Synthesis, crystal structure, spectroscopic properties and potential anti-cancerous activities of four unsaturated bis-norcantharimides

Article abstract of DOI:10.1016/j.molstruc.2016.02.093

Four unsaturated norcantharimide (UNCI) dimers were synthesized and characterized by elemental analysis, ESI-QTOF-MS, FT/IR, UV-Vis, ¹H and ¹³C NMR as well as single crystal X-ray diffraction. In addition, theoretical studies have been investigated to compare with the experimental findings. Introduction of various lengths of single bond link chains provides high conformational flexibility and thus unusual molecular and crystal structures for dimers. Two of the four dimers twist into helicate, but crystallize into centrosymmetric lattice; one adopts approximately centrosymmetric conformer, but packs into non-centrosymmetric polar space group (P2₁). Moreover, in vitro cytotoxic activities of four UNCI dimers and their corresponding saturated NCI dimers were evaluated. All four UNCI dimers are inactive and one NCI dimer shows modest cytotoxicity. These findings were compared with the relevant results in literature. It is found that the antitumor properties of UNCI/NCI dimers depend mainly on the length of link chains (the longer chain, the higher therapeutic efficacy) and have relationship with the double bond, which requires more experimental support.

Full text of DOI:10.1016/j.molstruc.2016.02.093

Accepted Manuscript
Synthesis, crystal structure, spectroscopic properties and potential anti-cancerous
activities of four unsaturated bis-norcantharimides
Shuang-Shuang Cheng, Yan Shi, Xiao-Na Ma, Dian-Xiang Xing, Lian-Dong Liu, Yun
Liu, Yun-Xue Zhao, Qi-Cheng Sui, Xue-Jie Tan
PII:
S0022-2860(16)30187-9
10.1016/j.molstruc.2016.02.093
MOLSTR 22302
DOI:
Reference:
To appear in: Journal of Molecular Structure
Received Date: 2 February 2016
Revised Date: 25 February 2016
Accepted Date: 25 February 2016
Please cite this article as: S.-S. Cheng, Y. Shi, X.-N. Ma, D.-X. Xing, L.-D. Liu, Y. Liu, Y.-X. Zhao, Q.-
C. Sui, X.-J. Tan, Synthesis, crystal structure, spectroscopic properties and potential anti-cancerous
activities of four unsaturated bis-norcantharimides, Journal of Molecular Structure (2016), doi: 10.1016/
j.molstruc.2016.02.093.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Synthesis, crystal structure, spectroscopic properties and potential anti-
cancerous activities of four unsaturated bis-norcantharimides
Shuang-Shuang Cheng a, Yan Shi a, Xiao-Na Ma a, Dian-Xiang Xing a, Lian-Dong Liu b, Yun Liu a, Yun-
Xue Zhao c, Qi-Cheng Sui a, Xue-Jie Tan a
a School of Chemistry and Pharmaceutical Engineering, Qilu University of Technology, Jinan, Shandong Province, 250353, P. R. China
b College of Chemistry, Chemical Engineering and Materials Science, Shandong Normal University, Jinan, Shandong Province, 250014, P. R. China
c Department of Pharmacology, School of Medicine , Shandong University, Jinan, Shandong Province, 250012, P. R. China
ARTICLE INFO
ABSTRACT
Article history:
Received
Received in revised form
Accepted
Four unsaturated norcantharimide (UNCI) dimers were synthesized and characterized by
elemental analysis, ESI-QTOF-MS, FT/IR, UV–Vis, 1H and 13C NMR as well as single crystal
X-ray diffraction. In addition, theoretical studies have been investigated to compare with the
experimental findings. Introduction of various lengths of single bond link chains provides high
conformational flexibility and thus unusual molecular and crystal structures for dimers. Two of
the four dimers twist into helicate, but crystallize into centrosymmetric lattice; one adopts
approximately centrosymmetric conformer, but packs into non-centrosymmetric polar space
group (P21). Moreover, in vitro cytotoxic activities of four UNCI dimers and their
corresponding saturated NCI dimers were evaluated. All four UNCI dimers are inactive and
one NCI dimer shows modest cytotoxicity. These findings were compared with the relevant
results in literatures. It is found that the antitumor properties of UNCI/NCI dimers depend
mainly on the length of link chains (the longer chain, the higher therapeutic efficacy) and have
relationship with the double bond, which requires more experimental support.
Available online
Keywords:
Unsaturated norcantharimide dimers
Synthesis
Characterization
Single crystal structure
Cytotoxic activity
1. Introduction
collagen, and platelet-activating factor-induced aggregation [8].
So, derivatives of modified CAI/NCI are potentially useful as
anticancer agents.
As archetypal small molecule protein phosphatase inhibitors,
[1] cantharidin (CAN) and norcantharidin (a demethylated form
of cantharidin, also called demethylcantharidin, so abbreviated as
DMC, Scheme 1) have been used worldwide as an anticancer
agent since 1264 for the treatment of hepatoma, leukemia,
pancreatic cancer, colon cancer, oral carcinoma, bladder cancer,
breast cancer, lung cancer and digestive tract tumors [2]. Their
ability to act against multidrug-resistant cells makes it an ideal
compound for individualized cancer treatment [3]. Similarly,
CAN possesses cytotoxicity to a series of normal cells, including
gastrointestinal tract, urethra and kidney [2d], which delayed
their use in the pharmaceutical industry. However, the organic
chemistry has provided new and more potent derivatives with
high activity against protein phosphatase enzyme and less
toxicity profiles.
Scheme 1. Some important core structures of cantharidine analogues
During the last five decades, thousands of analogues and
derivatives have been synthesized and thoroughly investigated
[2e], including DMC-platinum complexes [2b, 2c, 4],
(nor)cantharimide series (abbreviated as CAI/NCI, Scheme 1)
[2m, 5], anhydride ring-opened series (especially those with only
one free carboxylate) [1g, 6] and so on. These analogues have
demonstrated all kinds of antitumor activities and each has its
own specific activity. Due to the possibility to incorporate any
kinds of substituent in the nitrogen, as well as essentially
equipotent inhibitory activity of the serine/threonine protein
phosphatases 1 and 2A (PP1 and PP2A) with CAN (more potent
than DMC) [7], CAI/NCI series show higher anticancer
activities, and have been shown to inhibit xanthenes oxidase and
As we know, the type of heteroatoms in the bridge and in
the anhydride cycle are very important, but, the presence of
double bond (5,6-ene) has little effect on activity [5d, 7, 9]. 5,6-
dehydrocantharidin (the unsaturated analogue of CAN,
abbreviated as UCAN) and CAN have similar inhibition of PP2A
[2e]. 5,6-dehydronorcantharidin (the unsaturated analogue of
DMC, abbreviated as UDMC) and DMC still have similar
inhibition of PP1, PP2A and PP2B [1f]. This suggests that
UCAN and UDMC show similar anti-cancer and protein
phosphatase activity as that of CAN and DMC (others think that
the saturation of C5-C6 bond appears to affect the inhibitory
activity, but not so crucial [5d]). More importantly, UCAN and
UDMC are so easily synthesizable that they can often be used as
to have antiplatelet effects on thrombin, arachidonic acid,
———
Corresponding author. Tel.: +86 531 89631208; fax: +86 531 89631207; E-mail addresses: tanxuejie@163.com.

2
the starting material in the synthesis of CAN and DMC [10].
ACCEPTED MANUSCRIPT
Then, what’s the difference between 5,6-dehydronorcantharimide
(the unsaturated analogue of NCI, abbreviated as UNCI) and
NCI? Surprisingly, the derivatives based on UNCI have hardly
been explored in the literature, much less than that of UCAN and
UDMC analogues. Wang et al. found that the arylantimony
derivatives based on NCI and UNCI have similar in vitro
antitumor activities [11]. For example, the complexes I6, I7, and
II7, II8 in their paper have very high and similar antitumor
activities against some cancer cells (Scheme 2). Li et al. [12]
have investigated the antiproliferative activities of ten UNCI and
NCI derivatives (Scheme 2, 5a-5f, 7a-7d), which displayed
moderate and similar inhibitory activities against A549 and PC-3
cell lines with the IC50 values >250.0 ꢀmol/l (the IC50 of DMC
were 44.8 and 201.0 ꢀmol/l in their experiments). In one word,
since UNCI and its analogues not only have simple chemical
structures and less toxicity, but also retain the antitumor
activities, is it possible for them to provide enormous possibilities
for the science and industry of antitumor medicine? These
uncertainties encourage us for further investigations related to
these issues. We hope that these chemical modifications in the
structure of UNCI could be a real and rapid way in developing
new drug candidates.
Scheme 3. The structures of some (nor)cantharimide (NCI/CAI)
dimers reported in literature [15, 16, 17].
Another blank area is the crystal structure of the dimers.
There have been no reports about the crystal structure of any
NCI/CAI dimers. It is known that most small molecule drugs
(>90%) are delivered in crystalline form [18] and at least half of
marketed solid chemical drug substances exhibit polymorphism
[19]. Meanwhile, medicinal chemistry requires robust reliable
structures to accurately position key pharmacophoric units in the
correct chemical space. It is this positioning that gives rise
ultimately to the desired biological activity. To obtain a better
understanding of the solid-state properties of these substances, it
is necessary to identify and characterize crystal structures, and
even to perform a polymorphic screening and physiochemical
properties characterization, on potential drug candidates. This
will help in eventually selecting a suitable form for further
development and manufacturing. As part of our ongoing project
studying novel unsaturated analogues of DMC/NCI, we have
recently obtained systematical studies on a novel silver and
singly protonated UDMC complex [20]. We now present our
studies on the synthesis and thorough spectral and X-ray
crystallographic characterization of four UNCI dimers (shown in
Scheme 4). UDMC-DETA and UDMC-TETA are new; the rest
two dimers have appeared before [21], but neither detailed
structure information nor any properties have been reported. It is
a substantial challenge to identify all these dimers’ single-crystal
structures because they are usually isolated as thin powders,
which showed no tendency to crystallize. At last, all dimers are
evaluated for their in vitro cytotoxic activity against two cancer
cell lines, A549 (human lung cancer) and 4T1 (mouse breast
cancer). Some interesting structure-activity relationships were
observed.
Scheme 2 Two sets of similar compounds with similar antitumor
activities reported in reference [11] and [12]. Their name
abbreviations are maintained as in the original report.
On the other hand, the dimer structure is ubiquitous in
natural products and dimeric molecules would be expected to
show enhanced receptor affinity relative to their corresponding
monomeric counterparts [13]. Dimeric compounds have been
synthesized and studied for the treatment of cancer, HIV,
Alzheimer, malaria and various parasitic diseases [14].
McCluskey et al. [15] reported the synthesis and anticancer
activities of two NCI dimers (Scheme 3, compounds 1 and 2),
which displayed the highest levels of cytotoxicity against a series
of cell lines among 35 NCI derivatives that they synthesized. In
addition, Noda et al [16] isolated three CAI dimers (Scheme 3,
compounds 3 to 5) from the Chinese blister beetle, Mylabris
phalerate PALLAS (Meloidae). Their structures were determined
based on spectroscopic and chemical evidence. But their
cytotoxic activities were not demonstrated and have not been
reported up to the present. The fourth example of CAI dimer is
The abbreviations and corresponding systematic names for
four UNCI dimers formed between UDMC and butanediamine
(BDA), 1.6-hexamethylendiamine (HDA), diethylenetriamine
(DETA), triethylenetetramine (TETA) are as follows: UDMC-
BDA, 2,2'-(1,4-butanediyl)bis(3a,4,7,7a-tetrahydro-4,7-epoxy-
1,3-bishydroisoindole-1,3-dione);
UDMC-HDA,
2,2'-(1,6-
hexanediyl)bis(3a,4,7,7a-hexahydro-4,7-epoxy-1,3-
bishydroisoindole-1,3-dione); UDMC-DETA, 2,2'-(3-aza-1,5-
pentanediyl)bis(3a,4,7,7a-te-trahydro-4,7-epoxy-1,3-
bishydroisoindole-1,3-dione); UDMC-TETA, 2,2'-(3,6-diaza-
1,4-butanediyl)bis(3a,4,7,7a-te-trahydro-4,7-epoxy-1,3-
bishydroisoindole-1,3-dione).
(3aR,3'aR,4S,4'S,7R,7'R,7aS,7'aS)-rel-
2,2'-[dithiobis(1,3,4-
thiadiazole-5,2-diyl)]bis[hexahydro-3a,7a-dimethyl- 4,7-epoxy-
1H-isoindole-1,3(2H)-dione (Scheme 3, compound 6) reported
by Kok et al [17]. The compound showed cytotoxic potential on
the entire four cancer cell lines examined. The cytotoxic pattern
of the dimer on carcinoma cell lines was similar to that of similar
single state. The major difference between them was observed in
KG1a, where the dimer was still effective at 12.5 ꢀg/ml but a
higher concentration was required for that of similar single state.
Compared with these NCI and CAI dimers, we are unaware,
however, of any studies about the detailed structure and
potential biological activities based on UNCI dimmers.
2 Experimental Section
2.1 Materials, synthesis and measurements
Chemicals, cell culture reagents and media were purchased
from Aladdin-reagent Chemicals and were used without further
purification.
Synthesis of UDMC follows methods in the literature [22].
Condensation of diamines with UDMC in anhydrous toluene or
acetone gave the crude product UNCI dimers in around 20%

