Engineering nanoscale glyco-zeolitic-imidazolate frameworks: Insights into the mechanism of formation

Engineering nanoscale glyco-zeolitic-imidazolate frameworks: Insights into the mechanism of formation Rocío Rodríguez-Marín a,1 , Salvador R.G. Balestra b,1 , Said Hamad b , Elena M. S´ anchez-Fern´ andez a,** , Carolina Carrillo-Carri´ on c,* a Department of Organic Chemistry, Faculty of Chemistry, University of Sevilla, C/ Profesor García Gonz´ alez 1, 41012, Sevilla, Spain b Department of Physical, Chemical, and Natural Systems, University Pablo de Olavide, Ctra. Utrera km 1, 41013, Sevilla, Spain c Institute for Chemical Research (IIQ), CSIC-University of Seville, Avda. Am´ erico Vespucio 49, 41092, Sevilla, Spain ARTICLE INFO Keywords: Zeolitic-imidazolate frameworks Glycomimetics Glycolipids De novo encapsulation Electronic structure calculations ABSTRACT The efficient encapsulation of large carbohydrates into porous metal-organic frameworks (MOFs), and not simply attached to the MOF’s surface, is still challenging and underexplored. In this work we have investigated the scope of an optimized synthetic procedure following a biomimetic mineralization strategy for the encapsulation of a variety of therapeutic glycolipids within a Zeolitic-Imidazolate Framework-8 nanostructure (GlycoZIFs). In all cases, regardless of the glycosidic linkage nature of the glycolipid, we obtained uniform, crystalline and reproducible GlycoZIFs nanoparticles by using the same optimized experimental conditions, which demonstrate the versatility of our approach. Our experimental data revealed that the formation of glyco-micelles, by taking advantage of the surfactant-like character of these glycolipids, is key to promote the nucleation of ZIF-8 around, allowing thus a precise control of the spatial location and amount of glycodrug encapsulated in each ZIF-particle. In turn, the electronic structure calculations showed that there is a strong interaction between the hydroxyl groups in positions C3 and C4 of the glycone core of the glycolipid and the Zn atoms on the ZIF-8 surface, suggesting that those favourable glyco-ZIF interactions also played an important role to induce the ZIF-8 nucleation. Experimental control data and computational studies obtained with a protected glycolipid featuring O-acetyl groups supported that conclusion. 1. Introduction The implementation of efficient synthetic strategies aimed at designing glycomimetic drugs making use of a chemically versatile framework termed sp 2 -iminosugar, has given rise to an extensive collection of metabolically stable sp 2 -iminoglycolipids (sp 2 -IGLs) with immunomodulatory capabilities. They present remarkable activity against pathological complications derived from inflammatory events (i. e., diabetic retinopathy and nephropathy, asthma, and acute inflammation), proliferation of tumor cells or even infections caused by parasites [1]. In the context of inflammation-associated diseases, an extremely complex regulatory network takes place during the inflammatory process, involving two types of cellular responses, proinflammatory and antiinflammatory, which are activated by completely different signals. An imbalance of both responses leads to dysregulation of the immune system, triggering the inflammatory response. Several works have pointed out the role of some sp 2 -IGLs as modulators of the immune response, resulting in a decrease of the expression of proinflammatory cytokines along with the concomitant increase of antiinflammatory markers [1]. In the context of cancer, sp 2 -IGLs demonstrated to reduce both the proliferation and the migration of malignant cells supporting their abilities as antimitotic, proapoptotic, and antimetastatic agents. Likewise, the growth inhibition of intracellular forms of the parasite Leishmania donovani by some sp 2 -IGLs raised their potential as antileishmanial agents. The privileged structure of the sp 2 -iminosugars, carbohydrate-based analogues featuring a pseudoamide-type functionality with sp 2 -hybridation, makes it a suitable scaffold to perform N-, C-, O-, S-, Se-glycosidation reactions by different and efficient synthetic methodologies [2, 3]. In recent years, browsing the aglycone space of these amphiphilic * Corresponding author. ** Corresponding author. E-mail addresses: [email protected] (E.M. S´ anchez-Fern´ andez), [email protected] (C. Carrillo-Carri´ on). 