Amplification of Chirality by Adenosine Monophosphate-Capped Luminescent Gold Nanoclusters in Nematic Lyotropic Chromonic Liquid Crystal Tactoids.

Amplification of chirality across length scales is a key concept pertinent to many models aiming to unravel the origin of homochirality. Tactoids of lyotropic chromonic liquid crystals formed by DNA, dyes, and other flat ionic molecules in water in the biphasic nematic + isotropic regime turn out to be a particularly relevant system to investigate chirality transfer and amplification. Herein, we present experiments to determine the amplification of chirality by luminescent gold nanoclusters decorated with adenosine monophosphate inducing chiral nematic tactoids formed by disodium cromoglycate in water. Polarized optical microscopy investigations of the induced homochiral tactoids reveal that adenosine monophosphate shows a higher optical activity when bound to the surface of such gold nanoclusters in comparison to free adenosine monophosphate, despite a three-time lower overall concentration. Free adenosine monophosphate also induces the opposite chiral twist both in the bulk nematic phase as shown by induced thin film circular dichroism spectropolarimetry and in the tactoids in comparison to adenosine monophosphate bound to the gold nanocluster. Overall, these experiments demonstrate that lyotropic chromonic liquid crystal tactoids are powerful systems to image and quantify chirality amplification by key biological chiral molecules that would have played a role in the origin of homochirality.

Introduction

The scientific community at large is investing a great deal of effort to rationalize the origin of biological homochirality. The amplification of chirality on inorganic surfaces, several at the nanoscale, emerged as one of the potentially more important underlying concepts.[1] Homochirality, the single handedness of many key biological molecules, is ubiquitous in nature and no doubt the signature of life on our planet. All living organisms use almost exclusively l-amino-acids and d-sugars as building blocks for proteins and nucleic acids.[1,2] Mathematically or more precisely geometrically speaking, an object is said to be chiral if it cannot be precisely mapped onto its mirror image by any kind of rotation or translation (i.e., it is said to be nonsuperimposable).[3] Phenomenologically, chirality is observed at virtually all length scales in nature from massless subatomic particles[4] to our hands, plant tendrils,[5] and snail shells.[6]

The extent or length scale of chiral induction through space (i.e., the transmission of chirality from a chiral solute to its surrounding achiral hosts or environment) is in many cases extremely difficult or even impossible to measure. By taking advantage of another ubiquitous phenomenon in nature, the liquid crystalline state, we have a medium that is highly sensitive to chiral perturbations and permits visualization as well as quantification of the efficiency of chirality transfer.[7] Liquid crystals (LCs) are characterized by the self-assembly of predominantly anisometric building blocks forming soft condensed matter. LCs can form a wide variety of architectures (phases) that can be categorized by the degree of ordering in one-, two- and three-dimensional space and by the mode of formation, that is, by varying the temperature (thermotropic LCs) or by changing the concentration in a fluid/solvent (lyotropic LCs). Nematic LCs (N-LCs), typically composed of low-molecular weight rod-like organic compounds, are particularly useful to detect, measure, and visualize chirality on several levels, from atomic to macromolecular[8] to nanoscale particles capped with a chiral ligand shell.[9]

Experimental data shedding light on the spatial extent of chiral induction from a chiral nanoscale surface or nanoparticle are continuing to deliver astonishing results. Recent experiments from our group indicate that chiral ligand-capped gold nanoparticles (AuNPs), capped either with cholesterol[10] or with axially chiral binaphthyl derivatives,[11] outperform their organic molecular counterparts, inducing a tighter helical pitch, p, in the induced chiral nematic LC (N*-LC) phase at a one order of magnitude lower chiral molecule concentration. In addition, these chiral ligand-capped AuNPs performed this amazing feat consistently over larger NP distances, translating into larger chiral correlation lengths (transmission over larger distances). Such enhancement of through-space chirality found support from recent examples of demonstrated long-range through-space interactions between chiral molecules and plasmonic nanostructures[12,13] as well as enhanced anisotropy (or Kuhn’s dissymmetry) factors, g (g = Δε/ε, where Δε and ε are the molar circular dichroism (CD) and molar extinction coefficient, respectively) for chiral molecules in the vicinity of plasmonic nanostructures.[13,14] Further data from our group provided additional evidence for the hypothesis that desymmetrization of plasmonic nanostructures, that is, substituting AuNPs for gold nanorods (AuNRs), results in further through-space chirality enhancement and tighter p values at even lower overall concentrations of chiral organic molecules. The investigated cholesterol-capped AuNRs showed a very distinct aspect ratio-dependent chirality amplification, prompting their own helical assembly in the induced N*-LC host that we explained by a chiral feedback loop.[15]

Most key biological chiral molecules such as amino acids, sugars, nucleotides, DNA, and so forth are hydrophilic and dissolve/disperse well only in water or polar protic solvents. Chirality amplification of such molecules decorating the surface of metal nanomaterials can only be studied in aqueous lyotropic not the hydrophobic thermotropic LC systems (such as, for example, 5CB) used for the cholesterol- and binaphthyl-based chiral additives.

