Self-recognition of structurally identical, rod-shaped macroions with different central metal atoms during their assembly process

Article abstract of DOI:10.1021/ja400656j

Two rod-shaped macroanions, ((C₄H₉) ₄N)₇[Mo₆O₁₈NC(OCH₂) ₃XMo₆O₁₈(OCH₂)₃CNMo ₆O₁₈] (X = Mn^III (1), Fe^III (2)), with almost identical charge densities and morphologies except for their different encapsulated central metal atoms were each observed to self-assemble into blackberry -type supramolecular structures in their dilute solution, driven by the counterion-mediated attraction. Amazingly, the two macroions remained self-sorted and self-assembled into homogeneous assemblies in their mixed solutions, demonstrating a self-recognition behavior between two highly similar macroions during their assembly process, as confirmed by DLS, SLS, and TEM/EDS analysis. This self-recognition behavior can be explained by the slightly different charge distributions of the macroanions resulting from their different central atoms (confirmed by theoretical DFT calculations and dissociation experiments) and the high activation energy of the slow assembly process, which suppresses the formation of hybrid oligomers at the beginning of the self-assembly process. This work confirms that the long-range counterion-mediated electrostatic attraction is sensitive to the small difference in macroions and consequently offers the possibility for delicate selectivity and preference among different macroions. This phenomenon might be directly related to (and be the important reason for) some recognition behaviors in biological systems.

Full text of DOI:10.1021/ja400656j

Article
pubs.acs.org/JACS
Self-Recognition of Structurally Identical, Rod-Shaped Macroions
with Different Central Metal Atoms during Their Assembly Process
Panchao Yin,† Jin Zhang,‡ Tao Li,§ Xiaobing Zuo,§ Jian Hao,‡ Anna Marie Warner,†
Soma Chattopadhyay,⊥,# Tomohiro Shibata,⊥,# Yongge Wei,*,‡ and Tianbo Liu*,†,∥
†Department of Chemistry, Lehigh University, Bethlehem, Pennsylvania 18015, United States
‡Department of Chemistry, Tsinghua University, Beijing 100084, P.R. China
§X-ray Science Division, Advanced Photon Source, Argonne National Laboratory, Argonne, Illinois 60439, United States
⊥CSRRI-IIT, MRCAT, Sector 10, Building 433B, Argonne National Laboratory, Argonne, Illinois 60439, United States
#Physics Department, Illinois Institute of Technology, Chicago, Illinois 60616, United States
∥Department of Polymer Science, The University of Akron, Akron, Ohio 44325, United States
S
* Supporting Information
ABSTRACT: Two rod-shaped macroanions, ((C4H9)4N)7-
[Mo6O18NC(OCH2)3XMo6O18(OCH2)3CNMo6O18] (X =
MnIII (1), FeIII (2)), with almost identical charge densities
and morphologies except for their different encapsulated
central metal atoms were each observed to self-assemble into
“blackberry”-type supramolecular structures in their dilute
solution, driven by the counterion-mediated attraction.
Amazingly, the two macroions remained self-sorted and self-
assembled into homogeneous assemblies in their mixed
solutions, demonstrating a self-recognition behavior between
two highly similar macroions during their assembly process, as
confirmed by DLS, SLS, and TEM/EDS analysis. This self-
recognition behavior can be explained by the slightly different
charge distributions of the macroanions resulting from their different central atoms (confirmed by theoretical DFT calculations
and dissociation experiments) and the high activation energy of the slow assembly process, which suppresses the formation of
hybrid oligomers at the beginning of the self-assembly process. This work confirms that the long-range counterion-mediated
electrostatic attraction is sensitive to the small difference in macroions and consequently offers the possibility for delicate
selectivity and preference among different macroions. This phenomenon might be directly related to (and be the important
reason for) some recognition behaviors in biological systems.
bases of diverse phenomena such as cell signaling, the immune
response, and gene regulatory interaction.15−18
INTRODUCTION
■
Non-covalent interactions, e.g., hydrogen bonding,1 π−π
stacking,2 coordination,3 static charge interaction,4 and
solvophobic interaction,5 are responsible for various self-
assembly and self-recognition behaviors in physicochemical
and biological processes.6−8 In particular, counterion-mediated
like-charge attraction and water-bridged hydrogen bonding
were found to play significant roles in the folding, assembly,
and recognition behavior of proteins, RNA, and DNA.9,10
Complex biomolecules rarely function in isolation; hence, a
thorough understanding of a biological process is highly
dependent upon an examination of assemblies of biomolecules
and the interactions driving the assembly process.11−13 When
two different assemblies are in the same solution, they may
coexist and remain self-sorted, or they may hybridize.13,14 Self-
recognition phenomena during the self-assembly of biomole-
cules, in many cases a natural feature of biomacromolecules,
contribute to the formation of biomolecular complexes with
uniqueness and specific functionality, which are the chemical
An approach to understand the self-recognition processes
related to non-covalent interactions is to choose simplified
models where only limited types of interactions apply.
