Design, structure and biological activity of b-turn peptides
of CD2 protein for inhibition of T-cell adhesion
Liu Jining
1
, Irwan Makagiansar
3
, Helena Yusuf-Makagiansar
3
, Vincent T. K. Chow
2
, Teruna J. Siahaan
3
and Seetharama D. S. Jois
1
1
Department of Pharmacy and
2
Department of Microbiology, National University of Singapore, Singapore;
3
Department of
Pharmaceutical Chemistry, The University of Kansas, Lawrence, KS, USA
The interaction between cell-adhesion molecules CD2 and
CD58 is critical for an immune response. Modulation or
inhibition of these i nteractions has been shown t o be thera-
peutically useful. Synthetic 12-mer linear and cyclic pe ptides,
and cyclic hexapeptides based on rat CD2 protein, were
designed to modulate CD2–CD58 interaction. The synthetic
peptides effectively blocked the interaction between CD2–
CD58 proteins as demonstrated by antibody binding,
E-rosetting and heterotypic adhesion assays. NMR and
molecular modeling studies indicated that the synthetic

and k
on
that
supports dynamic binding with rapid counter-receptor
exchange. This creates an optimal intercellular membrane
distance (% 135 A
˚
) on opposing cell surfaces suitable for
TcR-pMHC or NK receptor–MHC interactions to foster
immune recognition. Hence, in the presence of human
CD2–CD58 interaction, T-cells recognize the correct
pMHC with a 50- to 100-fold greater efficiency than its
absence [4]. In addition, endothelial cells (EC) in rheuma-
toid arthritis (RA) have been shown to express elevated
levels of CD58, and RA lymphocytes in synovial fluid
(SF) express increased levels of CD2 and CD58 relative
to RA or normal peripheral blood lymphocytes [5,6].
Thus, the inhibition of CD2–CD58 interaction can
potentially be used for t he treatment o f autoimmune
diseases.
It has been shown that b lockade of t he CD2–CD58
interaction [7,8] and/or modulation of the CD2 costim-
ulatory pathway [9–12] can result in prolonged tolerance
towards allografts. The soluble CD58–Ig fusion protein
Amevive (LFA3TIP) has been used to treat psoriasis
[13]. However, antibodies are huge protein molecules and
therapeutic antibodies are nonhuman in origin, these
often elicit significant side-effects attributed to th eir
immunogenicity. The humanized versions of antibodies
BTI-322 [14] and MEDI-507 [15] have been tested for the

inhibition assays. In order to understand structure–func-
tion relationship of peptides, we also carried out detailed
NMR, molecular modeling and docking studies of pep-
tide-protein complexes. Our results indicate that the
designed peptides are useful for inhibition of the T-cell
adhesion mechanism.
Materials and methods
Peptides
The linear and cyclic peptides lER, lVY, cER, cVY, cEL
and cYT (Table 1) were designed and purchased from
Multiple Peptide Systems (San Diego, CA, USA). The pure
product was analyzed by HPLC and fast atom bombard-
ment mass spectrometry (FABMS). The HPLC chromato-
gram showed that the purities of peptides were more than
90%, and FABMS showed the correct molecular ion for the
peptides. The control peptide was synthesized u sing auto-
matic solid-phase peptide s ynthesizer (Pioneer, Perspective
Biosystem, Foster, CA, USA) using Fmoc chemistry with
PAL resin. T he Fmoc-protected amino acids were obtained
from Novabiochem. All the solvents used in the Pioneer
peptide synthesizer were obtained from Applied Biosystems.
Peptide was purified by preparative HPLC (Waters 600
HPLC system), on a reversed-phase C18 column (Inertsil,
10 · 250 mm, 5 lm, 300 A
˚
) w ith a linear gradient o f
solvent A (0.1% (v/v) trifluoroacetic acid/H
2
O) and solvent
B (0.1% (v/v) trifluoroacetic acid/acetonitrile). The peptides

of
penicillin/streptomycin. Caco-2 cells were used between
passages 5 0 and 60. Sheep blood in Alsever’s solution was
purchased from TCS Biosciences Ltd., Singapore.
CD2 detection and flow-cytometry assay
To detect CD2 expression, 10
6
Jurkat cells were washed
with 0.5% (w/v) BSA/NaCl/P
i
, a nd incubated with FITC-
CD2 monoclonal antibody (mAb) for 1 h at 37 °C. After
washing three times with 0.5% (w/v) BSA/10 m
M
Hepes/
NaCl/P
i
, the cells were fixed using 1% (v/v) paraformalde-
hyde/NaCl/P
i
andanalyzedwithaflowcytometer(FAC-
Scan apparatus, Becton Dickinson) equipped with the
CELL
QUEST
software. Ten thousand c ells were counted for every
sample during acquisition.
Inhibition of antibody binding
MOLT-3 cells were grown and activated with 0.2 l
M
of

