Farnesyl pyrophosphate synthase enantiospecificity with a chiral risedronate analog, [6,7-dihydro-5H-cyclopenta[c]pyridin-7-yl(hydroxy)methylene]bis(phosphonic acid) (NE-10501): Synthetic, structural, and modeling studies

Article abstract of DOI:10.1016/j.bmcl.2008.03.088

The complex formed from crystallization of human farnesyl pyrophosphate synthase (hFPPS) from a solution of racemic [6,7-dihydro-5H-cyclopenta[c]pyridin-7-yl(hydroxy)methylene]bis(phosphonic acid) (NE-10501, 8), a chiral analog of the anti-osteoporotic drug risedronate, contained the R enantiomer in the enzyme active site. This enantiospecificity was assessed by computer modeling of inhibitor-active site interactions using Autodock 3, which was also evaluated for predictive ability in calculations of the known configurations of risedronate, zoledronate, and minodronate complexed in the active site of hFPPS. In comparison with these structures, the 8 complex exhibited certain differences, including the presence of only one Mg²⁺, which could contribute to its 100-fold higher IC₅₀. An improved synthesis of 8 is described, which decreases the number of steps from 12 to 8 and increases the overall yield by 17-fold.

Full text of DOI:10.1016/j.bmcl.2008.03.088

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 2878–2882
Farnesyl pyrophosphate synthase enantiospecificity with a chiral
risedronate analog, [6,7-dihydro-5H-cyclopenta[c]pyridin-7-
yl(hydroxy)methylene]bis(phosphonic acid) (NE-10501):
Synthetic, structural, and modeling studies
a
a
a
b
`
Sylvine Deprele, Boris A. Kashemirov, James M. Hogan, Frank H. Ebetino,
Bobby L. Barnett,c Artem Evdokimovb and Charles E. McKennaa,*
aDepartment of Chemistry, University of Southern California, Los Angeles, CA 90089, USA
bProcter & Gamble Pharmaceuticals, Mason, OH 45040, USA
cChemistry Department, University of Cincinnati, Cincinnati, OH 45221, USA
Received 22 January 2008; revised 22 March 2008; accepted 31 March 2008
Available online 8 April 2008
Abstract—The complex formed from crystallization of human farnesyl pyrophosphate synthase (hFPPS) from a solution of racemic
[6,7-dihydro-5H-cyclopenta[c]pyridin-7-yl(hydroxy)methylene]bis(phosphonic acid) (NE-10501, 8), a chiral analog of the anti-oste-
oporotic drug risedronate, contained the R enantiomer in the enzyme active site. This enantiospecificity was assessed by computer
modeling of inhibitor–active site interactions using Autodock 3, which was also evaluated for predictive ability in calculations of the
known configurations of risedronate, zoledronate, and minodronate complexed in the active site of hFPPS. In comparison with
these structures, the 8 complex exhibited certain differences, including the presence of only one Mg2+, which could contribute to
its 100-fold higher IC50. An improved synthesis of 8 is described, which decreases the number of steps from 12 to 8 and increases
the overall yield by 17-fold.
ꢀ 2008 Elsevier Ltd. All rights reserved.
Bisphosphonate drugs have long been used for the treat-
ment of bone disorders such as Paget’s disease, cancer-
related hypercalcemia, and postmenopausal osteoporo-
sis,1 but the molecular mechanisms underlying their effi-
cacy are only recently beginning to be understood.2,3
The first compounds in this class of drugs were intro-
duced as stable analogs of pyrophosphate, an endoge-
nous bone-mediating agent. Bisphosphonates (BPs)
incorporate a hydrolytically inert P–C–P backbone that
mimics the pyrophosphate P–O–P backbone, providing
the bone mineral affinity of the natural compound, while
making possible substitution at the central carbon, en-
abling creation of more effective drugs.
