Selective synthesis and kinetic measurement of 1:1 and 2:2 cyclic compounds containing 1,4,7,10-tetraazacyclododecane and azobenzene units

Article abstract of DOI:10.1021/jo000926p

1:1 cyclic compounds 8a-c (51-55%) and 2:2 cyclic compounds 9a-c (20-49%) containing 1,4,7,-10-tetraazacyclododecane (cyclen) and azobenzene units were selectively synthesized under UV irradiation (330 nm < λ < 380 nm) and in the dark. Synthesis depended on the wavelength of irradiation light and the length of methylene chains of the linker between the cyclen and azobenzene units. A study of NMR and UV-vis spectra indicated that properties of 8a-c and 9a-c are closely related to their structural flexibility. Rate constants (k) and thermodynamic parameters (ΔG^?, ΔH^?, and ΔS^?) of 8a-c and 9a-c were studied in nonpolar media (benzene) and polar media (methanol). The cis to trans isomerization rates in the dark for these cyclic compounds increase with ring size or structural flexibility (8a < 8c < 8b < 9a < 9b < 9c). In principle, ΔS^? dominates ΔG^? in cyclic compounds.

Full text of DOI:10.1021/jo000926p

J . Org. Chem. 2000, 65, 8979-8987
8979
Selective Syn th esis a n d Kin etic Mea su r em en t of 1:1 a n d 2:2 Cyclic
Com p ou n d s Con ta in in g 1,4,7,10-Tetr a a za cyclod od eca n e a n d
Azoben zen e Un its
Wen-hao Wei, Takenori Tomohiro,* Masato Kodaka, and Hiroaki Okuno
Biomolecules Department, National Institute of Bioscience and Human-Technology, 1-1 Higashi,
Tsukuba, Ibaraki 305-8566, J apan
tomohiro@nibh.go.jp
Received J une 19, 2000 (Revised Manuscript Received October 24, 2000)
1:1 cyclic compounds 8a -c (51-55%) and 2:2 cyclic compounds 9a -c (20-49%) containing 1,4,7,-
10-tetraazacyclododecane (cyclen) and azobenzene units were selectively synthesized under UV
irradiation (330 nm < λ < 380 nm) and in the dark. Synthesis depended on the wavelength of
irradiation light and the length of methylene chains of the linker between the cyclen and azobenzene
units. A study of NMR and UV-vis spectra indicated that properties of 8a -c and 9a -c are closely
related to their structural flexibility. Rate constants (k) and thermodynamic parameters (∆Gq, ∆Hq,
and ∆Sq) of 8a -c and 9a -c were studied in nonpolar media (benzene) and polar media (methanol).
The cis to trans isomerization rates in the dark for these cyclic compounds increase with ring size
or structural flexibility (8a < 8c < 8b < 9a < 9b < 9c). In principle, ∆Sq dominates ∆Gq in cyclic
compounds.
In tr od u ction
preliminarily reported the photochemical synthesis of two
cyclic compounds 1:1 8a -c and 2:2 9a -c. Here we report
the synthesis in detail as well as the results of a study
of their rate constants and thermodynamic parameters
in cis to trans isomerization in the dark.
Trigger-regulated compounds have been extensively
applied to molecular devices based on functional/multi-
functional regulation,1a,b molecular capsules,1c,d self-
organized supramolecular systems,1e-g etc. Photoswitch
systems of azobenzene, thiophenefulgide, and nitrospiro-
pyran, for example, have been specially used as sensitive
triggers in molecular recognition and assembly chemis-
try.2,3 To mimic the enzymatic reaction of metal proteins,
we recently synthesized a series of photoswitchable
compounds comprising a tetraazacyclododecane (cyclen)
as a metal binding site and an azobenzene unit as a light
trigger.4 Cis-trans isomerization of azobenzene upon
irradiation by light may change the shape of the ligand
molecule, enabling it to change properties of metal
complexes. As Shinkai et al. reported on the synthesis
of cyclic compounds containing a crown ether and an
azobenzene unit, these cyclic compounds were prepared
by condensation reaction between acid chlorides and
amines or by reduction of two nitro groups followed by
diazotization.2 These cyclization reactions, however, gen-
erally gave very low yields. In a previous work,4 we
Resu lts a n d Discu ssion
P r ep a r a tion of Com p ou n d s 4 a n d 7. Compounds 4
and 7 were synthesized as shown in Scheme 1. 4,4′-
Dihydroxylazobenzene 1 was prepared based on the
method of Willsta¨tter and Benz.5 Compounds 2a -c were
obtained in excellent yields of 95-99% by a reaction of
Br(CH2)nNH2‚HBr (n ) 2, 3, and 4) with carbobenzoxy
chloride (ZCl). In the process of alternate addition of ZCl
and 3.5 M K2CO3, the reaction solution was maintained
at pH 6-7 to avoid formation of pale yellow byproducts.
Protecting groups Z of 3a -c were readily cleaved by
treatment with 33% HBr/AcOH yield ammonium salts
4a -c. 1,7-Diprotected cyclen 5 was derived from the
reaction of cyclen and ZCl.6 Diacetic acid 6 was precipi-
tated by adding 6 M HCl and was directly subjected to
elemental analysis without further purification. Activa-
tion of carboxylic acids of 6 with 1,3-thiazolidine-2-thione
(TTH) was done in the presence of DDC or 1-ethyl-3-(3-
dimethylaminopropyl)-carbodiimide hydrochloride (WSC).
Compounds 1-7 were characterized on the basis of 1H
and 13C NMR, FAB-MS, and/or TOF-MS and elemental
analysis except for 7, which contained impurities even
after purification twice by recrystallization and TLC.
Cycliza tion Rea ction s. Cyclization reactions between
4 and 7 were conducted under high-dilution conditions
(Scheme 1 and Table 1). A solution of 7 in THF and a
solution of 1.0 equiv of 4b (n ) 3) and Et3N in THF-
H2O were simultaneously added to a large volume of THF
at 2 mL/min. The product was isolated by preparative
TLC on an alumina plate. When the reaction was
conducted under UV irradiation (330 nm < λ < 380 nm),
(1) (a) Zelikovich, L.; Libman, J .; Shanzer, A. Nature (London) 1995,
394, 790-792. (b) Wu¨rthner, F.; Rebek, J ., J r. J . Chem. Soc., Perkin
Trans. 2 1995, 1727-1734. (c) Fujita, M.; Oguro, D.; Miyazawa, M.;
Oka, H.; Yamaguchi, K.; Ogura, K. Nature (London) 1995, 378, 469-
471. (d) Rebek, J ., J r. J . Chem. Soc., Chem. Commun. 2000, 637-643.
(e) Lehn, J .-M. Spramolecuar Chemistry. Concepts and Perspectives;
VCH: Weiheim, 1995. (f) Lawrence, D. S.; J iang, T.; Levett, M. Chem.
Rev. 1995, 95, 2229-2260. (g) De Silva, A. P.; Gunaratne, H. Q. N.;
Gunnlaugsson, T.; Huxley, A. J . M.; McCoy, C. P.; Rademacher, J . T.;
Rice, T. E. Chem. Rev. 1997, 97, 1515-1566.
(2) (a) Shinkai, S.; Minami, T.; Kusano, Y.; Manabe, O. J . Am. Chem.
Soc. 1983, 105, 1851-1856. (b) Shinkai, S.; Nakaji, T.; Nishida, Y.;
Ogawa, T.; Manabe, O. J . Am. Chem. Soc. 1980, 102, 5860-5865. (c)
Shinkai, S.; Honda, Y.; Kusano, Y.; Manabe, O. J . Chem. Soc., Chem.
Commun. 1982, 848-850.
