Tuning enantioselective drug adsorption in isoreticular homochiral metal–peptide frameworks through proximity pore interactions

Abstract

Pharmaceutical research emphasizes stereochemistry and the enantioseparation of racemic drugs because different enantiomers can exhibit varying pharmacological and toxicological properties when interacting with the body's metabolic pathways. Metal–organic frameworks (MOFs) are porous adsorbents that can be designed to possess homochirality within their structures, enabling tunable porosity and enantioselective adsorption of racemic drugs. In our study, we present the synthesis of five novel and homochiral Co-L-GG(R) MOFs (where L-GG = glycyl-L(S)-glutamic acid and R = bipyridyl (bipy) pillar ligands). These isoreticular MOFs were synthesized using the ligand extension strategy. This approach allowed us to systematically control the pore sizes of the MOFs, enabling fine-tuning of the enantioselective adsorption of racemic drugs, primarily DL-penicillamine (Pen). Our findings reveal that the pore size greatly influences enantioselective adsorption, where too large or too small pores hinder the proximity-driven dispersive interactions between the drug and the pore surface, resulting in poor enantioselective adsorption of Pen. We achieved an enantiomeric excess (ee) of 60.1% (L over D), increasing to a maximum 76.1% ee using Co-L-GGvinylbipy by tuning proximity interactions in saturated pores. These results were accomplished by controlling the drug saturation within the MOF pores, promoting favorable interactions.

---

Introduction

The constant need for improved pharmaceuticals has highlighted the crucial role of stereochemistry in pharmaceutical research. More than half of the drugs currently in use are chiral, and nearly ninety percent of them are marketed as racemates.¹ This can be problematic, as these enantiomers may exhibit variations in pharmacological activity, pharmacokinetics, and toxicity profiles.¹ Achieving the delivery of an enantiopure drug to the market is a vital but time-consuming and costly endeavor. Enantiomer separation is commonly achieved through costly methods such as complicated stereoselective synthesis or chromatographic techniques that require expensive chiral stationary phases.^1,2 Research on alternative methods to separate enantiomers is currently underway.^3,4

Porous materials have been widely employed for various separation methods, including gas capture and storage,^5–7 and more recently, the separation of chiral molecules as chiral stationary phases.^8,9 Enantiomer recognition for racemic drugs is commonly simplified under the three-point interaction theory (Fig. 1), which describes the preferential attraction of a chiral molecule to a specific chiral substrate.^10,11 This theory suggests that a minimum of three configuration-dependent interactions take place on a planar surface between a molecule's chiral receptor and a chiral substrate. Depending on the stereoisomerism of the guest molecule to this receptor, the attraction is energetically favored, promoting longer retention onto the chiral substrate.^10,12 While this theory has seen many modifications over the last years,² the three-point interaction model remains the most prominent theory for explaining the underlying principles of chiral recognition. This model is based on the premise that favorable thermodynamic interactions play a crucial role in the retention mechanism of chiral adsorbents.² Through a comparable mechanism, chiral pores accommodate one enantiomer over the other by offering a better fit between the guest molecule and the pore cavity. This specific orientation facilitates more favorable interactions (van der Waals, hydrogen-bonding) with the pore surface, creating an energetically favorable environment for one enantiomer over the other.^12,13

Fig. 1 Model of the three-point interaction theory for enantioselective adsorption. Interaction between L-enantiomer of a drug is energetically favored based on a complete three-point contact to the chiral substrate (left) compared to one partial interaction with the D-enantiomer to the substrate (right). R denotes an organic substituent.

The selection of a suitable porous material with functionality and tunability for enantiomer separation remains a challenge.¹⁴ Zeolites serve as exemplary porous materials with inherent molecular sieves, high porosity, surface area, and uniform structure.^9,15 While they are often utilized for their impressive hydrogen and cation uptake capabilities, incorporating optical activity and chirality into zeolites presents a significant challenge.^9,16 Metal–organic frameworks (MOFs) exhibit similar potential with their structural tunability, large internal porosity, and ability to engineer chiral recognition sites within the pores of MOFs with chiral substituents of different size, shape, and function.¹⁷ This design approach enables the diffusion of one specific stereoisomer through the chiral pore environment, while the other optically active isomer selectively binds to the chiral recognition sites within the framework.^13,18

Typical synthesis of homochiral MOFs results from post-synthetic modification, spontaneous enantiomeric resolution by chiral induction agents which are prone to enantiomorphs, and through direct synthesis by incorporating enantiopure linkers.^13,19 For example, peptides offer structural versatility, multiple coordination sites with metals, and inherent chirality due to their amino acid constituents.²⁰ With a vast range of peptides available, it is possible to generate homochiral MOFs with pores that can promote interactions with a specific enantiomer.^21–23 However, there are drawbacks associated with the synthesis of porous and robust peptide MOF structures. The flexibility and self-folding nature of peptide sequences, owing to their aliphatic nature, restrict access to their pore apertures, making them unsuitable for enantioselective chiral drug separation.^24–26

