Pd-catalyzed ring assembly by vinylation and intramolecular heck coupling: A versatile strategy towards functionalized azadibenzocyclooctynes

Article abstract of DOI:10.1002/chem.201202790

A new modular approach based on Pd-catalyzed C-C bond formation is presented for the assembly of a benzannulated azocine scaffold, the key intermediate in the synthesis of functionalized azadibenzocyclooctynes (aza-DIBOs). The intramolecular ring-closing Heck coupling was investigated by variation of the C-X bond. The reaction rate is limited by the initial oxidative addition step and the regiochemistry strongly depends on the auxiliary phosphine. Under optimized conditions, the 8-endo regioisomer was obtained in 71 % yield over two steps (with no protecting group chemistry) or in one pot, inclusive of C-N bond formation. The practical generation of the octyne triple bond of a prototypical N-benzoyl aza-DIBO, without the need for chromatographic purification, is also described. The structural features, including those of the ring-strained cyclic octyne, were elucidated by NMR spectroscopy and X-ray crystallographic analysis. The high reactivity of the N-benzoyl aza-DIBO synthesized is demonstrated in a strain-promoted azide-alkyne cycloaddition reaction with an alkyl azide (k=0.38 M^-1 s^-1). Copyright

Full text of DOI:10.1002/chem.201202790

DOI: 10.1002/chem.201202790
Pd-Catalyzed Ring Assembly by Vinylation and Intramolecular Heck
Coupling: A Versatile Strategy Towards Functionalized
Azadibenzocyclooctynes
Michael Jꢀger,[a, b] Helmar Gçrls,[c] Wolfgang Gꢁnther,[a] and Ulrich S. Schubert*[a, b]
Abstract: A new modular approach
ditions, the 8-endo regioisomer was ob-
tained in 71% yield over two steps
(with no protecting group chemistry)
for chromatographic purification, is
also described. The structural features,
including those of the ring-strained
cyclic octyne, were elucidated by NMR
spectroscopy and X-ray crystallograph-
ic analysis. The high reactivity of the
N-benzoyl aza-DIBO synthesized is
À
based on Pd-catalyzed C C bond for-
mation is presented for the assembly of
a benzannulated azocine scaffold, the
key intermediate in the synthesis of
functionalized azadibenzocyclooctynes
(aza-DIBOs). The intramolecular ring-
closing Heck coupling was investigated
À
or in one pot, inclusive of C N bond
formation. The practical generation of
the octyne triple bond of a prototypical
N-benzoyl aza-DIBO, without the need
demonstrated in
a strain-promoted
À
by variation of the C X bond. The re-
azide–alkyne cycloaddition reaction
Keywords: azocines · cyclization ·
dibenzocyclooctynes · Heck reac-
tion · strain promoted azide–alkyne
coupling
action rate is limited by the initial oxi-
dative addition step and the regio-
chemistry strongly depends on the aux-
iliary phosphine. Under optimized con-
with an alkyl azide (k=0.38mÀ1 sÀ1).
Introduction
In 2008, Boons et al. used hydroxyl-DIBO for the attach-
ment of dyes and demonstrated the labeling of living cells
by SPAAC.[5] Later work showed the facile preparation of a
variety of bioactive DIBO conjugates,[2,6] mainly based on
the aza-analogue of DIBO, which possesses an amino group
in the ring system (Scheme 1). In the original approach, the
DIBO skeleton was assembled in five steps (inclusive of the
necessary protection/deprotection steps) by Sonogashira
coupling and subsequent ring closure by reductive amination
(Scheme 1a).[6b] Alternatively, a more convenient access in
three steps from commercially available dibenzosuberenone
by ring expansion of the intermediate oxime was recently
Medium-sized unsaturated ring systems serve as a frame-
work for numerous biologically active molecules, for exam-
ple, the benzannulated analogues of seven-membered bal-
anol or eight-membered azocines.[1] In addition, cyclic oc-
tynes have recently attracted great interest as substrates in
strain-promoted azide–alkyne cycloaddition (SPAAC) re-
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
co-workers[4] revealed remarkable potential in bio-orthogo-
nal conjugation strategies. In particular, dibenzocyclooctyne
(DIBO) derivatives represent an attractive class of com-
pounds due to their inherent high reactivity, extraordinary
shelf-life, and the possibility to easily introduce further func-
tionality.
[a] Dr. M. Jꢀger, Dr. W. Gꢁnther, Prof. Dr. U. S. Schubert
Laboratory of Organic and Macromolecular Chemistry (IOMC)
Jena Center for Soft Matter (JCSM)
Friedrich-Schiller-University Jena
Humboldtstraße 10, 07743 Jena (Germany)
Fax : (+49)3641-948-202
E-mail: ulrich.schubert@uni-jena.de
[b] Dr. M. Jꢀger, Prof. Dr. U. S. Schubert
Dutch Polymer Institute (DPI)
John F. Kennedylaan 2
5612 AB Eindhoven (The Netherlands)
[c] Dr. H. Gçrls
Scheme 1. Current approaches to aza-DIBO derivatives from commer-
cially available starting materials (number in parentheses includes gener-
Institute of Analytical and Inorganic Chemistry (IAAC)
Humboldtstraße 2, 07743 Jena (Germany)
ꢀ
ation of the C C triple bond). a) Sonogashira reaction and reductive ami-
Supporting information (experimental details; NMR spectroscopic
data and analysis) for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201202790.
nation, b) ring expansion of the oxime intermediate, c) reductive amina-
tion and double Friedel–Crafts reaction, d) oxidative cleavage of the
Fischer indole intermediate.
