Peptide inhibitors of CDK2-cyclin A that target the cyclin recruitment-Site: Structural variants of the C-Terminal Phe

Article abstract of DOI:10.1016/S0960-894X(02)00508-5

A focused series of octapeptides based on the lead compound H-His-Ala-Lys-Arg-Arg-Leu-Ile-Phe-NH₂ 1, in which the C-terminal phenylalanine residue was replaced by α and/or β-modified variants, was synthesized using solid-phase chemistry. Both the L-threo-β-hydroxy-phenylalanine (β-phenylserine, Pse) and (2S)-phenylalaninol derivatives, as competitive binders at the cyclin-recruitment site, displayed potent inhibitory activity towards the CDK2-cyclin A complex. Unexpectedly, the D-threo-Pse derivatives also showed inhibitory activity.

Full text of DOI:10.1016/S0960-894X(02)00508-5

Bioorganic & Medicinal Chemistry Letters 12 (2002) 2501–2505
Peptide Inhibitors of CDK2-cyclin A that Target the Cyclin
Recruitment-Site: Structural Variants of the C-Terminal Phe
Gail E. Atkinson,a Angela Cowan,b Campbell McInnes,b Daniella I. Zheleva,b
Peter M. Fischerb and Weng C. Chana,*
aSchool of Pharmaceutical Sciences, University of Nottingham, University Park, Nottingham NG7 2RD, UK
bCyclacel Ltd., James Lindsay Place, Dundee DD1 5JJ, UK
Received 17 May 2002; accepted 4 July 2002
Abstract—A focused series of octapeptides based on the lead compound H-His-Ala-Lys-Arg-Arg-Leu-Ile-Phe-NH2 1, in which the
C-terminal phenylalanine residue was replaced by a and/or b-modified variants, was synthesized using solid-phase chemistry. Both
the l-threo-b-hydroxy-phenylalanine (b-phenylserine, Pse) and (2S)-phenylalaninol derivatives, as competitive binders at the cyclin-
recruitment site, displayed potent inhibitory activity towards the CDK2-cyclin A complex. Unexpectedly, the d-threo-Pse deriva-
tives also showed inhibitory activity. # 2002 Elsevier Science Ltd. All rights reserved.
Progression through the eukaryotic cell cycle is driven
by the orderly activation of cyclin-dependent kinases
(CDK) in association with appropriate cyclin subunits
and phosphorylation by CDK-activating kinase at a
conserved Thr residue.1,2 Each phase of the cell cycle is
characterized by the presence and activity of different
CDK-cyclin complexes, for example, CDK2-cyclin A
during the S phase. Crucially, the functions of these
complexes are negatively regulated by a host of reg-
ulatory proteins, including members of the CIP/KIP
recruitment-site (C-RS). Comprehensive SAR studies
resulted in the identification of the lead compound 1, in
which the native Ser is replaced by an Ala residue.5 In
addition, the C-terminal Phe residue appeared to be a
crucial determinant in the binding competence of 1,
since dramatic loss of activity was observed upon trun-
cation or even conservative replacement with Tyr or
Trp.
family, p21WAF1 and p27KIP1 3,4
Loss or mutation of
.
one or more of these negative-regulatory proteins would
result in cellular neoplastic phenotypes.2,4
The recognition of CDK-cyclin complexes by p21WAF1
occurs in part through interaction with a binding groove
in the cyclin subunit. This groove is centrally involved in
the recruitment or initial binding of macromolecular
CDK substrates such as pRb and E2F.3,4 Thus, selective
blockade of the cyclin recruitment-site would prevent
recognition and subsequent phosphorylation of CDK
substrates, and therefore offers a therapeutic approach
towards restoration of p21WAF1-like tumour suppres-
sion. Our recent efforts led to the discovery of an octa-
peptide 152HSKRRLIF159 located C-terminal in
p21WAF1, which displayed potent CDK-cyclin inhibi-
tory activity due to its capacity to bind to the cyclin
We herein report unexpected SAR data derived from a
focused series of peptides, in which the Phe residue in 1
is replaced by both a-modified and w-constrained Phe
analogues. These unique peptide derivatives were syn-
thesized using robust solid-phase chemistry and our
recently reported polymer-supported linker 2.6
*Corresponding author. Tel.: +44-115-9515080; fax: +44-115-
9515078; e-mail: weng.chan@nottingham.ac.uk
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(02)00508-5

2502
G. E. Atkinson et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2501–2505
A molecular model of the complex between 1 and cyclin
A was constructed,5,7 which showed that the LIF motif,
especially the Phe aromatic group, makes complimen-
tary contacts with the C-RS hydrophobic groove. In
order to gain further insights and to refine our under-
standing of the structure–activity profile of 1, con-
formationally w-constrained derivatives of Phe were
installed and evaluated for their binding capacity and
kinase-inhibitory activity towards CDK2-cyclin A.
