Chirality‐Matching in Au–S Oriented Assembly Induces Highly Selective Recognition of Glucose Enantiomers

ABSTRACT

Enantioselective distinction of glucose (Glu) is a core challenge in biochemistry, as its subtle difference in chiral configuration leads to distinct biological activities. To address this, we propose an “Au─S oriented assembly‐chiral matching amplification” strategy, D/L‐cysteine (Cys) serves as a chiral recognition moieties and is immobilized on gold nanoparticles (Au NPs) via Au─S bonds to construct a chiral interface with well‐defined enantioselectivity. Mechanistic studies reveal that the chiral microenvironment of Cys optimizes the orientation and adsorption of Glu on the electrode surface through stereomatching effects, which in turn regulates enantioselective differences in electron transfer efficiency, endowing the interface with excellent enantioselectivity. Specifically, Au‐D‐Cys exhibits a significantly stronger electrochemical response to D‐Glu than to L‐Glu, while Au‐L‐Cys shows the opposite trend. The constructed electrochemical sensor has a linear range of 1–558 mm for both D‐Glu and L‐Glu, with corresponding sensitivities of 19.95 and 12.87 µA mm ⁻¹ cm⁻², enabling accurate quantification and distinction. This work demonstrates that chiral configuration matching effectively enhances sensor performance, providing a new strategy for the recognition of multi‐chiral‐center enantiomers.

A chiral interface construction strategy based on cysteine‐functionalized gold nanoparticles (AuNPs) was proposed, where AuNPs were employed to enhance the ligand‐substrate charge transfer interaction. Combined with the spatial matching differences between the chiral center of cysteine and the hydroxyl groups of D/L‐glucose (Glu), the amplification of electrochemical signals from molecular configurations was achieved, thereby enabling the detection of D/L‐Glu.

1. Introduction

Chirality, a fundamental property of nature, constitutes a core factor in regulating the functions of biomolecules [1, 2, 3]. Enantiomers of chirality often exhibit significant activity differences in the pharmaceutical and biological metabolism fields [4, 5]. Taking glucose (Glu) as an example, D‐Glu is a key substrate for energy metabolism in living organisms, while L‐Glu has almost no metabolic activity [6, 7, 8, 9, 10]. Therefore, the accurate recognition of D‐/L‐Glu is of great scientific significance for elucidating biological metabolic mechanisms and developing targeted drugs. However, the recognition of D‐/L‐Glu is confronted with inherent bottlenecks. Both enantiomers contain four consecutive chiral centers (C₂–C₅), with mirror symmetry only exhibited by the hydroxyl groups attached to the chiral carbons [11]. They are identical in molecular formula, chemical bond types, and basic physicochemical parameters. More importantly, in a multi‐chiral center system, differences in individual chiral configurations are prone to being averaged out by the overall structure, causing the microscopic chiral differences between enantiomers to be unable to be converted into macroscopically measurable signals, thus creating the “signal transduction bottleneck” [12, 13]. Even though traditional techniques such as high‐performance liquid chromatography and mass spectrometry can achieve a certain degree of enantiomer separation, these methods can hardly effectively capture averaged chiral signals and thus cannot meet the requirements of high‐precision chiral recognition [14, 15, 16].

Electrochemical sensing technology, by virtue of a unique direct transduction mechanism of molecular interaction‐electrical signal output, has shown great application value across domains including biomedical diagnostics, environmental surveillance, and food safety assessment [17, 18, 19, 20, 21]. This technology boasts advantages including rapid response, simple operation, low detection cost, and wide dynamic range [22, 23, 24]. However, traditional electrochemical methods still have fundamental limitations in terms of recognizing glutamic acid enantiomers with multiple consecutive chiral centers [25]. Conventional electrode interfaces lack precise chiral recognition units and cannot convert subtle configurational differences of enantiomers with multiple chiral centers into clearly distinguishable electrochemical response signals [26, 27, 28].

