Selection of supramolecular chirality by application of

ARTICLES PUBLISHED ONLINE: 12 FEBRUARY 2012 | DOI: 10.1038/NCHEM.1264

Selection of supramolecular chirality by application of rotational and magnetic forces N. Micali1, H. Engelkamp2, P. G. van Rhee2, P. C. M. Christianen2, L. Monsu` Scolaro3 * and J. C. Maan2 Many essential biological molecules exist only in one of two possible mirror-image structures, either because they possess a chiral unit or through their structure (helices, for example, are intrinsically chiral), but so far the origin of this homochirality has not been unraveled. Here we demonstrate that the handedness of helical supramolecular aggregates formed by achiral molecules can be directed by applying rotational, gravitational and orienting forces during the selfassembly process. In this system, supramolecular chirality is determined by the relative directions of rotation and magnetically tuned effective gravity, but the magnetic orientation of the aggregates is also essential. Applying these external forces only during the nucleation step of the aggregation is sufficient to achieve chiral selection. This result shows that an almost instantaneous chiral perturbation can be transferred and amplified in growing supramolecular selfassemblies, and provides evidence that a falsely chiral influence is able to induce absolute enantioselection.

M

ost essential biological molecules exist only in one of the two possible enantiomers (mirror-image structures) and are either left- or right-handed. The origin of this single handedness (homochirality) in biological systems is unknown, despite its importance for the emergence of life itself1,2. The quest to unravel this mystery has triggered an intense search to identify possible external influences that are able to cause enantioselection3, such as circularly polarized light4–6, the electroweak interaction7,8, vortex motion9, stirring10–13, catalysis at prochiral crystal surfaces14,15 and combinations of external fields16,17. However, external forces can account only for a tiny enantiomeric excess18 and an additional mechanism is required for the chirality to be amplified into homochiral systems6,15,19–25. To date, the combined enantioselection by purely physical fields and subsequent amplification of chirality, in the absence of these fields, has not been shown. Here we demonstrate that the application of rotational and magnetic (both levitating and orientating) forces only during the initial stage of the self-assembly of achiral molecules in solution leads to selection of chirality in the nucleation phase, which is subsequently amplified in the growing assemblies (Fig. 1). Using this particular set of physical forces (rotation, levitation—also referred to as ‘effective gravitation’—and orientation) to create a hydrodynamic chiral flow in solution has several advantages. First, all forces are tuneable in size and direction: the gravitational and alignment forces are controlled by applying high magnetic fields, through magnetic levitation26,27 and magnetic orientation28–31, respectively. This permits us to investigate the separate effects of each of these forces on the chiral selection process. Second, the chiral sign of the flow can be chosen freely by the relative directions of the rotation and the effective gravity; that is, the chiral sign is inverted simply by inverting the effective gravity, without changing the rotation sense32. Third, this approach makes it very easy to impose the physical chiral bias only at the very beginning of the self-assembly process (the nucleation phase) and remove the bias during the subsequent growth step. This allows us to disentangle the chiral selection and propagation steps, which is very difficult (although possible) with the more commonly used chemical

chiral templates, for which disentanglement requires their complete removal prior to the growth step15. Fourth, the supramolecular chirality is measured ex situ by circular dichroism (CD) at the very end of the self-assembly process, in the absence of the external physical forces. As a result, solely the chirality of the aggregates at rest is measured, and any CD signals that may be present only during the rotation are avoided33–35. Using this approach we find that the handedness of the aggregates is directed by the chiral sign of the hydrodynamic flow, defined by the rotation sense and the direction of the effective gravity (tuned by magnetic levitation). Under rotation, but in the absence of a magnetic field, no chiral selection is observed, which provides evidence that a hydrodynamic chiral flow alone is insufficient and that the magnetic alignment of the aggregates is essential. Applying the chiral bias only at the very beginning of the self-assembly process is enough to achieve enantioselection. This result shows that an initial, small, macroscopic chiral perturbation may direct the handedness of nanoscopic aggregates, which can be transferred subsequently to the final supramolecular self-assemblies during the growth step. Furthermore, our combination of rotation and effective gravity is an example of a falsely chiral system36,37. The concept of true and false chirality was introduced by Barron to include dynamical effects in the treatment of chirality as an extension of Lord Kelvin’s definition36,37. Following this description, a falsely chiral system changes chiral sign under space inversion (parity), but also under time reversal—combined with any proper spatial rotation, as opposed to a truly chiral system, which is invariant under time reversal. In this work the two distinct enantiomeric states of the physical forces are interconverted by both space inversion and time reversal. In the case of thermodynamic equilibrium such a falsely chiral influence should not lead to enantioselection, because the two enantiomers are isoenergetic. However, here the chiral selection occurs in the nucleation phase of the self-assembly process and provides insight into how a falsely chiral influence that is under kinetic control and far from equilibrium is able to induce absolute enantioselection, and thereby confirms a long-standing theoretical prediction17,36,37.

1

CNR-IPCF Istituto per i Processi Chimico-Fisici, Messina, Italy, 2 High Field Magnet Laboratory, Institute for Molecules and Materials, Radboud University Nijmegen, The Netherlands, 3 Dipartimento di Chimica Inorganica, Chimica Analitica e Chimica Fisica, Universita` di Messina, and C.I.R.C.M.S.B., Italy. * e-mail: [email protected] NATURE CHEMISTRY | VOL 4 | MARCH 2012 | www.nature.com/naturechemistry

© 2012 Macmillan Publishers Limited. All rights reserved.