As anticipated, no acylation of the exposed b-hydroxyl
functionality was evident. Acidolytic treatment of 4
using a TFA-based cocktail afforded the peptide
diastereoisomers 5 and 6 in high yield and purity.
Gratifyingly, the two diastereoisomers were readily
separated and purified by RP-HPLC. The identity and
purity of peptidic products were established using two
different reversed-phase silica HPLC columns10 and
ES–MS.
w-Constrained aromatic amino acids, typically b-methyl
or-isopropyl analogues, have been successfully applied
to unravel the bioactive conformation of peptides.8 In
this context, our approach is to use the commercially
available (2S,3R;2R,3S)-Pse as w-constrained surrogates
of Phe, in which the side-chain has a preferred trans
conformation. Moreover, the flexibility of this approach
allows us to exploit the installed threo-Pse residue as
synthetic precursor to Z-dehydrophenylalanine (Dhp)
and O-modified peptide derivatives.
The stereo-identities of the synthesized peptides were
established via the chymotrypsin-mediated resolution of
Ac-dl-Pse-OMe 7,11 outlined in Scheme 2. Thus, a fine
suspension of the methyl ester 7 in phosphate buffer was
exposed to a-chymotrypsin for 20 h, after which the
unreacted Ac-d-Pse-OMe 9 was extracted from the
reaction mixture using EtOAc. Following work up of
the aqueous phase, the desired product Ac-l-Pse-OH 8
was obtained as a white solid. Next, each derivative 8
and 9 was subjected to acid hydrolysis, lyophilization
and, without further purification, was N-protected using
FmocOSu under mild basic conditions to afford the
pure enantiomers 10 (mp 127–129 ꢀC, [a]D22 +27.9ꢀ
(c=1.0, MeOH) and 11 (mp 127–129 ꢀC, [a]D22 ꢁ29.0ꢀ
(c=1.0, MeOH).12 The stereo-identity of peptide 6 was
then established by its co-elution on RP-HPLC with the
peptide synthesized using Fmoc-(2R,3S)-Pse-OH 11 as
the initial building block.
The synthetic strategy (Scheme 1) is based on the selec-
tive and efficient condensation (>98%) of Fmoc-dl-
Pse-OH to the polymer-bound 4-alkoxy-40-methyl-
benzhydryl chloride linker 26 to afford the tethered
derivative 3. The peptide chain was then assembled
using standard Fmoc solid-phase peptide synthesis
conditions,9 and the N-terminus was capped by treat-
ment with di-tert-butyl dicarbonate (Boc2O) to give
the polymer-bound peptide 4 in typically >90% yield.
The polymer-tethered peptide 4 is an ideal intermediate
for selective chemical modification, for example, acyla-
tion, of the Pse8-residue—this would allow convenient
access to peptides that could probe the effect of sub-
stituents at the b-position. Following initial evaluation
of a series of O-acylation conditions, it was established
that acetylation was readily accomplished using Ac2O in
the presence of DMAP (Scheme 3). Acidolytic treat-
ment, followed by RP-HPLC purification provided the
Pse(Ac)-peptide derivatives 12 and 13.
We next focused our attention on the regio-selective
dehydration of the polymer-bound peptide 4. This
chemical modification would result in the formation
of an a,b-dehydrophenylalanine (Dhp) residue, thus
Scheme 1. Reagents and conditions: (i) FmocPseOH, DIEA, CH2Cl2,
24 h; (ii) piperidine-DMF (1:4); (iii) Fmoc-amino acid-TBTU-HOBt-
DIEA (1:1:1:2; 4–8 equiv); (iv) Boc2O (4–8 equiv), DMF; (v) TFA-
H2O-iPr3SiH (90:8:2), 2 h; (vi) separation using RP-HPLC using
Kromasil C8 preparative column (10ꢂ150 mm).