Herein, an “Au─S oriented assembly‐chiral matching amplification” strategy is proposed for efficient, precise, and sensitive distinction of D‐/L‐Glu enantiomers. D‐/L‐cysteine (Cys), as chiral anchoring ligands, self‐assembles on gold nanoparticles (Au NPs) via Au─S bonds to form a dense, ordered chiral molecular recognition interface. Via chiral configuration matching between D‐/L‐Cys and Glu, this interface optimizes Glu adsorption at the catalytic active sites of Au NPs and modulates interfacial electron transfer efficiency, thereby transducing Glu configurational information into quantifiable electrochemical signals. Electrochemical results demonstrate that the system exhibits remarkable chiral recognition specificity toward D‐/L‐Glu, shows a favorable linear response over a wide concentration range (R ² = 1), and achieves higher detection sensitivity in homochiral combinations than in heterochiral counterparts. This work addresses the multi‐chiral‐center enantiomer recognition challenge, providing a new conceptual paradigm and experimental platform for chiral analysis.

2. Results and Discussion

2.1. Synthesis and Characterization of Chiral Au NPs

To obtain chiral Au NPs, we employed sodium borohydride (NaBH₄) in a rapid chemical‐reduction method to reduce hydrogen tetrachloroaurate (HAuCl₄), and then introduced D‐Cys or L‐Cys as chiral ligands, which enabled the successful synthesis of chiral Au NPs (denoted as Au‐D‐/L‐Cys) (Figure 1a) [29]. The morphological features of the as‐prepared chiral Au‐D‐/L‐Cys were investigated using transmission electron microscopy (TEM). As observed from Figure 1b,c, the Au‐D‐/L‐Cys are well‐dispersed, with a homogeneous size distribution and a distinct spherical morphology. Remarkably, the particle size and morphology of Au‐D‐/L‐Cys were nearly identical. By means of high‐angle annular dark‐field scanning TEM (HAADF‐STEM) and elemental‐mapping images, a homogeneous spatial distribution of Au and S elements was observed throughout the Au‐D‐/L‐Cys (Figure 1d). The presence of the S element provided direct evidence of the successful modification of Au NPs by Cys. The dynamic light scattering (DLS) results indicated that the average hydration diameters of AuNPs, Au‐D‐/L‐Cys were all approximately 18 nm, which is consistent with the particle size distribution from TEM analysis (Figure 1e and Figure S1). Chiral modification scarcely induces particle aggregation, thus maintaining favorable dispersion and colloidal stability of Au NPs. As seen in Figure 1f, the ultraviolet visible (UV–vis) absorption profile of Au‐D‐/L‐Cys had a surface plasmon resonance (SPR) peak at approximately 530 nm. Compared with bare Au NPs, the SPR peaks of Au‐D‐/L‐Cys show no significant shift or peak broadening. This result indicates that the loading of Cys does not significantly affect the static optical properties of the Au. X‐ray diffraction (XRD) results confirm that the main diffraction peaks of chiral Au‐D‐/L‐Cys match well with Au (PDF#04–0784, Figure 1g), and extra impurity signals stem from synthetic inorganic byproducts.

FIGURE 1.

(a) Schematic illustration of the synthesis for bare Au NPs, chiral Au‐D‐/L‐Cys. Corresponding TEM (top) and HRTEM (bottom) images of (b) Au‐D‐Cys and (c) Au‐L‐Cys. (d) HAADF‐STEM image and elemental mapping profiles for Au and S elements of Au‐D‐/L‐Cys. (e) DLS size distribution profiles, (f) UV–vis absorption spectra, and (g) XRD patterns of AuNPs, chiral Au‐D‐/L‐Cys, and D‐/L‐Cys.

