Introduction

Optically pure amino acids play a critical role in elucidating the fundamental principles of life, developing safe and effective pharmaceuticals, facilitating chemical synthesis, and advancing applications in fields such as food nutrition and industrial biotechnology1,2,3,4,5. Therefore, the enantioselective recognition and separation of amino acids are of great importance6,7,8. A variety of analytical and detection techniques have been established for determining the enantiomeric composition of chiral amino acids, including nuclear magnetic resonance (NMR), ultraviolet/visible (UV/vis) spectroscopy, circular dichroism (CD) spectroscopy, high-performance liquid chromatography (HPLC), and fluorescence/phosphorescence methods9,10,11,12,13. Among them, enantiomer recognition based on fluorescence intensity and wavelength has garnered increasing attention owing to its high sensitivity, cost-effectiveness, and ease of operation14,15,16,17,18,19,20. To date, various fluorescent probes based on chiral molecules, supramolecular assemblies, and inorganic nanomaterials have been developed for the detection of amino acid enantiomers21,22,23. However, most of these probes exhibit relatively long response times (typically exceeding 30 min) toward specific enantiomers, which hinders their broader applications24,25,26,27. Furthermore, although these probes demonstrate effectiveness in fluorescence-based chiral recognition, few possess the capability to enable practical chiral separation of amino acid enantiomers.

Enantiomer separation is one of the primary methods for obtaining highly optically pure amino acids with reliable stereochemical integrity28,29. To date, several separation strategies have been established, including chromatographic techniques, membrane separation technology, and selective crystallization and precipitation30,31,32. Among them, selective precipitation is the simplest and most cost-effective methods for achieving large-scale enantiomeric separation of amino acids30. If chiral fluorescent probes are introduced as chiral selectors to crystallize or precipitate specifically with a single enantiomer in racemic mixtures of amino acids, the chiral separation process can be visualized using fluorescence33. This method has the potential to greatly simplify the operational procedures of chiral separation while improving overall separation efficiency34. In addition, due to the good water solubility and low liposolubility of amino acids, chiral analysis and separation in pure water phase have significant advantages35,36,37,38. Recently, our team and other research groups have reported some fluorescent probes that can recognize enantiomers of free amino acids in aqueous solutions through the changes in fluorescence intensity or wavelength39,40,41,42. However, due to significant challenges in designing fluorescent probes that can selectively aggregate and precipitate with specific amino acid enantiomers, there are currently limited reports on simultaneous enantioselective detection and separation of amino acids in aqueous media using fluorescence-based methods43,44,45,46.

Herein, we strategically introduced a morpholine-containing quaternary ammonium ion moiety into the 2′-position of 1,1′-bi-2-naphthol (BINOL)-based chiral aldehydes, thereby synthesizing a cationic chiral fluorescent probe, (S)/(R)-Y3 (Fig. 1a). This modification not only enhances water solubility but also facilitates electrostatic interactions with amino acids. The fluorescent probe (S)/(R)-Y3 enables rapid detection of D/L-methionine (Met) within 100 s through distinct differences in fluorescence intensity (Fig. 1b). Moreover, this chiral recognition process can be dynamically visualized via luminescent color and intensity changes, which result from aggregation-induced emission (AIE) upon the probe’s binding to zinc ion and a specific enantiomer of Met. We elucidated the mechanism of chiral recognition and high enantioselectivity using CD, scanning electron microscopy (SEM), density functional theory (DFT) calculations, NMR, and mass spectrometry. Furthermore, based on differences in luminescent intensity and color, enantiomer pairs of 17 free amino acids were successfully distinguished, establishing this fluorescent sensing method as one of the most efficient approaches in reported luminescent chiral recognition studies38,47,48,49,50, with response times as short as tens of seconds (Fig. 1c). Notably, the interaction between the probe and a single amino acid enantiomer induces aggregation and subsequent precipitation, enabling the separation of the enantiomer through simple filtration, with the process being traceable via fluorescence (Fig. 1d). Comparison with chiral HPLC results confirms the probe’s excellent capability for chiral amino acid separation. This integrated detection-separation strategy not only delivers immediate visual and fluorescent signals but also allows for the direct isolation of highly optically pure single enantiomers via a simple filtration process. This study demonstrates the separation of pure amino acid enantiomers from racemic mixtures within minutes, achieving rapid and efficient results via a fluorescent probe strategy.

Fig. 1: Constructing fluorescent molecules for enantioselective recognition and chiral separation of free amino acids.
Fig. 1: Constructing fluorescent molecules for enantioselective recognition and chiral separation of free amino acids.
Full size image

a Fluorescent molecules designed for improving the chiral recognition sites and generating electrostatic aggregation through the integration of an aldehyde group and a morpholinium cation moiety, enabling synergistic covalent and electrostatic interactions. b The (S)-Y3 molecule rapidly discriminates between D- and L-amino acids in aqueous solution via the AIE effect (recognition time <100 s), offering several advantages in visual chiral recognition and separation. c A review of luminescent chiral sensing techniques for amino acids (with chiral recognition ratios of at least 3) demonstrates that the present approach outperforms all previously reported methods in terms of sensing speed and the range of applicable substrates. d Schematic diagram of enantioselective recognition and separation of amino acids by chiral fluorescence probe (S)-Y3, featuring a photograph of the aggregate precipitate taken under visible light. AAs amino acids.