i
, the cells were fixed using 2% (v/v) paraformal-
dehyde/NaCl/P
i
and analyzed with a flow cytometer
(FACScan, Becton Dickinson) equipped with
CELL QUEST
software. Binding of FITC-anti-CD58, following incuba-
tion with F
c
blocker (Biodesign International) was used
as a positive control. Median values of fluorescence
intensity were taken as the binding intensities. As many
as 10 000 cells were counted for e very sample during
acquisition, and each experiment was performed in
triplicate. The control histogram (cells without peptide
treatment) was placed within 100–101 on the log scale of
FL1-Height by adjusting the FL1 detector. The data
were represented as their relative inhibition or enhance-
ment to the positive control.
Table 1. Peptides used in this study that are derived from rat CD2
protein. The sequence number refers to the residues from th e second
position in the peptide to eleventh position. Pen1 an d Cys12 were
introduced for cyclization p urpose.
Code Name
Sequence number
in the native protein
lER PenERGSTLVAEFC 36–45
cER Cyclo(1,12) PenERGSTLVAEFC 36–45
lVY PenVYSTNGTRILC 85–94

Jurkat cell suspension (2· 10
6
per mL) was added to the
mixture, and incubated for another 15 min. The cells were
pelleted by centrifugation (200 g,5min,4°C) and then
incubated a t 4 °C f or 1 h. The cell pellet w as gently
resuspended, and the E-rosettes counted with a haemo-
cytometer [16]. Cells with five or more SRBCs bound
were counted as rosettes. At least 200 cells were counted
to determine t he percentage of E-rosette cells. The
inhibitory activity was calculated by the following Eqn (1):
inhibition ð%Þ
¼½ðnegative E- rosette %
peptid e
À negative E-rosette %
blank
Þ=
E-rosette %
blank
Â100 ð1Þ
where, negative E-rosette %
peptide
¼ (Jurkat cells without
formation of E-rosettes/total Jurkat cells) · 100%.
Lymphocyte-epithelial adhesion assay
Caco-2 cells were used between passages 50 and 60 and w ere
plated onto 48-well plates at % 2 · 10
4
cellsÆwell
)1

amino acid residue peptide sequences are
underlined.
Ó FEBS 2004 Design of peptides for T-cell adhesion inhibition (Eur. J. Biochem. 271) 2875
transferred to 96-well plates for reading with a microplate
fluorescence analyzer.
Cell viability assay
Peptides which exhibited effects on Jurkat/Caco-2 adher-
ence were tested by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphe-
nyl-tetrazolium bromide (MTT) assay [17] to determine if
their effects were d ue to frank toxicity. A final concentration
of 180 l
M
peptide was added to Caco-2 or Jurkat cells for 1
or 2 h, which is the time of exposure of Caco-2/Jurkat cells
during the adherence assay. The cell viabilities were
validated by incubating with 5 mgÆmL
)1
MTT at 37 °C
for 3 h. The MTT-labeled cells were lysed by d imethylsulf-
oxide and the absorbance was measured with a microplate
reader at a wavelength of 570 nm.
NMR spectroscopy
The samples for the NMR spectra of the peptide were
prepared by dissolving 3 mg of the peptides in 0.5 mL of
90% H
2
O/10% D
2
O. For pH titration experim ents, the pH
of the solution was varied by the a ddition of DCl or NaOD

(i,i+1)andd
NN
(i, i +1), 1.9–3.0, 2.2–3.6 and
3.0–5.0 A
˚
, respectively [23]. Side chain protons were not
stereospecifically assigned; hence, ROE/NOE restraints for
the side chain protons were calculated by considering
pseudoatoms [23].
Determination of peptide structures
Conformational space was searched for the peptides using
the
DISCOVER
program version 2000 (Accelrys Inc., San
Diego, CA, USA) to identify conformations consistent
with the experimental ROE and coupling constant data
[24,25]. Briefly, the linear peptide was subjected to MD
simulations in vacuo at 300K with ROE and disulfide bond
constraints [26]. The resulting structure w as cyclized by
forming disulfide bo nds. The cyclic structure obtained was
slowly heated t o 900 K in steps of 100 K d ynamics for 5 ps
duration at each step. At 900 K, MD simulations were
performed for 20 ps to explore several possible conforma-
tions that the peptide can acquire. The trajectory from high
temperature dynamics was analyzed for similarities be-
tween the structures by comparing the root mean square
deviations (rmsd) between each possible pair of structures,
and was divided into several conformational families. The
average structure was taken from each family and chosen
as starting structures for the calculation of corrected