2.2. Chiral Properties of Au‐D‐/L‐Cys

Constructing chiral recognition interfaces with well‐defined stereochemical environments is crucial for enantioselective recognition. As illustrated in Figure 2a, D‐/L‐Cys was selected as the chiral ligand. The thiol group (─SH) at the molecular terminus of Cys forms a stable Au─S covalent bond with Au NPs, thereby enabling the site‐specific ordered and oriented anchoring of Cys on the surface of Au NPs. Ultimately, a chiral recognition interface with a well‐defined stereochemical orientation, regulated by the chiral configuration of the ligand, was constructed. To confirm the conjugation site between Au and D‐/L‐Cys, the Fourier transform infrared (FTIR) spectra of Au‐D‐/L‐Cys were compared with those of free D‐Cys and L‐Cys. No absorption band associated with S─H stretching vibrations at 2520 cm⁻¹ was detected in the FTIR spectra of Au‐D‐/L‐Cys, which confirms that Cys is predominantly anchored to the surface of Au NPs, leading to the generation of an Au─S bond (Figure 2b) [30]. The X‐ray photoelectron spectroscopy (XPS) spectrum presented in Figure 2c exhibited two well‐defined peaks at 87.9 and 84.2 eV, which are characteristic of Au 4f_5/2 and Au 4f_7/2 signals, respectively. The Au 4f peaks of Au‐D‐/L‐Cys shifted to lower binding energies compared with those of Au NPs, which can be attributed to the electronic interaction between Au NPs and D‐/L‐Cys. In the S 2p spectrum, the characteristic peak at 163.0 eV was attributed to the Au─S bond, further confirming that Cys was successfully immobilized on Au NPs through covalent interaction (Figure 2d) [31]. Zeta potential measurements give average values of −26.4, −28.8, and −27.7 mV for bare Au NPs, Au‐D‐Cys, and Au‐L‐Cys (Figure 2e). The Cys modified Au NPs exhibited significantly larger absolute zeta potential values than that of bare Au NPs. Cys interacts with the surface of Au NPs through its functional groups, effectively regulating the surface charge distribution. To further investigate the optical properties of the Au‐D‐/L‐Cys, we used UV/Vis and circular dichroism (CD) spectroscopy. Au‐D‐/L‐Cys exhibited two absorption peaks at around 230 and 530 nm in the UV/Vis spectra (Figure S2). Among them, the absorption peak at 230 nm was assigned to D‐/L‐Cys ligands, while the peak at 530 nm corresponded to the typical SPR absorption peak of Au NPs. As shown in Figure 2f, the CD spectra acquired for Au‐D‐/L‐Cys presented an almost perfect mirror symmetry, with one prominent CD peak at 206 nm that is attributable to the surface‐anchored Cys molecules. Subsequently, the anisotropy factor (g‐factor), a dimensionless parameter for quantifying the magnitude of chiral optical activity, was calculated [32]. The calculated g‐factor of Au‐D‐/L‐Cys reached a value as high as 0.003 and 0.002 (Figure 2g). This was mainly ascribed to the anchoring of Cys enantiomers onto the surface of Au NPs, thus verifying that the as‐prepared Au‐D‐/L‐Cys exhibited chiroptical activity. Without the introduction of Cys enantiomers, bare Au NPs show no detectable CD signals or absorbance peaks in the vicinity of 230 nm, but their sizes characterized by TEM are analogous to those of chiral Au NP counterparts (Figure S3). Collectively, Cys is covalently anchored to the Au surface via an Au–S bond, endowing the hybrids with chiroptical activity.

FIGURE 2.

(a) Schematic illustration of Au NPs self‐assembling with D‐Cys and L‐Cys via intermolecular interactions. (b) FTIR spectra of D‐Cys, L‐Cys, Au‐D‐Cys, and Au‐L‐Cys. High‐resolution XPS spectra for (c) Au 2p, and (d) S 2p. (e) Zeta potential comparison, (f) CD and (g) g‐factor spectra of AuNPs, Au‐D‐Cys, and Au‐L‐Cys.

