Self-Assembly of M30L60 Icosidodecahedron

Article abstract of DOI:10.1016/j.chempr.2016.06.007

Self-assembly is invaluable in the construction of giant molecular structures via a bottom-up approach. Although living organisms naturally make the most use of self-assembly and freely handle the mechanism at will, scientists are still far behind the level of nature. Inspired by the elegant structures of virus capsids, we have previously constructed roughly spherical giant metal complexes with the symmetry of an octahedron, cuboctahedron, and rhombicuboctahedron with the compositions M₆L₁₂, M₁₂L₂₄, and M₂₄L₄₈, respectively. Here, we report our first successful synthesis of an M₃₀L₆₀ molecular icosidodecahedron that consists of ～100 components: 30 Pd(II) ions and 60 ligands that assemble into the largest well-defined spherical macromolecule to date (diameter of ～8.2?nm). Tuning the flexibility of the ligand was the key for successful self-assembly. A highly complex but symmetrically organized structure was identified through X-ray crystallographic analysis. The interior space of the molecular complex is large enough (157,000??³) to enclose proteins.

Full text of DOI:10.1016/j.chempr.2016.06.007

Article
Self-Assembly of M30L60 Icosidodecahedron
Daishi Fujita, Yoshihiro Ueda,
Sota Sato, Hiroyuki Yokoyama,
Nobuhiro Mizuno, Takashi
Kumasaka, Makoto Fujita
mfujita@appchem.t.u-tokyo.ac.jp
HIGHLIGHTS
Self-assembly of an M30L60
icosidodecahedron
Bridging ligand design for
flexibility tuning
Crystallization and fine X-ray
analysis of M30L60
ACCESSION NUMBERS
CCDC: 1482268
Bottom-up construction of giant structures by the self-assembly of a large number
of components (n = ꢀ100) has been a daunting challenge. Here, Fujita and
colleagues report the self-assembly of a spherical metal polyhedron, possessing a
hitherto unreported icosidodecahedron geometry with 30 vertices and 60 edges.
The authors succeeded in controlling the self-assembly by intensive tuning of the
ligand flexibility. X-ray crystallographic analysis confirmed that the complex is the
largest well-defined spherical molecular capsule, comparable with the size of a
typical protein.
Fujita et al., Chem 1, 91–101
July 7, 2016 ª 2016 Elsevier Inc.
http://dx.doi.org/10.1016/j.chempr.2016.06.007

Article
Self-Assembly of M30L60 Icosidodecahedron
Daishi Fujita,1,2,5 Yoshihiro Ueda,1,5 Sota Sato,3 Hiroyuki Yokoyama,1 Nobuhiro Mizuno,4
Takashi Kumasaka,4 and Makoto Fujita1,6,
*
SUMMARY
The Bigger Picture
The technology we report here is a
milestone for the development of
nanotechnology. There are two
opposite approaches to
Self-assembly is invaluable in the construction of giant molecular structures via a
bottom-up approach. Although living organisms naturally make the most use of
self-assembly and freely handle the mechanism at will, scientists are still far
behind the level of nature. Inspired by the elegant structures of virus capsids,
we have previously constructed roughly spherical giant metal complexes with
the symmetry of an octahedron, cuboctahedron, and rhombicuboctahedron
with the compositions M6L12, M12L24, and M24L48, respectively. Here, we report
our first successful synthesis of an M30L60 molecular icosidodecahedron that
consists of ꢀ100 components: 30 Pd(II) ions and 60 ligands that assemble into
the largest well-defined spherical macromolecule to date (diameter of
ꢀ8.2 nm). Tuning the flexibility of the ligand was the key for successful self-as-
sembly. A highly complex but symmetrically organized structure was identified
through X-ray crystallographic analysis. The interior space of the molecular com-
fabricating accurate nanometer-
sized structures: top-down and
bottom-up approaches. The top-
down approach, typically using
photolithography techniques, has
developed rapidly along with the
semiconductor industry and has
reached the level whereby they
can handle single-digit
3
˚
nanometers. However, this
plex is large enough (157,000 A ) to enclose proteins.
approach is already getting close
to the theoretical limit with no
further room for development;
control of the structure with
INTRODUCTION
Molecular self-assembly commonly underlies the creation of various biological mac-
romolecules. It enables the construction of giant structures of a size not practically
attainable by a single molecular unit. Ferritin1, COPII coat cage,2,3 and various virus
capsids4,5 are representatives from nature’s many examples. Inspired by this elegant
biology, scientists have long pursued a way to emulate and acquire equally impres-
sive molecular self-assembly constructs and strategies.6–14 The design of molecular
self-assembly constructs with fewer than ten components, or at most a couple of
dozen, has becomes routine;15–27 however, self-assembly from a number of compo-
nents comparable with that seen in nature (ꢀ100 components) is still inherently chal-
lenging. Part of this challenge lies in the fact that the process of self-assembly and
the final structure obtained strongly depend on the number and precise nature of
the initial components. For example, when trying to predict the final self-assembled
structure of a newly synthesized ligand using results obtained from another, almost
identical ligand possessing a structure with only subtle differences, these tiny differ-
ences become amplified during self-assembly, making the structure of the final
product difficult to forecast. As a result, rational design strategies and mechanistic
discussion regarding molecular self-assembly for synthetic macromolecules are
extremely important but still unexplored.
atomic-level accuracy, at least, will
not be possible. In this respect,
the bottom-up approach is highly
promising for the manipulation of
matter of interest with atomic-
level accuracy, although this
approach is still too premature for
building structures bigger than
multiple nanometers. Here, we
have built a giant (8 nm) molecular
architecture by self-assembling
ꢀ100 components. This report
pushes our technological
progress to the next level.
We have been tackling this uncharted area by utilizing self-assembly of divalent Pd2+
ions (M) and ditopic pyridyl ligands with a bend angle (L), which gives giant spherical
28
metal complexes with the composition MnL2n
.
They are regarded as the largest
class of molecules with unambiguous structures that have ever been synthesized.
The topology of MnL2n self-assembled cages can be described as polyhedra, where
the Pd2+ ions are regarded as vertices and the ligand as edges. Because Pd2+ has
Chem 1, 91–101, July 7, 2016 ª 2016 Elsevier Inc.
91