a
protons were used for calibration. After high temperature
dynamics with NOE constraints the folded peptide was
cyclized by p eptide bond to arrive at th e starting s tructures
for cyclic hexapeptides. In the case of ROESY spectrum
for 12 amino acid residue cyclic peptides (12-mers), the
corrected interproton distances were used for subsequent
calculation of the structure. Each structure obtained during
high temperature dynamics was then slowly cooled down
to 400 K. Each structure was then soaked with water
molecules, followed by MD simulations at 300 K with all
the ROE/NOE constraints. These structures were further
energy minimized w ith solvent molecule s using the steepest
descents and conjugate gradient methods until the rms
derivative was 0.3 kcalÆmol
)1
ÆA
˚
)1
. The resulting structures
were analyzed again by
MARDIGRAS
, and the final struc-
tures chosen wh en two criteria w ere fulfilled: (a) t he
conformation of backbone had an interproton error of less
than 0.2 A
˚
compared to upper and lower boundaries of
distances from ROE/NOE data and (b) the conformation
had / angles within 30° of the calculated /-values from

tal volumes were assigned to the protein atoms by the
auxiliary program,
ADDSOL
. Affinity grid files were gener-
ated using the other auxiliary program,
AUTOGRID
.The
dimension of the grid box was chosen to cover the whole
protein with grid-point spacing of 0.375 A
˚
and centered at
the positions describe below. As there are two major cavities
in the top and bottom of hCD58 besides the binding sites in
the hCD2–hCD58 complex, the starting positions of
peptides were generated at three sites on CD58 protein
surface (Fig. 2A). The parameters were set as the default
values of the
AUTODOCK
Lamarckian genetic algorithm.
First, a randomized rigid docking (blind docking) was
performed and the conformers with lowest energy or in
significant clusters were chosen to perform further docking
studies with flexible docking.
During flexible docking, the dihedrals of backbone of the
ligand were kept rigid, whereas the dihedrals of side chain
were allowed to rotate. After docking, all structures
generated were assigned to clusters based on a tolerance
of 1 A
˚
all-atom rmsd from the lowest energy structure. The

most widely used method to identify T-cells by CD2–
CD58 interaction. SRBCs express CD58 protein, while
Jurkat leukemic T -cells express CD2 protein o n t heir
surface. The ability of Jurkat cells to express CD2 was
Fig. 2. Ribbon diagram of crystal structure o f CD2–CD58 complex and crystal structure of rat CD2. (A) Ribbon diagram of crystal structure of
CD2–CD58 (LFA-3) complex. Starting positions of peptides for docking s tudies are shown in the figure. The residues of hCD 2 t hat are in b-turn
region are shown as red sticks. Tyr86 from CD2 is shown in green. Residues from C D58 that are important in the interaction of CD2–CD58 are
shown in t he following colors: Lys32, Glu25 (purple); Asp33, Lys29, Glu37 (blue); Lys30 (magenta). Residues, Asp33 and Lys29 were shown to be
important in binding to peptides from CD 2 in docking stud ies. (B) Crystal structure of ratCD2. Residues in the b-turn re gion are shown as sticks
and labeled.
Ó FEBS 2004 Design of peptides for T-cell adhesion inhibition (Eur. J. Biochem. 271) 2877
measured by flow-cytometry assay. Binding of Jurkat cells
to SRBCs due to CD2 and CD58 interaction results in the
formation of E-rosettes. The ability of each of the
designed CD2 peptides to inhibit CD2–CD58 interaction
was evaluated by inhibition of E-rosette formation
between Jurkat cells and SRBCs. As depicted in Fig. 4,
the CD2 peptides showed 30–40% inhibitory activity at
200 l
M
. When the concentration of the peptide was
decreased, the inhibitory activity of the peptide was
correspondingly decreased. Even at 50 l
M
, peptide cVY
displayed nearly 30% activity. Among the four peptides
(12-mers) studied, cVY showed the highest inhibitory
activity of 40% at 100 l
M
concentration. Both linear and

