ABSTRACT
Achieving precise control over supramolecular topology and nuclearity in coordination‐driven self‐assemblies through integration of multiple stimuli remains a challenge. We here introduce ligand LA , based on a flexible and reactive 1,2‐dicarbonyl benzil‐based backbone, able to adopt several conformations. A range of homo‐ and heteroleptic Pd(II) assemblies is formed with different nuclearities and topologies, connected through distinct transformation pathways. In homoleptic systems, folded and open conformations of LA afford the mononuclear complex PdLA 2 and the lantern‐shaped cage Pd2 LA 4. Incorporation of secondary ligands LC or LD induces more expanded LA conformations, yielding cis‐Pd2 LA 2 LC 2 or a rare isosceles triangular Pd3 LA 2 LD 4 ring. These heteroleptic architectures can be interconverted, and both can undergo guest‐induced retro‐cage‐to‐cage transformation to regenerate homoleptic species. Furthermore, LA can also be conformationally locked through condensation of its benzil backbone with 1,2‐phenylenediamine, producing the rigid quinoxaline ligand LB . This transformation can occur in a post‐assembly fashion, converting both homoleptic and heteroleptic LA ‐based structures into the PdLB 2 complex. The conformation and accessibility of LA reactive sites within the architectures, as well as the nature of the encapsulated guest, strongly influence assembly stability and reactivity in this condensation, reminiscent of how nature controls functional group reactivity through effects of nanoscopic confinement, conformational restriction, and allosteric regulation.
Keywords: chemical networks, coordination cages, post‐assembly modification, self‐assembly, supramolecular chemistry
Incorporation of a flexible and chemically reactive benzil backbone into the ligand structure allows it to access multiple conformations, enabling the formation of diverse homo‐ and heteroleptic Pd(II) assemblies with different nuclearities and topologies. This conformational versatility also supports multiple stimuli‐induced transformations driven by guest binding, secondary ligand incorporation, or condensation with phenylenediamine.

1. Introduction
Due to their intrinsic flexibility, biomolecular systems based on proteins, RNA, and DNA can undergo changes in shape and size in response to external stimuli, enabling them to adapt to dynamic environments, modulate their functions, and participate in multistate regulatory networks [1, 2, 3, 4]. Inspired by these natural systems, synthetic supramolecular architectures have been designed to emulate such adaptive, shape‐shifting behavior. Among them, metallosupramolecular self‐assemblies such as metallomacrocycles and coordination cages occupy a prominent position, as the dynamic nature of coordination bonds allows for reversible structural reorganization under various external stimuli (Figure 1a) [5, 6]. This responsiveness enables such artificial constructs to mimic biological processes, including allosteric binding [7], molecular recognition [8] and transport [9], enzyme‐like catalysis [10, 11], and multi‐state transformational networks [12].
FIGURE 1.

(a) Stimuli‐induced nuclearity and topology transformations of Pd(II) coordination cages. (b) Flexible ligand LA , investigated in this work, featuring a 1,2‐dicarbonyl benzil backbone that enables the formation of multiple Pd(II) assemblies with variable nuclearity, prone to multi‐stimuli‐responsive transformations.
One notable manifestation of size‐ and shape‐shifting behavior in metallosupramolecular architectures is a change in nuclearity and topology, that is, the number of metal atoms within an assembly and their connectivity via bridging organic ligands [13]. Many Pd(II) coordination cages are capable of altering their nuclearity/topology in response to external factors such as a change in component ratio, solvent, concentration, or temperature [14, 15, 16, 17, 18]. It was shown that the geometry of the organic ligand plays a crucial role in determining the size and shape of the resulting assembly [13, 19, 20]. While ligand conformational flexibility often enables structural adaptivity of supramolecular structure formation [21, 22, 23, 24] and facilitates interconversion between species of different nuclearities and topologies, it may also complicate the isolation of discrete assemblies [25, 26, 27, 28, 29]. In such cases, guest inclusion can be exploited to template architectures of specific dimensions or connectivity [30, 31, 32, 33]. In some instances, guest binding or exchange can even trigger interconversion between distinct assemblies [34], for example, in the reversible restructuring of azulene‐based lantern‐shaped Pd2L4 cages to Pd4L8 tetrahedra reported by us [35], or the ion‐dependent reversible interlocking of Pd2L4 cages to give Pd4L8 catenanes, first reported by Kuroda [36].
