved by treatment with liquid
HF (10 mL), p-cresol (1 mL) at -5 °C for 1-2 h. Analytical
HPLC was used for peptide purification (gradient 0 to 75% B
over 60 min). Peptide retention times (tR) and mass spectral
data are listed in Table 6.
Syn th esis of Cycle AcF [OP d Ch a WR] (27). The linear
peptide Ac-Phe-Orn-Pro-dCha-Trp-Arg was synthesized by Boc
chemistry on a 0.20-mmol scale using HBTU/DIEA activation
and in situ neutralization on a Boc-L-Arg(Tos)-PAM resin (338
mg, SV ) 0.591 mmol/g). Cleavage and deprotection of the
resin (457 mg) were achieved by treating the resin with HF
(10 mL) and p-cresol (1 mL) at -5 to 0 °C for 1-2 h, to give
crude peptide (160 mg, 90%). Cyclization involved stirring the
crude peptide (41 mg, 45 µmol), BOP (126 mg, 0.28 mmol),
and DIEA (158 µL, 0.9 mmol) in DMF (57 mL) for 15 h. The
solvent was removed in vacuo and the cyclic peptide purified
by rp-HPLC (18.8 mg, 47%): tR ) 43.7 min; MS [M + H]+ calcd
) 896.5, [M + H]+ exptl ) 896.5.
P u r ifica tion a n d Ch a r a cter iza tion . Crude peptides were
purified using preparative rp-HPLC using a Vydac C18
reverse-phase column (2.2 × 25 cm). Gradients of 1 mL/min
of solvent A to solvent B were employed and monitored at 214
nm. Fractions were collected and tested by electrospray mass
spectrometry (ES-MS) for the correct molecular weight, and
purity was checked by analytical rp-HPLC on a Waters Delta-
Pak PrepPak C18 reverse-phase column (0.8 × 10 cm) (gradi-
ent 0 to 75% B over 60 min). The acetonitrile was HPLC grade
(BDH Laboratories), and TFA was synthesis grade (Auspep).
Characterization data (rpHPLC retention times and ES-MS)
for cyclic peptides are shown in Table 7.
P ep tid e Cycliza tion . The general procedure for cyclization
of linear peptides involved dissolving the peptide (1 equiv) and
BOP (5 equiv) in DMF (10 mM peptide concentration) and
stirring vigorously, followed by the addition of DIEA (15 equiv).
Solutions were generally allowed to stir at room temperature
overnight, although in most cases the reaction was complete
within 2 h. DMF was removed under high vacuum at 30 °C
on a rotary evaporator and then purified by rp-HPLC. For
cyclic peptides containing a free N-terminus, a Fmoc group
was used as the temporary N-terminal protecting group during
the cyclization step. DMF was removed under high vacuum
at 30 °C on a rotary evaporator, and then the peptide was
treated with 30% piperidine/DMF for 1 h at room temperature
to remove the Fmoc group. This was followed by solvent
removal under high vacuum and purification by rp-HPLC.
Representative examples of the various methods of synthesis
of the cycles follows.
Syn th esis of Cycle Me[F WP d Ch a Wr ] (18). The linear
peptide MeF-W-P-dCha-W-r was synthesized by Boc chemistry
on a 0.1-mmol scale using HBTU/DIEA activation and in situ
neutralization on a Boc-D-Arg(Tos)-PAM resin (169 mg, SV )
0.77 mmol/g). Cleavage and deprotection of the resin (97 mg)
were achieved by treating the resin with HF (10 mL) and
p-cresol (1 mL) at -5 to 0 °C for 1-2 h, and the crude peptide
was purified by rp-HPLC (13.7 mg, 38%). Cyclization involved
stirring the crude peptide (7.25 mg, 7.6 µmol), BOP (12 mg,
26.6 µmol), and DIEA (13 µL, 0.08 mmol) in DMF (18 mL) for
15 h. The solvent was removed in vacuo and the cyclic peptide
purified by rp-HPLC (1.92 mg, 27%): tR ) 42.2 min; MS [M +
H]+ calcd ) 940.5, [M + H]+ exptl ) 940.7.
