Eur JMed Chem (1995) 30,387-394
`0 Elsevier, Paris
`
`387
`
`factor Xa and DX=9065a, a
`Molecular model of an interaction between
`novel fact& Xa inhibitor:
`contribution
`of the acetimidoylpyrrolidine
`moiety
`of the inhibitor
`to potency and selectivity
`for serine proteases
`
`S Katakura, T Nagahara, T Hara, S Kunitada, M Iwamoto
`
`Exploratory Research Laboratories
`
`Co Ltd, l-16-13, Kita-Kasai,
`2, Daiichi Pharmaceutical
`(Received 18 October 1994; accepted 21 December 1994)
`
`Edogawa-ku,
`
`Tokyo 134, Japan
`
`Summary - Molecular modeling
`has demonstrated
`inhibitor,
`a novel FXa
`factor Xa (FXa) and DX-9065a,
`of a complex between
`salt-bridge
`interaction
`of the amidinonaphthalene
`moiety of the inhibitor
`to the Sl site pocket of the enzyme and a crucial carboxyl
`group of the inhibitor
`for the FXa/thrombin
`selectivity.
`In the present study, we synthesized
`some DX-9065a
`derivatives and studied
`their interaction modes with FXa,
`trypsin and thrombin
`to evaluate the role of the acetimidoylpyrrolidine
`moiety. The docking study of
`the inhibitor
`in the FXa model provided
`a molecular model of the inhibitor-FXa
`complex showing
`a hydrophobic
`interaction
`of the
`pyrrolidine
`ring with an aryl binding
`site of FXa, and a hydrogen bond between
`the acetimidoyl
`group and the backbone carbonyl
`oxygen of Glu97. Comparison
`between
`the interaction mode of DX-9065a
`with FXa and those with
`trypsin and thrombin
`explained
`the inhibitory
`potencies of the enzymes.
`
`a
`
`molecular modeling / factor Xa / DX-9065a / thrombin / trypsin
`
`Introduction
`
`in-
`serine protease
`is a trypsin-like
`Factor Xa (FXa)
`volved
`in the cascade of blood coagulation. Although
`FXa is a promising
`target for anticoagulant
`agents [l],
`no synthetic
`inhibitor
`of FXa has been developed yet.
`In an earlier study, we reported a novel FXa
`inhibitor,
`(2S)-2-[4-[[(33)-
`acetimidoyl-3-pyrrolidinylloxylphe-
`nyl]-3-(7-amidino-2-naphthyl)propanoic
`acid hydro-
`chloride
`pentahydrate,
`DX-9065a
`(fig 1)
`[2]. This
`compound has a potent
`inhibitory
`activity
`for FXa and
`high selectivity
`for FXa over other serine proteases.
`In a previous paper, we investigated
`the binding
`mode of DX-9065a
`by computer molecular modeling
`[3]. A docking study of DX-9065a
`into a three-dimen-
`sional model of FXa, based on the X-ray crystallogra-
`phic data for trypsin
`[4],
`indicated
`that the amidino-
`naphthalene moiety of the
`inhibitor
`binds
`to the Sl
`site of FXa with a salt-bridge
`interaction.
`Further-
`more,
`the Glu192 of thrombin
`corresponding
`to the
`Gln192 of FXa evokes an electrostatic
`repulsion of the
`carboxyl
`group of
`the
`inhibitor,
`which explains
`a
`mechanism
`of
`the FXa/thrombin
`selectivity
`of DX-
`9065a. However, DX-9065a
`has two basic terminals
`in
`the structure, suggesting
`that
`the acetimidoylpyrroli-
`dine and
`the amidinonaphthalene
`moieties
`signifi-
`cantly participate
`in
`the potent
`inhibitory
`activity
`against FXa.
`
`moiety also may contri-
`The acetimidoylpyrrolidine
`bute
`to
`the selectivity
`of DX-9065a
`for FXa
`from
`thrombin.
`The
`interaction
`of the moiety with a site of
`thrombin may
`fix the direction of the carboxyl group
`of DX-9065a, which
`leads
`to an increase
`in the elec-
`trostatic
`repulsion between
`the carboxyl group of the
`inhibitor
`and Glu192 of
`the enzyme. Thereby, DX-
`9065a cannot bind
`to thrombin.
`In addition,
`a loss of
`such an electrostatic
`repulsion
`between
`trypsin and
`DX-9065a
`suggests
`that
`the carboxyl group does not
`contribute
`to the selectivity
`for FXa
`from
`trypsin, and
`the acetimidoylpyrrolidine
`moiety
`should be involved
`in the selectivity.
`a contribution
`In the present study, we determined
`of the acetimidoylpyrrolidine
`moiety of DX-9065a
`
`in
`
`Fig 1. Structure
`a conformational
`
`Torsion angles rotated during
`of DX-9065a.
`search are shown with arrows.
`
`MYLAN - EXHIBIT 1015
`
`

`
`388
`
`for FXa by a molecular
`the potency and selectivity
`modeling
`technique. First, we synthesized a series of
`compounds
`7a-d
`in order
`to evaluate
`the interactions
`of
`the acetimidoylpyrrolidine
`moiety
`with
`FXa,
`thrombin
`and
`trypsin. Second, diverse
`low-energy
`conformers
`of DX-9065a
`were generated
`by
`the
`conformational
`search using a Metropolis
`Monte-
`Carlo method. Finally, on the basis of the SAR of the
`series of inhibitors,
`docking
`studies of these confor-
`mers into FXa, thrombin
`and trypsin were performed.