an a-hydroxyl group.4,5 The presence of a nitrogen (N-
BPs) atom in a second a-substituent can confer greatly
enhanced anti-osteoporotic potency, which has been
associated with the inhibition of human farnesyl pyro-
phosphate synthase (hFPPS, EC 2.5.1.10). hFPPS cata-
lyzes the condensation of the isoprenoid dimethylallyl
pyrophosphate (DMAPP) with isopentenyl pyrophos-
phate (IPP).6,7 This forms geranyl diphosphate (GPP),
which then condenses with a second IPP to yield far-
nesyl pyrophosphate (FPP). Inhibition of hFPPS blocks
prenylation of several G proteins, impairing their reloca-
tion to the cellular membrane and ultimately leading to
apoptosis of the osteoclast.2
Both phosphonate groups are required for maximal
affinity to bone, which is increased by the presence of
Recent X-ray crystallographic structures of several
clinically used nitrogen-containing bisphosphonates
including risedronate [(1-hydroxy-2-pyridin-3-ylethane-
1,1-diyl)bis(phosphonic acid)] in complex with bacte-
rial and human FPPS8–10 suggest that the side-chain
nitrogen of the inhibitor is positioned in the complex
to form a hydrogen bond with one or more active site
Keywords: Osteoporosis; Bisphosphonates; Anti-tumor; Molecular
modeling; Crystallography; Chirality.
*
Corresponding author. Tel.: +1 213 740 7007; fax: +1 213 740
0930; e-mail: mckenna@usc.edu
0960-894X/$ - see front matter ꢀ 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.03.088

`
S. Deprele et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2878–2882
2879
residues. None of the drugs used in these studies are
chiral.
tiomer was ꢀ30.86 kcal/mol over a cluster of 73 confor-
mations, whereas the R enantiomer had a value of
ꢀ31.33 kcal/mol in a cluster of 80 trials. As seen in Fig-
ure 3a, (S)-8 docking predicts poor conformational
homology with risedronate in the active site. Figure 3b
shows that (R)-8 demonstrates homology with a critical
risedronate-binding component, the pyridinyl nitrogen.
Although these absolute energies computed by Auto-
dock 3 are considered unreliable, the relative energy dif-
ference, ꢁ0.5 kcal/mol predicting favored binding of the
R isomer, combined with qualitative visualization of the
bound enantiomers through docking, merited experi-
mental evaluation. Accordingly, we sought to crystallize
the enzyme complex resulting from the incubation of the
enzyme with racemic 8 and to determine its structure by
X-ray crystallography (Fig. 4).
To further clarify the contribution of the drug side-chain
nitrogen interactions to FPPS inhibitor binding, we
decided to examine by active site modeling and X-ray
crystallography the relative binding of the enantiomeric
forms of a chiral risedronate analog in which the rota-
tion of the pyridine ring about the bond joining it to
the linking methylene carbon is prevented by its incor-
poration into a rigid ring, (R,S)-[6,7-dihydro-5H-cyclo-
penta[c]pyridin-7-yl(hydroxy) methylene]bis(phosphonic
acid) (NE-10501, 8)11 (Fig. 1). The racemate of this com-
pound has an IC50 of 629.5 nM, thus even if one enan-
tiomer were wholly responsible for the inhibition, its
IC50 would exceed that of risedronate (IC50 = 5.7 nM)7
by almost two orders of magnitude. Earlier work dem-
onstrates the importance of conformational templates
in examining the activity of bisphosphonates.12
Synthetic studies: The synthesis of 8, which involved the
first entry into the pyrindinyl carboxylate 7 via photo-
chemical ring contraction of 7-hydroxyisoquinoline-8-
diazonium chloride 4 (Scheme 1), was previously accom-
plished in 12 steps with an overall yield of 0.2%.11 Our
impatience to proceed to our crystallographic studies
impelled us to seek improvements to the efficiency of this
route, focusing on intermediate 1, isoquinolin-7-ol. In
the original procedure, 1 was prepared by condensing
To facilitate these and future investigations, a signifi-
cantly improved synthesis of 8 was devised, which is re-
ported here together with the results of our structural
studies of the 8-hFPPS complex.