Enantioselective adsorption relies on the affinity between the pores of the homochiral MOF and chiral drugs, however, the impact of the size or shape of the molecule and pore aperture on the enantioselective adsorption of chiral drugs is not yet fully understood. Additionally, there is a lack of knowledge regarding the interactions between drugs and chiral pore surfaces that enable enantioselective adsorption. To explore the influence of confinement effects within pore-apertures for enantioselective adsorption, reticular synthesis can be employed to systematically control the pore cavities of a homochiral framework while preserving the same chiral environment and structure topology.^27–29 To our knowledge, studies exploring the impact of pore size and functionality in isoreticular homochiral MOFs on the enantioselective adsorption of polar chiral drugs are scarce.

Our study focuses on investigating how the pore size of homochiral isoreticular MOFs influences the chiral affinity of DL-penicillamine (Pen). Penicillamine, trademarked as cuprimine, is a chelating agent marketed in its D stereoisomer and is widely used to mitigate heavy metal toxicity in the human body. On the other hand, its non-superimposable analogue (L-Pen) is toxic and can inhibit the production of vitamin B₆.³⁰ We synthesized isoreticular MOFs comprised of Co(II), L-GG: glycyl-L(S)-glutamate, and either bipy: 4,4′-bipyridyl or bipy-type pillar linkers to test the enantioselective adsorption of racemic Pen. These MOFs are made of one-dimensional dipeptide ladders connected by bipy-type linkers of varying size and functionality, enabling the generation of homochiral MOFs with adjustable porosity (Fig. 2a).²⁹ We thereby employ reticular synthesis to investigate the relationship between structure and properties of various pore environments and their impact on enantioselective adsorption.

Fig. 2 (a) Scheme showing homochiral synthesis of Co-L-GG chiral building units spaced by bipy-type linkers to form porous peptide ladders framework. (b) (Left) A scheme showing one possible configuration where L-Pen selectively diffuses through an L-configured chiral pore. (Right) A scheme showing one possible configuration where partial reception occurs based on the D-enantiomer (c). Experimental schematic for enantioseparation using a packed Co-L-GG(R) MOF bed. The D-enantiomer is less strongly adsorbed within the mobile phase solvent (racemic Pen in methanol) whereas the L-enantiomer retention in the chiral substrate is energetically favored.

Experimental section

Synthesis

The synthesis of Co-L-GG(R) was carried out through a modified procedure based on the protocol of Co-L-GGbipy synthesis established by Stylianou et al.²⁹ as detailed in the Supplementary Information (SI). In short, a solution of cobalt(II) acetate tetrahydrate (Co(OAc)₂·4H₂O) precursor was added to a mixture of L-GG and deionized water (DI H₂O) and heated to 85 °C for 8 hours, followed by cooling to room temperature. Five novel isoreticular Co-L-GG(R) MOFs were synthesized by incorporating different bipy-type linkers.

Characterization

To determine the structure and confirm the isoreticular nature of the Co-L-GG(R) MOFs, various diffraction and simulation techniques were employed. Powder X-ray diffraction (PXRD) was collected on each MOF to determine bulk crystallinity and phase purity through Le Bail refinements. Single crystal X-ray diffraction (SC-XRD) coupled with least-squares refinement (using SHELXL) were employed to elucidate the crystal structure of Co-L-GGvinylbipy. Structures that could not be solved from SC-XRD were computationally simulated and their theoretical diffraction patterns were compared to the experimental PXRD.

Other techniques were used to compare the Co-L-GG(R) structures and properties. Fourier transform infrared spectroscopy was performed on each MOF to show incorporation of L-GG into the Co(II) centers. Thermogravimetric analysis was employed to compare the MOF's thermal stabilities and BET adsorption experiments were also performed to compare their adsorption capacities.

Enantioseparation experiments

Enantioseparation experiments were conducted using DL-penicillamine (Pen) as the target drug for separation (Fig. 2b). Glass Pasteur pipettes were packed with glass wool and activated isoreticular Co-L-GG(R) MOFs. Racemic DL-Pen dissolved in methanol was added dropwise to each column, allowing for diffusion through the packed column (Fig. 2c). This solution was cycled 3 times back into the column before the packed MOF beds were then washed with ethanol to remove surface-bound molecules and then air-dried. The saturated chiral adsorbents were then washed with pure anhydrous methanol under heat to provoke diffusion of any remaining adsorbed Pen within the pore channels. The resulting methanolic adsorbed solution was collected and subjected to circular dichroism (CD) analysis to determine the optical purity of Pen in the MOF bed and analyzed with ¹H NMR to determine total adsorption. The same process was repeated for (RS)-ibuprofen. More specific details on experimental procedure and analysis of data are provided in the SI.