2.3. Mechanism for Discriminating D‐/L‐Glu Enantiomers

To elucidate the mechanism underlying the chiral recognition of Au‐D‐/L‐Cys we selected the enantioselective interaction between Au‐D‐/L‐Cys and D‐/L‐Glu as a model system and systematically evaluated the distinction performance toward D‐/L‐Glu. CD spectra showed that the CD signals of the Au‐D‐/L‐Cys were markedly enhanced upon addition of Glu enantiomers. Furthermore, homochiral combinations (Au‐D‐Cys + D‐Glu and Au‐L‐Cys + L‐Glu) displayed slightly stronger CD peak intensities than the corresponding heterochiral combinations (Figure 3a). These spectral observations preliminarily confirm that the chiral recognition‐driven intermolecular interactions exhibit enantiomeric selectivity differences, and interactions between homochiral configurations were much more significant. Specifically, Cys molecules on Au‐D‐/L‐Cys can furnish key polar recognition moieties, including amino groups(−NH₂) and carboxyl groups (−COOH). As enantiomers, D‐/L‐Glu exhibit configuration differences at their chiral centers, resulting in distinct spatial orientations of −OH, thereby regulating the binding efficiency and interaction modes between these two components and the chiral interface. To further unravel the mechanism, zeta potential, and DLS measurements were implemented to verify the advantages of homochiral combinations from the dual dimensions of charge distribution and colloidal interface structure. Zeta potential assays revealed that the absolute zeta potential values of homochiral combinations were markedly higher than those of heterochiral combinations. This is attributed to the highly matched charge distribution orientation, which enhances intermolecular electrostatic interactions and facilitates the formation of a denser, more stable hydrogen bond network, thus better stabilizing the surface charge (Figure 3b and Table S1) [33]. DLS measurements demonstrated that the hydrodynamic diameter of homochiral combinations was significantly larger than that of heterochiral combinations (Figure 3c and Table S2). The essence of this phenomenon lies in the stereomatching effect between homochiral molecules. Specifically, the stereocomplementarity of homochiral molecules enables the chiral moieties of Au‐D‐/L‐Cys to form an oriented and specific hydrogen‐bonded network with Glu molecules. The hydrogen‐bonded network greatly weakens intermolecular steric repulsion. Meanwhile, the synergistic effect of hydrogen bonding and hydrophobic interactions promotes the ordered assembly of Glu on Au NPs surfaces. This effect enables high‐density Glu adsorption and further yields a thicker, more stable surface coating. Notably, the formation of this ordered surface layer is not a simple molecular accumulation but a direct manifestation of the optimization effect of the chiral microenvironment. It enhances the adsorption specificity of Glu in the vicinity of the catalytic sites of Au NPs and, further, modulates the electron cloud distribution on Au NP surfaces, thereby laying a structural foundation for the precise regulation of interfacial electron transfer efficiency. To unravel the differences in hydrogen bonding interaction between Au‐D‐/L‐Cys and D‐/L‐Glu, we performed a targeted analysis of the 3300–3500 cm⁻¹ region in FTIR spectra, which corresponds to N−H/O−H stretching vibrations. The spectral features of this region directly reflect the intensity of intermolecular hydrogen bonding [34, 35]. Experimental results showed that upon the addition of Glu, the absorption peaks in this region all exhibited significant redshift, broadening, and intensity enhancement, which confirmed the formation of intermolecular hydrogen bonds between Glu and Cys. A further comparison of different chiral combination systems revealed that due to the differences in stereoisomerism and chiral orientation of the hydrogen bonds formed between D‐/L‐Glu and Cys, the absorption peaks of homochiral combinations (such as Au‐D‐Cy + D‐Glu and Au‐L‐Cy + L‐Glu) exhibited more pronounced broadening and intensity increase. This phenomenon fully demonstrates that hydrogen bonding interactions possess distinct chiral dependence, meaning that homochiral combinations are capable of forming stronger hydrogen bond networks. In addition, the reduced or even quenched intensity of N─H bending vibrations in the 1400–1550 cm^{− 1} region further provided indirect evidence that strong hydrogen bonds inhibit N─H bond vibrations, offering multidimensional support for assessing hydrogen bond strength (Figure 3d,e) [36]. Electrochemical impedance spectroscopy (EIS) reveals the interactions of chiral recognition from the perspective of interfacial electron transfer kinetics. In Nyquist plots, the semicircle diameter in the high‐frequency region is correlated with charge transfer resistance (Rct). A smaller diameter means lower interfacial electron transfer resistance and a faster electron transfer rate [37]. As shown in Figure 3f, the Rct values of homochiral combinations were significantly lower than those of heterochiral combinations, proving that the electron‐transfer process is more facile. This difference is mainly attributed to the stereochemical matching effect between chiral molecules. Variations in the spatial conformation of different chiral combinations directly govern the quantity and binding strength of intermolecular hydrogen bonds [38]. D‐Cys and D‐Glu (as well as L‐Cys and L‐Glu) exhibit favorable intermolecular conformational matching, thereby efficiently facilitating interfacial electron transfer. In summary, the molecular basis of these phenomena is as follows, chiral guest molecules (D‐/L‐Glu) and chiral selectors (Au‐D‐/L‐Cys) form transient diastereomeric complexes via three‐point attraction or repulsion interactions between enantiomers. In homochiral combinations, superior stereochemical matching and reduced steric hindrance enable the−NH₂ and −COOH of surface‐bound cysteine to form a denser, more stable hydrogen‐bond network with the −OH of Glu. For heterochiral combinations, mismatched molecular configurations increase steric repulsion, thereby weakening both the number and stability of intermolecular hydrogen bonds. These intrinsic differences in binding modes and interaction energies collectively achieve highly efficient and selective chiral distinction of D‐/L‐Glu (Figure 3g).