Modeling studies: We first undertook modeling of three
known hFPPS complexes of clinical bisphosphonates
(minodronate, risedronate, and zoledronate) with Auto-
Dock 3 to provide an indication of the predictive power
of the program with these systems. AutoDock 3 has been
used to model risedronate complexes with avian, bacte-
rial, and recombinant human FPPS13 and has shown14,15
potential as a tool for the discovery of new FPPS-targeted
BPs. For all three drugs, the computer model generated a
consensus ensemble of bound conformers that coincided
reasonably well with the experimentally determined struc-
tures (Fig. 2).16 In these studies, IPP was not included in
the model because, while IPP is required for the overall
function of the enzyme in vivo, its presence is not required
for the binding between N-BPs and hFPPS. The IPP bind-
ing site is fully formed only after the allylic substrate or N-
BP is bound and hFPPS is triggered into a partially closed
conformation.10
3-methoxybenzaldehyde
with
aminoacetaldehyde
diethyl acetal in refluxing benzene, hydrogenation of
the imine product to the amine with PtO2 catalyst, tosy-
lation, cyclodehydration by HCl in dioxane, and
demethylation with BBr3 at ꢀ78 ꢁC. We found that
these tedious multiple transformations can be replaced
by an elegantly simple one-pot procedure using commer-
cially available starting materials, in which 3-(benzyl-
oxy)benzaldehyde is refluxed with aminoacetaldehyde
dimethyl acetal until the water formed has been com-
pletely removed using a Dean–Stark trap, whereupon
the condensation product is reacted in situ with trifluo-
roacetic acid–BF3–Et2O, giving 1 directly after weakly
alkaline aqueous work-up, in 54% yield.17
Conversion of 1 to the 8-nitro derivative 2 with NO2BF4
in sulfolane followed the prior procedure, with the inclu-
sion of a recrystallization step to improve purity.18 After
reduction with 10% Pd/C and acidification, the 8-amino
hydrochloride salt 3 was obtained.19
In all such docking analyses performed to date, the com-
pared inhibitors have been different compounds, with
one very recent exception where diastereomeric forms
of an inhibitor were examined.15 We therefore modeled
the individual enantiomers of 8 into the active site of
hFPPS (PDB ID# 1YV5) to determine whether the
model would predict a significant preference for binding
one over the other. For 100 docking trials with each
enantiomer, the average docked energy for the S enan-
Conversion of 3 to the diazonium chloride 4 in prepara-
tion for the subsequent photochemical ring contraction
to the pyrindinyl intermediate 5 was effected with t-butyl
nitrite as previously described,11 but EtOH–HCl was re-
placed by MeOH–HCl (generated by the reaction of the
dry solvent with acetyl chloride) to unify the reaction
solvent, resulting in an enhanced yield.20
N
N
Irradiation of 4 in a high-intensity UV reactor (Rayon-
et) appeared to depress the yield of 5, and we reverted to
the recommended 275 W sun lamp, an original example
of which was obtained from an online reseller.21 The
overall yield for the next two steps, reduction of 5 to 6
and saponification to 7 (isolated as the HCl salt), was in-
creased about threefold by adjustments to the original
procedures.22,23
H
H
O
O
O
O
HO P
HO
P OH
OH
HO P
HO
P OH
OH
OH
OH
(S)
(R)
NE-10501
Figure 1. Structures of the enantiomers of 8.

`
S. Deprele et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2878–2882
2880
Figure 2. Docking of minodronate, risedronate, and zoledronate into the active site of hFPPS, based on their respective enzyme complex crystal
structures (3B7L, 1YV5, and 2F8C). All compounds were docked as the nitrogen-protonated forms. Spheres are magnesium ions.