Furthermore, homoleptic cages (containing only one type of organic ligand) may undergo changes in size and nuclearity upon introduction of a second kind of ligand or through cage‐to‐cage transformations under consumption of another assembly [37]. The formation of heteroleptic structures by merging of homoleptic species can give rise to nuclearities and structural motifs that are inaccessible in assemblies based on only one kind of ligand [38, 39, 40]. Consequently, this approach has recently opened access to a wide range of new supramolecular architectures of low symmetry and unique topology [23, 41, 42, 43, 44, 45, 46].
Incorporation of addressable functional moieties such as photoswitchable [47], redox [48], or reactive [49] units into ligand backbones can provide additional control over assembly size and shape by inducing light‐, electron‐, or reagent‐driven changes in ligand geometry. For instance, the introduction of azo(het)aryl functionalities into lantern‐shaped Pd2L4 coordination cages enables their conversion, via ligand E–Z photoisomerization, into assemblies of different nuclearities, e.g. PdL2 as shown by Beves [50, 51] or Pd3L6 reported by von Krbek [52]. We introduced Pd(II) systems based on photoactive dithienylethene ligands, in which photoswitching converts Pd3L6 rings into Pd24L48 nanospheres [53], while a Pd2L4 lantern‐shaped cage can be transformed into a Pd2L3(solv)2 bowl [54]. Recently, Goeb and our group reported a rare case of redox‐induced topology switching between a Pd2L4 cage and a homoleptic Pd2L2(solv)4 macrocycle or a heteroleptic Pd2L2L′(solv)2 bowl using a redox‐active, π‐extended tetrathiafulvalene (exTTF)‐based ligand [55]. Examples of nuclearity and topology switching achieved through covalent post‐assembly modification also remain rare, owing to the limited number of chemical transformations capable of altering ligand geometry while preserving the integrity of the metallosupramolecular assembly [49, 56, 57]. To the best of our knowledge, concerning topology/nuclearity switching in Pd(II) systems, the only reported example was described by Pilgrim and co‐workers, who utilized an inverse electron‐demand Diels–Alder reaction on tetrazine‐based ligand backbones, followed by oxidation, to interconvert between Pd4L8 tetrahedra and Pd2L4 lantern‐shaped cages [58].
While the action of an isolated individual stimulus such as the addition of a ligand, guest, or chemical reagent or the irradiation with light has been reported to control the nuclearity and topology of Pd(II) supramolecular assemblies, the combination of several distinct stimuli within a single, interconnected transformation network remains challenging [12]. Herein, we introduce ligand LA , which contains a flexible and chemically reactive 1,2‐dicarbonyl benzil‐based backbone (Figure 1b). Unrestricted rotation around the central carbon–carbon bond allows LA to adopt several conformations, making it suitable for incorporation into a variety of homo‐ and heteroleptic Pd(II) assemblies with different nuclearities and topologies. Its flexibility allows it to progress between different conformational states in response to external stimuli, such as the addition of a guest or secondary ligand, thus inducing the formation of specific metallosupramolecular structures. In addition, the 1,2‐dicarbonyl group can undergo condensation with diamines, for example, 1,2‐phenylenediamine, to generate a rigid quinoxaline backbone, further altering ligand geometry. This reactivity provides an additional chemical stimulus for the post‐assembly transformation into new species. Taken together, the combined action of three different types of stimuli (ligand, guest, and reagent) allowed us to establish a multistate transformational network that connects a range of homoleptic and heteroleptic Pd(II) assemblies of different sizes and shapes.
2. Results and Discussion
2.1. Homoleptic Assembly Formation
Benzil‐based ligand LA with isoquinoline donor groups can react with 1,2‐phenylenediamine (PDA) under the transformation of its flexible 1,2‐dicarbonyl backbone into a rigid quinoxaline moiety in the resulting ligand LB (Scheme 1A) [59]. This transformation acts as a conformational lock and results in a drastic change in ligand shape, affecting its mode of interaction with metal cations. While flexible LA has a hinge‐like shape, allowing it to adopt several conformations suitable for forming various assembly topologies, rigid derivative LB has a pincer‐like shape, more favorable for metal chelation.