yield (Scheme 4). Though the dehydrative condensation of DMC
1769(s, ν C=O), 1705(vs, ν C=O), 1406(s, ν C-N), 1169(s, ν C-
ACCEPTED MANUSCRIPT
with primary amines is the main methodology for the synthesis of
NCI derivatives, no reports were found to synthesize NCI/UNCI
dimers with this method. It should be noted that the temperature
is a key factor in determining the end product. The yield will
drastically decrease with increasing temperature and the major
products will be various decomposition products and monomeric
derivatives. The kinetic features associated with various products
depend on the specific structure of the diamines, the reactant
concentration, the medium and the temperature. These researches
are ongoing and will be reported elsewhere.
O-C);
UV/Vis
(CH3CN)
λmax/nm
(ε/L·mol-1·cm-1):
214.0(3.5×105).
UDMC-HDA Elemental analysis: found (calc. for
C22H24N2O6): C, 64.15 (64.07%); H, 5.91 (5.87%); N, 6.83
(6.79%); HRMS (ESI): m/z calcd for C22H24N2O6+H+: 413.1713
[M+H+]; found: 413.1711; M.p.158.8-159.4℃,.1H NMR
(DMSO):δ (ppm) 6.533(s, 4H, olefinic protons), 5.106(s, 4H,
methine protons linked to bridge O, O-CH-), 3.297(s,4H,
methylene protons linked to imide N, N-CH2-), 2.894(s,4H,
methine protons, -CH-), 1.384(t, J=6.4Hz,4H, methylene protons,
-CH2-), 1.225(s,4H, methylene protons, -CH2-). 13C NMR
(DMSO): δ (ppm) 176.413(carbonyl carbons), 136.397(olefinic
carbons), 80.283(methine carbons linked to bridge O, O-CH-),
47.034(methine carbons, -CH-), 37.700(methylene carbons
linked to imide N, N-CH2-), 26.798(methylene carbons, -CH2-),
25.368(methylene carbons, -CH2-). FT-TR(cm-1,KBr): 3069(w, ν
C=C-H), 2995(m, ν C-H), 2930(m, ν C-H), 2859(w, ν C-H),
1771(s, ν C=O), 1697(vs, ν C=O), 1412(s, ν C-N), 1167(s, ν C-
The crude product was filtered, washed and then
recrystallized with various solvents (see Table S1 in the
Supporting Information). All crystals are colorless with
approximately long and thin bar shapes as shown in Fig. S1 (in
Supporting Information. All “S” numbered tables and figures are
in Supporting Information). They are soluble in organic solvents
such as methanol, ethanol, chloroform, acetonitrile, dimethyl
formamide and dimethyl sulfoxide, but not soluble in toluene,
acetone, ether, hexane and petroleum ether. The general reactions
are shown as follows:
O-C);
UV/Vis
(CH3CN)
λmax/nm
(ε/L·mol-1·cm-1):
207.0(2.2×105).
UDMC-DETA Elemental analysis: found (calc. for
C20H21N3O6): C, 60.22 (60.14%); H, 5.41 (5.30%); N, 10.63
(10.52%); HRMS (ESI): m/z calcd for C20H21N3O6+H+: 400.1509
[M+H+]; found: 400.1513; M.p.126.0-127.2℃. 1H NMR
(CDCl3): δ (ppm) 6.510(s, 4H, olefinic protons), 5.270(s, 4H,
methine protons linked to bridge O, O-CH-), 3.556(t, J=6.0Hz,
4H, methylene protons linked to imide N, N-CH2-), 2.859(s, 4H,
methine protons, -CH-), 2.794(t, J=6.0Hz, 4H, methylene protons
linked to secondary amino group, NH-CH2), 2.097(s,1H, -NH-).
13C NMR (CDCl3):
δ (ppm) 175.842(carbonyl carbons),
Scheme 4 The reaction sequence of four UNCI dimmers.
135.995(olefinic carbons), 80.323(methine carbons linked to
bridge O, O-CH-), 46.943(methine carbons, -CH-),
45.673(methylene carbons linked to imide N, N-CH2-),
38.175(methylene carbons linked to secondary amino group, NH-
CH2-). FT-TR(cm-1,KBr): 3333(s, ν N-H), 3084(w, ν C=C-H),
3014(w, ν C=C-H), 2953(m, ν C-H), 2812(m, ν C-H), 1767(s, ν
C=O), 1711(vs, ν C=O), 1402(vs, ν C-N), 1204(s, ν C-O-C);
UV/Vis (CH3OH) λmax/nm (ε/L·mol-1·cm-1): 211.0(2.2×105).
Elemental (C, H, N) analyses were carried out with a
Perkin–Elmer 2400 microanalyzer. Accurate-mass measurements
were acquired on an Agilent-6520 quadrupole-time of flight
tandem mass spectrometer. 1H and 13C NMR spectra were run on
a Bruker Avance 400 MHz instruments. The chemical shifts are
reported in parts per million (ppm) relative to tetramethylsilane,
SiMe4 (δ = 0 ppm), referenced to the chemical shifts of residual
solvent peak [deuterated dimethyl sulfoxide (DMSO-d6)]. UV–
Vis absorption spectra were recorded using a UV-1700
spectrophotometer (Shimadzu, Japan), in 1×10-5mol.L-1
acetonitrile or methanol solution. Infrared (IR) spectra were
obtained as KBr pellets with a Bruker tensor 27 FT–IR
spectrometer (Bruker, Germany). Melting points were
determined on a WRS-2A electrothermal digital melting point
apparatus (Shanghai precision & scientific instrument Co., Ltd,
China).
UDMC-TETA Elemental analysis: found (calc. for
C22H26N4O6): C, 59.84 (59.72%); H, 5.98 (5.92%); N, 12.73
(12.66%); HRMS (ESI): m/z calcd for C22H26N4O6+H+: 443.1931
[M+H+]; found: 443.1939; M.p.146.9-147.1℃. 1H NMR
(DMSO): δ (ppm) 6.535(s, 4H, olefinic protons), 5.107(s, 4H,
methine protons linked to bridge O, O-CH-), 3.379(t, J=6.8Hz,
4H, methylene protons linked to imide N, N-CH2-), 2.905(s, 4H,
methine protons, -CH-), 2.559(t, J=6.8Hz,4H, methylene protons
linked to secondary amino group, NH-CH2), 2.480(s,4H,
methylene protons linked to secondary amino group, NH-CH2),
The physico-chemical characterization results are listed
below:
1.625(s, 2H, -NH-). 13C NMR(DMSO):
176.419(carbonyl carbons), 136.391(olefinic
δ
(ppm)
carbons),
UDMC-BDA Elemental analysis: found (calc. for
C20H20N2O6): C, 62.57 (62.49%); H, 5.31 (5.24%); N, 7.33
(7.29%); HRMS (ESI): m/z calcd for C20H20N2O6+H+: 385.1400
[M+H+]; found: 385.1404; M.p.165.0-166.0℃. 1H NMR
(CDCl3): δ (ppm) 6.505(s, 4H, olefinic protons), 5.259(s, 4H,
methine protons linked to bridge O, O-CH-), 3.484(s,4H,
methylene protons linked to imide N, N-CH2-), 2.834(s,4H,
methine protons, -CH-), 1.549(s, 4H, methylene protons, -CH2-).
80.291(methine carbons linked to bridge O, O-CH-), 48.282
(methylene carbons linked to imide N, N-CH2-), 47.062(methine
carbons, -CH-), 46.062(methylene carbons linked to secondary
amino group, NH-CH2-), 37.942(methylene carbons linked to
secondary amino group, NH-CH2-). FT-TR(cm-1,KBr): 3337(s, ν
N-H), 3063(m, ν C=C-H), 3005(m, ν C=C-H), 2940(m, ν C-H),
2903(m, ν C-H), 2845(m, ν C-H), 1767(s, ν C=O), 1705(vs, ν
C=O), 1404(s, ν C-N), 1177(s, ν C-O-C); UV/Vis (CH3OH)
λmax/nm (ε/L·mol-1·cm-1): 212.0(2.6×105).
13C NMR (CDCl3):
δ (ppm) 175.687(carbonyl carbons),
136.020(olefinic carbons), 80.385(methine carbons linked to
bridge O, O-CH-), 46.898(methine carbons, -CH-),
37.665(methylene carbons linked to imide N, N-CH2-),
24.078(methylene carbons, -CH2-). FT-TR(cm-1,KBr): 3076(m, ν
C=C-H), 3011(m, ν C=C-H), 2963(m, ν C-H), 2930(m, ν C-H),
2.2 X-Ray Crystallographic Analysis
The X-ray diffraction measurements were made on a
Bruker APEX II CCD area detector diffractometer at 293K