Results and discussion

Structure description

Along with Co-L-GGbipy, five novel isoreticular MOFs were synthesized using different pillar ligands: 1,2-bis(4-pyridyl)ethane (bpe), 4,4′-vinylenebipyridine (vinylbipy), 4,4′-(1,4-phenylene)bispyridine (phenylbipy), 4,4′-bis(4-pyridyl)biphenyl (bpb), and meso-1,2bis(4-pyridinyl)ethane-1,2-diol (mbped) (see Fig. 3). Pink plate-shaped crystals were successfully grown for Co-L-GGbipy, Co-L-GGvinybipy, Co-L-GGmbped, and Co-L-GGphenylbipy. The Co-L-GGvinylbipy crystals were successfully characterized with single-crystal X-ray diffraction (SC-XRD) and provided a monoclinic (space group P2₁) structure with cell parameters: a = 5.5751(3) Å, b = 8.4418(4) Å, c = 35.879(2) Å, and β = 93.574(5)°. The unit cell volume was calculated as 1685.33(16) ų, while the accessible pore volume accounted for 19.7%, or 331.6 ų. The absolute L(S)-configuration was confirmed by a Flack parameter value of 0.118(14). The presence of the peptide ladder framework spaced by the vinylbipy linker was confirmed (Fig. S1). Single-crystal X-ray diffraction screening of Co-L-GGmbped and Co-L-GGphenylbipy yielded similar lattice parameters (Table S1), but the structures could not be determined because of the weak diffraction of the crystals. Attempts to grow crystals for Co-L-GGbpe and Co-L-GGbpb were unsuccessful as small crystals tended to agglomerate in solution, forming a spherical powder upon mounting.

Fig. 3 Powder X-ray diffraction pattern of Co-L-GG(R) MOFs where R = pillar ligands. Pillar ligands used for each structure are shown next to the diffraction pattern. Simulated pattern is derived from SC-XRD of Co-L-GGvinylbipy.

Crystal structures for Co-L-GGbpe and Co-L-GGmbped were computationally simulated by replacing the vinylbipy linker derived from SC-XRD structure with those of bpe and mbped (Fig. S2 and S3). Despite similar N⋯N distances along the bipyridyl linker, their accessible pore volumes determined by OLEX showed variation because of their structural differences. Although the vinylbipy linker is slightly smaller in length than bpe, the Co-L-GGvinylbipy MOF had a greater free pore volume at 331.6 ų (19.7% free) compared to Co-L-GGbpe with free volume of 294.0 ų (17.4% free). This is due to the torsion induced by the alkane group of bpe, which reduces accessible pore volume compared to the C=C bond in vinylbipy that prevents torsion of the linker. Co-L-GGmbped had the smallest free pore volume at 257.6 ų (15.3% free) because of the hydroxyl groups occupying accessible pore space. Each of their simulated diffraction patterns matched that of the bulk sample, confirming the desired MOF structure had been made (Fig. S2d and S3d).

Le Bail fit analysis was performed on the PXRD patterns collected from the bulk samples of Co-L-GGvinylbipy, Co-L-GGbpe, Co-L-GGmbped, Co-L-GGphenylbipy, and Co-L-GGbpb using GSAS II software (Fig. S4). The refined lattice parameters of the Co-L-GGvinylbipy bulk sample align well with the SC-XRD data. Co-L-GGvinylbipy, Co-L-GGbpe, and Co-L-GGmbped crystallize in the monoclinic space group P2₁. The MOF with the longest ligand (Co-L-GGbpb) crystallizes in the triclinic P1̄ space group. There is a major increase in the unit cell volume of Co-L-GGbpb compared to other synthesized bipy-type MOFs, which could possibly be due to the large size of bpb compared to other ligands. The MOF with the longest ligand, bpb, exhibited its smallest diffraction angle at 3.62 degrees, while the MOF using bipy, which has the smallest length, had a diffraction angle at 6.06 degrees.²⁹ These findings are consistent with previous literature reports and are indicative of isoreticular expansion.^29,31 Based on the similar crystal structure, unit cell parameters, PXRD patterns matching the simulated patterns derived from single-crystal data, and consistency with other reported analogues reported in the literature, we propose that these MOFs possess the same topology. Additional evidence supporting the formation of expanded isoreticular frameworks comes from their FTIR spectra and thermal stabilities provided by TGA (Fig. S5 and S6). All MOFs synthesized were non-porous to N₂ but porous to CO₂ as shown by their adsorption isotherms (Fig. S7).

Enantioselectivity of Co-L-GG(R) on racemic drugs

Enantiomeric excess was quantified using circular dichroism (CD), a spectroscopic technique that enables rapid optical characterization of chiral molecules.^32–34 To determine optical purity of resolved Pen, the molar CD, Δε [M⁻¹ cm⁻¹], was measured at λ_max = 227.3 nm, and the L-enantioenrichment was determined using the calibration equation provided with the L/D enantiomer concentration (Fig. 4a and b). It was found that the adsorbed Pen in the MOF crystals were L-enriched (Fig. S8), with the highest enantiomeric excess (L-Pen over D-Pen) of 60.1% achieved using Co-L-GGvinylbipy (Fig. 4c). Among the Co-L-GG(R) analogues, only Co-L-GGbipy and Co-L-GGmbped did not show any effect on the enantioselective adsorption of Pen. Co-L-GGbipy, with an accessible pore volume of 115 ų, was unable to adsorb Pen as this molecule is sterically excluded granted its larger molecular volume (V = 206 ų), resulting in a racemic mixture.

