Small conformationally restricted piperidine N-arylsulfonamides as orally active γ-secretase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2007.08.013

The design and development of a new class of small 2,6-disubstituted piperidine N-arylsulfonamide γ-secretase inhibitors is reported. Lowering molecular weight including the use of conformational constraint led to compounds with less CYP 3A4 liability com

Full text of DOI:10.1016/j.bmcl.2007.08.013

Bioorganic & Medicinal Chemistry Letters 17 (2007) 5330–5335
Small conformationally restricted piperidine N-arylsulfonamides
as orally active c-secretase inhibitors
Hubert Josien,a,* Thomas Bara,a Murali Rajagopalan,a Theodros Asberom,a
John W. Clader,a Leonard Favreau,c William J. Greenlee,a Lynn A. Hyde,b
Amin A. Nomeir,c Eric M. Parker,b Dmitri A. Pissarnitski,a Lixin Song,b
Gwendolyn T. Wong,b Lili Zhang,b Qi Zhangb and Zhiqiang Zhaoa
aDepartment of Chemical Research, Schering-Plough Research Institute, 2015 Galloping Hill Road, Kenilworth, NJ 07302, USA
bDepartment of Neurobiology, Schering-Plough Research Institute, 2015 Galloping Hill Road, Kenilworth, NJ 07302, USA
cDepartment of Drug Safety and Metabolism, Schering-Plough Research Institute, 2015 Galloping Hill Road,
Kenilworth, NJ 07302, USA
Received 2 July 2007; revised 8 August 2007; accepted 8 August 2007
Available online 15 August 2007
Abstract—The design and development of a new class of small 2,6-disubstituted piperidine N-arylsulfonamide c-secretase inhibitors
is reported. Lowering molecular weight including the use of conformational constraint led to compounds with less CYP 3A4 liability
compared to early leads. Compounds active orally in lowering Ab levels in Tg CRND8 mice were identified as potential treatments
for Alzheimer’s disease.
Ó 2007 Elsevier Ltd. All rights reserved.
Alzheimer’s disease (AD), the most common form of
neurodegenerative disorders, is progressively moving
to the forefront of our public health agenda as our pop-
ulation age.1 An incurable illness that mostly affects the
elderly, AD is characterized by a loss of cognitive func-
tions that ultimately results in death within 8–10 years of
onset. Although AD has many histological features,2
convergent lines of evidence suggest that aberrant pro-
duction of b-amyloid (Ab), aggregation, and/or plaque
deposition in the brain of affected individuals are central
to the evolution of the disease.3 Current avenues of
intervention seek to stop or reverse its course by inhib-
iting Ab production and/or aggregation.2,4 Ab, in its
form Ab40 or Ab42, the latter being the most amyloido-
genic, is the result of proteolytic cleavage of the Ab pre-
cursor protein (APP) by b-secretase and c-secretase.5
Despite its complexity and involvement in other regula-
tory pathways such as Notch processing, c-secretase has
been proposed as a valuable target for a drug-discovery
program.6 Structurally diverse c-secretase inhibitors
have now been reported7 and early clinical results seem
to suggest that a therapeutic window might exist for sev-
eral classes of compounds.8
We previously reported our explorations into a series of
2,6-disubstituted piperidine sulfonamide c-secretase
inhibitors.7b Those compounds proved to be efficacious
at lowering Ab levels in a transgenic mice model of AD
but unfortunately, further testing showed that many
analogs in this series, such as our lead 1, are also potent
inhibitors of the CYP 3A4 liver co-enzyme (Fig. 1).
OH
N
(I)
R1
N
F
O
N
N
S
lower mol. weight
and LogP
O
R3
R2
O
N
S
O
O
O
F
O
O
cLogP = 6.0511
X
Cl
1
Memb Aβ40 IC50 = 2.5 nM
CYP 3A4 = 0.06 μM
R
3 = small (cyclo)alkyl
Keywords: Alzheimer’s disease; c-Secretase; Inhibitor; CYP 3A4; Sul-
fonamide; Piperidine; Conformational constraint.
*
Figure 1. Compound 1 and strategy for smaller less lipophilic analogs.
(See above-mentioned references for further information).