peptide showed less than 5% inhibitory activity at three
different concentrations studied. These peptides were also
tested for their toxicity using the MTT assay [17]. All the
four peptides tested in the study resulted in 90–100%
viability indicating that these peptides were not toxic to cells
and the inhibition data observed were not due to changes in
the cells arising from peptide toxicity.
Cyclic hexapeptides. In order to understand the amino
acid residues i nvolved in the biological activity and to study
the effect of reducing the chain length of peptides on
Fig. 3. Inhibition or enhancement of FITC-labeled CD58-antibody
binding to MOLT-3 cells by synthetic peptides derived from CD2
examined by FACS. MOLT-3 cells were activated by 0.2 l
M
PMA to
induce CD58 expression. FITC-anti-CD58 w as added to t he peptide-
treated c ells, followed b y a further incubation. B inding of FITC-anti-
CD58, following incubation with F
c
blocker was used as a positive
control. Median values of fluorescence intensity were taken as the
binding intensities. As many as 10
4
cells were counted for every sample
during acquisition. The control histogram (cells without peptide
treatment) was placed within 100–101 on the log scale of FL1-height.
The data were r epresented as their relative inhibition or enhancement
to the p ositive control. Each data point repre sents the mean of tripli-
cate assay at different p eptide concen tration (lM).
Fig. 4. In hibition of E-rosette form atio n by syn thet ic peptides derived

,anincreasein
inhibitory activity compared to linear and c yclic 12-mer ER
peptides. However, the VY cyclic hexapeptide (cYT) lost
activity upon truncation. Similar trends were observed in
the heterotypic adhesion assay of cyclic hexapeptides
(Fig. 5). The cEL peptide showed increased activity (50%
at 90 l
M
), whereas cYT showed drastically diminished
inhibitory activity compared to 12-mer VY peptides.
NMR structure determination
The three-dimensional structures of the cyclic peptide were
determined based on the NMR data of the cyclic peptides.
The one dimensional
1
H NMR spectrum of the peptides
cVY, cER, cEL and cYT showed good dispersion of the
chemical shifts and the coupling patterns, indicative of a
stable major conformer at the experimental temperature.
The structure of peptide cER. NMR data of cER
indicated the possibility of the b-turn structure in peptide
cER. The d
NN
(i, i +1) cross peaks between Gly4-Ser5 and
the stronger d
aN
(i, i +1) cross peaks between Arg3-Gly4
suggesting a possible b-turn at Glu2-Arg3-Gly4-Ser5
(Fig. 6A). The two consecutive d
NN

peptide was 1.02 A
˚
, while that of residues at turn region
Glu2-Arg3-Gly4-Ser5 was 0.32 A
˚
, indicating the stable
nature of the b-turn conformation. The /, w angles around
Arg3-Gly4 and Val8-Ala9 of the structures showed the
possibility of a type II b-turn at Glu2-Arg3-Gly4-Ser5
residues and a type III b-turn at Leu7-Val8-Ala9-Glu10,
respectively [31]. Therefore, the structure of peptide cER
consists of two b-turns, located at the N- and C-termini. A
comparison of the b-turn structure of cER with the similar
region in the crystal structure of rat and human CD2 was
carried out. In the rat CD2 crystal structure, the b-turn
structure was exhibited by residues Arg37-Gly38-Ser39-
Thr40. The peptide cER displayed a b-turn structure with
shift in one re sidue compared to the protein from which it is
derived. In ratCD2, the type of b-turn observed at Arg37-
Gly38-Ser39-Thr40 i s a type II¢ b-turn whereas in cER
peptide t he b-turn is type II [31]. This is due to the position
of Gly amino acid in the b-turn which is flexible. In human
CD2, similar region (Fig. 1) has a b-turn is around Thr38-
Ser39-Asp40-Lys41 a nd the turn o bserved was type I
b-turn. An additional b-turn was observed in the cER
peptide structure at the Leu7-Val8-Ala9-Glu10 sequence
compared with the corresponding part in rat CD2
(Fig. 7 A).
The structure of peptide cVY. Several lines of NMR
evidence we re consistent with the existence of a b-turn in

was used to represent the final structure.
To check the convergence, the structures in each family were
superimposed on the average structure i n each family. All
structures presented a well-defined b-turn spanning residues
Ser4-Gly7 [31]. Lack of convergence was observed in the
first residue and the last three residues in the peptide
sequence. The average rmsd of the b ackbone atoms of
12 structures compared to the average structure was
0.98 ± 0.35 A
˚
, while the average rmsd at the residue
Ser4-Thr5-Asn6-Gly7 was 0.34 ± 0.06 A
˚
, indicating the
Fig. 6. Summar y of ROEs for pe ptides cER (A) and cVY (B). The
thickness of bars indicate the intensity o f ROE cross-peaks, and were
assigned as strong, medium a nd weak.
Ó FEBS 2004 Design of peptides for T-cell adhesion inhibition (Eur. J. Biochem. 271) 2879
stable nature of the b-turn conformation. A representative
structure o f p eptide cVY families are shown in the Fig. 7B.
The /, w angles around Thr5-Asn6 showed that the
structure of the peptide deviated slightly from the ideal
type I b-turn [31]. A comparison of the b-turn structure of
cVY with the similar region in the crystal structure of rat
and human CD2 was carried out. T he type-I b-turn
observed in cVY around S er4-Thr5-Asn6-Gly7 was similar
to that in rat CD2 crystal structure (Ser87-Thr88-Asn89-
Gly90). In human CD2, the similar region Asp87-Thr88-
Lys89-Gly90 exhibits a type I b-turn. Superimposition of b-
turn regions from rat CD2 and human CD2 crystal