Results

Molecular synthesis and fluorescence recognition properties

Morpholine and its quaternary ammonium cation were selected as suitable recognition and reactive units, and were introduced at the 2′-position of BINOL-based chiral aldehydes. Starting from readily available (S)-BINOL as the chiral precursor, a series of chiral fluorescent molecules—(S)-Y1, (S)-Y2, and (S)-Y3—were synthesized through an efficient and streamlined synthetic route involving multiple substitution reactions (Supplementary Figs. 1 and 2). The structures of these compounds were fully characterized using ¹H NMR, ¹³C NMR, and high-resolution mass spectrometry (HR-MS) (See Supplementary Material). Their corresponding enantiomers were prepared using the same synthetic protocol starting from (R)-BINOL.

Initially, we evaluated the chiral fluorescent recognition performance of probe (S)-Y3 toward 19 common chiral amino acids and observed that it displayed the highest selectivity and sensitivity for Met (Supplementary Fig. 3). Building on this result, Met was selected as a representative analyte for a comprehensive investigation of the probe’s optical properties, chiral recognition capability, and potential for enantiomeric separation. To investigate the optical behavior of (S)-Y3 and (R)-Y3 upon interaction with Met (Fig. 2a), the absorption and CD spectra of the probes (S)-Y3 and (R)-Y3 were measured after the addition of zinc ions and Met enantiomers in aqueous solution. The absorption spectra of the (S)/(R)-Y3 probes and their mixtures with amino acid enantiomers exhibited differentiated absorption peaks and absorbance values (Fig. 2b). The absorption spectra of the probe differ following interaction with L-Met and D-Met, indicating that the L- and D-enantiomers form diastereomeric complexes with the probe, which exhibit distinct electronic structures and spatial conformations. CD spectra of (S)-Y3 and its complexes, including (S)-Y3 + Zn²⁺ and (S)-Y3 + Zn²⁺ + D-/L-Met, all showed a positive-to-negative Cotton effect inversion in the 210–310 nm region. Mirror-image CD spectra of comparable intensity were observed for the corresponding (R)-Y3 (Fig. 2c). Furthermore, the fluorescent response of (S)-Y3 to the L- and D-enantiomers of Met was evaluated in a pH 8.8 buffer solution. As shown in Fig. 2d, the (S)-Y3 probe exhibited minimal fluorescence response to Zn²⁺ alone. Upon treatment of the (S)-Y3 + Zn²⁺ mixture with D-Met, only a slight fluorescence enhancement at 550 nm was observed. In contrast, the addition of L-Met resulted in a significant fluorescence enhancement at 530 nm. The enantioselective fluorescent enhancement ratio (ef) reached as high as 43.1 [ef = (IL–I₀)/(ID–I₀), where I₀ represents the fluorescence intensity of (S)-Y3 + Zn²⁺ at 530 nm, and IL and ID denote the fluorescence intensities at 530 nm after addition of L-Met and D-Met, respectively]. Additionally, (R)-Y3 showed almost no fluorescence response to L-Met but a strong enhancement at 530 nm upon addition of D-Met—exactly opposite to the behavior of (S)-Y3 (Fig. 2e). These results indicate that the host molecule (S)-Y3 and its enantiomer (R)-Y3 exhibit chiral selectivity toward L-Met and D-Met, respectively. Moreover, the fluorescence quantum yield of (S)-Y3 + Zn2+ + L-Met was 10.5%, which is significantly higher than that of (S)-Y3 + Zn2+ + D-Met (3.8%) (Fig. 2f and Supplementary Figs. 4 and 5). In contrast, the fluorescence quantum yield of (R)-Y3 + Zn2+ + D-Met was 8.0%, which is notably higher than that of (R)-Y3 + Zn2+ + L-Met (3.4%). These results from spectroscopic studies indicate that probe (S)/(R)-Y3 exhibits excellent chiral Met recognition capability. We further explored the role of Zn²⁺ in the process of the probe’s recognition of amino acids. In the absence of Zn²⁺, the probe was unable to effectively distinguish the enantiomers of Met (Supplementary Fig. 6). This indicates that Zn²⁺ actively contributes to the formation of a rigid microenvironment through synergistic coordination, which is crucial for effective chiral recognition.