4
1
for all UNCI dimers (Mo Ka radiation, graphite
UDMC are listed in Figs. S2 and S3. The H and 13C NMR
ACCEPTED MANUSCRIPT
monochromator, λ = 0.71073 Å). The structures were solved
by SHELXL-97. The absorption correction was done using
the SADABS program [23]. Software packages APEX II
(data collection), SAINT (cell refinement and data
reduction), SHELXTL (data reduction, molecular graphics
and publication material), DIAMOND (simplifying crystal
packing diagram) were also used [24-26]. All non-hydrogen
atoms were refined with anisotropic displacement
parameters, and hydrogen atoms were placed in idealized
positions and refined as rigid atoms with the relative
isotropic displacement parameters.
spectra of four UNCI dimers are listed in Figs. S4 - S11.
1
In the H NMR spectra of UDMC, the olefinic and methine
protons all appear as singlet at δ = 6.567, 5.338 and 3.364 ppm,
respectively (Fig. S2). When changing into UNCI dimers, they
still appear as singlet, but all shift upfield (Table S2). For
example, the olefinic protons of UDMC-BDA to UDMC-TETA
are observed at 6.505, 6.533, 6.510 and 6.535 ppm, respectively.
This can be attributed to the increasing shielding effect caused by
the decreasing electronegativity from O to N, which makes
electron density shift from imide N towards olefinic bond. In
addition, protons that are nearer to the substituted N atom are
shielded more. Similar shielding effect is observed in the 13C
NMR spectra, mainly supported by the significant shifts towards
lower values of ppm (Table S2). For example, the two kinds of
methine carbons are shielded by about 1.2 (the mean value of -
1.304, -1.168, -1.259 and -1.14) and 1.4 (the mean value of -
1.322, -1.424, -1.384 and -1.416) ppm respectively, relative to
that in UDMC. The olefinic carbons are slightly shielded by only
0.3 ppm (the mean value of -0.455, -0.078, -0.48 and -0.084),
because they are distant (at least 4.5 Å) from the N atom. As
expected, the carbonyl carbon atoms should be greatly shielded
due to their proximity to the N substitution (∼ 1.38 Å).
Surprisingly, relative to 169.3 ppm for UDMC, 13C resonance
frequencies shift towards higher values of 175.7, 176.4, 175.8
and 176.4 ppm in the dimer derivatives (from UDMC-BDA to
UDMC-TETA) (Table S2). This 6.8 ppm (the mean value of
6.359, 7.085, 6.514 and 7.091) downfield shift reflects significant
deshielding of the N substitution. Does N atom behave as
shielding or deshielding towards the C atoms? This seems
counterintuitive. It has been reported that intermolecular π-π
interactions are responsible for this kind of deshielding as well as
shielding of protons [34]. However, it seems that this supposition
could not be used to explain the deshielding and shielding of
carbon in UNCI dimeric (Table S3) and monomeric [22c, 35]
derivatives. In addition, an attempt was made to check
shielding/deshielding effect by calculating natural population
analysis (NPA) and Mulliken [36] charges, which are shown in
Table S4. However, calculated results show that N atom
produces strong shielding instead of deshielding effect on
carbonyl C. Of course, we are aware that atomic charges are not
an observable quantum mechanical and these calculated values
should be treated with a great deal of caution. After all, they are
somewhat associated with the trend for electric field and electron
population distributions. At last, we examined the crystal
structure of UDMC and UNCI derivatives carefully and the
coplanarity of O=C-N(O)-C=O group was analyzed
systematically (Table S5). It can be seen that the two carbonyl
groups are essentially coplanar with N/O atom and this
coplanarity favors the existence of p-π conjugation. But the
degree of coplanarity is decreased when O is changed into N. For
example, the mean deviation is only 0.0034 Å for UDMC, but
the values come to 0.0122-0.0387 Å after it is changed into
UNCI derivatives. The deviation from coplanarity resulting from
the N substitution means a reduction in the extent of p-π
conjugation, which could be expected to level out the π-electron
density, the signal for the carbonyl C atoms with π-electron-
density deficit shifting downfield. This supposition was
confirmed by similar reports in literatures [22c, 35], which are in
line with their downfield shifting. In fact, many other instances of
C=O group deshielding with breakdown of coplanarity and a
reduction of conjugation effects are met in other compounds,
such as the acyclic and alicyclic α, β-unsaturated ketones and
aldehydes [37], unsaturated conjugated ω-amino-1,3-diketones
and diesters [38]. Such an effect has never been explained before.
2.3 Computational Study
In this work, theoretical calculations were mainly used to
investigate spectral properties of all UNCI dimers.
In preliminary optimizations, the geometry of all dimers
were first extracted from their single-crystal X-ray structures and
then optimized by employing density functional theory (DFT)
method with the B3LYP exchange correlation functional
calculations [27, 28]. Frequency analyses have been made at the
same level to ascertain the nature of the optimized structures to
be the real minimum. Vertical electronic excitations based on
B3LYP/6-31+G(d) optimized geometries were computed using
the time-dependent density functional theory (TD-DFT)
formalism [29] at the B3LYP/6-31+G(d) level. UV-Vis spectra
as well as the assignment of vibrational modes (IR) were done on
the basis of the GaussView 5.0 package [30]. The major
contributions of the transitions were designated with the aid of
SWizard program [31] using the Gaussian distribution model
with the half-bandwidth of 500 cm-1 on the basis of TD-DFT
results. DFT and TD-DFT calculations were carried out using the
Gaussian03 program package [32] on a Sunway BlueLight
MPP supercomputer housed at the National Supercomputer
Center in Jinan, China.
2.4 Cytotoxicity Assays
Cell Culture. Human A549 (lung carcinoma) cells and
mouse breast cancer 4T1 cells were purchased from Shanghai
Cell Bank, Type Culture Collection Committee, Chinese
Academy of Sciences. The cells were cultured in F12K medium
supplemented with 10% heat inactivated fetal bovine serum
(FBS), 2 mM glutamine, 100 U/mL penicillin, and 100 ꢀg/mL
streptomycin and maintained at 37 °C in humidified atmosphere
of 5% CO2.
MTT assay for cell proliferation. MTT assay was
conducted and modified as described in the literature [33]. The
cells (3000 cells) were seeded on 96-well microtitre plates in
F12K medium with 10% FBS and incubated overnight. The cell
culture medium was replaced by the different dose of compounds
solution, and then the cells were cultured for another 72 h. The
MTT
[3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium
bromide] reagent was added to the cell supernatant for a final
concentration of 0.5 mg/mL of MTT. After 3h the cell culture
medium was removed. Formazan crystals in adherent cells were
dissolved in 200 ꢀL DMSO and the absorbance of the formazan
solution was measured. Each compound was tested in triplicate
and the experiments were repeated three times.
3 Results and Discussion
3.1 NMR characterization
In order to make the comparisons of NMR spectra between
1
UDMC and UNCI dimers, the H and 13C NMR spectra of

Briefly summarizing above, such a “remote-shielding vis
absorption bands undergo significant changes which can be a
direct identification of this chemical reaction. The starting
material (UDMC) shows two strong peaks at 1217 and 1088 cm-1
(Fig. S12), which arises from the C-O-C vibration in cyclic
anhydride, the former is due to symmetric vibration and the latter
is due to antisymmetric vibration [39]. But the corresponding
frequencies in products (four UNCI dimers) shift to higher
wavenumbers, i.e. 1412 - 1398 cm-1 for C-N-C symmetric
vibrations and 1203 - 1165 cm-1 for C-N-C anti-symmetric
vibrations. Both of them still have strong intensities. Krikorian et
al and other groups [45 and the references therein] have found
two similar spectral regions in simple cyclic imides, one at 1300-
1400 cm-1 and another at 1030-1250 cm-1. They believe that these
two bands arise from mixed vibrationally coupled modes, viz. in-
plane δ(N-H) and v(C-N-C), with the symmetric C-N-C stretch
contributing to the higher frequency and the antisymmetric C-N-
C stretch to the lower. There are no N-H groups in the case of
UDMC-BDA and UDMC-HDA, and the N-H bonds in UDMC-
DETA and UDMC-TETA are far from the cyclic imide N
atoms. So the two bands contain no contribution from the N-H
vibrational component. Similarly, Grzetic and Oomens [46] have
observed two strong bands at 1331 and 1235 cm-1 for glutarimide
ring, which were assigned to the C-N-C symmetric and
asymmetric stretching, excluding the N-H vibrational component.
The computed spectra for the four imide dimers are in good
accordance with the corresponding experimental spectra between
1000 and 1400 cm–1. But the scale factor should be changed into
0.9946. In one word, the C-N-C stretching modes are essentially
located in the range of 1412 - 1398 cm-1 for symmetric vibrations
and 1203 - 1165 cm-1 for anti-symmetric vibrations. Both are
very characteristic and most diagnostic for the assignment of N-
substituted UNCI derivatives, since all four UNCI dimers
present the same bands.
ACCEPTED MANUSCRIPT
near-deshielding” contradiction upon N substitution is believed
to be the result of the decrease of coplanarity in the p-π electron
conjugated system.
The secondary amino proton of UDMC-DETA resonated at
2.097 ppm and appeared as a broad singlet (Fig. S6). The
integration of which indicates three protons instead of one,
probably due to the water molecule in the crystal.
3.2 IR characterization
Based on the understanding of the relationship between
experimentally observed frequencies and those DFT/B3LYP
calculated results for the four strongest peaks (happen to be the
most characteristic bands) in the IR spectrum of UDMC, we
made assignments and analysis for the bands in the spectra of
imide dimers.
Symmetric and antisymmetric stretches of C=O bond occur
at 1859 and 1786 cm-1 in UDMC [39] (Fig. S12), but when
UDMC is converted into imide dimers, no peak can be found in
the region of 1860-1780 cm-1, which are typical for saturated
anhydrides in a 5-membered ring [40]. In their corresponding
imide dimers, one band is located between 1771 and 1767 cm-1
(symmetric) and a more intense band occurs between 1721 and
1697 cm-1 (antisymmetric) (Figs. S13 – S16). The apparent shift
of C=O vibration indicates the conversions of original anhydride
into imide. Similar vibrational frequency ranges are reported in
other references [41]. All the calculations for these imide dimers
predict this shift (Table S6). For example, according to the
values calculated with the DFT/B3LYP method, symmetric and
antisymmetric stretches of C=O bond occur at 1918 and 1849 cm-
1
in UDMC, but the corresponding frequencies in imide dimers
are lowered to 1871 – 1831 cm-1 for symmetric vibrations and
1834 - 1767 cm-1 for antisymmetric vibrations. If proper value of
the scale factor can be set (0.9631 in this work) [42], both the
trend and the values of the shift are in good agreement with
experimental results. The account of the shift reasonably lies in
the 5-membered ring strain [39]. For a cyclic anhydride, the
resonance within the -CO-O-CO- system causes it to be coplanar
with the two carbonyls on the opposite side of the cyclic system.
Similarly, the -CO-N-CO- system maintains the coplanar
resonance in the cyclic imide. But the coplanarity is not as good
as that in cyclic anhydride, which can be proved by their crystal
structure (Table S5). Poor coplanarity and nitrogen substitution
will decrease ring strain, thus shifting the two bands to lower
frequencies. As for the intensity differences between the two
bands (the symmetric stretching band is very weak relative to the
antisymmetric stretching band), Yang and co-workers [39, 43]
explained that the change in dipole moment of a 5-membered
cyclic anhydride system for a symmetric stretching mode is small
whereas that for an antisymmetric stretching mode is large. We
think it can also be applied to the 5-membered cyclic imide
system. Meanwhile, according to the DFT calculated results, the
symmetric stretching band is indeed very weak relative to the
antisymmetric stretching band, in line with the experimental
results.
In the FTIR spectrum of UDMC-TETA (Fig. S16), an
intense band of N-H stretching vibration at 3337 cm-1 is observed.
The computed value is 3499.03 cm-1, poorly reproduced by the
calculations. The sharp, smooth shape of this band proves no
interactions with any proton acceptors. But in the FTIR spectrum
of UDMC-DETA (Fig. S15), the formation of an intramolecular
hydrogen bond N3–H3B···O4 (Table S11, imine hydrogen
bonded to the bended oxygen bridge) increase the distance
between N and H, weakening the N-H bond and thus decreasing
the vibrational frequency (red shift to 3333 cm-1). Similarly, the
computed value 3515.02 cm-1 does not agree with the
experimental result. Despite the mismatch between experimental
and DFT calculated spectra above 3000 cm–1, the trend is
accordant. Therefore, we can conclude that the hydrogen bond is
the major contributor to the N-H band differences between
UDMC-TETA and UDMC-DETA. Weak hydrogen bonding
between the N3–H3B of the dimer bridge and the oxygen bridge
redshifts the N−H stretch slightly and broadens it somewhat.
In summary, these UNCI dimers can be easily distinguished
based on their IR spectra (i.e. the four strongest peaks plus N-H
stretching modes in UDMC-TETA and UDMC-DETA).
3.3 Comparison of UV-Vis absorption spectra of UNCI
dimers experimentally and theoretically
Again, the band shape of carbonyl antisymmetric vibration
(1721 - 1697 cm-1) is very diagnostic. The peak is broadened and
asymmetric with implied or distinct shoulder at lower frequency
side (Figs. S12 – S16), which arises from the overlap between
C=C stretching mode and C=O antisymmetric vibration mode
[44].
The observed electronic absorption spectra of four UNCI
dimers (10-5 mol L-1 solution in CH3CN or CH3OH) are shown in
Fig. S17 with the maximum peak centered at 214, 207, 211 and
212 nm, respectively. The assignment of these electronic
transitions can be performed with the help of theoretical
calculations, which provide an in-depth understanding about their
electronic structures and physical properties. This part presents
When C-O-C is converted into C-N-C in the formation of 5-
membered cyclic imide, another two kinds of characteristic