the Fig. 8A. The b-turn at Arg2-Gly3-Ser4-Thr5 was type
II¢ b-turn as observed in the case of rat CD2 crystal
structure. Superimposition of backbone atoms of the
residues in the b-turn region of rat and human CD2
(similar region) with cEL peptide b-turn region (Arg2-Gly3-
Ser4-Thr5) indicated t hat the rmsd of the backbone atoms
was 0 .67 A
˚
with rat CD2 and 1.2 A
˚
with human CD2. Thus,
the peptide mimics the b-turn region of the protein from
which it is derived from. The peptide model also showed
intramolecular hydrogen bonds between NH of Thr5 and
CO of Arg2, as well as NH of Arg2 and CO of Thr5.
Structure of cyclic hexapeptides – cYT peptide. The
NMR data of the cYT peptide wer e indicative of its flexible
nature. The chemical shift dispersion of amides was less
than 1 p.p.m., and the Gly5 H
a
enantiotopic protons had a
degenerate chemical shift usually indicative of flexible
structure. Amide region of t he NOESY data suggested
weak intensity N OE connectivities between Tyr1-Ser2, Ser2-
Thr3, Asn4-Gly5 and Gly5-Thr6. Most of the coupling
constants were in the range of 6–8 Hz. Ser2 NH showed a
temperature coefficient of chemical shift value of
2.2 p.p.b.ÆK
)1
which may be due to the hydrogen-bonded

surface of CD58 (Fig. 2A), i.e. (a) position 1, which is a
CC¢ sheet and the interface of CD2–CD58 interaction; (b)
position 2, the top cavity where the turn region from CD2
interacts with C D58; (c) position 3, the bottom c avity where
a turn region of CD2 interacts with C D58. First, a
randomized rigid docking (blind docking) was performed
and the conformers with lowest energy or in significant
clusters were chosen to perform further docking studies with
flexible docking.
Peptide cER–CD58 complex. The automated molecular
docking calculations produced several possible binding sites
and c onformations for the peptide. The c onformation
corresponding to the low energy of docking was chosen as
the possible binding site. The results from the docking
studies of cER peptide-CD58 protein are s hown in Table 2.
Although, differen t starting positions were chosen for the
cER peptide on the CD58 protein surface, the final low
energy docked conformers of the peptide were near the top
cavity region on the protein. Thus, the most probable
binding site of cER peptide on CD58 is possibly near the top
cavity. Table 3 lists the residues involved in intermolecular
hydrogen bonding in the cER peptide and CD58 protein
interface. It is very clear that most of the residues that
exhibit b-turn structure in the peptide (Glu2-Arg3-Gly4-
Ser5) were involved in hydrogen bonding with the protein
receptor (CD58). The Ser5 residue in the turn region of
peptide cER interacts with the key r esidue Asp33 of CD58
that is important in adhesion. Thr6, the flanking residue of
the b-turn region also forms a hydrogen bond with Asp33
which was shown to be important in CD2–CD58

Bottom cavity Top cavity 1 )15.5 1
2 )14.9 2
Table 3. Amino acid residues forming hydrogen bonds in the cER–
–
CD58
interface. The residues in the turn region of peptide cER and in CD58
which are imp ortant for the CD2–CD58 interac tion are shown in bold
italic typeface.
Peptide cER
()15.5 kcalÆmol
)1
) CD58
Residue Atom Residue Atom
Ser5 H
c
Lys30 O
Phe11 O Lys30 H
f
Ser5 HN Gln31 O
d
Arg3 O Gln31 H
e
Ser5 O Asp33 HN
Thr6 O Asp33 H
d
Arg3 NH1 Ser69 O
Phe11 O Ser70 H
c
Arg3 H
e