`
`Chemistry
`
`the method
`by
`6a-c were prepared
`Compounds
`[5]. Treatment
`of
`described previously
`(scheme
`1)
`in ethanol and
`compounds
`6b,c with hydrochloride
`then with an ammonia/ethanol
`solution
`gave
`the
`amidine
`derivatives,
`which were hydrolyzed
`with
`hydrochloric
`acid
`to afford compounds
`7b,c. Acid
`of
`hydrolysis
`the amidine
`intermediate
`synthesized
`from 6a gave the phenol derivative 7d. Compound
`7a
`
`pathway with a
`following
`the
`was prepared along
`of
`the amidino
`group with
`modification:
`protection
`tert-butoxy dicarbonate, alkali hydrolysis with sodium
`hydroxide and deprotection with trifluoroacetic
`acid.
`
`Molecular modeling
`
`of serine proteases
`
`structure model of the heavy
`three-dimensional
`The
`chain of human FXa was constructed on the basis of
`trypsin,
`a highly
`homologous
`serine protease,
`as
`described previously
`[3]. A recent X-ray
`crystallo-
`graphic analysis of FXa revealed
`the geometric
`simi-
`larity of the a-carbons of FXa
`to those of thrombin,
`except
`for an insertion
`loop
`in thrombin,
`5r60A
`to
`Thr60I
`[6]. On the basis of this report,
`the coordinates
`from Asp170
`to Ile176 of the FXa model, which were
`topologically
`different
`from
`those of thrombin, were
`replaced by the corresponding
`coordinates of throm-
`bin, Leul70
`to
`Ile176. Side chains of the resulting
`model were optimized
`manually
`to remove
`steric
`clashes, and
`then
`the FXa model was obtained
`by
`
`Et02C
`
`0
`
`3
`
`0
`
`~
`
`EGC
`
`OH
`
`(a) or (b)
`
`W
`
`4a R1= cyclopentyl
`4b R,= CH&H2NHBoc
`4e R,= CHS
`
`6a R,= cydopentyl
`6b R,= CH&H2NHBoc
`6c R,= CHB
`
`“9
`
`7b R,= CH2CH2NH2
`7c R,= CH3
`7d R,= H
`
`- HCI
`
`Scheme 1. (a) R,OH, diethyl azodicarboxylate, PPh,, THF; (b) R,I, K&OX, DMF; (c) DBU, THF/EtOH; (d) Pd/C, HZ, EtOH;
`(e) EtOH/HCl; EtOH/NH,; tert-butoxy dicarbonate, Et,N, DMF; (f) 1 N NaOH; TFA; (g) EtOH/HCl; EtOH/NH3; 4 N HCl, A.
`
`

`
`in 500 steps
`of these residues
`energy minimization
`without moving
`the other amino-acid
`residues.
`The
`three-dimensional
`structure model
`of human
`a-thrombin
`was derived
`from
`the coordinates
`for the
`structure
`labeled 1DWE
`[7] in the Brookhaven Protein
`Data Bank
`[8]. Water molecules
`in
`the coordinates
`were removed and the hydrogens were added
`to the
`backbone
`and side chains based on
`the standard
`average bond angles and lengths.
`registered
`trypsin
`The atomic coordinates of bovine
`as 1TPP
`[4] in the same data bank were employed as a
`template
`of
`the
`tertiary model
`of human
`trypsin
`(TRYI). After
`removal of water molecules
`and addi-
`tion of hydrogens,
`the amino acids of bovine
`trypsin
`were replaced with
`those of human
`trypsin
`[9] by
`using
`the mutation
`option
`of
`the Quanta protein
`module. All side chains of the model were optimized
`by SOO-step energy minimization.
`The numbering
`system of the enzymes was defined on the basis of
`topological
`equivalence with chymotrypsin.
`
`Results and discussion
`
`Structure-activity
`
`relationships
`
`of DX-9065a
`activities
`and anti-thrombin
`Anti-FXa
`and its derivatives were measured as amidolytic
`activi-
`ties for chromogenic
`substrates S-2222 and S-2338,
`respectively
`[2]. In a similar manner, anti-trypsin
`acti-
`vity was measured using S-2222
`(table
`I). Pyrrolidine
`derivative
`2 and
`its acetimidoyl
`derivative
`1 had
`almost same potency
`in FXa
`inhibitory
`activity
`[5].
`Substitution
`of the methyl group of 7c with a cyclo-
`pentyl group,
`forming
`compound
`7a, exhibited
`an
`eightfold
`increase
`in inhibitory
`potency. Replacement
`of the aminoethyl
`group of 7b with
`the pyrrolidine
`in
`ring gave compound
`2 showing a 35-fold
`increase
`potency. These results
`indicate
`importance
`of a termi-
`nal
`five-membered
`ring
`in
`the FXa
`inhibitors.
`The
`anti-FXa
`activity
`of pyrrolidine
`derivative
`2
`is 21
`times as potent as that of cyclopentane
`derivative 7a.
`Similarly,
`aminoethyl
`derivative
`7b
`is fivefold more
`potent
`than 7~. These results
`indicate
`that the terminal
`amino
`group of
`the
`inhibitors
`contributes
`to
`the
`increase
`in potency. Thus,
`the pyrrolidine
`ring may be
`not only a constrained amino group but also a group
`contributing
`to a hydrophobic
`interaction with FXa.
`The anti-trypsin
`activity of compounds
`1, 2, and
`7a-d,
`unlike
`their anti-FXa
`activity,
`is almost
`the
`same. The hydrophobicity
`and basicity of the terminal
`substituents of the inhibitors
`are unlikely
`to be asso-
`ciated with the binding affinity
`for trypsin. Therefore,
`the FXa/trypsin
`selectivities
`of
`these
`inhibitors
`are
`primarily
`dependent on the anti-FXa activity.