four coordination sites with square planar geometry, self-assembled MnL2n poly-
hedra are consequently limited to equilateral and tetravalent polyhedra. Isotropic
(ideally spherical) structures are favored because of the minimization of surface en-
ergy, resulting in the formation of Platonic and Archimedean solids. Given these
geometric constraints, only five such structures can be theoretically obtained and
are illustrated in the road map (Figure 1A), namely MnL2n polyhedral with n = 6,
12, 24, 30, and 60. So far, we have succeeded in synthesizing the M6L12 octahedron
(n = 6),29 M12L24 cuboctahedron (n = 12),30 and M24L48 rhombicuboctahedron
(n = 24).31
As the next logical step, we aimed to synthesize the M30L60 (n = 30) polyhedron.
However, M30L60 was not a simple extension of the previous ones (n % 24) for the
following reasons. First, the effect of kinetic trapping during the self-assembly pro-
cess is no longer negligible. This effect was not a critical problem when n = 6, 12, or
24 because intermediate, incomplete structures can still form and unform reversibly.
As the number of components participating in the self-assembly increases, however,
the formation of more meta-stable intermediates becomes possible, at least some of
which are kinetically trapped products. We propose that this type of mechanism
involving kinetically trapped intermediates is reminiscent of the folding funnel hy-
pothesis for protein folding.32 Second, the symmetry group is different. Even though
all of the previously synthesized polyhedra with n = 6, 12, or 24 have cubic symmetry,
polyhedra with n = 30 or 60 adopt dodecahedral symmetry that contains pentagon
faces. As shown in Figure 1B, the pentagon results in two different and ill-balanced
interior angles, 60ꢁ and 108ꢁ. Thus, rational ligand design to allow for this structural
difference is the key to obtaining self-assembled structures with n > 24. In this study,
we addressed the foregoing scientific hurdles with a finely designed flex-tuned
ligand, which allowed us to obtain the M30L60 icosidodecahedral self-assembled
metal complex, the structure of which was determined by single-crystal X-ray struc-
tural analysis. The polyhedron consists of precisely 90 subcomponents, the diameter
of which reaches 8.2 nm, which is comparable in size with biological macromolecules
such as proteins.
RESULTS AND DISCUSSION
Ligand Design and Self-Assembly
When designing MnL2n self-assembled systems, we have long focused on the ligand
bend angles as a guiding principle.33 In theory, the ideal bend angle q can be
calculated mathematically for any target polyhedral, for example, q = ꢀ90ꢁ for
M6L4, q = ꢀ120ꢁ for M12L24, and q = ꢀ135ꢁ for M24L48. For n % 24,29,30,34 this guiding
principle works well. However, we found that the self-assembly of thiophene-
1Department of Applied Chemistry, Graduate
School of Engineering, The University of Tokyo,
7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan
cored ligand (1, q = 149ꢁ), even with a q value close to the ideal bend angle for
2PRESTO (Precursory Research for Embryonic
Science and Technology), Japan Science and
Technology Agency, 7-3-1 Hongo, Bunkyo-ku,
Tokyo 113-0033, Japan
31
M30L60 (q = 150ꢁ), still formed M24L48 rather than M30L60
.
This observation sug-
gests that the effect of kinetic trapping becomes dominant at n > 24; the M24L48
complex 2 from ligand 1 is likely a kinetically trapped structure in an energy local
minimum close to M30L60 on the self-assembly potential profile.33
3Advanced Institute for Materials Research,
Tohoku University, 2-1-1 Katahira, Aoba-ku,
Sendai 980-8577, Japan
4Japan Synchrotron Radiation Research Institute
(JASRI/SPring-8), 1-1-1 Kouto, Sayo-cho,
Sayo-gun, Hyogo 679-5198, Japan
We thus sought a way to bring the trapped structure out of the local minimum. After
2 was prepared by mixing 1 and Pd(NO3)2 at 70ꢁC for 1 hr in dimethylformamide
(DMF) (Figure 2A), the DMF solution was further heated at 70ꢁC for 12 hr. Mass spec-
trometry showed the formation of the M30L60 species with significant intensity in
the spectrum (Figure 2B). This observation confirms the idea that the M24L48
complex 2 is the kinetic species prior to assembling into the thermodynamic product
M30L60. Prolonged heating of the solution did not change the M24L48/M30L60 ratio.
5Co-first author
6Lead Contact
*Correspondence:
mfujita@appchem.t.u-tokyo.ac.jp
http://dx.doi.org/10.1016/j.chempr.2016.06.007
92
Chem 1, 91–101, July 7, 2016