Nanoscaled polyoxometalate (POM) molecular macroions
can be considered as “inorganic polyelectrolytes” composed
of negatively charged metal oxide polyhedra with well-defined
molecular structures, uniform shapes, tunable charges, and
negligible intramolecular charge interaction, which are ideal
models for studying the intermolecular interaction of
polyelectrolytes.19−21 Our previous research enables us to
understand and control the counterion-mediated attraction
between the macroions through the study of the reversible self-
assembly of the macroions into stable, hollow, single-layered,
“blackberry”-type structures in polar solvents, with the only
Received: January 20, 2013
Published: February 27, 2013
© 2013 American Chemical Society
4529
dx.doi.org/10.1021/ja400656j | J. Am. Chem. Soc. 2013, 135, 4529−4536

Journal of the American Chemical Society
Article
Figure 1. Synthetic procedure for the two molecular rods, using ball−stick representation of the molecular structures of the building blocks and the
molecular rods. X, center heteroatom; A, hexamolybdate; B, the ring-shaped Anderson-type molecule.
driving forces being counterion-mediated attraction and
hydrogen bonding.20−28 Such behavior is the result of the
counterion−macroion interaction involving moderate counter-
ion association around macroions, a unique phenomenon
known only for nanoscaled hydrophilic ions.20,21,27,29 Very
recently, an interesting self-recognition process was observed in
the mixed dilute solution of two almost identical, 2.5-nm-size,
spherical polyprotic metal oxide-based macroions (abbrevia-
tions {Mo72Fe30} and {Mo72Cr30}, respectively).30 The
“Keplerate” clusters {Mo72Fe30} and {Mo72Cr30} both have
30 M(H2O) groups coordinated to the non-Mo metal atoms,
but different mobilities of their surface hydration layers as well
as different degrees of deprotonation. The two types of clusters
demonstrate self-recognition by forming two types of
homogeneous blackberries instead of mixed ones in their
mixed aqueous solution.30 The charge difference between the
two macroions (−7 and −5, respectively) is believed to be the
major reason, suggesting that the long-range, delicate electro-
static interaction is essential and critical for self-recognition
behavior.30 In contrast, the assemblies with close-contact
interactions (e.g., hydrophobic interactions) cannot achieve
this recognition; e.g., different types of surfactants usually form
hybrid micelles.31−33
It is interesting to notice that simple inorganic macroions can
demonstrate the level of self-recognition usually believed to
exist only among biomolecules.18 While the delicate electro-
static interaction is attributed to this behavior, it is critical to
test the universality of this behavior among various macroions
and, more importantly, to push the limit by studying highly
similar mixed macroions. Herein, we report the study of a
mixture of two highly charged rod-shaped macroions with
almost identical structures and charge densities, differing only
in their encapsulated central metal atoms. Time-resolved static
and dynamic light scattering, transmission electron microscopy,
and energy-dispersive X-ray spectroscopy were utilized to
analyze the self-recognition behavior between the two
molecular rods. Density functional theory calculation and
kinetic monitoring of the disassembly process of the formed
blackberry structures in the individual solutions of the two
molecular rods and their mixture solution, respectively, were
combined to uncover the mechanism for their self-recognition
behavior.
EXPERIMENTAL SECTION
■
Static Light Scattering (SLS). A commercial Brookhaven
Instrument LLS spectrometer equipped with a solid-state laser
operating at 532 nm was used for measurement of both SLS and
DLS. SLS experiments were performed at scattering angles θ between
20 and 120°, at 2° intervals. However, due to the large fluctuations in
scattered intensities at low scattering angles, we removed the data from
20 to 40° in the final analysis. Derived from the Rayleigh−Gans−
Debye equation,34 a partial Zimm plot was used to analyze the SLS
data to obtain the radius of gyration, Rg. The partial Zimm plot stems
from the following approximate formula: 1/I = C(1 + Rg2q2/3). Here
Rg is determined from the slope and intercept of a plot of 1/I vs q2.