Well-suited and well-understood lyotropic LC systems are based on lyotropic chromonic LCs (LCLCs).[16−19] In LCLCs, multiring aromatic compounds in a certain concentration interval in water self-assemble into columnar stacks (such as H-aggregates of some dye molecules) ranging from just a few to several tens or hundreds of molecules in height.[20,21] These stacks are formed by segregation of polar or ionic groups located at the periphery of the multiaromatic cores. Most LCLC molecules are dyes, nucleic acids, or drug molecules such as disodium cromoglycate (DSCG, an asthma and general anti-inflammatory drug under the trade name Cromolyn, Figure a).[22]

Figure 1

Materials studied: (a) DSCG, (b) AMP, and (c) AMP-capped luminescent gold nanoclusters. (d) Phase diagram of DSCG in water. The dashed red line indicates the DSCG concentration investigated (15 wt %). (e, f) Schematics of tactoids formed by DSCG in H2O in the biphasic N + Iso region: (e) before (racemic) and (f) after addition of a chiral additive (here, AMP or AMP-AuNCs (homochiral tactoids)).

LCLCs commonly form two LC phases, the nematic or N-phase and the columnar M-phase.[18] The N-phase is remarkably sensitive to chiral additives leading to an induced chiral N*-phase upon addition of enantioenriched amino acids, sugars, or amino acid-capped (e.g., l-cysteine) AuNPs among others.[23] Earlier experimental studies revealed that chirality in N*-LCLC phases manifests itself in two ways, a twist of molecules within the stacks[16] that can be detected by induced circular dichroism (ICD) and a twist among stacks that can be assessed by ICD and visualized by crossed polarized optical microscopy (POM).[23] Our data for l-cysteine-capped AuNPs unambiguously supported chirality amplification by AuNPs capped with l-cysteine versus l-cysteine alone. Because consistent characteristic fingerprint textures were difficult to obtain to measure the helical pitch, p, at lower chiral additive concentration (critical especially for nanomaterials with a tendency to aggregate in LCLC phases), we used the molar ellipticity, Θ, measured by ICD spectropolarimetry. We determined that Θ was about five orders of magnitude more intense for the l-cysteine-capped AuNPs with an overall l-cysteine concentration of 10–4 wt % in comparison to the sample of DSCG in water doped with 12 wt % neat l-cysteine.[24]

Results and Discussion

To further investigate biologically significant key chiral molecules such as adenosine monophosphate, AMP (a nucleotide and a key building block of universal redox cofactors such as NAD+, NADP+, and FAD,[25]Figure b), and the potential amplification of chirality after immobilization of such nucleotides on plasmonic nanostructures (AMP-capped gold nanoclusters, AMP-AuNCs, Figure c), we here present a viable alternative in assessing chirality transfer efficiency and amplification using the biphasic nematic + isotropic (N + Iso) region formed by DSCG in water at temperatures above the bulk N-phase first described by Lavrentovich and co-workers[19] and schematically shown in Figures d–f.

Lavrentovich[19] and others[26,27] established that N-LCLCs in the biphasic N + Iso region form elongated tactoids. The spatial confinement of the N-phase in its isotropic environment causes a structural twist even when the material is achiral or no chiral additive is added. There is simply an equal amount of tactoids with left- and right-handed structural twist, that is, the tactoid array is racemic as shown in Figure e. However, even small amounts of chiral guest molecules transform the racemic array into homochiral tactoids with an exclusively right- or left-handed twist that can be visualized by POM (Figure f).[19]

To apply this to AMP, either neat or capping a nanomaterial surface, we first synthesized AMP-capped AuNCs and characterized both the free ligand and the purified AuNCs. The synthesis of the Au(I)NCs followed a procedure reported by Pérez-Prieto[28] and all spectroscopic characterization data (UV–vis and photoluminescence spectra, Figures a,b) were in perfect agreement with those given in ref (28). In addition, we here provide solution CD spectra for the free ligand and the AuNCs (Figure c) as well as thermogravimetric analysis (TGA) data and transmission electron microscopy (TEM) images for the AMP-AuNCs (Figures d,e). TGA data were used as an experimental value to support theoretical calculations of the AuNC ligand weight fraction needed to later compare the optical activity data (calculated: 34.5 wt %, TGA: 31.6 wt %) using an established protocol and assuming an approximate composition of Au∼144L∼50[29] and the binding of the AMP moiety via the amino group and N7 of adenine (see Figure b) as described by Pérez-Prieto.[28] The solution CD spectra show a negative CD band for free AMP and positive bands for the AMP-AuNCs (characteristic set of resolved peaks centered around 300 nm but no plasmonic CD band). TEM image analysis revealed that the average core diameter, Dcore, of the AMP-AuNCs is about 1.8 ± 0.27 nm with fairly narrow size distribution (Figure f).