2150
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 2150 – 2157

FULL PAPER
reported (Scheme 1b).[7] A third approach is adapted from
the twofold Friedel–Crafts reaction of chlorinated cyclopro-
pene and an electron-rich bis-aryl ethane[6a] generated from
N-benzyl aniline (Scheme 1c).[8] Finally, the eight-membered
ring system can also be prepared in five steps by a Fischer-
type indole synthesis followed by oxidative cleavage
(Scheme 1d).[9] However, these methods require harsh reac-
tion conditions, expensive reagents, protecting group chem-
istry, and/or multiple steps to prepare the intermediate
À
Scheme 2. Retrosynthetic analysis of the DIBO framework by C C
À
bond-formation (steps A and C), C N bond formation (step B), N-func-
ꢀ
octyne precursor. The formation of the C C triple bond can
ꢀ
tionalization (step D), and C C triple-bond generation (step E).
be achieved by 1) addition of dibromine and double dehy-
drohalogenation,[6b,7] 2) UV-triggered extrusion of CO,[6a,8]
or 3) fluoride-assisted elimination from the vicinal trimethyl-
silyl-substituted triflic enol.[2] However, the introduction of
any further functionality always has to precede triple-bond
formation due to the instability of unsubstituted aza-DIBO
under ambient conditions. Hence, we set out to explore
modular assembly of the octene skeleton by Pd-catalyzed
ring formation to establish a practical protocol for the syn-
thesis of novel N-benzoyl-functionalized aza-DIBOs and
their use in SPAAC reactions.
A few related ring systems have been prepared by intra-
molecular Heck coupling, but the direct formation of aza-
DIBO is not reported. In 2000, Skrydstrup et al. reported
the formation of a eight-membered ring (up to 20% yield)
from an undesired side reaction during construction of the
seven-membered-ring system of balanol.[1a] Later studies re-
vealed a broad substrate scope for the cyclization of bromo-
and iodo-aryls; a variety of linkages between the aromatic
units (e.g. ether,[10] aminoacyl,[10a,11] sulfoxo,[10a] and amide[12]
groups) and a range of functional groups (e.g. nitro,[10c]
chloro,[10b] or methoxy)[10b] were tolerated. Notably, strongly
polar solvents favored formation of the 8-endo isomer, al-
though the experimental conditions differ among the re-
ports.
Results and Discussion
Retrosynthetic analysis: Our strategy for the construction of
the DIBO skeleton is based on three bond-forming steps
2
À
(Scheme 2). The C C bond formations between sp -hybri-
Precursor synthesis: The bifunctional precursor for ring for-
mation—equipped with both a vinyl and halide moiety—can
be readily prepared by vinylation of bromo-benzaldehyde
and subsequent reductive imine condensation with the rele-
vant 2-halo aniline (Scheme 3). 2-Vinylbenzaldehyde (1) can
be prepared by Suzuki coupling of vinylboroxine[13] or by a
Hiyama reaction with the inexpensive vinylsiloxane.[14] In
view of a large-scale preparation, the Pd catalyst loading
was lowered (from 5 to 0.5 mol%) and the isolation was
simplified by extraction with diethyl ether and filtration
through silica gel to yield 1 on a multigram scale. Subse-
quent reductive amination with a 2-halo aniline was initially
attempted with sodium borohydride. Although imine forma-
tion was found to be efficient, the desired amine was
dized carbon atoms (Scheme 2, steps A and C) are excellent
candidates for Pd-catalyzed reactions. The C N bond can be
established by alkylation or reductive amination (Scheme 2,
step B). Boonsꢃ route employs C N bond formation for ring
closure and requires the Z configuration of the intermediate
alkene, which had to be generated by Sonogashira coupling
and subsequent partial reduction.[6b] However, the related
Heck-coupling approach yields predominantly the E isomer,
À
À
À
therefore C N bond formation must precede the ring clo-
sure in this case. Hence, we targeted the formation of the
DIBO skeleton based on Heck coupling (Scheme 2, steps A
or C) and an intermediate reductive amination (Scheme 2,
step B).
Scheme 3. Synthesis of bis-functionalized N-benzyl anilines 2a and 2b and precursor 5. i) Siloxane/TBAF or boroxine/K2CO3, Pd0, phosphine. ii)
NaCNBH3, ZnCl2, MeOH. iii) Pd0, phosphine, base.
Chem. Eur. J. 2013, 19, 2150 – 2157
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2151

U. S. Schubert et al.
formed only in very low yield (<5%). Prolonged reaction
times, addition of excess reducing agent, activation with
acetic acid, or use of sodium cyanoborohydride did not im-
prove the yield significantly. However, activation with equi-
molar amounts of zinc chloride[15] furnished clean conver-
sion to the bifunctional precursors 2a and 2b in good yields.