Scheme 2. Reagents and conditions: HCl(g), MeOH, 5 ꢀC to rt, 18 h;
(ii) Ac2O (3 equiv), aq AcONa (4 equiv), 5 ꢀC to rt, 16 h; (iii) a-chy-
motrypsin (Type II from bovine pancreas), 0.05 M aq phosphate buf-
fer (pH 8), rt, 20 h; (iv) 6 M aq HCl, reflux, 5 h; (v) FmocOSu, aq
Na2CO3, THF, 5 h.

G. E. Atkinson et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2501–2505
2503
Scheme 4. Reagents and conditions: (i) Et3N (30 equiv), SOCl2 (2ꢂ4
equiv), THF-CH2Cl2, ꢁ78ꢀ to 5 ꢀC; (ii) TFA-H2O-iPr3SiH-EtSMe
(88:8:2:2), 2 h.
Scheme 3. Reagents and conditions: (i) Ac2O (10 equiv), DIEA (1.3
equiv), DMAP (0.1 equiv), DMF, 24 h; (ii) TFA-H2O-iPr3SiH (90:8:2),
2 h; (iii) RP-HPLC using Kromasil C8 preparative column.
sized by solid-phase procedure (Scheme 5), which relies
on the highly efficient condensation of N-Fmoc-phenyl-
alaninol (FmocPhe[ CH2OH]) to the linker-resin 2
(96%) and the chemical robustness of the benzhydryl
methyl ether bond in 15.
introducing elements of conformational rigidity and
unique conjugative properties. Recent studies6 revealed
that a polymer-bound N-acyl-threo-b-hydroxy-amino-
acyl entity, on treatment with SOCl2, underwent intra-
molecular cyclization to a cyclic sulfamidite,13 which,
under mild basic condition, can undergo stereospecific
E1-elimination to afford a Z-a,b-dehydroaminoacyl
residue. We have successfully applied this methodol-
ogy to effect in situ dehydration of the threo-Pse8-
residue within the polymer-bound peptide 4 to furnish,
following acidolysis, the desired Z-Dhp-containing pep-
tide derivative 14 (Scheme 4). Peptide 14 was purified by
RP-HPLC and obtained in an overall yield of 24%.
All the synthetic peptide derivatives were screened in
vitro for their ability to inhibit the kinase activity of
CDK2-cyclin A using retinoblastoma protein (pRb) as
substrate,5 and in competitive binding assay5 using the
CDK2-cyclin A-p21WAF1 peptide system (Table 1).
Both sets of screening results revealed similar structure–
activity profiles and consistency.
The (2S,3R)-Pse8 peptide 5 showed similar competitive
binding potency as 1 but was a better inhibitor of kinase
function. Analysis of the molecular model generated for
the cyclin A-peptide 1 complex (Fig. 1) suggests that the
(2S,3R)-Pse residue can be accommodated since the Phe
pro-S b-hydrogen projects into an unoccupied cyclin
groove volume that is large enough for a hydroxyl
group. Moreover, the steric constraint in the w1-space
imposed by the b-OH group may impose more favor-
able interaction of the phenyl ring with the cyclin
groove. We propose that the Pse side-chain makes
excellent complimentary contact with the cyclin groove
resulting in an enhancement of activity. This is pro-
foundly obvious with the (Ser2,l-Pse8)-derivative 17,
which showed a considerable enhancement in activity
compared to the corresponding Ser2-peptide.
Using the procedures outlined in Schemes 1–4, a corre-
sponding series of Ser2-peptide derivatives 17–22 (see
Table 1) was also synthesized. Though the Ser2-deriva-
tive of 1 is known to be ꢃ100-fold less active as an
inhibitor of CDK2-cyclin A,5 these derivatives 17–22
are necessary to establish that the observed structure–
activity profile is due to localized and not long-range
structural effects.