Dynamic Light Scattering (DLS). DLS measures the intensity−
intensity time correlation function by means of a BI-9000AT
multichannel digital correlator. The field correlation function |g(1)(τ)|
was analyzed by the constrained regularized CONTIN method35 to
yield information on the distribution of the characteristic line width Γ
from |g(1)(τ)| = ∫ G(Γ) e−Γτ dΓ.The normalized distribution function
of the characteristic line width, G(Γ), so obtained, can be used to
determine an average apparent translational diffusion coefficient, Dapp
= Γ/q2. The hydrodynamic radius Rh is related to D via the Stokes−
Einstein equation, Rh = kT/6πηD, where k is the Boltzmann constant
and η the viscosity of the solvent at temperature T. From DLS
measurements, we can obtain the particle-size distribution in solution
from a plot of ΓG(Γ) versus Rh. The Rh of the particles is obtained by
extrapolating Rh,app to zero scattering angle.
Transmission Electron Microscopy (TEM) and Energy-
Dispersive X-ray Spectroscopy (EDS). The TEM images were
taken on a JEOL JEM-2000 electron microscope operated at 200 kV
with EDS attachment (Oxford). Samples for the TEM analysis were
prepared by dropping a small volume of the solution sample onto a
carbon film on a copper grid.
1
1D NMR and 2D Diffusion Ordered H NMR Spectroscopy
(DOSY). All the NMR spectra were recorded on a Bruker Avance 500
1
spectrometer equipped with a BBO probe at 25 °C. For H NMR
experiments, 2.0 mg of compound 1 and 1.1 mg of TBA*I were each
dissolved in 1 mL of acetone-d6.
DOSY was performed on a Bruker 500 MHz spectrometer with the
magnetic field gradient g varying from 0 to 32 G/cm in 16−32 steps.
The length of the gradient, d, was from 6000 to 8000 ms, and the time
interval between two pulsed gradients, D, was from 0.1 to 0.15 s. All
spectra were taken at room temperature. After the data collection,
4530
dx.doi.org/10.1021/ja400656j | J. Am. Chem. Soc. 2013, 135, 4529−4536

Journal of the American Chemical Society
Article
Figure 2. (a) Superimposition of experimental solution small-angle (red circles) and wide-angle (green circles) X-ray scattering curves of 1 in
acetonitrile (2 mg/mL) and simulated scattering curve (in blue) using the program SolX with the molecular model (c). (b) PDDF of 1 in its
acetonitrile solution (red curve) obtained using the program GNOM, and PDDF calculated (green curve) from the molecular model (c) using SolX.
The broadening features in the PDDF obtained from GNOM could arise from the configurational ensemble in solution due to the free rotation of
the Mo−N−C bond. (c) The build molecular model for experimental data fitting.
FIDs were processed and analyzed with the NMR software TopSpin
2.0 provided by Bruker. Both T1/T2 relaxation and CONTIN
methods were used to fit the raw data. The observed proton signal I in
a standard DOSY spectrum is expressed through eq 1:
non-conjugative nature of the linkage between A and B units,
they retain their original charges of −2 and −3, respectively
(Figure 1). Molecular rods 1 and 2 have identical lengths and
diameters of ∼2.77 and 0.88 nm, respectively. The actual
structures of the molecular rods in solution (acetonitrile) have
been investigated through solution SAXS with a spatial
resolution of 2.9 Å.37 The two molecular rods show identical
SAXS curves (Figures 2a and S7 (Supporting Information)),
suggesting that they have identical molecular structure in
solution state. The calculated SAXS profile for the rod-shaped
molecular model fits well with the experimental data of 1,
indicating that the rod shape remains when the crystals of 1
dissolve in acetonitrile (Figure 2a).38 The radius of gyration of
the molecule, Rg = 0.73 nm, is calculated from the small-angle
region of SAXS using the Guinier equation. Further exploration
of the scattering data with the program GNOM generates pair
distance distribution functions (PDDFs) in real space, p(r).39
The PDDFs exhibit the features of rod-shaped molecular triads,
where the first peak centered at ∼0.25 nm describes the intra-
subunit (A or B) atom-pair distances, while the second peak at
∼0.90 nm and the third peak at ∼1.70 nm represent the
separations of A to B and A to A, respectively (Figure 2b,c).
Time-resolved solution X-ray scattering studies and scattering
measurements on aged samples suggest that the molecular rods
are quite stable, at least for 2 months (Figure S8). As a general
feature of the solution physical chemistry of macroions, the ion-
pairing of the molecular rods with the counterion, TBA, was
⎡
⎤
⎛
⎞
⎟
δ
3
2
2
⎜
I = I exp −(2πγδ) Δ −
Dg
⎢
⎥
⎦
0
⎝
⎠
⎣
(1)
where I0 is the reference intensity and γ the gyromagnetic ratio of the
proton. If only one diffusive component exists in the solution, a
straight line will occur in a plot of ln(I/I0) versus g2, and the apparent
diffusion coefficient D can be calculated from the slope using linear
regression analysis.