(3) (a) Willner, I. Acc. Chem. Res. 1997, 30, 347-356. (b) Behrendt,
R.; Renner, C.; Schenk, M.; Wang, F.; Wachtveil, J .; Oesterhelt, D.;
Moroder, L. Angew. Chem., Int. Ed. Engl. 1999, 38, 2771-2773. (c)
Asanoma, H.; Ito, T.; Yoshida, T.; Liang, X.; Komiyama, M. Angew.
Chem., Int. Ed. Engl. 1999, 38, 2393-2395.
(4) Wei, W.-H.; Tomohiro, T.; Kodaka, M.; Okuno, H. J . Chem. Soc.,
Perkin Trans. 1. 1999, 3397-3398.
(5) Willsta¨tter, R.; Benz, M. Chem. Ber. 1906, 339, 3492-3503.
(6) Kovacs, Z.; Sherry, A. D. Synthesis 1997, 759-763.
10.1021/jo000926p CCC: $19.00 © 2000 American Chemical Society
Published on Web 12/06/2000

8980 J . Org. Chem., Vol. 65, No. 26, 2000
Wei et al.
Sch em e 1a
a
Key: (i) KOH, 200 °C; (ii) Br(CH2)nNH‚HBr, ZCl, K2CO3, pH 6-7; (iii) (a) NaH, (b) 60 °C; (iv) HBr/AcOH; (v) ZCl, NaOH, pH 2-3;
(vi) (a) BrCH2COOH, KOH, (b) 1, Na2CO3, 45-65 °C; (vii) TTH, WSC, DMAP; (viii) 330 nm < λ < 380 nm; (ix) in the dark.
Ta ble 1. Ma in P r od u cts in Rea ction s of Com p ou n d s
irradiation and converted to the trans form in the dark.
The cis-acyclic compound (Scheme 1) readily closed the
ring to give 8a or 8b. However, the trans-acyclic com-
pound is in a disadvantageous configuration to conduct
the 1:1 intramolecular cyclization reaction. Both 2:2 cyclic
compounds 9a and 9b were primarily obtained, although
intramolecular cyclization is generally preferential to
intermolecular reaction in a high dilution. The CPK
model examination predicts that 1:1 intramolecular
cyclization can proceed even in the trans-acyclic com-
pound if methylene chain linkers are long enough.
Actually, in contrast to 4a and 4b (n ) 2 and 3),
compound 4c (n ) 4) reacted with 7 in the dark to yield
both 8c at 27% and 9c at 20% (Table 1). Under UV
irradiation, it yielded 8c at 55%.
4a -c w ith 7a
UV (330-380 nm)
8a 8b 8c
52 55
850 878
0.55 0.57
in the dark
product
9a
9b
9c + 8c
20 + 27
yield, % 51
47
1643
49
m/ zb
822
0.47
1669
0.30 0.32
1755
0.33
c
Rf
a
a n ) 2; b n ) 3; c n ) 4; n is the number of -(CH2)n-
(methylene chain linkers). FAB-MS/[M + H]+ (100%). c Alumina
TLC eluted with 80% EtOAc-MeCN.
b
1:1 cyclic compound 8b was the predominant product
isolated in a 52% yield. In the dark,7 however, 8b was
not detected on TLC and both 2:2 cyclic compound 9b at
49% and 1:1 “acyclic” product [Rf ) 0.18 MeCN/MeOH/
Et3N (8:1:0.1), m/z (FAB) 868.0 (M + H)+] at 8% were
obtained. Analogous results were obtained in the reaction
of 4a (n ) 2) and 7. UV-vis spectra of the acyclic product
(n ) 3) showed that the cis form was produced under UV
Rea ction Con d ition s for Cycliza tion . Yields of 8b
and 9b under different reaction conditions are shown in
Table 2. All reactions were conducted in high dilution at
0.1 mmol (reactant)/200 mL (solvent) at room tempera-
ture. Products were isolated by preparative TLC. Typical
results are shown in entries 1 and 2. Under UV irradia-
tion (entry 1), 8b was yielded at 52% and traces of 9b
(7) To completely convert the cis form to the trans form, solutions
were irradiated under light (λ > 460 nm) for 5 min and allowed to
stand in the dark at 30 °C for 8 h, or allowed to stand in the dark for
24 h at 30 °C before a solution of 4 was added in the dark. UV-vis
spectra showed that the trans form absorption intensity increased
slightly in the dark after visible irradiation (λ > 460 nm).
1
were observed on TLC. H NMR spectra for the cis-4b
showed small characteristic signals of trans-4b; namely,
trans to cis isomerization was not complete under UV

Synthesis of 1:1 and 2:2 Cyclic Compounds
J . Org. Chem., Vol. 65, No. 26, 2000 8981
Ta ble 2. Yield s of 8b a n d 9b u n d er Differ en t Rea ction
Con d ition s
reaction conditions
product (%)
entry methoda reactant Et3N (equiv) lightb
8b
52
9b
1
2
I
I
4b + 7
4b + 7
4b + 7
4b + 7
4b + 7
4b + 7
10d + 7
10 + 7
4b + 6
4b + 6
4b + 6
1
1
UV
trace
dark N.D.c 49
3
4
5
6
7
8
9
10
11
I
I
I
I
II
II
III
III
III
1
1
1.5
1.5
vis
W.L. 5-13 8-23
UV 44 trace
dark N.D. 36
UV 53 trace
dark N.D. 52
W.L. N.D. N.D.
UV
5
40
7
trace
dark N.D. trace
a
Method: (I) a solution of 4b + Et3N and of 7 were simulta-
neously added to THF-H2O; (II) same procedure as (I) using THF
instead of THF-H2O; (III) WSC was used for the coupling reaction
b
in DMF. Irradiation under 330 nm < λ < 380 nm (UV) and under
λ > 460 nm (vis); in the dark (dark); white light of common
laboratory illumination (W.L.). c No compound was detected by
d
TLC. Compound 10 is a free amine derived in advance from
ammonium salt 4b.
irradiation. This accounts for the formation of traces of
9b. Under visible irradiation (λ > 460 nm), about 10% of
4b was present in the cis form judging from UV-vis spec-
tra of 4b. Consistent with this prediction, under visible
irradiation (entry 3), both 9b (40%) and 8b (5%) were
obtained rather than only 9b (49%) as observed in the
dark (entry 2). When the reaction was conducted under
common laboratory illumination (entry 4), compounds 8b
and 9b were not selectively synthesized and yields were
rather low. Because of 4b solubility, solvent THF-H2O
and base Et3N were used for reactions. However, the use
of excess Et3N (entries 5 and 6) lowered yields of 8b and
9b, particularly for 9b, to 36% from 49%, since 7 was
easily hydrolyzed in alkaline solution. To improve the
reaction by method II of increasing reactant solubility
and limiting the effect of hydrolysis, free amine 10 de-
rived from ammonium salt 4b and THF as solvent were
used. As shown in entries 7 and 8, however, the yields of
8b and 9b hardly changed. In addition, ca. 30% of 4b was
lost in preparing 10. We attempted to synthesize com-
pound 8b by direct treatment of diacid acid 6 with 4b in
the presence of Et3N and WSC (entry 9), but the desired
products 8b and 9b were scarcely obtained (entries
9-11). This indicates that compound 7 bearing the TTH-
amide group is a suitable compound for the reaction.8
Sp ectr oscop ic Stu d y. Cyclic compounds 8a -c and
9a -c were characterized on the basis of 1H and 13C NMR,
FAB-MS or TOF-MS, elemental analysis, and UV-vis
spectra. NMR signals of 8b and 9b were assigned from
2D-NMR spectra of 1H-1H COSY and HMQC. The
wavelengths of electronic absorption maximum (λmax) for
8a -c and 9a -c are listed in Table 3 and typical UV-
vis spectra (8a and 9a ) are shown in Figure 1. Isomer-
ization of the azobenzene unit for 8a -c and 9a -c quickly
reached a photostationary state under visible irradiation
(λ > 460 nm) for cis to trans and under UV irradiation
(330 nm < λ < 380 nm) for trans to cis within 10 min at
room temperature. Even after visible irradiation (λ > 460
nm) for 10 min, cis forms (cis-8a , cis, cis-9a ) were not
completely converted to trans forms (trans-8a -vis, trans-
9a -vis) (Figure 1). When solutions were allowed to stand
F igu r e 1. UV-vis spectra for 8a and 9a in MeOH at 25 °C.