`The
`inhibitory
`activity
`of compounds
`7a-d against
`thrombin was negligible.
`
`1, 2, and
`
`Table I. Inhibitory activities of derivatives of DX-9065a on
`FXa, thrombin and trypsin.
`
`Compound
`
`R
`
`FXa
`
`Thrombin Trypsin
`
`- HCI 0.16
`
`>lOOO
`
`10
`
`N--f
`
`NH
`
`CH3
`
`H l HCI
`
`0.20
`
`>lOOO
`
`5.4
`
`4.8
`
`10
`
`7.5
`
`>lOOO
`
`>lOoo
`
`>looO
`
`>1000
`
`16
`
`1
`
`2
`
`7a
`
`D
`
`7b
`
`p&‘&m
`
`. HCI
`
`7c
`
`-CHB
`
`7d
`
`-I-I
`
`4.2
`
`7.3
`
`35
`
`26
`
`Conformational
`
`search of DX-9065a
`
`low-energy conformers effi-
`In order to sample diverse
`ciently, a Metropolis Monte-Carlo
`search procedure
`[ 10, 111 was applied
`to DX-9065a.
`The
`initial
`struc-
`ture of DX-9065a was taken
`from
`its X-ray crystallo-
`graphic
`structure
`[5]. To generate a conformer,
`six
`torsion angles o~--cI), (see fig 1) were altered randomly
`within 45” at a pseudo-temperature
`of SOOOK using
`Boltzmann
`jump search option of the Quanta program
`[ 121. The generated
`conformer, whose dihedral
`root
`mean square difference
`from
`the previous conformer
`was over 43”, was accepted as a new conformer,
`which was subsequently
`optimized
`by 100~step mini-
`mization without
`changing
`the ring conformation
`of
`the pyrrolidine.
`This procedure was
`iterated 5000
`times.
`the torsion angle w, of 180 + 10”
`Conformers with
`were
`removed
`since
`their
`acetimidoylpyrrolidine
`moieties were
`incapable
`of binding
`to FXa. Elimi-
`nation of conformers with energy
`levels more
`than
`3 kcal/mol
`above the minimum
`gave 144 conformers.
`The conformational
`space shows that the acetimidoyl-
`pyrrolidine
`moiety
`is predominantly
`located
`in
`two
`
`

`
`144
`resulting
`figure 2. The
`in
`as shown
`regions,
`into
`the FXa model
`to study
`conformers were docked
`the mode of interaction
`between
`the acetimidoylpyrro-
`lidine moiety of DX-9065a
`and FXa.
`
`Docking
`
`studies of DX-9045a
`
`and
`two basic, amidinonaphthalene
`has
`DX-9065a
`acetimidoylpyrrolidine
`groups. The previous model-
`ing study of the complex between FXa and DX-9065a
`demonstrated
`that
`the Sl pocket of FXa prefers
`the
`planar amidinonaphthalene
`group
`to the acetimidoyl-
`pyrrolidine
`group with a spatially
`expanded structure
`[5]. Therefore,
`the amidinonaphthalene
`part of DX-
`9065a was adjusted
`to the Sl site to form a hydrogen-
`bonded
`ion pair of
`the amidino
`group with
`the
`carboxyl
`group of Asp189 of
`the enzyme without
`steric collisions. The docking
`study showed that 10 of
`the 144 conformers had a binding mode such that their
`
`of 144 conformers with an energy
`Fig 2. Superimposition
`less than 3 kcal/mol above the minimum. The amidinonaph-
`thalene, benzene and acetimidoylpyrrolidine
`moieties are
`displayed.
`
`interacted with a site of FXa corres-
`rings
`pyrrolidine
`‘aryl binding
`site’
`[13] of thrombin.
`to the
`ponding
`ring was surrounded with
`‘Qr99,
`The pyrrolidine
`Phe174 and Trp215 without any steric hindrance. This
`result
`is consistent with
`the model
`proposed
`by
`Padmanabhan
`et al,
`in which
`the dansyl group of
`dansyl-GluGlyArg,
`an FXa-inhibitor,
`interacts with
`the hydrophobic
`site of FXa
`[6], suggesting
`the hydro-
`phobic
`interaction
`of the pyrrolidine
`ring with the aryl
`binding
`site. The pyrrolidine
`rings of the other confor-
`mers did not have such an interaction with FXa.
`One of the 10 conformers was selected as a putative
`active conformer,
`since
`the acetimidoyl
`nitrogen was
`expected
`to form a hydrogen bond with
`the backbone
`carbonyl oxygen of Glu97. The hydrogen bond could
`explain
`the importance
`of the terminal
`basic group of
`the
`inhibitors
`1, 2 and 7b
`in
`inhibitory
`potency.
`Energy-minimization
`of
`the complex
`between
`this
`conformer and the FXa model was performed
`in 500
`steps without moving
`the a-carbons of FXa,
`to give a
`plausible model of DX-9065a
`bound
`to the enzyme.
`As shown in figure 3, the carbonyl oxygen of Glu97,
`which was located near the aryl binding
`site, appeared
`to donate a hydrogen bond
`to the acetimidoyl
`nitro-
`gen. The distance between
`the
`two atoms was ob-
`served to be 2.8 A, which was appropriate
`for a hydro-
`gen bond. Alternatively,
`a salt-bridge
`interaction may
`be considered as an interaction
`between
`the terminal
`basic center of the inhibitor
`and the acidic side chain
`of Glu97. However,
`the hydrogen bond seemed prefer-
`able to the salt-bridge
`interaction
`since the orientation
`of the side chain of Glu97 was unfavorable
`for
`the
`tight
`interaction with
`the acetimidoyl
`group.