NMR Str u ctu r e Deter m in a tion . 1H NMR spectra were
recorded for 3 mg of 18 and 27 in 550 µL DMSO-d6 and
chemical shifts were referenced to the solvent signal. 2D 1H
NMR NOESY (relaxation delay 2.0 s, mix time 150-300 ms),
DQF-COSY, and TOCSY (mixing time 80 ms) experiments
were recorded in phase-sensitive mode. Acquisition time )
0.170 s, spectral width ) 12 000 Hz, with F1 ) 4096 and F2
)
700.
NMR data were processed using UXNMR software (Bruker)
on a Silicon Graphics Indy workstation. 2D NOE cross-peaks
were characterized into strong (1.8-2.7 Å), medium (1.8-3.5
Å), and weak (1.8-5.0 Å) based on their intensities. Backbone
Syn th esis of Cycle F [KP d Ch a WR] (20). The linear
peptide Fmoc-Phe-Lys-Pro-dCha-Trp-Arg was synthesized by
Boc chemistry on a 0.05-mmol scale using HBTU/DIEA activa-
tion on a Boc-L-Arg(Tos)-PAM resin (114 mg, SV ) 0.44 mmol/
g). Cleavage and deprotection of the resin (93 mg) were
achieved by treating with HF (10 mL) and p-cresol (1 mL) at
-5 to 0 °C for 1-2 h to give the crude peptide (28 mg, 94%).
Cyclization involved stirring the crude peptide (28 mg, 25
µmol), BOP (49 mg, 0.11 mmol), and DIEA (38 µL, 0.2 mmol)
in DMF (22 mL) for 15 h. The solvent was removed in vacuo
3
dihedral restraints were inferred from J NH-HR coupling con-
3
stants with φ restrained to -120 ( 40° for J NH-HR > 8 Hz,
-65 ( 25° for J < 5 Hz29 for L-amino acids, and +120 ( 40°
3
for J NH-HR > 8 Hz in dCha. Three-dimensional structures were
calculated using a simulated annealing and energy minimiza-
tion protocol in the program XPLOR 3.1.30 In the first step an
ab initio simulated annealing protocol31 was used, starting
from template structures with randomized φ and ψ angles and
extended side chains, to generate a set of 50 structures. The

Antagonists of the C5a Receptor
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 11 1973
(6) Bubeck, P.; Grotzinger, J .; Winkler, M.; Kohl, J .; Wollmer, A.;
Klos, A.; Bautsch, W. Site-specific mutagenesis of residues in
the human C5a anaphylatoxin which are involved in possible
interaction with the C5a receptor. Eur. J . Biochem. 1994, 219,
897-904.
(7) Federwisch, M.; Wollmer, A.; Emde, M.; Stuhmer, T.; Melcher,
T.; Klos, A.; Kohl, J .; Bautsch, W. Tryptophan mutants of human
C5a anaphylatoxin: a fluorescence anisotropy decay and energy
transfer study. Biophys. Chem. 1993, 46, 237-48.
(8) Toth, M. J .; Huwyler, L.; Boyar, W. C.; Braunwalder, A. F.;
Yarwood, D.; Hadala, J .; Haston, W. O.; Sills, M. A.; Seligmann,
B.; Galakatos, N. The pharmacophore of the human C5a ana-
phylatoxin. Protein Sci. 1994, 3, 1159-68.
(9) Chenoweth, D. E.; Hugli, T. E. Human C5a and C5a analogues
as probes of the neutrophil C5a receptor. Mol. Immunol. 1980,
17, 151-61.
(10) Siciliano, S. J .; Rollins, T. E.; DeMartino, J .; Konteatis, Z.;
Malkowitz, L.; VanRiper, G.; Bondy, S.; Rosen, H.; Springer, M.