Figure 1. Schematic Representation of MnL2n Self-Assembly
(A) The family of MnL2n polyhedra. Metals (M) and bridging ligands (L) are mapped onto the vertices
and edges of the polyhedral, respectively.
(B) An enlarged view around a vertex of the icosidodecahedron (M30L60). The two different interior
angles, 60ꢁ and 108ꢁ, are ill balanced with respect to the Pd2+ square planar coordination
geometry.
Instead, partial decomposition of the complexes was observed, and we were unable
to isolate and characterize the M30L60 complex by other spectroscopic or crystallo-
graphic methods.
To suppress the formation of M24L48 and facilitate the self-assembly of M30L60, we
reasoned that a strategy to accommodate pentagon faces into the assembly process
is of significant importance. The designed ligand should have flexibility to avoid ill-
balanced interior angles, but flexible ligands are more likely to be kinetically trapped
in smaller structures with fewer components. In order to balance the two competing
factors, we tuned the ligand flexibility by chemically varying the spacer moiety, i.e.,
the connector between the center thiophene and the side pyridyl groups.
Ligand 3a, equipped with an ethynylene spacer group, was first examined. When this
ligand was treated with 0.51 equivalent of [Pd(MeCN)4](BF4)2 in DMSO-d6 at 70ꢁC for
2 hr, peak broadening and downfield shifts of the aromatic protons were observed in
the 1H nuclear magnetic resonance (NMR) spectrum (Figure S23), a commonly
observed phenomenon in the formation of large structures.31 However, the diffusion
coefficient values of self-assembled products, evaluated by 1H diffusion ordered
spectroscopy (DOSY) NMR experiments, suggested that the self-assembled prod-
ucts did not converge into a single species. Instead, the reaction gave a mixture
of products with diffusion coefficients consistent with M12L24 or even smaller self-
assembled products.35 This observation implied that this ligand, being too flexible,
favors the formation of kinetic products yielding smaller polyhedral structures by dis-
torting the ligand frame to an arch shape.36
Chem 1, 91–101, July 7, 2016
93

Figure 2. Self-Assembly of MnL2n Complexes from Thiophene-Cored Ligand 1 and Pd2+ Ion
Reaction scheme (A) and mass spectrometric data (B) after heating the components at 70ꢁC for 1 hr
(top) and subsequently for 12 hr (bottom) in DMF.
We next focused on more rigid phenylene-based spacers. Ligands 3b–3d, shown in
Figure 3, were newly designed. Unfortunately, ligand 3b, with a simple phenylene
spacer, was not suitable for experiments because of its low solubility. For ligands
3c and 3d, the introduction of methyl groups sterically disturbs the p-conjugation
of the planar conformation and thus enhances their solubility in DMF. Whereas
ligand 3c only affords a complex mixture (Figure S25), the tetramethylphenylene
spacer 3d showed a promising NMR spectrum when heated with 0.51 equivalent
of Pd(BF4)2(CH3CN)4 in DMF-d7 at 70ꢁC for 12 hr (Figure 4). Three broad peaks
were observed in the 1H NMR spectrum in the range of 10–6 parts per million
(ppm) (Figure 4C). The peaks at 9.8, 7.8, and 7.0 ppm were assigned to PyHa,
PyHb, and thiophene proton, respectively. Downfield shifts of PyHa (Dd = 1.04
ppm) and PyHb (Dd = 0.49 ppm) indicated Pd2+ coordination, and the broad and
simple signals were consistent with the formation of a very large structure with
high symmetry. In this spectrum, all of the pyridine rings seemed to be equivalent,
which suggested the formation of the M12L24 cuboctahedron or the M30L60 icosido-
1
decahedron by symmetry considerations. H DOSY NMR also indicated the forma-
tion of a single product with a diffusion coefficient (D) of 6.6 3 10ꢂ11 m2 sꢂ1
(logD = ꢂ10.18) at 300 K in DMF-d7 (Figure 4D). The observed D value is much
smaller than that of the corresponding M12L24 complex (see Supplemental Experi-
mental Procedures), an observation that is also consistent with the construction of
the first M30L60 complex 4.
The M30L60 composition was confirmed by cold-spray ionization time-of-flight mass
spectrometry (CSI-TOF MS)37 (Figure S9). A series of prominent peaks of [4 – (BF4)m
+
(DMF)n] (with m values from 20 to 30 and n values from 0 to 8) was observed, and no
other peaks with significant intensity were detected. Although the resolution of the
spectrum was not that high and we could not clearly distinguish their isotope
94
Chem 1, 91–101, July 7, 2016