the CD58 protein. This supports the low biological
activity of cYT observed in the E-rosetting and hetero-
typic adhesion assays.
Discussion
Inhibition of CD2–CD58 interaction has important impli-
cations in controlling immune responses in autoimmune
diseases. I n this study, we d esigned 12-mer linear and cyclic
peptides (lVY, cVY, lER, and cER) as well as cyclic
hexapeptides (cEL a nd cYT) that were derived from the rat
CD2 sequence. Initially, the design of small molecular
inhibitors was based on the crystal structure of rat CD2 [33–
37]. The CD58 (LFA-3) binding ability o f CD2 is known to
reside in domain-1 of CD2 protein. CD2 peptide mapping
and mutagenesis indicated that the binding surface of CD2
consists of b-sheet formed by strands GFCC¢C¢¢.Thecrystal
structure of CD2 (Fig. 2B) indicated that the rather flat
b-sheet surface does not provide a complementary shape to
bind to CD 58, and h ence does not have well-defined
epitopes t o design small molecular inhibitors. The structure
of CD2 is similar to CD4 and other IgSF molecules. In the
D1 domain of CD4, the b-turn near CC¢ appears to be
important fo r binding to its receptor [38]. b-Turn peptides
based on C D4 have been shown to be effective in inhibiting
CD4 interactions [38]. Analysis of the crystal structure of
CD2 revealed that on either side of t he binding surface of
CD2, there are b-turns whic h stabilize the b-strands. Thus,
we hypothesized that these b-turns may serve as good
surface epitopes for the design of peptides to inhibit CD2–
CD58 interactions. Meanwhile, the crystal structure of
human CD2–CD58 became available [30]. Examination of

conformations
in the cluster
CC¢ sheet CC¢ sheet 1 )10.7 2
2 )10.0 5
Top cavity CC¢ sheet 1 )10.4 1
2 )10.0 1
Bottom Cavity CC¢ sheet 1 )9.7 2
4 )9.4 1
Table 5. Amino acid residues forming hydrogen bonds in the cVY–
CD58 in ter face. The residue s in the turn region of peptide cVY an d in
CD58 which are important for t he CD2–CD58 interactions are shown
in bold italic t ypeface.
Peptide cVY
()10.7 kcal/mol) CD58
Residue Atom Residue Atom
Tyr3 O Lys29 Hz
Thr5 O
c
Lys29 Hz
Asn6 H
d
Asp33 O
d
2882 L. Jining et al. (Eur. J. Biochem. 271) Ó FEBS 2004
important in CD2–CD58 interaction [3]. Therefore, we
hypothesized that the stable b-turn conformation mimick-
ing the native CD2 surface-binding region with CD58 may
be important for inhibitory activity of the peptide. Since
human CD2 shares sequence homology with rat CD2, we
designed the peptid e based on the b-turn region o f rat C D2

To stabilize the b-turn structure in the designed
peptides, cyclic versions of the peptides were synthesized.
We chose to cyclize the peptides by disulfide bonds with
the introduction of amino acids penicillamine (Pen) and
cysteine at the two ends of the peptide sequence.
Cyclization by disulfide bond is relatively easy and yields
good yield after purification. Penicillamine with two bulky
methyl groups is known to stabilize disulfide bonds [39].
Penisusedintheposition1becauseinpreviousworkwe
have been successful in improving conformational stability
of the cyclic peptides by using Pen at position 1 [24,40].
Initially, 12-mer peptides were designed (Table 1). After
preliminary examination of biological activity of 12-mer
linear and cyclic peptides, the peptides were truncated to
six amino acid residues in order to elucidate the minimum
number of critical amino acids necessary for biological
activity. These hexapeptides were cyclized by amide bonds
to stabilize the structure. In the hexapeptides, cyclization
by disulfide bond w as not d esigned since a ddition of
penicillamine and cysteine to form a disulfide bond will
increase the number of amino acids in the peptide. The
information obtained from truncating the peptides will
also be useful in the design of future generation pharma-
cophores. A control peptide (Table 1) was designed to
compare the importance of primary and secondary
structures in the designed peptides. The control peptide
sequence was chosen from the Ôhot-spotÕ [3] region of the
hCD2–hCD58 interface on t he CD2 p rotein. The
sequence w as then r eversed and the important amino
acidTyr86aswellasTyr81wasreplacedbyAlato

may be unique. MOLT-3 and Jurkat cells are derived
from lymphocytes, while sheep blood cells are erythro-
cytes. Furthermore, CD58 may have different epitopes to
bind to antibody in the antibody binding assay. Compar-
ing the concentration of peptides used in inhibition studies
by all the three methods, the antib ody binding inhibition
assay was the least sensitive and required higher concen-
trations of peptides to observe inhibition. Also, the
peptide lVY is a linear and may exhibit random structure.
Overall, the E-rosetting and heterotypic adhesion assays
provide concrete evidence that the designed peptides can
inhibit the mechanism of cell–cell adhesion.
Truncation of the 12-mer peptide cER to the cyclic
hexapeptide cEL resulted in higher biological activity. This
provides the evidence that amino acids in the b-turn region
and stable conformation of the peptide are important for
biological activity. Truncation of the cVY peptide to the
hexapeptide cYT resulted in drastically reduced biological
activity suggesting the lack of stable structure and amino
acids important for biological activity of the peptide .
This suggests that b-turn conformation of the residues in
the b-turn region may be important in the inhibitory
mechanism, in addition to the primary structure of the
peptide.
To confirm our hypothesis that the b-turn structures are
important for the inhibitory activities of the peptides, the
structures of the cyclic peptides were determined by NMR.
The results proved that the stable b-turn conformation
exists in the cyclic peptides. The b-turn regions in these
peptides appeared to be more stable with flexibility at the