Fig. 2: Optical properties in enantioselective recognition.
Fig. 2: Optical properties in enantioselective recognition.
Full size image

a The molecular structures of (S)/(R)-Y3 and D/L-Met. UV–vis absorption (b) and CD (c) spectra of (S)/(R)-Y3, (S)/(R)-Y3 + Zn2+, (S)/(R)-Y3 + Zn2+ + D-/L-Met in water (33 μmol/L, pH 8.8). d Fluorescence spectra of (S)-Y3, (S)-Y3 + Zn2+, (S)-Y3 + Zn2+ + D-/L-Met, (excited by 420 nm) at concentration of 33 μM in water (pH 8.8). e Fluorescence spectra of (R)-Y3, (R)-Y3 + Zn2+, (R)-Y3 + Zn2+ + D-/L-Met, (excited by 420 nm) at concentration of 33 μM in water (pH 8.8). f The fluorescence quantum yield of (S)/(R)-Y3, (S)/(R)-Y3 + Zn2+, (S)/(R)-Y3 + Zn2+ + D-/L-Met in water (pH 8.8). g Visualization of the luminescence color changes in (S)-Y3 + Zn²⁺ upon interaction with D-Met and L-Met under 365 nm UV irradiation following a specific reaction time. h Fluorescence intensity at different time points following the reaction of (S)-Y3 + Zn²⁺ with D-Met and L-Met. i Fluorescence spectra of (S)-Y3 + Zn2+ + L-Met in acetonitrile–water solvent with varying composition ratios. j The fluorescence intensity of (S)-Y3 + Zn2+ + L-Met under varying acetonitrile-to-water ratios presented in (i). k Column charts of fluorescence intensity at 530 nm of (S)-Y3 (33 μmol/L in buffer with different pH values, 1 equiv) + Zn2+ (in H2O, 2 equiv) reacting with L-/D-Met (25 equiv) dissolved in buffer with different pH values; blue line is ef values versus the pH values (λexc = 420 nm; slits: 3/3 nm; standing time: 15 min).

The enantioselective fluorescence response time of chiral fluorescent probes toward Met was further investigated, as shown in Fig. 2g. Based on the fluorescence visualization images obtained under UV excitation, probe (S)-Y3 + Zn2+ was observed gradually transitions from non-luminescent to yellow luminescent, and finally to bright green luminescence upon interaction with L-Met. Surprisingly, this transformation can occur within tens of seconds. In contrast, under the same time frame, there is almost no change in fluorescence intensity when the probe interacts with D-Met. This dynamic video of rapid chiral recognition is provided in the Supplementary Movies 1–4. Fluorescence spectroscopy further revealed that the fluorescence intensity ratio of the probe interacting with Met enantiomers reaches up to 8 within 300 s (Fig. 2h). Similarly, upon interaction between (R)-Y3 and D-Met, a comparable transition from non-luminescent to yellow and then to bright green luminescence occurs within tens of seconds (Supplementary Fig. 7). These results demonstrate that the corresponding D/L-Met enantiomers can be rapidly recognized by the (R)/(S)-Y3 chiral fluorescent probes.

Given the strong fluorescence emission of (S)-Y3 + Zn²⁺ + L-Met, we further investigated the AIE properties of the (S)-Y3 probe after reaction with Zn²⁺ and L-Met. The probe-amino acid complex was introduced into acetonitrile-water mixtures at varying ratios, as illustrated in the Fig. 2i. When the water content (Fw) increased from 10% to 70%, a slight decrease in fluorescence intensity was noted (Fig. 2j). However, upon increasing the water ratio from 70% to 90%, a significant enhancement in fluorescence intensity was observed—a hallmark of typical AIE behavior. In comparison, the (S)-Y3 + Zn²⁺ + D-Met complex displayed weaker AIE characteristics (Supplementary Fig. 8). Therefore, the AIE observed in the (S)-Y3 probe upon interaction with L-Met contributes to enhanced fluorescent enantioselectivity. This phenomenon can be attributed to restricted intramolecular motion within the (S)-Y3 + Zn²⁺ + L-Met complex in the aggregated state, which effectively suppresses non-radiative energy dissipation and promotes the radiative transition pathway, thereby enhancing fluorescence emission.

The influence of pH on the enantioselectivity of the probe was further investigated. L-Met or D-Met, dissolved in buffer solutions with pH values of 7.0, 7.4, 8.2, and 8.8, were respectively added to the mixtures of (S)-Y3 and Zn²⁺ in their corresponding buffers. As shown in the Fig. 2k, the probe (S)-Y3 exhibited no fluorescence response to either L-Met or D-Met under neutral conditions at pH 7.0. At pH 7.4, the probe displayed slight enantioselective fluorescence enhancement toward Met enantiomers, but the signal intensity was weak, with an ef value of only 1.8. When the pH was increased to 8.2 and 8.8, significantly stronger fluorescence was observed for the mixture containing L-Met compared to that with D-Met. Excellent enantioselectivity was achieved at pH 8.2 and 8.8, with ef values reaching 28.5 and 43.1, respectively. Regarding the influence of higher pH (e.g., pH 10.0 and 11.0), we examined the fluorescence spectra of (S)-Y3 + Zn²⁺ + L-Met under these conditions (Supplementary Fig. 9). The results showed that the enantioselective response to Met at pH 10.0 was significantly diminished compared to that at pH 8.8. To assess the stability of the imine bond formed between the aldehyde group of the probe and the amino group of the amino acid across varying pH levels, we monitored the formation and stability of the (S)-Y3 + Zn²⁺ + L-Met complex using HPLC (Supplementary Fig. 10). A clear and distinct peak corresponding to (S)-Y3 + Zn²⁺ + L-Met was observed at a retention time of approximately 7.4 min under neutral and alkaline conditions (pH 7.0, 7.4, 8.0, 8.8, and 10.0). In contrast, no such peak was detected under acidic conditions (pH 6.0). These findings indicate that the (S)-Y3 + Zn²⁺ + L-Met complex either fails to form or rapidly degrades under acidic conditions, while remaining stable and readily detectable under neutral to alkaline conditions.