1 Equal contribution. Contents lists available at ScienceDirect Materials Today Chemistry journal homepage: www.journals.elsevier.com/materials-today-chemistry/ https://doi.org/10.1016/j.mtchem.2025.102546 Received 24 September 2024; Received in revised form 30 December 2024; Accepted 22 January 2025 Materials Today Chemistry 44 (2025) 102546 Available online 31 January 2025 2468-5194/© 2025 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC license ( http://creativecommons.org/licenses/bync/4.0/ ). molecules by numerous structure-biological activity relationship studies have allowed us to select some sp 2 -IGL-based immunomodulatory glycodrugs bearing a hydrophobic tail, specifically a dodecyl lipid chain (C 12 H 25 ) attached to different unnatural glycosidic connectors (sulfoxides, sulfones, selenoureas, thioureas, among others) [4,5]. The large number of pathological processes in which sp 2 -IGLs can intervene, providing significant pharmacological benefits, makes them an outstanding group of stable chemical and structural carbohydrate mimetics with great added value. Two representative examples of this family of sp 2 -glycolipids selected by their remarkable therapeutic properties in inflammatory in vivo models are depicted in Scheme 1A (compounds 1 [6] and 2 [7]). The body of results achieved from different preclinical models spanning from in vitro, ex vivo and in vivo assays, certainly reflects the benefits triggered by a variety of sp 2 -IGLs against deleterious effects provoked by adverse inflammatory processes [1,8,9]. However, a constraining factor of these glycodrugs, and also frequently found in other amphiphilic pharmacological agents, is their low solubility in aqueous solutions and biological fluids, restricting their in vivo systemic bioavailability. Increasing drug doses that offset the poor biodistribution is not considered a suitable option due to the potential unwanted effects, hence drug solubility enhancement has become a challenging task for the scientific community over the years [10]. The use of nanoparticle (NP)-based technology that enables more efficient delivery of the bioactive loaded compound is well recognized as an appealing therapeutic approach. Among the diverse types of NPs used as nanocarriers (lipid-based NPs, polymerics, lipid-polymer hybrids, inorganics), metal-organic frameworks (MOFs)-based drug delivery systems (DDSs) have recently achieved prominent accomplishments in the biomedicine field when prepared on a nanometric scale (nanoMOFs) [11–13]. The unique features of some families of nanoMOFs (e.g., high drug loading capacity, good biocompatibility, tunable porosity and chemical composition, controlled and tunable biostability to minimize their in vivo accumulation) let them to be considered one of the most promising candidates for developing intracellular DDSs [13]. Although a great number of small therapeutic molecules has been succesfully loaded into MOFs to facilitate cell-uptake and attain improved therapeutic properties [14], the encapsulation of large biomolecules (e.g., proteins, nucleic acids, carbohydrates) in a precise and controlled way is not an easy endeavour. Symmetry defect and labile metal-ligand coordination are the main obstacles for synthesis of well-ordered crystalline materials incorporating biomolecules. Significant advances have been achieved in the last years for the encapsulation of proteins and nucleic acids [14–17]; however, the encapsulation of carbohydrates has been sparsely reported so far. To the best of our knowledge, only Falcaro and coworkers have incorporated a variety of glycosaminoglycan (GAG)-based drugs in metal-azolate frameworks [18]. They observed that after encapsulating the GAGs, the crystallinity and control on the particle size of the resulting biocomposites were seriously compromised, which may impose significant limitations in their further therapeutic applications. Using a different synthetic strategy, our group has recently reported the controlled encapsulation of a glycolipid within ZIF-8, specifically a selenium-based sp 2 -IGL (compound 1, Scheme 1A), with a precise control on the spatial distribution of the glycodrug [19]. Our method for the preparation of 1@ZIF nanoparticles was carried out under mild conditions at room temperature (RT) and in water, preventing thus the potential degradation or inactivation of the bioactive glycolipid mimetics during their encapsulation, and protecting them as they travel to target cells. In addition, by exploiting the pH-dependent degradation of the ZIF-8 structure, we attained an efficient intracellular delivery of the glycodrug under the acidic conditions inside endosomes upon uptake. This allowed us to demonstrate, for the first time, the synergistic antioxidant effect of 1@ZIF nanoparticles in human endothelial cells. Notably, 1@ZIF was also able to reverse, to a certain level, the oxidative stress induced in cells, which was beneficial for preserving critical endothelial functions such as angiogenesis and cell migration [19]. Encouraged by this recent achievement, herein