Small-Angle X-ray Scattering (SAXS). The SAXS experiments
were performed at station 12-ID-B with X-ray energy of 12 keV at the
Advanced Photon Source of the Argonne National Laboratory. The
sample-to-detector distance was about 2 m. A 2D CCD detector was
used to acquire images with typical exposure times around 1.0 s.
RESULTS AND DISCUSSION
■
Macroionic Features and Stability of the Molecular
Rods in Solution. Two molecular rods, ((C4H9)4N)7-
[Mo6O18NC(OCH2)3XMo6O18(OCH2)3CNMo6O18] (X =
MnIII (1), FeIII (2)), were synthesized and fully characterized
in previous literature.36 These two compounds were synthe-
sized by covalently linking two super-octahedron-shaped
hexamolybdates (A) onto the two sides of a ring-shaped
Anderson-type POM (B) (Figure 1). Basically, the two
molecular rods are identical except that the center heteroatom
of the ring-like molecule is different: MnIII for rod 1 and FeIII
for rod 2. Both molecular rods are highly charged (−7) with
tetrabutylammonium (TBA) as counterion, and hydrophilic
with the surface composed of molybdenum oxide. Due to the
1
1
investigated by H NMR and 2D diffusion-ordered H NMR
spectroscopy (DOSY).40 The four-proton peaks of TBA in
acetone-d6 solution of 1 were much broader than in acetone-d6
solution of tetrabutylammonium iodide (TBA*I), suggesting
4531
dx.doi.org/10.1021/ja400656j | J. Am. Chem. Soc. 2013, 135, 4529−4536

Journal of the American Chemical Society
Article
Figure 3. (a) CONTIN analysis of DLS results of the solutions of 1 (green dots), 2 (blue dots), and their mixture (red dots). (b,c) TEM images of
assemblies in the solutions of 1 and 2, respectively.
Figure 4. Self-assembly of molecular rods into blackberry structures (polyhedra code: blue, MoO6 or MoO5N; green, MnO6 or FeO6).
possible strong association of TBAs with large structures in 1’s
acetone solution (Figure S9). DOSY results indicated that
TBAs in the solution of 1 diffuse much more slowly (diffusion
coefficient D = 2.8 × 10−9 m2/s) than the “free” TBAs in
TBA*I solution (D = 4.4 × 10−9 m2/s), suggesting strong
association of TBAs with the molecular rods (Table S1). The
charge density of the molecular rods is −1.06 e/nm2, which is
higher than the range that might be expected for the
counterion-mediated static charge-regulated self-assembly of
macroions into blackberries.
Self-Assembly of the Molecular Rods into Black-
berries. The two rod-like molecules are quite soluble in
acetonitrile but do not self-assemble into large supramolecular
structures in dilute solution with the bulky and hydrophobic
TBA counterions, due to their high surface charge density.
Adding ZnCl2 to their acetonitrile solutions can trigger the self-
assembly process, since Zn2+ ions are more likely to interact
with rod-like macroanions than the TBA+ ions. This leads to
stronger counterion-mediated attraction and consequently
results in the self-assembly. Time-resolved SLS results indicated
that the scattered intensities at 90° scattering angle for both
types of rod-like macroionic solutions at room temperature
showed a linear increment with time and reached equilibrium at
∼40 days (Figure S1). DLS results indicated that the
hydrodynamic radii (Rh) of the assemblies in each individual
solution of two molecular rods did not change during the whole
self-assembly process (Figures S2 and S3). The combination of
DLS and SLS suggested that it was the number of the
assemblies, not the assembly size, that increased during the self-
assembly process. A typical CONTIN analysis from the DLS
study of individual solutions of the compounds indicated that
Rh of large assemblies showed no angular dependence, with
4532
dx.doi.org/10.1021/ja400656j | J. Am. Chem. Soc. 2013, 135, 4529−4536

Journal of the American Chemical Society
Article
narrow distribution and average values of 80 4 and 27
2
nm (Figure 3a), respectively, for the solutions of 1 and 2. SLS
study analyzed by Zimm plot indicated that the assemblies had
average Rg = 76 4 and 30 2 nm, respectively. The ratio of
Rh/Rg ≈ 1.0 suggests a hollow spherical structure for each of
the assemblies of 1 and 2, which is also confirmed by TEM
results (Figure 3b,c). The counterion (Zn2+)-mediated
attraction among the highly charged molecular rods is
responsible for the formation of the assemblies (blackberries),
which shows some obvious similarities with natural biomacro-
molecules, such as viral capsid proteins.41,42 Both the blackberry
and spherical viral capsid are nanoscaled single-layer hollow
shells, which are quite different from the vesicles of lipid
molecules (Figure 4). Moreover, the two assemblies demon-
strate similar kinetic features: rate-limiting oligomer formation
followed by rapid formation of the large assemblies from the
oligomers.41 The formed blackberry-type structures of 1 and 2
are of obviously different sizes, although the two macroanions
are almost identical in both shape and charge. It is believed that
the encapsulated central metal ions lead to distortion of the
Anderson-type POM framework and therefore affect the charge
distribution of unit B. The different charge distributions of the
two macroions play a vital role in the binding of added
counterion (Zn2+) on the surface of the macroions, which
finally determines the size of the assemblies. More details on
the mechanism will be discussed in the next section.