Trans-vis and cis spectra were measured after visible (λ > 460
nm) and UV (330 nm < λ < 380 nm) irradiation for 10 min.
trans-8a and trans,trans-9a spectra were measured after
reaching the thermal-stationary state in the dark.
in the dark for 5-16 h at 60 °C, absorption intensities
slowly increased and reached a thermal-stationary state
(trans-8a , trans,trans-9a ). Isomerization at 25 °C for
8a -c and 9a -c was neglected.
Especially noteworthy (Table 3) is that some spectro-
scopic properties of trans-8a -c (smaller ring) differ from
those of trans,trans-9a -c (larger ring), whereas those of
all cyclic compounds in the cis form corresponded well
to linear compound 10 (data not shown). These spectro-
scopic characteristics are summarized below.
1
(1) Each H NMR spectrum of trans-8a -c showed that
3,3′,5,5′-H (meta protons to azo) gave a broad signal in
CDCl3 at 25 °C, whereas 2,2′,6,6′-H (ortho protons) gave
a doublet (Figure 2). With linear compound 10 and large
ring compounds 9a -c in the trans,trans form, both
3,3′,5,5′-H and 2,2′,6,6′-H protons appeared as sharp
doublets. At a higher temperature (55 °C), 3,3′,5,5′-H of
8b turned into a doublet at 6.97 ppm due to the magnetic
equivalence of these protons. At a lower temperature
(-20 °C), however, it became 3 doublet signals (6.83, 6.95,
and 7.08 ppm with 1:2:1 intensity ratio). 2,2′,6,6′-H
protons also gave 3 doublet signals (7.81, 7.84, and 7.92
ppm) at -20 °C. Namely, aromatic protons of the azoben-
zene unit are not magnetically equivalent at a lower
temperature, and 3,3′,5,5′-H is subjected to a larger effect
than 2,2′,6,6′-H. This clearly indicates that trans-8b has
a rigid conformation. Restricted rotational movement of
the azobenzene unit causes nonequivalence of 3,3′,5,5′-H
resulting in the broad signal at 25 °C. CPK model studies
support this conclusion. Note that signals of the meth-
ylene linker protons between the cyclen and azobenzene
1
unit split into two spin systems in H-1H COSY spectra
of trans-8b at -20 °C, suggesting that the whole molec-
ular shape deviates from C2 symmetry.
(2) A downfield shift (0.36-0.41 ppm) of O-CH2 protons
(azobenzene-O-CH2-) was observed for trans-8a -c com-
pared to corresponding protons of trans,trans-9a -c,
indicating that O-CH2 protons tend to exist near a plane
of the azobenzene ring. The deshielding effect of the
(8) (a) Nagao, Y.; Miyasaka, T.; Seno, K.; Fujita, E. J . Chem. Soc.,
Perkin Trans. 1 1984, 2439-2446. (b) J acques, V.; Mesbahi, M.;
Boskovic, V.; Desreux, J . F. Synthesis 1995, 1019-1026.

8982 J . Org. Chem., Vol. 65, No. 26, 2000
Ta ble 3. Typ ica l Da ta on NMR a n d Absor p tion λm a x of UV-Vis Sp ectr a
Wei et al.
1:1 cyclic compound
2:2 cyclic compound
linear
8a
8b
8c
9a
9b
9c
10b
NMMa (ppm)
6.95 br
7.79 d
4.29 br
116.6
3,5-H
2,6-H
O-CH2
3,5-C
trans
trans
trans
trans
trans
7.07 br
7.86 d
4.38 br
118
69.2
6.98 br
7.85 d
4.28 br
116.8
67.7
6.91 d
7.74 d
4.02 br
114.6
65.8
6.80 d
7.73 d
3.89 br
114.5
6.81 d
7.73 d
3.88 br
114.4
65.6
6.89 d
7.86 d
4.01 br
114.4
O-CH2
67.2
65.6
65.2
UV-vis, λmaxc (nm)
MeOH
C6H6
trans
cis
trans
cis
364
446
366
445
362
363
449
366
446
356
447
356
448
358
447
356
355
445
360
447
448
365
447
447
359
447
358
447
a
In CDCl3; br ) broad, d ) doublet, J H ) 8.3-8.9 Hz; 3,5-H (C) and 2,6-H to azobenzene unit; O-CH2 of azobenzene-O-CH2; after
reaching a thermal-stationary state in the dark, spectra were recorded at 25 °C for the trans (trans-8a -c and trans,trans-9a -c); and
b
after UV irradiation (330 < λ < 380 nm) for 10 min for the cis form (cis-8a -c and cis,cis-9a -c). Linear compound 10 is a free amine
derived from ammonium salt 4b. c Another cis-λmax at 310-312 nm in C6H6 and one at 315-318 nm in MeOH are omitted for clarity.
compared to corresponding carbons of trans,trans-9a -c.
As chemical shifts in 13C NMR are generally dominated
by electron density on carbon atoms, the above results
suggest a decrease of the electron density on 3,3′,5,5′-C
and O-CH2.
(4) λmax values of trans-8a -c were larger by 6-8 nm
(viz., red shift) than those of trans,trans-9a -c both in
benzene (359 nm) and in methanol (356 nm, Figure 1).
λmax values of all cyclic compounds in the cis form and
trans,trans-9a -c were close to those of linear compound
10. As mentioned above, due to the restricted conforma-
tion of trans-8a -c, the p-orbital of nonbonding electrons
of the adjacent oxygen overlaps more efficiently with the
π-orbitals of the azobenzene. The resulting “additional
conjugation” causes a red shift. The red shift was also
supported by INDO/1 calculation (ZINDO of the CAChe
system, Oxford Molecular Group).
Thus, these interesting properties of NMR and UV-
vis spectra are closely related to the structural flexibility
of cyclic compounds. Structural flexibility is reconsidered
below to interpret cis-trans isomerization kinetics.
Kin etic Mea su r em en t a n d Th er m od yn a m ic P a -
r a m et er s Ca lcu la t ion . Here we focus on thermal
isomerization from cis to trans in the dark (Figure 1).
We measured rate constants (k) and derived free energy
(∆Gq), enthalpy (∆Hq), and entropy (∆Sq) of activation in
the formation of trans-8a -c and trans,trans-9a -c in a
polar solvent (methanol) and a nonpolar solvent (ben-
zene). k values were determined from the increase in the
absorption maximum of trans forms. The final absor-
bance of the trans form at 60 °C was used as the value
corresponding to 100% conversion. It was also confirmed
that back-transformation from trans to cis by the light
of the spectrophotometer was negligible during our
measurement time. In the case of the 2:2 products 9a -
c, there are two azobenzene moieties in a molecule. The
molar extinction coefficient for azobenzene moieties was
nearly double that of corresponding 1:1 products 8a -c
(Figure 1). 1H NMR spectra for 9c on the halfway through
isomerization clearly showed 4 doublets of aromatic
protons whose chemical shifts [6.63 (d), 6.79 (d), 6.88 (d),
7.78 (d)] were different from those of corresponding
signals for all-cis [6.66 (d), 6.81 (d)] and all-trans isomers
[6.81 (d), 7.73 (d)]. The intensity of these signals was
increased after heating of all-cis isomer, and then de-
creased with an increase of all-trans isomer. No further
spectral changes were observed in NMR and electronic
absorption spectra of compounds 9a -c after UV irradia-
F igu r e 2. Temperature dependence of 3,3′,5,5′-H and 2,2′,-
6,6′-H 1H NMR signals of 8b in the trans,trans form.
azobenzene ring current results in the downfield shift of
O-CH2 protons. The NOE difference spectrum (1-D) of
8a in the trans form at room temperature showed that
irradiation of O-CH2 protons clearly enhanced the signal
of 3,3′,5,5′-H on the azobenzene ring and irradiation of
latter protons also enhanced O-CH2 protons. These
results support the concept that O-CH2 protons are
located near 3,3′,5,5′-H. The same proton enhancements
were not observed for 8a in the cis form or 9a in the
trans,trans form. The interaction of O-CH2 protons with
3,3′,5,5′-H was also observed in the NOESY spectrum (2-
D) of trans-8b at 0 °C. Spectra (2-D) did not show any
interaction of Z group protons with other protons, indi-
cating that the effect of Z groups on the azobenzene unit
is precluded.