201

ARTICLES

NATURE CHEMISTRY a

–

DOI: 10.1038/NCHEM.1264

O3S

N H NH +

+ HN

=

H N

–

O3S

or

–

0

–50

100 10–1

100 10–1

10–2

10–2

10–3 10–3

Number concentration

50

c Spectral amplitude

SO3

b 100

CD (mdeg)

Self-assembly under hydrodynamic chiral flow and magnetic field

10–4 0.1 450 Wavelength (nm)

d 0.02 Linear dichroism

0.4 Absorbance

1 RH (μm)

500

0.2

LD signal

400

0.01

0

0.0 400

0.00 450 Wavelength (nm)

500

0

5

1

2 3 B2 (T2)

10 15 Magnetic field (T)

4

20

Figure 1 | Characterization of the supramolecular aggregates. a, Molecular formula of the achiral porphyrin TPPS3 , schematically represented by the yellow– green platelet. The TPPS3 monomers self-assemble (red arrow) into chiral aggregates under an applied chiral hydrodynamic flow. b, In the UV/vis spectrum (bottom black curve) TPPS3 J-aggregates exhibit an absorption at 490 nm, which is red-shifted in comparison to the monomer peak located at 433 nm. J-aggregates of opposite handedness show mirror CD spectra (top red and blue curves). c, The size (red curve) and number concentration (blue curve) distributions obtained by quasi-elastic light-scattering measurements on equilibrated solutions of TPPS3 aggregates. The red and blue arrows indicate that the left and right vertical axes correspond to the curves for, respectively, the size and number concentration. d, LD of TPPS3 aggregates induced by a magnetic field demonstrate the magnetic alignment of equilibrated aggregates. Inset: at low fields the LD signal scales with B 2 (red line is a guide to the eye).

Results and discussion As a spectroscopic probe, we used the J-aggregates of an achiral molecular building block, tris-(4-sulfonatophenyl)phenylporphyrin (TPPS3) (Fig. 1a). The TPPS3 molecules aggregate into assemblies through an interplay of electrostatic and p-stacking interactions, and it is reported that stirring TPPS3 solutions can cause mirror-symmetry breaking in their supramolecular helical assemblies10–12. Aggregation is triggered in an aqueous solution by lowering the pH and increasing the ionic strength. Immediately after preparation (up to two hours) the aggregates are very small (,40 nm), as evidenced by a very low light-scattering intensity. The full aggregation process requires three days to complete. The final equilibrated aggregates exhibit an absorption band at 490 nm and their chirality is revealed by exciton-split CD bands (Fig. 1b)38. Elastic and quasi-elastic light-scattering data also show that these final TPPS3 aggregates remain relatively small (gyration radius ≥0.1 mm, hydrodynamic radius ≤0.8 mm) (Fig. 1c, see also Supplementary Fig. S1) and are much smaller than the structures investigated in previous studies10– 12 . These nanosize aggregates are too small to sediment and therefore do not align under the influence of normal gravity, which permits us to disentangle the role of the orientating and gravitational forces and investigate their separate effects on chiral selection. 202

The experiment is outlined in Fig. 2. Seven cylindrical vessels were placed at different positions (z) inside a magnet (Bcentre ¼ 0, 16 or 25 T) and rotated at fixed frequency (15 Hz) for a period t, which ranged from 30 to 120 minutes, at a temperature of 23 8C (Fig. 2a). Such rotation of cylindrical vessels results in a solidbody rotation or rotational vortex motion of the solution39, and so avoids the creation of pseudovortices. Clockwise and anticlockwise (viewing from the top) rotation was used (Fig. 2a). The alignment force was determined by the strength (B(z)) of the applied magnetic field and results from the anisotropy in the magnetic susceptibility x along the different molecular axes of the porphyrin (Supplementary Fig. S2 and Table S1)28–31. Solutions in the vessels at different vertical positions (z) inside the magnet experience different magnetic field strengths B(z) and field gradients B′ (z)(dB/dz) (Fig 2b, black line). This leads to a magnetic levitation force proportional to x B(z)B′ (z), as well as a z-dependent effective gravity Geff(z) ¼ Gn(1 þ B(z)B′ (z)|x |/m0r) (Fig 2b, red line), with r being the density of the material, m0 the magnetic constant and Gn normal gravity32. For example, when Bcentre ¼ 25 T, the vessel at z ¼ 0 does not experience an extra magnetic force (Geff ¼ Gn , Fig. 2b, black dots); the three vessels with z , 0 undergo an enhanced effective gravity (Geff . Gn, Fig. 2b, blue dots); the NATURE CHEMISTRY | VOL 4 | MARCH 2012 | www.nature.com/naturechemistry


NATURE CHEMISTRY b

ω

a

Geff (Gn) –1

0

1

2

3

4

150

100

Near levitation

Normal gravity

–50

50

0

0 –50

50 0 –50 480

–50

500

480 500 Wavelength (nm)

480

500

b –100

CW rotation

0.01

Δg

–100

Enhanced effective gravity

CD (mdeg)

z (mm)