Figure 3. (a) S-8 docking predicts poor conformational homology with risedronate in the active site of hFPPS (1YV5). (b) R-8 docking gives overlap
of bound risedronate’s key binding element, the pyridyl nitrogen, with 80% of the calculated conformations.
able synthetic pathway at this point, yields of only
ꢁ11% were achieved consistently. We first attempted
to substitute the method of Lecouvey et al.,24 in which
2 equiv of tris(trimethylsilyl)phosphite is reacted with a
carboxylic acid or anhydride. We were unable to obtain
pure 6,7-dihydro-5H-2-pyrinidine-7-acid hydrochloride,
and therefore attempted to generate the TFA anhydride
of 7. This proved accessible in situ, but reaction with
tris(trimethylsilyl) phosphite was unsuccessful. We then
returned to the original reaction and using slightly dif-
ferent conditions,25 succeeded in doubling the yield of
the bisphosphonic acid to 23%. Substitution of toluene
for the reaction solvent provided a further gain, and
we finally were able to obtain pure 8 from 7 in 38%
yield.26 For the total synthesis, the overall yield is 17-
fold higher than the previous method, and begins from
an inexpensive, commercially available starting material.
Figure 4. The crystal structure of NE-10501 in complex with hFPPS
(PDB ID: 2RAH). R enantiomer was preferentially bound. Large
Crystallography: Compound 8 thus prepared, was co-
crystallized with hFPPS. X-ray crystallography revealed
that only the R-isomer (PDB ID: 2RAH) was detectable
in the active site of hFPPS. However, although the over-
all configuration of the active site complex resembles
that of other inhibitor-bound FPPS structures reported
to date,27 the complex differed in that surprisingly, only
a single Mg2+ was present. Changes in the side-chain
locations of Asp 103 and Gln 171 are also apparent.
spheres, Mg2+; small spheres, H2O.
The last step of the original synthesis relied upon the
standard but sometimes refractory PCl3/H3PO3 phos-
phorylation of a carboxylic acid, in this instance 7-car-
boxy-6,7-dihydro-5[H]-cyclopenta[c]pyridinium
ride, 7 to create the final bisphosphonate product, 8.
chlo-
Although the previous report11 demonstrated a reason-

`
S. Deprele et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2878–2882
2881
Scheme 1. Reagents and conditions: (a) i—aminoacetaldehyde dimethylacetal, toluene, 110 ꢁC, 6 h; ii—TFA anhydride, BF3ÆEt2O, <10 ꢁC, 4 days;
iii—Ether, NH4OH to pH 9; (b) i—sulfolane, nitronium tetrafluoroborate, rt, 6 h; ii—EtOH/Et2O (1:1); (c) i—10% Pd/C, H2, ethanol; ii—Ether,
HCl; (d) MeOHÆHCl, t-BuONO 0 ꢁC ! rt, 4 h; (e) NaHCO3, MeOH, hv, 0 ꢁC, 3 h; (f) 10% Pd/C, MeOH, rt, 5 h; (g) i—NaOH, 58 ꢁC, 4 h; ii—EtOH,
HCl; (h) i—H3PO3, PCl3, toluene, 110 ꢁC, 4 h; ii—HCl, 100 ꢁC, overnight.
Redocking of 8 into the experimentally determined ac-
tive site complex still gave docking scores favoring the
binding of the R over S enantiomer, although with a
smaller bias (data not shown), provided that a water
molecule was inserted into the position normally occu-
pied by the Mg2+, which interacts with Asp 103 and a
bisphosphonate P–O oxygen (This water is present in
the experimental structure but requires manual inclusion
in the docking program). The presence of the rigid 8
pyridinyl ring appears to block the incorporation of this
Mg2+ into the complex, decreasing its stability.
Acknowledgments
This work was supported with funding by Procter &
Gamble Pharmaceuticals, Inc. Charles E. McKenna
was a consultant for Procter & Gamble Pharmaceuti-
cals, Inc., during the period of this research. Data were
collected at Southeast Regional Collaborative Access
Team (SER-CAT) 22-ID (or 22-BM) beamline at the
Advanced Photon Source, Argonne National Labora-
tory. Supporting institutions may be found at
www.ser-cat.org/members.html. Use of the Advanced
Photon Source was supported by the U.S. Department
of Energy, Office of Science, Office of Basic Energy Sci-
ences, under Contract No. W-31-109-Eng-38.