(3) Similar to 1H NMR, 13C NMR signals of 3,3′,5,5′-C
on the azobenzene ring and O-CH2 carbons in trans-8a -c
showed downfield shifts (2.2-3.4 ppm and 1.6-3.2 ppm)

Synthesis of 1:1 and 2:2 Cyclic Compounds
J . Org. Chem., Vol. 65, No. 26, 2000 8983
Ta ble 4. F ir st-Or d er Ra te Con sta n ts (k) for Isom er iza tion fr om Cis to Tr a n s
k × 105, s-1
in benzene
2:2 cyclic compound
in MeOH
2:2 cyclic compound
1:1 cyclic compound
linear
10a
1:1 cyclic compound
linear
T (°C)
8a
8b
8c
9a
9b
9c
8a
8b
8c
9a
9b
9c
10
40
50
60
1.91
5.47
14.0
5.09
15.9
44.0
3.39
10.4
30.0
6.55
20.6
59.0
7.80
24.5
70.6
8.21
26.1
75.3
12.3
36.6
104
2.15
6.64
19.4
8.20
25.6
67.0
6.95
20.9
60.1
8.67
26.8
77.3
10.6
32.7
95.0
11.0
34.5
100
13.3
40.1
112
a
Linear compound 10 is a free amine derived from ammonium salt 4b.
tion (converted to all-cis isomers) or heating (to all-trans
isomers). These indicated that cis, trans isomers should
exist during the isomerization from all-cis (cis, cis) to all-
trans (trans, trans) isomers. Note that all of kinetic data
for compounds 9 described in the following discussion are
based on one azobenzene moiety. The isomerization
obeyed first-order kinetics in both solvents (Table 4). ∆Hq
and ∆Sq were calculated from linear plots (Figure 3, r >
0.998) of ∆Gq against T. The resulting values for 8a -c,
9a -c, and 10 are summarized in Table 5 and shown in
Figure 4. Benzene is a nonpolar solvent (viz., no perma-
nent dipole moment), although it can be polarized by an
outer electric field. The solvation of benzene should be
much lower than that of methanol since the donor
number (DN) is 0.1 and the acceptor number (AN) of
benzene is 8.2, whereas the DN of methanol is 19 and
the AN of methanol is 41.3.9 The kinetic properties of a
benzene solution are thus considered to be more domi-
nated by intrinsic structures, whereas those of a metha-
nol solution are assumed to reflect both intrinsic molec-
ular characteristics and solvation.
Kin etics in Ben zen e. In benzene solution, k of cyclic
compounds tends to increase with the ring size (8a < 8c
< 8b < 9a < 9b < 9c) (Table 4), i.e., ∆Gq tends to
decrease with ring size, meaning that isomerization is
easier for more flexible compounds. To reveal what factor
dominates the reaction rate, we estimated the individual
contribution of ∆Hq and T∆Sq to ∆Gq (Figure 4). The
change of ∆Hq and T∆Sq is relatively monotonic similar
to ∆Gq. The dependence of ∆Gq on the ring size appears
to be governed by T∆Sq for cyclic compounds, since both
∆Hq and T∆Sq tend to increase with ring size. This result
on T∆Sq is reasonable, because the increase in structural
restriction of transition states based on ground states is
expected to be less in larger-ring compounds than in
smaller-ring compounds and therefore T∆Sq is assumed
to move positively as the ring size increases (Figure 4).
In other words, the rate acceleration of 9a -c compared
to 8a -c is attributable to their ∆Sq terms, which are less
negative than those of 8a -c (Table 5), since the larger
2:2 ring is more flexible than the smaller 1:1 ring. The
length of the methylene chain (n ) 2, 3, 4) also affects
isomerization (Table 5). The magnitudes of ∆Sq terms are
in the order 8a < 8b < 8c for 1:1 compounds and 9a <
9b < 9c for 2:2 compounds. These orders are consistent
from the standpoint of structural flexibility. It is also
intriguing to compare ∆Hq and ∆Sq values between cyclic
(8a -c and 9a -c) and linear (10) molecules, which may
correspond to a cyclic compound with an infinite ring size.
Compound 10 gives the largest reaction rate of all
compounds, caused by a decrease in ∆Hq (Figure 4 and
Table 4).
F igu r e 3. ∆Gq vs T in (a) benzene and (b) MeOH. Linear
compound 10 is a free amine derived from ammonium salt 4b.
Kin etics in Meth a n ol. In a polar solvent like metha-
nol, the effect of solvation is considerable and may
appreciably influence the reaction rate. In a methanol
solution, k of cyclic compounds tends to increase with ring
size as in a benzene solution (Table 4), i.e., ∆Gq tends to
decrease with ring size. One point different from the
benzene system is that the ∆Hq and ∆Sq values of 8a
deviate from monotonic curves (Figure 4). Clearly, how-
ever, the dependence of k on ring size is dominated by
(9) (a) Gutmann, V.; Wychera, E. Inorg. Nucl. Chem. Lett. 1966, 2,
257-260. (b) Mayer, U.; Gutmann, V. Adv. Inorg. Chem. Radiochem.
1975, 17, 189-230.

8984 J . Org. Chem., Vol. 65, No. 26, 2000
Ta ble 5. En th a lp y (∆Hq) a n d En tr op y (∆Sq) of Activa tion for Isom er iza tion fr om Cis to Tr a n s in Ben zen e a n d MeOH
in benzene in MeOH
2:2 cyclic compound 2:2 cyclic compound
Wei et al.
1:1 cyclic compound
linear
10
1:1 cyclic compound
linear
8a
8b
8c
9a
9b
9c
8a
8b
8c
9a
9b
9c
10
∆Hq (kcal/mol)
∆Sq (eu)
20.0
21.7
21.9
22.1
22.2
22.3
21.7
22.1
21.1
21.7
21.9
22.1
22.2
21.4
-7.9
-16.2 -8.99 -8.87 -7.21 -6.61 -6.02 -6.80 -9.17 -9.82 -8.32 -6.94 -6.19 -5.79
F igu r e 4. ∆Gq, ∆Hq, and T∆Sq vs compounds in benzene and
MeOH (T ) 298 K).
F igu r e 5. Compensation relationship between ∆Hq and ∆Sq
in (O) benzene and ([) MeOH.
∆Sq such as for a benzene solution, except for 8a , whose
reaction rate is governed by ∆Hq rather than ∆Sq. As
observed in benzene, 10 also gives the largest rate of all
compounds in methanol (Table 4) due to the decrease in
∆Hq (Figure 4).
temperatures are those above which an increase in ∆Sq
results in a decrease in ∆Gq and below which an increase
in ∆Sq leads to an increase in ∆Gq. With the present
reaction temperature, the former case holds for both
solvent systems (Figure 4). Since 10 has a type different
from other azobenzene compounds, it probably deviates
from the compensation relationship in benzene. The
result that 8a shows unusual kinetic behavior (Figure
4) and deviates from the compensation relationship in
methanol (Figure 5) suggest the possibility that methanol
may solvate to 8a in a manner different from other
compounds.