Corresponding author. Tel.: +1 908 740 6557; fax: +1 908 259
2492; e-mail: hubert.josien@spcorp.com
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.08.013

H. Josien et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5330–5335
5331
Table 1. Membrane c-secretase inhibition and CYP3A4 profile across
6-substituted piperidine series using standard carbamates (I, X = 4-Cl)a
Drug–drug interaction due to CYP inhibition is a con-
cern, especially in the context of an aging population of-
ten under multiple treatments.9 We thus sought to
identify and address the factors contributing to this lia-
bility, while preserving oral Ab lowering activity. Inter-
action with CYP 3A4 has been linked to lipophilicity
and the presence of basic amines,10 two hallmarks of
our series in addition to its quite high molecular weight
(ꢀ600). We thus envisioned a series of analogs (I) where
the aryl moiety at R3 has been substituted with smaller
alkyls and where the amine has been reengineered, all
the while preserving the structurally important cyclopro-
pyl carbamate moiety.12 We chose R3 > Me since earlier
results suggested that analogs bearing this moiety would
likely be less potent than others.12
Compound R3
NR1R2
Memb
Ab40
IC50 (nM) (lM)
CYP
3A4c
b
11
12
13
14
15
3,5-diF–Ph
c-Pr
0.3
2.8
2.5
3.5
3.5
<0.8
2.2
0.7
0.8
—
N
N
n-Pr
Et
CF3CH2
1d
3,5-diF–Ph
c-Pr
Et
2.4
27
12
14
0.06
0.5
N
N
N
OH
16d
17d
18d
0.4
<0.3
CF3CH2
19
20
21
22
3,5-diF–Ph
c-Pr
n-Pr
17
46
47
56
1.3
3.5
2.4
11
Compounds of formula (Ia) were prepared according to
a modification of our original scheme, using Stille cou-
pling to install the alkyl side chain and Kulinkovich
reaction13 to functionalize the methyl ester into the
cyclopropyl carbinol 4. Conversion to the carbamate
was performed using previously reported methods
(Scheme 1).7b
OH
Et
a All compounds are racemic.
b Values are means of two experiments.
c Values determined after 30 min pre-incubation with compound.
d clogP = 6.05 (1); 4.92 (16); 5.02 (17); 4.81 (18).11
Preparation of cyclopropyl analogs (Ib) required selec-
tive functionalization of diester 5 into aldehyde 7 as pre-
cursor of the left-hand side cyclopropyl. Oxidation of
alcohol 8 to give ester 9 set the stage for the final steps
as before. Similarly, addition of trifluoromethyltrimeth-
ylsilane on aldehyde 7 followed by Barton–McCombie
deoxygenation14 provided the 2,2,2-trifluoroethyl ana-
logs (Ic) after functionalization (Scheme 2).
a,b
c,d
O
O
OH
()n
N
H
Br
N
()n
N
S
O
O
O
O
2
3
Ar
4
R1
e,f
O
N
()n
N
S
R2
O
O
O
Series were first evaluated for the impact of R3 substitu-
tion on c-secretase inhibitory potency and CYP 3A4 lia-
bility, using standard amines and 4-chlorophenyl
sulfonamide from our previous SAR (Table 1).7b Our
initial results indicated that: (i) significant improvement
in CYP profile was observed with small alkyls at R3
versus the corresponding 3,5-difluorophenyl. It was
accompanied by a drop in c-secretase potency to some
degree, but compounds in the single-digit nanomolar
range could still be obtained; (ii) linear groups were
slightly favored over cyclopropyl in lower molecular
eight series (e.g., 17 vs 16), while the opposite was true
in higher molecular weight series (12 vs 14)15; (iii) while
lipophilicity played a role in CYP liability, comparison
of non-basic to (un)hindered basic amines also showed
that the right-hand side contributed significantly to
CYP liability.