we set out to explore whether the synthetic methodology applied for the encapsulation of the selenium-containing glycomimetic 1 in ZIF-8 was extensible to other iminosugar-based drugs, aimed to investigate the versatility of our optimized approach for the synthesis of a collection of diverse GlycoZIFs with noteworthy therapeutic applications. The selected sp 2 -IGLs (compounds 1–5, Scheme 1A) feature different glycosidic linkage-types (i.e., selenourea, sulfur atom and sulfone functionality), and also present differences in their glycone moiety, specifically the lack of hydroxyl group in C2 position in compounds 3 and 4 (termed 2-deoxy-sp 2 -IGLs), and the blocking of hydroxyl groups in C3/C4 positions by O-acetyl groups (compound 5). With this set of glycolipids, we will be able to evaluate the influence of both the aglycone nature and the saccharide residue on the physico-chemical properties of the synthesized GlycoZIFs. Combining experimental and computational studies, we plan to investigate the encapsulation mechanism of the glycolipid into the ZIFstructure and to define the key structural features required for the successful formation of the GlycoZIFs, which will allow us to predict, in the next future, the possible extension of the method to other families of carbohydrate-based drugs. Scheme 1. (A) General representation of the sp 2 -iminosugar-based glycodrugs, and specific chemical structures of the glycomimetics studied: sp 2 -IGLs (1, 2), 2deoxy-sp 2 -IGLs (3, 4) and di-O-acetylated-2-deoxy-sp 2 -IGL (5). (B) Schematic representation of the one-pot synthesis of GlycoZIFs nanoparticles, indicating the main objectives planned in this work. R. Rodríguez-Marín et al. Materials Today Chemistry 44 (2025) 102546 2 2. Materials and methods 2.1. Synthetic procedure for the preparation of glycodrugs Full structural characterization of new 2-deoxy-sp 2 -IGLs is described in detail in the Supporting information (Scheme S1, Figs. S1–S3). The following compounds: (1S)-(N ′ -dodecylselenoureido)-5N,6O-oxomethylidenenojirimycin (1) [6], (1R)-1-dodecylsulfonyl-5N,6O-oxomethylidenenojirimycin (2) [20], 3,4-di-O-acetyl-5N,6O-(oxomethylidene) nojirimycin iminoglycal [21] and (1R)-3,4-di-O-acetyl-2-deoxy-1-S-dodecyl-5N,6O-oxomethylidene-1-thionojirimycin (5) [22], were prepared according to previously reported procedures. The two novel 2-deoxyα -glycomimetics, 3 and 4, were successfully synthesized following the optimized synthetic protocol reported by our group for the stereoselective synthesis of iminosugar 2-deoxy-thioglycosides from the bicyclic sp 2 -iminoglycal carbamate promoted by cerium(IV) ammonium nitrate in the presence of dodecane-1-thiol [22]. Next, oxidation of the sulfur atom of the resulting di-O-acetylated-2-deoxyα -S-glycoside (5) employing an excess of meta-chloroperbenzoic acid (MCPBA) provided the α -glycosylsulfonyl derivative. Subsequent deprotection reactions under basic conditions of the oxidized sp 2 -IGL and its precursor thioether led to the target fully unprotected compounds 3 and 4 in excellent yields (see ESM for details, Scheme S1). The chemical structures of the as-synthesized 2-deoxy-sp 2 -IGLs were confirmed by nuclear magnetic resonance (NMR), mass spectrometry (MS) and elemental analysis (see ESM). As expected, analysis of the coupling constants (J 1,2a ~ 5–7 Hz) revealed the exclusive formation of the α -anomers in both cases due to the pronounced anomeric effect present in this family of compounds [1]. 2.2. Synthesis of GlycoZIFs The corresponding GlycoZIFs were prepared following the experimental procedure reported recently by our group [19]. Briefly, 3 mL of an aqueous solution of zinc nitrate (Zn(NO 3 ) 2 ⋅6H 2 O; 25 mM) was added over 3 mL of an aqueous solution of 2-methylimidazol (HmIM; 1.3 M) at RT under continuous stirring (350 rpm). Next, 0.6 mL of a methanolic solution of the corresponding sp 2 -IGL (1–4; 10 mM) was added dropwise under stirring. The resulting mixture was stirred for 2 min and left to stand for 2 h. The appearance of a homogeneous turbidity over time is indicative of the formation of the GlycoZIFs particles. After 2 h, the particles were collected by centrifugation (12,000 rcf, 15 min) and washed three times with methanol (MeOH). Finally, the particles were redispersed in MeOH at a concentration of 10 mg/mL and stored in the fridge (4 ◦C) until use. The as-prepared GlycoZIFs were named 1@ZIF, 2@ZIF, 3@ZIF, 4@ZIF. The same experimental conditions were employed using the protected sp 2 -IGL (1R)-3,4-di-O-acetyl-2-deoxy-1-S-dodecyl-5N,6O-oxomethylidene-1-thionojirimycin (5) to afford 5@ZIF. 2.3. Synthesis of ZIF-8 particles as controls Two types of ZIF-8 control particles were synthesized, ZIF-8 control 1 (ZIF-8/Cnt1) and