Com p a r ison of Kin etics in Ben zen e a n d Meth a -
n ol. It was reported that the rate of isomerization from
cis- to trans-azobenzene is enhanced by polar media.10
Shinkai et al. reported that the rate of this isomerization
in o-dichlorobenzene is greater by a factor of 20 than that
in benzene.2b They also reported, however, that the rate
was somewhat smaller in methanol than in benzene.2b
This result is not compatible with our observation that
all cyclic and linear molecules studied in this experiment
are transformed faster in methanol than in benzene,
although the rate enhancement was 2 times at most
(Table 4). Note that ∆Hq contributes favorably to rate
acceleration in methanol, except for 8a , whose isomer-
ization is enhanced mainly by the ∆Sq term. In 8c and
9a -c, ∆Sq also contributes favorably in addition to ∆Hq.
Com pen sation Relation sh ip. We plotted ∆Hq against
∆Sq to confirm whether the isomerization proceeded
under the same mechanism (Figure 5). Moderately good
linearity was obtained both in benzene (r ) 0.994) and
in methanol (r ) 0.950). Since 10 in benzene and 8a in
methanol deviate largely from linearity, they were not
included in the calculation. Linear relations are ex-
pressed by eqs 1 and 2
Con clu sion s
We successfully regulated the cyclization reaction
photochemically. The 1:1 cyclic compound 8a -c (51-55%
of yields) and 2:2 9a -c (20-49%) were selectively
synthesized under UV light irradiation (330 nm < λ <
380 nm) and in the dark. NMR and UV-vis spectra,
kinetic data of cis to trans isomerization, and cyclization
yields are all consistently explained by the relative
structural flexibility of 8a -c and 9a -c. For metal
binding, we preliminarily observed that the Cu2+ complex
of the 1:1 cyclic ligand (8b) undergoes photocontrolled
catch and release of anionic ligands. A further study is
under way and will be published elsewhere.
Hq ) 225∆Sq + 23.7 (in benzene)
Hq ) 269∆Sq + 23.7 (in methanol)
(1)
(2)
Exp er im en ta l Section
Gen er a l P r oced u r e a n d Ma ter ia ls. All chemicals and
solvents were used without further purification unless other-
wise noted. All cyclization reactions of 8a -c and 9a -c were
conducted under N2. 4-Bromobutylamine hydrobromide, Br-
(CH2)4NH2‚HBr was prepared via a 2-step reaction as reported
where isokinetic temperatures are 225 and 269 K. These
(10) Zalukaev, L. P.; Voronkov, M. G.; Moiseeva, L. V.; Afanasev,
S. V. Dokl. Akad. Nauk SSSR. 1976, 230, 136.

Synthesis of 1:1 and 2:2 Cyclic Compounds
J . Org. Chem., Vol. 65, No. 26, 2000 8985
by Brown and Gulick.11 1H and 13C NMR spectra were acquired
at 270 and 67.8 MHz, respectively. NOE difference spectrum
(1-D) and NOESY (2-D) were acquired at 500 MHz. Mass
spectra were recorded by MALDI TOF and FAB methods.
Elemental analysis was obtained from the Analytical Center
in the National Institute of Materials and Chemical Research,
J apan. A 500 W super-high-pressure Hg lamp was used
through a UV-D36B filter (Toshiba) for UV irradiation (330
nm < λ < 380 nm) and through a Y-46 filter (Toshiba) for
visible irradiation (λ > 460 nm). Aluminum oxide 60 F254
neutral (Merck) and silica gel 60 F256 (Merck) were used for
TLC.
Mea su r em en t s of NMR Sp ect r a of Tr a n s a n d Cis
F or m s. NMR spectra of trans and cis forms were recorded as
follows: Samples (10-12 mg) dissolved in 0.6 mL of solvent
in a NMR tube were allowed to stand in the dark for 5-16 h
at 60 °C. After reaching a thermal-stationary state, spectra
in the trans form were measured at room temperature.
Solutions in a UV cell with a cap were irradiated by UV light
(330 nm < λ < 380 nm) for 10 min at 25 °C. After irradiation,
cis spectra of NMR were measured immediately.
Mea su r em en ts of Kin etic Ra te Con sta n ts. The sample
was sealed in a quartz UV cell with a cap. The concentration
of the sample was decided by a maximum absorption intensity
of 0.9-1.0. Rate constants of isomerization from cis to trans
in the dark were determined by the increase of intensity at
the wavelength of maximum absorption of trans from the cis
photostationary state. This state was rapidly reached under
UV irradiation (330 nm < λ < 380 nm) for 10 min. In the
process of irradiation, the temperature of samples was kept
the same as the workup temperature. The workup time was 1
h for each temperature.
4,4′-Dih yd r oxya zoben zen e 1. Compound 1 was prepared
based on the method of Willsta¨tter and Benz.5 A mixture of
KOH (50 g, 760 mmol), p-nitrophenol (10 g, 72 mmol), and
water (10 mL) was heated to 120 °C and left to stand for 1 h.
When the temperature slowly rose to 195-200 °C, the reaction
vigorously started to give a brown viscous liquid with a large
number of bubbles developing. After the reaction was com-
pleted, products were dissolved in water. A dark-red solution
was acidified to pH 3 with concentrated HCl and extracted
with ether. Combined ether extracts were dried over Na2SO4
overnight. Ether was removed to dryness under reduced
pressure. Residue was recrystallized from 50% (v/v) ethanol
aqueous solution to give yellow crystals of compound 1. 4.0 g,
43% yield. Rf ) 0.32 (silica, hexane-EtOAc 1:1). TOF-MS: m/z
215 (M + H)+. δH (DMSO-d6): 10.10 (s, 2H), 7.71 (d, J ) 8.91,
4H); 6.90 (d, J ) 8.90, 4H). δC (DMSO-d6): 160.1, 145.6, 124.3,
115.9.
N-Ben zyloxyca r bon yl-2-br om oeth yla m in e 2a . 2-Bromo-
ethylamine hydrobromide, Br(CH2)2NH2‚HBr (2.05 g, 10 mmol),
was dissolved in a solution of water (5 mL) and dioxane (5
mL). The solution was treated with alternating addition of a
solution of carbobenzoxy chloride, ZCl (2.10 g, 12 mmol) in
dioxane (5 mL), and of 3.5 M K2CO3 with vigorously stirring
at room temperature. The pH was maintained at 6-7 by
dropwise addition. After ZCl was added over 0.5 h, it was
stirred under pH 7-8 for 1 h. Excess ZCl was hydrolyzed with
2 M aqueous NaOH (1 mL) with stirring for 2 h. The mixture
was extracted with ether. The combined ether was washed
with 1 M NaOH once and with water twice and dried over
Na2SO4 overnight. A colorless oil of 2a was isolated with a
silica gel column. 2a was slowly crystallized, 2.45 g, 95% yield.
Rf ) 0.43 (silica, hexane-EtOAc 10:1). TOF-MS: m/z 259 (M
+ H)+. Anal. Calcd for C10H12BrNO2: C, 46.53; H, 4.69; N, 5.43;
Br, 30.96. Found: C, 46.57; H, 4.54; N, 5.40; Br, 30.72. δH
(CDCl3): 7.37-7.32 (m, 5H), 5.12 (s, 2H), 3.60 (q, J ) 5.61,
2H), 3.46 (t, J ) 5.61, 2H). δC (CDCl3): 156.3, 135.8, 128.4,
128.1, 127.8, 66.4, 39.0, 31.4.