Self-Recognition Behavior of the Two Molecular
Rods. Impressively, the two inorganic molecular rods show
self-recognition behavior by self-assembling into two types of
individual blackberry structures instead of mixed ones in their
mixed solution. The macroionic shape, surface composition,
charge, and counterions are expected to play vital roles in the
self-assembly process, based on our previous research.21
However, for the two molecular rods, all of those factors look
identical, making the possible self-recognition challenging and
intriguing. It thus provides a good opportunity to explore the
extreme conditions required for this phenomenon. A mixed
acetonitrile solution of 1 and 2 (1:1 molar ratio, 0.25 mg/mL
each) was prepared to study their self-recognition behavior.
The resulting two separated modes in the CONTIN analysis of
DLS study of the mixed solution indicated the presence of two
differently sized large species (Figure 3a). The two peaks,
within the error limit of the CONTIN analysis, correspond to
those assemblies from the individual solutions containing either
molecular rod 1 or 2 with the same concentration and added
ZnCl2 amount, indicating that the molecular rods might
recognize each other and assemble into their individual
homogeneous blackberry structures (Figures 3a and 5).
Figure 5. Model of the self-recognition during the self-assembly of two
molecular rods in their mixed solution (polyhedra code: blue, MoO6
or MoO5N; purple, MnO6; yellow, FeO6).
like structures with two different sizes coexist (Figure 6a−c),
with sizes close to those of the two corresponding
homogeneous blackberries assembled from the molecular
rods, respectively, confirming the existence of two large species
in DLS results on the mixed solution. Moreover, EDS results
for the individual blackberries suggested that the dominant
metal elements for larger vesicular assemblies (size ∼160 nm)
and smaller vesicular assemblies (size ∼60 nm) are Mn and Fe,
respectively, besides Mo and Zn (Figure 6d,e), which confirms
that molecules 1 and 2 assembled into the larger and smaller
vesicles, respectively. The intensities of Mn or Fe peaks in EDS
studies are comparatively weak due to their low concentrations
in the compounds/assemblies (molar ratio of Mn/Fe to Mo is
1:18). However, the obvious difference between Fe and Mn
signals for each measurement is convincing enough to support
the model of self-recognition. To rule out the possibility that
the EDS results are random, another 18 vesicular assemblies
with different sizes were selected for EDS analysis, which were
consistent with the previous results and supported the model of
homogeneous blackberry structures (Figure S4).
Mechanism of Self-Recognition. Our previous study on
the self-assembly of macroions suggests that the formation of
oligomers at the beginning of the assembly process is the rate-
limiting step with a high activation energy barrier, which
enables the preferential formation of self-sorted oligomers in
the mixture of 1 and 2 (which might be slightly energetically
favored over 1−2 mixed dimers).30 In other words, the
formation process is slow compared to the self-assembly of
amphiphiles; however, it provides the opportunity for
recognition among macroions with minor differences. The
heteroatoms encapsulated inside the inorganic molecular rods
are the only difference between the two molecules and are
expected to lead to the self-recognition behavior (Figures 1 and
5). DFT calculation on Milliken charge distribution indicates
that the central Mn atom of molecule 1 is more positively
charged than the central Fe atom of molecule 2, and
consequently the perimeter of the planar molecular unit (B)
of 1 is more negatively charged than that of 2, making their net
charges still identical for the two macroanions (Tables S2−S4).
Additionally, time-resolved SLS results indicated that the two
molecular rods self-assembled at similar speeds in their
individual acetonitrile solutions with the same rod concen-
tration (0.5 mg/mL, Figure S1), which rules out the possibility
that the self-recognition is primarily due to kinetic effects (i.e.,
that one type of molecular rods assembles faster). Meanwhile,
the mixed solution of 1 and 2 (1:1 molar ratio, concentration
0.25 mg/mL each) showed a much slower overall growth in
scattered intensity in SLS, i.e., much slower overall self-
assembly process than the individual solution of 1 or 2, which
confirms that the molecular rods favor self-sorting in their
mixture solution.