sessed with respect to converted starting material. In the ab-
sence of any phosphine ligand, incomplete conversion of 2a
(53–64% recovered starting material) to 3 as the major
product is observed. In contrast, the use of triarylphosphines
results in faster apparent conversion rates and some forma-
tion of the desired Heck product, which we tentatively
assign to stabilization of the active catalyst. Hence, we were
interested to explore chelating diphosphines to see if forma-
tion of the bridged side product is suppressed. A general im-
provement was found when ferrocenyl-bridged or ethane-
bridged diphosphines (L3 and L4, respectively, Scheme 3)
were used. Although the conversion of 2a was still incom-
plete, the Heck coupling products 4 and 5 were formed in
8–11 and 15–17% yield, respectively (Table 1, entries 5 and
6). More importantly, the alkyl-bridged ligand L4 disfavored
the formation of 3. Consequently, we tested the cyclization
in the presence of an excess of monodentate tri-tert-butyl-
phosphine (L5, Scheme 3), which also resulted in the prefer-
ential formation of 5 (21%). Hence, we focused next on
monodentate dialkyl-aryl phosphines L6 and L7 (Scheme 3),
which are versatile ligands in Pd-catalyzed coupling reac-
tions.[17] The significantly increased yields of 5 (42–49%) are
attributed to the higher conversion of 2a (Table 1, entries 8
and 9). At this point, the influence of temperature and re-
Intramolecular Heck coupling: The crucial step in the for-
mation of the cyclic octene is regioselective ring closure by
Heck coupling. Initially, the reaction conditions described
above for preferential formation of the 8-endo product were
adopted for substrate 2a and a series of commercial phos-
phine ligands were screened (Table 1, entries1–9). Use of
[PdACHTUNGTRENNUNG(dba)2] (dba=dibenzylidene acetone) as precatalyst
without additives gave low conversion and small amounts of
side product 3 but no detectable Heck coupling products
(Table 1, entry 1). Surprisingly, compound 3 was not the an-
ticipated biphenyl coupling product, noted by Skrydstrup
et al. for an ester-bridged analogue.[1a] Instead, NMR spec-
troscopic analysis indicated that bridging within the aza-
octene ring had occurred, judged from the absence of the
NH proton and presence of only four quaternary carbon
atoms (see Figures 23–27 in the Supporting information).
The regiochemistry of the additional cyclization step led to
the N-fused indoline derivative (in contrast to the possible
fused six/four-membered ring), which was identified by com-
parison with reported NMR spectroscopic data.[16] Upon re-
action in the presence of tetrabutylammonium bromide
(TBABr), a substantial amount of 3 and traces of the 7-exo
(4) and 8-endo (5) products were obtained (Table 1,
entry 2). With the addition of triphenylphosphine (L1) the
dominant formation of bridged product 3 was found, along-
side 7-exo product 4 (13%) (Table 1, entry 3). Interestingly,
when the tolylphosphine analogue (L2, Scheme 3) was used
the ratio of Heck coupling products altered and the forma-
tion of 5 (11%) was favored (Table 1, entry 4), despite the
competing bridging reaction. To narrow the range of alter-
native phosphine ligands, the coupling yields were reas-
AHCTUNGTREGaNNNU ction time was investigated. Lowering the temperature to
608C resulted in incomplete conversion (Table 1, entry 10),
whereas 808C was sufficient for cyclization (Table 1,
entry 11). Performing the reaction at 1208C (microwave
heating) resulted in quantitative conversion within 1 h with
comparable yields to the best thermal conditions (Table 1,
entry 12). However, no influence on the product distribution
was observed over the temperature range studied (60–
1208C). Following the course of the reaction proved contin-
uous conversion of the starting material (Figure 1) but
showed that the products are primarily formed in the early
stages (<4 h). A potential halide inhibition of the active cat-
alyst (similar to the addition of TBABr; Table 1, entry 2) or
concurrent decomposition of the products formed may ex-
Table 1. Cyclization of vinyl halides.[a]
Entry
2
Conc [mM]
Pd [mol%]
Additive/ligand[b]
Temperature [8C]
Time [h]
Recovered 2 [%]
3 [%]
4 [%]
5 [%]
1
2
3
4
5
6
7
8
2a
2a
2a
2a
2a
2a
2a
2a
2a
2a
2a
2a
2b
2b
2a
2a
2a
25
25
25
25
25
25
25
25
25
33
33
33
33
33
25
500
570
5
5
5
5
5
5
5
5
5
5
5
5
5
5
1
0.5
1
None
TBABr
L1
L2
L3
L4
L5
L6
L7
L7
L7
L7
L7
L7
L7
L7
L7
100
100
100
100
100
100
100
100
100
60
4
4
4
4
4
4
4
4
4
3.5
3.5
1
3.5
3.5
16
4.5
7
64
53
33
<1
29
41
52
11
1
66
13
<1
69
63
52
24
<1
6
36
52
63
35
5
<1
2
13
1
8
11
4
3
1
2
2
3
3
2
2
3
3
<1
3
5
11
15
17
21
42
49
18
43
56
12
13
27
49
71
6
5
9
31
9
26
24
4
10
11
12
13
14
15
16
17
80
120[c]
23
60
6
100
130
120[c]
25
11
10
[a] Determined by GC versus anisole as an internal standard. [b]Pd/P ratio=1:2. [c] Microwave heating for 1 h.
2152
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 2150 – 2157

Functionalized Azadibenzocyclooctynes
FULL PAPER
(Table 1, entry 17), in line with the high yields based on con-
verted starting material (see above). In addition, the poten-
À
tial intermolecular C C bond formation remains insignifi-
cant even at increased concentrations. In summary, the opti-
mized conditions of the Heck coupling allow the time saving
and economical preparation of 5.
