2-Aminoethoxydiphenyl borate as a prototype drug for a group of structurally related calcium channel blockers in human platelets

Article abstract of DOI:10.1124/mol.105.015701

We have synthesized a series of 2-aminoethoxydiphenyl borate (2-APB, 2,2-diphenyl-1,3,2-oxazaborolidine) analogs and tested their ability to inhibit thrombin-induced Ca²⁺ influx in human platelets. The analogs were either synthesized by adding various substituents to the oxazaborolidine ring (methyl, dimethyl, tert-butyl, phenyl, methyl phenyl, and pyridyl) or increasing the size of the oxazaborolidine ring to seven- and nine-membered rings. NMR analysis of the boron-containing analogs suggests that each of them exist as a ring structure through the formation of an N→B coordinate bond (except for the hexyl analog). The possibility that these boron-containing compounds formed dimers was also considered. All compounds dose-dependently inhibited thrombin-induced Ca²⁺ influx in human platelets, with the 2,2-diphenyl-1,3,2-oxazaborolidine-5-one derivative having the weakest activity at 100 μM, whereas the (S)-4-benzyl and (R)-4-benzyl derivatives of 2-APB were approximately 10 times more potent than the parent 2-APB. Two nonboron analogs (3-methyl and 3-tert-butyl 2,2-diphenyl-1,3-oxazolidine) were synthesized; they had approximately the same activity as 2-APB, and this implies that the presence of boron was not necessary for inhibitory activity. All of the compounds tested were also able to inhibit thrombininduced calcium release. We concluded that extensive modifications of the oxazaborolidine ring in 2-APB can be made, and Ca²⁺-blocking activity was maintained. Copyright

Full text of DOI:10.1124/mol.105.015701

0026-895X/06/6901-247–256$20.00
MOLECULAR PHARMACOLOGY
Copyright © 2006 The American Society for Pharmacology and Experimental Therapeutics
Mol Pharmacol 69:247–256, 2006
Vol. 69, No. 1
15701/3070902
Printed in U.S.A.
2-Aminoethoxydiphenyl Borate as a Prototype Drug for a
Group of Structurally Related Calcium Channel Blockers in
Human Platelets
Yuliya Dobrydneva, Christopher J. Abelt, Beth Dovel, Celina M. Thadigiri, Roy L. Williams,
and Peter F. Blackmore
Department of Physiological Sciences, Eastern Virginia Medical School, Norfolk, Virginia (Y.D., P.F.B.); Department of
Chemistry and Biochemistry, Old Dominion University, Norfolk, Virginia (B.D., C.M.T., R.L.W.); and Department of Chemistry,
The College of William and Mary, Williamsburg, Virginia (C.J.A.)
Received June 15, 2005; accepted October 7, 2005
ABSTRACT
We have synthesized a series of 2-aminoethoxydiphenyl borate flux in human platelets, with the 2,2-diphenyl-1,3,2-oxazaboro-
(2-APB, 2,2-diphenyl-1,3,2-oxazaborolidine) analogs and lidine-5-one derivative having the weakest activity at 100 ␮M,
tested their ability to inhibit thrombin-induced Ca2ϩ influx in whereas the (S)-4-benzyl and (R)-4-benzyl derivatives of 2-APB
human platelets. The analogs were either synthesized by add- were approximately 10 times more potent than the parent
ing various substituents to the oxazaborolidine ring (methyl, 2-APB. Two nonboron analogs (3-methyl and 3-tert-butyl 2,2-
dimethyl, tert-butyl, phenyl, methyl phenyl, and pyridyl) or in- diphenyl-1,3-oxazolidine) were synthesized; they had approxi-
creasing the size of the oxazaborolidine ring to seven- and mately the same activity as 2-APB, and this implies that the
nine-membered rings. NMR analysis of the boron-containing presence of boron was not necessary for inhibitory activity. All
analogs suggests that each of them exist as a ring structure of the compounds tested were also able to inhibit thrombin-
through the formation of an N3B coordinate bond (except for induced calcium release. We concluded that extensive modifi-
the hexyl analog). The possibility that these boron-containing cations of the oxazaborolidine ring in 2-APB can be made, and
compounds formed dimers was also considered. All com- Ca2ϩ-blocking activity was maintained.
pounds dose-dependently inhibited thrombin-induced Ca2ϩ in-
2-Aminoethoxydiphenyl borate (2-APB; systematic name: nature of SOCE and how the SOCE channels are regulated,
2,2-diphenyl-1,3,2-oxazaborolidine, compound 1) was origi-
nally described as an inhibitor of inositol 1,4,5-trisphosphate
receptors (IP3R) in a variety of cells (Maruyama et al., 1997)
and has been extensively used as a probe for examining
calcium influx processes. 2-APB was inappropriately used in
several studies to imply that IP3Rs were involved in store-
operated Ca2ϩ entry (SOCE) (Ma et al., 2000; van Rossum et
al., 2000). For a review of our current understanding of the
see Parekh and Putney (2005). It has been suggested that
SOCE is mediated by members of the transient receptor
potential canonical (TRPC), family although this has not
been accepted universally (Parekh and Putney, 2005).
Upon further investigation, it became evident that 2-APB,
along with inhibition of IP3Rs, was also inhibiting SOCE
directly in several cell types (Bakowski et al., 2001; Braun et
al., 2001; Broad et al., 2001; Diver et al., 2001; Dobrydneva
and Blackmore, 2001; Dobrydneva et al., 2001, 2002; Gregory
et al., 2001; Iwasaki et al., 2001; Ma et al., 2001; Prakriya
and Lewis, 2001; Bootman et al., 2002). The inhibition of
SOCE seemed to be at an extracellular site by a mechanism
not involving intracellular IP3Rs. In our study showing that
2-APB blocked SOCE in human platelets (Dobrydneva and
Blackmore, 2001), we tested whether several other structur-
ally related compounds were also inhibitors of SOCE. We
The laboratory of P.F.B. was supported by grants from The Jeffress Memo-
rial Trust and The Commonwealth Health Research Board. Y.D. received
support from American Association of University Women Educational Foun-
dation and the United States Department of Agriculture.
Some of the findings in this manuscript were presented at two meetings of
the Virginia Academy of Science (Dovel et al., 2002; Thadigiri et al., 2003) and
at Experimental Biology (Dobrydneva et al., 2001, 2002).
Article, publication date, and citation information can be found at
http://molpharm.aspetjournals.org.
doi:10.1124/mol.105.015701.
ABBREVIATIONS: 2-APB, 2-aminoethoxydiphenyl borate; SOCE, store-operated Ca2ϩ entry; SERCA, sarcoplasmic/endoplasmic reticulum
calcium ATPase; TRPC, transient receptor potential canonical; TRPV, transient receptor potential vanilloid; TRPM, transient receptor potential
melastatin; N3B, nitrogen boron coordinate bond; d6-DMSO, deuterated dimethyl sulfoxide; IP3R, 1,4,5-trisphosphate receptor; AM, acetoxy-
methyl ester.
247

248
Dobrydneva et al.
found that the nonboron analog of 2-APB, 2,2-diphenyltetra- ciated protein-25 (Rosado and Sage, 2001, 2004a,b; Redondo
hydrofuran, and diphenylboronic anhydride also inhibited et al., 2004).
both calcium mobilization and calcium influx (Dobrydneva
The aim of the present study was to synthesize a series of
and Blackmore, 2001). Two other nonboron analogs, pheny- boron-containing and nonboron-containing analogs of 2-APB
toin and diphenhydramine, were very weak inhibitors of to determine whether we could obtain compounds more po-
calcium influx. From this limited structural analysis, it tent than the parent 2-APB at inhibiting Ca2ϩ influx in
seemed that compounds with calcium-blocking activity pos- human platelets. We anticipated that this series of com-
pounds would shed light onto the structure of the pharma-
cophore that was responsible for calcium-blocking activity.
We succeeded in synthesizing several analogs that were as
effective as the parent 2-APB, and two boron-containing com-
pounds that were more potent than 2-APB at inhibiting Ca2ϩ
influx in platelets.
sessed diphenyl groups; also, the presence of a saturated
five-membered ring structure was needed. The presence of
boron in the structure was not an absolute requirement,
although boron and nitrogen were necessary for ring forma-
tion through an N3B coordinate bond (Fig. 1). It has also
been proposed that 2-APB forms a dimer (Fig. 1) (van Ros-
sum et al., 2000). 2-APB, at high concentrations, has also
been shown to activate TRPV1, TRPV2, and TRPV3, three
heat-gated members of the transient receptor potential va-
nilloid ion-channel subfamily (Hu et al., 2004). 2,2-Diphe-
nyltetrahydrofuran and diphenylboronic anhydride (Dobryd-
neva and Blackmore, 2001) were also able to modulate the
activity of these channels (Chung et al., 2005).
