Modification of the N-terminus of peptidomimetic protein tyrosine phosphatase 1B (PTP1B) inhibitors: Identification of analogues with cellular activity

Article abstract of DOI:10.1016/S0960-894X(02)01065-X

Low molecular weight peptidomimetic compounds based on O-malonyl tyrosine and O-carboxymethyl salicylic acid are potent inhibitors of PTP1B. Modifications of the N-terminal Boc-Phe moiety were undertaken in an effort to improve physical chemical properties and to achieve cellular activity. Although Phe ultimately proved to be the optimal N-terminal amino acid, several viable replacements for the Boc group were identified, two of which afforded analogues that were effective at enhancing the insulin-stimulated uptake of 2-deoxyglucose by L6 myocytes.

Full text of DOI:10.1016/S0960-894X(02)01065-X

Bioorganic & Medicinal Chemistry Letters 13 (2003) 971–975
Modification of the N-Terminus of Peptidomimetic Protein
Tyrosine Phosphatase 1B (PTP1B) Inhibitors: Identification of
Analogues with Cellular Activity
Scott D. Larsen,a,* F. Craig Stevens,a Thomas J. Lindberg,a Paul M. Bodnar,a
Theresa J. O’Sullivan,a Heinrich J. Schostarez,a Barbara J. Palazukb
and John E. Bleasdaleb
aMedicinal Chemistry Research, Pharmacia Corporation, 333 Portage St., Kalamazoo, MI 49007, USA
bCell & Molecular Biology, Pharmacia Corporation, 333 Portage St., Kalamazoo, MI 49007, USA
Received 14 October 2002; accepted 2 December 2002
Abstract—Low molecular weight peptidomimetic compounds based on O-malonyl tyrosine and O-carboxymethyl salicylic acid are
potent inhibitors of PTP1B. Modifications of the N-terminal Boc-Phe moiety were undertaken in an effort to improve physical
chemical properties and to achieve cellular activity. Although Phe ultimately proved to be the optimal N-terminal amino acid,
several viable replacements for the Boc group were identified, two of which afforded analogues that were effective at enhancing the
insulin-stimulated uptake of 2-deoxyglucose by L6 myocytes.
# 2003 Elsevier Science Ltd. All rights reserved.
In Type II diabetes mellitus, the most prevalent form of
the disease, tissues develop resistance to the actions of
insulin even though, in most instances, the insulin
receptors in those tissues are structurally normal and
are in near normal abundance.1 One strategy to combat
this insulin resistance therapeutically is to maintain
insulin receptors (IR) in the active tyrosine-phosphory-
lated form by inhibiting enzymes that catalyze IR
dephosphorylation. Based on substantial evidence that
protein tyrosine phosphatase 1B (PTP1B) catalyzes IR
dephosphorylation and is involved physiologically and
pathologically in terminating insulin signaling, this
enzyme has emerged as an attractive therapeutic target.2
A molecular mechanism that describes the interaction of
PTP1B with activated IR has been proposed3 and a
variety of small molecule inhibitors of PTP1B have been
described, some of which potentiate insulin action on
cells or in animal models of diabetes.4
chemical properties, resulting in the identification of
small molecular weight competitive inhibitors incorpor-
ating carboxyl groups as surrogates for the tyrosine
phosphate group.6,7 Although some of these compounds
exhibited Ki values in the submicromolar range (e.g., 1),
evidence for insulin-sensitizing activity in whole cells
was lacking, with the exception of a triester prodrug 2
that effected modest augmentation of insulin-stimulated
2-deoxyglucose uptake into myocytes.5 Recognizing
that multiple carboxyl groups were likely precluding cell
penetration, attempts were made to jettison the
N-terminal carboxyl, leading to the discovery that the
simple N-Boc analogue 3 retained significant activity
against PTP1B.6 An investigation of carboxyl bioiso-
steres was concurrently undertaken, ultimately resulting
We recently reported the discovery that the simple tri-
peptide Ac-NH-Asp-Tyr(SO3H)-Nle-NH2 is a highly
effective inhibitor of PTP1B (Ki=5 mM).5 An analogue
program was subsequently pursued with the goal of
attenuating peptidic character and improving physical
*Corresponding author. Tel.: +1-269-833-7713; fax: +1-269-833-
2516; e-mail: scott.d.larsen@pharmacia.com
0960-894X/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0960-894X(02)01065-X

972
S. D. Larsen et al. / Bioorg. Med. Chem. Lett. 13 (2003) 971–975
in the identification of a tetrazole-containing analogue 4
with modest cellular activity.8
compounds were assayed for their ability to inhibit a
C-terminal truncated, soluble form of recombinant
human PTP1B.12 Specificity was determined by con-
current assays against two structurally dissimilar phos-
phatases, LAR and SHP-2. As a measure of the
potential to enhance insulin sensitivity in cells, all active
compounds were further evaluated for their ability to
augment insulin-stimulated uptake of 2-deoxyglucose
(2-DOG) into L6 myocytes.12
The results of several X-ray co-crystal structures5,6,8
indicated that the N-terminus was likely to be amenable
to further manipulation. Inhibitor electron density was
relatively diffuse in that area, suggesting that projection
into solvent was occurring to some extent, and a guani-
dine moiety (Arg 47) was not being fully engaged by the
Boc group.