S
S
N
N
N
N
N
N
3a
3b
CH3
H3C
H3C
H3C
CH3
CH3
S
S
N
N
CH3
H3C
CH3
H3C
H3C
CH3
3c
Figure 3. Newly Designed Flexible Ligands 3a–3d
3d
distributions, it is rather surprising that such a giant M30L60 complex was stable
enough to be detected by mass spectrometry at all.
X-Ray Crystallographic Analysis
The 3D coordination geometry of the M30L60 was clearly demonstrated by X-ray crys-
tallographic analysis (Figure 5). After years of attempts, we found that a single crystal
of 4 could be grown under extremely slow vapor diffusion (2–3 months) of AcOiPr
into a DMSO solution of 4. Although the crystals obtained were mechanically fragile
and unstable upon solvent evaporation, we overcame this hurdle by improving our
crystal-handling techniques (e.g., handling the crystals under refrigerated tempera-
tures). Since crystals of 4 contained significantly larger volumes of solvent (revealed
to be 90% of volume after analysis), small structured mosaics within the crystals
tended to form upon freezing. Thus, we opted to collect data at room temperature,
as it yielded better results with lower mosaicity than when collected at 100 K. In the
˚
end, we obtained high-quality data up to 2.0 A resolution by synchrotron X-ray radi-
ation (BL38B1 and BL41XU at SPring-8 and BL1A at KEK PF-AR). The diffraction data
resembled those of protein crystals rather than conventional metal complexes.
ꢁ
˚
˚
˚
Large unit cell volume (a = 122.3 A, b = 78.3 A, c = 78.8 A, and b = 121.8 ) and
extremely large void space (90%) made the crystallographic data unique. The typical
procedure of small-molecule X-ray analysis and statistical criteria no longer worked.
Therefore, the model building was carried out through ligand/solvent model align-
ment on the 2Fo – Fc map and the maximum entropy method38,39 electron-density
map in a manner analogous to protein crystallography. The refinement model of
the spherical complex was calculated by the CGLS command in SHELX,40 instead
of LS, which is typically used for small molecules.
The X-ray structure shows a flawless icosidodecahedral geometry as designed
(Figure 5). The cloudless electron-density map surely proves the existence and the
connectivity of the icosidodecahedral structure (Figure 5A). Although bond lengths
and angles do not perfectly conform to the abstract geometric mathematical struc-
ture, as might be expected, the overall topology of the structure is perfectly consis-
tent. Because of its flexibility, the framework assembled by ligand 3d is slightly
sagged (Figure 5B) so that all Npyridine-Pd(II)-Npyridine angles involved in triangle
and pentagon faces are adjusted to be around 90ꢁ, an ideal angle for Pd(II) square
planar geometry (Figure 5C). The diameter of the circumscribed sphere abstracted
around the complex is 8.2 nm, and the inscribed sphere is 6.7 nm. It is notable
that solvents occupy 90.0% of the total cell volume. As these solvents are not struc-
tured in the void, no distinguishable electron-density peak was observed, except the
ones weakly coordinated to the aromatic rings of the ligands (Figure 5D). The
Chem 1, 91–101, July 7, 2016
95

Figure 4. Self-Assembly of M30L60 Icosidodecahedron
(A) The reaction scheme. Ligand 3d was treated with Pd2+ ion in a 2:1 ratio at 70ꢁC for 12 hr in
DMF-d7.
(B) 1H NMR spectrum of ligand 3d. Signals a, b, and c denote PyHa, PyHb, and thiophene protons,
respectively.
–
(C) 1H NMR spectrum of complex 4 (BF4 salt). Aromatic signals shifted to downfield and heavily
broadened.
(D) 1H DOSY spectrum of sphere 4 (500 MHz, DMF-d7, 300 K). The spectrum indicates a single
product with a diffusion coefficient (D) of 6.6 3 10ꢂ11 m2
s
ꢂ1 (logD = ꢂ10.18). All the NMR spectra
(500 MHz) were measured as DMF-d7 solution at 300 K.
coordination of DMSO to aromatic rings (lone-pair-p interactions) is a known phe-
nomenon.41 Most of the atoms other than oxygen were severely disordered in this
case. The sliced view of the spherical complex (Figure 5E) visually affirms the state-
ments mentioned above and that our model building is sufficient. Even though there
is no comparable example in the literature, considering such a huge void space
96
Chem 1, 91–101, July 7, 2016