To investigate the interaction mechanism between (S)-Y3 and Met that triggers its enantioselective fluorescent response, structurally similar compounds (S)-Y1 and (S)-Y2 (Fig. 1a) were used as control agents and allowed to react with L-Met and D-Met under identical conditions. Both (S)-Y1 and (S)-Y2 exhibited negligible fluorescent responses toward either L- or D-Met dissolved in pH 7.0, 7.4, 8.2, and 8.8 buffer (Supplementary Figs. 11 and 12). The results indicated that neither of the control compounds exhibited any significant fluorescence change when combined with L- or D-Met in either neutral or alkaline buffer solutions. Given the potential influence of the limited water solubility of (S)-Y2 on its interaction with amino acids in aqueous media, we further evaluated its fluorescent response toward L- and D-Met in DMSO/H2O mixtures. Although (S)-Y2 showed fluorescence enhancement in response to both enantiomers in pH 7.0, 7.4, 8.2, and 8.8 buffers, no enantioselectivity was detected (Supplementary Fig. 13). Moreover, L- and D-Met solutions at pH 7.0 and 7.4 elicited stronger fluorescence than those at pH 8.2 and 8.8. The high enantioselectivity of (S)-Y3 toward chiral Met arises not from a single functional group, but from the synergistic interplay among the aldehyde, morpholinium, and BINOL chiral scaffolds. The aldehyde group serves as the covalent anchor for the amino acid and directs its orientation in the formed covalent complex, the morpholinium moiety enhances binding affinity and fine-tunes the local spatial and electronic environment, and the BINOL scaffold enables precise chiral discrimination.

Enantioselective aggregation

Next, we further investigated the aggregation behavior and chiral recognition properties of the probes after interacting with chiral amino acids through visual monitoring under natural light. As shown in Fig. 3a, upon the addition of L-Met to the aqueous solution of (S)-Y3 + Zn2+, the solution initially exhibited a clear yellow color under visible light conditions. Within 30 s, gradual solid precipitation began to form, and the solution transformed into a suspension within 120 s. Under identical conditions, the (S)-Y3 + Zn2+ solution containing D-Met remained clear without any signs of aggregation or precipitation. Similarly, (R)-Y3 + Zn2+ rapidly formed a solid precipitate when reacted with D-Met within 100 s, whereas no aggregation was observed in the L-Met solution (Supplementary Fig. 14). These observations align well with the fluorescence spectra and visual results, demonstrating that the chiral fluorescent probes (S)/(R)-Y3 can swiftly recognize specific Met enantiomers and induce solid precipitation upon aggregation. Subsequently, the solid precipitate from the (S)-Y3 + Zn2+ + L-Met and (R)-Y3 + Zn2+ + D-Met mixture was filtered out, resulting in a colorless and transparent filtrate (Fig. 3b and Supplementary Fig. 15), which changes from bright green luminescence to non-emissive (appearing colorless under ambient light). And the fluorescence spectroscopy revealed that the fluorescence intensity of the filtrate was nearly zero.

Fig. 3: Characterization for selective aggregation and chiral recognition.
Fig. 3: Characterization for selective aggregation and chiral recognition.
Full size image

a Visualization of the color and state changes in the (S)-Y3 + Zn2+ solution with D-Met and L-Met under natural light at various interaction time points. b Visual images of the aqueous solution of (S)-Y3 (99 μM, 1 equiv) + Zn2+ (2 equiv) + L-Met (25 equiv) and the filtered solution under natural light and UV light irradiation (365 nm). Fluorescence spectra of (S)-Y3 + Zn2+ + L-Met and its filtered solution. c SEM images and EDS of the mixutre (S)-Y3 (0.67 mmol/L, 1 equiv) + Zn2+ (2 equiv) treated by L-Met (25 equiv) and D-Met (25 equiv). d 1H NMR spectra of the (S)-Y3 (0.95 mmol/L), (S)-Y3 (0.95 mmol/L, 1 equiv) + Zn2+ (2 equiv), (S)-Y3 (0.95 mmol/L, 1 equiv) + Zn2+ (2 equiv) + L-Met (25 equiv), (S)-Y3 (0.95 mmol/L, 1 equiv) + Zn2+ (2 equiv) + D-Met (25 equiv). Zn(OAc)2 were dissolved in D2O, (S)-Y3, L- and D-Met were dissolved in pH 8.8 BICINE buffer prepared by D2O. And the aggregations in the mixtures of (S)-Y3 + Zn2+ + L-Met were filtered, followed by being dissolved in DMSO-d6 before 1H NMR measurement. e HR-MS of the reaction mixture of (S)-Y3 (33 μmol/L in pH 8.8 BICINE buffer, 1 equiv) with L-Met (in pH 8.8 BICINE buffer, 25 equiv) and Zn(OAc)2 (in H2O, 2 equiv), the precipitation is dissolved in MeOH.