The calcium channel responsible for store-mediated cal-
cium entry in human platelets is believed to be TRPC1
(Brownlow and Sage, 2003). The regulation of this entry
process seems to be very complex and possibly involves the
conformational coupling between the IP3R in the endoplas-
mic reticulum and the store-operated calcium channels lo-
cated in the plasma membrane (Rosado and Sage, 2005).
There is evidence for the involvement of synaptosome-asso-
ciated protein-25, pp60src, the actin cytoskeleton, and extra-
cellular regulated kinase in this process (Rosado and Sage,
2001, 2004a,b; Redondo et al., 2004). Patch-clamp experi-
ments have shown that 2-APB acts extracellularly on various
TRP channels (Xu et al., 2005), so it reasonable to assume
that 2-APB also acts extracellularly on TRPC1 in platelets.
However, 2-APB could also be interacting with any of the
other components that regulate TRPC1 activity such as ex-
tracellular regulated kinase, pp60src, and synaptosome-asso-
Materials and Methods
Materials. The following were obtained from Sigma-Aldrich (St.
Louis, MO): 2-aminoethoxydiphenyl borate, 2-(methylamino)etha-
nol, 2-(t-butylamino)ethanol, 2-amino-1-phenylethanol, 4-amino-1-
butanol, 6-amino-1-hexanol, diphenylborinic anhydride, the sodium
salt of glycine, ethanolamine, dimethyl sulfoxide, benzophenone,
phosphorus pentachloride, and EGTA. 3,3-Diphenyl dihydrofuran-2-
one (trivial name: ␣,␣-diphenyl-␥-butyrolactone) was from Acros Or-
ganics (Fairlawn, NJ). All organic solvents were from Fisher Scien-
tific Co. (Pittsburgh, PA). Fura-2/AM was from Invitrogen (Carlsbad,
CA).
Synthesis of 2-APB Analogs. All melting points were deter-
mined with a Thomas Hoover Capillary Melting Point Apparatus
(Thomas Scientific, Swedesboro, NJ) and are uncorrected. 1H NMR
spectra were determined in a UnityPlus-400 MHz instrument oper-
ating at a field strength of 399.89 MHz or a Varian Mercury Vx-400
spectrometer (Varian, Inc., Palo Alto, CA). All carbon, hydrogen, and
nitrogen analyses were performed by Desert Analytics (Tucson, AZ).
A general method was developed for the synthesis of 2-APB and the
various 2-APB analogs, which is described for each compound. The
method involved a reaction between diphenylborinic anhydride and
a corresponding amino alcohol. The synthesis of 2,2-diphenyl-1,3,2-
oxazaborolidine (1) and (S)-4-benzyl-2,2-diphenyl-1,3,2-oxazaboroli-
dine (11) are shown in Fig. 2. All of the reaction solutions were
allowed to stir in an ice bath until precipitation occurred (no more
than 1 h), then they were chilled in an ice bath undisturbed for an
additional 15 min, which increased the amount of precipitation. All
of the products were purified by recrystallization using ethyl acetate
as the solvent, except for 2,2,5-triphenyl-1,3,2-oxazaborolidine (8),
which was recrystallized using benzene. All of the products were
white solids with sharp melting points. Monomeric structures with
an internal N3B coordinate bond are shown for all compounds
synthesized (Fig. 3). Because we have depicted 2-APB and its deriv-
atives as ring structures in this study, then the commonly used name
2-aminoethoxydiphenyl borate does not describe this ring system.
We have therefore used the 1983 IUPAC Hantzsch-Widman system
to name the ring forms of 2-APB and its derivatives. The numbering
order of the heteroatoms in the oxazaborolidine ring is oxygen num-
ber 1, nitrogen number 2, and boron number 3; hence, the numbering
of this ring is 1,3,2-oxazaborolidine in 2-APB.
2,2-Diphenyl-1,3,2-oxazaborolidine (1). Ethanolamine (0.02 g)
was added to 0.10 g of diphenylborinic anhydride in 4 ml of acetoni-
trile. The solution was stirred for 1 h without any signs of precipi-
tation. The solution was refrigerated for 15 h in which the product
precipitated and was filtered to give 0.056 g of compound 1 (75%
yield). The melting point was 193 to 194°C (literature, 190–192°C).
1H NMR (d6-DMSO): 2.83 (p, CH2), 3.77 (CH2), 6.07 (broad t, NH2),
7.01 (m, 2H), 7.14 (m, 4H), and 7.39 (m, 4H). The broad triplet at 6.07
caused by the NH2 rapidly exchanged with the addition of D2O.
3-Methyl-2,2-diphenyl-1,3,2-oxazaborolidine (2). 2-Methyl-
aminoethanol (0.05 g) was added to 0.20 g of diphenylborinic anhy-
Fig. 1. Structures of 2-APB: open-chain form, five-membered (1,3,2-
oxazaborolidine) ring form created when an N3B coordinate bond is
formed, and dimeric form produced when two N3B coordinate bonds are
formed between two molecules of 2-APB. Dimeric 2-APB could also exist
with the formation of only one N3B coordinate bond (structure not
shown).

2-APB Analogs Inhibit Calcium Influx in Human Platelets
249
dride in 8 ml of acetonitrile. The product precipitated to give 0.066 g
of compound 2 (46% yield); melting point, 196 to 197.5°C. 1H NMR
(d6-DMSO): ␦ 2.16 (d, CH3), 2.64 (p, CH), 3.03 (p, CH), 3.75 (q, CH),
4.01 (q, CH), 6.67 (m, NH), 7.09 (t, 2H), 7.16 (t, 4H), and 7.46 (dd,
4H). Analysis: calculations for C15H18BNO: C, 75.33; H, 7.60; N, 5.86;
found, C, 75.32; H, 7.56; N, 5.68. The multiplet at 6.67 was for the
NH rapidly exchanged with the addition of D2O.
3,3-Dimethyl-2,2-diphenyl-1,3,2-oxazaborolidine (3). Diphe-
nylborinic anhydride (0.01 g) was stirred with 0.05 ml of dimethyl-
aminoethanol in 10 ml of acetonitrile. The product slowly precipi-
tated from solution to give 0.195 g of compound 3 (82% yield); melting
point, 165 to 166°C. 1H NMR (d6-DMSO): 2.50 (s, 6H, CH3), 2.92 (t,
CH2), 4.10 (t, CH2), 7.10 (m, 2H), 7.18 (m, 4H), and 7.65 (m, 4H).
3-tert-Butyl-2,2-diphenyl-1,3,2-oxazaborolidine (4). 2-tert-
Butylaminoethanol (0.07 g) was added to 0.20 g of diphenylborinic
anhydride in 8 ml of acetonitrile. The product precipitated to give
0.125g of compound 4 (76% yield); melting point, 133.1 to 134.0°C. 1H
NMR (d6-DMSO): ␦ 1.16 (s, CH3), 2.69 (t, CH2), 3.55 (t, CH2), 6.91 (tt,
NH), 6.98 (t, 2H), 7.06 (t, 4H), 7.44 (dd, 4H). Analysis: calculations
for C18H24BNO: C, 73.85; H, 8.62; N, 4.31; found: C, 73.31; H, 8.76;
N, 4.68. The peak at 6.91, assigned to the NH, rapidly exchanged
with the addition of D2O.
2,2,5-Triphenyl-1,3,2-oxazaborolidine (5). 2-Amino-1-phe-
nylethanol (0.08 g) was added to 0.20 g of diphenylborinic anhydride
in 8.0 ml of acetonitrile. The product precipitated to give 0.072 g of
compound 5 (41% yield); melting point, 194.9 to 196.0°C. 1H NMR
(d6-DMSO): ␦ 3.25 (q, CH2), 4.85 (dd, CH), 6.31 (t, NH2), 7.03 (t, 2H),
7.08 (t, 4H), 7.16 (dd, 4H), 7.25 (t, 2H), 7.34 (t, 4H), 7.49 (dd, 4H).