Table 1 summarizes a survey of amino acid replace-
ments for Phe (13) within the malonate template 7.
Diverse residues were evaluated, in addition to residues
judged to have the potential to interact with the guani-
dine moiety of Arg 47. Removing the aromatic ring of
the phenylalanine (14) attenuated activity somewhat,
suggesting that the aromatic ring is making some lipo-
philic interaction with Arg 47. Analogues replacing the
aromatic ring with aliphatic residues (e.g., 21, 24, 25)
resulted in similar reductions in activity, as did those
that moved the aromatic ring farther from the amide
chain (15, 22, 27). Replacing the phenyl ring with indole
(28) or phenol (26) was more promising, but neither
analogue had significantly improved activity vs the lead
9, suggesting that additional pi-stacking or electrostatic
interactions with Arg 47 were not being achieved.
Employing chemistry developed in our earlier work, the
N-termini of two different templates (O-malonyl tyro-
sine and O-carboxymethyl salicylic acid) were probed
via modification of the amino acid residue or the term-
inal N-substituent (Schemes 1 and 2). Amines 5 and 9,
prepared as previously described,6 were neutralized or
deprotected, respectively, under standard conditions
prior to EDC-mediated coupling with various Boc-pro-
tected amino acids. The resulting amides 6 and 10 could
be directly saponified to acids 7 and 11, respectively.
Alternatively, 6 and 10 could be N-deprotected as
above, and the resulting free amines subsequently acy-
lated or sulfonylated under standard conditions,
affording analogues 8 and 12. Most of the reagents used
to functionalize the free amines were commercially
available. Exceptions included the serine and homo-
serine derivatives used to prepare 37, 39 and 40,9 the
heterocyclic acids used to prepare 42, 44 and 45,10 and
the pyrrolidone phenylalanine derivative used to pre-
pare 52.11 As described in our earlier work, all new
A preliminary investigation of N-termini was also
undertaken within the malonate template (8, Table 2).
Alterations included homologation (32), conversion of
the Boc carbamate to a urea (29), replacement of the
carbonyl with sulfonyl (30, 33) and replacement of the
t-butyl with aminoalkyl (31) or alkoxyalkyl (34). Only
the alkoxyalkyl group offered some improvement,
although again the magnitude does not suggest the
establishment of a new binding interaction. Interest-
ingly, some of these alterations appeared to improve the
specificity for inhibition of PTP1B relative to LAR.
A more limited set of amino acid changes was examined
in the O-carboxymethyl salicylic acid template (11,
Table 3). Here, removal of the phenyl group (36) was
even more detrimental than in the malonate template,
and replacement with hydrogen bond donor or acceptor
groups was not productive. Replacement of the phenyl
with pyridine (41) similarly reduced activity. Only tyr-
osine (35) was a viable surrogate for phenylalanine,
mirroring the result observed in Table 1 for the mal-
onate template.
Scheme 1. Reagents and conditions: (a) l-Boc-NHCH(G1)CO2H,
EDC, TEA; (b) HCl/HOAc; (c) G2Cl, TEA or G2OH, EDC, TEA; (d)
LiOH, THF/H2O.
Table 4 summarizes an extensive investigation of the
N-terminal substituents of the O-carboxymethyl sal-
icylic template 12. Compounds 42 – 45 represent
attempts to replace the N-terminal carboxyl group of 1
with less acidic heterocyclic bioisosteres,10 a modifi-
cation that could be expected to enhance cell perme-
ability. All proved to inhibit PTP1B more effectively
than the Boc derivative 3, with mercaptotetrazole 43
being the most potent (Ki=0.7 mM), but cellular activity
remained elusive. A highly lipophilic stearyl derivative
46 was prepared to try to achieve passive diffusion into
cells, an approach that has proven effective for a pepti-
dic PTP1B inhibitor,13 but intrinsic activity was lost.