In summary, AutoDock 3 modeling studies were consis-
tent with the favored conformations of risedronate,
zoledronate, and minodronate in the hFPPS active site
obtained by X-ray crystallography and indicated that
the enzyme should preferentially bind the R enantiomer
from a solution of racemic [6,7-dihydro-5H-cyclo-
penta[c]pyridin-7-yl(hydroxy)methylene] bis(phosphonic
acid) 8, for which a more efficient synthesis (8 vs 12
steps, 17· higher yield) was devised. X-ray crystallogra-
phy confirmed the enantiomeric selectivity, but an unu-
sual active site complex structure was revealed,
incorporating a single Mg2+. The results suggest a new
structural basis to interpret the lower affinity of hFPPS
for 8 relative to the three clinical drugs.
References and notes
1. Fleisch, H.; Russell, R. G.; Francis, M. D. Science 1969,
165, 1262.
2. Green, J. R. Oncogene 2004, 9, 3.
3. Reszka, A. A.; Rodan, G. A. Mini-Rev. Med. Chem. 2004,
4, 711.
4. Fleisch, H. A.; Russell, R. G.; Bisaz, S.; Muhlbauer, R. C.;
Williams, D. A. Eur. J. Clin. Invest. 1970, 1, 12.
5. Russell, R. G.; Muhlbauer, R. C.; Bisaz, S.; Williams, D.
A.; Fleisch, H. Calcif. Tissue Res. 1970, 6, 183.

`
S. Deprele et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2878–2882
2882
6. Martin, M. B.; Arnold, W.; Heath, H. T., 3rd; Urbina, J. A.;
9.27 (s, 1H), 8.60 (d, J = 6 Hz, 1H), 8.25 (d, J = 9 Hz, 1H),
8.16 (d, J = 6 Hz, 1H), and 7.74 ppm (d, J = 9 Hz, 1H).
Oldfield, E. Biochem. Biophys. Res. Commun. 1999, 263, 754.
7. Dunford, J. E.; Thompson, K.; Coxon, F. P.; Luckman, S.
P.; Hahn, F. M.; Poulter, C. D.; Ebetino, F. H.; Rogers,
M. J. J. Pharmacol. Exp. Ther. 2001, 296, 235.
8. Hosfield, D. J.; Zhang, Y.; Dougan, D. R.; Broun, A.;
Tari, L. W.; Swanson, R. V.; Finn, J. J. Biol. Chem. 2004,
279, 8526.
9. Kavanagh, K. L.; Guo, K.; Dunford, J. E.; Wu, X.; Knapp,
S.; Ebetino, F. H.; Rogers, M. J.; Russell, R. G. G.;
Oppermann, U. Proc. Nat. Acad. Sci. U.S.A. 2006, 103, 7829.
10. Rondeau, J.-M.; Bitsch, F.; Bourgier, E.; Geiser, M.;
Hemmig, R.; Kroemer, M.; Lehmann, S.; Ramage, P.;
Rieffel, S.; Strauss, A.; Green, J.; Jahnke, W. Chem. Med.
Chem. 2006, 1, 267.
11. Ebetino, F. H.; Soyke, E. G., Jr.; Dansereau, S. M.
Heteroat. Chem 2000, 11, 442. The structure of the title
compound is depicted erroneously here.
19. 7-Hydroxyisoquinolin-8-aminium
chloride
(2 g,
10.5 mmol) and 10% Pd/C (0.4 g) were mixed in EtOH
(50 mL) under H2. After filtration, the solution was
evaporated and the residue dissolved in MeOH (10 mL).
After addition to ether (100 mL) and 4 M HCl in dioxane
(7.5 mL), the product precipitated: light brown solid
1
(1.26 g, 77%). H NMR (DMSO-d6): d 9.78 (s, 1H), 8.18
(d, J = 6.7 Hz, 1H), 8.10 (d, J = 6.7 Hz, 1H), 7.65 (d,
J = 8.4 Hz, 1H) and 7.29 ppm (d, J = 8.4 Hz, 1H).