The critical evidence for the self-recognition comes from the
combination of TEM and EDS studies. The TEM images of the
assemblies in the mixed solution indicate that hollow vesicle-
4533
dx.doi.org/10.1021/ja400656j | J. Am. Chem. Soc. 2013, 135, 4529−4536

Journal of the American Chemical Society
Article
Figure 6. (a) TEM image of the assemblies in the mixed solutions (red rectangle, large assembly; green rectangle, small assembly). (b) Zoom-in
image of the large assembly. (c) Zoom-in image of the small assembly. (d) EDS results of the large assembly. (e) EDS results of the small assembly.
Figure 7. (a) Time-resolved SLS results of the three solutions after adding TBA*EDTA (green, 1’s solution; blue, 2’s solution; red, mixture
solution). (b) Model of the reversible self-assembly/disassembly process.
Owing to the difference in their charge distributions, 1 is
slightly more negative on its surface and able to interact with
counterion much more strongly to strengthen the counterion-
mediated attraction, which consequently results in the smaller
curvature of the assemblies of 1, e.g., larger sizes for blackberry
structures, which has been nicely proved by our experiments
(Figure 5).
Due to the anhydrous solution environment, zinc chloride
favors binding onto the oxo ligand-enriched surface of the
molecular rods, which triggers the blackberry formation for
rods 1 and 2 and attributes their difference in charge
distribution to the size difference of their assemblies. In
contrast to the solvation model of ions in aqueous solution,
zinc chloride prefers to directly bind/coordinate onto the
surface oxo ligands of POMs since solvent molecules
(acetonitrile) are comparatively weak ligands, which can be
directly proved by the extended X-ray absorption fine structure
experiments (see Supporting Information).43,44 Thus, the
negative charge density of the oxo ligand of the molecular
rods, which determines the binding strength of oxo−zinc
association, has a significant effect on the blackberry size and
the self-recognition behavior. As discussed above, the terminal
oxo ligands of 1 carry higher negative charges than those of 2,
which strengthens 1−ZnCl2 binding and therefore decreases
the curvature of assemblies, i.e., increases the blackberry size.
Moreover, in the anhydrous acetonitrile solution of the
molecular rods, tiny amount of zinc chloride could become
partially positively charged by replacement of the chloride
ligand with solvent molecule or surface oxo ligand of the
clusters, and for that reason we titrated an excess amount of
ZnCl2 (molar ratio of ZnCl2 to POM is ∼37) to provide
enough counterions for the molecular rods.43 Binding of Zn
complex on the surface of molecular rods can lower the net
charge density to a level appropriate for self-assembly. 1 is
expected to be more negatively charged on its surface and
therefore able to bind more Zn complex than 2, which lowers
the net charge of 1 more and makes its blackberries larger. The
ability of 1 to bind strongly with more zinc ions than 2 is
further confirmed by the following study, which monitors the
disassembly of blackberries in the solutions of the molecular
rods after removing ZnCl2.
Dissociation of the Blackberries and Confirmation of
the Role of Counterions. Time-resolved SLS results of the
above individual acetonitrile solutions of 1 and 2 and their
mixed solutions (with ZnCl2 added) at equilibrium indicated
that scattered intensity decreased exponentially from values
higher than 1000 kcps to ∼100 kcps (scattered intensity for
benzene, 113 kcps) when tetrabutylammonium
4534
dx.doi.org/10.1021/ja400656j | J. Am. Chem. Soc. 2013, 135, 4529−4536

Journal of the American Chemical Society
Article
ethylenediaminetetraacetate ((TBA)4*EDTA, see Supporting
Information; molar ratio of (TBA)4*EDTA to added Zn2+ is
∼1) was added to each of the solutions, respectively, suggesting
the continuous disassembly of the blackberry structures. EDTA
is a strong chelating agent that can quickly coordinate with
metal ions almost stoichiometrically. In our case, EDTA can
extract the Zn2+ from the blackberry structures, which will
become thermodynamically unstable due to the loss of the
counterion-mediated attraction, and consequently disassemble
into monomeric molecular rods. The rate-determining step for
the disassembly of the blackberry structures is the diffusion of
Zn2+ from inside shell structures into bulk solution, which is
highly dependent on the interaction between molecular rods
and Zn2+. The half-life of the disassembly process could be used
to estimate the robustness of the macroion−Zn2+ interaction
and the blackberry structures. In agreement with the DFT
calculation results that 1 can interact more strongly with Zn2+,
1’s solution showed a much longer half-life (2893 s) than that
of 2’s solution (597 s) (Figure 7a). The molecular rods self-
assembled in self-sorted ways in their mixed solution and the
two types of homogeneous blackberry structures should have
different level of robustness according to the above discussions.