Scheme 2. Reagents and conditions: (a) TFA/CH2Cl2; (b) l-Boc-
NHCH(G1)CO2H, EDC, TEA; (c) HCl/HOAc; (d) G2Cl, TEA or
G2OH, EDC, TEA; (e) LiOH, THF/H2O.

S. D. Larsen et al. / Bioorg. Med. Chem. Lett. 13 (2003) 971–975
Table 1. Replacement of Phe in O-malonyl tyrosines (7)
973
Compd
G1
% PTP1B inh @ 100, 10, 1 mM (Ki)
% LAR inha
% SHP-2 inha
% ctrl 2-DOG uptakea,b
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Bn
Me
CH2SBn
91, 48, 7 (9 mM)
83, 36, 7 (14 mM)
83, 16, 3
27
0
0
4
34
0
0
8
3
3
0
0
4
0
7
18
10
3
0
3
8
4
6
0
5
0
8
0
0
5
0
9
88
88
89
62
84
74
70
100
84
79
110
CH2-2-napth
CH2-1-napth
(CH2)2SMe
(CH2)2SOMe
CH2C6F5
CH2i-Pr
CH2OBn
CH2CONH2
iPr
CH2c-C6H11
CH2p-HO-Ph
CH2CH2Ph
CH2-3-indolyl
90, 29, 3
86, 33, 3
82, 25, 5
67, 21, 7
80, 25, 5
76, 24, 4
71, 19, 3
85, 40, 9
72, 18, 3
88, 18, 1
90, 50, 10
77, 17, 2
89, 44, 10
93
96
104
92
aAssayed at 100 mM.
bUptake of 2-deoxyglucose into L6 myocytes in the presence of half-maximal concentration of insulin, expressed as a percent of the control.
Table 2. Replacement of Boc in O-malonyl tyrosine inhibitors (8)
Compd
G2
% PTP1B inh @ 100, 10, 1 mM (Ki)
% LAR inha
% SHP-2 inha
% ctrl 2-DOG uptakea,b
13
29
30
31
32
33
34
Boc
CONHt-Bu
SO2Me
91, 48, 7 (9 mM)
86, 40, 9
86, 40, 9
27
21
0
0
2
10
6
8
14
8
10
11
88
93
93
89
87
98
104
CO(CH2)2NEt2
CO(CH2)2NHBoc
SO2CH2Ph
92, 54, 12
93, 62, 14
62, 16, 4
6
6
CO(CH2)2OMe
94, 65, 19
aAssayed at 100 mM.
bUptake of 2-deoxyglucose into L6 myocytes in the presence of half-maximal concentration of insulin, expressed as a percent of the control.
Table 3. Replacement of Phe in O-carboxymethyl salicylates (11)
Compd
G1
% PTP1B inh @ 10, 1 mM (Ki)
% LAR inha
% SHP-2 inha
% ctrl 2-DOG uptakea,b
3
Bn
CH2p-HO-Ph
Me
CH2OMOM
CH2OH
CH2CH2OMOM
CH2CH2OH
CH22-pyr
78, 30 (2.0 mM)
81, 32
43, 8
0
0
4
2
1
0
0
2
2
5
5
108
115
76
70
84
113
74
106
35
36
37
38
39
40
41
51, 16
41, 11
42, 6
45, 11
50, 10
4
6
8
1
aAssayed at 100 mM.
bUptake of 2-deoxyglucose into L6 myocytes in the presence of half-maximal concentration of insulin, expressed as a percent of the control.