Analysis: calculations for C20H20BNO: C, 79.75; H, 6.69; N, 4.65;
found, C, 80.27; H, 6.62; N, 4.65. The triplet peak at 6.31 rapidly
exchanged with the addition of D2O.
Fig. 2. The general scheme for the synthesis of the diphenyl oxazaboro-
lidine analogs is shown. Two examples are shown. A, the synthesis of the
parent 2,2-diphenyl-1,3,2-oxazaborolidine (1) is shown, and the starting
reactants are diphenylborinic anhydride and ethanolamine. B, the syn-
thesis of (S)-4-benzyl-2,2-diphenyl-1,3,2-oxazaborolidine (11) is shown,
and the starting reactants in this synthesis are diphenylborinic anhy-
dride and (S)(Ϫ)2-amino-3-phenyl-1-propanol.
Fig. 3. Structures of compounds (1
through 14) synthesized in this study.
The method for the synthesis of each
compound is detailed in the text.
Compound 15 (3,3-diphenyl dihydro-
furan-2-one) was obtained from
a
commercial source (ACROS). Com-
pounds 5, 8, 9, 11, and 12 contain a
chiral carbon atom. Compounds 5, 8,
and 9 are racemic mixtures.

250
Dobrydneva et al.
2,2-Diphenyl-1,3,2-oxazaborepane (6). 4-Amino-1-butanol
3-Methyl-2,2-diphenyl-1,3-oxazolidine (13). Silver oxide (2.0 g,
(0.053 g) was added to 0.20 g of diphenylborinic anhydride in 8.0 ml 8.6 mmol) was added to a mixture of dichlorodiphenylmethane (2.0 g,
of acetonitrile. The product precipitated from solution to give 0.020 g 9.4 mmol) and 2-(methylamino)ethanol (1.4 ml, 17 mmol). The mix-
of compound 6 (27% yield); melting point, 185.60 to 186.1°C. 1H NMR ture was warmed on a hot plate until a violent reaction ensued
(d6-DMSO): ␦ 1.58 (p, CH2), 1.67 (p, CH2), 2.71 (p, CH2), 3.59 (t, CH2), (Scheme 1). The mixture was cooled, and the solid residue was
5.82 (t, NH2), 7.01 (t, 2H), 7.12 (t, 4H), 7.45 (dd, 4H). Analysis: washed repeatedly with hexanes. The combined filtrates were
calculations for C16H20BNO: C, 75.95; H, 7.98; N, 5.58; found: C, washed with water, dried over sodium sulfate, filtered, and concen-
trated in vacuo. The residue was distilled under high vacuum. The
distillate was recrystallized from hexanes giving 0.87 g of compound
13 (3.6 mmol, 42% yield). NMR spectra were in agreement with
published data (Azzena et al., 1993).
76.47; H, 8.04; N, 5.59.
2,2-Diphenyl-1,3,2-oxazaboronane (7). 6-Amino-1-hexanol
(0.073 g) was added to 0.20 g of diphenylborinic anhydride in 8.0 ml
of acetonitrile. The product precipitated to give 0.228g of compound
7 (32% yield); melting point, 70.0 to 71.0°C. 1H NMR (d6-DMSO): ␦
1.28 (p, CH2), 1.35 (p, CH2), 1.42 (p, CH2), 1.63 (p, CH2), 2.63 (q,
CH2), 3.33 (broad s, NH2), 3.46 (t, CH2), 7.08 (t, 2H), 7.17 (t, 4H), 7.53
(dd, 4H). Analysis: calculations for C18H24BNO: C, 72.73; H, 8.26; N,
11.57; found: C, 72.33; H, 8.08; N, 11.04. The peak at 3.33 rapidly
exchanged with the addition of D2O.
2,2,5-Triphenyl-benzo[3,4]-1,3,2-oxazaborolidine (8). Sodium
borohydride (0.107 g) was added to 0.50 g of 2-acetylpyridine in 2.0
ml of methanol to yield 0.293 g of 2-(1-hydroxyethyl)pyridine after
stirring for 30 min. The precipitate was filtered out of the solution
using vacuum filtration and dried. 2-(1-Hydroxyethyl) pyridine
(0.283 g) was added to 0.825 g of diphenylborinic anhydride in 31.0
ml of acetonitrile. The product precipitated to yield 0.473 g of the
compound 8 (yield 69.2%); melting point, 200 to 201.1°C (literature
melting point, 199–200°C; Farfan et al., 1992).
5-Methyl-2,2-diphenyl-benzo[3,4]-1,3,2-oxazaborolidine (9).
A 0.277-g sample of 2-(1-hydroxyethyl)pyridine was stirred with
0.518 g of diphenylborinic anhydride in 30.0 ml of acetonitrile. The
product precipitated to give 0.381 g of compound 9 (72.8% yield);
melting point, 164.2 to 164.9°C (literature melting point, 164–165°C;
Farfan et al., 1992).
2,2-Diphenyl-1,3,2-oxazaborolidine-5-one (10). Diphenyl-
borinic anhydride (0.2 g), dissolved in 20.0 ml of acetonitrile, was
stirred at with 0.058 g of the sodium salt of glycine. The acetonitrile
was removed under nitrogen to give 0.238 g of compound 10 (45%
yield); melting point, 235 to 238°C. 1H NMR (d6-DMSO): 3.47 (t,
CH2), 7.09 (t, NH2), 7.17 (m, 2H), 7.23 (m, 4H), 7.39 (m, 4H). Anal-
ysis: calculations for C14H14BNO2: 70.43; H, 5.80; N, 5.86; found: C,
70.12; H, 5.83; N, 5.80. The triplet at 7.09 rapidly exchanged with the
addition of D2O.
3-tert-Butyl-2,2-diphenyl-1,3-oxazolidine (14). Silver acetate
(1.7 g, 10.2 mmol) was added to a mixture of dichlorodiphenylmeth-
ane (1.0 g, 4.7 mmol) and 2-(tert-butylamino)ethanol (5.5 g, 47
mmol). The mixture was warmed on a hotplate for 30 min during
which a brown solid appeared (Scheme 2). The reaction was allowed
to cool, and the solid was washed with hexanes (100 ml). The hexane
filtrate was washed three times with water (100 ml each), dried over
sodium sulfate, and concentrated in vacuo. Crystals formed slowly
upon standing, which were placed on an absorbant tissue to remove
the benzophenone side-product that remains in the melt. The prod-
uct 14 thus obtained was 230 mg (0.82 mmol, 17% yield) with a
melting point of 66 to 68°C. 1H NMR (CDCl3): ␦ 7.59 (m, 4H), 7.27 (m,
6H), 3.76 (t, J ϭ 6.5 Hz, 2H), 3.31 (t, J ϭ 6.5 Hz, 2H), 0.92 (s, 9H); 13
C
NMR (CDCl3): ␦ 144.5, 129.6, 127.5, 127.3, 99.7, 63.0, 54.5, 47.6,
29.9. The compound was sublimed under vacuum for analysis. Anal-
ysis: calculations for C19H23NO: C, 81.10; H, 8.24; N, 4.98; found: C,
80.98; H, 8.33; N, 4.81.
2,2-Diphenyl-1,3-oxazolidine (16). We were unable to isolate 2,
2-diphenyl-1,3-oxazolidine (compound 16), which is the carbon ana-
log of the ring form of 2-APB, where the boron is replaced by a
carbon. Results from the parent oxazolidine are therefore not re-
ported. The compound, however, was successfully prepared by the
condensation of ethanol amine with benzophenone imine. Neverthe-
less, the oxazolidine 16 exists as an equilibrium mixture with its
ring-opened 17 tautomer (Fig. 4). Proton NMR analysis suggests a
ratio of 14:86 in favor of the ring-opened form. The evidence for the
ring-opened structure is the nonequivalence of the phenyl rings. The
phenyl rings in the ring-open tautomer are diastereomeric because
the imine double bond. The 1H NMR shows four sets of strong
aromatic resonances at 7.62, 7.46, 7.34, and 7.17 ppm. The ring-
closed structure should show only two strong signals for the sets of
(S)-4-Benzyl-2,2-diphenyl-1,3,2-oxazaborolidine (11). (S)(Ϫ)-
2-Amino-3-phenyl-1-propanol (0.50 g) was dissolved in 11.0 ml of
acetonitrile and treated with a solution of 1.14 g of diphenylborinic
anhydride in 9.0 ml of acetonitrile. The reaction was stirred for 1
week, and then the solvent was removed with nitrogen gas to give
0.556 g of compound 11 (60% yield); melting point, 129 to 135°C. 1H
NMR (d6-DMSO): 2.734 (dd, 1H), 2.94 (dd, 1H), 3.35 (m, 1H), 3.55 (t,
1H), 3.81 (t, 1H), 5.80 (t, 1H), 6.42 (t, 1H), 7.05 to 7.45 (mm, 15H).