Ar
(Ia)
Scheme 1. Reagents: (a) AllylSnBu3/vinylSnBu3, PdCl2(PPh3)2; (b) H2,
Pd/C; (c) ArSO2Cl, Et3N; (d) Ti(O-i-Pr)4, EtMgBr; (e) p-NO2PhO-
C(O)Cl, pyridine; (f) R1R2NH.
a,b,c
d,e
O
OTBS
O
O
OTBS
N
H
N
OHC
N
S
Ar
O
O
O
O
O
O
5
6
7
f,g,h
i,j
k
R1
N
OH
O
N
S
Ar
N
S
Ar
N
R2
O
S
Ar
O
O
O
O
O
O
O
8
9
(Ib)
In our most orally active series 1,7b ethyl at R3 (17)
offered the most optimal properties (potency, CYP
liability) and the series was progressed further, introduc-
ing various arylsulfonamides and de novo amines (a few
R3 = c-Pr were also prepared for comparison). Pharma-
cokinetic (PK) data for individual compounds were
measured in the rat following oral administration at
10 mpk (Table 2).
R1
N
l,m,n
o,p
F3C
OTBS
7
F3C
O
N
S
Ar
N
S
Ar
R2
O
O
O
O
O
(Ic)
10
Scheme 2. Reagents: (a) NaBH4; (b) H2, Pd/C; (c) TBSCl, imidazole;
(d) 4-Cl–PhSO2Cl, Et3N; (e) DIBAH; (f) Ph3P@CH2; (g) CH2I2,
Et2Zn; (h) TBAF; (i) NaIO4, RuCl3; (j) SOCl2, MeOH; (k) steps d–f; (l)
CF3TMS, BF3ÆOEt2; (m) n-BuLi, PhOC(S)Cl; (n) Bu3SnH, AIBN; (o)
TBAF; (p) steps i–k.
We steered our efforts toward keeping the molecular
weight as low as possible (<550). In our previous

5332
H. Josien et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5330–5335
Table 2. In vitro and AUC data in rat for series (I)a
(I)
R1
N
O
R3
R2
N
S
O
O
O
X
R3
Et
Et
X
NR1R2
Memb Ab40
IC50 (nM)
Cell Ab40
IC50 (nM)
CYP 3A4c (lM)
d
AUC0–6h (hng/mL)
Compound
b
b
OH
23
24
4-Cl
4-Cl
3.2
28
4.3
1.1
3.3
0
N
N
190
—
OH
O
25
26
Et
Et
4-Cl
4-Cl
9.0
13
20
10
1.7
1.0
81
N
N
OH
1869
N
OH
27
Et
4-Cl
12
14
<0.3
3414
N
N
N
N
O
H
28
29
30
14
31
32
33
34
35
36
37
38
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
4-Cl
17
15
1.2
2.2
2.5
0.8
4.6
3.9
5.8
1.6
1.7
1.6
2.6
1.2
1784
892
939
476
405
1780
102
323
1332
91
OH
N
N
N
N
N
4-Cl
6.9
3.5
3.5
10
8.3
7.7
9.1
25
O
H
N
4-Cl
N
N
N
OH
4-Cl
N
4-F
OH
N
4-F
25
N
N
N
OH
4-F
16
115
20
N
3,4-di-F
3,4-di-F
3,4-di-F
3,5-di-F
3,5-di-F
9.6
27
O
H
N
N
N
N
OH
6.9
81
45
N
580
296
OH
29
144
N
N

H. Josien et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5330–5335
5333
Table 2 (continued)
d
AUC0–6h (hng/mL)
Compound
R3
X
NR1R2
Memb Ab40
IC50 (nM)
Cell Ab40
IC50 (nM)
CYP 3A4c (lM)
b
b
N
N
N
39
40
c-Pr
c-Pr
4-Cl
4-Cl
16
12
32
39
0.6
0.4
491
—
O
H
N
OH
a All compounds are racemic.
b Values are means of two experiments.
c Values determined after 30 min pre-incubation with compound.
d Measured over 0–6 h after 10 mpk oral dosing in rat.
series, the N-(2-hydroxyethyl)-piperazine amine moiety
resulted in many interesting compounds,7b but those
were also among the worst offenders in terms of CYP
3A4 liability (compounds 1, 16–18). As previously men-
tioned, the presence of a basic amine can lead to CYP
3A4 liability.10 We thus implemented several strategies
to replace the above moiety with improved surrogates.
In one of those, the basic amine was replaced with a
substituted carbon but the PK of the resulting com-
pound (25) proved to be quite low, as observed in other
non-basic compounds (23 and 24) and in related series
as well.7b In another approach, the length of the hydro-
xyl-ethyl tether was modified (28) to beneficial outcome
in terms of CYP profile. The resulting compound also
had good PK, although cellular potency remained in
the double-digit nanomolar range.