0

Inverted effective gravity

50

ACW rotation

50

Vertical position z (mm)

100

CW rotation

–2

B

B = 0 T, Geff = Gn B > 0 T, Geff > 0 B > 0 T, Geff < 0

CD (mdeg)

a

ARTICLES

DOI: 10.1038/NCHEM.1264

0.00

–150 10

c

15 20 Magnetic field (T)

–0.01

25

ACW rotation

Δg

0.01 0.00 –0.01

(1.00 ± 0.00)Gn

(0.92 ± 0.01)Gn

(0.67 ± 0.02)Gn

B=0T Geff = Gn

(0.26 ± 0.05)Gn

B>0T Geff < 0

B > 0 T, Geff > 0

c CW rotation

Δg

0.01

5 mm (–0.32 ± 0.09)Gn

(–0.67 ± 0.11)Gn

(–1.06 ± 0.14)Gn

vessels with z . 0 experience an inverted effective gravity (Geff , 0, Fig. 2b, red dots), with the very top vessel located close to the magnetic levitation point (Geff ¼ 0, Fig. 2d, green dot), where the magnetic force upwards cancels the gravitational force downwards. The paraboloidal shape of the solution at various values of Geff is illustrated in Fig. 2c and roughly follows the v2/Geff dependence expected for a rotating solution under effective gravity39. The curvature increases with decreasing Geff , is maximal around Geff ¼ 0 when all the solution resides at the wall and is inverted for negative Geff. Blank experiments with Geff ¼ Gn were performed on rotating samples, with the same experimental set-up, but at B ¼ 0. After each experiment the samples were extracted from the magnet and allowed to stand for three days at 23 8C. After this period each solution was transferred from the vessel to a quartz

ACW rotation

Figure 2 | Experimental procedure. a, The set-up consists of a tube that holds seven vials each with equal volumes of an aggregating TPPS3 solution (shown in green). These vials were placed at different positions z (black scale) inside the magnet and rotated with angular frequency v ¼ 2pf, (orange arrow, f ¼ 15 Hz) for time periods from t ¼ 30 to t ¼ 120 minutes. The magnetic field was oriented vertically (B, red arrow). The local effective gravity Geff experienced by the samples is shown by blue (enhanced positive Geff ), black (normal gravity), red (inverted Geff ) and green (near zero Geff or levitation) arrows. After time t the vials were allowed to stand (outside the magnet) for three days at the same temperature. b, Magnetic field profile (black line) and Geff(z) for water (red line) as a function of the position z. The red and black vertical arrows indicate that the bottom and top scales correspond, respectively, to B(z) and Geff(z). The position of the samples is indicated by dots: blue, positive Geff; black, normal gravity Gn; red, negative Geff; green, near zero Geff. c, Photographs showing the influence of effective gravity (decreasing from top left to bottom right) on the surface of spinning water (15 Hz).

0.01

Δg

(0.05 ± 0.06)Gn

0.00 –0.01

0.00 –0.01 –50

0 B2Geff –1 (T2 s2 m–1)

50

Figure 3 | Selection of supramolecular chirality. a, Typical CD spectra showing the correlation between the handedness of the final TPPS3 aggregates, spinning directions (clockwise (CW), upper panels, and anticlockwise (ACW), lower panels) and effective gravity (Geff . 0, middle panels, and Geff , 0, right panels), with (middle and right panels) and without (left panels) a magnetic field. b, The chirality parameter Dg measured for clockwise (upper panels) and anticlockwise (lower panels) rotation, with Bcentre ¼ 0 (left panels) and B . 0 (middle and right panels), under positive (middle panels) and negative (right panels) effective gravity conditions. Without a magnetic field, no correlation between rotation and handedness was observed (left panels). With a magnetic field, the handedness was inverted by inverting the rotation direction under the same gravity (for example, for Geff . 0, compare the upper and lower middle panels) or by inverting the effective gravity under the same rotation sense (for example, for clockwise rotation compare the upper middle and right panels). c, Dg measured for both clockwise (upper panel) and anticlockwise (lower panel) rotation as a function of B 2(z)/Geff(z). Green triangles, t ¼ 30 minutes; red circles, t ¼ 60 minutes; blue squares, t ¼ 120 minutes. Red lines are linear fits to the data points for t ¼ 60 minutes.

cell for the ultraviolet–visible (UV/vis) absorbance and CD measurements, a procedure that avoids any artefacts that may arise from the material sticking to the wall of the vessels11 and that measures the true supramolecular chirality of the TPPS3 aggregates at rest.