Analysis: calculations for C21H22BNO: C, 80.0; H, 6.98; N, 4.45;
found: C, 79.89; H, 6.95; N, 4.54. The two triplets at 5.80 and 6.42 are
assigned to the two nonequivalent NH protons rapidly exchanged
with the addition of D2O.
(R)-4-Benzyl-2,2-diphenyl-1,3,2-oxazaborolidine (12). (R)(ϩ)-
2-Amino-3-phenyl-1-propanol (0.50 g) was dissolved in 30.0 ml of
acetonitrile and treated with a solution of 1.14 g of diphenylborinic
anhydride in 9.0 ml of acetonitrile. The solution was stirred at for 4
days when the solvent was removed with nitrogen gas to give 0.510
g of compound 12 (55% yield); melting point, 120 to 125°C. 1H NMR
(d6-DMSO): 2.73 (dd, 1H), 2.94 (dd, 1H), 3.35 (t, 1H), 3.81 (t, 1H),
5.80 (t, 1H), 6.42 (t, 1H), 7.05 to 7.45 (mm, 15H). Analysis: calcula-
tions for C21H22BNO: C, 80.0; H, 6.98; N, 4.45; found: C, 79.66; H,
7.18; N, 5.25. The two triplets at 5.80 and 6.42 are assigned to the
two nonequivalent NH protons rapidly exchanged with the addition
of D2O.
Scheme 1.
Synthesis of Oxazolidine Analogs. Dichlorodiphenylmethane
was prepared from benzophenone and phosphorus pentachloride ac-
cording to the method of Staudinger and Freudenberger (1943).
Scheme 2.

2-APB Analogs Inhibit Calcium Influx in Human Platelets
251
four equivalent ortho and meta hydrogens, respectively. In the ali- a final concentration of 2.0 mM, and then test compounds were
pahtic region, there are two pairs of triplets. The major pair occurs at dissolved in Me2SO (various concentrations in 2.5 ␮l), and thrombin
3.84 and 3.48 ppm, whereas the minor pair occurs at 3.85 (obscured 0.01 units/ml were added. The measurements of [Ca2ϩ]i was per-
by the major isomer) and 3.15 ppm. The methylenes adjacent to the
nitrogen are the upfield resonances of each set. The imine structure
formed at room temperature in a SPEX ARCM spectrofluorometer
(SPEX Industries, Edison, NJ) using excitation wavelengths of 340
of the ring-opened isomer gives rise to the shift from 3.15 to 3.48 and 380 nm and an emission wavelength of 505 nm. Calibration was
ppm. Finally, the chemical shifts of the aliphatic protons of the performed as described previously (Dobrydneva and Blackmore,
ring-closed methyl derivative 13 are similar to those of the minor 2001). [Ca2ϩ]i was calculated by using the SPEX dM3000 software.
isomer. Precedent for ring-opening tautomerization has been re- The data in Fig. 5 show a typical experiment. To calculate the
ported. Thus, the oxazoline from acetophenone and ethanolamine percentage of inhibition of thrombin effects by the various analogs,
exists as a 50:50 mixture with the ring-opened form (Konkova et al., the [Ca2ϩ]i before thrombin addition was measured; then, the [Ca2ϩ
]
i
1999). It is reasonable to suggest that the conjugation provided by was measured after the thrombin effect had reached a maximum
the extra phenyl ring in benzophenone results in a greater propor- value (which was approximately after 1 min). The difference was
tion of the imine form.
then compared with the effect observed in the absence of any inhib-
Blood Donors and Platelet Preparation. All donors were itor (control), which represents 0% inhibition.
healthy nonsmoking volunteers (age 20–60 years) who had not con-
sumed any medication known to affect platelet function (e.g., calci-
um-channel blockers and aspirin) for at least 10 days before the
Results and Discussion
study. Venous blood was collected into 1/10 volume of (74.8 mM
sodium citrate, 38.1 mM citric acid, and 123 mM dextrose, pH 6.4)
(Baxter, McGaw Park, IL). The blood was centrifuged at 250g for 10
min at room temperature to obtain platelet-rich plasma. The plate-
let-rich plasma was centrifuged at 550g for 12 min to sediment the
platelets. The platelets were then resuspended in a modified Ty-
rode’s physiological salt solution (145 mM NaCl, 4 mM KCl, 1 mM
MgSO4, 0.5 mM Na2HPO4, 10 mM sodium-HEPES and 6 mM glu-
cose, pH 7.4) containing 1.0 mM EGTA to prevent spontaneous
aggregation during the various experimental manipulations by bind-
Chemistry. All of the 2-APB analogs that we synthesized
were obtained in 27 to 76% yields after purification by re-
crystallization. The reactions were temperature-dependent,
leading to high yields at cool temperatures and low yields at
warm temperatures. It is believed that the low yields at
warm temperatures could be caused by the acetonitrile re-
acting with the diphenylboronic anhydride before the amine
could be added. A possible product may be the acetonitrile
adduct of diphenylboronic anhydride, formed when the un-
paired electrons of acetonitrile forms an N3B coordinate
bond (Fig. 6). This finding illustrates one possible pitfall
when synthesizing boron-containing compounds in the pres-
ence of solvents (e.g., acetonitrile) or reactants with unpaired
electrons that could form coordinate bonds. At warm temper-
atures, the solution of acetonitrile and anhydride developed a
golden color, whereas at cooler temperatures, the solution
remains colorless, which suggests some involvement of the
solvent acetonitrile. To avoid the contamination of the prod-
uct produced when acetonitrile and diphenylboronic anhy-
dride react, the reaction vessel was placed in an ice bath for
the duration of the reaction, and the recrystallizing solvents
ing extracellular Ca2ϩ
.
Platelet Loading with Fura-2 and Measurement of [Ca2؉]i.
Intracellular calcium measurements [Ca2ϩ]i used the fluorescent dye
Fura-2, which involved incubating the platelets with the cell-per-
meant acetoxymethyl ester (Fura-2/AM). Suspensions of human
platelets (isolated as described above) were incubated with 2 ␮M
Fura-2/AM for 1 h at room temperature (on a rocking platform).
Excess Fura-2/AM was removed by centrifugation (500g for 10 min),
and the platelets were suspended in modified Tyrode’s buffer without
added EGTA. Aliquots of platelet suspension (0.5 ml) were added to
1-ml cuvettes containing a Teflon-coated stirrer bar (CHRONO-LOG,
Havertown, PA). To measure intracellular Ca2ϩ mobilization, Ca2ϩ
was not added to the platelet suspension; test compounds were added
at approximately 10 s, then at 20 s, 0.01 units/ml thrombin was
added. To evaluate Ca2ϩ influx, approximately 30 s before [Ca2ϩ
]
i
measurements were performed, Ca2ϩ was added back to the buffer to
Fig. 5. The effect of (S)-4-benzyl-2,2-diphenyl-1,3,2-oxazaborolidine (11)
on the ability of thrombin to increase [Ca2ϩ]i in the presence of extracel-
lular calcium. Extracellular calcium 2.0 mM was added to the platelet
suspension 30 s before data collection was started. Approximately 10 s
after data collection was commenced, various concentrations (1, 10, and
100 ␮M) of (S)-4-benzyl-2,2-diphenyl-1,3,2-oxazaborolidine (11) were
added to the platelet suspension. Thrombin (0.01 units/ml) was added
approximately 10 s later. A representative experiment is shown, The
ability of thrombin to increase [Ca2ϩ]i was inhibited in a dose-dependent
fashion by (S)-4-benzyl-2,2-diphenyl-1,3,2-oxazaborolidine (11).
Fig. 4. Structures of the proposed tautomeric forms of the nonboron
analog of 2-APB, 2,2-diphenyl-1,3-oxazolidine (16), and 2-(diphenylmeth-
yleneamino)ethanol (17). The formation of compounds 16 and 17 involved
condensing ethanol amine with benzophenone imine after deamination of
the 2-(aminodiphenylmethylamino)ethanol intermediate.