Figure 2

(a) UV–vis spectrum of AMP (red) and AMP-AuNCs (green), inset shows a weak plasmon band at 545 nm for the AuNCs, (b) emission spectrum of AuNCs in water recorded at λexc = 300 nm (green) and its excitation spectrum (blue) at λem = 477 nm, (c) solution CD spectra in water of AMP (red) and AMP-AuNCs (green), (d) TGA plot for AMP-AuNCs, (e) TEM image of AuNCs, (f) size histogram from image analysis, and (g) photograph of AMP-AuNCs in water (excitation at 366 nm).

Next, we prepared mixtures of AMP and AMP-AuNCs in 15 wt % DSCG in water. The concentration of each chiral additive in these mixtures was adjusted to 0.5, 1, 5, and 7 wt %. In the first set of experiments, performed at room temperature (20 °C), we studied the induced N*-LCLC bulk phase, first by POM and then for specific mixtures also by ICD spectropolarimetry. POM images of the bulk N*-phases (Figure ) provide several important information. First, the highly colorful birefringent textures suggest the presence of the induced N*-phase, and several regions within each image suggest the formation of fingerprint textures (at least locally). For the AMP doped mixtures (Figures a–d), the mixture with the highest AMP concentration shows a texture indicative of a columnar or biphasic N* + M phase,[18] which might be aided by insertion of adenine into the DSCG stacks or intermolecular hydrogen bonding between adenine moieties coordinated by ionic interactions to adjacent DSCG stacks (see Figure S1 in Supporting Information); all other POM images for lower AMP concentrations show textures typical for induced N*-LCLC phases.

Figure 3

Polarized optical photomicrographs (crossed polarizers) of mixtures of 15 wt % DSCG in water (a–d) doped with AMP at: (a) 0.5, (b) 1, (c) 5, (d) 7 wt %, and (e–h) doped with AMP-AuNCs at: (e) 0.5 wt % in rectangular capillary under an applied magnetic field of 0.05 T (additional images are provided in Supporting Information, Figure S2), (f) 1, (g) 5, and (h) 7 wt %. Images (f–h) show mixtures between glass slides treated to induce planar anchoring (cell gap: 180 μm). Yellow arrows in (f–h) indicate AuNC aggregation.

POM images for the AMP-AuNC-doped DSCG–water mixtures, even after alignment in an external magnetic field and using flat rectangular capillaries did only produce fingerprint textures in a few areas of each image, making a direct measurement of the helical pitch of the induced N*-phase challenging (Figures e–h). However, the darker regions in the POM images for mixtures with AMP-AuNC concentrations ≥1 wt % increasingly indicated AuNCs aggregation (particulate aggregates highlighted by yellow arrows in Figures f–h) as discussed earlier. Therefore, optical measurements on the tactoids formed in the N + Iso region and ICD spectropolarimetry of the N*-LCLC bulk were only performed for the two mixtures of DSCG in water doped with 0.5 wt % of AMP or AMP-AuNCs. The differences in AMP concentration in these mixtures are 2.7 × 10–2 mol % for AMP and 9.9 × 10–3 mol % for AMP capping the AuNCs, on the basis of the calculations referred to above.

The ICD spectra shown in Figure are characterized by broad bands centered around 340 nm that appear to be from two overlapping bands at 310 and 350 nm, observed in the UV–vis spectrum of the N-phase formed by DSCG in water[24] similar to ICD spectra reported earlier for induced lyotropic N*-LCLC phases of DSCG in water.[24]

Figure 4

ICD spectra of DSCG in H2O (15% by weight) doped with 0.5 wt % of: (a) AMP and (b) AMP-AuNCs. Solid lines are linear dichroism spectra at the indicated sample rotation angles (in 45° intervals), dotted lines are the sum ICD spectra.