ZIF-8 control 2 (ZIF-8/Cnt2), following the same experimental procedure described above but replacing the sp 2 -IGL solution by the addition of 0.6 mL of MeOH in the case of ZIF-8/Cnt1 or 3 mL of an aqueous solution of cetyltrimethylammonium bromide (CTAB; 2.0 ×10 −3 M) in the case of ZIF-8/Cnt2. In both cases, the purification step was identical as that performed with the corresponding GlycoZIFs. ZIF-8/Cnt1 and ZIF-8/Cnt2 particles were finally redispersed in MeOH at a concentration of 10 mg/mL and stored in the fridge (4 ◦C) until use. 2.4. General information for characterization of GlycoZIFs Transition Electron Microscopy (TEM) images were acquired using a JEOL TEM 1400 operated at 200 kV. Samples were prepared by drying a diluted dispersion of the particles on 200 mesh copper grids coated with Formvar/carbon film. Measurements of dynamic light scattering (DLS) and zeta-potential (ζ-potential) were performed using a Malvern Zetasizer Nano ZSP equipped with a 10 mW He–Ne laser operating at a wavelength of 633 nm and fixed scattering angle of 173◦. For DLS analysis, diluted samples were loaded into a quartz cuvette and three measurements, each consisting of twelve data runs, were taken at RT after an equilibration step of 120 s. The ζ-potential of the nanoparticles dispersed in Milli-Q water was measured with laser Doppler anemometry (LDA) by using the same Malvern Zetasizer Nano ZSP instrument. High-performance liquid chromatography-mass spectrometry (HPLC-MS) analysis were done using a Waters Alliance 2695 HPLC coupled to an ESI-ion trap mass spectrometer instrument (Bruker AmaZon). Samples were analyzed using 0.1 % formic acid eluting gradients at a flow rate of 0.3 mL/min. Spectra were registered in both positive and negative modes in the m/z 100–2000 range. Powder X-Ray Diffraction (PXRD) was performed using a Bruker D8Advance Diffractometer. X-ray radiation of Cu K α was used, and the measurement range was from 5◦to 70◦(2θ) with a step of 0.02◦(2θ). 1 H NMR spectra of the GlycoZIFs were recorded in CD 3 OD using a 400 MHz Bruker Avance III HD spectrometer. N 2 sorption isotherms (77 K) of powder samples were carried out in a Micromeritics Tristar II 3020 system. Before analysis, samples were degassed under vacuum for 18 h at 120 ◦C. The apparent surface areas were calculated from the Barrett–Emmett–Teller (BET) method in the pressure interval P/P o =0.01–0.3 (being P o the saturation pressure). Pore volume and external surface area were calculated by the t-plot method. The pore size distributions of mesopores were calculated from the desorption branch of the isotherm using the Barrett–Joyner–Halenda (BJH) method. Thermogravimetric Analysis (TGA) of powder samples was performed using a Thermal Advantage SDT-600 instrument with a general heating profile from 30 to 650 ◦C and using a heating rate of 5 ◦C/min under air in a flow of 100 mL/min. 2.5. Determination of critical micelle concentration (CMC) of glycodrugs The CMC of the studied sp 2 -IGL compounds was determined by DLS. Measurements were carried out in a glass cuvette at 25 ◦C and repeated three times. A series of solutions ranging from 0.02 to 1.0 mM was prepared in Milli-Q water from a methanolic stock solution of sp 2 -IGL (10 mM). The intensity values of scattered light as a function of concentration for the glycomimetics were depicted (see ESM for details). The scattering intensities detected for sp 2 -IGL concentrations below CMC gave an approximately constant value corresponding to that of water. The intensity started to show a linear increase with concentration at the CMC, since the number of micelles increased in the solution. The intersection of best fit lines drawn through the data points corresponds to the CMC value. Further evidence for micelle formation came from the correlation function curves, where intercepts of correlation functions became much higher (>0.7) after micelles formation. 2.6. Computational methods Tight-Binding Density Functional Theory calculations were carried out using the DFTB+(v. 24.1) code [23], within the GFN2-xTB Hamiltonian [24], which is the first parametrized tight-binding method to include electrostatic interactions and exchange-correlation effects up to second order in the multipole expansion, as well as the D4 Grimme dispersion model [25]. As a model for the ZIF-8 surface we used the (100) surface, which is the most stable. The surface dipole was removed by transferring two imidazolate atoms from the top to the bottom of the model surface. The planar models of the micelles were created using the Packmol code [26]. For all systems studied, in order to accommodate the reactive functional groups of the glycolipid molecules near the reaction sites, we first R. Rodríguez-Marín