252
Dobrydneva et al.
were changed to ethyl acetate and benzene. All of these amine side chain in 2-APB by adding two extra carbons (ethyl
2-APB analogs were in agreement with the carbon, hydrogen, to butyl) did not prevent ring formation on the basis of the
and nitrogen analyses and exhibited the predicted functional NMR data. However, extending the length of the side chain
group absorptions in the infrared. The NMR findings of these by another two carbons (butyl to hexyl) prevented the forma-
compounds were in good agreement with the assigned struc- tion of a ring, on the basis of the NMR data. The reason for
tures.
this is not clear; however, the longer side chain decreases the
One of the intriguing aspects of the structural analysis of probability that the nitrogen’s unpaired electrons will be in
2-APB was observed in the NMR data for this series of close proximity to the boron and hence lessen the likelihood
compounds. 2-APB can be visualized as either an open-chain that a coordinate bond will be formed. The formation of hexyl
form or a ring structure (Fig. 1) because of the potential dimers would also be less likely to occur because either one or
coordinate covalent bond that might arise between the elec- two coordinate bonds would have to be formed. As seen in
tron-rich amine and the electron-deficient boron. The exis- Table 1, hexyl-APB (compound 7), butyl-APB (6), and 2-APB
tence of the ring form is indeed confirmed by the crystal (1) all have the same ability to inhibit thrombin, despite
structure of 2-APB (Retting and Trotter, 1976) which shows hexyl-2-APB (7) not forming a ring. However, we should
a heterocyclic form in staggered arrays, with each molecule caution extrapolating NMR data performed in dimethylsulf-
linked to two others through hydrogen bonds. An examina- oxide and biological data in which cells are incubated in an
tion of the NMR of 2-APB in d6-DMSO shows that the NH2 aqueous buffered salt solution at pH 7.4. Under these aque-
protons are indeed in a highly deshielded position (6–7 ppm), ous conditions, we do not know the proportion of open-chain,
which is also suggestive of a ring rather than of an open- ring, and dimer forms of 2-APB and its analogs.
chain structure. In contrast, the NH2 peak for ethanolamine
The presence of a chiral center gives rise to the interesting
is found at 3.3 ppm. This series of 2-APB analogs was de- splitting pattern of the adjacent protons. A distinctive struc-
signed to examine this ring concept for 2-APB and also to tural feature of the two stereoisomers (compounds 11 and 12)
determine whether a five-membered ring was needed or produced in this study was observed in their NMR spectra.
whether larger rings would be active (compounds 6 and 7). As might be expected, the two R- and S-enantiomers would
The objective was to determine whether chain extension necessarily have the same basic NMR patterns, and indeed,
could resolve the question of the coordination of the NH2 this is observed in d6-DMSO. The NMR data further suggest
group to boron. The N-methyl (compounds 2 and 3), and that the two NH2 protons in the R- and S-compounds are
N-butyl (4) analogs, as well as the phenyl (5) and butyl (6) essentially nonequivalent. This nonequivalence produces a
analogs, all showed the NH and NH2 protons to be highly triplet arising from the mutual splitting of the two NH2
deshielded in the 5.9- to 6.8-ppm region in d6-DMSO. This protons and subsequent splitting by the adjacent proton on
would suggest that these analogs were also in a similar ring the optically active center. The protons adjacent to the oxy-
structure because of the proximity of the NH and NH2 groups gen in these two compounds are also apparently nonequiva-
to the electron-deficient boron. However, NMR data for the lent as well and are split into a set of triplets. The protons at
hexyl analog (compound 7) in d6-DMSO showed the NH2 the chiral center of this oxazaborolidine ring system are also
protons to be at 3.3 ppm, which is in keeping with an open- nonequivalent and are therefore split into a set of doublets of
chain structure. Thus, extending the length of the ethanol- doublets.
TABLE 1
The effect of 2-APB derivatives and nonboron-containing analogs on
thrombin-induced increases in ͓Ca2ϩ͔ in the presence of extracellular
i
calcium in platelets (predominantly calcium influx)
Abbreviated trivial names are shown for simplicity, together with a designated
number for each compound (Fig. 3).
Inhibition of Thrombin Response
at Compound Concentration
Compound Name
(Number)
100 ␮M
10 ␮M
1 ␮M
%
2-APB (1)
98 Ϯ 6
96 Ϯ 1
99 Ϯ 1
98 Ϯ 1
96 Ϯ 1
97 Ϯ 3
98 Ϯ 0
47 Ϯ 8
31 Ϯ 10
45 Ϯ 6
43 Ϯ 7
52 Ϯ 6
48 Ϯ 9
39 Ϯ 9
74 Ϯ 3
55 Ϯ 10
Ϫ2 Ϯ 3
86 Ϯ 8
82 Ϯ 7
37 Ϯ 9
55 Ϯ 5
17 Ϯ 8
62 Ϯ 6
16 Ϯ 4
7 Ϯ 5
15 Ϯ 8
11 Ϯ 4
10 Ϯ 2
8 Ϯ 5
15 Ϯ 3
14 Ϯ 7
3 Ϯ 12
N.D.
34 Ϯ 11
48 Ϯ 5
15 Ϯ 13
19 Ϯ 10
Ϫ7 Ϯ 9
36 Ϯ 8
Monomethyl APB (2)
Dimethyl APB (3)
t-Butyl APB (4)
Phenyl APB (5)
Butyl APB (6)
Hexyl APB (7)
a
Phenyl pyridyl APB (8)
Methyl pyridyl APB (9)
Glycine ester APB (10)
(R)-Benzyl APB (11)
(S)-Benzyl APB (12)
Methyloxazolidine (13)
t-Butoxazolidine (14)
Diphenylbutyrolactone (15)
Diphenyltetrahydrofuranb
—
98 Ϯ 1
16 Ϯ 1
97 Ϯ 5
91 Ϯ 6
53 Ϯ 7
83 Ϯ 5
59 Ϯ 4
69 Ϯ 9
Fig. 6. Proposed structure of the acetonitrile adducts of diphenylborinic
acid anhydride. Product A is a diacetonitrile adduct of diphenylborinic
anhydride, and product B is a monoacetonitrile adduct of diphenylborinic
acid that could result from the hydrolysis of adduct A. The structure
shown in C is the result of combining acetonitrile with boron trifluoride,
and this adduct is commercially available (Sigma-Aldrich), whereas ad-
duct D (which has been crystalized) is the result of reaction between
imidazole with boron trifluoride.
N.D., not determined.
a Stimulates calcium influx by itself at 100 ␮M.
b From Dobrydneva and Blackmore (2001).

2-APB Analogs Inhibit Calcium Influx in Human Platelets
253
The rationale for the synthesis of the 2,2-diphenyl-1,3,2- cluded phenyl-APB (5), dimethyl-APB (3), monomethyl-APB
oxazaborolidine-5-one (compound 10) from the condensation (2), t-butyl-APB (4), butyl-APB (6), hexyl-APB (7), and meth-
of diphenylborinic anhydride and the sodium salt of glycine yl-pyridyl APB (9). Dose responses are shown for 2-APB and
was taken from the possibility that greater activity might be butyl-APB in Fig. 7. Each of these compounds produced be-
achieved by creating a better leaving group within the 2-APB tween 96 and 99% inhibition at 100 ␮M, between 31 and 74%
structure. This would assume that the activity of this class of inhibition at 10 ␮M, and between 3 and 16% inhibition at 1
compounds may be related to an interaction of some nucleo- ␮M. Phenyl-pyridyl APB (8) stimulated Ca2ϩ influx at 100
philic moiety on the receptor/channel protein with the highly ␮M by itself; however, at 10 and 1 ␮M, it was inhibitory. The
polarized 2-APB ring system. This might lead to the subse- response of 100 ␮M phenyl pyridyl APB to increase [Ca2ϩ
]
i
quent displacement of the coordinately bound NH2 group was approximately 50% of the increase observed with 0.01
from the 2-APB moiety and provide for a reversible attach- units/ml thrombin (Fig. 7). The kinetics of the increase in
ment to the receptor/channel site. By introducing the ester [Ca2ϩ]i was similar to that observed with thrombin and much
moiety into the basic 2-APB structure, this type of interac- faster than that seen initially with the sarcoplasmic/endo-
tion was expected to be enhanced and may produce a more plasmic reticulum calcium ATPase (SERCA) inhibitor thap-
effective interaction with the SOCE receptor protein. How- sigargin. Thapsigargin initially produces a very small and
ever, this compound was a very weak inhibitor of thrombin- slow increase in [Ca2ϩ]i, (also Fig. 4 in Dobrydneva and
induced Ca2ϩ influx at 100 ␮M (see below).