Hiyama–Heck cascade: Inspired by the successful ring for-
À
mation, we set out to combine both C C bond formations
in a one-pot protocol, that is, the monovinylation step and
the cyclization. To the best of our knowledge, one-pot Pd-
catalyzed incorporation of the vinyl fragment into the
DIBO skeleton is not reported. To control both steps the
halide on the aniline subunit was varied (X=Cl, Br, I) to
tune the rate of each oxidative addition step, in particular to
prevent unproductive twofold vinylation. The desired pre-
cursors 2c–e were readily synthesized as described above in
excellent yield (92–93%) and the optimized cyclization con-
ditions for the initial screening were applied with DMF as
the solvent and addition of tetrabutylammonium fluoride
(TBAF) for the activation of vinylsiloxane 6a (Table 2, en-
tries 1–3). However, GC and GC-MS analysis of the re-
*
Figure 1. GC-analysis of the intramolecular Heck coupling of 2a (
)
under the conditions of the initial ligand screening (Table 1, entries 1–9)
&
*
to give the bridged (3; ), 7-exo (4; &) and 8-endo (5; ) products.
plain this experimental observation. To test this hypothesis,
isolated product 5 was subjected to identical reaction condi-
tions (L1 or L7 as ligand); no decomposition and no trans-
formation to 3 was observed. Hence, the bridging is likely to
occur during the Pd catalysis as a competing reaction path-
way and involves formal addition of the secondary amino
group to the double bond. To the best of our knowledge, the
related intermolecular reaction is not reported. More impor-
tantly, compound 5 was formed in up to 73% yield (based
on converted starting material) after 3 h. Next, we tested
the more reactive iodo substrate 2b at 23 and 608C
(Table 1, entries 13 and 14). Interestingly, a similar product
distribution was found, which we assigned to the rate-limit-
AHCTUNGTREGaNNUN ction mixture showed only traces of 5, attributed to ineffi-
cient vinylation of substrates 2c and 2d (ꢁ6%) and signifi-
cant dehalogenation (up to 67%) in the case of 2e. It was
apparent that the selected reaction conditions (DMF at high
temperatures) were incompatible for vinylation by siloxanes
and may potentially hamper the cyclization, as previously
found for TBA salts (Table 1, entry 2). To elucidate the
scope of the vinylation step, we attempted the conversion of
2e without TBAF. Use of potassium fluoride in a DMF/
water mixture furnished the vinylation product (35%),
along with debrominated byproduct (17%) (Table 2,
entry 4). The subsequent intramolecular coupling step oc-
À
ing oxidative addition step. Hence, the subsequent C C
bond formation leads to the same product distribution irre-
spective of the halide source or the reaction temperature. Fi-
nally, to establish a practical large-scale protocol, the scope
of the reaction was investigated with respect to catalyst
loading and substrate concentration (Table 1, entries 15–17).
As expected, a decreased catalyst loading required longer
reaction times and conversion remained incomplete even
after 16 h at 1008C, but a simultaneous increase in reaction
temperature to 1308C counterbalanced this effect and al-
lowed efficient conversion with a twentyfold increase in sub-
strate concentration. Finally, the best results were obtained
for the microwave-assisted protocol to form 5 in 71% yield
AHCTUNGTERNcNGUN urred in the presence of a stronger base (K2CO3) to yield 5
(18%) along with comparable amounts of 3 (18%; Table 2,
entry 5). Motivated by the feasibility of the double cross-
coupling, we attempted to avoid the use of fluoride salts for
activation of the siloxanes and use aqueous potassium car-
bonate instead. Unfortunately, compound 5 was only ob-
tained as the minor product (8%), along with the debromi-
Table 2. Double cross-coupling.[a]
Entry
2
Vinyl donor
(equiv)
Base
Conc
[mM]
Time
[h]
Recovered 2
[%]
Dehalogenation[b]
Divinyl
[%][b]
Monovinyl
[%][b]
3
[%]
5
[%]
G
1
2
3
4
5
6
7
8
2c
2d
2e
2e
2e
2e
2e
2c
2d
2e
2e
2e
6a (1.2)
6a (1.2)
6a (1.2)
6a (4.0)
6a (1.2)
6a (1.2)
6b (2.0)
6c (1.2)
6c (1.2)
6c (1.2)
6c (1.2)
6c (1.2)
TBAF
TBAF
TBAF
KF
KF and K2CO3
K2CO3
K2CO3
K2CO3
K2CO3
K2CO3
K2CO3
K2CO3
30
30
30
160
470
160
470
30
30
30
470
50
4
4
4
1.5
20
2
20
9
9
9
44
20
20
24
21
ND[c]
ND
67
17
2
0
0
0
0
1
0
0
15
28
7
6
1
1
35
14
ND
ND
ND
ND
18
24
ND
1
4
6
20
ND
1
0
0
<1
18
8
<1
<1
<1
<1
<1
<1
<1
<1
<1
11
14
19
15
2
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
48
28
61
73
9
10
11
12
1
ND
33 (52)[d]
[a] L7, 1208C (thermal heating). [b] Identified by GC and GC-MS. [c] Not determined. [d] Based on intermediate 2e.