Figure 5. X-Ray Crystallographic Analysis of Complex 2d
Model building was carried out on the 2Fo – Fc and maximum entropy method electron-density
map. Hydrogen atoms are not displayed for clarity.
(A) The entire view of the X-ray crystallographic structure of M30L60. A molecular structure with
flawless icosidodecahedral symmetry was unveiled.
(B–D) An enlarged view around a ligand of M30L60. The modeled ligand structure clearly fits the
observed electron density. (B) The ligand becomes slightly sagged upon sphere formation. (C) As a
result, the N-Pd-N angles involved in the triangular and pentagon faces are adjusted to be close to
Chem 1, 91–101, July 7, 2016
97

Figure 5. Continued
an ideal 90ꢁ. (D) Counter ions are ordered at the apical positions of the Pd(II) centers. Electron-
density peaks attributed to DMSO molecules, which are coordinated to the ligands’ aromatic rings
(lone-pair-p interactions). Only oxygen atoms were modeled because of their severe disordering.
(E) A sliced image of the complex with an electron-density map. No distinguishable electron-
density peak was observed in the void space.
(90%), we perceive that the R factor (25.6%) and free R factor (27.0%) based on the
observed structure factor Fo are quite reasonable.
Conclusions
In conclusion, we have succeeded in the self-assembly of an M30L60 spherical com-
plex 4 and revealed its icosidodecahedral structure by X-ray crystallography. The
exclusive formation of M30L60 was supported additionally by 1D 1H NMR, 1H
DOSY NMR, and CSI-MS analysis. Considering the fact that ligands other than 3d
did not afford M30L60, even with the identical bend angle, fine-tuning of the flexibility
was key for achieving the self-assembly of the M30L60 structure from a large number
of components. The outer diameter of 4 reached more than 8 nm, and its interior
3
˚
space was as large as 157,000 A , undoubtedly one of the largest classes of well-
defined synthetic molecules ever handled. The structure is highly promising as a
scaffold for encapsulation of various nanomaterials such as metal nanoparticles42
or large proteins.39,43
EXPERIMENTAL PROCEDURES
Synthesis of 3d
A mixture of 2,5-bis(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)thiophene (101 mg,
0.30 mmol), 4-(4-iodo-2,3,5,6-tetramethylphenyl)pyridine (253 mg, 0.75 mmol), [1,10-
bis(diphenylphosphino)ferrocene]dichloropalladium complex with dichloromethane
(31 mg, 0.038 mmol), and K2CO3 (414 mg, 3.0 mmol) in 1,2-dimethoxyethane (DME;
10 mL) and H2O (1.0 mL) was stirred at 110ꢁC for 14 hr under an Ar atmosphere. After
cooling to the ambient temperature, the reaction mixture was filtered through a Celite
pad, and the filtrate was evaporated in vacuo. The residue was purified by column chro-
matography on silica gel (ethyl acetate/hexane, 20:80 to 50:50) to give 3d (85 mg, 57%)
as a white solid.
Melting point: 279ꢁC–281ꢁC.
1H NMR (500 MHz, CDCl3, 300 K): d 8.68 (d, J = 5.0 Hz, 4H), 7.21 (d, J = 5.0 Hz, 4H),
6.92 (s, 2H), 2.13 (s, 12H), 1.88 (s, 12H). 1H NMR (500 MHz, DMF-d7, 300 K): d 8.74 (d,
J = 6.0 Hz, 4H), 7.27 (d, J = 6.0 Hz, 4H), 6.99 (s, 2H), 2.21 (s, 12H), 1.96 (s, 12H). Diffu-
sion coefficient (500 MHz, DMF-d7, 300 K): logD = ꢂ9.3, D = 5.01 3 10ꢂ10 m2 sꢂ1
.
13C NMR (125 MHz, CDCl3, 300 K): d 150.9 (C), 150.1 (CH), 143.0 (C), 139.6 (C), 134.6
(C), 134.0 (C), 131.1 (C), 126.6 (CH), 124.9 (CH), 18.1 (CH3), 18.0 (CH3).
IR: 2926, 1594, 1538, 1451, 1407, 1375, 1283, 1259, 1212, 1069, 1011, 979, 863,
804 cmꢂ1
.
High-resolution mass spectrometry (electron-spray ionization time-of-flight): calcu-
lated for m/z C34H34N2S [M + H]+ 503.2531, found 503.2515.
Elemental analysis: calculated for C34H34N2S C, 81.23; H, 6.82; N, 5.57, found C,
80.94; H, 6.82; N, 5.57.
98
Chem 1, 91–101, July 7, 2016