(2.19 g, 10 mmol) and ZCl (2.10 g, 12 mmol) to give a colorless
oil of compound 2b, 2.70 g, 99% yield. Rf ) 0.47 (silica,
hexane-EtOAc 10:1). TOF-MS: m/z 273 (M + H)+. Anal. Calcd
for C11H14BrNO2: C, 48.55; H, 5.18; N, 5.15; Br, 29.36.
Found: C, 48.60; H, 5.10; N, 5.13; Br, 29.43. δH (CDCl3): 7.36-
7.31 (m, 5H), 5.12 (s, 2H), 3.43 (t, J ) 6.43, 2H), 3.37 (q, J )
6.43, 2H), 2.07 (t, J ) 6.43, 2H). δC (CDCl3): 156.3, 136.1,
128.4, 128.1, 127.6, 66.4, 39.0, 32.2, 30.5.
N-Ben zyloxyca r bon yl-3-br om obu tyla m in e 2c was pre-
pared in a manner analogous to 2a from Br(CH2)4NH2‚HBr
(2.33 g, 10 mmol) and ZCl (2.10 g, 12 mmol) to give a colorless
oil of compound 2c, 2.73 g, 95% yield. Rf ) 0.50 (silica, hexane-
EtOAc 10:1). TOF-MS: m/z 287 (M + H)+. Anal. Calcd for
C
12H16BrNO2: C, 50.37; H, 5.63; N, 4.89; Br, 27.92% Found:
C, 50.40; H, 5.62; N, 5.77; Br, 27.83. δH (CDCl3): 7.36-7.31
(m, 5H), 5.12 (s, 2H), 3.42 (t, J ) 6.60, 2H), 3.23 (q, J ) 6.43,
2H), 1.87-1.69 (m, 2H). δC (CDCl3): 156.4, 136.5, 128.5, 128.1,
127.0, 66.7, 40.1, 33.1, 29.8, 28.7.
4,4′-Di[γ-(N-b en zyloxyca r b on yl)a m in oet h a n oxy]a zo-
ben zen e 3a . To a solution of 1 (0.321 g, 1.5 mmol) in anhy-
drous DMF (5 mL) was added a white suspension of NaH
(0.180 g, 4.5 mmol, 60% content) in anhydrous DMF (10 mL).
The hydrogen produced was removed under reduced pres-
sure until bubbles no longer developed from the yellow
suspension. To this, N-benzyloxycarbonyl-2-bromoethylamine
2a (0.936 g, 3.6 mmol) was added in anhydrous DMF (10 mL).
After being stirred for 4 h at 60 °C, the mixture was cooled
and poured into cold water (60 mL). Yellow crystals of 3a
were recrystallized from a mixture solvent of 2-propanol and
ethanol, 0.579 g, 68% yield. TOF-MS: m/z 570 (M + H)+. Anal.
Calcd for C32H32N4O6: C, 67.59; H, 5.67; N, 9.85. Found: C,
67.50; H, 5.73; N, 9.85. δH (DMSO-d6): 7.82 (d, J ) 8.91, 4H),
7.36-7.31 (m, 10H), 7.10 (d, J ) 8.91, 4H), 5.04 (s, 4H), 4.10
(t, J ) 5.61, 4H), 3.42 (q, J ) 5.61, 4H). δC (DMSO-d6): 160.6,
156.3, 146.2, 137.1, 128.3, 127.7, 124.1, 115.0, 66.7, 65.3,
38.6.
4,4′-Di[γ-(N-ben zyloxyca r bon yl)a m in op r op a n oxy]a zo-
ben zen e 3b was prepared in a manner analogous to 3a from
1 (0.321 g, 1.5 mmol) and 2b (0.987 g, 3.6 mmol) to give yellow
crystals of 3b, 0.664 g, 74% yield. TOF-MS: m/z 598 (M + H)+.
Anal. Calcd for C34H36N4O6: C, 68.44; H, 6.08; N, 9.39.
Found: C, 68.39; H, 6.11; N, 9.43. δH (DMSO-d6): 7.86 (d, J )
8.91, 4H), 7.36-7.31 (m, 10H), 6.97 (d, J ) 8.91, 4H), 5.11 (s,
4H), 4.10 (t, J ) 5.94, 4H), 3.44 (q, J ) 6.43, 4H), 2.05 (t, J )
6.10, 4H). δC (DMSO-d6): 160.9, 156.3, 146.2, 137.3, 128.4,
127.8, 124.2, 115.0, 65.7, 65.3, 37.3, 29.1.
4,4′-Di[γ-(N-b en zyloxyca r b on yl)a m in ob u t a n oxy]a zo-
ben zen e 3c was prepared in a manner analogous to 3a from
1 (0.321 g, 1.5 mmol) and 2c (1.038 g, 3.6 mmol) to give yellow
crystals of 3c, 0.667 g, 71% yield. TOF-MS: m/z 626 (M + H)+.
Anal. Calcd for C36H40N4O6: C, 69.21; H, 6.45; N, 8.97.
Found: C, 69.22; H, 6.50; N, 8.84. δH (CDCl3): 7.86 (d, J )
8.91, 4H), 7.36-7.31 (m, 10H), 6.97 (d, J ) 8.91), 5.11 (s, 4H),
4.05 (t, J ) 6.60, 4H), 3.30 (q, J ) 6.43, 4H), 1.84-1.73 (m,
4H). δC (CDCl3): 160.8, 156.4, 147.0, 136.5, 128.5, 128.1, 124.3,
114.5, 67.7, 66.7, 40.7, 26.8, 26.4.
4,4′-Di(γ-a m in oeth a n oxy)a zoben zen e Dih yd r obr om id e
4a . A saturated solution (1.0 g, 4.45 mmol) of 33% dry
hydrogen bromide in glacial acetic acid was added to 3a (0.284
g, 0.5 mmol) in a reaction flask with a calcium chloride dry
tube. Immediately after the addition of the reagent, carbon
dioxide began to develop. The mixture was allowed to stand
at room temperature with occasional shaking for 1 h and
washed with ether. A scarlet solid dried with desiccator 4a
was obtained in 88% yield, 0.203 g. TOF-MS: m/z 301 (M +
H)+. Anal. Calcd for C16H20N4O2‚2HBr: C, 41.58; H, 4.36; N,
12.12. Found: C, 41.30; H, 4.46; N, 12.02. δH (DMSO-d6): 8.15
(br, 6H), 7.85 (d, J ) 8.91, 4H), 7.12 (d, J ) 8.91, 4H), 4.30 (t,
J ) 5.61, 4H), 3.27 (q, J ) 5.61, 4H). δC (DMSO-d6): 160.1,
146.5, 124.2, 115.3, 64.6, 38.31.
N-Ben zyloxycar bon yl-3-br om opr opylam in e 2b was pre-
pared in a manner analogous to 2a from Br(CH2)3NH2‚HBr
4,4′-Di(γ-a m in op r op a n oxy)a zob en zen e d ih yd r ob r o-
m id e 4b was prepared in a manner analogous to 4a from 3b
(0.298 g, 0.5 mmol) and 33% HBr/AcOH (1.0 g, 4.45 mmol) to
give a black solid of 4b, 0.236 g, 96% yield. TOF-MS and FAB-
MS: m/z 329 (M + H)+. Anal. Calcd for C18H24N4O2‚2HBr: C,
(11) Brown, R. F.; Gulick, N. M. J . Am. Chem. Soc. 1955, 77, 1079-
1081.

8986 J . Org. Chem., Vol. 65, No. 26, 2000
Wei et al.
44.10; H, 4.93; N, 11.43. Found: C, 44.43; H, 5.20; N, 11.06.