may potentially bind to the CD58 protein as indicated by
the docking studies. The lowest energy docked structure was
at the top cavity. On the other hand, cVY peptide has
relatively high energy compared to cER, but the position of
docking is near the CC¢ sheet which involves many salt
bridges and hydrogen bonds in CD2–CD58 interaction.
Thus, in terms of inhibition of cell adhesion, cVY has more
potential since it may directly interrupt the binding mech-
anism of CD2–CD58. Inhibition of cell adhesion involves
interrupting the interaction of key residues at t he protein–
protein interface. If we consider hydrogen bonding interac-
tion between the p eptide and CD58, the cER peptide f orms
hydrogen bonds with Asp33 in the lowest energy docking
position. Asp33 is one of the key residues in CD2–CD58
protein–protein interaction [3]. The peptide cVY is invol-
ved in hydrogen bonding interaction with Asp33 and Lys29
on CD58 in the lowest docked energy structure. Both
Asp33 and Lys29 are very important residues in CD2–
CD58 interaction. Mutational studies have indicated that
removal of these two interactions can result in loss of
CD2–CD58 interaction. This correlates with higher inhib-
itory activity of the cVY peptide compared to the cER
peptide.
In the case of cyclic hexapeptide cEL, the lowest
energy docked structure showed that cEL hydrogen
bonds with Lys34, whereas cYT is not involved in
hydrogen bonding with key residues that are important
for CD2–CD58 interaction. This correlates with the
higher inhibitory activity of the cEL peptide and the very
low inhibitory activity of the cYT peptide compared to

Acknowledgements
We gratefully acknowledge the Super Computer and Visualization
Unit, National University of Singapore for the use of computational
facilities; Prof. Arthur J. Olson, the Scripps Research Institute , USA,
for providing us with the AutoDock software. This research was funded
by an a cademic research grant ( R-148-000-026-1 12) from the National
University of Singapore, S ingapore.
References
1. Bierer, B.E. & Burakoff, S.J. (1988) T cells adhesion molecules.
FASEB J. 2, 2584–2590.
2. Wang, J.H. & Reinherz, E. (2000) Structure basis of cell–cell
interactions in the i mmune system. Curr. Opin. Structural Biol. 10,
656–661.
3. Kim, M., Sun, Z.Y.J., Byron, O., Campbell, G., Wagner, G.,
Wang, J H. & Reinherz, E.L. (2 001) Molecular d issection of the
CD2-CD58 c ounter–recep tor interface identifies CD2 Tyr86 an d
CD58 Lys34 residue s as the functional Ôho t spotÕ. J. Mol. Biol. 312,
711–720.
4. Davis, S.J. & van der Merwe, P.A. ( 1996) The structure and ligand
interactions of CD2: implicatio ns for T cell function. Immunol.
Today 17, 1 77–187.
5. Bachmann, M.F., Barner, M. & Kopf, M. (1999) CD2 sets
quantitative thresholds in T cell activation. J. Exp. Med. 190,
1383–1391.
6. Mojcik, C.F. & Shevach, E. (1997) Adhesion mo lecules: a rheu-
matologic perspective. Arthritis Rheumatism 6, 991–1004.
7. Kaplon, R.J., Hochman, P.S., Mic hler, R .E., Kwiatkowsk i, P.A.,
Edwards, N.M., Berger, C.L., Xu, H., Meier, W ., Walln er, B .P.,
Chisholm, P . & Marboe, C .C. (1995) Short c ourse single agent
therapy with an LFA-3-IgG1 fus ion protein prolongs primate