(II)
(III)
Ar
O2S
Ar
Ar
O2S
O2S
N
N
N
O
O
R3
R1R2N
H
R1R2N
O
O
O
R
H
O
H
H
H
H
H
OR
H
R
H
3N
H
H
N
N
OCOR
SO2
Ar
H
OR
H
H
SO2
Ar
SO2
Ar
X-ray7b
(Ar = 4-Cl-Ph; R3 = 3-F-Ph;
R = 2-NO2-Ph-O)
Figure 2. Conformational analysis and 2,6-diaxial lock.
The best results were obtained when introducing steric
hindrance near the piperazine amine, in parallel to what
has been reported by other investigators.10 Introduction
of a substituent directly on the piperazine ring (26) gave
promising data, but our most optimal compounds were
obtained via introduction of a gem-dimethyl next to the
amine, or the use of a bridged piperazine derivative.16
Compounds 29 and 30 showed reasonable PK while
featuring nanomolar cellular c-secretase inhibition
and >2 lM CYP 3A4 potency. Replacement of the
4-chloro-arylsulfonamide with fluorinated arylsulfona-
mides, by contrast, did not significantly alter CYP liabil-
ity or PK and often resulted in somewhat less potent
analogs, an unsurprising result considering that those
modifications do not significantly affect lipophilicity,
molecular weight, and/or basicity at the right-hand side
amine.
(1:4
OH
a,b
c,d
OTBS
OH
OHC
N
S
Ar
N
S
Ar
N
S
Ar
41
O
O
O
O
O
O
7
42ab
Ar
O2S
Ar
O2S
Ar
O2S
Ar
O2S
g
e,f
N
N
N
N
+
+
O
R1R2N
H
H
R1R2N
HO
O
H
H
OH
O
O
43a
43b
(II)
(III)
Scheme 3. Ar = 4-Cl–Ph. Reagents and condition: (a) CH3PPh3Br,
n-BuLi; (b) TBAF; (c) Dess–Martin periodinane; (d) AllylMgBr; (e)
Grubbs’s second generation catalyst; (f) H2, 1 atm, PtO2; (g) Scheme 1
steps e–f.
In the course of this study, NMRs of intermediates and
products were collected to confirm that the chair-like
conformation of the piperidine ring and di-axial orienta-
tion of the 6-substituent and 2-cyclopropylcarbamate
observed in previous series were maintained.7b We also
advantageously had a closer look at the X-ray structure
of a related intermediate (Fig. 2, R3 = aryl) in which we
noticed: (i) close proximity of the cyclopropyl ring
and R3 group; (ii) alignment of the S–C(aryl) and
C(cyclopropyl)–O bonds of the arylsulfonamide and
carbamates, respectively, two important contributors
of c-secretase inhibition.
Those observations led us to environ a series of bridged
analogs ‘linking’ the 6-ethyl and 2-cyclopropyl side
chains (Fig. 2). Modifications brought to those moieties
might also bring additional benefits in the form of better
CYP 3A4 profile and/or metabolism. As shown below,
although two different isomers (II) or (III) could be
designed by branching out of the cyclopropyl, only (II)
could conceivably lock the bond alignment observed in
the X-ray structure.
Access to series (II) and (III) was relatively straightfor-
ward from intermediate 7 (Scheme 3, Ar = 4-Cl–Ph).

5334
H. Josien et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5330–5335
Key steps involved allyl Grignard condensation on an
aldehyde derived from 41 followed by a ring closure
metathesis using Grubs second generation catalyst
(yields were substantially lower using the first genera-
tion).17 The unsaturated diastereoisomeric alcohols were
separated at this stage and, after reduction of the alkene
bond under reduced pressure, processed to the carba-
mates according to previous methods. The Grignard
addition proceeded in a 4:1 ratio in favor of the
preferred series (II). Key intermediates were unambigu-
ously assigned by NMR.
Accordingly, further substitution of the methylene
attached to the bridge might conceivably improve
potency.