Chem. Eur. J. 2013, 19, 2150 – 2157
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2153

U. S. Schubert et al.
nated product (11%) and 3 (24%; Table 2, entry 6). Again,
inefficient vinylation leads to severe side reactions, thus, we
considered the fluoride-free activation of vinyltrimethoxysi-
lane 6b with sodium hydroxide, as reported for the high-
yielding Hiyama–Heck reactions under ligand-free condi-
tions.[18] However, no improvement of the vinylation step
was observed when we applied: 1) the original ligand-free
conditions, 2) activation by potassium fluoride, or 3) our op-
timized cyclization conditions (Table 2, entry 7). Hence, the
versatility of the Hiyama–Heck route appears to be limited
by incompatible conditions for each C–C coupling step.
the one-pot preparation of 5 (Scheme 4, right). Because re-
ductive amination with borohydride reagents in the presence
of Pd0 led to partial hydrogenation of the vinyl groups,[19] in-
termediate 2e was prepared alternatively by nucleophilic
substitution of commercially available 2-bromobenzyl bro-
mide[20] and the reaction conditions for ring formation were
adjusted. The desired compound 5 forms in a remarkable
overall yield of 33% (52% based on intermediate 2e;
Table 2, entry 12). The lower yields arise from double alky-
lation of the amino group in the first step, which influences
the stoichiometry in a similar fashion as previously discussed
for the twofold vinylation of 2d. The corresponding product
bears a vinylbenzyl substituent at the nitrogen atom (see
Figures 33–36 in the Supporting Information), which may be
transformed into an alkyne to afford an aza-DIBO function-
alized with a terminal alkyne group. Further optimization is
beyond the scope of this study but is likely to increase the
overall yield.
Suzuki–Heck cascade: At this stage, the related Suzuki–
Heck route seemed a viable alternative because similar con-
ditions for both C–C coupling steps can be applied, demon-
strated for the vinylation with boroxine.[13] Initial experi-
ments with commercially available vinylboroxine 6c
(Scheme 4) showed the efficient vinylation of 2c–e followed
by cyclization to 5 (Table 2, entries 8–10). The product dis-
tribution can be explained by consideration of the reactivity
Cyclooctyne formation: After ring assembly was achieved,
À
of the C X bond—the rates decrease in the order I>Br>
functionalization/protection of the amino group and genera-
ꢀ
Cl—and the fact that vinylation is significantly faster than
cyclization. This leads to efficient monovinylation of the
more reactive halide in 2c (I) or 2e (Br), followed by
tion of the C C triple bond were considered. The intermedi-
ate 5 is readily converted into a functionalized amide, fol-
lowed by addition of dibromine, and subsequent twofold de-
hydrohalogenation to form the alkyne.[6b] However, purifica-
tion by normal-phase chromatography (on either silica or
alumina) is complicated due to the high polarity of the
amide group compared to the subtle changes in the non-
polar skeleton (in agreement with the available 1H NMR
spectra).[7] Because of the flexibility of the commonly used
alkyl linker, the isolation of the final alkyne by crystalliza-
tion was not reported and the materials were used without
further purification.
À
slower reaction of the remaining C X bond to effectively
furnish the overall cyclization product in 48 and 61% yield,
respectively. In the case of 2c, which has two C Br bonds,
À
divinylation occurs before cyclization, which consumes the
available vinyl groups and also prevents the cyclization,
shown by the significantly lower yield of 5 (28%). In con-
trast to the Hiyama–Heck route, only traces of the dehalo-
genation products and small amounts of regioisomer 4 and
bridged coupling product 3 were formed. Notably, the gener-
ality of the regioselective Suzuki–Heck cascade for ring-as-
sembly is demonstrated by 2c, which proceeds by vinylation
of the aniline subunit and subsequent cyclization. In line
with previous results, large-scale preparation from 2e at in-
creased reagent concentration and reduced catalyst loading
Hence, we decided to replace the alkyl chain by a more
rigid aromatic unit to circumvent this problem. The benzoic
acid analogue 7 is smoothly generated in good isolated yield
(81%) and is cleanly converted to dibromo adduct 8
(Scheme 5). The 1H NMR spectrum of 8 (Figure 2, top)
shows a complex pattern with significant line broadening of
some peaks, which is attributed to two conformers (65:35)
based on NOESY and HSQC analysis (see Figures 44–45 in
the Supporting Information). The EXSY cross-peak in the
NOESY spectrum between both species is attributed to con-
allowed the preparation of
entry 11).
5 in 73% yield (Table 2,
One-pot preparation: The modular character of the ap-
proach may be extended to include C N bond formation in
À
Scheme 4. Double cross-coupling. i) NaCNBH3, ZnCl2, MeOH. ii) Siloxane, fluoride, Pd0, phosphine. iii) Boroxine/K2CO3, Pd0, phosphine. iv) K2CO3;
then boroxine, Pd0, phosphine.
2154
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 2150 – 2157

Functionalized Azadibenzocyclooctynes
FULL PAPER
crystal contains both enantiomers with the absolute R,R-
and S,S-configuration at C1 and C2. Finally, the 1H NMR
spectrum of a redissolved crystal gave an identical ratio of
the two species and supports the assignment of conformers
in solution. The final step to obtain octyne 9 involves two-
fold dehalogenation, achieved by treatment with potassium
tert-butoxide in THF. The 1H NMR spectrum (Figure 2,
bottom) shows a similar pattern to cyclooctene 7, but lacks
the vinylic proton signals. The methylene protons display
sharp resonances, which indicates rigidity in the strained cy-
clooctyne ring. The structure of 9 was also investigated by
X-ray analysis (Figure 3). The DIBO skeleton is flattened,
with an angle of 19.28 between the mean planes of the ben-
zene subunits. It is interesting to compare this value to the
available structural data of related cyclooctynes.[22] Only two
organic structures are reported that contain a sulfoxo
group[23] or two sulfur atoms[24] in the ring. Although the ad-
jacent carbon atoms are sp3 hybridzed, similar torsion angles
of the triple bond of 16.8 and 21.88, respectively, are ob-
served. The triple bond in 9 (1.204(4) ꢄ) is slightly longer
than for the sulfur analogues (1.191 ꢄ) and comparable to
non-strained tolane (1.198(3) ꢄ).[25] However, ring strain is
clearly present, indicated by the bond angles between adja-
cent carbon atoms in the ring [C2-C1-C7=151.0(3)8; C1-C2-
C3=155.4(3)8]. The values are significantly smaller than for
both cyclic analogues (160.3–164.68) and are comparable to
copper-complexed cyclooctyne compounds (152–1558).[26]
The spatial separation between the peripheral bromo func-
ꢀ
Scheme 5. N-Functionalization and C C triple-bond formation. i) Acid
chloride, CH2Cl2, NEt3, À58C. ii) Br2, CH2Cl2, 08C, 2 h. iii) KOtBu, THF,
À488C, 3.5 h.