SCHEME 1.

Transformations of ligand LA and its multiple Pd(II)‐based self‐assemblies. (A) Condensation of LA with 1,2‐phenylenediamine (PDA); self‐assembly of Pd2 LA 4/PdLA 2 (B), PdLB 2 (C), Pd2 LA 2 LC 2 (D), and Pd3 LA 2 LD 4 (E) species; (F) interconversion of Pd2 LA 2 LC 2 and Pd3 LA 2 LD 4 heteroleptic assemblies; (G) guest encapsulation in cage Pd2 LA 4; (H) guest‐induced transformation of Pd2 LA 2 LC 2 and Pd3 LA 2 LD 4 heteroleptic assemblies into G@Pd2 LA 4; post‐assembly condensation of homoleptic (I) and heteroleptic (J) Pd(II) species with PDA.
The reaction of ligand LA (2 eq.) with [Pd(CH3CN)4](X)2 (1.1 eq.) (X = BF4 – or OTf–) in CD3CN at room temperature (RT) almost immediately produced two species distinct from the initial ligand (Scheme 1B). Over time, one species gradually converted into the other, indicating the presence of kinetic and thermodynamic products (Figure S17). After 5 days at room temperature (RT), the kinetic product had almost completely disappeared, giving an equilibrium product ratio of approximately 11:1 for the OTf– counterion (Figure 2a). For the assembly with the BF4 – counterion, the NMR signals of the thermodynamic product were highly broadened at room temperature (complicating the determination of the product ratio) but became noticeably sharper upon heating to 70°C, closely matching the signal positions of the thermodynamic product formed with triflate (Figure S10). These observations suggest that the assembly with BF4 – leads to a more dynamic structure, whereas the larger and more coordinative OTf– anion more effectively stabilizes a distinct conformation of reduced structural flexibility [60]. In both systems, ESI‐MS analysis revealed ions corresponding to a Pd2 LA 4 cage and a mononuclear PdLA 2 complex (Figure 2b and Figure S12), consistent with the presence of two products in solution. Additionally, 1H DOSY measurements indicated two species with hydrodynamic radii (r H) of approximately 10 Å for the major thermodynamic product and approximately 8 Å for the minor kinetic one (Table S5), corresponding to larger Pd2 LA 4 and smaller PdLA 2 structures, respectively (Figure 2a). Notably, the use of alternative solvents such as CD3NO2 and DMSO‐d6 leads to further stabilization and exclusive formation of the Pd2 LA 4 cage structure, as confirmed by ESI‐MS (Figure S12) and NMR analyses (Figure S10). Interestingly, in DMSO‐d6 , a twofold splitting of the ligand signals is observed, accompanied by an increase in the hydrodynamic radius to 11.1 Å. A similar behavior was previously observed by us in an acridone‐based Pd(II) helical cage, where such an effect was caused by a solvent‐induced (also from CD3CN to DMSO‐d6 ) transition from a helicate to a broken‐symmetry mesocate‐like structure [22]. Indeed, DFT optimization of the Pd2 LA 4 structure starting from a mesocate conformation leads to a “half‐helical” cage geometry, in which one Pd(isoquinoline)4 unit is twisted while the other remains untwisted (Figure S58). This geometry is consistent with the observed twofold signal splitting in the NMR spectrum and features an increased Pd–Pd separation (15.1 Å), in agreement with the larger hydrodynamic radius in DSMO‐d6 solution.
FIGURE 2.

(a) 1H NMR spectra (CD3CN) of LA (top) and the Pd2 LA 4/PdLA 2 mixture (with OTf– counter‐ions; bottom) with DOSY traces for Pd2 LA 4 and PdLA 2; (b) ESI mass spectrum of the Pd2 LA 4/PdLA 2 mixture with OTf– counter‐ions; (c) 1H NMR spectra (CD3CN) of LB (top) and PdLB 2 (bottom) with DOSY trace for PdLB 2; (d) ESI mass spectrum of PdLB 2; (e, f) X‐ray structures (Pd2 LA 4 and PdLB 2), DFT models (PdLA 2 and Pd2 LB 4) and calculated free energy differences between corresponding cages and mononuclear complexes.