6
DFT/TD-DFT calculations of UNCI dimers’ structure,
Fig. S24, where HOMO is localized over the bridging –NH– and
–CH2– groups, away from the terminal bicyclo [2.2.1] skeletons.
The largest orbital contributions of HOMO arise from the 2p
orbital of bridging N atom (around 69.78%) mixed with fewer
characters of the 2p orbital of four bridging C atoms (around
12.38%). The LUMOs are localized on the terminal bicyclo
[2.2.1] skeletons and show a predominant character of π*C=O and
π*C=C orbitals. So this band is largely originated from nN → π*
transitions. In brief, the band observed at 211.0 nm,
corresponding to the calculated band at 206.2 nm, can be
assigned as the π → π* transitions.
ACCEPTED MANUSCRIPT
electronic states and the main features for electronic absorption
spectrum in gas phase as well as in proper solvents (the same
solvents as those in experiments). Calculation results are in line
with experimental data for all four dimmers (Figs. S17 – S19),
but the degree of match-up is not very good in solvents with
respect to those in gas phase. So the following analyses are
mainly based on the gas phase, in which the lowest 30 singlet -
singlet spin-allowed excitation states (all up to an energy of ~6.0
eV or ~190 nm) were taken into account for all calculations.
The geometry of four dimers used for calculations
corresponds to their single-crystal X-ray structures. The
optimized geometrical images and their single-crystal X-ray
structures are compared and listed in Fig. S20. We notice that
they have the same conformations and the same point groups.
UDMC-TETA adopts approximate Ci symmetry. According
to Table S7 and Fig. S18, in the range of 240–200 nm, there are
four strong excitation states (the S1 at 225.9 nm with f = 0.1064,
the S2 at 215.5 nm with f = 0.0451, the S3 at 212.2 nm with f =
0.0796, and S4 at 205.1 nm with f = 0.0698). It can be seen from
the plots (Figs. S25-S28) that the HOMO levels of S1, S2 and S3
are spread over the bridging -(CH2)2-NH-(CH2)2-NH-(CH2)2-
To compare with experimental results, gas-phase UV-Vis
absorption spectra are calculated. Excitation energies, oscillator
strengths and corresponding electronic transition compositions
for the simulated absorption bands are listed in Table S7. Within
the near UV–Vis range, the strongest absorption oscillator
strengths are found at 206.8, 205.8, 206.2 and 212.2 nm
respectively, with small deviations between 0.2–7.2 nm. On the
basis of these calculated match-up results, the transition
mechanisms can be interpreted.
group, mainly over the
N atoms (are of 2p orbital
characterization), while the HOMO levels of S4 are localized
over C=C bonds and bridging O atoms. The LUMOs (from
LUMO+2 to LUMO+11) overall of these excited states are
almost distributed over the whole molecule except some of the H
atoms. It’s worthy to be noted that the largest orbital
contributions of two main LUMOs (LUMO+6 and LUMO+9)
arise from NH (13.13% and 37.07%, respectively) and CH2
(40.81% and 28.55%, respectively) fragments in the -(CH2)2-NH-
(CH2)2-NH-(CH2)2- bridge, which can be regarded as σN-H* and
σC-H* orbitals. So the observed band at 212 nm, corresponding to
the calculated absorption band at 205-230 nm, which results from
the superposition of four excitation states, can be assigned as
many complicated transitions, mainly including nN → σN-H*, nN
→ σC-H*, nN → π* mixed with πC=C → π* transitions.
For UDMC-BDA (Ci point group), the main electron
transitions of 214.0 nm (calculated result is 206.8 nm) come from
transitions
involving
the
HOMO/HOMO-1,
and
LUMO+3/LUMO+2, LUMO+1/LUMO orbitals (Table S7).
HOMO and HOMO-1 are doubly degenerate with energy
separated by only 0.0016 ev, both located in the C=O bonds (Fig.
S21). LUMO+3 and LUMO+2, LUMO+1 and LUMO are
another two sets of doubly degenerated orbitals with similar
energies (separated by 0.0027 and 0.0079 ev, respectively), all
localized over C=C bonds mixed with C=O bonds (Fig. S21).
Therefore, this band can be assigned as the πC=O → π*C=C mixed
with πC=O → π*C=O transitions.
Thus far, only the calculated results in gas phase have been
demonstrated and those in solvents have been omitted, because
all match the experimental results well and the predicted
transitions are similar. However, the spectral discrepancies
between gas phase and solution are apparent for UDMC-TETA.
In fact, the intrinsic transition mechanisms are different too. For
example, there is only one strong peak above 200 nm. The
calculated excitation energy, oscillator strength and the
assignments of the transitions as well as the corresponding MO
contour plots are shown in Fig. S29. Similar as that in gas phase,
the initial states of the transition are spread over the bridging -
(CH2)2-NH-(CH2)2-NH-(CH2)2- group, mainly over the N atoms
(are of 2p orbital characterization). But the final states are mainly
localized on the cyclic imides, i.e. the π-p-π groups. So the
excitation state at 217.8 nm with f = 0.2237, which mainly comes
from the nN → π* transition, is responsible for the observed band
centered at 212.0 nm.
Like UDMC-BDA, UDMC-HDA adopts Ci point group
symmetry with the HUMO, HUMO-1 and HUMO-2 (the initial
states of these transitions) being triply degenerate and localized
over C=C and C=O bonds (Fig. S22). The LUMO follows a pair
of degenerate LUMO+2 and LUMO+3 MOs, forming the final
states of these transitions, which have a significant contribution
from C=C and C=O bonds and negligible electron density on
other atoms. Apparently, 207.0 nm (calculated 205.8 nm) is from
π → π* transitions.
Due to a lack of symmetry in UDMC-DETA (C1 point
group), the ground-state electronic structure contains little
degeneracy and half parts of the molecule contain different MOs
(Figs. S23 and S24). According to the calculated results, there
are two strong peaks above 200 nm, one at 206.2 nm and another
at 238.1 nm. The intensity of the former is about three times
stronger than the latter and they show different mechanisms. As
the initial states of the transition at 206.2 nm, the high-lying
occupied orbitals (HOMO-1 and HOMO-2) are quasi-
degenerated orbitals with similar energy (-0.26865 and -0.26962
a.u. respectively), and mostly contributed by C=C and C=O
bonds. LUMO+1, LUMO+2 and LUMO+3 orbitals (the final
states of the transition at 206.2 nm) are still mainly composed of
C=C and C=O bonds in antibonding arrangement. To some
extent, this band represents π → π* transitions. But the band
around 238.1 nm is dominated by the single electron excitations
from HOMO to LUMOs (LUMO+5, LUMO+4, LUMO+3). The
distributions of the electronic states in these MOs can be seen in
In summary, both the dimer structures and the absorption
spectra reproduced from theoretical calculations fit the
experimental results well, so attempts to understand the nature of
electronic transitions were carried out through comparison. Some
important notes: 1) The transition mechanisms are different
though showing some similarities; 2) The main bands in Ci
symmetry compounds UDMC-BDA and UDMC-HDA are
dominated by π → π* transitions, where πC=O → π* plays the
predominant role in UDMC-BDA; 3) UDMC-DETA has the
lowest C1 symmetry, and the main band can be assigned as the π
→ π* transitions; 4) UDMC-TETA adopts approximate Ci
symmetry and has the most complicated transition mechanism,
which is most likely due to the introduction of another -(CH2)2-

analysis. No crystal structure of NCI/UNCI dimers can be found
NH- group, increasing the flexibility of tAheCmColEecPulTeEaDnd MANUSCRIPT
reducing the predictability of the resulting UV-Vis properties
.
because they are not easy to form single crystals. Here, we report
the single crystal data for four UNCI dimers.
3.4 Crystal Structures of UNCI dimmers - Molecular
Structure
A summary of the crystal data, experimental details and
refinement results are given in Table 1. Molecular structures
with atomic numbers of four UNCI dimers are depicted in Fig. 1.
As can be seen from the figures, each polycyclic imide skeleton
has the exo-conformation, which is more stable than the endo-
structure and inevitably becomes the overwhelmingly major
products under thermodynamic control [22c, 48].
The crystal structure of UDMC has been reported three
times in 1972 [22a], 1998 [22b] and 2008 [22c]. Hundreds of its
N-substituted derivatives (UNCI) have been synthesized but only
about 22 of their structures (CCDC [47], Version 5.36, updated
to May. 2015) have been confirmed through the X-ray diffraction
Table 1 Crystallographic data and structure refinements summary for four UNCI dimers
Dimers
UDMC-BDA
UDMC-HDA
UDMC-DETA
UDMC-TETA
Chemical formula
C20H20N2O6
384.38
C22H24N2O6
412.43
C20H21N3O6·H2O
417.41
C22H26N4O6
442.47
bar/colorless
monoclinic
P 21
Mr
Crystal habit
Crystal system
Space group
a /Å
block/colorless
monoclinic
P 21/c
bar/colorless
monoclinic
P 21/n
bar/colorless
triclinic
P -1
11.362(3)
5.8986(17)
13.186(4)
90.00
5.246(3)
28.643(17)
6.546(4)
90.00
8.765(3)
9.709(3)
11.940(4)
76.723(4)
78.423(4)
74.682(4)
943.1(5)
2
5.234(4)
6.563(5)
30.52(2)
90.00
b /Å
c /Å
α /°
β /°
98.318(4)
90.00
94.401(10)
90.00
92.618(12)
90.00
γ /°
V /Å3
874.4(4)
2
980.8(10)
2
1047.2(13)
2
Z
Dcalc. /g·cm–3
ꢀ /mm–1
T /K
1.460
1.397
1.470
1.403
0.109
0.102
0.113
0.104
298
298
298
298
F(000)
Rint
404
436
440
468
0.0405
0.0692
0.0615
0.0372
0.0758
0.1878/3012
0.878
R1 [I>2σ(I)]
wR2/reflections
S
0.0631
0.0545
0.0432
0.1335/1713
1.050
0.1057/1733
0.870
0.1118/3269
1.062
entirely differ. For UDMC-BDA, the propeller-like twisting
between the UNCI units imparts chirality to the dimer, while
UDMC-HDA exhibits an antigauche-gauche conformation
without induction of any chirality (Fig. 2).
UDMC-DETA is a “weird” molecule in the following two
parts. 1) The distance between two UNCI moieties is much
shorter than that in other dimers (Fig. 2). It’s only 5.3541(13) Å.
We doubted the stability of this uncommon “U” type conformer
and attributed its existence to the water molecule nearby. But
when the experimental structure without H2O was optimized
using DFT/B3LYP method, the same kind of “U” type conformer
was maintained in the final result. We began to realize that this
conformer is stable with and without H2O nearby. In fact, in
order to confirm this supposition, we have checked many kinds
of conformers. Conformation energy profiles for each single
bond rotation are illustrated in Fig. S30, focusing on
the energy variation of the gas phase UDMC-DETA as groups
revolve around every single bond connection in the bridge chain.
It was found that the energy barriers are very different for these
single bond rotations (from 0.33 eV to 321.66 eV), which are not
really “free”, so the “U” type conformer is somewhat stable when
it is established. On the other hand, most dihedral angles have
wide adaptive range, which leads to a large number of relatively
stable conformations. Fig. S31 shows the comparison of two
most stable conformers, along with the optimized “U” type
conformer derived from the crystal structure, as well as the fourth
Figure 1. ORTEP view of four UNCI dimers with the atom
numbering scheme. Displacement ellipsoids for non-H atoms are
drawn at the 30% probability level. (a) UDMC-BDA, “A” represents
the symmetry code of “−x, 1-y, 2−z” (For simplification, only one
form of the disordered C10-C10A is shown); (b) UDMC-HDA, “A”
represents the symmetry code of “3−x, -y, 1−z”; (c) UDMC-DETA;
(d) UDMC-TETA.
Both UDMC-BDA and UDMC-HDA adopt Ci point group
symmetry and both have an inversion center between the center
two C atoms. Both contain one half-molecule in the asymmetric
units and both have two formula units in the unit cell (Z=2). The
former belongs to the space group P21/c, while the latter belongs
to P21/n. Though the difference in the space group is due to the
cell choice, the bridge -(CH2)n- (n=4 and 6) structure and the
packing mode (will be discussed later) of the two UNCI dimers