The enantioselective aggregation of (S)-Y3 toward Met was further investigated through SEM and energy dispersive spectrum (EDS). As depicted in Fig. 3c, the solid aggregates (precipitate) isolated from the L-Met-containing system appeared as flat sheets, while those forming from the D-Met containing mixture appeared as much smaller-sized bands. Combining the SEM results with the fluorescent response and the enantioselective aggregation, it could be speculated that comparing with D-amino aicds, the stereochemical configuration of L-enantiomers of amino acids could better match with (S)-Y3 to form assemblies in a larger size, thus resulting in enhanced emission due to the AIE effect. The EDS results demonstrate that the sulfur atom of methionine is uniformly distributed throughout the aggregates, indicating the formation of a well-defined and homogeneous mixed aggregate between the probe and Met, rather than self-aggregation of either individual component.

In order to further understand the interaction between (S)-Y3 + Zn2+ and Met, 1H NMR and HR-MS analyses were performed on the reaction mixtures. Figure 3d depicted the 1H NMR results of (S)-Y3, (S)-Y3 + Zn2+, (S)-Y3 + Zn2+ + L-Met, and (S)-Y3 + Zn2+ + D-Met. (S)-Y3 gave a singlet at δ 10.22 for its aldehyde proton, a singlet at δ 8.60 for the naphthyl proton ortho to the aldehyde group. When (S)-Y3 mixed with 2 equiv. Zn2+, these two singlets shifted to δ 10.19 and 8.64, respectively, which attributed to the coordination with Zn2+. Upon addition of 25 equiv. L- or D-Met to the mixture of (S)-Y3 + Zn2+, the singlet at δ 10.22 disappeared, which indicated that aldehyde group of (S)-Y3 reacted with amino group of Met. In addition, comparing with (S)-Y3 + Zn2+ + D-Met, (S)-Y3 + Zn2+ + L-Met showed more peaks for naphthyl ring proton at δ 7 ~ 8, which could be inferred that (S)-Y3 + Zn2+ and L-Met might form complexes with different aggregation degrees. We further established a 1:1 stoichiometric ratio between probe (S)-Y3 and Met in the complex under precipitation conditions, as determined by 1H NMR spectroscopy (Supplementary Fig. 16). Moreover, the molecular ion peak of 318.2325 was found in the HR-MS of (S)-Y3 + Zn2+ + L-Met (Fig. 3e), but not appeared in the mixture of (S)-Y3 + Zn2+ + D-Met (Supplementary Fig. 17). The peak at m/z 318.2325 could be attributed to the product (S)-Y3 + Zn2+ + L-Met, which could be the monomer of electrostatic aggregation.

Chiral recognition and optical sensing mechanism

DFT calculations were employed to further investigate the aggregation behavior and luminescence mechanism of the chiral amino acid recognition by the (S)-Y3 probe. The stability of the molecular structure formed upon binding of (S)-Y3 + Zn2+ with Met enantiomers was assessed using an independent gradient model based on Hirshfeld partition analysis (IGMH)51 (Fig. 4a and Supplementary Fig. 18). This analysis indicates the presence of various intramolecular weak interactions between the zinc-containing units and both D- and L-Met amino acid residues (blue layered region), which contribute to the enhanced structural stability of both (S)-Y3 + Zn2+ + L-Met and (S)-Y3 + Zn2+ + D-Met complexes. To investigate the electrostatic interactions between the (S)-Y3 + Zn2+ complex and L-Met or D-Met molecules, we further calculated the electrostatic potential of two adjacent (S)-Y3 + Zn2+ + L-Met complexes (Fig. 4b and Supplementary Fig. 19). The results show that, in the case of two adjacent (S)-Y3 + Zn2+ + L-Met complexes, the morpholinium quaternary ammonium cation exhibits a more pronounced positive charge compared to (S)-Y3 + Zn2+ + D-Met. Likewise, the carboxyl group in (S)-Y3 + Zn2+ + L-Met displays a stronger negative charge than its counterpart in (S)-Y3 + Zn2+ + D-Met. This greater spatial separation of positive and negative charges enhances the electrostatic attraction between adjacent (S)-Y3 + Zn2+ + L-Met complexes. Furthermore, the total energy of the dimeric (S)-Y3 + Zn2+ + L-Met system is 2.9 kcal/mol lower than that of the corresponding (S)-Y3 + Zn2+ + D-Met dimer. This lower energy state, combined with the presence of well-defined positively charged quaternary ammonium and negatively charged carboxyl groups, accounts for the increased propensity of (S)-Y3 + Zn2+ + L-Met to form aggregates. The explore on the chiral recognition mechanism reveals that beyond the covalent imine-forming interaction between the probe molecule and amino acids, intramolecular and intermolecular non-covalent interactions also yield an effective synergistic effect in chiral recognition52. Furthermore, computational results indicate that the total energy of (S)-Y3 + Zn2+ + L-Met is 17.2 kcal/mol lower than that of (S)-Y3 + Zn2+ + D-Met (Fig. 4c). Additionally, in the structure of (S)-Y3 + Zn2+ + L-Met, the dihedral angle of binaphthalene is 59.6°, significantly smaller than the 108.9° observed in (S)-Y3 + Zn2+ + D-Met. This suggests that L-Met introduces less steric hindrance than D-Met upon binding with (S)-Y3. A smaller dihedral angle is also more favorable for intermolecular aggregation.