To date, we are not aware of any studies that appraise
the effect of C-terminal homochiral b-amino alcohols on
the bioactivity of peptides that target C-RS. Hence, the
(2S)-phenylalaninol8-derivatives 16 and 22 were synthe-
Table 1. Inhibition (IC50 mM)a of CDK2-cyclin A complex determined using competitive binding assays and a functional kinase assay (pRb as
substrate)
Peptide derivatives
Competitive binding assays
CDK2/cyclin A
Kinase assay
Cyclin A
1
0.043ꢄ0.010
0.074ꢄ0.014
0.12ꢄ0.04
0.23ꢄ0.06
0.19ꢄ0.05
0.060ꢄ0.008
0.022ꢄ0.001
4.3
0.041ꢄ0.011
0.028ꢄ0.007
0.11ꢄ0.08
0.24ꢄ0.11
0.16ꢄ0.07
0.054ꢄ0.019
0.018ꢄ0.005
n.d.
0.039ꢄ0.012
0.002
0.30
0.32ꢄ0.09
0.20ꢄ0.25
0.070ꢄ0.011
0.024ꢄ0.011
3.4
l-Pse8-OH, 5
d-Pse8-OH, 6
l-Pse(Ac)8-OH, 12
d-Pse(Ac)8-OH, 13
Z-Dhp8-OH, 14
l-Phe8[ CH2OH], 16
Ser2
Ser2,l-Pse8-OH, 17
Ser2,d-Pse8-OH, 18
Ser2,l-Pse(Ac)8-OH, 19
Ser2,d-Pse(Ac)8-OH, 20
Ser2,Z-Dhp8-OH, 21
Ser2,l-Phe8[ CH2OH], 22
0.25ꢄ0.06
0.76ꢄ0.06
1.32ꢄ0.16
1.82ꢄ0.59
0.20ꢄ0.09
n.d.
0.19ꢄ0.13
0.71ꢄ0.87
0.80ꢄ0.14
0.60ꢄ0.28
0.22ꢄ0.17
0.31ꢄ0.16
0.56ꢄ0.45
1.20
n.d.
n.d.
1.3ꢄ0.6
0.43ꢄ0.12
aValues are means of three experiments; n.d.=not determined.

2504
G. E. Atkinson et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2501–2505
The introduction of Dhp residue (peptides 14 and 21)
led to a potency enhancement in the context of the Ser2
peptide. This appeared to indicate that a reduction in
the degree of rotational freedom of the phenyl ring is
tolerated. The other significant observation was that the
introduction of C-terminal 2S-phenylalaninol produced
a small improvement in potency. Here, it is likely that
the a-hydroxyl group makes a unique H-bond contact
with the guanidino function of Arg250 in cyclin A (see
Fig. 1).
These encouraging results not only highlight a generic
methodology for probing structure–activity relation-
ships using b-hydroxyamino acid building blocks, but
also provide a framework for the design of peptidomi-
metics and small molecules that target the cyclin
recruitment-site.
Scheme 5. Reagents and conditions: (i) FmocPhe[ CH2OH], DIEA,
ClCH2CH2Cl-THF (3:1), 24 h; (ii) piperidine-DMF (1:4); (iii) Fmoc-
amino acid-TBTU-HOBt-DIEA (1:1:1:2; 4–8 equiv); (iv) Boc2O (4–8
equiv), DMF; (v) TFA-H2O-iPr3SiH (90:8:2), 2 h.
Acknowledgements
The most striking and unexpected observation is that
the (2R,3S)-Pse8-, i.e., d-threo-Pse-containing peptides 6
and 18 retained good potency, although they are some-
what less potent compared to the (2S,3R)-Pse8 peptides
5 and 17, respectively. We are currently exploring the
implications of these results.
We are grateful to the BBSRC, UK and Cyclacel Ltd.
for the funding of a CASE studentship (to GEA).
References and Notes
Furthermore, the above hypothesized complimentary
interactions between the peptide ligand and C-RS is
highly fastidious, since the presence of a bulky b-acet-
oxy group within Pse8-residue in peptides 12, 13, 19 and
20 resulted in decreased binding competency and kinase
inhibitory activity. Interestingly, the potency order
appears to be reversed for the (2S,3R)- and (2R,3S)-
forms in both pairs of peptide derivatives, although the
differences may not be significant. Nevertheless these
results suggest that the b-acetoxy group precludes the
complementary interactions more in the context of the
l-derivatives.