`In addition,
`there was no hydrophobic
`site of FXa
`interacting
`with
`the benzene
`ring attached
`to
`the
`phenoxypyrrolidine
`moiety of all
`the 144 conformers.
`The ring might be a spacer to fix the pyrrolidine
`ring
`on the favorable
`location.
`two
`consists of
`Trypsin
`also has a site, which
`hydrophobic
`Leu99 and Trp215
`and a hydrophilic
`Lys175, corresponding
`to the aryl binding
`site of FXa.
`A docking
`study of DX-9065a
`in
`the trypsin model
`showed that the inhibitor
`could bind
`to the enzyme
`in
`a similar manner
`to FXa (fig 4). However,
`the hydro-
`phobic binding
`of the acetimidoylpyrrolidine
`moiety
`with the site of trypsin may not be very strong as the
`site is less hydrophobic
`than
`that of FXa. This could
`account
`for
`the almost
`identical
`antitrypsin
`activities
`of 7a and 7c. In addition,
`an electrostatic
`repulsion
`between
`the pyrrolidine
`nitrogen
`and
`the basic side
`chain of Lys175
`could weaken a hydrogen
`bond
`between
`the pyrrolidine
`nitrogen
`and
`the carbonyl
`oxygen of Glu97, explaining
`the almost
`identical anti-
`trypsin activity of the compounds with or without
`the
`terminal
`basic nitrogen.
`Thus, comparison
`between
`the binding modes of DX-9065a with trypsin and FXa
`
`

`
`391
`
`Fig 3. Model structure of the complex between FXa (thin lines) and DX-9065a (thick lines).
`
`Hg 4. Model structure of the complex between trypsin (thin lines) and DX-9065a (thick lines).
`
`selectivity,
`pyrrolidine
`
`of the FXa/trypsin
`the acetimidoyl
`
`can give an explanation
`which
`is ascribed
`to
`moiety.
`the molecular
`In
`the previous paper, we proposed
`mechanism
`of
`the FXa/thrombin
`selectivity
`of DX-
`9065a.
`An
`electrostatic
`repulsion
`between
`the
`carboxyl group of DX-906Sa
`and that of GIu192 of
`thrombin
`causes
`the unfavorable
`enzyme-inhibitor
`ac-
`contact and
`thereby extinguishes
`the
`inhibitory
`tivity against
`thrombin
`[3]. Both
`thrombin
`and FXa
`have
`the aryl binding
`site which was formed with
`
`study of DX-
`Ile1’74 and Trp215. A docking
`Leu99,
`indicated
`that the site accommodated
`the pyrro-
`9065a
`ring without
`steric clashes.
`In addition,
`throm-
`lidine
`bin also had Glu97 which seemed capable of forming
`a hydrogen
`bond with
`the amino
`terminal
`of DX-
`9065a. The
`interaction between
`the acetimidoylpyrro-
`lidine moiety and the enzyme would
`fix the orienta-
`tion of
`the carboxyl
`group of DX-9065a,
`which
`resulted
`in the electrostatic
`repulsion of the carboxyl
`group
`to Glu192
`of
`thrombin
`(fig 5). However,
`compounds
`7c and 7d without
`the aeetimidoylpyrro-
`
`

`
`392
`
`against
`activity
`inhibitory
`little
`group have
`lidine
`indicates
`that the carboxyl group
`thrombin. This result
`of DX-9065a
`is solely
`related
`to
`the electrostatic
`repulsion
`leading
`to the FXa/thrombin
`selectivity.
`In
`this paper,
`the conformational
`search was per-
`formed by
`the Metropolis
`Monte-Carlo
`method
`to
`sample
`various
`low-energy
`conformers
`efficiently.
`This method generated a plausible
`low-energy confor-
`mer of DX-9065a whose pyrrolidine
`and acetimidoyl
`moieties
`interacted with
`the aryl binding
`site and the
`Glu97 of FXa, respectively. This conformer
`could not
`be found by the previous
`systematic
`conformational
`analysis
`of DX-906Sa
`[3]. Thus,
`the Metropolis
`Monte-Carlo method can efficiently
`provide plausible
`low-energy
`conformers when compounds
`have many
`rotatable bonds, as DX-9065a
`does.
`In
`this
`procedure,
`with
`conformers
`the
`144
`of
`low potential
`energy were selected by a criterion
`3 kcal/mol
`above the minimum.
`Alternatively,
`a crite-
`rion of 10 kcal/mol
`above
`the minimum
`offered
`the
`2197 conformers. Ninety percent of the acetimidoyl-
`pyrrolidine
`group of these conformers
`had the same
`conformational
`space as those of the 144 conformers.
`The remaining were found
`to clash with FXa or had
`no hydrophobic
`interaction with FXa
`in the docking
`study
`(data not shown). Thus,
`the 144 conformers
`were sufficient
`for the evaluation
`of the binding mode
`of the acetimidoylpyrrolidine
`moiety with FXa.