NATURE CHEMISTRY | VOL 4 | MARCH 2012 | www.nature.com/naturechemistry


203

ARTICLES

L

Geff or

L

Geff

CD (mdeg)

NATURE CHEMISTRY

DOI: 10.1038/NCHEM.1264

50 0 –50

480 500 Wavelength (nm)

L

Geff or

L

Geff

CD (mdeg)

TPPS3 50 0 –50 480 500 Wavelength (nm)

Figure 4 | Correlation between the observed chirality and the applied physical forces. TPPS3 (yellow–green platelet) self-assembles in the presence of an initial external chiral influence generated by rotational and magnetic forces. The rotation is characterized by the angular momentum (L), the direction of which is set by clockwise or anticlockwise rotation (viewing from the top). The effective gravity (Geff) is tuned by a magnetic levitation force. Depending on the symmetry of the combined forces the system grows into left-handed or right-handed aggregates, identified by their CD spectra (yellow helices, shown here for illustrative purposes; the handedness corresponding to each set of forces has not been determined). In particular, a parallel arrangement of L and Geff (blue box), achieved by clockwise rotation and Geff . 0 or by anticlockwise rotation and Geff , 0, leads to negative CD signals (blue curve, lower right panel); vice versa, an antiparallel arrangement of L and Geff (red box), achieved by anticlockwise rotation and Geff . 0 or by clockwise rotation and Geff , 0, leads to positive CD signals (red curve, upper right panel).

Figure 3 summarizes the correlations between the different applied external forces and the observed chirality. In particular, Fig. 3a shows typical CD spectra for different chiral signs of the hydrodynamic flow. Without a magnetic field (left panels), both clockwise and anticlockwise rotations led to CD spectra that exhibited relatively low intensities and signs uncorrelated to the rotation sense. In a magnetic field (middle and right panels), the CD intensity increased drastically and the handedness was determined by the relative directions of spinning and effective gravity. We quantified the sign and magnitude of the supramolecular chirality by the dissymmetry factor Dg ¼ g(491nm) – g(486nm) , where g is the ratio of the CD to the conventional absorption (see Methods section)38. Figure 3b shows Dg for a large number of experiments (52 at B ¼ 0 and 96 at B . 0). The sign of Dg for almost all the samples prepared in a magnetic field (90 out of 96, 94%) correlated with the chiral sign of the flow. Dg was negative for clockwise rotation with Geff . 0 and it was still negative for anticlockwise rotation with Geff , 0; vice versa, Dg was positive for anticlockwise rotation with Geff . 0 and for clockwise rotation with Geff , 0. Thus, the sign of Dg was changed by inverting the direction of either the rotation sense or the effective gravity. Inverting both leaves the chiral sign unchanged. The handedness of the aggregates is, therefore, determined exclusively by the relative directions of the rotational and gravitational forces applied, which is shown schematically in Fig. 4. The rotation is characterized by angular momentum L, the direction of which is set by clockwise or anticlockwise rotation (viewed from the top). The direction of the effective gravity vector Geff is varied by the magnetic levitation. Antiparallel L and Geff (Fig. 4, red box) leads to one chirality and parallel L and Geff (Fig. 4, blue box) results in the opposite chirality. This finding is a unique example in which supramolecular chirality is proved to be influenced by gravity and one of the very few cases40,41 in which gravity is shown to affect a chemical self-assembly process at all. At B ¼ 0 and Geff ¼ Gn small values of Dg were observed, with a bias towards positive Dg for both rotation directions (left panels, Fig. 3b), probably because of a small, unidentified chemical chiral contamination within the starting materials. This bias, however, was overruled completely by the applied chiral set 204

of physical forces, which underlines the efficiency of our experimental approach. The absence of any correlation between the chiral signs of the CD spectra and the flow at B ¼ 0 demonstrates that the chiral flow alone is not sufficient to direct the supramolecular chirality. An additional requirement for enantioselection is the alignment of the aggregates by the magnetic field. Molecules with a large anisotropy in their magnetic susceptibility x tend to be aligned by a magnetic field with their axis of lowest x along the field to minimize their magnetic energy, Um ¼ –12x B 2/m0. However, for individual molecules, the change in energy is small relative to the thermal energy. Only a sufficiently large group of ordered molecules can overcome the thermal randomization and align. As a result, aggregates that consist of molecules with a strongly anisotropic x , such as (phthalo)cyanines29,42 porphyrins30,31, coronenes43 and thiophenes44, can be oriented by magnetic fields28 (Supplementary Information). Figure 1d shows a separate magnetic orientation experiment on an equilibrated TPPS3 aggregate solution positioned in the centre of the magnet. The magnetic field-induced linear dichroism (LD) curve (measured at a wavelength of 488 nm) exhibits the typical magnetic alignment behaviour (see also Supplementary Fig. S3). At low fields the LD signal (proportional to the degree of orientation) increases with B 2 (Fig. 1d, inset), which saturates at higher fields when the alignment is complete29,30,42. The direction of B does not influence the results, either of the LD experiments or of the chiral selection, as expected for magnetic alignment. In this context it is important that the magnetic field itself does not play a fundamental role45, as opposed to, for example, magnetochiral photochemistry16 and magnetochiral dichroism46,47, where the chiral sign depends on the absolute orientation of the field with respect to the light propagation. In our study the magnetic field orientation relative to the directions of rotation and effective gravity is not relevant. The magnetic field merely provides the tunable levitation and alignment forces, which both scale with B 2. Levitation and alignment created by any other means are expected to lead to an identical chiral selection process. It is crucial that absorption-, CD-, light scattering- and B-induced LD signals of the aggregates are not or are barely NATURE CHEMISTRY | VOL 4 | MARCH 2012 | www.nature.com/naturechemistry