Blackmore, 2001), then after approximately 1 min, there is a
2-APB Derivatives Block Thrombin-Induced Ca2؉ El- rapid and large increase in [Ca2ϩ]i, probably caused by the
evation. In the present study, we examined the capacity of generation of thromboxane, because the response was largely
the various analogs to inhibit the ability of thrombin to blocked by treating the platelets with aspirin (data not
increase [Ca2ϩ]i in platelets in the presence of extracellular shown). Thus, we do not believe that phenyl pyridyl APB was
calcium (Table 1) and in the absence of extracellular calcium increasing [Ca2ϩ]i by a mechanism involving the inhibition of
(Table 2). When extracellular calcium was present, approxi- SERCA because the kinetics and magnitude of increase in
mately 80 to 90% of the increase in [Ca2ϩ]i induced by low [Ca2ϩ
] are completely different from that seen with the
i
concentrations of thrombin (0.01 units/ml) was caused by SERCA inhibitor thapsigargin. We have not examined this
calcium influx (Dobrydneva and Blackmore, 2001). This was phenomenon any further. There were two analogs that
because in the absence of extracellular calcium, the increase showed a greater potency than 2-APB to inhibit thrombin:
in [Ca2ϩ]i induced by thrombin was approximately 10 to 20% these were (R)- and (S)-benzyl-APB (compounds 11 and 12)
of that observed when extracellular calcium was present.
Modification of the oxazaborolidine ring in the 2-APB an-
(Fig. 8).
Previous studies in platelets showed that 2-APB inhibited
alogs does not alter activity substantially, except for the calcium influx channels (Diver et al., 2001; Dobrydneva and
glycyl ester (compound 10), which had a small 16% inhibitory Blackmore, 2001). 2-APB also inhibited calcium mobilization
activity at 100 ␮M and no activity at lower concentrations in platelets (Maruyama et al., 1997; Diver et al., 2001; Do-
(Table 1). Most of the 2-APB (compound 1) analogs that we brydneva and Blackmore, 2001). We therefore also examined
synthesized possessed approximately the same ability to in- the effects of the 2-APB analogs on thrombin-induced cal-
hibit thrombin-induced [Ca2ϩ]i increase (Table 1). This in- cium mobilization (Table 2) to determine whether they were
TABLE 2
The effect of 2-APB derivatives and nonboron-containing analogs on
thrombin-induced increases in ͓Ca2ϩ͔ in the absence of extracellular
calcium in platelets (internal calciumi mobilization)
Abbreviated trivial names are shown for simplicity together with a designated
number for each compound (Fig. 3).
Inhibition of Thrombin Response
at Compound Concentration
Compound Name
(Number)
100 ␮M
10 ␮M
1 ␮M
%
2-APB (1)
91 Ϯ 2
90 Ϯ 2
86 Ϯ 6
89 Ϯ 2
96 Ϯ 2
91 Ϯ 4
90 Ϯ 3
88 Ϯ 6
54 Ϯ 9
16 Ϯ 1
16 Ϯ 2
5 Ϯ 3
25 Ϯ 4
20 Ϯ 1
23 Ϯ 4
24 Ϯ 5
16 Ϯ 4
0 Ϯ 5
36 Ϯ 5
3 Ϯ 3
0 Ϯ 4
4 Ϯ 8
5 Ϯ 4
6 Ϯ 3
8 Ϯ 8
10 Ϯ 8
0 Ϯ 5
N.D.
Monomethyl APB (2)
Dimethyl APB (3)
t-Butyl APB (4)
Phenyl APB (5)
Butyl APB (6)
Hexyl APB (7)
Phenyl pyridyl APB (8)
Methyl pyridyl APB (9)
Glycine ester APB (10)
(R)-Benzyl APB (11)
(S)-Benzyl APB (12)
Methyloxazolidine (13)
t-Butoxazolidine (14)
Diphenylbutyrolactone (15)
Diphenyltetrahydrofuranb
a
—
29 Ϯ 16
100 Ϯ 1
95 Ϯ 6
24 Ϯ 9
94 Ϯ 6
13 Ϯ 7
51 Ϯ 7
49 Ϯ 9
59 Ϯ 11
14 Ϯ 9
37 Ϯ 2
0 Ϯ 3
12 Ϯ 7
13 Ϯ 6
0 Ϯ 4
16 Ϯ 6
N.D.
Fig. 7. Effect of phenyl pyridyl APB (8), thrombin, and thapsigargin to
increase [Ca2ϩ]i in platelets in the presence of extracellular calcium. The
concentration of phenyl pyridyl APB (8) was 100 ␮M, thapsigargin was
100 nM, and two concentrations of thrombin (0.005 and 0.01 units/ml)
were used. Representative traces are shown.
33 Ϯ 7
30 Ϯ 11
N.D., not determined.
a Stimulates calcium release by itself at 100 ␮M.
b From Dobrydneva and Blackmore (2001).

254
Dobrydneva et al.
more selective at inhibiting either process. When one com- Methylation (2), dimethylation (3), and t-butylation (4) or
pares the inhibitory effects of each compound at a 10 ␮M fusing of a benzo ring to the oxazaborolidine ring at this
concentration, 2-APB was the most effective compound at nitrogen (8 and 9) had very little influence on activity. The
inhibiting internal release with the exception of (S)-benzyl addition of a phenyl group at position 5 of the oxazaborolidine
APB (compound 12), whereas at a 1 ␮M concentration, 2-APB ring did not influence activity at a concentration of 10 ␮M (5
was clearly the most effective inhibitor, with a 36% inhibition and 8). At 100 ␮M, compound 8 by itself promoted calcium
(Table 2). When extracellular calcium was present and the influx; however, at 10 ␮M, compound 8 had no effect on
inhibitory effects of each compound at a 10 ␮M concentration calcium influx by itself, but it produced a good inhibition
were compared (Table 1), 2-APB possessed an inhibitory (77%) of thrombin, so that this compound was active. The
activity that was similar to that of many of the compounds addition of a carbonyl group at position 5 of the oxazaboroli-
(e.g., compounds 3, 4, and 6); however, several analogs were dine ring (10) drastically reduced activity.
clearly more effective (e.g., compounds 8, 9, 11, 12, and 14).
Substitution at position 4 of the oxazaborolidine with a
Several of the analogs were also more effective at the 1 ␮M benzyl group (compounds 11 and 12) actually increased the
concentration (e.g., compounds 11, 12, and 14). Thus, many of calcium-blocking activity of 2-APB by approximately 1 order
the analogs synthesized had smaller inhibitory effects on of magnitude. This is a curious finding because carbon 4 is a
internal release (Table 2) compared with 2-APB, whereas chiral center, which would place the benzyl groups in both
several analogs were more effective at inhibiting calcium compounds in different orientations relative to the ox-
influx (e.g., compounds 11 and 12). Despite extensive modi- azaborolidine ring. This lack of stereoselectivity between
fications to 2-APB, many of the compounds are able block compounds could imply that there must be a large hydropho-
both calcium influx and internal release, at least in platelets. bic binding pocket that can tolerate binding of the benzyl
From this limited number of analogs examined, we propose groups when presented in two different orientations (Fig. 9).
a common pharmacophore that is responsible for inhibiting
Another feature of the boron-containing compounds pre-
thrombin-induced calcium influx. This pharmacophore con- sented in this study that should be considered is their ability
sists of two phenyl groups linked to either tetrahedral carbon to dimerize. However, because comparable calcium-blocking
or boron. This is confirmed by the fact that replacing the activity was also observed with carbon analogs of 2-APB
boron with a carbon such as in compounds 13 and 14 resulted (compounds 13 and 14), which cannot dimerize, dimers do not
in inhibitory activity comparable with that seen with most of contribute to their inhibitory activity. We showed previously
the boron-containing analogs. Therefore, it does not seem that diphenyl tetrahydrofuran (nonboron analog of 2-APB
that the presence of boron is critical for activity.