20. A solution of 0.1 M methanolic HCl (31 mL) was added to
7-hydroxyisoquinolin-8-aminium chloride (0.5 g) under
N2 at 0 ꢁC, followed by 1 mL t-butylnitrite. The solution
was stirred for 4 h and then added to cold Et2O (200 mL).
The product after drying was a light brown solid (353 mg,
68%). 1H NMR (DMSO-d6) d 9.09 (s, 1H), 8.52 (d,
J = 5 Hz, 1H), 8.00 (d, J = 9 Hz, 1H), 7.90 (d, J = 5 Hz,
1H) and 7.00 ppm (d, J = 9 Hz, 1H).
21. A solution of 7-hydroxyisoquinoline-8-diazonium chloride
(355 mg, 1.7 mmol) and sodium bicarbonate (271 mg,
3.22 mmol) in dry MeOH (500 mL), at 0 ꢁC, was irradi-
ated with a 275-W sun lamp for 3 h. The solvent was
removed and the residue extracted with CH2Cl2. The
extracts were evaporated: red–orange solid (225 mg, 76%).
1H NMR (DMSO-d6): d 8.81 (s, 1H), 7.67 (d, J = 4 Hz,
1H), 7.59 (d, J = 6 Hz, 1H), 7.54 (d, J = 6 Hz, 1H), 6.39 (d,
J = 4 Hz, 1H) and 3.72 ppm (s, 3H).
12. Ebetino, F. H.; McOsker, J. E.; Borah, B.; Emge, T. J.;
Crawford, R. J.; Berk, J. D. Osteoporosis, Third Interna-
tional Symposium on Osteoporosis 1990, 1344.
´
13. Ebetino, F. H.; Roze, C. N.; McKenna, C. E.; Barnett, B.
L.; Dunford, J. E.; Russell, R. G. G.; Mieling, G. E.;
Rogers, M. J. J. Organomet. Chem. 2005, 690, 2679.
14. Mao, J.; Mukherjee, S.; Zhang, Y.; Cao, R.; Sanders, J.
M.; Song, Y.; Zhang, Y.; Meints, G. A.; Gao, Y. G.;
Mukkamala, D.; Hudock, M. P.; Oldfield, E. J. Am.
Chem. Soc. 2006, 128, 14485.
`
15. Hogan, J. M.; Deprele, S.; Kashemirov, B. A.; Barnett, B.;
22. Methyl 7[H]-cyclopenta[c]pyridine-7-carboxylate (440 mg,
2.51 mmol) and 10% Pd/C (1.07 g) in MeOH were kept
under H2 for 5 h. The mixture was filtered and the solvent
removed: light brown oil (311 mg, 70%). 1H NMR
(DMSO-d6): d 7.67 (s, 1H), 7.55 (d, J = 4.7 Hz, 1H), 6.50
(d, J = 4.7 Hz, 1H), 3.38 (t, J = 8 Hz, 1H), 2.85 (s, 1H),
2.11 (m, 1H) and 1.50 ppm (m, 2H).
Evdokimov, A.; Dunford, J.; Russell, R. G. G.; Ebetino,
F. H.; McKenna, C. E. 234th ACS National Meeting;
2007, MEDI-073.
16. The crystal structure of risedronate in complex with
hFPPS was obtained from the Protein Data Bank (PDB
ID: 1YV5). This file was edited to remove water molecules
and risedronate. Partial charges and solvation parameters
were added using AutoDock Tools, and the magnesium
ions were manually assigned a charge of +2. The grid map
was calculated using AutoDock 3.0.5 (http://auto-
dock.scripps.edu/). Ligands were constructed using Spar-
tan ’02 (Wavefunction, Inc.: Irvine, CA) and partial
charges and protons were added, and torsions defined,
using AutoDock Tools. Docking simulations were per-
formed in AutoDock 3.0.5 using a Lamarckian genetic
algorithm employing a rigid protein and a flexible ligand
(100 independent runs per enantiomer). Calculations were
carried out on a Linux operating system and the resulting
structure files were analyzed using VMD 1.8.5, a visual-
ization program (http://www.ks.uiuc.edu/Research/vmd/).