NATURE CHEMISTRY

Absorbance at 490 nm

0.10

ARTICLES

DOI: 10.1038/NCHEM.1264

Chiral selection during nucleation

Chiral amplification during growth

0.05

Rotation + magnetic levitation and alignment

101

102

103 Time (s)

104

105

Figure 5 | Proposed model for the chiral selection and amplification. The chiral perturbation (rotation and magnetic alignment and levitation) is applied during the early stage of the aggregation process (blue arrow). At this stage, the achiral TPPS3 monomers (yellow–green platelets) selfassemble into small nanoassemblies (short yellow helices), which within a couple of days grow into larger chiral structures (yellow helices). The plot shows a typical kinetic absorbance trace (at 490 nm, black dots), fitted by a single exponential function (red line).

detectable immediately after removing the chiral perturbation. These signals are measured after three days, when the solution is equilibrated (Fig. 5). Considering the experimental evidence, we propose the following schematic model for the enantioselection mechanism. In the nucleation period preceding the growth of the aggregates, nanoassemblies form in solution. At this stage the hydrodynamic flow gives a chiral twist to the nuclei. The magnetic field orients the nuclei along the rotation axis, and thereby decreases the influence of the randomizing Brownian motion that tends to destroy the chirality. Once the chiral seeds are formed, they serve as templates for chiral growth, even after the chiral influence is removed, which leads to amplification of the supramolecular chirality. Within this model the Dg values obtained are expected to depend strongly on the experimental conditions (such as rotation speed, size of Geff and degree of magnetic orientation) and on experimental factors (such as the dimensions of the vessels and the exposure time t). The rotation speed and vial dimensions were constant in our experiments. The magnetic orientation and Geff were varied simultaneously, because the different vessels in the magnet experienced different field strengths and field gradients (Fig. 2a,b). To account for both effects, we plotted (Fig. 3c) Dg of all the experiments shown in Fig. 3b as a function of B 2(z)/Geff(z). All the data points roughly follow a linear trend, with a slope determined by the spinning sense, as indicated by the red lines that correspond to the t ¼ 60 minutes data (even though the stochastic nature of the process is expected to cause a substantial amount of scattering among the data). Application of the external forces during a time as short as 30 minutes (Fig. 3c, green triangles) is sufficient to obtain enantioselection, and the effect increases with time t. We attribute the 1/Geff dependence to the influence of the actual chiral hydrodynamic flow, because the paraboloidal surface of a spinning liquid is inversely proportional to gravity (Fig 2c)39. If the macroscopic chiral hydrodynamic flow was responsible for sculpting the aligned growing aggregates, this would lead to a 1/Geff effect. To test this point, we performed an experiment using completely filled vials, which eliminated a paraboloidal flow, with all other parameters left unaltered. The resulting CD spectra are highly asymmetric and show no correlation between the rotation sense and Geff (Supplementary Fig. S4).

The B 2 dependence confirms the crucial role of the magnetic alignment of the aggregates and explains the low Dg values and the absence of chiral selection at B ¼ 0. It extends through the magnetic field range up to 25 T, because in the nucleation phase the aggregates are too small to be aligned completely28–31,42, in contrast to the larger final aggregates after three days, which exhibit full alignment around 10 T (Fig. 1d). As no saturation of Dg at the highest B 2/Geff values was observed, higher amplitudes are expected with the application of even higher magnetic fields. Figure 3c provides an indication as to why, in some of the experiments, the chiral sign of the aggregates did not match the sign of the chiral flow: either the B2/Geff value is low (B2/Geff ¼ 9 T2 s2 m21) or the variation of Geff within the vessel is large (B2/Geff ¼ –23 T2 s2 m21, Supplementary Information), which, in combination with the statistical spread in the data, explains the deviation. For all experiments with t ¼ 120 minutes and/or using vessels at the middle three positions of the magnet, when B and B 2/Geff are large, the chiral selection is complete, which shows that 100% enantioselectivity of supramolecular chirality is possible by the application of a chiral set of sufficiently large rotational and magnetic forces. Our experiment provides full control over the relative directions of L (rotation) and Geff , and thereby the chiral sign of the flow. The two distinct enantiomeric situations are given by, respectively, the parallel and antiparallel configurations of L and Geff (Fig. 4), which is a typical example of a falsely chiral system. As mentioned previously, a falsely chiral system changes chiral sign under space inversion (parity), but also under time reversal combined with any proper spatial rotation, as opposed to a truly chiral system, which is invariant under time reversal. In our system, L is a time odd, axial vector unaffected by the operation of parity, but changes sign under time reversal. Geff is a time even, polar vector, which is invariant under time reversal and changes sign under space inversion. The parallel and antiparallel L and Geff are, therefore, interconnected by both space inversion and time reversal, which makes it a falsely chiral system that, under thermodynamic equilibrium, should not lead to enantioselection36,37. However, because here the chiral selection occurs during the nucleation phase of the self-assembly process, the system is under kinetic control, far from equilibrium, and time-reversal symmetry is broken. Under such conditions it has been predicted that a falsely chiral influence might lead to absolute enantioselection17,36,37,45. In summary we exploited a combination of physical forces to generate a macroscopic hydrodynamic chiral flow that directs the handedness of nanoscale aggregates. We show that orientation of the aggregates is indispensable to activate the mechanism for chiral mirror-symmetry breaking. The results presented here provide insight into how an extremely small18 falsely chiral17,36,37 influence can be expressed and amplified in a supramolecular system. Even if they provide further experimental support for the hypothesis on the important role of vortices in prebiotic chemistry, whether these observations are related to the onset of the homochirality of life remains an open question.