Fig. 4: Fluorescence sensing and chiral recognition mechanisms.
Fig. 4: Fluorescence sensing and chiral recognition mechanisms.
Full size image

a The IGMH isosurface and molecular structures of (S)-Y3 + Zn2+ + L-Met and (S)-Y3 + Zn2+ + D-Met. b The calculated distribution of ESP in 2[(S)-Y3 + Zn2+ + L-Met] at M06-2×/6-31 G(d,p) level. c Proposed and modeled structures of (S)-Y3 + Zn2+ + L-Met and (S)-Y3 + Zn2+ + D-Met. The energy differences between complexes formed with the enantiomeric pairs are provided in kcal/mol (computational details are presented in the Supporting Information). d Geometry comparison diagram between the S0 and S1 states of (S)-Y3 + Zn2+ + L-Met and (S)-Y3 + Zn2+ + D-Met, along with corresponding RMSD values (structures in the S0 and S1 states are shown in red and blue, respectively).

The theoretical feasibility of this fluorescent enhancement mechanism was confirmed through frequency analysis and structural optimization. Specifically, the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) were calculated for (S)-Y3, (S)-Y3 + Zn²⁺, (S)-Y3 + Zn²⁺ + L-Met, and (S)-Y3 + Zn²⁺ + D-Met, respectively (Supplementary Fig. 20). Notably, in (S)-Y3 + Zn²⁺ + D-Met, the frontier molecular orbitals exhibit localized excited-state features, with both the HOMO and LUMO confined to the binaphthyl moiety. In contrast, in the case of (S)-Y3 + Zn²⁺ + L-Met, the LUMO is distributed across the whole binaphthyl framework and the imine bond, whereas the HOMO is predominantly confined to the benzene moiety. This distinct spatial distribution of the molecular orbitals facilitates a more effective intramolecular charge transfer within the excited state of (S)-Y3 + Zn²⁺ + L-Met, which may explain the enhanced fluorescence intensity that has been observed. To investigate the geometric changes between the S0 and S1 states (Fig. 4d), root mean square displacement (RMSD) values were computed for (S)-Y3 + Zn²⁺ + D-Met and (S)-Y3 + Zn²⁺ + L-Met. The RMSD value for (S)-Y3 + Zn²⁺ + L-Met (0.1664 Å) is significantly lower than that for (S)-Y3 + Zn²⁺ + D-Met (0.2020 Å). This suggests that the (S)-Y3 + Zn²⁺ + L-Met complex experiences little structural distortion and a lower likelihood of non-radiative energy loss, leading to improved fluorescence during chiral recognition.

Fluorescent response of (S)-Y3 toward 19 free amino acids

To demonstrate the universality of rapid and high-contrast fluorescence-based chiral sensing by the (S)-Y3 probe, we applied a chiral fluorescent sensing system in aqueous solution to a series of chiral amino acids, as shown in Fig. 5a. All 19 natural chiral amino acids were allowed to react with the (S)-Y3 probe in the presence of zinc ion, and the fluorescence enhancement ratios of each enantiomer were recorded (Fig. 5b). Among these, 17 pairs of amino acid enantiomers could be effectively distinguished in aqueous solution based on either fluorescence intensity or wavelength (Fig. 5c and Supplementary Fig. 21). Most amino acid enantiomers exhibited a significant enantioselective fluorescence enhancement ratio (ef > 3) within 200 s after interaction with the (S)-Y3 probe. The relative fluorescent intensities of (S)-Y3 + Zn²⁺ and (S)-Y3 + Zn²⁺ + L-/D-amino acids are presented in Fig. 5c and the corresponding results of its enantiomer (R)-Y3 were shown as Supplementary Fig. 22. Notably, from the perspective of fluorescence intensity, this probe demonstrated significantly enhanced chiral recognition capability for Met, asparagine (Asn), leucine (Leu), and phenylalanine (Phe) compared to the other 15 amino acids. Furthermore, based on differences in fluorescence intensity and color, the probe showed high chiral recognition ability for 17 amino acids. The probe exhibited excellent chiral discrimination on amino acids containing aliphatic, hydroxyl, acidic, and amide functional groups, both visually and spectrally. In addition to exhibiting significant enantioselective fluorescence enhancement for the enantiomers of Met, Asn, Leu, and Phe, (S)-Y3 also displayed observable aggregation differences between their D- and L-enantiomers, which can be easily detected by the naked eye. When the (S)-Y3 + Zn²⁺ mixture interacts with the D-enantiomers of these four amino acids, the solution remains clear and transparent, regardless of whether it is left to stand for 15 or 30 min (Supplementary Fig. 23). In contrast, when interacting with any of their L-enantiomers, particle suspension becomes visible within 15 min, and after 30 min, aggregates settle at the bottom of the container. These observations indicate that the products formed by the reaction of (S)-Y3 + Zn²⁺ with L-amino acids are more prone to aggregation than those formed with D-amino acids. Having previously examined the precipitation reaction kinetics of (S)-Y3 with Met on a second timescale (Fig. 2a), we extended our investigation to include three additional amino acids—Asn, Leu, and Phe—to assess their rapid interaction dynamics. (S)-Y3 began to precipitate within 15 s upon mixing with L-Phe, 80 s with L-Asn, and 140 s with L-Leu. (Supplementary Fig. 24). Therefore, the (S)-Y3 probe can not only serve as a fluorescent tool for rapid and effective chiral recognition of 17 amino acids but also enable enantioselective aggregation and precipitation of four specific amino acids.