8
conformer with imposing C2 point group symmetry. As can be
the C5-C6 double bond and non-bonding p electrons of the
bridging oxygen atom [22a].
ACCEPTED MANUSCRIPT
seen, the “U” type conformer is one of the most stable
conformers (the energy separation ≈ 0.00234 eV with the most
stable one). 2) Ci point group symmetry constrains can’t be
placed on the dimer and only C2 symmetry can be imposed to it,
but the energy is about 0.125 eV higher than that of the most
stable conformer (Fig. S31). Since chirality is the absence of
inversion symmetry, it can be deduced that this mode of “twist
disfavour” may impart chirality into the conformationally
restricted dimer, which can be the origin of chirality in
biarylamine units [49] and in some metal complexes [50].
The last interesting thing to be noted is that, as highly
twisted structures, UDMC-BDA and UDMC-DETA dimers are
in fact helicate since they disfavor the twist [53]. As expected,
UDMC-HDA and UDMC-TETA should have the same kinds of
tendency in helicate formation with high induction of helicity.
The thermodynamic conditions required for the formation of a
single helicate remains elusive.
In brief summary, when the flexible single bond chains
connect the rigid UNCI moieties into dimers, various conformers
will be formed, including zigzag chains, “S” type and “U” type
helicate chains, and often accompanied with the induction of
helical chirality, which enable the possibilities of structure
diversity and property multiplicity.
3.5 Crystal Structures of UNCI dimmers - Packing Structure
In the packing structure of the four dimers, no valuable π-π
stacking interactions can be found and the dominant force is
hydrogen bonding. The presence of various hydrogen bonding
leads to quite interesting supramolecular architectures.
Figure 2. Simplified molecular structure of four UNCI dimers,
illustrating different conformers of the bridge chains. Polycyclic
imide skeleton except N is simplified by its center gravity (red balls
at each end of the chains) and all H atoms have been omitted for
clarity. (a) UDMC-BDA; (b) UDMC-HDA; (c) UDMC-DETA
(water molecule has been omitted for clarity); (d) UDMC-TETA.
UDMC-TETA has the longest distance of 15.7178(67) Å
between two UNCI moieties (Fig. 2), which is mainly because of
the largest number of bridging atoms (8 atoms). As expected, the
bridging chain has the maximum conformational flexibility,
which can lead to subtle or dramatic changes in crystal structure,
tending to reduce the predictability of the resulting assemblies. In
the case of our studied crystal structure, the asymmetric unit
contains one complete UDMC-TETA dimer and the dimer
presents all antigauche conformation for three C-C single bonds.
Moreover, the dimer has an inversion centre through the centre of
the molecule. Surprisingly, such a centrosymmetric dimer
crystallizes in the noncentrosymmetric polar, monoclinic space
group P2(1) (No. 4), which is the only one of the four dimers
crystallizing in a noncentrosymmetric space group. It is known
that the presence of a centrosymmetric supramolecular synthon
strongly tends toward centrosymmetric crystals [51]. However, a
few examples [51, 52] indicate that the bulk crystal chirality can
come from spatial disposition rather than the presence of chiral
molecules themselves. In the case of UDMC-TETA, is the
crystal chiral or achiral? This will be discussed in its packing
structure.
In order to explore the skeleton changes when UDMC forms
UNCI dimers, four dihedral angles (which can be regarded as
one feature of similar skeletons) have been compared in Table
S8 on the basis of their crystal structures. It can be seen that the
skeleton remains unchanged before and after the UNCI dimers’
formation and the influence of environment (for example,
hydrogen bonds) may cause a range-wide variation of some
dihedral angles. For example, the angles between planes C and D
can vary from 123.7° to 134.1° because of the presence or
absence of hydrogen bonds. In most cases, plane D still does not
bisect the angle between planes B and C. A tentative explanation
for this might be the repulsive effect between the π-electrons of
Figure 3. (a) 2D structure of UDMC-BDA formed by C-H…O
interactions (shown red, green and blue dotted lines for three H-
bonds respectively), view along the
a axis. The middle one third
is illustrated by the simplified dimers as used in Fig. 2, aiming
to show different orientations (with blue and green colors, same
in the blow). The left one third is further simplified by the
center of gravity of all dimers, showing the two dimensional
(3,6) topological diagram. (b) 3D structure formed by the
packing of different 2D layers, view along the
b axis.
UDMC-BDA is a 2D infinite layer framework formed by
three kinds of C-H…O interactions (Table S9). Every hydrogen
bond can link the dimer units into the same 2D layer structure,
which extend parallel to the crystallographic bc plane. As can be
seen in Fig. 3a, the repeated 2D rhomboid structures are
composed of an interleaving arrangement of the M- and the P-
forms of enantiomers. Although the dimer has a centrosymmetric
space group (P 21/c) showing no chirality, the packing structure
furnishes both the possible enantiomers, in which each dimer
interacts with two adjacent dimers of opposite chirality by means
of a moderate C-H…O hydrogen bonding interaction. If each
UDMC-BDA dimer is simplified as a node, then their
surrounding H-bonds can make them be simplified as a 6-
connecting node, and the supramolecular layer has a two
dimensional (3,6) net (Fig. 3a). But no hydrogen bond

the aforementioned 1D cage-like chain, view perpendicular to the
interactions can be found between adjacent laAyerCs C(FEig.P3Tb)E. DSo MANUSCRIPT
extending direction, i.e. b axis. (i) End-on view of the crosssection of
another channel inside the cage-like chain, view along the extending
the layer structure is mainly maintained by van der Waals forces.
direction, i.e.
c axis. (j) Packing diagram illustrating the
aforementioned 1D cage-like chain, view perpendicular to the
extending direction, i.e. c axis. (k) Schematic illustration of the 3D
structure view along the c axis.
As the least symmetric one among the four UNCI dimers,
UDMC-DETA has the maximum probability to crystallize in a
noncentrosymmetric lattice instead of its final space group P-1.
So the current result may be contingent. But the packing structure
analysis indicates that one water molecule makes
centrosymmetric space group inevitable. Table S11 shows that
five kinds of intermolecular hydrogen bonds occur in the crystal
structure (O7—H7A···N3 is also an intermolecular H-bond
between water and UDMC-DETA, but it occurs in the same
asymmetric unit. Though it is excluded from “five intermolecular
H-bonds”, this water-involved H-bond plays the most important
role because the construction of most supramolecular dimers
relys on it). Surprisingly, four kinds of them help to link
noncentrosymmetric UDMC-DETA dimers into four kinds of
centrosymmetric supramolecular dimers (Figs. 5a-5e), which in
fact constitute the supramolecular synthons used to build up
crystal architecture. This strongly indicates a loss of chirality
during the presence of a centrosymmetric supramolecular
synthon. These synthons are stacked together, first producing
cage-like chains stretching along two directions (b and c axis),
and then leading to a 2D layer network lying the crystallographic
bc plane (Fig. 5f). At last, the fifth kind of intermolecular
hydrogen bonds (C12—H12···O3 (x+1, y, z)) bridge two adjacent
layers to afford a 3D framework. Interestingly, two channels are
formed within the 2D layer. One runs along the b axis and
another along the c axis (Figs. 5f -7j), both inside the cage-like
chains and running through the center of the assembly. The
former is like an infinite rectangular tube with an effective
crosssection 7.337×3.611 Å based on the shortest atom
separation on opposite walls (Figs. 5g, 5h). The latter is also like
an infinite rectangular tube but with two misaligned bridge O
nearly separated it, in this end-on view the crosssection of the
channel is approximately 7.795×3.723 Å with the narrowest neck
about 1.912 Å (based on the shortest O…O separation) (Figs. 5i,
5j). Water molecules lie outside the channels and only be used to
stabilize them through hydrogen bonds.
Figure 4. (a) Infinite 1D chain of UDMC-HDA formed through C–
H…O H-bonds (shown red dotted lines), view perpendicular to the
extending direction, i.e. c axis. (b) 2D structure formed by
intermolecular C–H…O H-bonds, view perpendicular to the lying
plane, i.e. (1 0 -1) crystal face. (c) 3D structure formed by the
aforementioned (in (a) and (b)) two kinds of C–H…O H-bonds, view
along the b axis. Only three discontinuous 2D layers (horizontal) and
one chain (sloping) are shown for clarity. (d) Schematic illustration
of the 3D structure view along the b axis.
There are two kinds of intermolecular C–H…O hydrogen
bonds in UDMC-HDA (Table S10), both involves the same
y,
carbonyl O. But one H-bond (C3—H3···O2(x, z-1)) link the
dimers into 1D chains (Fig. 4a) and the other one (C6—
H6···O2(x-1/2, -y+1/2, z-1/2)) link the dimers into 2D layers (Fig. 4b).
Then layers are further pillared by chains to generate 3D
networks (Figs. 4c and 4d). The dimers pack with a herringbone
arrangement, so the layers show a zigzag alignment with each
other parallel to (1 0 -1) crystal face. It seems that this packing is
somewhat loose and contains large cavities, yielding the lowest
crystal density of 1.397 g·cm–3 among the four UNCI dimers.
As is well known, the rational design and prediction of
structures in solid state with the help of proper synthons
formation are the main objectives of crystal engineering. In this
sense, given that there is no water molecule in the crystal,
UDMC-DETA could potentially result in structures with
noncentrosymmetric lattice. But our attempts to prepare such
crystal without water met with failure.
Figure 5. (a) Centrosymmetric supramolecular dimer containing two
noncentrosymmetric UDMC-DETA dimers (a pair of enantiomers,
shown in different colors) formed via H-bond O7—H7B···O6(–x+1, -y,
-z). (b) Centrosymmetric supramolecular dimer formed via H-bond
C5—H5···O7 (x, y+1, z). (c) Centrosymmetric supramolecular dimer
formed via H-bond C6—H6···O5 (–x+1, -y+1, -z+1). (d) Centrosymmetric
supramolecular dimer formed via H-bond C15—H15···O7 (–x+1, -y+1, -
z). (e) Four kinds of centrosymmetric supramolecular dimers present
the basic supramolecular synthons, constituting the first step from
molecules to crystals. (f) 2D structure formed by aforementioned
supramolecular synthons, view perpendicular to the spreading plane,
i.e. the crystallographic bc plane. Two arrows are used to show the
formation of two kinds of channels. (g) End-on view of the
crosssection of one channel inside the cage-like chain, view along
the extending direction, i.e. b axis. (h) Packing diagram illustrating