without a nitrogen in the tetrahydrofuran ring) possessed the
All of the most active analogs possessed oxygen at position ability to inhibit thrombin-induced calcium influx (Table 1
1 of the oxazolidine or oxazaborolidine ring. An exception was and Dobrydneva and Blackmore, 2001). This compound nev-
compound 15, which had oxygen in a different position in the ertheless was not as effective as 2-APB at 100 ␮M; however,
ring. Compound 15 was still inhibitory, although it was at lower concentrations, it was more effective than 2-APB
clearly less effective than compounds containing oxygen at (Table 1). Thus, the absence of nitrogen and boron did not
position 1; it possessed 59% inhibitory activity at 100 ␮M and have a deleterious effect on activity. We also tested another
only 17% inhibitory activity at 10 ␮M and no activity at 1 ␮M nonboron analog, 3,3-diphenyl dihydrofuran-2-one (com-
(Table 1). In addition, the replacement of the oxygen at pound 15). This compound was not as effective as 2-APB and
position 1 with a carbonyl group, such as in phenytoin, re- many of its analogs (Table 1). This compound contains an
sulted in only a 22% inhibition at 100 ␮M (Dobrydneva and oxygen at position 4 and a carbonyl at position 3 (relative to
Blackmore, 2001). All of the active analogs possessed nitro- oxygen in 1,3-oxazolidine); therefore, the location of the oxy-
gen at position 3 of the oxazolidine or oxazaborolidine ring. gen and/or the carbonyl group in the ring seems critical for
activity. In addition, the absence of nitrogen in the ring may
contribute to weak activity.
Because the 2-APB analogs possess the potential to dimer-
ize and cyclize because of the formation of an N3B coordi-
Fig. 9. Structure of diphenyl oxazaborolidine and oxazolidine rings;
Fig. 8. Dose response of 2-APB (1), butyl APB (6), (S)-benzyl APB (11),
and (R)-benzyl APB (12) to inhibit the ability of thrombin (0.01 units/ml)
to increase [Ca2ϩ]i in platelets in the presence of extracellular calcium.
Data are mean Ϯ S.E.M. from four separate experiments.
showing the position of each modification used in the present study.
Atoms in the oxazaborolidine or oxazolidine are numbered 1 through 5
(Table 3). Substituents to the oxazaborolidine and oxazolidine rings are
labeled R1 through R5 (Table 3).

2-APB Analogs Inhibit Calcium Influx in Human Platelets
255
nate bond, it is also possible that they could also form N3B although this effect is cell-specific (Soulsby and Wojcikiewicz,
coordinate bonds with various amino groups in the calcium 2002). In addition, 2-APB also inhibits IP3Rs ubiquitination
channel. The reaction between acetonitrile and diphenyl- and proteasomal degradation in some cells (Soulsby and
borinic anhydride (Fig. 6) illustrates this phenomenon. A Wojcikiewicz, 2002). 2-APB has also been shown to inhibit
variety of nitrogen adducts of BF3 have been previously char- smooth endoplasmic reticulum calcium ATPase pumps in
acterized crystallographically, and these include amines, im- some cells, thus causing a slow increase in [Ca2ϩ]i (Bootman
idazole (Barber et al., 2005; Fig. 6D), pyridine, and acetoni- et al., 2002). In addition, 2-APB prevents mitochondria from
trile (Swanson et al., 1969; Fig. 6C) adducts (Barber et al., releasing Ca2ϩ (Prakriya and Lewis, 2001). The best effect of
2005). It is therefore conceivable that the boron analogs 2-APB so far characterized is its ability to inhibit store-
synthesized in the present study can also form coordinate mediated calcium influx in a variety of cells, including plate-
bonds with amino acids (either N3B or O3B). There is lets (Dobrydneva and Blackmore, 2001; Bootman et al.,
precedence for the boron-containing organic molecule bort- 2002). There are many TRP channels that are inhibited by
ezomib forming a coordinate bond with the N-terminal thre- 2-APB in the 10 to 100 ␮M range, and these include TRPC1,
onine residue of the 26S proteasome (Pekol et al., 2005). If TRPC3, TRPC5, TRPC6, TRPV6, TRPM3, TRPM7, TRPM8,
N3B coordinate bonds were to occur in the cell, then these
boron-containing compounds would perhaps bind more
tightly to the calcium channel than their nonboron contain-
ing counterparts because of coordinate bond formation. Table
3 summarizes all of the modifications to 2-APB that were
prepared and what influence they had on activity. It should
be pointed out that in the present study, we have not at-
tempted to alter the diphenyl groups to determine whether
such modifications would influence activity.
This study shows that 2-APB can undergo significant mod-
ifications that have little effect on its ability to inhibit throm-
bin-induced Ca2ϩ influx in human platelets. However, we did
find that two benzyl derivatives (compounds 11 and 12) were
more potent than the parent 2-APB compound. This gives us
encouragement that further modifications to the 2-APB mol-
ecule can increase its potency even further. We did find
previously that diphenyltetrahydrofuran (Dobrydneva and
Blackmore, 2001; Table 1) produced approximately 70% in-
hibition of the thrombin response at 100 ␮M; however, the
IC50 value was approximately 1 ␮M. Thus, diphenyltetrahy-
drofuran was the most potent inhibitor that we have found so
far, although it was not as effective as 2-APB and many of its
analogs, which produce approximately 100% inhibition at
100 ␮M (Table 1).
and TRPP2 (Xu et al., 2005). Other members of the TRP
channels are either activated or unaffected 2-APB. 2-APB
also has been shown to modulate the store-operated calcium
release-activated Ca2ϩ channels (e.g., Braun et al., 2001) and
the Mg2ϩ-inhibited cation channel (Kozak et al., 2002;
Prakriya and Lewis, 2002).
It is of interest to compare the effects of the 2-APB analogs
used in the current study on other parameters that influence
[Ca2ϩ]i in a variety of cell types. Different 2-APB analogs
may have selective inhibitory or stimulatory effects on the
various TRP channels, and they are likely to be very useful in
elucidating the nature of the TPR channels present in vari-
ous cells.
References
Azzena U, Melloni G, and Nigra C (1993) Reductive cleavage of N-substituted
2-aryl-1,3-oxazolidines: generation of alpha-amino-substituted carbanions. J Org
Chem 58:6707–6711.
Bakowski D, Glitsch MD, and Parekh AB (2001) An examination of the secretion-like
coupling model for the activation of the Ca2ϩ release-activated Ca2ϩ current Icrac
in RBL-1 cells. J Physiol (Lond) 532:55–71.
Barber RA, Bull AEA, Charmant JPH, Norman NC, and Orpen AG (2005) N-
Imidazole-boron trichloride adduct. Acta Crystallogr E61:o553–o554.
Bootman MD, Collins TJ, Mackenzie L, Roderick HL, Berridge MJ, and Peppiatt CM
(2002) 2-Aminoethoxydiphenyl borate (2-APB) is
a reliable blocker of store-
operated Ca2ϩ entry but an inconsistent inhibitor of InsP3-induced Ca2ϩ release.
FASEB J 16:1145–1150.
Braun F-J, Broad LM, Armstrong DL, and Putney JW Jr (2001) Stable activation of
single CRAC-channels in divalent cation-free solutions. J Biol Chem 276:1063–
1070.
As indicated by Bootman et al. (2002), 2-APB influences
calcium homeostasis in a variety of cells by several different
mechanisms. These include the original finding that 2-APB
inhibits calcium release via IP3Rs (Maruyama et al., 1997),
Broad LM, Braun FJ, Lievremont JP, Bird GS, Kurosaki T, and Putney JW Jr (2001)
Role of the phospholipase C-inositol 1,4,5-trisphosphate pathway in calcium re-
lease-activated calcium current and capacitative calcium entry. J Biol Chem
276:15945–15952.
Brownlow SL and Sage SO (2003) Rapid agonist-evoked coupling of type II Ins (1,4,5)
TABLE 3
P3 receptor with human transient receptor potential (hTRPC1) channels in human
Summary of the modifications in the oxazaborolidine or oxazolidine
ring (Fig. 9) that influence the ability of thrombin to stimulate Ca2ϩ
influx in human platelets
The introduction of carbonyl groups or oxygen into the five-membered ring tends to
decrease activity. The introduction of a benzyl on carbon 4 increased activity of the
parent 2-APB molecule. There were many modifications and substitutions on the
ring that could be tolerated and which had minimal effect on activity.
platelets. Biochem J 375:697–704.