S. Two-site binding of C5a by its receptor: an alternative binding
paradigm for G protein-coupled receptors. Proc. Natl. Acad. Sci.
U.S.A. 1994, 91, 1214-8.
(11) DeMartino, J . A.; Van Riper, G.; Siciliano, S. J .; Molineaux, C.
J .; Konteatis, Z. D.; Rosen, H.; Springer, M. S. The amino
terminus of the human C5a receptor is required for high affinity
C5a binding and for receptor activation by C5a but not C5a
analogues. J . Biol. Chem. 1994, 269, 14446-50.
(12) Kawai, M.; Quincy, D. A.; Lane, B.; Mollison, K. W.; Luly, J . R.;
Carter, G. W. Identification and synthesis of a receptor binding
site of human anaphylatoxin C5a. J . Med. Chem. 1991, 34,
2068-71.
(13) Kawai, M.; Quincy, D. A.; Lane, B.; Mollison, K. W.; Or, Y.-S.;
Luly, J . R.; Carter, G. W. Structure-function studies in a series
of carboxyl-terminal octapeptide analogues of anaphylatoxin
C5a. J . Med. Chem. 1992, 35, 220-3.
simulated annealing protocol consisted of 20 ps of high-
temperature molecular dynamics (1000 K) with a low weight-
ing on the repel force constant and NOE restraints. This was
followed for a further 10 ps with an increased force constant
on the experimental NOE restraints. The dihedral force
constant was increased prior to cooling the system to 300 K
and increasing the repel force constant over 15 ps of dynamics.
Refinement of these structures was achieved using the con-
jugate gradient Powell algorithm with 1000 cycles of energy
minimization32 and a refined force field based on the program
CHARM.33 Structures were displayed using INSIGHT (Biosym
Technologies, San Diego, CA). All structural analysis described
in the text was done using structures calculated without
explicit H-bond restraints. The final structures were examined
to obtain pairwise rms differences over the backbone heavy
atoms (N, CR, and C).
Recep tor Bin d in g Assa y. Assays were performed with
fresh human PMNs, isolated as previously described,18 and a
buffer of 50 mM HEPES, 1 mM CaCl2, 5 mM MgCl2, 0.5%
bovine serum albumin, 0.1% bacitracin, and 100 µM PMSF
(phenylmethanesulfonyl fluoride). Buffer, unlabeled human
recombinant C5a (Sigma) or test peptide, [125I]C5a (∼50 pM)
(New England Nuclear, MA), and PMNs (0.2 × 106) were
added sequentially to a Millipore Multiscreen assay plate.
After incubation for 60 min at 4 °C, the samples were filtered
and the plate was washed once with buffer (100 µL). Filters
were dried, punched, and counted in an LKB γ-counter.
Nonspecific binding was assessed by the inclusion of 1 mM
peptide or 100 nM C5a which typically resulted in 10-15%
total binding.
(14) Mollison, K. W.; Krause, R. A.; Fey, T. A.; Miller, L.; Wiedeman,
P. E.; Kawai, M.; Lane, B.; Luly, J . R.; Carter, G. W. Hexapeptide
analogues of C5a anaphylatoxin reveal heterogeneous neutrophil
agonism/antagonism. FASEB J . 1992, 6, A2058.
(15) Ember, J . A.; Sanderson, S. D.; Taylor, S. M.; Kawahara, M.;
Hugli, T. E. Biologic activity of synthetic analogues of C5a
anaphylatoxin. J . Immunol. 1992, 148, 3165-73.
(16) Sanderson, S. D.; Ember, J . A.; Kirnarsky, L.; Sherman, S. A.;
Finch, A. M.; Taylor, S. M. Decapeptide agonists of human C5a:
the relationship between conformation and spasmogenic and
platelet aggregatory activities. J . Med. Chem. 1994, 37, 3171-
80.