Fig. 4 (a) Circular Dichroism spectra of Penicillamine at various enantiomer concentrations. (b) Calibration plot generated from CD with corresponding enantioenrichment (red) of adsorbed Pen following resolution in packed Co-L-GG(R) MOFs. (c) Adsorbed quantity of L-Pen and D-Pen within various isoreticular Co-L-GG(R) MOFs ordered from increasing free pore volume. Co-L-GGbipy is omitted as there was no adsorption. Enantiomeric excess is calculated for each case.

Motivated by these findings, (RS)-ibuprofen (Ibu) was similarly tested for enantioseparation. Circular dichroism calibration was conducted for Ibu (Fig. S9a and b) and the same resolution experiments were performed. Due to similar configuration of S-Ibu to the chiral pore cavity resulting from the Co-L(S)-GG(R) framework, it was hypothesized that the adsorbed solution would exhibit (S)-enantioenrichment as observed with L-Pen adsorption. However, the CD spectra indicated a racemic mixture of (RS)-ibuprofen (enantiomeric excess of 0%) for each Co-L-GG(R) MOF (Fig. S9c). This outcome can likely be attributed to the larger size of Ibu (V = 333 ų) compared to Pen (V = 206 ų) which hinders the diffusion of Ibu through the pores and surface interactions responsible for enantioselective adsorption. To enable enantioselective adsorption for bulkier guest molecules such as Ibu, the ligand extension strategy can be expanded towards engineering metal–organic cages or expanded framework analogues with similar pore environments. This can be achieved by changing the chiral building units such as using longer peptides or using polytopic pyridyl linkers. These architectures provide precise tunability over cavity dimensions and interior functionality, enabling the retention of chiral and hydrogen-bonding environments while increasing accessible volume.

Influence of ligand extension on enantioselective adsorption

Homochiral MOFs with high surface areas and multiple active chiral pore sites are desired for the resolution of racemic drugs.³⁵ The extent of enantioseparation relies on both the size of the drug and the accessible pore channels within the MOF. All synthesized MOFs were porous to CO₂ but not towards N₂, strongly indicating that these structures are ultra-microporous. Such restricted porosity enables strong confinement and molecular size-exclusion, consistent with the observed selective adsorption behavior and the proposed separation mechanism. Among the six MOFs, Co-L-GGvinylbipy exhibited the highest ability to enantioselectively adsorb Pen, with an enantiomeric excess (L-Pen over D-Pen) of 60.1%. The arrangement of increasing free pore volumes in Fig. 4c were first ordered based on available structures and derived free pore volumes (MOFs using bipy, mbped, bpe then vinylbipy). The sequence of Co-L-GGphenylbipy and Co-L-GGbpb free volumes were arranged by observing the total adsorption of Pen relative to the other structures. The adsorption capacity follows the trend of linker length (vinylbipy < phenylbipy < bpb), which aligns with the general consensus that extended linkers lead to greater pore volume in MOFs (Fig. 4c).³¹

Although the total adsorption shows an increase with pore size, the enantioselectivity does not exhibit the same trend. Circular dichroism results revealed that increasing the pore size leads to a decrease in enantioselective adsorption (60.1% ee vinylbipy vs. 34.8% ee phenylbipy vs. 29.2% ee bpb). This is thought to arise from a decrease in van der Waals interactions and hydrogen bonding between the hydroxyl and amine groups of Pen and the oxygen atom of the carboxylate group in the peptide within the MOF structure, resulting in a reduction in overall adsorption. Smaller pores thus promote more energetically favored interactions between L-Pen and chiral pore-wall surface. This trend breaks when comparing Co-L-GGbpe (V = 294.0 ų, ee = 47.4%) and Co-L-GGvinylbipy (V = 331.6 ų, ee = 60.1%), which have comparable pore volumes and identical chiral environments. The difference in ee between these MOFs suggests that the conformation of the guest Pen adsorbed in the pore plays a crucial role in influencing the interactions between Pen and the active pore surface. In the case of Co-L-GGbpe, where the pores are smaller than optimal, Pen is forced into a constrained orientation within the pore, limiting the proximity-driven dispersive interactions with the pore surface.²⁸ On the other hand, in pores that are too large relative to the guest molecule, these interactions are weakened due to dilution by an increased quantity of adsorbed molecules. The Co-L-GGvinylbipy demonstrates a “just right” pore size that allows Pen to adopt a conformation that maximizes the interactions with the pore surface, resulting in enhanced enantioselectivity as shown in Fig. 5.