Methods

Materials and methods. In a typical run, a 3 × 1026 M solution of TPPS3 (Midcentury, sodium salt) was acidified with aqueous H2SO4 (98%, Fluka) to a final concentration of 0.0133 M (pH 1.9) and eventually aggregation was triggered by adding a 3 M solution of NaCl (SigmaUltra) up to a concentration of 0.1 M in a total volume of 7 ml (water used for preparing stock solutions and for diluting samples was high-purity doubly distilled from Galenica Senese). This stock solution was quickly divided into various aliquots (0.9 ml) in small 1.5 ml vials with screw caps and Teflon septa, placed in a double-walled metallic holder tube with a thermostat (23+0.1 8C) that fitted into the polar expansion of the magnet (see Fig. 2a). The holder was rotated inside the magnet for 30–120 minutes at a frequency of 15 Hz. In different experiments, the rotation direction (clockwise or anticlockwise, as observed from the top), the centre magnetic field (0 T, 16 T or 25 T) and the time were varied,



205

ARTICLES

NATURE CHEMISTRY

keeping all other parameters constant. After a period of time (t) the samples were extracted from the magnet and from the holder and allowed to stand for three days at 23 8C. Then the solutions were transferred to a quartz cell (Hellma, 1 cm path length) for spectroscopic measurements. UV/vis absorption spectra were recorded on a spectrophotometer Varian Cary 50 and CD spectra were obtained on a spectropolarimeter Jasco model J-810. To quantify the sign and magnitude of the supramolecular chirality we used the dissymmetry factor Dg ¼ g(491nm) – g(486nm) , where g is the ratio of the CD to the conventional absorption ( g ¼ DA/A, where A is the absorbance of the sample and DA ¼ AL – AR , with AL and AR the absorbance of left- and right-handed circular polarized light, respectively)38. Description of the magnets used and determination of the effective gravity. The magnets used for the experiments were a 32 mm bore 20 T duplex Bitter magnet and a 32 mm bore 33 T Florida-Bitter magnet at the High Field Magnet Laboratory at the Radboud University Nijmegen. The field strength was set by the electrical current. To determine the effective gravity the following procedure was used. The profile of the magnetic field B(z) was measured independently using a Hall probe. From the profile, the corresponding field gradient dB(z)/dz was calculated. The normalized effective gravity for water was then defined by Geff/Gn ¼ (1 þ B(z)B′ (z)/1,360). Gn¼ 9.81 m s22 is the normal gravitational acceleration and B(z)B′ (z) ¼ –1,360 T2 m21 is the experimentally determined value at which water levitates in air. At Bcentre ¼ 25 T the effective gravity for water in this set-up varied from –2Gn to 4Gn (Supplementary Fig. S2). The samples prepared at the levitation point were excluded from the analysis as in that situation the water surface is not paraboloidal because water is expelled to the wall of the vial. Imaging of the TPPS3 solution during rotation as a function of effective gravity. The TPPS3/water solution was filmed inside a vessel rotating at a fixed frequency (15 Hz) as a function of effective gravity. The vessel was mounted at the point of maximum B(z)B′ (z) at 82 mm above the field centre of a 50 mm bore 31 T FloridaBitter magnet. At this position B(z)B′ (z) ¼ –4.946 B 2centre m21, from which the effective gravity Geff ¼ Gn(1 þ B(z)B′ (z)/1,360) was calculated. Using two mirrors and a zoom lens the image was taken by a charge-coupled device video camera. Snap shots are shown in Fig. 2c. The Supplementary Movie shows the full sweep of the centre magnetic field Bcentre from 0 to 25 T. Elastic and quasi-elastic light-scattering experiments. These were performed at small and wide angles using an He–Ne laser and home-made apparatus already described in the literature48. The particle-size analysis was performed on equilibrated samples three days after preparation, using a method described in the literature48,49. LD measurements. LD was measured in a 32 mm bore 20 T Duplex Bitter magnet using a standard polarization modulation technique50. Different lines of a Spectra Physics Argon ion laser (model 2017) were used. The equilibrated TPPS3 J-aggregate solution was contained in a 5 mm thick optical cell (Hellma) with the temperature kept constant at 23+0.1 8C by a water-based temperature controller. The LD was measured by slowly sweeping the magnetic field between 0 and 20 T. The small signal of the pure solvent caused by the experimental set-up was used as a background; a positive LD signal meant that the absorbance with light polarized parallel to the magnetic field was higher than that with light polarized perpendicular to the magnetic field.