1
Figure 2. Aromatic region of the H NMR spectra (300 MHz, CD2Cl2) of:
top) 8 with assignment of the minor (white boxes) and major (grey
boxes) conformers; bottom) 9.
formational exchange because of the opposite phase sign to
the NOE cross-peak. This behavior is in line with studies on
the conformational behavior of nonfunctionalized dibenzo-
cyclooctane.[21] The X-ray structure of 8 (Figure 3) further
supports the assignments and shows the most important
structural features: 1) the anti arrangement of the bromine
substituents, 2) the transoid arrangement of the amide bond
ꢀ
tionality and the center of mass of the C C triple bond in 9
is approximately 7.9 ꢄ. Notably, this distance should be (ap-
proximately) preserved upon rotation along the amide bond
as a consequence of the rigid linker.
À
À
À
(O1 C16 N1 C4), and 3) the boat conformation in the
solid state. Due to the centrosymmetric space group, the
SPAAC reaction of cyclooctyne 9: The reactivity of octyne
9 was tested in a SPAAC reaction with an azide (Figure 4,
top). The reaction was performed at 297 K in CDCl3 be-
cause compound 9 was insoluble in deuterated methanol,
which is the common solvent for kinetic studies. Hence, we
decided to utilize w-azido-undecanol (10), a hydrophobic
aliphatic azide with a terminal hydroxyl group to add fur-
ther functionality. The progress of the reaction was followed
by 1H NMR spectroscopy by monitoring the well-resolved
CH protons of the cyclooctyne moiety and the terminal
methylene units of the azide (Figure 5). To ensure the cor-
rect correlation of peak area and concentration, the relaxa-
tion time for the protons was determined (t=1.3 s) and the
delay time between the scans in the kinetic experiment pro-
longed. The reaction was expected to follow second-order
kinetics with different starting concentrations. The analytical
solution of the rate law predicts a linear relationship be-
tween reaction time and the logarithm of the corrected con-
centrations, to account for the different starting concentra-
tions (Figure 4, bottom).[27] The concentrations of the alkyne
and azide were extracted independently from the NMR
spectroscopic data and an overlay of the plots shows the
same decrease of both species with an average rate constant
of k=0.377Æ0.03mÀ1 sÀ1 (see Figure 50 in the Supporting In-
Figure 3. X-ray structures of: top) 8; bottom) 9. Thermal ellipsoids drawn
at 50% probability.
Chem. Eur. J. 2013, 19, 2150 – 2157
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2155

U. S. Schubert et al.
dure was extended to combine both C–C coupling steps (vi-
nylation and cyclization), which required an efficient mono-
vinylation reaction prior to intramolecular cyclization. Initial
attempts by a Hiyama–Heck cascade to exploit inexpensive
vinylsiloxane gave unsatisfactory yields, attributed to incom-
patible reaction conditions that led dominantly to dehaloge-
nation. Instead, the Suzuki–Heck methodology furnished 5
in up to 73% yield. The modularity of the cascade reaction
À
is supported by variation of the C X bond because the cy-
clooctene ring also forms efficiently from the vinylic aniline
À
subunit. Finally, a one-pot synthesis that included C N bond
formation demonstrated the versatility of this modular strat-
egy. Although the isolated yield (33%) was lower and may
be optimized further, the presented modular approach
allows direct synthesis of 5, despite the multitude of possible
side reactions.
ꢀ
We addressed the efficient generation of the C C triple
bond of N-benzoyl DIBO 9 by N-acylation, addition of di-
bromine, and subsequent twofold dehydrohalogenation. The
intermediate dibromine adduct shows hindered flexibility of
the ring skeleton, assigned to two conformers based on 2D
NMR spectroscopic data, and supported by X-ray analysis.
The final cyclic octyne is an air- and room temperature
stable, microcrystalline compound. Notably, the X-ray crys-
tal structure of this compound class gives experimental
access to structural parameters, for example, the angular
Figure 4. Top) SPAAC reaction of 9 (7.8 mm) and w-azido-undecanol 10
(10.0 mm) in CDCl3 at 297 K; bottom) kinetic analysis based on the peak
areas of the octyne or azide methylene groups, respectively.
ꢀ
strain of the C C triple bond (151–1558), the dihedral tor-
sion angle of the benzannulated rings (198), and a set spatial
separation of 7.9 ꢄ between the functional group of the
linker and the alkyne. Finally, the reactivity of 9 was demon-
strated in a SPAAC reaction with an aliphatic azide to
reveal a high second-order rate constant of k=0.38mÀ1 sÀ1.