Vapor diffusion of diethyl ether into the reaction mixture with BF4 – counter‐ions in CD3CN yielded crystals suitable for X‐ray diffraction, allowing the self‐assembled structure to be determined in the solid state. As expected, the crystal structure corresponds to the Pd2 LA 4 cage assembly, which adopts a helical shape (Figure 2e). This finding also explains the signal broadening observed in the NMR spectra of Pd2 LA 4 with BF4 – counter‐ions, as the helicate cage is likely to equilibrate between structures with opposite twisting senses as well as a mesocate form in solution [22, 61]. In the solid state, cage Pd2 LA 4 crystallizes in the achiral P‐1 space group, and the unit cell contains both P‐ and M‐helical species, each encapsulating one acetonitrile and one BF4 – ion in their cavities (Figure S57). The azimuthal angle defined by the N‐Pd vectors of a single ligand in the helical assembly was found to be approximately 135°, while the Pd–Pd separation in the prolate structure is approximately 13.7 Å, thus rather matching the 1H DOSY NMR results obtained for the larger (and major) species obtained in solution that seems to additionally associate with outside‐bound anions.
The NMR and ESI‐MS data show that the conformational flexibility of LA comes with free rotation along the C─C bond in its 1,2‐dicarbonyl unit. Hence, LA can act as either a bis‐monodentate or a bidentate ligand, leading to the formation of Pd2 LA 4 cages and PdLA 2 chelates in solution. The equilibrium constant between these species with OTf– counter‐anions in CD3CN at 295 K was estimated to be 1.8 × 105 M−1, with a free energy difference of about –30 kJ/mol favoring the formation of the cage structure. The higher stability of the helical cage structure was also supported by DFT calculations (Figure 2e; geometry optimization at the r2scan‐3c [62]/CPCM(CH3CN) [63] level and single‐point energy calculations at the ωB97X [64]‐D4 [65]/def2TZVPP [66]/SMD [67]‐DRACO [68](CH3CN) level with thermostatistical corrections by the mRRHO approach [69]; see Section S7, Supporting Information). For the mononuclear assembly, the calculations suggest that the trans‐PdLA 2 isomer is about 35 kJ/mol more energetically favorable than a hypothetical cis‐isomer (Table S6).
In contrast to LA , the reaction of LB featuring a rigid quinoxaline backbone with Pd(II) in CD3CN produces only one product (Scheme 1C), exhibiting sharp signals in the 1H NMR spectrum whose positions more closely resemble those of the mononuclear PdLA 2 species rather than the Pd2 LA 4 cage (Figure 2c). The 1H DOSY measurements revealed an r H value of 8.7 Å for this species, which is larger than for the PdLA 2 chelate but smaller than for the Pd2 LA 4 cage. Meanwhile, ESI‐MS clearly indicates the presence of ions corresponding exclusively to the mononuclear PdLB 2 complex (Figure 2d). Furthermore, single‐crystal X‐ray diffraction confirmed the assumed trans‐PdLB 2 structure in the solid state (Figure 2f). According to DFT calculations, for LB the formation of the mononuclear trans‐PdLB 2 structure is indeed more favorable than that of the Pd2 LB 4 cage (ΔG calcd = +9 kJ/mol, Figure 2f), most probably due to a larger ligand strain energy (+13 kJ/mol per ligand) that the rigid and planar quinoxaline backbones would experience in a cage assembly. The change of counter‐ion (OTf–), solvent (CD3NO2 or DMSO‐d6 ), or guest addition (aliphatic or aromatic disulfonates) does not affect the self‐assembly outcome.
2.2. Heteroleptic Assembly Formation
We next investigated the ability of the studied ligands to form heteroleptic assemblies in the presence of Pd(II) cations (Scheme 1D,E). As a bis‐monodentate ligand, the open conformation of LA is expected to have converging metal‐binding vectors due to its isoquinoline donor groups. Therefore, as shape‐complementary partners, we selected 2,7‐di(pyridin‐3‐yl)‐fluoren‐9‐one (LC ) and 3,6‐di(pyridin‐4‐yl)‐9,10‐dimethoxyphenanthrene (LD ), which feature diverging binding vectors. These ligands have previously enabled the clean formation of cis‐Pd2 A 2 B 2‐type heteroleptic assemblies when combined with converging ligands containing isoquinoline donor groups and rigid acridone or carbazole backbones [22, 70]. Moreover, using LC in combination with a flexible chiral ligand, we have recently achieved, for the first time, a low‐symmetry isosceles triangular Pd3 A 2 B 4 topology [41].