Received 24 August 2011; accepted 4 January 2012; published online 12 February 2012

References 1. Wagnie`re, G. H. On Chirality and the Universal Asymmetry: Reflections on Image and Mirror Image (Wiley-VCH, 2007). 2. Luisi, P. L. The Emergence of Life: From Chemical Origins to Synthetic Biology (Cambridge Univ. Press, 2006). 3. Mason, S. F. Origins of biomolecular handedness. Nature 311, 19–23 (1984). 4. Flores, J. J., Bonner, W. A. & Massey, G. A. Asymmetric photolysis of (RS)-leucine with circularly polarized UV light. J. Am. Chem. Soc. 99, 3622–3625 (1977). 5. Bailey, J. et al. Circular polarization in star-formation regions: implications for biomolecular homochirality. Science 281, 672–674 (1998). 6. Noorduin, W. L. et al. Complete chiral symmetry breaking of an amino acid derivative directed by circularly polarized light. Nature Chem. 1, 729–732 (2009). 7. Berger, R. & Quack, M. Electroweak quantum chemistry of alanine: parity violation in gas and condensed phases. ChemPhysChem 1, 57–60 (2001). 8. Kondepudi, D. K. & Nelson, G. W. Weak neutral currents and the origin of biomolecular chirality. Nature 314, 438–441 (1985). 9. Aquilanti, V. & Maciel, G. S. Observed molecular alignment in gaseous streams and possible chiral effects in vortices and in surface scattering. Orig. Life Evol. Biosph. 36, 435–441 (2006). 10. Ribo, J. M., Crusats, J., Sagues, F., Claret, J. & Rubires, R. Chiral sign induction by vortices during the formation of mesophases in stirred solutions. Science 292, 2063–2066 (2001). 206

DOI: 10.1038/NCHEM.1264

11. D’Urso, A., Randazzo, R., Lo Taro, L. & Purrello, R. Vortexes and nanoscale chirality. Angew. Chem. Int. Ed. 49, 108–112 (2010). 12. Escudero, C., Crusats, J., Dıéz-Pe´rez, I., El-Hachemi, Z. & Ribo´, J. M. Folding and hydrodynamic forces in J-aggregates of 5-phenyl-10,15,20-tris(4-sulfophenyl) porphyrin. Angew. Chem. Int. Ed. 45, 8032–8035 (2006). 13. Arteaga, O. et al. Emergence of supramolecular chirality by flows. ChemPhysChem 11, 3511–3516 (2010). 14. Alexander Kuhn, A. & Fischer, P. Absolute asymmetric reduction based on the relative orientation of achiral reactants. Angew. Chem. Int. Ed. 48, 6857–6860 (2009). 15. Kawasaki, T., Kamimura, S., Amihara, A., Suzuki, K. & Soai, K. Enantioselective C–C bond formation as a result of the oriented prochirality of an achiral aldehyde at the single-crystal face upon treatment with a dialkyl zinc vapour. Angew. Chem. Int. Ed. 50, 6796–6798 (2011). 16. Rikken, G. L. J. A. & Raupach, E. Enantioselective magnetochiral photochemistry. Nature 405, 932–935 (2000). 17. Avalos, M. et al. Absolute asymmetric synthesis under physical fields: facts and fictions. Chem. Rev. 98, 2391–2404 (1998). 18. Mead, C. A. & Moscowitz, A. Some comments on the possibility of achieving asymmetric synthesis from achiral reactants in a rotating vessel. J. Am. Chem. Soc. 102, 7301–7302 (1980). 19. Viedma, C. Chiral symmetry breaking during crystallization: complete chiral purity induced by nonlinear autocatalysis and recycling. Phys. Rev. Lett. 94, 065504 (2005). 20. Palmans, A. R. A. & Meijer, E. W. Amplification of chirality in dynamic supramolecular aggregates. Angew. Chem. Int. Ed. 46, 8948–8968 (2007). 21. Randazzo, R., Mammana, A., D’Urso, A., Lauceri, R. & Purrello, R. Reversible ‘chiral memory’ in ruthenium tris(phenanthroline)-anionic porphyrin complexes. Angew. Chem. Int. Ed. 47, 9879–9882 (2008). 22. Lauceri, R., Raudino, A., Scolaro, L. M., Micali, N. & Purrello, R. From achiral porphyrins to template-imprinted chiral aggregates and further. Self-replication of chiral memory from scratch. J. Am. Chem. Soc. 124, 894–895 (2002). 23. Onouchi, H., Miyagawa, T., Morino, K. & Yashima, E. Assisted formation of chiral porphyrin homoaggregates by an induced helical poly(phenylacetylene) template and their chiral memory. Angew. Chem. Int. Ed. 45, 2381–2384 (2006). 24. Pasternack, R. F., Giannetto, A., Pagano, P. & Gibbs, E. J. Self-assembly of porphyrins on nucleic-acids and polypeptides. J. Am. Chem. Soc. 113, 7799–7800 (1991). 25. Bellacchio, E. et al. Template-imprinted chiral porphyrin aggregates. J. Am. Chem. Soc. 120, 12353–12354 (1998). 26. Beaugnon, E. & Tournier, R. Levitation of organic materials. Nature 349, 470 (1991). 27. Berry, M. V. & Geim, A. K. Of flying frogs and levitrons. Eur. J. Phys. 18, 307–313 (1997). 28. Maret, G. & Dransfeld, K. in Topics in Applied Physics Vol. 57 (ed. Herlach, F.) 143–204 (Springer, 1985). 29. Boamfa, M. I., Christianen, P. C. M., Engelkamp, H., Nolte, R. J. M. & Maan, J. C. Magnetic fields as an investigation technique and manipulation tool for phthalocyanine molecular aggregates. Adv. Funct. Mater. 14, 261–265 (2004). 30. Helmich, F. et al. Dilution-induced self-assembly of porphyrin aggregates: a consequence of coupled equilibria. Angew. Chem. Int. Ed. 49, 3939–3942 (2010). 31. Kitahama, Y., Kimura, Y. & Takazawa, K. Study of internal structure of mesotetrakis (4-sulfonatophenyl) porphine J-aggregates in solution by fluorescence microscope imaging in a magnetic field. Langmuir 22, 7600–7604 (2006). 32. Heijna, M. C. R. et al. Magnetically controlled gravity for protein crystal growth. Appl. Phys. Lett. 90, 264105 (2007). 33. Tsuda, A. et al. Spectroscopic visualization of vortex flows using dye-containing nanofibers. Angew. Chem. Int. Ed. 46, 8198–8202 (2007). 34. Wolffs, M. et al. Macroscopic origin of circular dichroism effects by alignment of self-assembled fibers in solution. Angew. Chem. Int. Ed. 46, 8203–8205 (2007). 35. Crusats, J., El-Hachemi, Z. & Ribo´, J. M. Hydrodynamic effects on chiral induction. Chem. Soc. Rev. 39, 569–577 (2010). 36. Barron, L. D. Reactions of chiral molecules in the presence of a time-noninvariant enantiomorphous influence: a new kinetic principle based on the breakdown of microscopic reversibility. Chem. Phys. Lett. 135, 1–8 (1987). 37. Barron, L. D. True and false chirality and absolute asymmetric-synthesis. J. Am. Chem. Soc. 108, 5539–5542 (1986). 38. Berova, N., Nakanishi, K. & Woody, R. W. Circular Dichroism: Principles and Applications (Wiley-VCH, 2000). 39. Kundu, P. K. & Cohen, I. M. Fluid Mechanics, 4th edn (Academic Press, 2008). 40. Tabony, J. & Lob, D. Gravitational symmetry breaking in microtubular dissipative structures Proc. Natl Acad. Sci. USA 89, 6948–6952 (1992). 41. Glade, N., Beaugnon, E. & Tabony, J. Ground-based methods reproduce spaceflight experiments and show that weak vibrations trigger microtubule selforganisation. Biophys. Chem. 121, 1–6 (2006). 42. Shklyarevskiy, I. O. et al. Determination of the molecular arrangement inside cyanine dye aggregates by magnetic orientation, J. Phys. Chem. B 108, 16386–16391 (2004). NATURE CHEMISTRY | VOL 4 | MARCH 2012 | www.nature.com/naturechemistry