UV (MeOH) for trans form: λmax 356 nm (ꢀ 26 600 dm3 mol-1
cm-1); for cis form: λmax319 nm (ꢀ 10 900), 445 nm (ꢀ 3600 dm3
mol-1 cm-1). δH (CD3OD) for trans form: 7.72 (d, J ) 8.91,
4H), 6.94; (d, J ) 8.91, 4H), 4.01 (t, J ) 6.30, 4H), 2.74 (t, J )
6.90, 4H), 1.85 (q, J ) 6.80, 4H); for cis form: 6.77 (d, J )
8.91, 4H), 6.75; (d, J ) 8.91, 4H), 3.89 (t, J ) 6.30, 4H), 2.70
(t, J ) 6.90, 4H), 2.00 (q, J ) 6.80, 4H). δC (CD3OD) for trans
form: 160.5, 146.2, 124.1, 115.1, 65.1, 36.2, 26.7; for cis form:
157.3, 146.8, 122.1, 114.6, 64.9, 36.2, 26.7.
4,4′-Di(γ-a m in obu ta n oxy)a zoben zen e d ih yd r obr om id e
4c was prepared in a manner analogous to 4a from 3c (0.312
g, 0.5 mmol) and 33% HBr/AcOH (1.0 g, 4.45 mmol) to give a
black solid of 4c, 0.218 g, 84% yield. TOF-MS: m/z 357.5 (M
+ H)+. Anal. Calcd for C20H28N4O2‚2HBr: C, 46.35; H, 5.45;
N, 10.81. Found: C, 46.34; H, 5.28; N, 11.02. δH (DMSO-d6):
7.85 (br, 6H), 7.83 (d, J ) 8.91, 4H), 7.10 (d, J ) 8.91, 4H),
4.08 (t, J ) 6.60, 4H), 2.87 (q, J ) 6.43, 4H), 1.90-1.74 (m,
4H). δC (DMSO-d6): 160.7, 146.1, 124.1, 115.0, 67.3, 40.7, 25.6,
23.8.
2:2 Cyclic P r od u cts 9a -c. When cyclization reactions
were conducted in the dark, 9a -c yellow solids were obtained.
Note that to completely convert the cis form to trans form,
before solutions of 4a -c were added, solutions were allowed
to stand in a darkroom for 24 h at 30 °C, or irradiated with
light (λ > 460 nm) for 5 min and then allowed to stand in a
darkroom at 30 °C for 8 h.
1:1 Cyclic P r od u cts 8a . 51% yield. Rf ) 0.47 (EtOAc-
MeCN 8:1). TOF-MS and FAB-MS: m/z 822 (M + H)+. Anal.
Calcd for C44H52N8O8‚2H2O: C, 61.67; H, 6.59; N, 13.08.
Found: C, 61.73; H, 6.50; N, 12.85. UV (MeOH) for trans form:
λmax 364 nm (ꢀ 22 700 dm3 mol-1 cm-1); for cis form: λmax315
nm (ꢀ 6800), 446 nm (ꢀ 2300 dm3 mol-1 cm-1). δH (CDCl3): 7.86
(d, J ) 8.25, 4H), 7.38-7.31 (m, 10H), 7.07 (br, 4H), 5.10 (s,
4H), 4.38 (br, 4H), 3.36 (br, 4H), 3.16 (br, 4H), 2.76 (br, 8H),
2.41 (br, 8H). δC (CDCl3): 170.3, 162.1, 156.0, 147.0, 136.3,
128.6, 128.2 128.0, 124.9, 122.5 (small), 118.0, 114.2 (small),
69.2, 67.3, 56.6, 51.4, 45.5 (br), 41.5.
1:1 Cyclic P r od u cts 8b. 52% yield. Rf ) 0.55 (EtOAc-
MeCN 8:1). FAB-MS and TOF-MS: m/z 850.0 (M + H)+. Anal.
Calcd for C46H56N8O8‚2H2O: C, 62.43; H, 6.83; N, 12.66.
Found: C, 62.44; H, 6.55; N, 12.41. δH (CDCl3) for trans form:
7.85 (d, J ) 8.25, 4H), 7.36-7.31 (m, 10H), 6.98 (br, 4H), 5.10
(s, 4H), 4.28 (br, 4H), 3.35 (br, 4H), 2.86 (br, 12H), 2.47 (br,
8H), 1.95 (br, 4H); for cis form: 7.38-7.20 (m, 10H), 6.81 (d,
J ) 8.91, 4H), 6.73 (d, J ) 8.91, 4H), 4.94 (s, 4H), 3.97 (br,
4H), 3.37 (br, 12H), 3.13 (br, 4H), 2.77 (br, 8H), 1.94 (br, 4H).
δC (CDCl3) for trans form: 170.5, 161.2, 156.1, 147.0, 136.5,
128.6, 128.4, 128.3, 128.1, 124.8, 116.8, 67.7, 67.4, 56.8, 52.0,
45.8 (br), 35.7, 28.1; for cis form: 170.5, 157.7, 156.5, 147.1,
136.0, 128.6, 128.4 128.3, 128.1, 122.4, 114.4, 67.4, 66.0, 58.6,
54.6, 47.4 (br), 36.1, 28.6.
1:1 Cyclic P r od u cts 8c. 55% yield. Rf ) 0.57 (EtOAc-
MeCN 8:1). TOF-MS and FAB-MS: m/z 878.0 (M + H)+. Anal.
Calcd for C48H60N8O8‚2H2O: C, 63.14; H, 7.06; N, 12.27.
Found: C, 63.41; H, 6.87; N, 12.51. δH (CDCl3): 7.79 (d, J )
8.25, 4H), 7.36-7.31 (br, 10H), 6.95 (br, 4H), 5.08 (s, 4H), 4.29
(br, 4H), 3.35 (br, 4H), 2.90 (br, 12H), 2.56 (br, 8H), 1.69 (br,
8H). δC (CDCl3): 171.5, 159.8, 156.7, 146.7, 136.6, 128.6, 128.3,
127.9, 124.3, 122.6 (small), 116.6, 114.3 (small), 67.5, 67.2,
56.8, 55.0, 47.2 (br), 38.6, 26.4, 25.5.
1,7-Dicar boxym eth yl-4,10-di(ben zyloxycar bon yl)-1,4,7,-
10-tetr a a za cyclod od eca n e 6. Compound 5 was prepared by
the method of Kovacs and Sherry.6 A solution of KOH (0.753
g, 11.4 mmol, 85% content) in MeOH (6 mL) was cooled to 0
°C and carefully added to a solution of bromoacetic acid (1.584
g, 11.4 mmol) in MeOH (10 mL) at 0 °C on an ice bath. The
temperature did not rise above 5 °C during addition. The
resulting solution was added to a suspension of 5 (2.201 g, 5
mmol) and anhydrous K2CO3 (1.575 g, 11.4 mmol) in MeOH
(50 mL). The mixture was stirred first at 45 °C for 7 h, and
then at 65 °C overnight. The solvent was removed to near
dryness under reduced pressure. The resulting white solid was
dissolved in water (50 mL) and acidified to pH 1 with 6 M
HCl. The supernatant liquid was decanted, and the residue
was washed with chilled water and dried under reduced
pressure to give 6, 2.464 g, 89% yield. FAB-MS: m/z 557.6 (M
+ H)+ and TOF-MS: m/z 558 (M + H)+. Anal. Calcd for
C
28H36N4O8‚H2O: C, 58.53; H, 6.67; N, 9.75. Found: C, 58.57;
H, 6.70; N, 9.68. δH (CDCl3): 7.33 (br, 10H), 5.21 (br m, 4H),
3.82 (br m, 8H), 3.23 (br, 12H). δC (CDCl3): 170.6, 155.7, 135.5,
128.2, 127.9, 67.2, 58.8, 57.0, 46.7.