Therefore, the half-life of the blackberry dissociation in the
mixed solution should be longer than that of the solution
containing only 2 due to the existence of 1 (and its more robust
assemblies), but shorter than that of the solution containing
only 1 because of the lower concentration of 1’s blackberries.
This has been fully proved by our experimental results (half-life
dissociation for the mixed solution, 1585 s), confirming that the
molecular rod 1 can bind zinc complex more strongly than 2
does. Interestingly, the SLS and DLS results of the these three
solutions suggested that large assemblies with sizes similar to
those of their respective original solutions were observed when
ZnCl2 was added to each of the solutions again, which not only
confirms the stability of monomers during the self-assembly/
disassembly process but also provides us a way to reversibly
control the self-assembly/disassembly of macroions in their
solutions (Figure 7b and experimental details in Supporting
Information).
AUTHOR INFORMATION
■
Corresponding Author
tliu@uakron.edu; yonggewei@tsinghua.edu.cn
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
T. Liu acknowledges support from the NSF (CHE1026505)
and Lehigh University. T. Li, S. Chattopadhyay, T. Shibata, and
X. Zuo are thankful to Dr. Vladislav Zyryanov for designing the
sample cells used for measuring the liquid samples and Dr. J. T.
Miller of Argonne National Laboratory for fruitful discussions
and the use of the Advanced Photon Source. The Advanced
Photon Source, an Office of Science User Facility operated for
the U.S. Department of Energy (DOE) Office of Science by
Argonne National Laboratory, is supported by the U.S. DOE
under Contract No. DE-AC02-06CH11357. MRCAT oper-
ations are supported by the DOE and the MRCAT member
institutions. Y. Wei acknowledges support from NFSC Nos.
21225103 and 20921001, and Tsinghua University Initiative
Foundation Research Program No. 20101081771.
REFERENCES
■
(1) Hanessian, S.; Simard, M.; Roelens, S. J. Am. Chem. Soc. 1995,
117, 7630.
(2) Salonen, L. M.; Ellermann, M.; Diederich, F. Angew. Chem., Int.
Ed. 2011, 50, 4808.
(3) Langner, A.; Tait, S. L.; Lin, N.; Rajadurai, C.; Ruben, M.; Kern,
K. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 17927.
(4) Feng, J.; Ruckenstein, E. J. Chem. Phys. 2006, 124, 124913.
(5) Snyder, P. W.; Mecinovic,
Harder, M.; Mack, E. T.; Lockett, M. R.; Her
́
J.; Moustakas, D. T.; Thomas, S. W.;
́
oux, A.; Sherman, W.;
Whitesides, G. M. Proc. Natl. Acad. Sci. U.S.A. 2011, 108, 17889.
(6) Lehn, J.-M. Angew. Chem., Int. Ed. 1990, 29, 1304.
(7) So, C. R.; Kulp, J. L.; Oren, E. E.; Zareie, H.; Tamerler, C.; Evans,
J. S.; Sarikaya, M. ACS Nano 2009, 3, 1525.
(8) Harada, A.; Kobayashi, R.; Takashima, Y.; Hashidzume, A.;
Yamaguchi, H. Nat. Chem. 2011, 3, 34.
(9) Heilman-Miller, S. L.; Thirumalai, D.; Woodson, S. A. J. Mol. Biol.
2001, 306, 1157.
(10) Roh, J. H.; Behrouzi, R.; Briber, R. M.; Guo, L.; Thirumalai, D.;
CONCLUSION
■
Two types of almost identical macroionic molecular rods, with
the only difference being the central atom of their central
cluster unit (Mn versus Fe), were observed to self-assemble
into hollow spherical blackberry structures and showed self-
recognition behavior by assembling into two types of
homogeneous blackberries instead of mixed ones in their
mixed solution. By controlling the addition and removal of the
counterions, Zn2+, the self-assembly/disassembly of the
molecular rods can be controlled. This self-recognition
achieved between highly similar macroions indicates that the
long-range counterion-mediated electrostatic attraction is
sensitive to the small difference in macroions and consequently
offers the possibility for delicate selectivity and preference
among different macroions. This conclusion could be helpful to
understand the recognition behavior and self-assembly of
biomolecules and direct the design of polyelectrolyte materials.
Woodson, S. A. Biophys. J. 2009, 96, 575a.
(11) Lehn, J.-M. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 4763.
(12) Olson, A. J.; Hu, Y. H. E.; Keinan, E. Proc. Natl. Acad. Sci. U.S.A.
2007, 104, 20731.