Self-Assembly of 4
To a solution of ligand 3d (5.06 mg, 10 mmol) in DMF-d7 (1.01 mL) was added a
solution of Pd(BF4)2(CH3CN)4 in DMF-d7 (0.10 mL, 52.3 mM, 5.2 mmol), and the
resulting mixture was stirred at 70ꢁC for 12 hr. 1H NMR spectroscopy confirmed
the quantitative formation of the sphere complex 4.
1H NMR (500 MHz, DMF-d7, 300 K): d 9.75 (br, 240H), 7.75 (br, 240H), 6.98 (br, 120H),
2.13 (br, 720H), 1.89 (br, 360H), 1.69 (br, 360H); these two signals showed coales-
cence at 320 K. Diffusion coefficient (500 MHz, DMF-d7, 300 K): logD = ꢂ10.3,
D = 5.01 3 10ꢂ11 m2 sꢂ1 13C NMR (150 MHz, DMF-d7, 300 K): d 156.4, 152.9,
.
143.7, 139.1, 135.5, 131.6, 129.6, 128.5, 18.5. The 13C NMR spectrum was difficult
to measure because of the severe broadness of the signals in DMF-d7, and we finally
obtained the spectrum after 232,000 acquisition attempts on a 600 MHz NMR
system equipped with a gradient cold probe (53040HCNVC).
Crystallographic Analysis of 4
Colorless crystals of M30L60 were obtained by vapor diffusion of isopropyl acetate
into a DMSO solution of 4 (with tetrafluoroborate as the counter anion). The
preliminary diffraction studies were performed at the BL41XU beamline at SPring-8
(Rayonix MX225H&E CCD detector) and the BL1A beamline at KEK PF-AR (Dectris
PILATUS 2M-F PAD detector) at 296 K. The final diffraction data were measured at
the BL38B1 beamline at SPring-8 (Rayonix MX225H&E CCD detector). HKL200044
was used for the processing and data reduction in the C2 space group. The structure
was solved and refined by SHELX40 in the C2/m space group. These data did not
have a high enough resolution for calculation of the hydrogen positions. The crystal
lattice contained huge voids (90.0%) with a lot of disordered solvents and anions.
˚
Because of the very moderate diffraction power of the crystal (resolution 2.0 A), sol-
vent molecules could not be located as realistic molecular structures, and instead,
oxygen atoms (DMSO) were located for the residual electron density with full or
3
˚
partial occupancies until a plateau of ca. 0.6 e/A was reached in the maximum en-
tropy density map. The ligands showed very large thermal motion (maybe as a result
of positional disorder, which however, could not be modeled), and thus a multitude
of bond distance and thermal parameter restraints (DFIX, DANG, and FLAT) to the
ligands, counter anions, and Pd-N distances had to be applied in order to avoid
chemically unreasonable bond distances and angles. Pd atoms were refined aniso-
tropically, and all other atoms were refined isotropically.
ACCESSION NUMBERS
The M30L60 structure reported in this article has been deposited in the Cambridge
Crystallographic Data Centreunder accession number CCDC: 1482268.
SUPPLEMENTAL INFORMATION
Supplemental Information includes Supplemental Experimental Procedures, 26 fig-
ures, one table, and one crystallographic data file and can be found with this article
online at http://dx.doi.org/10.1016/j.chempr.2016.06.007.
AUTHOR CONTRIBUTIONS
Conceptualization, M.F.; Methodology, D.F. and Y.U.; Investigation, D.F., Y.U., S.S.,
and H.Y.; Formal Analysis, D.F., S.S., N.M., and T.K.; Writing – Original Draft, D.F.
and Y.U.; Review & Editing, D.F. and M.F.; Funding Acquisition, D.F. and M.F.; Su-
pervision, M.F.
Chem 1, 91–101, July 7, 2016
99