Table 4 summarizes in vivo results obtained after dosing
selected analogs to young transgenic pre-plaque Tg
CRND8 mice model of AD. Reduction in plasma Ab40
levels was measured after 3 h following oral or sub-cuta-
neous administration. In this model, compound 29 led
to near-complete abolition of plasma Ab40 levels when
administered at 30 mpk either orally or sub-cutaneously.
Close analog 31 also significantly lowered plasma Ab40
levels when administered sub-cutaneously. Efficacy in
lowering brain Ab40 levels was not measured since we
had found a better related series but it is expected to be sig-
nificant based on high brain penetration (one example).
Assessment of the new series in the c-secretase mem-
brane assay confirmed the X-ray analysis (Table 3): we
were pleased to observe that series (II) bearing the stan-
dard 4-piperidinopiperidine produced a potent analog
(44), whereas series (III) where arylsulfonamide and
carbamate were locked away from the X-ray conforma-
tion was essentially inactive (45). In a related study,
trimmed 5-membered analogs of (II) and (III) were also
synthesized and did not show any potency.18 Series (II)
was further explored and provided several analogs in the
low to mid-single-digit nanomolar activity such as 46.
However, the CYP 3A4 profile did not differ signifi-
cantly from the one observed in the des-cyclic series.
Although this series appears to be somewhat less potent
than its non-bridged predecessor (I, R3 = Et), on closer
look it is actually more related to series (I, R3 = Me).
In summary by lowering the molecular weight of an
early lead and modifying its right-hand side basic amine,
we were able to substantially lessen CYP3A4 liability
while retaining significant Ab40 lowering capability.
Further improvement in that direction including the
recourse to reengineering of the core will be reported
in the near future.
Acknowledgments
The authors thank Drs. A. Buevich, T.-M. Chan, and A.
Evans for their help in confirming structures by NMR as
well as for many helpful discussions. Scale-up of inter-
mediates by Dr. J. Wong and M. Liang is also greatly
appreciated.
Table 3. Membrane c-secretase inhibition and CYP3A4 profile across
bridged piperidine series using standard carbamates (Ar = 4-Cl–Ph)a
Compound Core NR1R2
Memb Ab40 CYP
IC50b (nM)
3A4c (lM)
44
45
46
47
II
15
1.1
—
N
N
N
References and notes
1. (a) McDowell, I. Aging 2001, 13, 143; (b) Debaggio, T.
Losing my Mind: an Intimate Look at Life with Alzhei-
mer’s; Free Press: Simon and Schuster, 2002.
2. Pallas, M.; Camins, A. Curr. Pharmacol. Design 2006, 12,
4389.
III
II
2425
57
N
N
N
0.5
0.6
OH
3. (a) Hardy, J.; Selkoe, D. J. Science 2002, 297, 353; (b)
Selkoe, D. J. J. Am. Med. Assoc. 2000, 283, 1615.
4. Conde, S. Exp. Opin. Therap. Pat. 2002, 12, 503.
5. (a) Lundkvist, J.; Naeslund, J. Curr. Opin. Pharmacol.
2007, 7, 112; (b) Wolfe, M. S. J. Med. Chem. 2001, 44,
2039; (c) Vassar, R. J. Mol. Neurosci. 2001, 17, 151.
6. (a) Evin, G.; Sernee, M. F.; Masters, C. L. CNS Drugs
2006, 20, 351; (b) Churcher, I.; Beher, D. Curr. Pharma-
ceut. Design 2005, 11, 3363; (c) Josien, H. Curr. Opin. Drug
Disc. Dev. 2002, 5, 513.
7. (a) Narlawar, R.; Perez, R.; Blanca, I.; Baumann, K.;
Schubenel, R.; Haass, C.; Steiner, H.; Schmidt, B. Bioorg.
Med. Chem. Lett. 2007, 17, 176; (b) Asberom, T. A.; Zhao,
Z.; Bara, T. A.; Clader, J. W.; Greenlee, W. J.; Hyde, L.
A.; Josien, H. B.; Li, W.; McPhail, A. T.; Nomeir, A. A.;
Parker, E. M.; Rajagopalan, M.; Song, L.; Wong, G. T.;
Zhang, L.; Zhang, Q.; Pissarnitski, D. A. Bioorg. Med.
Chem. Lett. 2007, 17, 511, and references herein.