et al. Materials Today Chemistry 44 (2025) 102546 3 carried out restrained Molecular Dynamics (MD) simulations, using the Plumed code [27], introducing linear and harmonic restraints on the distance between the potentially reactive O atoms (those in the OH groups in system 1 and in the O-acetyl groups in system 2). The restraining energy (δE) is given by: δE=∑ijk 2(dij −a)2+m(dij −a), where k=25 kJ/(mol⋅ ˚ A2), m=25 kJ/(mol ⋅ ˚ A), a=2.7˚ A, and d is the distance between the reactive O atom and the superficial Zn atom. In order to facilitate molecular reassembling, the temperature was set to 450 K. The pressure was 1 bar. In all MD simulations the timesteps were set to 0.5 fs, and Nose-Hoover barostats (with a 10 fs time constant) and thermostats (with a 100 fs time constant) were employed to carry out the NPT MD simulations. The simulations were run for 10 ps. The last configurations were energy-minimised with RFO algorithms. Once the reactive functional groups were close to the reaction sites, we removed the restraints and performed NPT MD simulations, for 100 ps, starting from the energy-minimised structures, at a pressure of 1 bar and temperature of 298 K. 3. Results and discussion 3.1. Synthesis of ZIF-based nanostructures containing glycodrugs (GlycoZIFs) Being fully aware that solubility issues in in vivo systems of the aforementioned amphiphilic sp 2 -glycodrugs may diminish their extraordinary intrinsic therapeutic activity towards different pathologies, we consider in this work the possibility of expanding the ZIF-8based nanotechnological approach reported for compound 1. Extending the methodology to other sp 2 -IGLs would allow us to improve the intracellular delivery of each of these glycodrugs, and thereby maximize their therapeutic benefits. To explore the scope of the encapsulation approach, we selected four sp 2 -IGLs as target candidates (1–4, Scheme 1A), being 1 the compound already used in our previous work [19]. Since additional characterization of 1@ZIF have been performed in this work, we have included it throughout the manuscript. From the chemical point of view, the α -glycomimetics examined in this study showcase notable differences in their structural features not only regarding the α -glycosidic connector-type (S, SO 2 or SeC(NH) 2 ), but also in the glycone core due to the replacement of the hydroxyl group located in C2 position by a hydrogen atom in the piperidine residue, compounds 1 and 2 vs 2-deoxy-derivatives 3 and 4 (Scheme 1A). Once we prepared the selected sp 2 -IGLs, the corresponding GlycoZIFs were successfully synthesized by following a de novo or biomimetic mineralization encapsulation approach as previously reported for compound 1 [19]; see the detailed experimental procedure in section 2.2. The as-prepared GlycoZIFs were termed 1@ZIF, 2@ZIF, 3@ZIF and 4@ZIF, which stand for the particles containing the sp 2 -IGLs 1–4. We did not observe visual changes in the encapsulation process with the different glycolipids, detecting in all cases the gradual increase in turbidity over time as the GlycoZIF crystals grew. Furthermore, HPLC analyses of the supernatants after purification of GlycoZIFs confirmed the efficient encapsulation of the glycodrugs in all cases, obtaining very similar values of encapsulation efficiencies (96–98 %) and loading capacities (3–4 wt%). Also, the termed ZIF-8 control 1 (ZIF-8/Cnt1) and ZIF-8 control 2 (ZIF-8/Cnt2) were prepared in absence of glycodrug for comparison. For ZIF-8/Cnt1, the synthesis was performed employing the same optimized experimental conditions but replacing the corresponding sp 2 -IGL by methanol, to keep the amount of methanol constant in the final reaction mixture. Note that the glycodrug stock solution was prepared in methanol for solubility reasons. In this case, larger particles were obtained due to the absence of the amphiphilic glycolipid that acts as a modulator agent during the crystal formation. However, the size of the particles is decisive in their biological performance, since it affects the cellular uptake process and intracellular stability, so it is desirable to have control particles (without loaded drug) with a size similar to that of the therapeutic nanosystem. In this line, ZIF-8/Cnt2 was also prepared, where the sp 2 -IGL was replaced by the surfactant CTAB as a size controlling agent, allowing the particle size to be controlled (<100 nm) so that it was similar to that of GlycoZIFs. Next, the resulting GlycoZIFs and control particles were analyzed to determine their physico-chemical properties using different techniques: DLS, TEM, PXRD, NMR, N 2 physisorption and TGA. 3.2. Morphological and structural characterization of GlycoZIFs The hydrodynamic diameter (d h ) of the