The sum ICD spectrum for DSCG-water doped with 0.5 wt % AMP shows a broad negative band with a maximum sum molar ellipticity value of ΘAMP = −1.3 × 108 °·cm2·dmol–1 (about three orders of magnitude lower than the l-cysteine-capped AuNPs reported earlier (Θ = 1.5 × 1011 °·cm2·dmol–1)),[24] whereas the mixture doped with 0.5 wt % AMP-AuNCs shows several, more defined positive bands centered around the same wavelengths (here, the 310 and 350 nm bands are better resolved with maximum ΘAMP-AuNC = +2.8 × 1014 °·cm2·dmol–1 (also one order of magnitude lower than Θ = +4.5 × 1015 °·cm2·dmol–1)). The difference in sign indicates that AMP and AMP-AuNCs induce N*-LCLC phases with opposite helical twist (both within and among the DSCG stacks). The additional band at shorter wavelength (280 nm) could be related to the AMP-AuNC adenine decorating the helical DSCG stacks (related to the band in solution CD spectrum of AMP-AuNCs in Figure c). Because the exact contributions to these ICD signals for the AMP-AuNC-doped DSCG–water mixtures are difficult to ascertain, we then proceeded to study the biphasic N + Iso region by elevating the temperature during POM investigations to about 37.5 °C. The formed tactoids were then evaluated as described by Lavrentovich and co-workers.[19]

The sign of left- or right-twisted tactoid (twist of the director) is first established by rotating the analyzer with respect to its otherwise crossed 90° position with the polarizer by an angle ±γ (extinction angle) to find the two maxima of light transmission (Figure ).[19] This extinction angle γ for AMP was 103 ± 1° (right-twisted tactoids) and the AMP-AuNCs 75 ± 1° (left-twisted tactoids), supporting the opposite sign of the induced bulk N*-LCLC phase determined by ICD measurements. Plotting the transmitted intensity of the polarized light as a function of the extinction angle γ between polarizer and analyzer, and considering the width of the tactoids (measured by POM) leads to the twist angle τ (as described in more detail in Experimental Section). The optical activity η (in °/m × wt %) is then calculated by η = |τ|/cd, where c is the concentration and d is the thickness of the tactoids.[19]

Figure 5

Optical photomicrographs of tactoids formed by DSCG (15 wt % in H2O) doped with the two chiral additives at 0.5 wt % each on cooling from the isotropic liquid phase at 37.5 °C: (a) AMP causes dark tactoid centers at γ = 103 ± 1°, (b) AMP-AuNCs cause dark tactoid centers at γ = 75 ± 1°. Data of transmission vs γ and fits (for details see Experimental Section) for (c) AMP and (d) AMP-AuNCs in DSCG–water mixtures. A similar set of data using a white light source is given in Supporting Information (Figure S3).

The values for τ, η, and the actual concentration of AMP in both cases (doped either with AMP or with AMP-AuNCs) are given in Table .

Table 1

Twist Angle (τ), Optical Activity (η), and Overall Concentration of AMP in DSCG–H2O Mixtures (in mol %) Doped with 0.5 wt % of the Chiral Additives

additive	|τ| (°)	η (°/m × wt %)	cAMP (mol %)
AMP	13 ± 1	1.15 ± 0.09 × 106	2.7 × 10–2
AMP-AuNC	15 ± 1	1.30 ± 0.09 × 106	9.9 × 10–3

These values show that a nearly three-time lower AMP concentration in the case of the AMP-AuNC-doped sample leads to slightly higher values in the optical activity. This evidently indicates that the AuNCs amplify chirality, and therefore, the chirality transfers to the N-LCLC tactoids formed in the biphasic region of DSCG in water.

Conclusions

We have demonstrated that in addition to other key biological molecules such as cholesterol and amino acids like l-cysteine, AMP chirality is amplified after attachment to a plasmonic nanomaterial surface. We have also shown that straightforward POM techniques can be used in both thermotropic and lyotropic LC systems to visualize and quantify chirality amplification. Here, specifically nematic LCLC tactoids formed in the biphasic N + Iso region of DSCG in water were imaged and used to calculate and compare the optical activity of neat AMP and AMP capping the surface of luminescent AuNCs. It appears that this amplification of chirality through space by chiral molecules affixed to nanostructures that disperse in the host medium is a universal phenomenon and that condensed LC phases, thermotropic, lyotropic, or in the form of tactoids in LCLCs, are a superb medium to visualize these amplification effects. It seems likely that the amplification is a result of the chiral molecules acting together, creating significant local chirality as demonstrated earlier by calculations of a pseudoscalar chirality index.[15] The N*-LCLC tactoids described here are a particularly sensitive platform to assess chirality amplification, even for relatively weak chiral inducers such as AMP, because they are already chiral (yet racemic) before the addition of the chiral additives. The added chirality simply tips the balance toward one of the enantiomeric (enantiopure) tactoid forms, leading to higher optical activity values than those obtained from CD data in isotropic solutions or those deduced from the helical pitch derived from fingerprint textures of the bulk N*-LCLC phase.[19] Hence, tactoids formed in the biphasic region of chromonic LCs are the most sensitive choice for detecting chirality amplification, particularly for weak chiral additives such as AMP (as suggested by the lower molar ellipticity values from CD experiments in the bulk N*-LCLC phase) for which the helical twisting power (HTP or βM) cannot be acquired by conventional POM studies (including Grandjean–Cano wedge cells)[30] because the helical pitch is too large. In each case, free AMP and AMP-AuNCs, the adenine moiety, which does not participate in the process, either bound as a bidentate ligand to the AuNC surface or likely inserted into the DSCG stacks, facilitates interactions between the active chiral ribose units and the DSCG stacks via a combination of ionic and hydrogen bonding interactions.