The reaction of LA , LC , and [Pd(CH3CN)4](BF4)2 in a 1:1:1 ratio at 70°C for 90 h resulted in the formation of a new species (Scheme 1D), showing one distinct set of signals for both LA and LC ligands, different from those of the free ligands and their corresponding homoleptic assemblies (Figure 3a). The same product could also be obtained through the reaction of the corresponding homoleptic cages (Figure S30). 1H DOSY measurements revealed an r H value of 10.6 Å for this species (Figure 3a), which is larger than that of the homoleptic Pd2 LA 4 cage. We were successful in obtaining an X‐ray structure of the cis‐Pd2 LA 2 LC 2 cage that showed a Pd–Pd distance increase to 14.4 Å as compared to the homoleptic Pd2 LA 4 helicate, consistent with the DOSY results (Figure 3b). Furthermore, ESI‐MS analysis confirmed the proposed stoichiometry, showing ions corresponding to [Pd2 LA 2 LC 2 + nBF4]4– n (n = 0–2; Figure 3c). Notably, altering the component ratio of LA , LC , and Pd(II) did not lead to the formation of any other new discrete species.
FIGURE 3.

(a) 1H NMR spectra (CD3CN) of LA (top), LC (middle), and Pd2 LA 2 LC 2 (bottom) species with DOSY trace for Pd2 LA 2 LC 2; (b) X‐ray structure of heteroleptic cage Pd2 LA 2 LC 2 with Pd–Pd distance. (c) ESI‐MS spectrum of cage Pd2 LA 2 LC 2. Minor peaks belong to traces of homoleptic [Pd2 LA 4]4+ (m/z = 517.1), [Pd2 LA 4 + BF4]3+ (719.1), and [Pd2 LC 2(CX3CN)4 + 2BF4]2+ (X = H, D; 610.1) and heteroleptic [Pd2 LA 3 LC + BF4]3+ (632.4).
Meanwhile, conducting the reaction of Pd(II) cations with ligands LA and LD in a 1:1:1 ratio resulted in a mixture, with some NMR signals corresponding to Pdn LA 2n (n = 1, 2) species, while others were distinct from those of free LD or Pdn LD 2n (n = 3, 4) assemblies. Further addition of LD and Pd(II) to the reaction mixture resulted in a cleaner NMR spectrum and a decrease in the intensity of the Pd2 LA 4/PdLA 2 signals. Based on these observations, we conducted the reaction using a stoichiometry of LA , LD , and Pd(II) in a 2:4:3 ratio, which allowed the exclusive formation of a single clean assembly (Scheme 1E and Figure 4a). Moreover, the same species can also be obtained through cage‐to‐cage transformation from the corresponding homoleptic cages Pdn LA 2n (n = 1, 2) and Pdn LD 2n (n = 3, 4). However, in this case, the assembly proceeds much more slowly and less cleanly as compared to the direct reaction of Pd(II) with the ligands (Figure S31).
FIGURE 4.

(a) 1H NMR spectra (CD3CN) of LA (top), LD (middle), and Pd3 LA 2 LD 4 (bottom) with the associated DOSY trace for Pd3 LA 2 LD 4; (b) fragment of 1H‐1H NOESY spectrum (CD3CN) of Pd3 LA 2 LD 4 with selected cross‐peaks confirming heteroleptic assembly and the depiction of corresponding inter‐ligand contacts from the DFT model; (c) ESI mass spectrum of Pd3 LA 2 LD 4; (d) DFT models of possible Pd3 LA 2 LC 4 isomers and calculated free energy differences, supporting the isosceles triangular ring structure.