Fig. 5: A rapid fluorescence method for the detection of chiral free amino acids.
Fig. 5: A rapid fluorescence method for the detection of chiral free amino acids.
Full size image

a The chemical structure of the fluorescent probe (S)-Y3 used for the detection of nineteen chiral amino acids in water. b Fluorescence enhancement ratio for chiral amino acids grouped into various structural categories. Gln and His are represented by distinct colors due to their unique fluorescence properties. c The chemical structures of the tested chiral amino acids, along with photographs illustrating the visual fluorescence intensity and color differences observed when using the (S)-Y3 probe under identical experimental conditions (pH 8.8).

Enantiomer separation and visualization

Given the excellent enantioselective recognition, aggregation capacity, and rapid solid precipitation properties of (S)/(R)-Y3 toward chiral amino acids, we explored its potential as a probe for visualizing chiral separation. The complete process of chiral separation coupled with fluorescence visualization and HPLC analysis is illustrated in Fig. 6a. The imine bonds formed between the probe and specific enantiomer of Met are susceptible to acid hydrolysis, facilitating the release of free amino acid enantiomer. Initially, L-Met enantiomers with enantiomeric excesses (ee, ([L]-[D])/([D] + [L])) of 50%, 0%, and −50% were introduced into a solution containing (S)/(R)-Y3 and zinc ions to evaluate the probe’s efficiency in separating chiral amino acid mixtures. As illustrated in Fig. 6b–d, the fluorescence intensity significantly decreased after the solid complexes formed by the probe and the corresponding amino acid enantiomers were removed via filtration. Specifically, the fluorescence intensity decreased by 12.4-fold in the 50% ee group, 13.6-fold in the 0% ee group, and 15.6-fold in the −50% ee group. Further visualization under natural light revealed the presence of yellow precipitates before filtration, which exhibited strong green fluorescence under 365 nm UV light. After filtration, the solution became colorless and transparent under natural light, and under UV irradiation, the fluorescence color shifted from bright green to dark orange (Supplementary Fig. 25). Control experiments indicated that the orange fluorescence originated from the interaction between opposite configuration amino acids and the probe. These findings demonstrate that the (S)/(R)-Y3 probe can effectively separate specific amino acid enantiomers through aggregation-induced precipitation, with the separation process being visually traceable via fluorescence changes.

Fig. 6: Chiral separation analysis and fluorescence visualization.
Fig. 6: Chiral separation analysis and fluorescence visualization.
Full size image

a Scheme of probe (S)/(R)-Y3 for chiral separation and fluorescence visualization of D/L-amino acids. Fluorescence spectra of amino acid mixtures with 50% ee Met (b), 0% ee Met (c), and −50% ee Met (d) before and after filtration, following the addition of (S)/(R)-Y3 probe and Zn²⁺ (probe concentration: 0.20 mM, Met: 25 equiv, Zn²⁺: 2 equiv, λexc = 420 nm, slits: 3/3 nm). Inset images: Visualization of amino acid mixtures with different ee values (50% ee Met, 0% ee Met, and −50% ee Met) after the addition of (S)/(R)-Y3 probe and Zn²⁺ under natural light and 365 nm UV light. HPLC chromatograms obtained after dissociation of precipitates formed by the interaction between the (S)/(R)-Y3 probe and chiral amino acid mixtures: e 50% ee Met, 0% ee Met, and −50% ee Met; f 70% ee Phe, 60% ee Phe, and −70% ee Phe; (g) 80% ee Leu and −80% ee Leu. HPLC analyzed on the chiral-phase column of (CHIRALPAK ZWIX(−), size: 4.0 mm × 150 mmL × 3 μm, mobile phase for Met and Phe: MeOH (with 50 mM formic acid and 25 mM diethylamine); mobile phase for Leu: MeOH (with 50 mM formic acid and 25 mM diethylamine)/CH3CN = 60/40).