1,7-Di(1′-N-acetyl-1,3-th iazolidin e-2-th ion e)-4,10-di(ben -
zyloxyca r bon yl)-1,4,7,10-tetr a a za cyclod od eca n e 7. To a
solution of 6 (0.557 g, 1 mmol) and 1,3-thiazolidine-2-thione,
TTH (0.262 g, 2.2 mmol) in CH2Cl2 (8 mL) cooled to 0 °C on
an ice-bath, was added 1-ethyl-3-(3-dimethylaminopropyl)-
carbodiomide hydrochloride, WSC (0.478 g, 2.5 mmol) and
4-(N,N′-dimethyl)aminopyridine, DMAP (0.020 g). The mixture
was stirred at 0 °C for 5 h and then at room temperature
overnight. The yellow solution was diluted with CH2Cl2 (20
mL), washed twice with water, and dried over anhydrous
MgSO4. The solvent was removed under reduce pressure. A
residue was dissolved in CHCl3 (5 mL) and a pale yellow
compound 7 was precipitated with addition of a large volume
of ethyl ether (150 mL), 0.426 g, 56% yield. TOF-MS: m/z 760
(M + H)+. δH (CDCl3): 7.34-7.30 (m, 10H), 5.11 (s, 4H), 4.48,
4.35 (br, 4H), 3.42 (br, 8H), 3.28 (br, 4H), 2.91 (br, 8H). δC
(CDCl3): 201.3, 173.1, 156.5, 136.9, 128.4, 127.9, 127.8, 66.9,
60.0, 55.6, 29.0, 54.3, 47.4.
Gen er a l P r oced u r es for 1:1 (8a -c) a n d 2:2 (9a -c)
Cyclic P r od u cts. To THF (350 mL), a solution of 7 (0.190 g,
0.25 mmol) in THF (75 mL), and of 4b (0.123 g, 0.25 mmol)
and Et3N (0.069 mL, 0.5 mmol) in 75 mL of THF-H2O (1:2)
were simultaneously added with ace dropping funnels at 2.0
mL/min under vigorous stirring at room temperature. After
addition, the solution was continuously stirred for 1 h. Solvents
were removed under reduced pressure; the remaining residue
was dissolved in CH2Cl2 (50 mL). The organic phase was
washed twice each subsequently with 1 M K2CO3, H2O, 1 M
HCl, and H2O and dried over Mg2SO4 overnight. Products were
isolated by preparative TLC on aluminum oxide.
2:2 Cyclic P r od u cts 9a . 47% yield. Rf ) 0.30 (EtOAc-
MeCN 8:1). TOF-MS and FAB-MS: m/z 1643.0 (M + H)+. Anal.
Calcd for C88H104N16O16‚4H2O: C, 61.67; H, 6.59; N, 13.08.
Found: C, 61.83; H, 6.45; N, 12.90. UV (MeOH) for trans,-
trans form: λmax 356 nm (ꢀ 51 700 dm3 mol-1 cm-1); for cis,cis
form: λmax317 nm (ꢀ 15 700), 447 nm (ꢀ 5200 dm3 mol-1 cm-1).
δH (CDCl3): 7.71 (d, J ) 8.89, 8H), 7.38-7.31 (m, 10H), 6.91
(d, J ) 8.38, 8H), 5.06 (s, 8H), 4.02 (br, 8H), 3.42 (br, 24H),
3.17 (br, 8H), 2.77 (br, 16H). δC (CDCl3): 170.3, 161.1, 156.0,
147.0, 136.3, 128.6, 128.3, 128.2, 124.4, 122.4 (small), 114.6,
114.5 (small), 67.5, 65.8, 56.7, 51.3, 45.4 (br), 38.6.
2:2 Cyclic P r od u cts 9b. 49% yield. Rf ) 0.32 (EtOAc-
MeCN 8:1). TOF-MS and FAB-MS: m/z 1669.0 (M + H)+. Anal.
Calcd for C92H112N16O16‚4H2O: C, 62.43; H, 6.83; N, 12.66.
Found: C, 62.50; H, 6.42; N, 12.30. δH (CDCl3) for trans,trans-
form: 7.73 (d, J ) 8.90, 8H), 7.36-7.31 (m, 10H), 6.80 (d, J )
8.58, 8H), 5.09 (s, 8H), 3.89 (br, 8H), 3.53, 3.41 (br, 24H), 3.19
(br, 8H), 2.79 (br, 16H), 1.96 (br, 8H); for cis,cis form: 7.30-
7.27 (m, 20H), 6.82 (d, J ) 8.91, 8H), 6.69 (d, J ) 8.91, 8H),
5.03 (s, 8H), 3.89 (br, 8H), 3.42, 3.33 (br d, 24H), 3.15 (br, 8H),
2.77 (br, 16H), 1.92 (br, 8H). δC (CDCl3) for trans,trans form:
170.9, 160.5, 156.9, 147.0, 136.1, 128.7, 128.6, 128.3, 128.1,
124.4, 114.5, 67.7, 65.6, 56.4, 54.8, 47.7 (br), 35.7, 28.1; for cis,-
cis form: 171.0, 157.6, 156.8, 146.8, 136.1, 128.7, 128.4, 128.1,
122.7, 114.5, 67.5, 65.8, 58.6, 55.5, 47.8 (br), 36.3, 29.1.
2:2 Cyclic P r od u cts 9c. 20% yield. Rf ) 0.33 (EtOAc-
MeCN 8:1). TOF-MS and FAB-MS: m/z 1755.0 (M + H)+. Anal.
Calcd for C96H120N16O16‚4H2O: C, 63.14; H, 7.06; N, 12.27.
Found: C, 63.48; H, 6.91; N, 12.60. δH (CDCl3): 7.73 (d, J )
8.91, 8H), 7.36-7.31 (m, 10H), 6.81 (d, J ) 8.91, 8H), 5.06 (s,
8H), 3.88 (br, 8H), 3.50, 3.35 (br, 24H), 3.17 (br, 8H), 2.87 (br,
16H), 1.75 (m, 16H). δC (CDCl3): 171, 160.3, 156.9, 146.8,
136.3, 128.6, 128.3, 127.9, 124.1, 122.8 (small), 114.4, 67.1,
65.6, 56.8, 55.0, 47.0 (br), 38.5, 26.4, 25.4.
1:1 Cyclic P r od u cts 8a -c. When cyclization reactions
were conducted under UV irradiation with a 500-W super-
high-pressure Hg lamp through a color glass filter (330 nm <
λ < 380 nm). 8a -c yellow solids were obtained.

Synthesis of 1:1 and 2:2 Cyclic Compounds
J . Org. Chem., Vol. 65, No. 26, 2000 8987
4,4′-Di(γ-a m in op r op a n oxy)a zoben zen e 10. To a solution
of compound 4b (0.245 g, 0.5 mmol) dissolved in H2O-MeOH
(25 mL, 10:1), toluene (50 mL) and Et3N (0.7 mL, 5 mmol) were
added. The mixture was vigorously shaken in a separatory
funnel. The water phase was extracted with toluene. The
combined toluene extracts were washed with water and dried
over granular CaO for 2 days. Toluene was removed under
the protection of nitrogen gas to give a yellow solid of 10. 0.117
g, 71% yield. 10 was stored in desiccator-filled nitrogen gas.
δH (CDCl3): 7.86 (d, J ) 8.91, 4H), 6.89 (d, J ) 8.91, 4H), 4.01
(t, J ) 6.30, 4H), 2.93 (t, J ) 6.90, 4H), 1.98 (q, J ) 6.80, 4H).
δC (CDCl3): 160.6, 146.4, 124.2, 114.4, 65.2, 36.0, 26.6.
Ack n ow led gm en t. We thank Drs. Kazuhiko Naka-
mura and Tadashi Nemoto for their invaluable advice
on NOE and NOESY NMR. This work was financed in
part by the J apan Science and Technology Corporation
(J ST).
Su p p or tin g In for m a tion Ava ila ble: 1H NMR spectra of
9c including all-cis (cis,cis) to all-trans (trans,trans), and cis,-
trans isomers. This material is available free of charge via the
Internet at http://pubs.acs.org.
J O000926P