Single crystals of this assembly were obtained; however, they exhibited very weak diffraction, even under synchrotron irradiation. Therefore, to elucidate the structure, we employed detailed NMR and MS analyses combined with DFT calculations (Figure 4). The 1H DOSY measurements confirmed that all NMR signals of the heteroleptic assembly belong to a single species with an r H of 12.8 Å (Figure 4a), while 1H–1H NOESY analysis revealed NOE interactions between the protons in the ortho positions of the isoquinoline and pyridine moieties of LA and LD (Figure 4b), further confirming the incorporation of both ligands into a single molecule. The ESI mass spectrum revealed the presence of [Pd3 LA 2 LD 4 + nBF4]6–n (n = 1–4) ions, definitively confirming a rare Pd3 LA 2 LD 4 heteroleptic stoichiometry (Figure 4c). For such a Pd3 LA 2 LD 4 composition, three isomers are conceivable [41, 46], exhibiting C2v , Cs , and C2 (with two enantiomers) symmetries (Figure 4d). The 1H NMR spectrum of Pd3 LA 2 LD 4 shows a single set of signals for the LA ligand and a twofold splitting of the LD signals, resulting in three sets of ligand signals with a 1:1:1 integral ratio (Figure 4a). This pattern is only consistent with the C2v ‐symmetric isosceles triangular topology, in which the symmetry is broken only for the LD ligands, as they are neighbored on one side by identical LD ligands and on the other side by LA ligands. In contrast, the other two isomers with Cs and C2 symmetries would produce at least a twofold splitting of the LA signals and a fourfold splitting of the LC signals, which was not observed. Geometrical analysis of the modelled isosceles triangular Pd3 LA 2 LD 4 structure indicates that the longest inter‐fragment distances are approximately 25–26 Å, consistent with the 1H DOSY measurements. Furthermore, DFT calculations indicate that this topology is the most energetically favorable structure among all Pd3 LA 2 LD 4 isomers (by –53 and –73 kJ/mol, respectively, Figure 4d) and also more stable than a binuclear heteroleptic cis‐Pd2 LA 2 LD 2 cage (by –8 kJ/mol for the transformation: 4 × Pd2 LA 2 LD 2 + Pd4 LD 8 ⇌ 4 × Pd3 LA 2 LD 4; Figure S59 and Table S7). Interestingly, for the LC ligand (which was previously employed by us to access a comparable isosceles triangular Pd3 A 2 B 4 topology [41]), the experimentally found binuclear heteroleptic Pd2 LA 2 LC 2 complex is significantly more stable than a tentative Pd3 LA 2 LC 4 species, both according to DFT calculations (by –188 kJ/mol for the 3 × Pd3 LA 2 LC 4 ⇌ 3 × Pd2 LA 2 LC 2 + Pd3 LC 6 transformation; Figure S59 and Table S7) and according to the experimental observations.
Remarkably, both heteroleptic cages Pd2 LA 2 LC 2 and Pd3 LA 2 LD 4 can be partially interconverted by adding the corresponding ligands LC or LD and additional Pd(II) (Scheme 1F). Addition of 2 eq. of LC to Pd3 LA 2 LD 4 induces partial conversion to Pd2 LA 2 LC 2, giving a 3:2 ratio of Pd3 LA 2 LD 4 to Pd2 LA 2 LC 2 after 18 h at RT, which is then not changing anymore over time (Figure S32). Conversely, adding 1 eq. of Pd(II) and 4 eq. of LD to Pd2 LA 2 LC 2 yields the same ratio, 3:2 in favor of Pd3 LA 2 LD 4 after 57 h at RT. In both cases, free ligands are released, and the formation of additional unidentified species is observed (Figure S33). Subsequent heating at 70°C does not noticeably promote the transformation further but instead decreases the concentration of both cages and results in the formation of more side products.
We also tested the formation of Pd(II) heteroleptic assemblies involving the quinoxaline‐backboned ligand LB and either LC or LD ; however, in these cases, only the corresponding homoleptic structures formed in narcissistic coexistence, as these ligands appear to be not shape‐complementary to each other.
2.3. Guest Binding and Guest‐Induced Transformations
Next, we examined the guest‐binding properties of the obtained assemblies using three disulfonate dianions of increasing size: ethane‐1,2‐disulfonate (G1), propane‐1,3‐disulfonate (G2), and benzene‐1,4‐disulfonate (G3) (Scheme 1G,H).