`inter-
`The calculated
`energy of a drug-receptor
`action based on steric and electrostatic
`interactions
`may
`indicate
`favorable binding
`sites of the drug
`to
`receptors and/or enzymes
`[14]. However,
`a drug-
`
`could not be assessed exactly by
`interaction
`receptor
`because of the neglect of the entro-
`such a calculation
`to AG, which
`is predominant
`in some
`pic contribution
`drugs
`[15]. Since DX-9065a
`has a flexible
`structure
`and possesses some hydrophobic
`interactions
`with
`FXa,
`the enzyme-inhibitor
`interaction may be incor-
`rectly estimated by the energy calculations. Therefore,
`compounds
`7a-d were synthesized
`to evaluate such
`an entropic
`interaction.
`A combination
`of their SAR
`and the results of computer graphic
`techniques helped
`in the prediction
`of the putative binding
`sites. These
`results will not only
`improve
`the understanding
`of the
`enzyme-inhibitor
`interactions
`but also contribute
`to
`the research into new drugs.
`
`Experimental
`Chemistry
`
`protocols
`
`Melting points were determined on a Biichi 520 apparatus in
`glass capillary tubes and are uncorrected. Column chromato-
`graphy was performed on Merck silica gel 60 (7&230 mesh
`ASTM) and on Diaion HP-20 (highly porous polymer type
`synthetic adsorbent: Mitsubishi Chemical Industries). Thin-
`layer chromatography (TLC) was performed on Merck TLC
`aluminum sheets precoated with silica gel 60 Fz4 and detected
`by UV quenching at 254 nm or by spraying with phospho-
`molybdic acid or ninhydrin. All analytical samples were found
`to be homogeneous on TLC. rH-NMR spectra were recorded
`on a Jeol JNM-EX400 spectrometer, and chemical shifts are
`given in ppm (6) from tetramethylsilane, which was used as the
`internal standard. Mass spectra were performed using a Jeol
`JMS-AX505W (EI) or a Jeol JMS-I-IX110 (FD, FAB) spectro-
`meter. Analyses indicated by the symbols of the elements or
`functions were within 50.4% of theoretical values.
`
`w 189
`
`Fig 5. Model structure of the complex between thrombin (thin lines) and DX-9065a (thick lines).
`
`

`
`4a
`Ethyl 2-[4-[(cyclopentyl)oxy]phenyl]-2-oxoacetate
`To a stirred solution of ethvl 2-f4-hvdroxvohenvl1-2-oxoacetate
`(3.88 g, 20 mmol), cyclopentLno1 11.72-g,
`20-mmol)
`and tri-
`phenylphosphine
`(5.50 g, 21 mmol)
`in dry THF
`(30 ml) was
`added diethyl
`azodicarboxylate
`(3.65 g, 21 mmol)
`at room
`temperature.
`The
`reaction mixture was stirred
`for 18 h and
`concentrated.
`The
`resulting
`residue was purified by silica-gel
`chromatography
`using
`toluene
`to afford an oil (4.50 g, 85.8%).
`lH-NMR
`(CDCl,)
`6: 1.41 (3H,
`t, J = 7.3 Hz), 1.55-2.00
`(8H,
`m), 4.42 (2H, q,.l = 7.3 Hz), 4.80-4.90
`(br s, 1H
`, 6.90 (2H, d,
`J = 8.8 Hz), 7.96 (2H, d, J = 8.8 Hz). HRMS M+) calcd
`for
`?
`C15H,804: 262.1205;
`found: 262.1218.
`
`Ethyl 2-[4-[[2-(ter?-butoxycarbonylamino)ethyl]o~]phenyl]-2-
`oxoacetate 46
`from ethyl 2-(4-hydroxyphenyl)-2-
`Compound 4b was prepared
`oxoacetate and 2-(tert-butoxycarbonylamino)ethanol
`according
`to the procedure
`for preparation
`of 4a. Purification was carried
`out using silica-gel
`chromatography
`(toluene/acetone
`= 93:7)
`to
`give an oil (yield 38.5%).
`lH-NMR
`(CDCl,) 6: 1.42 (3H,
`t, J =
`7.3 Hz), 1.45 (9H, s), 3.50-3.63
`(2H, m), 4.05-4.15
`(2H, m),
`4.43 (2H, q, J = 7.3 Hz), 4.95-5.05
`(lH, br), 6.96 (2H, d, J =
`8.8 Hz), 8.00 (2H, d, J = 8.8 Hz). MS (FD) m/z: 337 M+.
`
`4c
`Ethyl 2-(4-methoxyphenyl)-2-oxoacetate
`To a stirred solution of ethyl 2-(4-hydroxyphenyl)-2-oxoacetate
`(3.88 g, 20 mmol) and methyliodide
`(2 ml) in dry DMF
`(50 ml)
`was added anhydrous K,CO,
`(2.90 g, 21 mmol). The mixture
`was heated at 80°C
`for 30 min, cooled
`to room
`temperature,
`and neutralized with concentrated HCl. The resulting mixture
`was poured
`into water and extracted with
`toluene. The organic
`layer was dried and concentrated
`to give an oil (3.67 g, 88.2%).
`tH-NMR
`(CDCQ
`6: 1.40 (3H, t, J = 7.0 Hz), 3.87 (3H, s), 4.42
`(2H, q, J = 7.0 Hz), 6.95 (2H, d, J = 9.0 Hz), 7.98 (2H, d, J =
`9.0 Hz). HRMS
`(M+) calcd
`for C,,H,,O,:
`208.0736;
`found:
`208.0719.
`
`bromide 5
`[(7-(Cyano-2-naphrhyl)methyl]~riphenylphosphonium
`This was prepared according
`to the described previously proce-
`dure [l].