The presented modular strategy is expected to facilitate
fast and efficient large-scale synthesis of DIBO derivatives
by a series of Pd-catalyzed cascade reactions employing
commercially available anilines, benzaldehydes, and benzoic
acids. In addition, the bromo group may be further trans-
formed by utilizing advances in efficient Pd-catalyzed cross-
coupling protocols to introduce further functionality. Hence,
this approach is not only of relevance to SPAAC reactions
or the assembly of biologically active scaffolds, but may also
be exploited in the future to introduce regioselectivity and/
or enhance chemoselectivity, for example, by placing bulky
groups on the adjacent aromatic units or by manipulation
with electron-donating or -withdrawing substituents.
Figure 5. 1H NMR spectra (400 MHz, CDCl3) with proton assignments of
9 (top), 10 (middle), and the SPAAC reaction mixture after 40 min
(bottom). Arrows indicate the consumption of 9 (white) and 10 (grey),
and the appearance of the SPAAC product (black) during the course of
the reaction.
formation), very close to the values observed for the related
N-alkoxy-functionalized aza-dibenzocyclooctynes.[6b]
Conclusion
A new modular strategy for the synthesis of a prototypical
functionalized N-benzoyl aza-DIBO was developed that in-
volved: 1) assembly of the ring system in 5, and 2) genera-
tion of crystalline N-benzoyl alkyne 9.
Retrosynthetic analysis of the intermediate cyclooctene 5
suggested construction by two C–C coupling reactions and
Experimental Section
À
one C N bond formation from abundant, commercially
available starting materials. The scope of this strategy relies
on the regioselectivity of the intramolecular Heck reaction
for cyclization. After initial optimization of the reaction
conditions (phosphine ligand, aryl halide, temperature), the
undesired side reactions were suppressed to favor the 8-
endo product (up to 71% yield). Remarkably, additional
bridging between the vinyl and amino group was observed
to directly afford a N-fused indoline derivative. The proce-
CCDC-876510 (8) and CCDC-876511 (9) contain the supplementary crys-
tallographic data for this paper. These data can be obtained free of
charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
General procedure for reductive amination (2): A flask was charged with
2-substituted benzaldehyde (1 equiv), 2-substituted aniline (1.1–
1.4 equiv), and methanol (0.7–4.8 mLmmolÀ1). Zinc(II) chloride
(1.1 equiv) was added and the suspension was stirred at RT for 10 min.
Sodium cyanoborohydride (1.1–1.2 equiv) was added and the mixture
2156
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 2150 – 2157

Functionalized Azadibenzocyclooctynes
FULL PAPER
was heated at 808C for 45 min. The reaction mixture was allowed to cool
to RT and diluted with a saturated aqueous solution of sodium carbonate
(30 mL) and brine (200 mL). The suspension was extracted with diethyl
ether (ꢅ3) and the organic layers were combined, dried over sodium sul-
fate, filtered, and adsorbed onto silica. Excess solvent was removed
under reduced pressure and the crude product was purified by flash
column chromatography on silica gel (SNAP KP-Sil, 50 g, hexane/CH2Cl2
90:10–0:100).
[1] a) M.-P. Denieul, B. Laursen, R. Hazell, T. Skrydstrup, J. Org.
Chem. 2000, 65, 6052–6060; b) K. C. Majumdar, R. K. Nandi, S. Sa-
manta, B. Chattopadhyay, Synthesis 2010, 985–990.
[2] J. C. Jewett, C. R. Bertozzi, Chem. Soc. Rev. 2010, 39, 1272–1279.
[3] H. C. Kolb, M. G. Finn, K. B. Sharpless, Angew. Chem. 2001, 113,
2056–2075; Angew. Chem. Int. Ed. 2001, 40, 2004–2021.
[4] N. J. Agard, J. A. Prescher, C. R. Bertozzi, J. Am. Chem. Soc. 2004,
126, 15046–15047.
[5] X. H. Ning, J. Guo, M. A. Wolfert, G.-J. Boons, Angew. Chem. 2008,
120, 2285–2287; Angew. Chem. Int. Ed. 2008, 47, 2253–2255.
[6] a) A. A. Poloukhtine, N. E. Mbua, M. A. Wolfert, G.-J. Boons, V. V.
Popik, J. Am. Chem. Soc. 2009, 131, 15769–15776; b) M. F. Debets,
S. S. van Berkel, S. Schoffelen, F. P. J. T. Rutjes, J. C. M. van Hest,
F. L. van Delft, Chem. Commun. 2010, 46, 97–99.
[7] A. Kuzmin, A. Poloukhtine, M. A. Wolfert, V. V. Popik, Bioconju-
gate Chem. 2010, 21, 2076–2085.
[8] F. Starke, M. Walther, H.-J. Pietzsch, ARKIVOC 2010, 11, 350–359.
[9] J. C. Jewett, E. M. Sletten, C. R. Bertozzi, J. Am. Chem. Soc. 2010,
132, 3688–3690.
[10] a) L. A. Arnold, W. Luo, R. K. Guy, Org. Lett. 2004, 6, 3005–3007;
b) K. C. Majumdar, I. Ansary, B. Sinha, B. Chattopadhyay, Synthesis
2009, 3593–3602; c) B. C. G. Sçderberg, J. W. Hubbard, S. R. Rector,
S. N. OꢃNeil, Tetrahedron 2005, 61, 3637–3649.
[11] L. A. Arnold, R. K. Guy, Bioorg. Med. Chem. Lett. 2006, 16, 5360–
5363.
General procedure for intramolecular Heck coupling (5): A microwave
vial was charged with 2 (1 equiv), [PdACHTNUGRTENUNG(dba)2] (0.005–0.05 equiv), phos-
phine (0.01–0.1 equiv), base (2 equiv), and DMF (2 mLmmolÀ1) that con-
tained anisole as a GC-standard. The vial was purged with N2 for 5 min
and the reaction mixture was heated at 60–1308C (see Table 1). The
crude products were purified by flash column chromatography on silica
gel (hexane/CH2Cl2 or toluene).