Binding of all three guests by the homoleptic Pd2 LA 4 cage in CD3CN proceeds in the slow exchange regime in a 1:1 ratio as indicated by NMR and ESI‐MS data (Figures S34–S38) as well as X‐ray crystallography for the host‐guest complexes with G2 and G3 (Figure 5 and Figure S57). Furthermore, the addition of guests results in the disappearance of any residual mononuclear PdLA 2 species. Similar to what was observed for assembly formation with different counteranions (BF4 – or OTf–), the NMR profiles of the host‐guest complexes were found to strongly depend on the nature of the anionic guest. Binding of the smallest disulfonate guest G1 resulted in broadened signals, thus indicating a highly dynamic structure. At the same time, binding of G2 and G3 resulted in mostly sharp NMR profiles with a noticeably down‐field shifted signal corresponding to the a protons of the isoquinoline moieties, indicating hydrogen bond formation between host and guest [71]. Guest displacement experiments indicate that G1 can be fully displaced by G2 or G3 in Pd2 LA 4, while the addition of G2 to G3@Pd2 LA 4 results in a 2:3 G2@Pd2 LA 4 to G3@Pd2 LA 4 equilibrium mixture (Figure S37). These results are also in line with DFT calculations, suggesting the following guest affinity row for Pd2 LA 4 cage: BF4 – < OTf– < G1 < G2 < G3 (Figure 5b), thus correlating with guest size [72]. Notably, the flexible and dynamic nature of LA allows the Pd2 LA 4 cage to expand and contract, adjusting its size depending on the encapsulated guest. This behavior is evidenced by Pd–Pd distance variations observed in the solid‐state structures of the host–guest complexes (13.0 Å for G2 vs. 13.6 Å for G3), as well as by corresponding differences in the 1H DOSY‐based hydrodynamic radii (9.6 Å for G2 vs 9.8 Å for G3, Table S5).
FIGURE 5.

(a) Partial 1H NMR spectra (CD3CN) of (BF4)2@[Pd2 LA 4], (OTf)2@[Pd2 LA 4], G1@[Pd2 LA 4], G2@[Pd2 LA 4], G3@[Pd2 LA 4]; (b) optimized DFT models and X‐ray structures of host‐guest complexes with various guests as well as theoretically calculated relative guest binding energies.
For the heteroleptic cages Pd2 LA 2 LC 2 and Pd3 LA 2 LD 4, binding of G1 proceeds rather in a fast‐exchange regime. Pd2 LA 2 LC 2 exhibits exo‐binding of G1 indicated by the shift of only outside pointing protons (Figure S39), which also agrees with the x‐ray data indicating a very small cavity, allowing only solvent molecules to fit inside (Figure S55). At the same time, Pd3 LA 2 LD 4, with a larger cavity, mainly shows shifting of inside pointing protons, indicating encapsulation (Figure S40). Both assemblies gradually precipitate upon the addition of more than 1 eq. of G1, preventing us from determining association constants. Precipitation was also observed when 1 eq. or more of G2 or G3 was added. In contrast, the addition of only 0.5 eq. of G2 or G3 did not induce immediate precipitation but led to the appearance of NMR signals corresponding to the G2/G3@Pd2 LA 4 host‐guest complexes (Figures S41 and S42). The strong affinity of homoleptic Pd2 LA 4 for G2 and G3, combined with the conformational flexibility of ligand LA , thus enables a retro‐cage‐to‐cage transformation in which the heteroleptic assemblies revert back to the homoleptic Pd2 LA 4 cage (Scheme 1H) [14]. As both heteroleptic cages contain only half of the homoleptic cage (corresponding to two LA units), 0.5 eq. of the guests should be the stoichiometrically required amount for this transformation. At room temperature, the retro‐cage‐to‐cage transformations are very slow. At 70°C, heating Pd2 LA 2 LC 2 with 0.5 eq. of guest for about 4 days results in approximately 50% conversion of Pd2 LA 2 LC 2 to G3@Pd2 LA 4 and about 40% conversion to G2@Pd2 LA 4, consistent with the established guest‐affinity trends of Pd2 LA 4. Further heating leads to gradual precipitation of both systems. For Pd3 LA 2 LD 4, formation of G2/G3@Pd2 LA 4 complexes is also observed by NMR and ESI‐MS (Figures S43–S45), but precipitation occurs even more rapidly upon heating, not allowing estimation of the conversion values for the retro‐cage‐to‐cage transformations in this case.