FIGURE 3.

(a) Comparative CD spectra of the four chiral combinations. (b) Zeta potential comparison of Au, Au‐D‐Cys, and Au‐L‐Cys upon interaction with D‐Glu and L‐Glu. (c) Intensity distribution of nanoparticle diameters for different chiral combinations. FT‐IR transmittance spectra characterizing molecular interactions on (d) Au‐D‐Cys and (e) Au‐L‐Cys chiral surfaces after incubation with D‐/L‐Glu enantiomers. (f) EIS plots of different chiral combinations, with the inset depicting the equivalent circuit model. (g) Schematic of chirality recognition (left) and hydrogen bonding in homochiral and heterochiral interactions (right).

2.4. Enantioselective Recognition of D‐/L‐Glu by Au‐D‐/L‐Cys

To explore the enantioselectivity recognition properties of the Au‐D‐/L‐Cys, the enantiomers D‐Glu and L‐Glu were used as target molecules (Figure 4a). Cyclic voltammetry (CV) curves acquired in distinct D‐/L‐Glu solutions clearly reflect current variations induced by chiral recognition, demonstrating favorable electrochemical responses of Au‐D‐/L‐Cys to D‐/L‐Glu (Figure 4b). Specifically, Au‐D‐Cys exhibits specific recognition for D‐Glu and Au‐L‐Cys exhibits specific recognition for L‐Glu, likely attributed to promoted electron transfer via homochiral intermolecular interactions. The CV curves of D‐Glu and L‐Glu almost coincide at bare Au electrodes (Figure S4). The principal component analysis (PCA) distinction plot enables effective separation of D‐Glu and L‐Glu sample solutions (Figure 4c). Based on the differences in electrochemical signals, the kinetic processes of Au‐D‐/L‐Cys in chiral electrochemical recognition were studied by conducting CV cycling tests at various scan rates (10−100 mV·s⁻¹) in add D‐/L‐Glu solutions (Figure 4d and Figure S5). The square root of the peak current shows a good linear relationship with the scan rate, indicating that the oxidation reaction of D‐/L‐Glu at the Au‐D‐/L‐Cys chiral interface is dominated mainly by diffusion dynamics (Figure S6). The sensitivity of the prepared Au‐D‐/L‐Cys for D‐/L‐Glu recognition and detection was evaluated using current‐time curve (i–t). The I–t curves showed that as D‐/L‐Glu concentration increased stepwise from 1 to 558 µm, current of the sensor rose in a stepped manner. The demonstration of the sensitive response of the sensor to D‐/L‐Glu concentration changes and ability to satisfy quantitative detection requirements. Meanwhile, the response differences of Au‐D‐/L‐Cys electrodes to D‐/L‐Glu intuitively reflect their chiral selective recognition ability (Figure 4e). Further analysis of the current density‐concentration linear curve revealed that the sensor exhibits a good linear response to D‐/L‐Glu over a wide concentration range (with all R² values close to 1), indicating the stability and reliability in quantitative detection (Figure 4f). When the slope of the linear calibration curve was used to evaluate sensitivity, the Au‐D‐Cys electrode exhibited a high sensitivity of 19.95 µA mm ⁻¹ cm⁻² toward D‐Glu, while a much lower value of 5.14 µA mm ⁻¹ cm⁻² was obtained for L‐Glu. In comparison, the Au‐L‐Cys electrode delivered a sensitivity of 12.87 µA mm ⁻¹ cm⁻² for L‐Glu and 6.78 µA mm ⁻¹ cm⁻² for D‐Glu (Table S3). The result confirms the core mechanism that “homochiral recognition is superior to heterochiral recognition”, meaning that when the chiral configurations match, interaction of the electrode with homochiral Glu is stronger, thus enabling the electrode to exhibit higher sensitivity. To further clarify the origin of chiral selectivity, control experiments of pristine Au NPs were conducted under identical conditions (Figure S7). The results show that the i–t signals and current density‐concentration linear curves of Au NPs toward D‐/L‐Glu nearly completely overlap, with sensitivity slopes of 3.43 and 3.37 µA mm ⁻¹ cm⁻², respectively, and all linear correlation coefficients higher than 0.99 (Table S4). This confirms that pure Au NPs lack intrinsic chiral recognition capacity, and the distinct chiral response of Au‐D‐/L‐Cys electrodes originates from the surface‐bound chiral ligands (D‐/L‐Cys). To further explore the practical feasibility of this sensing system, mixed D‐/L‐Glu solutions with various molar ratios were tested. Clear and well‐separated clusters in the PCA plots confirm that the developed chiral sensor can effectively distinguish enantiomer mixtures with different proportions (Figure S8) Beyond chiral detection, the anti‐interference capability and sensing selectivity of the Au‐D‐/L‐Cys electrode were further evaluated. When common biological interferents including sodium chloride (NaCl), uric acid (UA), dopamine (DA), and ascorbic acid (AA), along with structurally analogous sugars including D‐galactose (D‐Gal), sucrose (SUC), D‐arabinose (D‐Ara), and fructose (Fru), were sequentially introduced, no obvious current fluctuations were observed. Only upon the addition of the target analyte D‐/L‐Glu did a significant current response occur (Figure 4g and Figure S9). These results confirm the high specificity of the sensor toward D‐/L‐Glu and the strong anti‐interference capability against various interfering substances in complex matrices. To further validate the practical applicability of the sensor in complex matrices, standard addition recovery experiments were performed with commercial 50% Glu oral solution serving as the real sample matrix. The recovery rates range from 98% to 110%, falling within the acceptable error range for analytical methods (Figure S10 and Table S5). The indicates that the sensor has excellent selectivity for D‐/L‐Glu, can effectively resist interference from such substances in complex matrices, and lays a solid foundation for the accurate detection of Glu in actual biological samples.