GlycoZIFs particles, 1@ZIF, 2@ZIF, 3@ZIF and 4@ZIF, dispersed either in MeOH (Fig. 1A) or Milli-Q water (Fig. 1B) was measured by DLS analysis. It is worthwhile noting that the redispersion of the GlycoZIFs in water did not significantly modify the hydrodynamic size. The results (Table S1, ESM) revealed that all GlycoZIFs, regardless of both their structural differences and the solvent employed, presented very similar hydrodynamic sizes (ca. 80 nm), and importantly, rather low polydispersity index (PDI), which was indicative of a homogenous population of particles. Comparison of intensity, volume and number-weighted size distributions obtained for the different GlycoZIFs particles showed small differences as expected due to the narrow size distributions (Fig. S4). The surface charge of these GlycoZIFs dispersed in water and in a phosphate buffer (PB) solution (0.01 M, pH =7.4) was also studied by ζ-potential measurements (Fig. 1C). The values for ζ-potential of the GlycoZIFs dispersed in water were ca. 10 mV, while GlycoZIFs dispersed in a PB solution (0.01 M, pH =7.4) exhibited negative charge, ca. −30 mV, likely due to the coordination of the HPO 4 2− to the Zn 2+ in the GlycoZIF surface. No significant changes in the surface charge of GlycoZIFs compared to the control nanosized ZIF-8 particles (ZIF-8/Cnt2) in any of the tested media confirmed that glycodrug molecules were indeed encapsulated within the ZIF-8 structure and not adsorbed on the surface of the particles. In order to check possible differences in the morphology (shape and size) and the homogeneity of the GlycoZIF nanoparticles depending on the sp 2 -IGL encapsulated, the particles were examined under the microscope. Representative TEM images of 1@ZIF, 2@ZIF, 3@ZIF, 4@ZIF and the controls (ZIF-8/Cnt1 and ZIF-8/Cnt2) are shown in Fig. 2. These results revealed that the sp 2 -IGLs molecules were incorporated forming micelles which are located roughly in the center of the structure. This fact seems to indicate that, under the optimized experimental conditions, the micelles present in the medium during the synthesis (i.e., aqueous mixture containing the precursors) act as seeds for the subsequent formation of a ZIF-8 shell around them. Notably, small differences in the structural features of the sp 2 -IGLs led to some changes in the final morphology of the GlycoZIFs particles. In this regard, some GlycoZIFs had a cubic morphology with rounded corners (1@ZIF, 2@ZIF), while others presented a quasi-spherical shape (3@ZIF, 4@ZIF). Control ZIF8/Cnt1 particles showed the typical dodecahedral shape, while the control ZIF-8/Cnt2 presented a cubic shape, as expected when using CTAB as surfactant agent [19]. This finding is not surprising since previously described results have already shown that different morphologies can be achieved by using different modulators agents (surfactants or organic amines) or by varying the concentration of these modulators in the mother solution [28]. Next, the crystallinity of the GlycoZIFs was analyzed by PXRD (Fig. 3), confirming the characteristic ZIF-8 single-phase sodalite topology crystallinity for all the GlycoZIFs regardless of the structural differences of the encapsulated sp 2 -IGL. As observed, all the diffraction peaks fit well with the simulated ZIF-8 and with both ZIF-8 controls, which reveal that the encapsulation of sp 2 -IGL micelles within the structure did not affect significantly the crystallinity of the ZIF-8 shell grown around them. However, a slight broadening of some diffraction peaks in all GlycoZIFs was also evident. This fact could be related to some structural defects within the nanostructure as result of the incorporation of the glycomimetics, as well as due to the smaller nanoparticle size of the GlycoZIFs (d h ~ 75–89 nm) compared to ZIF-8/Cnt1 (d h ~ R. Rodríguez-Marín et al. Materials Today Chemistry 44 (2025) 102546 4 509–515 nm) and ZIF-8/Cnt2 (d h ~ 103–110 nm) (Table S1). Likewise, irrespective of either the functionality at the glycosidic linkage or the presence/absence of the hydroxyl group located in C2 of the glycone moiety, 1 H NMR analyses allowed us to corroborate that these α -glycomimetics were not adsorbed on the surface of the nanoparticles but encapsulated inside the ZIF-8 framework, as previously reported for the hybrid 1@ZIF [19]. To this end, 1 H NMR spectra of compounds 2, 3 and 4, dissolved in CD 3 OD, and their corresponding GlycoZIFs (2@ZIF, 3@ZIF and 4@ZIF) dispersed in CD 3 OD, were recorded at 400 MHz and compared to those 1 H NMR spectra obtained Fig. 1. DLS number distributions of d h of GlycoZIFs and control ZIF-8 particles as dispersed in (A) MeOH or (B) Milli-Q water. (C) ζ-potential of GlycoZIFs and control particles dispersed in