974
S. D. Larsen et al. / Bioorg. Med. Chem. Lett. 13 (2003) 971–975
Table 4. Replacement of Boc in O-carboxymethyl salicylates (12)
Compd
G2
% PTP1B inh @ 10, 1 mM (Ki)
% LAR inha
% SHP-2 inha
% ctrl 2-DOG uptakea,b
3
Boc
COCH2S(1,2,4-triazol-3-yl)
COCH2(5-HS-1H-tetrazol-1-yl)
COCH2S(1,2,3-triazol-5-yl)
COCH2SO(1,2,4-triazol-3-yl)
CO(CH2)16CH3
78, 30 (2.0 mM)
82, 36
91, 57 (0.7 mM)
88, 49
85, 40
0
0
0
0
0
0
0
0
0
0
10
0
0
0
0
0
0
6
7
6
0
5
2
6
108
80
89
86
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
10
5
17
54
13
2
27
19
25
5
11
38
18
12
17
32
34
20
28
25
92
3, 1
77, 28
83, 37
78, 23
75, 19
73, 15
71, 24
COMe
CO(CH2)2OMe
CO(CH2)3OH
95
89
108
88
81
95
103
73
73
136
73
108
96
77
CO(CH2)2NHBoc
CO(CH2)2NHSO2Me
NHG2=2-pyrrolidon-1-yl
CO2Ph
85, 43 (0.87 mM)
88, 47
87, 45
74, 22 (1.4 mM)
70, 19
86, 43
88, 46
85, 39
85, 39 (0.59 mM)
87, 46 (0.67 mM)
COCH2-3-indolyl
COCH2Ph
CO(CH2)2Ph
CO(CH2)3Ph
COCH2-3-HO-Ph
COCH2-4-HO-Ph
COCH2-4-Me-Ph
COCH2-4-CF3-Ph
COCH2-4-MeO-Ph
130
141
aAssayed at 100 mM.
bUptake of 2-deoxyglucose into L6 myocytes in the presence of half-maximal concentration of insulin, expressed as a percent of the control.
The activity of simple N-acetyl derivative 47 was essen-
tially identical to the lead 3, suggesting that the tert-
butyl group is not participating in binding to PTP1B to
any significant extent, a finding that is consistent with
the structural data obtained for 3 in our previous work.6
Installation of a 2-methoxypropanoyl group on the
terminal nitrogen (48) did not afford the same magni-
tude in potency boost as was observed in the malonate
template (34). Inclusion of more polar groups or
homologating the Boc group (49–51) were similarly
ineffective at improving potency or realizing cellular
activity. A pyrrolidone analogue (52) was prepared to
reduce molecular weight and to investigate the effect of
removing the N-terminal hydrogen. No significant effect
on activity was observed, consistent with earlier struc-
tural data that indicated an absence of hydrogen bond-
ing between the N-terminal NH of 3 and PTP1B.
than the phenylacetyl derivative 55, some evidence for
augmentation of insulin-stimulated 2-DOG uptake into
L6 myocytes was observed (136% control, single
experiment). Variously substituted aryl rings were then
examined (58–62), including weakly acidic phenols with
the potential to interact electrostatically with Arg 47.
Although all of these analogues ultimately proved to be
essentially equipotent to 55 as inhibitors of PTP1B, two
analogues (61 and 62) are noteworthy because, despite
poor adsorptive coefficients measured in Caco-2 cells (<2
nm/s), each reproducibly stimulated insulin-dependent
glucose transport by L6 myocytes (Table 4). The 41%
increase observed with 62 is almost half the increase in
glucose transport expected when the insulin concentration
is increased from 10 nM (at which the inhibitors are tested)
to a saturating insulin concentration (300 nM). Although
some erosion in specificity versus SHP-2 is observed upon
replacement of the Boc group with arylacetyl groups, the
specificity remains greater than 100-fold.
Replacement of the tert-butyl moiety of 3 with phenyl
(53) provided the first neutral N-terminal O-carboxy-
methyl salicylate analogue possessing activity sig-
nificantly better than the lead. Nevertheless, cellular
activity with this compound still was not observed.
Compounds 54 and 55, bearing arylacetyl groups in lieu
of Boc, were similarly active. Lengthening the alkyl
chain by one (56) or two (57) carbons attenuated
enzyme activity, suggesting that a one carbon tether is
optimal for binding. Interestingly, although the phenyl
propanoyl analogue 56 possessed less intrinsic activity
In conclusion, the N-termini of peptidomimetic PTP1B
inhibitors 3 and 13 were modified by alteration of the
amino acid residue or the N-terminal substituent in an
attempt to modify the physical properties without
diminishing binding affinity. Although many changes
were tolerated with retention of enzyme-inhibitory
activity, only the replacement of the Boc group with
arylalkanoyl groups yielded compounds with superior
activity and measurable cellular activity.