General procedure for Pd-catalyzed cascade reactions (5): A microwave
vial was charged with
2
(1 equiv), [PdACTHNGUTERN(UNG dba)2] (0.01 equiv), L7
(0.02 equiv), base (2 equiv) and 6 (1.2–4.0 equiv). After addition of DMF
(2 mL mmolÀ1) and water (0.1 mL mmolÀ1) containing anisole as GC-
standard, the vial was sealed, purged with N2, and heated to 1208C for
the given times (see Table 2). The crude product was purified by flash
column chromatography on silica eluting with hexane/CH2Cl2 or toluene.
Compound 7: Compound 5 (0.650 g, 3.14 mmol) was dissolved in dry
CH2Cl2ACHTUNGTRENNUNG (40 mL) and triethylamine (4 mL) was added. The mixture was
cooled to À58C, then 4-bromobenzoyl chloride (0.765 g, 3.49 mmol) was
added. The reaction mixture was washed with 2m aqueous NaOH (ꢅ2),
2m aqueous HCl (ꢅ2), and brine (ꢅ2). The organic layer was dried over
sodium sulfate, filtered, and adsorbed onto silica by removal of excess
solvent under reduced pressure. The crude product was purified by flash
column chromatography on silica gel (hexane/ethyl acetate 90:10) to
afford 7 (0.994 g, 2.55 mmol, 81%).
[12] A. Aoyama, H. Aoyama, M. Makishima, Y. Hashimoto, H. Miyachi,
Heterocycles 2009, 78, 2209–2216.
[13] C. Slugovc, B. Perner, F. Stelzer, K. Mereiter, Organometallics 2004,
23, 3622–3626.
[14] S. E. Denmark, C. R. Butler, Org. Lett. 2006, 8, 63–66.
[15] S. Kim, C. H. Oh, J. S. Ko, K. H. Ahn, Y. J. Kim, J. Org. Chem. 1985,
50, 1927–1932.
[16] S. J. Mahoney, E. Fillion, Chem. Eur. J. 2012, 18, 68–71.
[17] T. E. Barder, S. D. Walker, J. R. Martinelli, S. L. Buchwald, J. Am.
Chem. Soc. 2005, 127, 4685–4696.
Compound 8: A solution of 7 (0.440 g, 1.127 mmol) in CH2Cl2 (20 mL)
was cooled to 08C. Dibromine (0.250 g, 1.564 mmol) was added and the
reaction mixture was stirred for 2 h. Excess dibromine and solvent were
removed under reduced pressure and 8 (600 mg, 1.09 mmol, 97%) was
used in the next step without further purification.
[18] A. Gordillo, E. de Jesus, C. Lopez-Mardomingo, Chem. Commun.
2007, 4056–4058.
Compound 9: A solution of 8 (0.500 g, 0.909 mmol) in dry THF (2 mL)
was cooled to À488C. A solution of potassium 2-methylpropan-2-olate
(0.218 g, 1.943 mmol) in dry THF (2 mL) was slowly added. The mixture
was stirred for 3.5 h at À488C then allowed to warm to RT. Excess sol-
vent was removed under reduced pressure at RT, an aqueous solution of
ammonium chloride was added, and the mixture was extracted with
chloroform. The organic layer was dried over sodium sulfate, filtered,
and excess solvent was removed under reduced pressure. The crude prod-
uct was dried under high vacuum to yield 9 (0.353 g, 0.798 mmol, 88%).
[19] Identified by GC-MS and NMR spectroscopy.
[20] R. A. Earley, M. J. Gallagher, J. Chem. Soc. C 1970, 158–163.
[21] I. Alkorta, J. Elguero, Struct. Chem. 2010, 21, 885–891.
[22] CSD search (2011–12–07) Version 5.32: cycloctyne, non-metal atoms
were chosen for the remaining ring.
[23] J. K. Crandall, T. A. Ayers, J. Org. Chem. 1992, 57, 2993–2995.
[24] F. M. Kerton, G. F. Mohmand, A. Tersteegen, M. Thiel, M. J. Went,
J. Organomet. Chem. 1996, 519, 177–184.
[25] A. Mavridis, I. Moustakalimavridis, Acta Crystallogr. Sect. B: Struct.
Sci. 1977, 33, 3612–3615.
[26] G. Grçger, U. Behrens, F. Olbrich, Organometallics 2000, 19, 3354–
3360.
SPAAC reaction: The reaction was performed by mixing solutions of
alkyne 9 and azide 10 in CDCl3 (7.8 mm and 10.0 mm, respectively) at
297 K and was monitored by 1H NMR spectroscopy. The concentrations
of 9 and 10 were calculated from the peak areas of the well-resolved
methylene groups versus the OCH2 group as an internal standard.
[27] H. Mauser, in Formale Kinetik, Bertelsmann Universitꢀtsverlag,
Dꢁsseldorf, 1974, pp. 34–35.
Received: August 3, 2012
Published online: December 19, 2012
Acknowledgements
The authors thank the Dutch Polymer Institute (DPI) for financial sup-
port. Furthermore, we thank Anja Baumgꢀrtel for MALDI-TOF MS
measurements and Esra Altuntas for ESI-TOF MS measurements.
Chem. Eur. J. 2013, 19, 2150 – 2157
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2157