(13) Rebek, J. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 10423.
(14) Caulder, D. L.; Raymond, K. N. Angew. Chem., Int. Ed. 1997, 36,
1440.
(15) Mann, S. Nature 1988, 332, 119.
(16) Martin, B.; Becourt, C.; Bienvenu, B.; Lucas, B. Blood 2006, 108,
270.
(17) Misevic, G. N.; Guerardel, Y.; Sumanovski, L. T.; Slomianny, M.
C.; Demarty, M.; Ripoll, C.; Karamanos, Y.; Maes, E.; Popescu, O.;
Strecker, G. J. Biol. Chem. 2004, 279, 15579.
(18) Cellular and Biomolecular Recognition; Jelinek, R., Ed.; Wiley-
VCH: Weinheim, 2009.
(19) Cronin, L.; Muller, A. Chem. Soc. Rev. 2012, 41, 7333.
(20) Liu, T. Langmuir 2010, 26, 9202.
(21) Yin, P.; Li, D.; Liu, T. Chem. Soc. Rev. 2012, 41, 7368.
(22) Kistler, M. L.; Bhatt, A.; Liu, G.; Casa, D.; Liu, T. B. J. Am.
Chem. Soc. 2007, 129, 6453.
ASSOCIATED CONTENT
* Supporting Information
■
S
(23) Liu, G.; Liu, T. B.; Mal, S. S.; Kortz, U. J. Am. Chem. Soc. 2006,
128, 10103.
(24) Liu, T. J. Am. Chem. Soc. 2003, 125, 312.
(25) Liu, T. J. Am. Chem. Soc. 2002, 124, 10942.
All experimental details, Tables S1−S5, and Figures S1−S10.
This material is available free of charge via the Internet at
http://pubs.acs.org.
4535
dx.doi.org/10.1021/ja400656j | J. Am. Chem. Soc. 2013, 135, 4529−4536

Journal of the American Chemical Society
Article
(26) Liu, T.; Imber, B.; Diemann, E.; Liu, G.; Cokleski, K.; Li, H. L.;
Chen, Z. Q.; Muller, A. J. Am. Chem. Soc. 2006, 128, 15914.
(27) Yin, P.; Li, D.; Liu, T. Isr. J. Chem. 2011, 51, 191.
(28) Liu, T.; Diemann, E.; Li, H.; Dress, A. W. M.; Muller, A. Nature
2003, 426, 59.
(29) Wang, Y. F.; Weinstock, I. A. Dalton 2010, 39, 6143.
(30) Liu, T.; Langston, M. L. K.; Li, D.; Pigga, J. M.; Pichon, C.;
Todea, A. M.; Muller, A. Science 2011, 331, 1590.
̈
(31) Tah, B.; Pal, P.; Mahato, M.; Talapatra, G. B. J. Phys. Chem. B
2011, 115, 8493.
(32) Brito, R. O.; Marques, E. F.; Gomes, P.; Falcao, S.; Soderman,
̃
̈
O. J. Phys. Chem. B 2006, 110, 18158.
(33) Tian, M.; Fan, Y.; Ji, G.; Wang, Y. Langmuir 2012, 28, 12005.
(34) Hiemenz, P. C.; Rajagopalan, R. Principles of Colloid and Surface
Chemistry; Marcel Dekker: New York, 1997.
(35) Provencher, S. W. Comput. Phys. Commun. 1982, 27, 229.
(36) Zhang, J.; Hao, J.; Wei, Y.; Xiao, F.; Yin, P.; Wang, L. J. Am.
Chem. Soc. 2009, 132, 14.
(37) Zhang, R.; Thiyagarajan, P.; Tiede, D. M. J. Appl. Crystallogr.
2000, 33, 565.
(38) Svergun, D. I.; Koch, M. H. J. Rep. Prog. Phys. 2003, 66, 1735.
(39) Svergun, D. J. Appl. Crystallogr. 1992, 25, 495.
(40) Stigter, D. Biophys. J. 1995, 69, 380.
(41) Zhang, J.; Li, D.; Liu, G.; Glover, K. J.; Liu, T. J. Am. Chem. Soc.
2009, 131, 15152.
(42) Zlotnick, A.; Aldrich, R.; Johnson, J. M.; Ceres, P.; Young, M. J.
Virology 2000, 277, 450.
(43) Libus, W.; Puchalska, D.; Szuchnicka, T. J. Phys. Chem. 1968, 72,
2075.
(44) Evans, J. C.; Lo, G. Y. S. Spectrochim. Acta 1965, 21, 1033.
4536
dx.doi.org/10.1021/ja400656j | J. Am. Chem. Soc. 2013, 135, 4529−4536