either water or PB (0.01 M, pH =7.4). R. Rodríguez-Marín et al. Materials Today Chemistry 44 (2025) 102546 5 after dissolving the GlycoZIF-based nanostructures under the presence of diluted sulfuric acid. The same protocol was followed with ZIF-8/Cnt1. Before the acid dissolution of the particles, 1 H NMR spectra of all the GlycoZIFs showed two signals corresponding to the protons of HmIM and no signals attributable to the protons of either the glycone core or the aliphatic linear chain. This result agrees with the incorporation of these glycomimetics within the rigid and solid ZIF-8 shell, being their resonance prevented when a magnetic field is applied. However, when 2@ZIF, 3@ZIF and 4@ZIF nanostructures were dissolved, due to the breaking of Zn–N bonds under acid experimental conditions, representative signals of protons belonging to the released sp 2 -IGLs were identified. As a representative example, 1 H NMR analyses for 2@ZIF are depicted in Fig. 4. 1 H NMR spectra for 3@ZIF, 4@ZIF and ZIF-8/Cnt1 are shown in the ESM (Figs. S5–S8). These results for all the sp 2 -IGLs under study clearly reveal their successful encapsulation into the ZIF-8 crystalline structure. To evaluate the effect on the porosity of ZIF-8 after the encapsulation of 1, 2, 3 and 4, N 2 sorption isotherms of the GlycoZIFs were measured (Fig. 5). Whereas ZIF-8/Cnt1 and ZIF-8/Cnt2 particles presented a reversible type I isotherm, typical for microporous materials, GlycoZIFs displayed type I/type IV isotherms with a hysteresis loop at high relative pressures, suggesting the presence of mesoporous. The Brunauer−Emmett−Teller (BET) apparent surface area (S BET ) decreased notably in GlycoZIFs (Table S2), which is attributed to the incorporation of sp 2 -IGLs within the structure. Anyway, all the GlycoZIFs showed very similar textural properties (micropore area (S micro ), external surface area (S ext ), micropore volume (V micro ) and mesopore volume (V meso ), see Table S2 for details), confirming again the similar behaviour of the different studied glycolipids. The significant decrease in the micropore area accompanied with the notable increase of external surface area may be attributed to the presence of glyco-micelles inside the framework. Note that the inclusion of drugs as individual compounds (not forming assemblies or micelles) does not usually lead to such a large increase in the external surface area. BJH pore size distributions curve of GlycoZIFs (Fig. S9 and Table S2) showed that mesoporous were present within the structure. Therefore, these results seem to indicate that the sp 2 -IGLs were encapsulated into ZIF-8 in the form of micelles, consistent with TEM observations, and thereby produced some mesoporosity while preserving the microporosity of ZIF-structure to a large extent. Next, the thermal behaviour of the GlycoZIFs was also evaluated (Fig. S10). The TGA curve of control ZIF-8/Cnt1 particles matched well with the reported literature [29], with a sharp weight loss at around 440 ◦C that corresponds to the transformation from ZIF-8 into ZnO. Notably, the encapsulation of sp 2 -IGLs led to changes in TGA profiles of the corresponding GlycoZIFs, showing an early mass loss at ~300 ◦C, which could be attributed to the decomposition of sp 2 -IGLs molecules encapsulated within the ZIF-structure. Finally, the release of glycodrugs from the corresponding GlycoZIFs was studied under two biological conditions, i.e., extracellular and intracellular environments, which differ in the pH of the surrounding media. To do this, the GlycoZIFs were incubated for 24 h in either Tris buffer solution at pH =7.4 (emulating extracellular conditions) or in acetate buffer solution pH =4.5 (emulating conditions inside the endosomes), and we quantified the amount of glycodrug (1, 2, 3 or 4) Fig. 2. Representative TEM images of GlycoZIFs: 1@ZIF, 2@ZIF, 3@ZIF, 4@ZIF, and control particles: ZIF-8/Cnt1 and ZIF-8/Cnt2. Note that in some cases to more clearly observe the micelle inside the particle it is necessary to take the image out of focus. R. Rodríguez-Marín et al. Materials Today Chemistry 44 (2025) 102546 6 delivered to the medium by HPLC. As shown in Fig. S11, there was a successful release at pH =4.5 for all GlycoZIFs, reaching between 74 and 79 % after 24 h, while the amount of glycodrug released was minimal (<10 %) at pH =7.4. These results were expected due to the pHdependant stability of the ZIF-8, as the Zn–N coordination bonds are broken at acidic pH, resulting in the dissolution of the particles and subsequent release of the encapsulated cargo. Taking advantage of this pH-responsiveness of the GlycoZIFs, we could achieve a controlled intracellular delivery of the target glycodrugs, solving their solubility limitations and increasing the amount of glycodrug reaching the interior of cells, which is anticipated to result in better therapeutic efficiency. 