10
Taken together, all four UNCI dimers pack as layers in the
ACCEPTED MANUSCRIPT
crystal, and the layers are connected only by van der Waals
forces in UDMC-BDA, or by H-bonds as well (in UDMC-HDA,
UDMC-DETA and UDMC-TETA). It appears that the
interstitial space between the layers cannot be efficiently filled
and the distances between adjacent layers are significantly
different (Fig. S32). This leads to a very different crystal density
(from 1.397 to 1.470 g·cm–3) and all lower than that of UDMC
(1.550 g·cm–3). It’s worthy to be noted that two UNCI dimers
(UDMC-BDA, UDMC-DETA) have chirality but the presence
of dimer units with opposite chirality makes the crystal achiral.
Another interesting phenomenon is that approximately
centrosymmetric
UDMC-TETA
crystallizes
in
a
noncentrosymmetric space group P21. It can be deduced from our
analysis that the molecular conformation and crystal structure
have lots of possibilities, those reported here are just random
ones.
Figure 6. (a) Infinite 1D chain of UDMC-TETA formed through
y-1/2, -z)
C6–H6…O2(-x,
H-bonds (shown red dotted lines), view
perpendicular to the extending direction, i.e. b axis. (b) 2D structure
formed by four kinds intermolecular C–H…O H-bonds, view
perpendicular to the lying plane, i.e. (1 0 -3) crystal face. The four
kinds of H-bonds are shown in different colors: blue C3—H3···O2(x,
3.6 Antitumor evaluation ─Inhibition of lung/breast cancer
cell growth
To study the growth inhibitory effects of four UNCI dimers
on lung cancer and breast cancer cells, we treated human A549
and mouse 4T1 cells with compounds and examined the growth
of cells with MTT assay. To gain further evidence for their
antitumor activities, the anti-proliferative activities of the
saturated analogues DMC-BDA and DMC-HDA, which were
prepared from UDMC-BDA and UDMC-HDA by catalytic
hydrogenation, were also determined. Meanwhile, the
experiments were carried out with UDMC, DMC and cisplatin
for comparison.
z); red C6—H6···O2(-x,
-z); green C12—H12···O6(x,
;
y-1,
y-1/2,
y+1, z)
purple C15—H15···O6(–x+3, y+1/2, -z+1). (c) 3D structure formed by all
five C–H…O H-bonds, view along the b axis. Only three
discontinuous 2D layers (horizontal) and one chain (sloping) are
shown for clarity. (d) Schematic illustration of the 3D structure view
along the b axis.
There are five kinds of intermolecular hydrogen bonds in
UDMC-TETA crystal (Table S12). Four involve carbonyl
oxygens and they are approximately centrosymmetric, while the
other one contains only one bridge oxygen (C13—H13···O4(x-1, y,
z)) in half part of the dimer. That’s the reason why the dihedral
angles of the skeleton are not symmetrically equal for UDMC-
TETA (see “3.4 Crystal Structures of UNCI dimmers -
Molecular Structure”). Each type of H-bonds results in the
formation of 1D chains (Fig. 6a). The centrosymmetric four H-
bonds work together to link UDMC-TETA dimers into 2D
supramolecular layers (Fig. 6b). Within every layer, the dimers
pack with a herringbone arrangement (similar as that in UDMC-
HDA), so the layers show a zigzag alignment with each other
parallel to (1 0 -3) crystal face (Figs. 6b, 6c). Subsequently, the
H-bonds involving bridge oxygens further connect adjacent
layers into 3D network (Figs. 6c, 6d). Still as that in UDMC-
HDA, the crystal has the second lowest density (1.403 g·cm–3),
indicating this packing mode may be loose.
Table 2. Inhibition of A549 and 4T1 cells growth by four
UNCI dimers, compared with similar compounds
.
compounds
A549 cells growth
4T1 cells growth
inhibition, IC50 (ꢀM)
inhibition, IC50 (ꢀM)
UDMC-BDA
UDMC-HDA
UDMC-DETA
UDMC-TETA
DMC-BDA
DMC-HDA
UDMC
> 100
> 100
> 100
> 100
> 100
> 100
> 100
> 100
> 100
> 100
94.0 ± 2.0
> 100
89.0 ± 1.0
> 100
DMC
49.0 ± 0.9
6.6 ± 0.6
46.0 ± 0.7
0.5 ± 0.08
In the present structural analysis, another key feature is that
such an approximately centrosymmetric dimer crystallize in a
noncentrosymmetric polar space group (P21), which should be
attributed to the “approximate” symmetry instead of “strict” one.
But as we know, if the symmetry-related parts coexist in the
asymmetric unit, they cannot be symmetrically identical. Then,
does the slight asymmetry affect crystal structure? Or is the
minor asymmetry influenced by crystal architecture? The
question remains elusive for us now. Anyway, in the packing
structure, we have not found chiral cavities, which can be formed
through the packing of the molecules. And the detection of
optical activity proves that UDMC-TETA is achiral. A few
words about this topic: The space group P21 is often considered
chiral space group (one of the 65 noncentrosymmetric Sohncke
space group) [52h, 54], but the space group itself is achiral since
it does not form one member of an enantiomorphous pair, even
though a crystal structure in P21 can be chiral [52a].
cisplatin
The cytotoxic activities as 50% inhibitory concentration
(IC50) values are shown in Table 2. It can be seen that in A549
and 4T1 cells, all four UNCI dimers and one NCI dimers as well
as UDMC are inactive (show no noteworthy cytotoxicity at 100
ꢀM drug dose). Only DMC-HDA shows modest cytotoxicity.
McCluskey et al. [15] have reported two NCI dimers, one is
propyl-linked bisnorcantharimide, another is dodecyl-linked
bisnorcantharimide (Scheme 3, compounds 1 and 2), and their
cytotoxic effects against a panel of nine human cancer cell lines
were investigated by MTT assay. Even though the nine cancer
cell lines are different from the two in our experiment, the results
are somewhat similar. Propyl-linked NCI dimer shows very
weak cytotoxicities (Inhibition ratio (%) at 100 ꢀM drug
concentration ranges from 10 ± 7 to 52 ± 41), much less potent
than the dodecyl-linked analogue (GI50 ranges from 8.3 ± 0.7

ꢀM to 60 ± 6
changed into UNCI dimers, the signals of olefinic and methine
protons in H NMR spectra all shift upfield, due to the shielding
effect of substituted N atom. Similar shielding effect is observed
in the 13C NMR spectra for olefinic and methine carbons. But the
carbonyl carbon is always deshielded with respect to their
starting material UDMC, which does not agree with traditional
theory and DFT calculated results. In this paper, we explained the
shielding/deshielding contradiction, which is attributed to the
decrease of coplanarity in the O=C-N(O)-C=O p-π conjugated
system.
ꢀM). Some interesting AstCruCctuEreP-aTctEivDity MANUSC1RIPT
relationships are disclosed: of all the dimers, twelve CH2 groups
linked dimer shows the best cytotoxic activity, six CH2 groups
linked dimer shows only moderate cytotoxicity, while three and
four CH2 groups linked dimers are likewise inactive. In this
sense, we could observe that the longer chain, the higher
therapeutic efficacy for NCI dimers. Long chain linked dimers
appear to be endowed with cytotoxic activity, though the reason
is still unclear.
S. H. Li et al. [12] have reported the antiproliferative
activities of a series of ten UNCI and NCI derivatives (Scheme
2, compounds 5a-f, 7a-d), which displayed moderate and similar
inhibitory activities against A549 and PC-3 cell lines. But the
IC50 values of UNCI series are a little higher than that of their
NCI analogues. Another important reference is J. S. Li’ paper
[11], in which they studied in vitro antitumor activities of fifteen
arylantimony derivatives based on UNCI/NCI. The structures of
four most cytotoxic derivatives (I6, I7, II7 and II8) are listed in
Scheme 2. The inhibition ratio against six cancer cell lines at 10
ꢀg/ml drug concentration for I6 range from 6.9 to 74.5%, while
from 38.6 to 87.6% for II7, indicating NCI derivatives (II7) have
stronger inhibitory activity than UNCI analogues (I6). At the
same time, the inhibition ratio range from 84.8 to 97.3% for I7,
and from 33.6 to 92.2% for II8, indicating UNCI derivatives (I7)
have higher inhibitory activity than NCI analogues (II8). When
considering the results of our present research, DMC-HDA
shows modest cytotoxicity, while UDMC-HDA is inactive. It
may be deduced that the double bond between C5, C6 positions
may slightly inhibit cytotoxic activity in most cases.
Inspection of the four experimental IR spectra and
comparison with the starting material and with their
corresponding theoretical spectra leads to the conclusion that the
four most intense peaks are diagnostic and they can characterize
the formation of the cyclic imide rings: 1771 - 1767 cm-1 (C=O
of cyclic imide anti-symmetric stretching), 1721 - 1697 cm-1
(C=O of cyclic imide symmetric stretching), 1412 - 1398 cm-1
(C-N-C symmetric vibrations) and 1203 - 1165 cm-1 (C-N-C anti-
symmetric vibrations).
UV-Vis absorption spectra of the four UNCI dimers were
reproduced from TD-DFT calculations, which match the
experimentally obtained spectra. Attempts to understand the
nature of electronic transitions were carried out through
comparison. All together our data, the main bands in the dimers
are dominated by π → π* mixed with n → π* transitions and the
detailed transition mechanisms are different.
The single crystals of a series of UNCI dimers enable us to
compare their molecular and crystal structures systematically.
Comparison of the molecular assemblies in the crystals clearly
demonstrated that only small structural differences in a molecule,
that is, the number of CH2 or NH groups in link chains, cause a
significant change in the assembly in the crystalline state.
UDMC-HDA and UDMC-TETA exhibit common antigauche-
gauche conformation without induction of any chirality, but
UDMC-BDA and UDMC-DETA present “S” type or “U” type
helicate configurations with chirality. When packing into
crystals, both helicates crystallize into centrosymmetric lattice
with the loss of chirality, but approximately centrosymmetric
UDMC-TETA packs into noncentrosymmetric polar space group
(P21) though still without chirality. In their crystal packing
structures, supramolecular synthons via conventional hydrogen
bonds have been analyzed and the common feature is that 2D
layer structures are formed. There have been no reports in which
such subtle link group effects in crystals of dimeric molecules
were systematically analyzed and, therefore, the present system
can contribute to the design of desired functional crystalline
materials.
Presumably in vitro antitumor properties of UNCI/NCI
dimers depend mainly on the length of link chains and have
something to do with the double bond (5,6-ene), but this
hypothesis requires more UNCI and NCI dimer analogues with
exact structure information as well as additional biologically
evaluations. These works as well as detailed investigations
focusing on cytotoxicity in several different types of cancer cells
(some CAN analogues show more specific inhibitory and
cytotoxic activities on both the Hep3B HCC and the KG1a AML
cell lines, but their cytotoxic effects were weaker on both A549
lung cancer and MDA-MB231 breast cancer [17]) and the
mechanisms of action are ongoing.
It is still earlier to summarize the situation in the context of
UDMC/DMC dimers modification. Because, we still have only
learned what not to do if one wishes to improve its anti-tumor
activities (for example, one should avoid to change the bridging
O atom) rather than what one must aim for in order to ensure
improved bioactivities, for example, the bonding properties
(saturated or unsaturated), the length of link chains, the
solubility, charge, chirality, crystal structure, conformation (exo-,
endo-, etc.) and the degree of polymerization (dimer or
monomer). But our preliminary data provide entity to study these
structural factors associated with biological activities.
Cell viability assay demonstrated that these four UNCI
dimers were ineffective death inducers in human A549 and
mouse 4T1 cells and control experiments with their saturated
analogues indicated that only DMC-HDA shows modest
cytotoxicity. By comparing with the relevant results obtained
from the literature, a preliminary conclusion suggests that the
antitumor properties of UNCI/NCI dimers depend mainly on the
length of link chains (the longer chain, the higher therapeutic
efficacy) and have something to do with the double bond (in
most cases, NCI derivatives suppress tumor growth more
effectively than UNCI analogues). To clarify the biological role
of link chains, we next plan to synthesize various and longer
chains linked UNCI/NCI dimers based on this model.
Meanwhile, a two-step assembly process is invoked as a way of
rationalizing the observed structures and we will pursue the
crystal structures of corresponding saturated dimers in order to
shed more light on the precise similarities/differences in their
4 Conclusions
In this paper we report the synthesis and characterization of
four UNCI dimers, in which two UNCI units are linked with four
and six CH2 groups or -(CH2)2-NH-(CH2)2- and -(CH2)2-NH-
(CH2)2-NH-(CH2)2- chains. To compare with the experimental
results, UV-Vis, IR spectra and the relationships between
conformation and energy were investigated by theoretical
calculations.
The structural elucidation and the complete NMR
assignment of the four dimers were performed. When UDMC is