(17) Sanderson, S. D.; Kirnarsky, L.; Sherman, S. A.; Vogen, S. M.;
Prakesh, O.; Ember, J . A.; Finch, A. M.; Taylor, S. M. Decapep-
tide agonists of human C5a: the relationship between conforma-
tion and neutrophil response. J . Med. Chem. 1995, 38, 3669-
75.
(18) Finch, A. M.; Vogen, S. M.; Sherman, S. A.; Kirnarsky, L.; Taylor,
S. M.; Sanderson, S. D. Biological active conformer of the effector
region of human C5a and modulatory effects of N-terminal
receptor binding determinants on activity. J . Med. Chem. 1997,
40, 877-84.
(19) Drapeau, G.; Brochu, S.; Godin, D.; Levesque, L.; Rioux, F.;
Marceau, F. Synthetic C5a receptor agonists. Pharmacology,
metabolism and in vivo cardiovascular and hematologic effects.
Biochem. Pharmacol. 1993, 45, 1289-99.
An ta gon ism Assa y. Antagonist assays were assessed by
monitoring myeloperoxidase release as follows. Cells were
isolated as previously described18 and incubated with cytocha-
lasin B (10 µg/mL, 10 min, 37 °C). Hank’s balanced salt
solution containing 0.1% gelatin and test peptide was added
onto a 96-well plate (total volume 100 µL/well), followed by
25 µL of cells (4 × 106/mL). To assess the capacity of each
peptide to antagonize C5a, cells were incubated for 10 min at
37 °C with each peptide, followed by addition of C5a (100 nM)
and further incubated for 10 min. Then 50 µL of phosphate
buffer (0.1 M, pH 6.8) was added to each well, followed by the
addition of 25 µL of a fresh 1:1 mixture of dimethoxybenzidine
(5.7 mg/mL) and H2O2 (0.51%). The reaction was stopped at
20 min by addition of 2% sodium azide (25 µL). Absorbances
were measured at 450 nm in a Rainbow plate reader and
corrected for control values (no peptide).
Ack n ow led gm en t. We thank Trudy Bond in the
Centre for amino acid analyses, the Australian Research
Council for a Senior Research Fellowship (D.J .C.), and
the National Health and Medical Research Council for
grant support.
(20) Konteatis, Z. D.; Siciliano, S. J .; Van Riper, G.; Molineaux, C.
J .; Pandya, S.; Fischer, P.; Rosen, H.; Mumford, R. A.; Springer,
M. S. Development of C5a receptor antagonists. Differential loss
of functional responses. J . Immunol. 1994, 153, 4200-4.
(21) Wong, A. K.; Finch, A. M.; Pierens, G. K.; Craik, D.; Taylor, S.
M.; Fairlie, D. P. Molecular probes for G-protein coupled
receptors. Conformationally constrained antagonists derived
from the C-terminus of human plasma protein C5a. J . Med.
Chem. 1998, 41, 3417-22.
(22) Zhang, X.; Boyar, W.; Galakatos, N.; Gonnella, N. C. Solution
structure of a unique C5a semi-synthetic antagonist: implica-
tions in receptor binding. Protein Sci. 1997, 6, 65-72.
(23) Pellas, T. C.; Boyar, W.; van Oostrum, J .; Wasvary, J .; Fryer, L.
R.; Pastor, G.; Sills, M.; Braunwalder, A.; Yarwood, D. R.;
Kramer, R.; Kimble, E.; Hadala, J .; Haston, W.; Moreira-
Ludewig, R.; Uziel-Fusi, S.; Peters, P.; Bill, K.; Wennogle, L. P.
Novel C5a receptor antagonists regulate neutrophil functions
in vitro and in vivo. J . Immunol. 1998, 160, 5616-21.
(24) Kessler, H. Conformation and biological activity of cyclic pep-
tides. Angew Chem., Int. Ed. 1982, 21, 512-23.
Refer en ces
(1) J ose, P. J .; Moss, I. K.; Maini, R. N.; Williams, T. J . Measurement
of the chemotactic complement fragment C5a in rheumatoid
synovial fluids by radioimmunoassay: role of C5a in the acute
inflammatory phase. Ann. Rheum. Dis. 1990, 49, 747-52.