`
`3-(7-cyano-2-naphthyl)-2-[4-((cyclopen~l)oxy]phenyl]-
`Ethyl
`propanoate
`6a
`To a stirred solution of 4a (1.30 g, 5.0 mmol) and 5 (2.60 g,
`5.0 mmol)
`in dry THF
`(20 ml)/EtOH
`(20 ml) was added 1,8-
`diazabicyclo[5.4.0]undec-7-ene
`(DEW)
`(835 mg, 5.5 mmol) at
`room
`temperature. The reaction mixture was stirred
`for 18 h at
`room
`temperature
`and concentrated.
`The
`residue was purified
`by silica-gel
`chromatography
`using
`toluene/AcOEt
`(9:l)
`to
`give a mixture of E- and Z-ethyl 2-(7-cyano-2-naphthyl)-3-[4-
`[[cyclopentyl]oxy]phenyl]acrylate
`as an oil
`(1.9 g). The mix-
`ture of the above E- and Z-olefins
`(1.9 g) in dry THF
`(80 ml)/
`EtOH
`(50 ml) was hydrogenated
`over 5% Pd/C (900 mg) at
`atmospheric
`pressure
`for 5 h. The resulting mixture was filtered
`and concentrated.
`The
`residue was purified
`by silica-gel
`column
`chromatography
`using
`toluene
`to provide a colorless
`solid
`(1.60 g, 78.1%), mp 85-86”C.
`lH-NMR
`(CDCl,)
`6: 1.10
`(3H,
`t, J = 7.3 Hz), 1.55-2.00
`(8H, m), 3.17 (lH, dd, .I = 13.7
`and 6.8 Hz), 3.54 (lH,
`dd, J = 13.7 and 8.3 Hz), 3.80-3.90
`(lH, m), 3.i)5+.15‘(2H,
`m), 4.65-5.75
`(lH, m), 6.80 (2H, d,
`J = 8.8 Hz). 7.20 (2H. d.J = 8.8 Hz). 7.42 (1H. d. J = 8.3 Hz).
`ij, 7.76 jlH; d; J = 8.3 Hzj;
`7.53 (lH, h; J = 8.3 k); 7.60 (lH,
`7.84 (lH, d, J = 8.3 Hz), 8.11 (lH,
`s). HRMS
`(M+) calcd
`for
`found: 413.2014. Anal C,,H,,NO,
`(C,
`(&H27N03:
`413.1991;
`H, W
`
`2-[4-[[2-(tert-butonycarbonylamino)ethyl3-
`Ethyl
`(7-cyano-2-naphthyl)propanoate
`6b
`for preparation
`Prepared
`from 4b according
`to the procedure
`4a as a colorless
`solid
`(yield 25.6%), mp 94-95°C.
`lH-NMR
`(CDC&) 6: 1.10 (3H,
`t, J = 7.3 Hz), 1.44 (9H, s), 3.18 (lH, dd,
`J = 13.7 and 6.8 Hz), 3.45-3.60
`(3H, m), 3.85-3.93
`(lH, m),
`3.95-4.15
`(4H, m), 4.97-5.10
`(lH,
`br s), 6.82
`(2H, d,
`J = 8.3 Hz), 7.23 (2H, d, J = 8.3 Hz), 7.40 (lH, d,.l = 8.3 Hz),
`7.51 (lH, d, J = 8.3 Hz), 7.60 (lH, s), 7.74 (lH, d,J = 8.3 Hz),
`7.82
`(lH,
`d, J = 8.3 Hz), 8.08
`(lH,
`s). Anal C29H32NZ05
`CC, H, W.
`
`of
`
`propanoate
`
`Ethyl 3-(7-cyano-2-naphthyl)-2-(4-methoxyphenyl)
`6c
`of
`for preparation
`to the procedure
`from 4c according
`Prepared
`6a as a colorless solid
`(yield 57.9%), mp 91-92°C.
`1H-NMR
`(CDClJ
`6: 1.11 (3H, t,J = 7.3 Hz), 3.18 (lH, dd,J = 13.7 and
`6.8 Hz), 3.55
`(lH, dd, J = 13.7 and 8.3 Hz), 3.79
`(3H, s),
`3.80-3.90
`(lH, m), 4.00-4.15
`(2H, m), 6.84 (2H, d, J = 8.8
`Hz), 7.23 (2H, d, J = 8.8 Hz), 7.42 (lH, d, J = 8.3 Hz), 7.54
`(lH, d,J = 8.3 Hz), 7.61 (lH, s), 7.77 (lH, d,J = 8.3 Hz), 7.85
`(lH, d,J = 8.3 Hz), 8.12 (lH, s). Anal C&H21N03
`(C, H, N).
`
`3-(7-Amidino-2-naphthyl)-2-[4-[(cyclopentyl)oxy]phenyl]pro-
`panoic acid hydrochloride
`hemihydrate
`7a
`(10 ml)/
`A solution of 6a (920 mg, 2.2 mmol)
`in dry CHzClz
`EtOH
`(30 ml1 was saturated with HCl eas with
`ice cooline. and
`left to’stand ?or 18 h at room
`temperzture. After distill&
`off
`the solvents and HCl,
`the resulting
`residue was dissolved
`in
`EtOH
`(40 ml), saturated with ammonia
`gas with
`ice cooling
`and left to stand for 18 h at room
`temperature. The mixture was
`concentrated,
`and
`the resulting
`residue was dissolved
`in dry
`DMF
`(20 ml), To the solution were added
`tert-butoxy
`dicarbo-
`nate (483 mg, 2.2 mmol) and triethylamine
`(446 mg, 4.4 mmol).
`The mixture was stirred
`for 3 h at room
`temperature,
`acidified
`with concentrated HCl and concentrated. The resulting
`residue
`was extracted with AcOEt.