1. (a) Poon, R. Y. C.; Yamashita, K.; Adamczewski, J. P.;
Hunt, T.; Shuttleworth, J. EMBO J. 1993, 12, 3123. (b)
Fischer, R. P.; Morgan, D. O. Cell 1994, 78, 713. (c) Morgan,
D. O. Nature 1995, 374, 131.
2. (a) Malumbres, M.; Barbacid, M. Nature Rev. Cancer 2001,
1, 222. (b) Dynlacht, B. D. Nature 1997, 38, 149.
3. (a) Polyak, K.; Lee, M. H.; Erdjument-Bromage, H.; Koff,
A.; Roberts, J. M.; Tempst, P.; Massague, J. Cell 1994, 78, 59.
(b) Russo, A. A.; Jeffrey, P. D.; Patten, A. K.; Massague, J.;
Pavletich, N. P. Nature 1996, 382, 325. (c) Ball, K. L.; Lain, S.;
Fahraeus, R.; Symthe, C.; Lane, D. P. Curr. Biol. 1997, 7, 71.
(d) Brown, N. R.; Noble, M. E. M.; Endicott, J. A.; Johnson,
L. N. Nature Cell Biol. 1999, 1, 438. (e) Rank, K. B.; Evans,
D. B.; Sharma, S. K. Biochem. Biophys. Res. Comm. 2000, 271,
469.
4. (a) Endicott, J. A.; Noble, M. E. M.; Tucker, J. A. Curr.
Opin. Struct. Biol. 1999, 9, 738. (b) Pavletich, N. P. J. Mol.
Biol. 1999, 287, 821.
5. Fischer, P. M.; Zheleva, D. I.; McInnes, C.; Gavine, A.;
Zhelev, N. Z.; Lane, D. Clin. Cancer Res. 2001, 7 (Suppl),
3821. S. McInnes, C.; Andrews, M. J. I.; Zheleva, D. I.; Lane,
D. P.; Fischer, P. M. Curr. Med. Chem. Anti-Cancer Agents
2002, submitted for publication.
6. Atkinson, G. E.; Fischer, P. M.; Chan, W. C. J. Org. Chem.
2000, 65, 5048.
7. The model was generated using standard molecular dock-
ing techniques and the structure of CDK2-cyclin A-p27KIP1
complex (RCSB file 1JSU) 3(b) as the starting point.
8. (a) Gibson, S. E.; Guillo, N.; Tozer, M. J. Tetrahedron
1999, 55, 585. (b) Hruby, V. J.; Balse, P. M. Curr. Med. Chem.
2000, 7, 945.
9. Chan, W. C.; White, P. D., Eds.; Fmoc SolidPhase Peptide
Synthesis: A Practical Approach; Oxford University Press:
Oxford, 2000.
10. Hypersil PepC18, linear elution gradient 20–60% B in
25 min at 1.2 mL minꢁ1; Vydac C8, linear elution gradient 10–
Figure 1. Molecular model of cyclin A-peptide 1 complex.5,7 To aid
viewing of the model, only a portion of the cyclin groove is shown.
The pro-S b-H is highlighted in yellow, the cyclin groove Arg250 resi-
due’s side-chain in blue.

G. E. Atkinson et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2501–2505
2505
60% B in 20 min at 1.3 mL minꢁ1. Columns dimension:
4.6ꢂ150 mm; effluent was monitored at 220 nm; A: 0.06% aq
TFA; B: 0.06% TFA in MeCN–H2O (9:1).
11. Chenevert, R.; Letourneau, M.; Thiboutot, S. Can. J.
Chem. 1990, 68, 961.
12. Compound 10: ꢀH ([2H6]-DMSO, 250 MHz) 7.88 (2H, d,
J=7.5 Hz), 7.67 (1H, d, J 7.2 Hz), 7.63 (1H, d, J 7.5 Hz), 7.31–
7.43 (9H, m): aromatic Hs; 5.80 (1H, br s, OH); 5.16 (1H, d, J
3.3 Hz, CbH); 4.29 (1H, dd, J 3.3, 9.4 Hz, CaH); 4.01–4.16 (3H,
m, Fmoc CH, CH2).
13. Stohlmeyer, M. M.; Tanaka, H.; Wandless, T. J. J. Am.
Chem. Soc. 1999, 121, 6100.