(2) Valazquez, P.; Cribbs, D. H.; Poulos, T. L.; Tenner, A. J .
Aspartate residue 7 in amyloid beta protein is critical for
classical complement pathway activation: implications for alzhe-
imers disease pathogenesis. Nature Med. 1997 3, 77-9.
(3) Engler, R. L.; Roth, D. M.; del Balzo, U.; Ito, B. R. Intracoronary
C5a induces myocardial ischemia by mechanisms independent
of the neutrophil: leukocyte filters desensitize the myocardium
to C5a. FASEB J . 1991, 5, 2983-91.
(4) Hammerschmidt, D. E.; Weaver, L. J .; Hudson, L. D.; Craddock,
P. R.; J acob, H. S. Association of complement activation and
elevated plasma C5a with adult respiratory distress syndrome.
Pathophysiological relevance and possible prognostic value.
Lancet 1980, 1, 947-9.
(5) Mollison, K. W.; Mandecki, W.; Zuiderweg, E. R.; Fayer, L.; Fey,
T. A.; Krause, R. A.; Conway, R. G.; Miller, L.; Edalji, R. P.;
Shallcross, M. A.; et al. Identification of receptor-binding resi-
dues in the inflammatory complement protein C5a by site-
directed mutagenesis. Proc. Natl. Acad. Sci. U.S.A. 1989, 86,
292-6.
(25) Dyson, H. J .; Wright, P. E. Defining solution conformations of
small linear peptides. Annu. Rev. Biophys. Chem. 1991, 20, 519-
38.
(26) Rose, G. D.; Gierasch, L. M.; Smith, J . A. Turns in peptides and
proteins. Adv. Protein Chem. 1985, 37, 1-109.

1974 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 11
Finch et al.
(27) Short, A.; Wong, A. K.; Finch, A. M.; Haaima, G.; Shiels, I. A.;
Fairlie, D. P.; Taylor, S. M. Effects of a new C5a receptor
antagonist on C5a- and endotoxin-induced neutropenia in the
rat. Br. J . Pharmacol. 1999, 126, 551-4.
(28) Abbenante, G.; Fairlie, D. P.; Gahan, L. R.; Hanson, G. R.;
Pierens, G. K.; van den Brenk, A. L. Conformational control by
thiazole and oxazoline rings in cyclic octapeptides of marine
origin. Novel macrocyclic chair and boat conformations. J . Am.
Chem. Soc. 1996, 118, 10384-8.
(31) Nilges, M.; Gronenborn, A. M.; Bru¨nger, A. T.; Clore, G. M.
Determination of three-dimensional structures of proteins by
simulated annealing with interproton distance restraints. Ap-
plication to crambin, potato carboxypeptidase inhibitor and
barley serine proteinase inhibitor 2. Protein Eng. 1988, 2, 27-
38.
(32) Clore, G. M.; Bru¨nger, A. T.; Karplus, M.; Gronenborn, A. M.
Application of molecular dynamics with interproton distance
restraints to three-dimensional protein structure determination.
A model study of crambin. J . Mol. Biol. 1986, 191, 523-51.
(33) Brooks, B. R.; Bruccoleri, R. E.; Olafson, B. D.; States, D. J .;
Swaminathan, S.; Karplus, M. CHARMM: A program for
macromolecular energy minimization and dynamics calculations.
J . Comput. Chem. 1983, 4, 187-217.
(29) Pardi, A.; Billeter, M.; Wu¨thrich, K. Calibration of the angular
dependence of the amide proton-C alpha proton coupling con-
stants, 3J HN alpha, in a globular protein. Use of 3J HN alpha
for identification of helical secondary structure. J . Mol. Biol.
1984, 180, 741-51.
(30) Bru¨nger, A. T. X-PLOR Manual Version 3.1; Yale University:
New Haven, CT, 1992.
J M9806594