ACKNOWLEDGMENTS
This work was supported by the JST-ACCEL program and the PRESTO program,
partly supported by KAKENHI (25102007). The synchrotron X-ray crystallography
was performed at the BL1A beamline at the KEK Photon Factory (2015G097) and
at the BL38B1 and BL41XU beamlines at SPring-8 (2014A0042 and 2015B0120).
Received: April 27, 2016
Revised: June 1, 2016
Accepted: June 9, 2016
Published: July 7, 2016
REFERENCES AND NOTES
1. Lawson, D.M., Artymiuk, P.J., Yewdall, S.J.,
Smith, J.M., Livingstone, J.C., Treffry, A.,
Luzzago, A., Levi, S., Arosio, P., and Cesareni,
G. (1991). Solving the structure of human H
ferritin by genetically engineering
self-assembly of finite two- and three-
dimensional ensembles. Chem. Rev. 111,
6810–6918.
nanoscale cuboctahedra by coordination
chemistry. Nature 398, 796–799.
23. Moulton, B., Lu, J., Mondal, A., and Zaworotko,
M.J. (2001). Nanoballs: nanoscale faceted
polyhedra with large windows and cavities.
Chem. Commun. 863–864.
13. Young, N.J., and Hay, B.P. (2013). Structural
design principles for self-assembled
coordination polygons and polyhedra. Chem.
Commun. 49, 1354.
intermolecular crystal contacts. Nature 349,
541–544.
24. Eddaoudi, M., Kim, J., Wachter, J.B., Chae,
H.K., O’Keeffe, M., and Yaghi, O.M. (2001).
Porous metal-organic polyhedra: 25 A
cuboctahedron constructed from 12 Cu2(CO2)
4 paddle-wheel building blocks. J. Am. Chem.
Soc. 123, 4368–4369.
2. Stagg, S.M., Gurkan, C., Fowler, D.M.,
¨
14. Smulders, M.M.J., Riddell, I.A., Browne, C., and
Nitschke, J.R. (2013). Building on architectural
principles for three-dimensional
LaPointe, P., Foss, T.R., Potter, C.S., Carragher,
B., and Balch, W.E. (2006). Structure of the
Sec13/31 COPII coat cage. Nature 439,
234–238.
metallosupramolecular construction. Chem.
Soc. Rev. 42, 1728–1754.
3. Stagg, S.M., LaPointe, P., Razvi, A., Gurkan, C.,
¨
Potter, C.S., Carragher, B., and Balch, W.E.
(2008). Structural basis for cargo regulation of
COPII coat assembly. Cell 134, 474–484.
15. Saalfrank, R.W., Stark, A., Peters, K., and
Schnering von, H.G. (1988). The first
25. Li, J.-R., and Zhou, H.-C. (2010). Bridging-
ligand-substitution strategy for the preparation
of metal-organic polyhedra. Nat. Chem. 2,
893–898.
‘‘adamantoid’’ alkaline earth metal chelate
complex: synthesis, structure, and reactivity.
Angew. Chem. Int. Ed. Engl. 27, 851–853.
4. Grimes, J.M., Burroughs, J.N., Gouet, P.,
Diprose, J.M., Malby, R., Zie´ ntara, S., Mertens,
P.P., and Stuart, D.I. (1998). The atomic
structure of the bluetongue virus core. Nature
395, 470–478.
26. Pasquale, S., Sattin, S., Escudero-Ada´n, E.C.,
Martı´nez-Belmonte, M., and de Mendoza, J.
(2012). Giant regular polyhedra from calixarene
carboxylates and uranyl. Nat. Commun. 3, 785.
16. Fujita, M., Oguro, D., Miyazawa, M., Oka, H.,
Yamaguchi, K., and Ogura, K. (1995). Self-
assembly of ten molecules into nanometre-
sized organic host frameworks. Nature 378,
469–471.
5. Zhang, X., Boyce, M., Bhattacharya, B., Zhang,
X., Schein, S., Roy, P., and Zhou, Z.H. (2010).
Bluetongue virus coat protein VP2 contains
sialic acid-binding domains, and VP5
resembles enveloped virus fusion proteins.
Proc. Natl. Acad. Sci. USA 107, 6292–6297.
27. Bilbeisi, R.A., Ronson, T.K., and Nitschke, J.R.
(2013). A self-assembled [FeII12L12] capsule
with an icosahedral framework. Angew. Chem.
Int. Ed. Engl. 52, 9027–9030.
17. MacGillivray, L.R., and Atwood, J.L. (1997). A
chiral spherical molecular assembly held
together by 60 hydrogen bonds. Nature 389,
469–472.
28. Harris, K., Fujita, D., and Fujita, M. (2013). Giant
hollow MnL2n spherical complexes: structure,
functionalisation and applications. Chem.
Commun. 49, 6703.
6. Steed, J., and Gale, P., eds. (2012).
Supramolecular Chemistry: From Molecules to
Nanomaterials (John Wiley).
18. Stang, P.J., Olenyuk, B., Muddiman, D.C., and
Smith, R.D. (1997). Transition-metal-mediated
rational design and self-assembly of chiral,
nanoscale supramolecular polyhedra with
unique T symmetry. Organometallics 16, 3094–
3096.
29. Suzuki, K., Tominaga, M., Kawano, M., and
Fujita, M. (2009b). Self-assembly of an M6L12
coordination cube. Chem. Commun. 1638–
1643.
7. Conn, M.M., and Rebek, J. (1997). Self-
assembling capsules. Chem. Rev. 97, 1647–
1668.
19. Caulder, D.L., Powers, R.E., Parac, T.N., and
Raymond, K.N. (1998). The self-assembly of a
predesigned tetrahedral M4L6 supramolecular
cluster. Angew. Chem. Int. Ed. Engl. 37, 1840–
1843.
8. Caulder, D.L., and Raymond, K.N. (1999).
Supermolecules by design. Acc. Chem. Res. 32,
975–982.
30. Tominaga, M., Suzuki, K., Kawano, M.,
Kusukawa, T., Ozeki, T., Sakamoto, S.,
Yamaguchi, K., and Fujita, M. (2004). Finite,
spherical coordination networks that self-
organize from 36 small components. Angew.
Chem. Int. Ed. Engl. 43, 5621–5625.
9. Fujita, M., Tominaga, M., Hori, A., and
Therrien, B. (2005). Coordination assemblies
from a Pd(II)-cornered square complex. Acc.
Chem. Res. 38, 369–378.
20. Abrahams, B.F., Egan, S.J., and Robson, R.
(1999). A very large metallosupramolecular
capsule with cube-like 43 topology
assembled from twelve Cu(II) centers and
eight tri-bidentate tri-anionic ligands
derived from 2,4,6-triphenylazo-1,3,5-
trihydroxybenzene. J. Am. Chem. Soc. 121,
3535–3536.
31. Sun, Q.F., Iwasa, J., Ogawa, D., Ishido, Y., Sato,
S., Ozeki, T., Sei, Y., Yamaguchi, K., and Fujita,
M. (2010). Self-assembled M24L48 polyhedra
and their sharp structural switch upon subtle
ligand variation. Science 328, 1144–1147.
10. Saalfrank, R.W., Maid, H., and Scheurer, A.
(2008). Supramolecular coordination chemistry:
the synergistic effect of serendipity and rational
design. Angew. Chem. Int. Ed. Engl. 47, 8794–
8824.
32. Pain, R.H., ed. (1994). Mechanisms of Protein
Folding (IRL Press).
21. Takeda, N., Umemoto, K., Yamaguchi, K., and
Fujita, M. (1999). A nanometre-sized
11. Forgan, R.S., Sauvage, J.-P., and Stoddart, J.F.
(2011). Chemical topology: complex molecular
knots, links, and entanglements. Chem. Rev.
111, 5434–5464.
33. Fujita, D., Yokoyama, H., Ueda, Y., Sato, S., and
Fujita, M. (2015). Geometrically restricted
intermediates in the self-assembly of an M12l24
cuboctahedral complex. Angew. Chem. Int.
Ed. Engl. 54, 155–158.
hexahedral coordination capsule assembled
from 24 components. Nature 398, 794–796.
12. Chakrabarty, R., Mukherjee, P.S., and Stang,
P.J. (2011). Supramolecular coordination:
22. Olenyuk, B., Whiteford, J.A., Fechtenkotter, A.,
¨
and Stang, P.J. (1999). Self-assembly of
100 Chem 1, 91–101, July 7, 2016