17. 3-(Benzyloxy)benzaldehyde (53.4 g, 252 mmol) and amino-
acetaldehyde dimethylacetal were refluxed for 6 h in toluene
(550 mL) under N2. TFA (106 mL, 758 mmol) and BF3ÆE-
t2O (93 mL, 740 mmol) were added dropwise, at 10 ꢁC.
After 4 d, the product was rinsed with ether and dissolved in
H2O. The pH was increased to 9 with NH4OH, precipitating
a light brown solid, which was dried under vacuum (19.36 g,
23. Methyl 6,7-dihydro-5[H]-cyclopenta[c]pyridine-7-carbox-
ylate (450 mg, 2.54 mmol) in 1.0 M NaOH (2.6 mmol) in
MeOH (26 mL) was stirred at 58 ꢁC for 4 h. The solvent
was removed and the residue dissolved in 10 mL EtOH.
The addition of 4.0 M HCl (1.25 mL) in dioxane, followed
by cold Et2O gave a solid, dried under vacuum (460 mg,
1
90%). H NMR (DMSO-d6): d 12.04 (s, 1H), 8.75 (s, 1H),
8.63 (d, 1H), 7.81 (d, 1H), 4.28 (t, J = 7.2 Hz, 1H), 3.11 (m,
2H) and 2.32 ppm (m, 2H).
24. Lecouvey, M.; Mallard, I.; Bailly, T.; Burgada, R.;
Leroux, Y. Tetrahedron Lett. 2001, 42, 8475.
25. Kieczykowski, G. R.; Jobson, R. B.; Melillo, D. G.;
Reinhold, D. F.; Grenda, V. J.; Shinkai, I. J. Org. Chem.
1995, 60, 8310.
26. 7-Carboxy-6,7-dihydro-5[H]-cyclopenta[c]pyridinium chlo-
ride (140 mg, 0.7 mmol) and phosphorous acid (180.6 mg,
2.2 mmol) were stirred at 80 ꢁC in toluene until the solids
melted. PCl3 (0.2 mL, 2.29 mmol) was added dropwise and
the temperature raised to 110 ꢁC. The mixture was stirred
for 4 h and the toluene decanted. The system was refluxed
in 1.0 M HCl (2 mL) overnight. Water was removed in
vacuo and the black solid product was rinsed with acetone
and MeOH: white solid (130 mg, 38%). 1H NMR (DMSO-
d6): d 8.39 (s, 1H), 7.85 (d, 1H), 7.24 (d, 1H), 3.54 (m, 1H),
2.75 (m, 2H) and 1.92 ppm (m, 2H) 31P NMR (D2O): d.
17.5 (d, Jp–p = 35 Hz) and 16 ppm (d, Jp–p = 35 Hz).
27. The protein was supplemented with 10 mM MgCl2 and
mixed with 1 mM inhibitor. After incubation, crystalliza-
tion was performed using the hanging drop method, using
a 1:1 ratio of protein solution and well solution, which
consisted of 0.2 M ammonium sulfate, 15% isopropanol,
15% ethylene glycol, and 0.1 M sodium acetate, at pH 4.5.
1
54%). H NMR (DMSO-d6): d 10.15 (s, 1H), 9.08 (s, 1H),
8.27 (d, J = 6 Hz, 1H), 7.81 (d, J = 9 Hz, 1H), 7.67 (d,
J = 6 Hz, 1H), 7.32 (d, J = 9 Hz, 1H), 7.24 ppm (s, 1H).
18. NO2BF4 (13 g, 98 mmol) in sulfolane, was added to
isoquinolin-7-ol (10 g, 69 mmol) at 0 ꢁC. After stirring
under N2 in sulfolane (50 mL, 525 mmol) for 6 h at rt,
quenching with MeOH and evaporation, the residue was
treated with 100 mL of 1:1 EtOH/Et2O. The sulfolane
layer was washed with Et2O, yielding a brown solid that
was dissolved in H2O and recrystallized with Et2O:8-
1
nitroisoquinolin-7-ol (9 g, 69%). H NMR (DMSO-d6): d