FIGURE 4.

(a) Schematic illustration of the current response principle for chiral recognition of D‐/L‐Glu based on Au‐D‐/L‐Cys. (b) CV curves of Au‐D‐Cys (upper) and Au‐L‐Cys (lower) interacting with D‐Glu and L‐Glu. (c) Principal component analysis (PCA) plot. (d) CV curves of Au‐D‐Cys with D‐Glu at various scan rates (10 to 100 mV/s). (e) I–t curves of Au‐D‐Cys and Au‐L‐Cys interacting with D‐Glu and L‐Glu at different concentrations (1 to 558 µm). (f) Current density‐concentration linear curves. (g) Selectivity evaluation of the chiral sensor toward various potential interfering substances.

3. Conclusions

To summarize, we constructed a chiral recognition interface via an “Au─S oriented assembly‐chiral matching amplification” strategy for highly efficient, selective recognition of D‐/L‐Glu enantiomers. D‐/L‐Cys acted as chiral anchoring ligands that underwent self‐assembly on the AuNP surface via Au─S bond, forming a dense, ordered chiral recognition interface. The chiral centers of Cys enable stereoselective interactions with D‐/L‐Glu via spatial configuration matching and stereospecific hydrogen bonding. This optimizes Glu adsorption on Au NPs catalytic sites, regulates interfacial electron transfer efficiency, and achieves efficient transduction of chiral configurational information into quantifiable electrochemical signals. Electrochemical characterizations validated that the as‐constructed interface was successfully integrated into an electrochemical sensor, enabling accurate quantification and robust enantioselective distinction of D‐/L‐Glu. This work establishes a distinctive strategy to achieve highly selective identification of multi‐chiral‐center enantiomers, which advances the mechanistic insights into nanoscale chiral recognition processes.

Author Contributions

The manuscript was written through the contributions of all authors. All authors have approved the final version of the manuscript.

Funding

The authors gratefully acknowledge the financial support of the National Natural Science Foundation of China (22174110, 22127803, 52470193, 22304096), the project ZR2023QB006 supported by Shandong Provincial Natural Science Foundation, supported by the China Postdoctoral Science Foundation under Grant Number 2023M741853, the Qingdao Postdoctoral Science Foundation (QDBSH20230202031), the Plan for Youth Innovation Team of Colleges in Shandong Province, Qingdao Science and Technology Benefit Demonstration Project: 26‐1‐5‐xdny‐25‐nsh.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: smll73913‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that supports the findings of this study are available in the supplementary material of this article.