Based on the chiral separation of amino acids and fluorescence visualization results, we further filtered and collected the precipitate formed by the aggregation of the (S)/(R)-Y3 probe after its reaction with specific Met, Phe, and Leu enantiomers. As shown in Fig. 6e–g, the peak times and positions of (S)/(R)-Y3, D/L-Met, D/L-Phe, and D/L-Leu were initially identified using chiral HPLC, confirming the effectiveness of chromatographic separation for chiral Met, Phe, Leu, and (S)/(R)-Y3. The aggregation precipitate was then dissociated into free amino acids using 0.1% formic acid in water, and the enantiomeric purity of the amino acids Met, Phe, and Leu in the precipitate was detected by chiral HPLC. The results of Chiral HPLC confirm probe (S)-Y3 achieves precise isolation of enantiomerically pure L-Met from aqueous solutions initially containing L-Met at 50% ee and 0% ee, respectively. Furthermore, (S)-Y3 also achieves effective separation of enantiomerically pure L-Phe and L-Leu from their enantiomer mixtures (70% and 60% ee for L-Phe; 80% ee for L-Leu), respectively. In contrast, probe (R)-Y3 facilitates the separation of enantiomerically pure D-configured amino acids from enantiomer mixtures of −50% ee for L-Met, −70% ee for L-Phe, and −80% ee for L-Leu. These results demonstrate that the application of the chiral probe (S)/(R)-Y3 enables the effective separation of chiral amino acid mixtures through a process involving aggregation-induced precipitation, followed by simple filtration and acidification-mediated dissociation. This approach allows for the successful isolation of amino acids with a single enantiomeric configuration.

Discussion

In conclusion, a binaphthyl-based fluorescent probe, (S)/(R)-Y3, containing a morpholinyl quaternary ammonium cation moiety was designed and synthesized. The incorporation of this cationic fragment serves dual purposes: enhancing water solubility and facilitating electrostatic interactions with the carboxyl groups of amino acid residues. Notably, (S)-Y3 rapidly recognizes L-Met within 100 s and forms bright aggregates upon interaction. Similarly, its enantiomer (R)-Y3 selectively interacts with D-Met within the same time frame, also forming luminescent aggregates. This phenomenon can be attributed to the formation of a complex between the probe and specific chiral amino acids that exhibits pronounced AIE characteristics, leading to fluorescence activation. This chiral recognition process involving (S)/(R)-Y3 is accompanied by a distinct fluorescence color transition from yellow to green, enabling dynamic visualization of chiral discrimination. Beyond Met, the probe (S)-Y3 achieves fluorescent chiral differentiation of 16 additional amino acids through variations in both fluorescence intensity and emission color. Moreover, the probe can rapidly aggregate and precipitate with specific enantiomers of Met, Asn, Leu, and Phe. Single amino acid enantiomers can thus be separated via a simple filtration method, with the process being visually traceable through fluorescence changes. The efficiency of chiral separation based on these fluorescent probes was further confirmed by HPLC analysis. This study not only offers a fluorescent tool for efficient chiral recognition and separation of amino acids but also provides valuable insights into chiral chemistry research.

Of course, our research work still has the following aspects that need to be broken through for the quantitative application of amino acid chiral separation: 1. The current system’s reliance on Zn²⁺ indeed restricts its direct applicability in certain biological or environmental contexts. Future efforts will focus on developing metal-free detection systems, such as tuning the electronic structure of probes to stabilize imine intermediates without metal coordination. 2. While the pH condition (~8.8) is essential for achieving high selectivity in this study, it inherently limits the method’s broader applicability. To address this, future probe designs will explore incorporating functional groups with enhanced buffering capacity or self-regulating pH properties to expand the operational pH range. 3. Macroscopic separation has so far been demonstrated only for a subset of amino acids, indicating a need for improved universality. This limitation reflects the inherent specificity of the probe’s recognition mechanism. Future strategies include constructing probe libraries with varied chiral architectures to target diverse amino acid classes, or developing solid-phase extraction materials by immobilizing probes onto solid supports for reusable and scalable separation processes.

Methods

Materials

All chemicals and solvents were of reagent grade and used as received, unless otherwise noted. Various amino acids were obtained from commercial suppliers and were used without further purification. The molecular structures were confirmed using 1H NMR, 13C NMR, and high-resolution ESI mass spectroscopy.

General methods

NMR data were collected on Bruker 400 MHz NMR Spectrometer in DMSO-d6, CDCl3, or D2O. Chemical shifts (δ) were reported in the unit of ppm using tetramethylsilane as the internal standard. High-resolution mass spectra were acquired on an Agilent Q-TOF6300 instrument, and the percentage deviation between the calculated and measured values is less than 5‰. The steady-state fluorescence emission spectra were recorded with a Horiba FluoroMax-4 spectrofluorometer. Fluorescence quantum yields were measured by using an integrating sphere on a HAMAMATSU Quantaurus-QY C11347-11. CD spectra were acquired using the JASCO J815 spectrophotometer. The enantiomer separation result was determined by HPLC analysis performed on Agilent 1260 Infinity II Series chromatographs with a UV–Vis spectrometer or an Evaporative Light Scattering Detector model 3000KS and a CHIRALPAK ZWIX(−) chiral column with a length of 15 cm. Zn(OAc)2 was used as the Zn2+ source. Deionized water was used for the fluorescent analysis.

Geometric optimization

Cartesian coordinates of optimized L/D-Met, (R)/(S)-Y3, (S)-Y3 + Zn2+ + L/D-Met, 2[(S)-Y3 + Zn2+ + L/D-Met] in the S0 or S1 state calculated by the DFT, M06-2X-def2svp, Gaussian 16 program.