Unfortunately, the inherent luminescence of DSCG at excitation wavelengths suitable for excitation of the AuNCs[28] prevented further fluorescence confocal microscopy studies elucidating the spatial distribution of the AMP-AuNCs in the induced N*-LCLC phase (bulk or tactoids). Future experiments will now focus on the other four nucleobases and their mono- to triphosphates, particularly in the light of their remarkable self-assembly including the formation of LC phases via base pair formation and end-to-end stacking.[31−35]

Experimental Section

Materials

DSCG, tetrachloroauric(III) acid (HAuCl4), poly(ethylene glycol) PEG (MW = 3350 g·mol–1), AMP, and the mild reducing agent 4-(2-hydroxyethyl)-1-piperazine-ethanesulfonic acid (HEPES)[36] used for the AMP-AuNC synthesis were purchased from Sigma-Aldrich.

Methods

UV–vis, solution CD, and induced thin film CD measurements were performed using an OLIS Cary spectrophotometer suite (spectrophotometer grade solvents were used). ICD measurements were performed using an OLIS Cary spectrophotometer. Thin films for ICD spectropolarimetry were prepared between two quartz substrates separated by ∼10 μm Kapton tape spacers. Samples were rotated in 45° intervals from 0 to 315° (in the plane normal to the incident light beam) to differentiate CD absorption from linear dichroism and birefringence. TGA was done with TA Instruments TGA Q500 (New Castle, DE, USA). The heating rate was set at 10 °C·min–1. Photoluminescence spectra were collected using a Varian Cary Eclipse with variable excitation wavelengths. For sample preparation, DI water (resistivity 18.2 MΩ, Barnstead Nanopure) was used. TEM was done using an FEI Tecnai TF30 ST TEM instrument at an accelerating voltage of 300 kV. Samples were prepared by dip-coating a carbon-coated copper grid (400 mesh) in an aqueous solution (∼1 mg·mL–1) of the AuNCs. For the investigation of induced bulk N*-LCLC phase, an Olympus BX-53 polarizing optical microscope equipped with a Linkam LTS420E heating/cooling stage was employed. Substrates were either precleaned glass substrates coated with the polyimide SE-7511L (Nissan Chemical Industries, Ltd) and rubbed unidirectionally to align the axis of tactoids (cell gap: 180 μm using Mylar spacers) or rectangular capillaries (nominal gap: 250 μm, Vitrocom, NJ). All mixtures were filled by capillary force and sealed carefully with epoxy glue to prevent any evaporation. The mixtures were heated to the isotropic liquid phase, then cooled down to reach the coexistence N + Iso region, and overtime the tactoids stabilized (at 37.5 °C). The twist angle of the director in the middle of tactoids and the azimuthal orientation of tactoid height were measured by POM (Olympus BX-53 equipped with Linkam LTS420E). PEG (0.8 wt %, MW = 3,350 g·mol–1) was added to stabilize the tactoids (keeping the concentration of DSCG in water at 15 wt % and the concentration of the chiral additives at 0.5 wt % with respect to the total mass of DSCG and H2O). The cell thickness was adjusted to 180 μm by Mylar spacer films. The optical transmission was measured with respect to the extinction angle, γ, between polarizer and analyzer (between 0 and 180°), which is then fit with the Mauguin approximation[37] because the twist elastic constant is about one order of magnitude smaller than those for splay and bend[38]where , β = γ – τ, τ is the angle of the tactoid director twist, λ = 550 nm is the wavelength of light, Δn = −0.026 is the birefringence, and d = 22.5 μm is the height of the tactoids.[39] By measuring the intensity of the transmitted light (using Ocean Optics USB4000), the twist angle of the tactoids is obtained.