S. D. Larsen et al. / Bioorg. Med. Chem. Lett. 13 (2003) 971–975
975
References and Notes
made from l-Boc-homoserine methyl ester (Howarth, N. M.;
Wakelin, L. P. G. J. Org. Chem. 1997, 62, 5441) by protection
with MOMCl (DIEA, DCM) followed by saponification. 40:
prepared from TBDMS-protected l-Boc-homoserine (Alt-
mann, K.-H.; Chiese, C. S.; Garcia-Echeverria, C. Bioorg.
Med. Chem. Lett. 1997, 7, 1119).
1. Kruszynska, Y. T.; Olefsky, J. M. J. Invest. Med. 1996, 44,
413.
2. Ukkola, O.; Santaniemi, M. J. Int. Med. 2002, 251, 467.
3. Salmeen, A.; Andersen, J. A.; Myers, M. P.; Tonks, N. K.;
Barford, D. Molec. Cell 2000, 6, 1401.
10. Drysdale, M. J.; Pritchard, M. C.; Horwell, D. C. J. Med.
Chem. 1992, 35, 2573.
4. (a) Blaskovich, M. A.; Kim, H.-O. Expert Opin. Ther. Pat.
2002, 12, 871. (b) Johnson, T. O.; Ermolieff, J.; Jirousek, M. R.
Nat. Rev.—Drug Discov. 2002, 1, 696.
11. 52 was prepared as in Scheme 2 from N-(2-oxo-1-pyrroli-
dinyl)-l-phenylalanine (Ocain, T. D.; Deininger, D. D. US
5,023,338) instead of l-Boc-NHCH(G1)CO2H.
5. Bleasdale, J. E.; Ogg, D.; Palazuk, B. J.; Jacob, C. S.;
Swanson, M. L.; Wang, W.-Y.; Thompson, D. P.; Conradi,
R. A.; Mathews, W. R.; Laborde, A. L.; Stuchly, C. W.;
Heijbel, A.; Bergdahl, K.; Bannow, C. A.; Smith, C. W.; Lil-
jebris, C.; Schostarez, H. J.; May, P. D.; Stevens, F. C.; Lar-
sen, S. D. Biochemistry 2001, 40, 5642.
6. Larsen, S. D.; Barf, T.; Liljebris, C.; May, P. D.; Ogg, D.;
O’Sullivan, T. J.; Palazuk, B. J.; Schostarez, H. J.; Stevens,
F. C.; Bleasdale, J. E. J. Med. Chem. 2002, 45, 598.
7. Larsen, S. D.; May, P. D.; Bleasdale, J. E.; Liljebris, C. L.;
Schostarez, H. J.; Barf, T. Preparation of substituted phenyl-
alanine derivatives as protein tyrosine phosphatase inhibitors.
WO 9911606, 11 March 1999.
8. Liljebris, C.; Larsen, S. D.; Ogg, D.; Palazuk, B. J.; Bleas-
dale, J. E. J. Med. Chem. 2002, 45, 1785.
9. 37: prepared from MOM-protected l-Boc-serine, which
was made by saponification of the corresponding methyl ester
(Hanessian, S.; Ninkovic, S. J. Org. Chem. 1996, 61, 5418). 39:
prepared from MOM-protected l-Boc-homoserine, which was
12. Initial rates of PTP-catalyzed hydrolysis of p-nitrophenyl
phosphate (pNPP) were determined as described5 and, where
appropriate, Ki values were computed using the direct linear
method of Cornish-Bowden (Enzpack 3, Biosoft, Cambridge,
UK). The various PTP enzymes used were purified recombi-
nant enzyme constructs of human origin (PTP1B, SHP-2,
LAR).5 Glucose transport by L6 myocytes (myotubes) was
monitored as the initial rate of uptake of 2-[3H]deoxyglucose
using l-[14C]glucose as a marker of the extracellular space.
Effects of PTP1B inhibitors on glucose transport were exam-
ined at a concentration of insulin (10 nM) shown previously to
elicit an approximate 50% maximal uptake.5 Data were nor-
malized to cell lysate protein recovered and are expressed as
(net insulin-dependent glucose transport in the presence of
PTP1B inhibitor–net insulin-dependent glucose transport in
the absence of PTP1B inhibitor)Â100%. A value of 120% or
higher was considered a significant increase.
13. Kole, H. K.; Liotta, A. S.; Kole, S.; Roth, J.; Montrose-
Rafizadeh, C.; Bernier, M. J. Biol. Chem. 1996, 271, 31619.