`The organic
`layer was washed,
`dried and concentrated.
`The
`residue was purified by silica-gel
`column
`chromatography
`using acetone/CHzClz
`(1:9)
`to afford
`ethyl
`3-[7-(N-tert-butoxycarbonylamidino)-2-naphthyl]-2-[4-
`[(cyclopentyl)oxy]phenyl]propanoate
`as an oil (530 mg). To a
`stirred solution of the oil (530 mg, 2.2 mmol)
`in THF
`(5 ml)/
`EtOH
`(5 ml)&0
`(5 ml) was added 1 N NaOH
`(1.0 ml,
`1.0 mmol). The reaction mixture was stirred
`for 20 h at room
`temperature.
`The mixture was acidified with 10% citric acid
`and extracted with AcOEt/toluene
`(1:l). The organic
`layer was
`dried and concentrated.
`The resulting
`residue was dissolved
`in
`trifluoroacetic
`acid (5 ml) and stirred
`for 1 h at room
`tempera-
`ture. The solvent was removed, and the residue was purified by
`HP-20 column chromatography
`(acetonitrile/H,O
`(7:93)). After
`addition
`of small
`amount
`of concentrated
`HCl
`to desired
`fractions,
`the solvent was
`removed
`to give a colorless
`solid
`(150 mg, 15.3%).
`lH-NMR
`(DMSO-d,)
`6: 1.45-2.00
`(8H, m),
`3.14 (lH, dd, J = 14.0 and 6.8 Hz), 3.46 (lH, dd, J = 14.0 and
`8.8 Hz), 3.96 (lH,
`t,J = 7.3 Hz), 4.75 (lH, br), 6.82 (2H, d,J =
`8.5 Hz), 7.25 (2H, d, J = 8.5 Hz), 7.60 (lH, d,J = 8.3 Hz), 7.78
`(lH, d,.i = 8.8 Hz), 7.84 (lH, s), 7.94 (lH, d,J = 8.3 Hz), 8.07
`(lH, d, J = 8.8 Hz), 8.39 (lH, s), 9.29 (2H, br), 9.51 (2H, br),
`12.34
`(lH,
`br). MS
`(FAEJ) m/z: 403
`(M
`+ H)+. Anal
`C2SH26N203.HC1.1/2H20
`(C, H, N, Cl).
`
`3-(7-Amidino-2-naphthyl)-2-(4-[(2-aminoethyl)oxy]phenyl]-
`propanoic
`acid dihydrochloride
`hydrate 7b
`(10 ml)/
`A solution of 6b (600 mg, 1.2 mmol)
`in dry CH,Cl,
`EtOH
`(40 ml) was saturated with HCl gas under
`ice cooling
`and left to stand for 18 h at room
`temperature. After distilling
`
`

`
`were executed by the Charm/m minimi-
`Energy minimizations
`zation option of the Quanta program
`(method: adopted basis
`Newton-Raphson;
`energy gradient
`tolerance: 0.001 kcal/mol&).
`
`Pharmacology
`
`Assay of anti-trypsin
`
`activity
`
`activities were measured by using chromogenic
`Anti-trypsin
`substrate S-2222
`(Kabi Vitrum)
`and human
`trvnsin
`(Athens
`Research and Technology,
`Inc): DMSO
`(50 ulj-or
`a ‘DMSO
`solution of the inhibitor
`and 1 mM S-2222
`(100 pl) were mixed
`with 0.1 M Tris-0.2 M NaCl buffer at pH 8.4 (300 ~1). The
`reaction was started with
`the addition of 1 @ml
`human
`trypsin
`solution
`(50 pl), and the mixture was
`incubated
`for 10 min at
`37°C. The
`reaction was
`terminated with
`the addition
`of 60%
`AcOH
`(100 ul), and the optical densities
`(OD) were measured
`(405 nm): anti-trypsin
`activity
`(inhibition %) = 1 - (OD of the
`inhibitor/
`OD of saline control)
`x 100. The
`IG,
`value was
`obtained
`by plotting
`the
`inhibitor
`concentration
`against
`the
`anti-trypsin
`activity
`(inhibition
`%) on statistical
`probability
`paper.
`
`References
`
`394
`
`in
`residue was dissolved
`off the solvents and HCl, the resulting
`EtOH
`(50 ml). The solution was saturated with NH, gas with
`ice cooling and left to stand for 18 h at room
`temperature. The
`mixture was concentrated,
`and the resulting
`residue was dis-
`solved in 4 N HCI and heated at reflux
`for 1 h. After cooling
`to
`room
`temperature,
`the precipitate was collected by filtration
`to
`give a colorless solid
`(200 mg, 34.5%).
`tH-NMR
`(DMSO-d,)
`8: 3.10-3.20
`(3H, m),. 3.48 @H, dd, 3 = 14.2 and 8.3 Hzy,
`3.97-4.06
`(1H. ml. 4.12-4.20
`f2H. m‘l. 6.92
`f2H. d. J =
`8.8 Hz), 7.30 (iH, d;.l=
`8.8 Hz), 7.56-7.64
`(lH, m), 7.77-7.86
`(2H, m), 7.94 (lH, d, J = 8.8 Hz), 8.07 (lH, d, J = 8.8 Hz),
`8.25-8.45
`(3H, br), 8.43 (lH, s), 9.41 (2H, br), 9.61 (2H, br).
`MS (FAB) m/z: 378 (M + H)+. Anal
`(;,HuN,03*2HC1-Hz0
`(C,
`H, N); Cl: calcd 15.14; found, 16.11.