12
[4] (a) K. K. W. To, Y. P. Ho, S. C. F. Au-Yeung, J. Chromatogr. A
.
supramolecular structures. We expect that the unambiguous
ACCEPTED MANUSCRIPT
2002
Res. Rev., 2003
Y. P. Ho, S. C. F. Au-Yeung, Bioorg. Med. Chem. 2004
(d) K. K. W. To, Y. P. Ho, S. C. F. Au-Yeung, Cancer Lett
223, 227; (e) C. W. Yu, K. K. W. Li, S. K. Pang, S. C. F. Au-Yeung,
Y. P. Ho, Bioorg. Med. Chem. Lett. 2006 16, 1686; (f) S. C. F. Au-
Yeung, P. S. K. Pang, Y. P. Ho, Nat. Rev. Drug Discov., 2006 , 5;
(g) S. K. Pang, C. W. Yu, S. C. F. Au-Yeung, Y. P. Ho, Biochem.
Biophys. Res. Commun 2007 363, 235; (h) S. K. Pang, C. W. Yu, S.
C. F. Au-Yeung, Y. P. Ho, Oncol. Reports 2008 20, 1269.
[5] (a) A. Tan, B. Koc, E. Sahin, N. H. Kishali, Y. Kara SYNTHESIS
2011, 7, 1079; (b) P. Y. Lin, S. J. Shi, H. L. Shu, H. F. Chen, C. C.
Lin, P. C. Liu, L. F. Wang, Bioorganic Chem 2000 28, 266; (c) I. J.
Tseng, S. Y. Sheu, P. Y. Li, J. A. Lee, K. L. Ou, L. W. Lee, J Exp
Clin Med 2012 , 280; (d) C. E. P. Galvis, L. Y. V. Méndez, V. V.
Kouznetsov, Chem. Biol. Drug. Des 2013 82, 477; (e) B. Sauer, J.
Gilbert, J. A. Sakoff, A. McCluskey, Lett. Drug Des. Discov. 2009
6, 1; f) M. J. Robertson, C. P. Gordon, J. Gilbert, A. McCluskey, A.
Jennette, J. A. Sakoff, Bioorg Med Chem 2011 19, 5734.
[6] A. McCluskey, M. C. Bowyer, E. Collins, A. T. R. Sim, J. A. Sakoff,
M. L. Baldwin, Bioorg. Med. Chem. Lett 2000 10, 1687.
,
947, 319; (b) Y. P. Ho, S. C. F. Au-Yeung, K. K. W. To, Med.
23 (5), 633; (c) K. K. W. To, X. Wang, C. W. Yu,
12, 4565;
2005
structure will help to illustrate the whole structure-activity
relationship deeply.
,
,
Acknowledgments
,
,
The authors are thankful to the Science & Technology
Development Projects of Shandong Province in China (No.
2011GGX10706 and 2012GGX10702) and the National Natural
Science Foundation of PR China (NSFC, No. 21103100) for
financial support. Generous support in computing resources from
the National Supercomputer Center in Jinan, China, is greatly
appreciated.
,
,
5
,
,
,
,
，
,
,
,
Supplementary Material
,
, 4
,
,
Supplementary Information available: Tables S1 to S12,
Figures S1 to S32 mentioned in the text. Crystallographic
information files of four UNCI dimers. CCDC < 1446287,
1446288, 1446289 and 1446290> contains the supplementary
crystallographic data for < UDMC-BDA, UDMC-HDA, UDMC-
DETA and UDMC-TETA >. These data can be obtained free of
charge from The Cambridge Crystallographic Data Centre via
http://www.ccdc.cam.ac.uk/data_request/cif , or from the Cambridge
Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ,
UK; fax: (+44) 1223-336-033; or e-mail: deposit@ccdc.cam.ac.uk.
,
(
,
,
,
,
[7] A. McCluskey, C. Walkom, M. C. Bowyer, S. P. Ackland, E.
Gardiner, J. A. Sak.
[8] W. Tsauer, J. G. Lin, P. Y. Lin, F. L. Hsu, H. C. Chiang, Anticancer
Res,1997
[9] (a) A. McCluskey, S. P. Ackland, M. C. Bowyer, M. L. Baldwin, J.
Garner, C. C. Walkom, J. A. Sakoff, Bioorg Chem 2003 31, 68; (b)
A. McCluskey, A. T. Sim, J. A. Sakoff, J Med Chem 2002 45
1151; (c) M. E. Hart, A. R. Chamberlin, C. Walkom, J. A. Sakoff, A.
McCluskey, Bioorg. Med. Chem. Lett, 2004 14, 1969; (d) A.
McCluskey, M. A. Keane, L. M. Mudgee, A. T. Sim, J. Sakoff, R. J.
Quinn, Eur J Med Chem 2000 35, 957.
[10] M. C. Chang, E. T. Tsai, J. Y. Wu, H. F. Liao, Y. J. Chen, C. D. Kuo,
Anti-Cancer Drugs 2015 26, 5, 508.
[11] G. C. Wang, J. Xiao, L. Yu, J. S. Li, J. R. Cui, R. Q. Wang, F. X.
Ran, J Organomet Chem 2004 689, 1631.
[12] G. G. Tu, J. F. Zhan, Q. L. Lv, J. Q. Wang, B. H. Kuang, S. H. Li,
Med. Chem. 2014 10, 376.
[13] P. S. Portoghese, J. Med. Chem. 1992
, 17, 2095.
,
,
,
,
,
,
References and notes
,
[1] (a) I. Bertini, V. Calderone, M. Fragai, C. Luchinat, E. Talluri, J. Med.
Chem. 2009
Sci. U.S.A 1992
Gombosova, B. Linck, H. Luss, F. U. Muller, T. Muller, P. Nacke,
,
52, 4838; (b) Y. M. Li, J. E. Casida, Proc. Natl. Acad.
,
,
.
,
89, 11867; (c) J. Knapp, P. Boknik, S. Huke, I.
,
,
W. Schmitz, U. Vahlensieck, J. Neumann, Gen. Pharmacol
31, 729; (d) A. McCluskey, J.A. Sakoff, Mini. Rev. Med. Chem
2001
Am. Chem. Soc, 2003
Kobayashi, N. Hirukawa
(g) S. G. Stewart, T. A. Hill, J. Gilbert, S. P. Ackland, J. A. Sakoff,
A. McCluskey Bioorganic & Medicinal Chemistry 2007 15, 7301.
,
1998
,
,
,
,
1
, 43; (e) Y. Baba, N. Hirukawa, N. Tanohira, M. Sodeoka, J.
125, 9740; (f) M. Sodeoka,Y. Baba,S.
Bioorg. Med. Chem. Lett 1997 , 1833
,
35, 1927.
,
,
[14] B. Gervais, Current Medicinal Chemistry, 2006, 13, 131.
[15] T. A. Hill, S. G. Stewart, S. P. Ackland, J. Gilbert, B. Sauer, J. A.
Sakoff, A. McCluskey, Bioorg Med Chem, 2007,15, 6126.
[16] T. Nakatani, K. Jinpo, N. Noda, Chem. Pharm. Bull, 2007, 55 (1),
,
,
7
；
ꢀ
,
,
[2] (a) T. Efferth, R. Rauh, S. Kahl, M. Tomicic, H. Bo¨chzelt, M. E.
92.
Tome, M. M. Briehl, R. Bauer, B. Kaina, Biochemical
[17] S. H. L. Kok, C. H. Chui, W. S. Lam, J. Chen, J. C. O. Tang, F. Y.
Lau, G. Y. M. Cheng, R. S. M. Wong, A. S. C. Chan, Int J Mol Med
2006 17, 151.
[18] N. Variankaval, A. S. Cote, M. F. Doherty, American Institute Of
Chemical Engineers 2008 54, 1682.
[19] G. P. Stahly, Cryst. Growth. Des 2007
[20] (a) F. Y. Wang, X. B. Jiang, X. J. Tan, D. X. Xing, Acta Crystallogr
C67, 2011, m218; (b) S. P. Li, S. Zhang, X. Jin, X. J. Tan, J. F. Lou,
X. M. Zhang, Y. X. Zhao, Eur. J. Med. Chem 2014, 86, 1; (c) X. Jin,
X. J. Tan, X. M. Zhang, M. Y. Han, Y. X. Zhao, Bioorg. Med. Chem.
Lett 2015 25, 4477.
[21] V. Rao, S. Navath, M. Kottur, J. R. McElhanon, D. V. McGrath,
Tetrahedron Lett 2013 54, 5011.
[22] (a) S. Baggio, A. Barriola, P. K. de Perazzo, J.Chem.Soc.,Perkin
Trans 1972 , 934; (b) B. M. Ramirez, G. D. de Delgado, W. O.
Velasquez, G. M. Perdomo, Z. Kristallogr.- New Crystal
Structures 1998 213, 197; (c) Y. W. Goh, B. R. Pool, J. M. White,
J.Org.Chem.,2008 73 151.
Pharmacology
Ho, Journal of Inorganic Biochemistry
K.K.W. To, S.C.F. Au-Yeung, X. Wang, G. Lin, X. Han, J. Med.
Chem. 2001 44, 2065; (d) G. S. Wang, J Ethnopharmacol, 1989 26
147; (e) L.P. Deng, J. Dong, H. Cai, W. Wang Current Medicinal
Chemistry 2013 20, 159 (f) F. Peng, Y. Q. Wei, L. Tian, L. Yang,
X. Zhao, Y. Lu, Y. Q. Mao, B. Kan, S. Lei, G. S. Wang, Y. Jiang, Q.
R. Wang, F. Luo, L. Q. Zou, J. Y. Liu, Cancer Res. Clin. Oncol.
2002 128, 223; (g) Y.J. Chen, C.D. Kuo,Y.M. Tsai, C.C. Yu, G.S.
Wang, H.F. Liao, Anti-Cancer. Drugs 2008 19, 55; (h) Y. N. Chen,
J. C. Chen, S. C. Yin, G. S. Wang, W. Tsauer, S. F. Hsu, S. L. Hsu,
Int. J. Cancer 2002 100, 158; (i) L. A.Williams, W. Moller, E.
Merisor, W. Kraus, H. Rosner, West. Ind. Med.J., 2003 52, 10; (j) J.
E. Huh, K. S. Kang, C. Chae, H. M. Kim, K. S. Ahn, S. H. Kim,
Biochem Pharmacol., 2004 67, 1811; (k) J. Z. Wu, Z. Q. Situ, J. Y.
Chen, B. Liu, W. Wang Chin. Med. J., 1992 105, 1026; (l) S. H.
Kok, C. H. Chui, W. S. Lam, J. Chen, J. C. Tang, F. Y. Lau, G. Y.
Cheng, R. S. Wong, A. S. Chan, Int. J. Mol. Med. 2006 17, 945;
(m) S. H. Kok, C. H. Chui, W. S. Lam, J. Chen, J. C. Tang, F. Y.
Lau, G. Y. Cheng, R. S. Wong, A. S. Chan,Int. J. Mol. Med., 2006
18, 1217; (n) W. Li, L. Xie, Z. Chen, Y. Zhu, Y. Sun, Y. Miao, Z.
Xu, X. Han, Cancer Sci., 2010 101, 1226; (o) L. C. Nicholls, D.
Teare, Br. Med. J. 1954 , 1384; (p) P. Y. Lin, S. J. Shi, F. L. Hsu,
C. F. Chen, J. Chin. Chem. Soc. (Taipei, Taiwan) 1998 45, 323.
[3] R. Rauh,S. Kahl,H. Boechzelt, R. Bauer,B. Kaina, T. Efferth, Chinese
Medicine 2007 , 1.
,
2006
,
69 ,811; (b) X. Wang, S. C.F. Au-Yeung, Y.P.
,
,
2007 101,909; (c) Y.P. Ho,
,
,
,
,
,
,
,
,
,
.
, 7, 1007.
,
,
;
.
,
,
,
,
,
,
,
,
,
,
,
,
,
, 2
,
,
,
,
,
,
,
,
,
[23] G. M. Sheldrick, SADABS, Bruker AXS Inc., Madison, Wisconsin,
USA, 2004.
,
[24] G. M. Sheldrick, Acta Crystallogr, Sect. A, 2008, 64 112.
[25] Bruker APEX II (version 2.0-1), SAINT (version 7.23A), XPREP
(version 2005/2), SHELXTL (version 6.1). Bruker AXS Inc.,
,
,
2
Madison, Wisconsin, USA, 2005
[26] K. Brandenburg, H. Putz, DIAMOND. Crystal Impact GbR, Bonn,
Germany, 2005
[27] A. D. Becke, J. Chem. Phys. 1993
.
,
.
,
, 2
,
98, 5648.