Fig. 5 Schematic illustration of L-Pen adsorbed in (a) Co-L-GGbpe and (b) Co-L-GGvinylbipy. The torsion of the bpe linker reduces accessible pore volume and forces the alignment of L-Pen in such a manner that inhibits surface interactions, reducing enantiomeric excess.

As previously described, all peptide MOFs except for Co-L-GGbipy and Co-L-GGmbped exhibited enantioselective adsorption of the smaller drug (Pen), while the larger drug ((RS)-Ibu) failed to selectively resolve through the chiral pore cavities. When Co-L-GG (no bipyridine linker) and L-GG were similarly tested to resolve Pen, no significant enantioseparation was observed, highlighting the importance of the homochiral MOF framework and pore aperture size. Co-L-GGbipy, with an accessible pore volume of 115 ų,²⁹ was unable to adsorb Pen, resulting in a racemic mixture. It is possible that the hydroxyl groups of Co-L-GGmbped occupy free volume within the pores, enhancing strong binding interactions of both enantiomers or by hindering diffusion of Pen in Co-L-GGmbped. ¹H-NMR analysis did not show any evidence of Pen after the racemic solution passed through the Co-L-GGmbped column and after the initial methanol wash. One possible hypothesis is that the polar hydroxyl groups, with their strong electron-donating character, enhance the affinity between the polar Pen and the chemical environment of the pore apertures. Consequently, the release of the adsorbed molecules is not energetically favored in methanol, resulting in a resolved solution devoid of Pen. Similar observations have been reported by Yongwu et al., where biphenol groups in chiral MOFs induced a microenvironment that promoted adsorption.²⁵ To completely release Pen, the saturated Co-L-GGmbped was heated overnight and washed once more with pure anhydrous methanol. ¹H-NMR analysis confirmed the presence of Pen (Fig. S10c) and CD confirmed that the solution was racemic with an enantiomeric excess of −0.15% (essentially racemic).

To further demonstrate the relationship between pore size and adsorption, ¹H NMR spectra were collected before and after each test for each MOF bed to quantify the total amount of resolved drugs (Fig. S10). While the solution showed enrichment of L-Pen, it is important to note that as the pore size increases, the ratio of D-Pen to L-Pen also increases, leading to a decrease in the L/D enantioselectivity ratio (Fig. 4c and Table S11). This indicates the impact of increased pore volume on the total adsorption and its effect on enantioselective adsorption by reducing the proximity interactions between Pen and the chiral pore surface. When comparing the total adsorbed Pen for Co-L-GGbpe and Co-L-GGvinylbipy, the quantity of Pen remains relatively constant across the racemic mixture and pure D and L-Pen solutions (Fig. S11). This is attributed to the complete saturation of drugs through physical adsorption and the strong affinity of Pen for Co binding via metal–chelator interactions, which hinder diffusion. However, in the case of Co-L-GGbpb, Pen can easily pass through its larger pore cavities with fewer interactions with the surface due to the increased pore volume. Unlike Co-L-GGbpe and Co-L-GGvinylbipy, the total adsorption of D-Pen in Co-L-GGbpb (0.40 mol_Pen/mol_MOF) is much lower than the observed saturation capacity (1.0 mol_Pen/mol_MOF) (Fig. S11), suggesting that the interactions, determined by the proximity of the drug to chiral surfaces, are limited in larger volume pores. The larger pores in Co-L-GGbpb may accommodate more guest Pen molecules, thereby decreasing the proximity–dispersive interactions with the pore surface required for physisisorption. On the other hand, the adsorption of pure L-Pen is greater for the racemic mixture, indicating that the pore shape is selective to L-Pen and that the saturation limit for this concentration has been reached.

Influence of pore capacity for enantioselective adsorption

To maximize the proximity-driven interactions of Pen with the pore surface of MOFs and enhance enantioselective adsorption, we conducted saturation experiments with Co-L-GGvinylbipy. Solutions of Pen at various enantiomer concentrations were gently stirred with equimolar amounts of the MOF for 48 h to ensure complete saturation, and the adsorbed Pen was collected for CD analysis.