3.3. Mechanistic hypothesis for the formation of GlycoZIFs All the above characterization results together demonstrate that the studied sp 2 -IGLs 1–4 present an analogous behaviour with regard to their incorporation into the ZIF-8 structure during crystal formation, also giving rise to GlycoZIFs with very similar physical-chemical properties. On this basis, we could infered that the common sp 2 -iminosugar polar head and the hydrophobic alkyl chain are key structural motifs that allow the successful preparation of GlycoZIFs’ structures. Notably, the nature of the α -glycosidic connector or the lack of the –OH group located in C2 position of the glycone core were found to be irrelevant for the GlycoZIFs formation. Based on the experimental results achieved so far, we aimed to disclose the molecular mechanism involved in the nucleation and growth stages during the formation of GlycoZIFs particles. We observed that the concentration of the glycolipid was a key factor for achieving their proper encapsulation [19], which seems to be associated to the presence/absence of glycolipid micelles in the reaction mixture. Low glycolipid concentrations (<0.2 mM) did not lead to the formation of GlycoZIFs, but ZIF-8 particles were obtained in which, curiously, changes in size and morphology were observed when compared with the control ZIF-8 sample (ZIF-8/Cnt1, prepared in the absence of glycolipid). This finding indicated that glycolipid is capable Fig. 3. PXRD patterns of 1@ZIF, 2@ZIF, 3@ZIF, 4@ZIF and control particles (ZIF-8/Cnt1, ZIF-8/Cnt2). Magnifications of the 2θ range to clearly visualize some broadened peaks in the GlycoZIFs patterns are presented as insets. Simulation of ZIF-8 (Crystallography Open Database: 7111970) with hkl planes corresponding to each peak is also shown. R. Rodríguez-Marín et al. Materials Today Chemistry 44 (2025) 102546 7 of acting as a size/shape-controlling agent, just as reported for other surfactant molecules such as CTAB. In contrast, when the glycolipid was added to the reaction media at a final concentration of 0.9 mM, at which the compound is mostly as micelles, we obtained GlycoZIFs particles as those presented along the work, incorporating the glycolipid with a good encapsulation efficiency and showing the presence of one glyco-micelle per ZIF-particle as obseved under the microcoscope. At such concentration, the glycolipids would be mostly as micelles in the media, H 2 O:MeOH (10:1), according to CMC estimated from DLS measurements (discussed later). These results point out that the glyco-micelles trigger the spontaneous formation of ZIF-8 shell, likely acting as nucleation seeds in a similar way to what was reported with inorganic nanoparticles (e.g., Au NPs and Pd NPs) in the formation of core-shell NP/ZIF-8 composites [30,31]. Therefore, we hypothesize that the mechanism begins with the fast formation of micelles when the glycolipid is added to the reaction mixture containing the ZIF-8 Fig. 4. 1 H NMR spectra (400 MHz, CD 3 OD) of: (a) 2@ZIF, (b) dissolved 2@ZIF, (c) free sp 2 -IGL 2 bearing a sulfonyl group as glycosidic linkage. R. Rodríguez-Marín et al. Materials Today Chemistry 44 (2025) 102546 8 precursors. As depicted in Fig. 6, Zn +2 ions are concentrated to the glyco-micelle surface due to Zn–OH coordinating bonds, producing a local supersaturation that favors ZIF-8 crystal growth around those micelles. As a result, GlycoZIFs crystalline nanoparticles are formed in just a few minutes, visually observable by the appearance of slight turbidity instantly after the addition of any of the glycolipids, whereas the particles size increases further with time until reaching the maximum size after 2 h, according to previous experimental studies [19]. To obtain some experimental evidence that supports the proposed mechanistic hypothesis, we conducted the following different experiments. First, to determine if sp 2 -IGLs were able to form micelles as proposed in the reaction media, we addressed the determination of CMC values of 1, 2, 3 and 4 using a method based on DLS measurements (see ESM, Table S3). According to the results, we could thus confirm that at the concentrations used in the synthesis of the GlycoZIFs, sp 2 -IGLs Fig. 5. N 2 adsorption (filled circles) and desorption (empty circles) isotherms of the as-prepared GlycoZIFs and control ZIF-8 particles. Fig. 6. Proposed mechanism for the formation of GlycoZIFs particles, showing the role of the glyco-micelles as nucleation seeds to promote the growth of the ZIF-8 shell around. R. Rodríguez-Marín et al. Materials Today Chemistry 44 (2025) 102546 9