[28] C. Lee, W. Yang, R. G. Parr, Phys. Rev. B: Condens. Matter Mater.
[45] P. J. Mcneilly, T. A. Andrea, S. E. Krikorian, Spectrochim. Acta A,
ACCEPTED MANUSCRIPT
Phys. 1988
[29] (a) R. Bauernschmitt, R. Ahlrichs, Chem. Phys. Lett. 1996
(b) R. E. Stratmann, G. E. Scuseria, M. J. Frisch, J. Chem. Phys.,
1998 109, 8218; (c) M. E. Casida, C. Jamorski, K. C. Casida, D. R.
Salahub, J. Chem. Phys 1998 108 , 4439.
[30] A. B. Nielsen, A. J. Holder, GaussView, User´s Reference
GAUSSIAN, Inc.: Pittsburgh, PA, USA 1997-1998
,
37 ,785.
1992
[46] J. Grzetic, J. Oomens, J. Am. Soc. Mass Spectrom
[47] F. H. Allen, Acta Crystallogr. B 2002 58, 380.
[48] (a) M. W. Lee, W. C. Herndon, J. Org. Chem. 1978
, 48, 1437.
,256, 454;
,
2013
, 24, 1228.
,
,
,
,
43, 518; (b) P.
,
,
X. Iten, A
2202.
.
A. Hofmann, C. H. Eugster, Helv. Chim. Acta
,
1979
,
62
,
,
,
.
[49] C. J. Sanders, K. M. Gillespie, D. Bell, P. Scott, J. Am. Chem.Soc.
2000 122, 7132.
[50] (a) X. Sun, D. W. Johnson, D. L. Caulder, K. N. Raymond, E. H.
Wong, J. Am. Chem. Soc. 2001 123, 2752; (b) A. M. Garcia-Deibe,
J. S. Metalobos, M. Fondo, M. Vazquez, M. R. Bermejo, Inorg.
Chim. Acta 2004 357, 2561; (c) X. Sun, D. W. Johnson, K. N.
Raymond, E. H. Wong, Inorg. Chem. 2001 40, 4504; (d) Y. P. Cai,
C. Y. Su, C. L. Chen, Y. M. Li, B. S. Kang, A. S. C. Chan, W. Kaim,
Inorg. Chem. 2003 42, 163; (e) W. Schuh, H. Kopacka, K. Wurst,
P. Peringer, Eur. J. Inorg. Chem. 2002, 2202; ( ) M. Albrecht, I.
Janser, H. Houjou, R. Frohlich, Chem.-Eur. J. 2004 10, 2839; (g
M. Albrecht, Chem.-Eur. J. 1997 , 1466; (h) S. P. Argent, H.
Adams, T. Riis-Johannessen, J. C. Jeffery, L. P. Harding, W. Clegg,
R. W. Harrington, M. D. Ward, Dalton Trans. 2006, 4996.
[31] (a) S. I. Gorelsky, SWizard Program Revision 4.6; University of
Ottawa: Ottawa, Canada, 2010; http://www.sg-chem.net/. (b) S. I.
,
Gorelsky, A. B. P. Lever, J. Organomet. Chem. 2001
,
635, 187.
,
,
[32] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A.
Robb, J. R. Cheeseman, J. A. Jr. Montgomery, T. Vreven, K. N.
Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V.
Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G. A.
Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda,
J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai,
M. Klene, X. Li, J. E. Knox, H. P. Hratchian, J. B. Cross, V. Bakken,
C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev,
A. J.Austin, R. Cammi, C. Pomelli, J. W. Ochterski, P. Y. Ayala, K.
Morokuma, G. A. Voth, P. Salvador, J. J. Dannenberg, V. G.
Zakrzewski, S. Dapprich, A. D. Daniels, M. C. Strain, O. Farkas, D.
K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V.
Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski, B. B.
Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. L.
Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A.
Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen,
M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian 03, revision C.02;
,
,
,
,
,
,
,
f
,
,
)
,
, 3
,
[51] B. Moulton, M. J. Zaworotko, Chem. Rev. 2001, 101, 1629.
[52] (a) C. Janiak, A. Chamayou, A. K. M. R. Uddin, M. Uddin, K. S.
Hagen, M. Enamullah, Dalton Trans.
W. Fu, Q. Ye, R. G. Xiong, Cryst. Growth Des
Ye, T. Hang, D. W. Fu, G. H. Xu, R. G. Xiong , Cryst. Growth Des.
2008 , 3501; (d) C. J. Sumby, J. Fisher, T. J. Prior, M. J. Hardie,
Chem. Eur. J 2006 12, 2945; (e) Q. Sun, Y. Bai, G. He, C. Duan, Z.
Lin, Q. Meng Chem. Commun., 2006, 2777; (f) R. W. Saalfrank, H.
,
2009, 3698; (b) T. Hang, D.
.
2009 , 2026; (c) Q.
, 9
,
8
.
,
Gaussian, Inc.: Wallingford, CT, 2004
[33] Y. Zhao, W. Liu, Y. Zhou, X. Zhang, P. V. Murphy, Bioorg. Med.
Chem. Lett 2010, 20, 7540.
[34] M. Majumder, N. Sathyamurthy, Theor Chem Acc
[35] (a) L. A. Mitchell, J. M. Stanley, L. E. De Hoyos, B. J. Holliday,
Acta Crystallogr.,Sect.C (Cr.Str.Comm.), 2013, 69 638; (b) T.
Fallon, A. C. Willis, M. N. Paddon-Row, M. S. Sherburn,
J.Org.Chem., 2014 79 3185.
[36] A. E. Reed, F. J. Weinstock, Chem. Phy, 1985
[37] D. H. Marr, J. B. Stothers, Can. J. Chem. 1965
.
,
Maid, F. Hampel, K. Peters, Eur. J. Inorg. Chem. 1999, 1859; (g) O.
Mamula, A. Von Zelewsky, T. Bark, G. Bernardinelli, Angew.
.
,
2012
,
131, 1092.
Chem., Int. Ed. Engl. 1999
Aoyama, J. Am. Chem. Soc. 1999
Seward, M. Zaworotko, J. Angew. Chem., Int. Ed. Engl. 1999
492; (j) C. Kaes, M. W. Hosseini, C. E. F. Rickard, B. W. Skelton,
A. H. White, Angew. Chem., Int. Ed. Engl 1998 37, 920; (k) D.
,
38, 2945; (h) T. Ezuhara, K. Endo, Y.
121, 3279; (i) K. Biradha, C.
38,
,
,
,
,
,
,
,
,
83, 735.
Whang, J. Heo, C. A. Kim, K. Kim, Chem. Commun. 1997, 2361; (l)
E. Psillakis, J. C. Jeffery, J. A. McCleverty, M. D. Ward, Dalton
Trans. 1997, 1645; (m) M. A. Withersby, A. J. Blake, N. R.
Champness, P. Hubberstey, W. S. Li, M. Schroder, Angew. Chem.,
,
, 43, 596.
[38] E. P. Prokof'ev, Zh. A. Krasnaya, Russian Chemical Bulletin
1978 27 10 2037.
[39] X. Gu, C. Q. Yang, Res. Chem. Intermed., 1998
,
,
,
,
,
24, 979.
Int. Ed. Engl. 1997
Fujita, J. X. Lu, Chem. Eur. J. 2000
Schmitz, A. A. Arif, P. Stang, J. Inorg. Chem. 2000
D. Ranford, J. J. Vittal, D. Q. Wu, X. D. Yang, Angew. Chem., Int.
Ed. Engl. 1999 38, 3498; (q) W. Jaunky, M. W. Hosseini, J. M.
,
36, 2327; (n) M. C. Hong, W. P. Su, R. Cao, M.
, 427; (o) W. W. Ellis, M.
39, 2547; (p) J.
[40] N. B. Colthup, L. H. Daly, S. E. Wiberley, Introduction to Infrared
and Ramon Spectroscopy, Academic Press, New York. 1975, pp 297,
308.
, 6
,
[41] (a) P. T. McKitfrick, J. E. Katon, Appl. Spectrosc. 1990
,
44
，812;
,
(b) H. Niu, M. Huang, S. Qi, E. Han, G. Tian, X. Wang, D. Wu,
Polymer, 2013, 54, 1700; (c) M. Zhang, H. Niu, J. Chang, Q. Ge, L.
Cao, D. Wu, Polym. Eng. Sci. 2015, 2615-2625; (d) I. N. Tarabara,
A. O. Kas’yan, M. Yu. Yarovoi, S. V. Shishkina, O. V. Shishkin, L.
Planeix, A. De Cian, N. Kyritsakas, Fischer, J. Chem. Commun.
1999, 2313.
[53] J. M. Lehn, A. Rigault, J. Siegel, J. Harrowfield, B. Chevrier, D.
Moras, Proc. Natl. Acad. Sci. U. S. A. 1987, 84, 2565.
I. Kas’yan, Russ. J. Org. Chem. 2004
,
40, 992. In (b), (c) and (d), the
[54] (a) K. Matsuda, S. Yamamoto, M. Irie, Tetrahedron Lett.
2001 42(41), 7291-7293; (b) W. G. Lu J. Z. Gu L. Jiang M. Y.
T. B. Lu, Cryst. Growth & Des. , 192; (c) Z. Chen, Y.
W. Xiong, L. Wang, Liang, M. Shao,
2009
P.S. Subramanian Polyhedron, 2009, 28, 245; (e) H. Wu, X. Dong,
,
authors believe asymmetic stretching lies in higher wavenumbers,
which maybe wrong according to our calculated results and other
relevant references.
,
，
2007, 8
F.
，
，
Tan
Su,
，
,
[42] X. J. Tan, H. Z. Liu, C. Z. Ye, J. F. Lou, Y. Liu, D. X. Xing, S. P. Li,
CrystEngComm
,
, 11, 318; (d) P. M. Selvakumar, E. Suresh,
S. L. Liu, L. Z. Song , Polyhedron
，2014
,
71, 119.
,
[43] C. Q. Yang, G. Zhang, Res. Chem. Intermed., 2000
[44] P. A. Okafor, B. Huxel, J. O. Iroh, J. Appl. Polym. Sci. 2015, DOI:
,
26, 515.
H. Liu, J. Ma, Y. Liu, Y. Liu, J. Yang
19.
, Inorg. Chim. Acta, 2011, 373,
10.1002/APP.42673.
.

ACCEPTED MANUSCRIPT
Highlights
1) Four unsaturated norcantharimide dimers were synthesized and well characterized.
2) Single crystal structure studies have been carried out, which is the first report about the crystal
structures of cantharidin derivative dimers.
3) Introduction of various lengths of single bond chains provides high conformational flexibility,
which can afford unusual structures and interesting biological activities.
4) In vitro antitumor evaluations indicate that the antitumor properties of dimers depend on the
length of link chains.