A positive trend was observed for enantioselectivity as the concentration of Pen loading increased. The uptake of the drug and L-enantiomer within Co-L-GGvinylbipy and Co-L-GGbpb was plotted to explore the relationship with enantioselective adsorption (adsorbed L-Pen/adsorbed DL-Pen) in pores of different sizes as the MOF approached maximum drug loading and increased proximal interactions (Fig. 6a). Fig. 6b and c demonstrate that enantioselective adsorption increased as the concentration of Pen increased and was highest near the maximum saturation capacity. This can be attributed to the complete saturation of the pores and the maximization of the interactions between L-Pen and the chiral pore cavity. The maximum pore saturation for Co-L-GGvinylbipy was determined to be 0.75 mol_Pen/mol_MOF (Fig. 6a), and an intrinsic enantioselectivity constant for Co-L-GGvinylbipy toward Pen was calculated as 0.45 (L-selectivity%/saturation%) (Fig. 6b). For simplicity, this value represents a single constant as the slope of a linear regression that describes the dependency of enantioselective adsorption of L-Pen as a function of pore saturation in Co-L-GGvinylbipy. This value loosely follows a linear regression (R² > 0.92) and likely depends on several factors to be explored for future studies. At complete saturation, Co-L-GGvinylbipy exhibited an enantioselectivity of 86.8% and an enantiomeric excess of 76.1% for L-Pen over D-Pen (Fig. 6b). To validate CD results, chiral liquid-chromatography mass-spectrometry (LC-MS) was employed for quantifying the enantiomer extraction results, revealing an L-enantioselectivity of 75.9 ± 2.9% (Fig. S12). In contrast, although Co-L-GGbpb had a higher saturation limit at approximately 1.2 mol_Pen/mol_MOF, its enantioselective affinity for L-Pen (as determined by the slope) was significantly lower at 0.21 (L-selectivity%/saturation%), suggesting that pore saturation had less influence on enantioselective adsorption (Fig. 6c). This indicates that the ability of Co-L-GGbpb to selectively adsorb L-Pen is not as influenced by saturating the pore environment compared to Co-L-GGvinylbipy. This can be attributed to dilution from the presence of more D-Pen in the pore environment of Co-L-GGbpb, which inhibits access to active stereogenic and dispersive interactions along the pore surface.

Fig. 6 (a) Adsorption of racemic DL-Pen in equimolar amounts Co-L-GGvinylbipy (red) and Co-L-GGbpb (black) at various racemic concentrations to determine the MOF's saturation capacity for DL-Pen (line) and L-Pen (dash), (b) graph of enantioselectivity Co-L-GGvinylbipy and enantiomeric excess of adsorbed L-Pen as a function of pore saturation (c). Graph of enantioselectivity Co-L-GGbpb and enantiomeric excess of adsorbed L-Pen as a function of pore saturation.

Recyclability

Recyclability tests for Co-L-GGvinylbipy were conducted to assess stability of the MOF's enantioselective character. Anhydrous methanol was passed through Co-L-GGvinylbipy after the separation experiment, and the MOF was washed with ethanol to remove surface-bound molecules. The MOF was then reactivated to de-solvate the pores. The recycling tests were performed three times on the packed MOF bed of Co-L-GGvinylbipy and the enantiomeric excess remained relatively constant at 76.1 ± 6%. The structural integrity of the peptide MOFs was confirmed by PXRD, suggesting that the metal-chelating effects of Pen did not have any impact on the structure of Co-L-GG(R) (Fig. S13).

Conclusion

Our study highlights the impact of pores in isoreticular homochiral peptide MOFs on the enantioselective adsorption of racemic Pen, with the Co-L-GG(R) favoring L-Pen adsorption. The pillar ligand length and functionality influenced the available pore volumes in the MOF structures, which proportionally increased with adsorption capacity for Pen. The relationship between the pore volume and enantioselective adsorption of L-Pen is complex. MOFs with the larger pores (structures using phenylbipy and bpb linkers) showed decreased ee due to reduced interactions between Pen and the MOF pore surface, caused by dilution of guest molecules in the pore environment. Pores smaller than optimal (Co-L-GGbpe) decreased ee, likely due to forced molecular orientation that limits dispersive interactions within the pores. Pores too small (Co-L-GGbipy) hindered diffusion of the guest molecule entirely, resulting in a racemic mixture. The mbped ligand extension also decreased ee due to the presence of hydroxyl-containing linkers which inhibit Pen resolution. Thus, while chiral pores are significant for enantioselective adsorption, the pore size must also be adjusted to optimize surface interactions with the guest molecules. Saturation capacities of Co-L-GGvinylbipy and Co-L-GGbpb were investigated, demonstrating that enantioselective adsorption in optimally sized pores (vinylbipy linkers) is more influenced by pore saturation compared to larger pores. These peptide MOFs show excellent stability and recyclability for up to three trials.

This work builds upon current works using porous adsorbents for chiral separations by exploring the dependence of pore-size and saturation on chiral affinity between guest molecule and host site. For the first time, we expand this family of isoreticular peptide MOFs to also incorporate hydroxyl groups to explore how chiral adsorption is impacted by the presence of functional groups. While more research is needed in this area with pillar ligand choice, results from this study show a promising design strategy for systematic tuning and optimization of chiral separations for not only traditionally smaller chiral molecules, but also chiral drugs of increasing sizes that possess both desired and undesired pharmacological properties. The separation of DL-penicillamine in this pilot study serves as an example of the potential to separate harmful and beneficial stereoisomers among a wide range of pharmaceutical drugs. The promising chiral reactivity of Co-L-GG(R) MOFs opens possibilities for exploring isoreticular asymmetric catalysis, which can be utilized to fine-tune reactions requiring catalysts with highly active Lewis acidic sites and stereoselective pore environment.^36–39 In addition, stereoselective size-exclusion can be advantageous in promoting not only enantiopure drugs but also enantiopure side-products during the synthesis process; thus, reducing the need for drug purification steps and increasing accessibility to the public.^40,41