NATURE CHEMISTRY

ARTICLES

DOI: 10.1038/NCHEM.1264

43. Shklyarevskiy, I. O. et al. High anisotropy of the field-effect transistor mobility in magnetically aligned discotic liquid–crystalline semiconductors. J. Am. Chem. Soc. 127, 16233–16237 (2005). 44. Shklyarevskiy, I. O. et al. Magnetic deformation of self-assembled sexithiophene spherical nanocapsules. J. Am. Chem. Soc. 127, 1112–1113 (2005). 45. Barron, L. D. Can a magnetic field induce absolute asymmetric synthesis? Science 266, 1491–1492 (1994). 46. Rikken, G. L. J. A. & Raupach, E. Observation of magneto-chiral dichroism. Nature 390, 493–494 (1997). 47. Kitagawa, Y., Segawa, H. & Ishii, K. Magneto-chiral dichroism of organic compounds. Angew. Chem. Int. Ed. 50, 9133–9136 (2011). 48. Villari, V. & Micali, N. Light scattering as spectroscopic tool for the study of disperse systems useful in pharmaceutical sciences. J. Pharm. Sci. 97, 1703–1730 (2008). 49. Micali, N., Villari, V., Consoli, G. M. L., Cunsolo, F. & Geraci, C. Vesicle-tomicelle transition in aqueous solutions of amphiphilic calixarene derivatives. Phys. Rev. E 73, 051904 (2006). 50. Kemp, J. C. Polarized Light and its Interaction with Modulated Devices (HINDS International, Hillsboro, 1987).

Acknowledgements We thank P.W. Albers for technical assistance with the magnet set-up. This work was supported by EuroMagNET II under EU Contract No. 228043, PRIN 2008- 2008A9C4HZ and 20088NTBKR (Ministero dell’Istruzione, dell’Universita` e della Ricerca) and by the Stichting voor Fundamenteel Onderzoek der Materie financially supported by the Nederlandse Organisatie voor Wetenschappelijk Onderzoek.

Author contributions N.M. and L.M.S. initiated the project. N.M., H.E., P.C.M.C. and L.M.S. designed and realized the experimental set-up for chiral selection. N.M., H.E. and L.M.S. performed the chiral selection experiments. P.G.R. and P.C.M.C. developed the LD set-up and P.G.R. performed these experiments. N.M., H.E., P.G.R., P.C.M.C. and L.M.S. analysed the results. H.E., P.C.M.C. N.M. and L.M.S. co-wrote the paper. All the authors discussed the results and commented on the manuscript.

Additional information The authors declare no competing financial interests. Supplementary information accompanies this paper at www.nature.com/naturechemistry. Reprints and permission information is available online at http://www.nature.com/reprints. Correspondence and requests for materials should be addressed to L.M.S.



207