`
`acid
`
`3-(7-Amidino-2-naphthyl)-2-(4-methoxyphenyi)propanoic
`hydrochloride
`hemihydrate
`7c
`to the proce-
`Compound
`7c was prepared
`from 6c according
`dure
`for
`the preparation
`of 7b. Purification
`was carried out
`using HP-20 column chromatography
`(acetonitrile/H,O
`(7:93))
`to give a colorless amorphous
`solid
`(yield 47.5%), mp 195°C
`(dec).
`tH-NMR
`(DMSO-d,)
`6: 3.15
`(lH,
`dd, J = 13.7 and
`6.8 Hz), 3.47 (lH, dd, J = 14.0 and 8.5 Hz), 3.71 (3H, s), 3.90-
`4.00 (lH, m), 6.86 (2H, d, J = 8.8 Hz), 7.27 (2H, d,J = 8.8 Hz),
`7.60 (lH, d,.l=
`8.3 Hz), 7.75 (lH, d, J = 8.8 Hz), 7.84 (lH, s),
`7.94 (lH, d,J = 8.8 Hz), 8.07 (lH, d, J = 8.8 Hz), 8.36 (lH, s),
`9.20 (2H, br), 9.46 (2H, br), 12.34
`(lH, br). MS
`(FAH) m/z:
`349 (M + H)+. Anal ~,H,&03.HCl.0.5H,0
`(C, H, N, Cl).
`
`acid
`
`3-(7-Amidino-2-naphthyl)-2-(4-hydroxyphenyl)propanoic
`hydrochloride
`7d
`to the proce-
`from 6d according
`Compound
`7d was prepared
`solid
`(yield
`dure
`for
`the preparation
`of 7b as a colorless
`6: 3.13 (lH,
`62.4%), mp 225°C
`(dec).
`tH-NMR
`(DMSO-d,)
`dd, J = 14.0 and 6.8 Hz), 3.44 (lH, dd, J = 14.0 and 8.5 Hz),
`3.89 (lH,
`t, J = 7.8 Hz), 6.69 (2H, d, J = 8.8 Hz), 7.14 (2H, d,
`J = 8.8 Hz), 7.59 (lH, d,J = 8.3 Hz), 7.76 (lH, d, J = 8.3 Hz),
`7.82 (lH, s), 7.94 (lH, d,J = 8.3 Hz), 8.06 (lH, d,J = 8.3 Hz),
`8.37 (lH, s), 9.21 (2H, br), 9.36 (lH,
`s), 9.46 (2H, br), 12.27
`[lCHHbr& bit
`(FAEi) m/z: 335 (M + H)+. Anal C&-11&J203.HCI
`99,
`*
`
`Energy calculations
`
`computer
`on an Iris 320 GTX
`All procedures were performed
`from Silicon Graphics
`Incorporation.
`Energy calculations were
`carried out using
`the force field of version 2.2 of the Charm/m
`program
`[16]. The cut-off distance and the distance-dependent
`dielectric
`constant E were set to 8.0 A and 4 rij, respectively.
`
`1
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`9
`10
`
`11
`12
`
`13
`14
`15
`16
`
`M
`
`(1993)
`
`Biochem
`
`J, Bode W, Huber
`
`R
`
`T, Iwamoto
`
`M
`
`Biophys
`
`Res
`
`(1983)
`
`Actn
`
`K, Katakura
`
`S, Iwamoto
`
`M (1994)J
`
`Med
`
`KP, Tulinsky
`
`A et al
`
`(1993)
`
`J Mel Biol232,
`
`U, Huber
`
`R, Stone
`
`SR, Hofsteenge
`
`J (1989)
`
`Thromb Res 68,507-512
`PG (1993)
`JJ, de Groat
`Sixma
`H, Yokoyama
`Y, Nagahara
`A,
`lshihara
`Ham T, Yokoyama
`(1994)
`Thromb Haemostasis
`71,314-319
`Katakura
`S, Nagahara
`T, Hara T, lwamato
`Commun
`197,%5-972
`J, Deisenhofer
`Marquart
`M, Walter
`Cqwtallogr
`Sect B 39.480-490
`Nagahara
`T, Yokoyama
`Y, lnamura
`Chem 37,1200-1207
`Padmanabhan
`K, Padmanabhan
`947-966
`I, Baumann
`Bode W, Mayr
`EMBO
`J 8,3467-3479
`Upton,
`Laboratory,
`National
`Brookhaven
`Protein
`Data Bank,
`Y, Ogawa M et al (1986) Gene 41.305-310
`Emi M, Nakamura
`Rosenbluth
`AW,
`Rosenbluth
`MN,
`Teller
`Metropolis
`NA,
`(1953)J
`Chem Phys 21,1987
`Li Z, Sheraga
`HA
`(1987)
`Proc NatlAcad
`Molecular
`Simulation
`Inc, New
`England
`USA
`Biochem
`EwJ
`J (1990)
`Bode W, Turk D, Stiirzebecher
`13,73C-748
`Camp Chew
`Leach
`AR, Kuntz
`ID (1992)J
`9,408-411
`in Z’harmacolSci
`Hitzemann
`R (1988)
`Trends
`Brooks
`BR, Bruccoleri
`RE, Olafson
`BD, States DJ, Swaminathan
`M (1983)
`J Comput
`Chem 4,187-217
`
`NY, USA
`
`AH,
`
`Teller
`
`EJ
`
`Sci USA 84,6611&X5
`Executive
`Park, Burlington,
`
`MA,
`
`193,175-182
`
`S, Karplus