Author contributions

K. C. S. and J. H. conceived the research ideas and overall direction of the project. J. H. led the experimental work. A. K. Y. and A. G. carried out solid-state characterization to determine the structure and porosity of the materials. A. V. and M. A. S. analyzed the MOF powders and identified the unit cell parameters. A. P. C. and J. B. evaluated the enantiopurity of the drugs using circular dichroism, while M. A. C. assessed drug purity through mass spectrometry. A. H. developed the hypothetical structure for one of the MOFs presented herein. J. H. and K. C. S. wrote the manuscript with contributions from all authors.

Conflicts of interest

There are no conflicts to declare.

Data availability

All data supporting the findings of this study are available within the supplementary information (SI). Supplementary information: discussion of materials, characterization protocols, synthetic procedure, experimental protocol, crystal structures, crystallography data, Le Bail refinements, FTIR spectra, TGA plot, CO₂ adsorption isotherm, Circular Dichroism Spectra, ¹H-NMR, and LC-MS data. See DOI: https://doi.org/10.1039/d5tb02047f.

Additional data and materials are available from the corresponding author, Dr Kyriakos C. Stylianou, upon request.

CCDC 2288909 contains the supplementary crystallographic data for this paper.⁴²

Acknowledgements

K. C. S. thanks the Department of Chemistry at Oregon State University (OSU) for support through start-up funding. J. H. would like to thank the OSU Honors College for funding through the Experiential Scholarship, the Department of Chemistry at OSU for access to research facility, and the Baio Group in the Chemical, Biological and Environmental Engineering Department at OSU for access to their Circular Dichroism. A. K. Y. acknowledges support from the Department of Chemistry for the Milton Harris fellowship (2023). A. P. C. acknowledges support from the National Science Foundation Ascend Postdoctoral Fellowship under grant number 2137997. Part of this research was conducted at the Northwest Nanotechnology Infrastructure, a National Nanotechnology Coordinated Infrastructure site at Oregon State University which is supported in part by the National Science Foundation (grant NNCI-2025489) and Oregon State University. The work done at Oregon State University (M.A.S) was supported by NSF Grant No. DMR-2025615. This project was funded in part with Oregon State Lottery funds administered by Business Oregon through a High Impact Opportunity Project (M. A. C.).