ease. Curr. Opin. Immunol. 13, 683–686.
14. Przepiorka, D., Phillips, G.L., Ratanatharathorn, V., Michele,
C.F., Sehn, L.H., Antin, J.H., Lebherz, D ., Awwad, M., Hope, J.
& McClain, J.B. (1998) A phase II study of BTI-322, a monoclonal
anti-CD2 antibody, for treatment o f steroid-resistant acute graft-
versus-host disease. Blood 92, 4066–4071.
15. Branco, L., Barren, P., Mao, S.Y., Pfarr, D., Kaplon, R.,
Psotema, C., Langermann, S., Koe nig, S. & Johnson, S. (1999)
Selective deletion of antigen-specific, activated T cells by a
humanized mAb to CD2 (MEDI-507) is mediated by NK cells.
Transplantation 68, 1588–1596.
16. Albert-Wolf, M., Meuer, S.C. & Wallich, R. (1991) Dual func-
tion of recombin ant hum an CD58: inhi bition of T-c ell adhesion
and activation via the CD2 pathway. Int. Immunol. 3, 1335–
1347.
17. Mosmann, T. (1983) Rapid colorimetric assay for cellular growth
and survival: application to proliferation and cy totox icity assays.
J. Immunol. Meth. 5, 55–63.
18. Bax, A. & Davis, D.G. (1985) MLEV-17-base d two-dimensional
homonuclear magnetization transfer spectroscopy. J. Magn.
Reson. 65, 355–360.
19. Rance, M., Sorensen, O.W., Bodenhausen , G., Wagner, G., Ernst,
R.R. & Wu
¨
thrich, K. (1983) Improved spectral resolution in
COSY
1
H NMR spectra of protein via double qua ntum filtering.
Biochem. Biophys. Res. Commun. 117, 479–485.
20. Bax, A. & Davis, D.G. (1985) Practical aspects of two-dimensional

(1990) Stereochemical modeling of disulfide bridges. Criteria f or
introduction into proteins by site-directed mutagenesis. Int. J.
Peptide Protein Res . 36, 147–155.
27. Liu, H., Banville, D.L., Basus, V.J. & James, T.L. (1995) Deriving
accurate interproton distances for ROESY spectra with limited
knowledge of scalar coupling constants via the carnival algorithm:
an iterative complete-relaxation-matrix approach. J. Magn. Reson.
107, 51–59.
28. Bystov, V.F. (1976) Spin-spin coupling and the conformational
states of peptide systems. Progr. NMR Spectrosc. 10, 41–81.
29. Morris, G.M., Goodsell, D.S., Halliday, R.S., Huey, R., Hart,
W.E., Belew, R.K. & Olson, A.J. (1998) Automatic docking using
a Lamarckian genetic algorithm and an empirical binding free
energy function. J. Co mpu. Chem. 19 , 1639–1662.
30.Wang,J.H.,Smoluar,A.,Tan,K.,Liu,J.H.,Kim,M.,Sun,
Z.Y.J., Wagner, G. & Reinherz, E.L. (1999) Structure of a he t-
erophilic adhesion complex between human CD2 and CD58
(LFA-3) counter-receptors. Cell 97 , 791–803.
31. Rose, G.D., Gierasch, L.M. & Sm ith, J .A. (1985) Turns in pep-
tides and proteins. Adv. In Protein Chem. 37, 1–97.
32. Hetenyi, C. & Van Der Spoel, D. (2002) Efficient docking of
peptides to proteins without prior knowledge of t he binding site.
Protein Sci. 11, 1729–1737.
33. Ikemizu, S., Sparks, L.M., van der Merwe, A., Stuart, D.I.,
Jones, E.Y. & Davis, S.J. (1999) Crystal structure of the CD2-
binding domain of CD58 (lymphocyte function-associated
antigen 3) at 1.8 A
˚
resolut io n. Proc. Natl Acad. Sci. USA 96, 4289–
4294.

Supplementary material
The following material is available from http://blackwell
publishing.com/products/journals/suppmat/ejb/ejb4198/
ejb4198.htm
Appendix 1. Chemical shift, coupling constants and tem-
perature dependence of amide proton resonance data at
298K for peptides, cER, cVY, cEL.
Ó FEBS 2004 Design of peptides for T-cell adhesion inhibition (Eur. J. Biochem. 271) 2885
Appendix 2. The backbone dihedral angles (in deg.) at
residues Glu2-Arg3-Gly4-Ser5 and at residues Ser4-Thr5-
Asn6-Gly7 for conformations of peptide cVY.
Appendix 3. Peptides cEL and cYT: CD58 docking results
starting from the potential binding sites out of 100 runs.
Fig. S1. Amide region of 500 MHz ROESY spectra of cER
cyclic peptide in 90% H
2
O/10% D
2
O.
Fig. S2. Amide region of 500 MHz ROESY spectrum of
cVY cyclic peptide in 90% H
2
O/10% D
2
O.
Fig. S3. Amide region of 500 MHz ROESY spectrum of
cEL cyclic peptide in 90% H
2
O/10% D
2