34. Yokoyama, H., Ueda, Y., Fujita, D., Sato, S., and
Fujita, M. (2015). Finely resolved threshold for
the sharp M12L24/M24L48 structural switch in
multi-component MnL2n polyhedral
assemblies: X-ray, MS, NMR, and
37. Yamaguchi, K. (2003). Cold-spray ionization
mass spectrometry: principle and applications.
J. Mass Spectrom. 38, 473–490.
41. Zukerman-Schpector, J., and Tiekink, E.R.T.
(2014). On the role of DMSO-O(lone pair)/
p(arene), DMSO-S(lone pair)/p(arene)
and SO/p(arene) interactions in the
crystal structures of dimethyl sulphoxide
(DMSO) solvates. CrystEngComm 16, 6398–
6407.
38. Takata, M. (2008). The MEM/Rietveld method
with nano-applications – accurate charge-
density studies of nano-structured materials by
synchrotron-radiation powder diffraction. Acta
Crystallogr. A 64, 232–245.
ultracentrifugation analyses. Chem. Asian J. 10,
2292–2295.
42. Suzuki, K., Sato, S., and Fujita, M. (2009).
Template synthesis of precisely monodisperse
silica nanoparticles within self-assembled
organometallic spheres. Nat. Chem. 2, 25–29.
35. Sato, S., Iida, J., Suzuki, K., Kawano, M., Ozeki,
T., and Fujita, M. (2006). Fluorous nanodroplets
structurally confined in an organopalladium
sphere. Science 313, 1273–1276.
39. Fujita, D., Suzuki, K., Sato, S., Yagi-Utsumi, M.,
Yamaguchi, Y., Mizuno, N., Kumasaka, T.,
Takata, M., Noda, M., Uchiyama, S., et al.
(2012). Protein encapsulation within synthetic
molecular hosts. Nat. Commun. 3, 1093–1097.
43. Fujita, D. (2014). Challenges to large molecular
36. Fujita, D., Takahashi, A., Sato, S., and Fujita, M.
(2011). Self-assembly of Pt(II) spherical
complexes via temporary labilization of
the metal-ligand association in 2,2,2-
trifluoroethanol. J. Am. Chem. Soc. 133,
13317–13319.
encapsulation. Pure Appl. Chem. 1, 3–11.
44. Otwinowski, Z., and Minor, W. (1997).
Processing of X-ray diffraction data collected in
oscillation mode. Methods Enzymol. 276,
307–326.
40. Sheldrick, G.M. (2015). Crystal structure
refinement with SHELXL. Acta Crystallogr. C
Struct. Chem. 71, 3–8.
Chem 1, 91–101, July 7, 2016 101