OH
II
104
N
a All compounds are racemic.
b Values are means of two experiments.
c Values determined after 30 min pre-incubation with compound.
Table 4. In vivo profile following acute dosing for selected compounds
in vivo efficacy
Compound
Memb Ab40
IC50 (nM)
Tg CRND8 mice reduction
in plasma Ab40 (30 mpk, 3 h) (%)
a
29
6.9
ꢁ96b
ꢁ91c
ꢁ85c,d
ꢁ59c
30
31
3.5
10
8. (a) Best, J. D.; Smith, D. W.; Michael, A.; Best, J. D.; Smith,
D. W.; Reilly, M. A.; O’Donnell, R.; Lewis, H. D.; Ellis, S.;
Wilkie, N.; Rosahl, T. W.; Laroque, P. A.; Boussiquet-
Leroux, C.; Churcher, I.; Atack, J. R.; Harrison, T.;
Shearman, M. S. J. Pharmacol. Exp. Ther. 2007, 320, 552;
aValues are means of two experiments.
b Oral dosing.
c Sub-cutaneous dosing.
d Brain concentration/plasma concentration = 3:1.

H. Josien et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5330–5335
5335
(b) Siemers, E. R.; Quinn, J. F.; Kaye, J.; Farlow, M. R.;
Porteinsson, A.; Tariot, P.; Zoulnouni, P.; Galvin, J. E.;
Holtzman, D. M.; Knopman, D. S.; Satterwhite, J.;
Gonzales, C.; Dean, R. A.; May, P. C. Neurology 2006,
66, 602.
15. In a
related series, 20,20-difluorocyclopropyl analogs
proved to be less potent (synthesis not shown).
16. The precursor of the gem-dimethyl amine was prepared
according to the following scheme:
O
1) Et2AlCN
9. (a) Zhang, Z.-Y.; Wong, Y. N. Curr. Drug Metab. 2005, 6,
241; (b) Zhou, S.; Chan, E.; Li, X.; Huang, M. Therap.
Clin. Risk Manag. 2005, 1, 3.
OTBS
+
NH
BocN
N
OTBS
2) MeMgBr
3) TFA
HN
10. (a) Lewis, D. F. V.; Lake, B. G.; Dickins, M. J. Enz. Inhib.
Med. Chem. 2006, 21, 127; (b) Bu, H.-Z. Curr. Drug
Metab. 2006, 7, 231; (c) Riley, R. J.; Parker, A. J.; Triggs,
S.; Manners, C. N. Pharmaceut. Res. 2001, 18, 652.
11. clogPs were calculated using Sybyl 7.1 from Tripos, using
clogP version 4.3 from BioByte, and were within <49 error.
12. Pissarnitski, D. A.; Asberom, T. A.; Bara, T. A.; Buevich,
A. V.; Clader, J. C.; Greenlee, W. J.; Gusik, H. S.; Josien,
H. B.; Li, W.; McEwan, M.; McKittrick, B. A.; Nechuta,
T. L.; Parker, E. M.; Sinning, L.; Smith, E. M.; Song, L.;
Vaccaro, H. A.; Voigt, J. H.; Zhang, L.; Zhang, Q.; Zhao,
Z. Bioorg. Med. Chem. Lett. 2007, 17, 57.
preparation of the bridged piperazine amine used synthetic
routes derived from Cignarella, G.; Nathansohn, G.;
Occelli, E. J. Org. Chem. 1961, 26, 2747, followed by N-
alkylation with BrCH2CH2OTBS and N-debenzylation.
17. Grubbs, R. H.; Trnka, T. M. In Ruthenium in Organic
Synthesis; Murahashi, S.-I., Ed.; Wiley-VCH: Germany,
2004; p 153.
18. Synthesis not shown. In such scaffolds, S–C(aryl) and
C(cyclopropyl)–O bonds of arylsulfonamide and carba-
mates, respectively, are not aligned:
Ar
O2S
13. Kulinkovich, O.; Sviridov, S. V.; Vasilevski, D. A.
Synthesis 1991, 234.
14. Barton, D. H. R.; McCombie, S. W. J. Chem. Soc., Perkin
Trans. 1 1975, 16, 1574.
H
H
N
H
N
H
O
OR
H
R1R2N
SO2
Ar
O
H