Chung MK, Guler AD, and Caterina MJ (2005) Biphasic currents evoked by chemical
or thermal activation of the heat-gated ion channel, TRPV3. J Biol Chem 280:
15928–15941.
Diver JM, Sage SO, and Rosado JA (2001) The inositol trisphosphate receptor
antagonist 2-aminoethoxydiphenylborate (2-APB) blocks Ca2ϩ entry channels in
human platelets: cautions for its use in studying Ca2ϩ influx. Cell Calcium 30:
323–329.
Dobrydneva Y and Blackmore PF (2001) 2-Aminoethoxydiphenyl borate directly
inhibits store-operated calcium entry channels in human platelets. Mol Pharmacol
60:541–552.
Dobrydneva Y, Williams R, and Blackmore PF (2001) 2-Aminoethoxydiphenyl bori-
nate (2APB): a direct inhibitor of Ca2ϩ influx channels in human platelets (Ab-
stract). FASEB J 15:A926.
Dobrydneva Y, Williams RL, and Blackmore PF (2002) 2-APB (2-aminoethoxyphenyl
borate) as a prototype drug for a group of structurally related calcium channel
blockers in human platelets (Abstract). FASEB J 16:A936.
Dovel B, Williams RL, Dobrydneva Y, and Blackmore PF (2002) The design, synthe-
sis and evaluation of novel calcium channel blockers in human platelets. VA J Sci
53:86.
Position in
More
Oxazaborolidine
Active
Less Active
Active
or Oxazolidine Ring
1
Oxygen
Carbonyl
2
Boron, carbon
Nitrogen
Carbon
3
Carbon, carbonyl
Oxygen
Carbonyl, oxygen
4
5
Substituent group
R1
Carbon
Methyl, dimethyl,
t-butyl
Pyridyl
Farfan N, Castillo D, Joseph-Nathan P, Contreras R, and Szentpaly L (1992)
Through-bond modulation of N3B ring formation shown by NMR and X-ray
diffraction studies of borate derivatives of pyridyl alcohols. J Chem Soc Perkin
Trans II 4:527–532.
R2
R3
R4
R5
Benzyl
Ethyl, butyl
Methyl, phenyl
Gregory RB, Rychkov G, and Barritt GJ (2001) Evidence that 2-aminoethyl diphe-
nylborate is a novel inhibitor of store-operated Ca2ϩ channels in liver cells and acts

256
Dobrydneva et al.
through a mechanism which does not involve inositol trisphosphate receptors.
Biochem J 354:285–290.
Hu HZ, Gu Q, Wang C, Colton CK, Tang J, Kinoshita-Kawada M, Lee LY, Wood JD,
and Zhu MX (2004) 2-Aminoethoxydiphenyl borate is a common activator of
TRPV1, TRPV2 and TRPV3. J Biol Chem 279:35741–35748.
Iwasaki H, Mori Y, Hara Y, Uchida K, Zhou H, and Mikoshiba K (2001) 2-Aminoe-
thoxydiphenyl borate (2-APB) inhibits capacitative calcium entry independently of
the function of inositol 1,4,5-trisphosphate receptors. Recept Channels 7:429–439.
Konkova SG, Badasyan AE, Khachatryan AKh, and Sargsyan MS (1999) Some
questions of ring-chain tautomerism in hydroxy-containing imines. Khimicheskii
Zhurnal Armenii 52:16–21.
boroxazolidine and the orthorhombic form of B,B-diphenylboroxazolidine. Can
J Chem 54:3131–3141.
Rosado JA, Lopez JJ, Harper AG, Harper MT, Redondo PC, Pariente JA, Sage SO,
and Salido GM (2004a) Two pathways for store-mediated calcium entry differen-
tially dependent on the actin cytoskeleton in human platelets. J Biol Chem
279:29231–29235.
Rosado JA, Redondo PC, Sage SO, Pariente JA, and Salido GM (2005) Store-operated
Ca2ϩ entry: vesicle fusion or reversible trafficking and de novo conformational
coupling? J Cell Physiol 205:262–269.
Rosado JA, Redondo PC, Salido GM, Gomez-Arteta E, Sage SO, and Pariente JA
(2004b) Hydrogen peroxide generation induces pp60src activation in human plate-
lets: evidence for the involvement of this pathway in store-mediated calcium entry.
J Biol Chem 279:1665–1675.
Kozak JA, Kerschbaum HH, and Cahalan MD (2002) Distinct properties of CRAC
and MIC channels in RBL cells. J Gen Physiol 120:221–235.
Rosado JA and Sage SO (2001) Role of the ERK pathway in the activation of
store-mediated calcium entry in human platelets. J Biol Chem 276:15659–15665.
Soulsby MD and Wojcikiewicz RJ (2002) 2-Aminoethoxydiphenyl borate inhibits
inositol 1,4,5-trisphosphate receptor function, ubiquitination and downregulation,
but acts with variable characteristics in different cell types. Cell Calcium 32:175–
181.
Staudinger H and Freudenberger H (1943) Thiobenzophenone. Org Synth Coll Vol
2:573–574.
Swanson R, Shriver DF, and Ibers JA (1969) Nature of the donor-acceptor bond in
acetonitrile-boron trihalides. The structure of boron trifluoride and boron trichlo-
ride complexes of acetonitrile. Inorg Chem 8:2182–2189.
Thadigiri C, Williams RL, Dobrydneva Y, and Blackmore PF (2003) Synthesis and
evaluation of novel calcium channel blockers in human platelets. VA J Sci 54:79.
van Rossum DB, Patterson RL, Ma HT, and Gill DL (2000) Ca2ϩ entry mediated by
store depletion, S-nitrosylation and TRP3 channels. Comparison of coupling and
function. J Biol Chem 275:28562–28568.
Xu SZ, Zeng F, Boulay G, Grimm C, Harteneck C, and Beech DJ (2005) Block of
TRPC5 channels by 2-aminoethoxydiphenyl borate: a differential, extracellular
and voltage-dependent effect. Br J Pharmacol 145:405–414.
Ma HT, Patterson RL, van Rossum DB, Birnbaumer L, Mikoshiba K, and Gill DL
(2000) Requirement of the inositol trisphosphate receptor for activation of store-
operated Ca2ϩ channels. Science (Wash DC) 287:1647–1651.
Ma HT, Venkatachalam K, Li HS, Montell C, Kurosaki T, Patterson RL, and Gill DL
(2001) Assessment of the role of the inositol 1, 4, 5-trisphosphate receptor in the
activation of transient receptor potential channels and store-operated Ca2ϩ entry
channels. J Biol Chem 276:18888–18896.
Maruyama T, Kanaji T, Nakade S, Kanno T, and Mikoshiba K (1997) 2APB, 2-ami-
noethoxydiphenyl borate,
a membrane-penetrable modulator of Ins(1,4,5)P3-
induced Ca2ϩ release. J Biochem 122:498–505.
Parekh AB and Putney JW (2005) Store-operated channels. Physiol Rev 85:757–810.
Pekol T, Daniels JS, Labutti J, Parsons I, Nix D, Baronas E, Hsieh F, Gan LS, and
Miwa G (2005) Human metabolism of the proteasome inhibitor bortezomib: iden-
tification of circulating metabolites. Drug Metab Dispos 33:771–777.
Prakriya M and Lewis RS (2001) Potentiation and inhibition of Ca2ϩ release-
activated Ca2ϩ channels by 2-aminoethyldiphenyl borate (2-APB) occurs indepen-
dently of IP3 receptors. J Physiol 536:3–19.
Prakriya
M and Lewis RS (2002) Separation and characterization of currents
through store-operated CRAC channels and Mg2ϩ-inhibited cation (MIC) chan-
nels. J Gen Physiol 119:487–507.
Redondo PC, Harper AG, Salido GM, Pariente JA, Sage SO, and Rosado JA (2004) A
role for SNAP-25 but not VAMPs in store-mediated Ca2ϩ entry in human platelets.
J Physiol 558:99–109.
Address correspondence to: Dr. Peter F. Blackmore, Department of Phys-
iological Sciences, Eastern Virginia Medical School, P.O. Box 1980, Norfolk,
VA 23501. E-mail: blackmpf@evms.edu
Rettig SJ and Trotter J (1976) Crystal and molecular structure of B,B-bis(p-toly)-