References

1. L. A. Nguyen, H. He and C. Pham-Huy, Int. J. Biomed. Sci., 2006, 2, 85–100.
2. M. Lämmerhofer, J. Chromatogr. A, 2010, 1217, 814–856.
3. K. Santra, D. Bhowmick, Q. Zhu, T. Bendikov and R. Naaman, J. Phys. Chem. C, 2021, 125, 17530–17536.
4. K. Roszak and A. Katrusiak, Org. Lett., 2023, 25, 37–41.
5. L. Gan, M. T. Nord, J. M. Lessard, N. Q. Tufts, A. Chidambaram, M. E. Light, H. Huang, E. Solano, J. Fraile, F. Suárez-García, C. Viñas, F. Teixidor, K. C. Stylianou and J. G. Planas, J. Am. Chem. Soc., 2023, 145, 13730–13741.
6. D. H. Le, R. P. Loughan, A. Gładysiak, N. Rampal, I. A. Brooks, A.-H. A. Park, D. Fairen-Jimenez and K. C. Stylianou, J. Mater. Chem. A, 2022, 10, 1442–1450.
7. A. K. Yadav, A. Gładysiak, A.-Y. Song, L. Gan, C. R. Simons, N. M. Alghoraibi, A. H. Alahmed, M. Younes, J. A. Reimer, H. Huang, J. G. Planas and K. C. Stylianou, JACS Au, 2024, 4, 4833–4843.
8. S.-M. Xie, M. Zhang, Z.-X. Fei and L.-M. Yuan, J. Chromatogr. A, 2014, 1363, 137–143.
9. J. Yu and R. Xu, J. Mater. Chem., 2008, 18, 4021–4030.
10. Y. Jeong, H. W. Kim, J. Ku and J. Seo, Sci. Rep., 2020, 10, 16166.
11. T. D. Booth, D. Wahnon and I. W. Wainer, Chirality, 1997, 9, 96–98.
12. T. Duerinck and J. F. M. Denayer, Chem. Eng. Sci., 2015, 124, 179–187.
13. J. Navarro-Sánchez, A. I. Argente-García, Y. Moliner-Martínez, D. Roca-Sanjuán, D. Antypov, P. Campíns-Falcó, M. J. Rosseinsky and C. Martí-Gastaldo, J. Am. Chem. Soc., 2017, 139, 4294–4297.
14. Y. Liu, W. Xuan and Y. Cui, Adv. Mater., 2010, 22, 4112–4135.
15. P. Peluso, V. Mamane and S. Cossu, J. Chromatogr. A, 2014, 1363, 11–26.
16. M. E. Davis, ACS Catal., 2018, 8, 10082–10088.
17. J. S. Seo, D. Whang, H. Lee, S. I. Jun, J. Oh, Y. J. Jeon and K. Kim, Nature, 2000, 404, 982–986.
18. X. Kuang, Y. Ma, H. Su, J. Zhang, Y.-B. Dong and B. Tang, Anal. Chem., 2014, 86, 1277–1281.
19. W. Gong, Z. Chen, J. Dong, Y. Liu and Y. Cui, Chem. Rev., 2022, 122, 9078–9144.
20. F. G. Cirujano, N. Martín, N. Almora-Barrios and C. Martí-Gastaldo, Catal. Sci. Technol., 2021, 11, 6053–6057.
21. J. Dong, Y. Liu and Y. Cui, J. Am. Chem. Soc., 2021, 143, 17316–17336.
22. S. L. Heinz-Kunert, A. Pandya, V. T. Dang, P. N. Tran, S. Ghosh, D. McElheny, B. D. Santarsiero, Z. Ren and A. I. Nguyen, J. Am. Chem. Soc., 2022, 144, 7001–7009.
23. C. Martí-Gastaldo, D. Antypov, J. E. Warren, M. E. Briggs, P. A. Chater, P. V. Wiper, G. J. Miller, Y. Z. Khimyak, G. R. Darling, N. G. Berry and M. J. Rosseinsky, Nat. Chem., 2014, 6, 343–351.
24. C. Martí-Gastaldo, J. E. Warren, K. C. Stylianou, N. L. O. Flack and M. J. Rosseinsky, Angew. Chem., Int. Ed., 2012, 51, 11044–11048.
25. Y. Peng, T. Gong, K. Zhang, X. Lin, Y. Liu, J. Jiang and Y. Cui, Nat. Commun., 2014, 5, 4406.
26. M. S. Zavakhina, I. S. Khan, M. O. Barsukova, A. A. Sapianik, D. G. Samsonenko, D. N. Dybtsev and V. P. Fedin, Russ. Chem. Bull., 2018, 67, 1268–1272.
27. D. N. Dybtsev, A. L. Nuzhdin, H. Chun, K. P. Bryliakov, E. P. Talsi, V. P. Fedin and K. Kim, Angew. Chem., Int. Ed., 2006, 45, 916–920.
28. Z.-G. Gu, S. Grosjean, S. Bräse, C. Wöll and L. Heinke, Chem. Commun., 2015, 51, 8998–9001.
29. K. C. Stylianou, L. Gómez, I. Imaz, C. Verdugo-Escamilla, X. Ribas and D. Maspoch, Chem. – Eur. J., 2015, 21, 9964–9969.
30. Y. Wang, Q. Han, Q. Zhang, Y. Huang, L. Guo and Y. Fu, Anal. Methods, 2013, 5, 5579–5583.
31. M. Eddaoudi, J. Kim, N. Rosi, D. Vodak, J. Wachter, M. O'Keeffe and O. M. Yaghi, Science, 2002, 295, 469–472.
32. D. Leung and E. V. Anslyn, Org. Lett., 2011, 13, 2298–2301.
33. M. O. Okuom, R. Burks, C. Naylor and A. E. Holmes, J. Anal. Methods Chem., 2015, 2015, 865605.
34. N. Rahman and S. Khan, ACS Omega, 2019, 4, 4252–4258.
35. P. Dhurjad, C. S. Dhalaram, N. Ali, N. Kumari and R. Sonti, Chirality, 2022, 34, 1419–1436.
36. J. Ho, M. T. Nord, J. P. Stafford and K. C. Stylianou, Catal. Sci. Technol., 2022, 12, 6998–7002.
37. L. Ma, J. M. Falkowski, C. Abney and W. Lin, Nat. Chem., 2010, 2, 838–846.
38. K. Tanaka, M. Kinoshita, J. Kayahara, Y. Uebayashi, K. Nakaji, M. Morawiak and Z. Urbanczyk-Lipkowska, RSC Adv., 2018, 8, 28139–28146.
39. K. S. Jeong, Y. B. Go, S. M. Shin, S. J. Lee, J. Kim, O. M. Yaghi and N. Jeong, Chem. Sci., 2011, 2, 877–882.
40. L. Ma, C. Abney and W. Lin, Chem. Soc. Rev., 2009, 38, 1248–1256.
41. F. Song, C. Wang, J. M. Falkowski, L. Ma and W. Lin, J. Am. Chem. Soc., 2010, 132, 15390–15398.
42. CCDC 2288909: Experimental Crystal Structure Determination, 2025 DOI:10.5517/ccdc.csd.cc2gtssw.

---