2.4. Post‐Assembly Condensation With 1,2‐Phenylenediamine
Finally, we examined the possibility of a post‐assembly transformation of the 1,2‐dicarbonyl backbone of LA within the homo‐ and heteroleptic assemblies by condensation with stoichiometric amounts of 1,2‐phenylenediamine (PDA) in CD3CN (Scheme 1I and J, Figures S46–S56).
The uncomplexed ligands LA and LB are only partially soluble in CD3CN (maximal concentration approximately 0.3 mM ≈ 10% of the concentration used in the self‐assemblies), but free LA still reacts with PDA at room temperature slightly faster than when it is part of an assembly within the initial reaction phase (Figure S56). For the homoleptic PdLA 2 and Pd2 LA 4 systems, we observed slight dependence of the reaction rate on the assembly nuclearity. At RT, during the early stages of the reaction, the freshly formed kinetic PdLA 2 species undergoes condensation with PDA producing PdLB 2 faster than the thermodynamic Pd2 LA 4 cage product (Figure S56). However, in parallel with the condensation, PdLA 2 also dimerizes to the Pd2 LA 4 cage structure (representing the faster process), and the PDA‐related reaction rates for the two assemblies converge as the conversion progresses.
During the reaction with PDA, minor release of free LA was observed by NMR in both cases (which was subsequently consumed), whereas no formation of free LB was detected. In addition, several transient intermediate or side species were detected by NMR, including a heteroleptic mononuclear PdLALB complex, a Pd(II)‐bound PDA species, and other unidentified species that were gradually consumed during the reaction (Figures S47 and S48). Analysis of the reaction mixture of Pd2 LA 4 with PDA by ESI‐MS revealed several possible reaction intermediates (Figure S49). Consistent with the NMR results, ions corresponding to the PdLALB complex were observed, whereas no signals corresponding to heteroleptic cage species Pd2 LA x LB y (x + y = 4) were detected. Moreover, signals corresponding to mono‐condensation products of PDA with the dicarbonyl group of LA coordinated to a single Pd(II) center were observed, namely [PdLA 2 + nPDA − nH2O]2+ (n = 1, 2) ions. In contrast, no signals corresponding to mono‐condensation products of the free ligand and cage structures were detected. These results indicate that the reaction most likely proceeds via the mononuclear Pd(II) assemblies, which exist in equilibrium with the cage structure and adopt a more suitable syn‐conformation concerning the central dicarbonyl moiety for condensation with PDA. Theoretical calculations further suggest that the formation of even a single LB ligand within the Pd2 LA 4 cage would introduce significant strain (Figure S60), substantially reducing cage stability and promoting its dissociation into mononuclear species, which were detected as reaction intermediates.
Interestingly, the counterion (or encapsulated guest) within the Pd2 LA 4 cage can significantly affect the condensation process. The formation rate of PdLB 2 decreases as the affinity of the Pd2 LA 4 cage for the anionic guest increases, following the sequence BF4 – > G1 > G2 > G3 (Figure S56). Noteworthy, for the G1@Pd2 LA 4 and G2@Pd2 LA 4 host‐guest complexes with aliphatic disulfonates, we observed slower reaction rates than for the Pd2 LA 4 host, but they were accompanied by a more pronounced release of free LA and minor formation of uncomplexed LB (Figures S50 and S51). In addition, signals corresponding to free Pd2 LA 4 host were detected after addition of PDA indicating the guest release during the reaction. In contrast, for the G3@Pd2 LA 4 complex, which exhibits the slowest condensation rate, only trace amounts of free LA were observed, with no formation of free LB or Pd2 LA 4 (Figure S52).
Heteroleptic cages Pd2 LA 2 LC 2 and Pd3 LA 2 LD 4 undergo condensation with PDA also noticeably slower than the homoleptic Pd2 LA 4 assembly. Notably, while Pd3 LA 2 LD 4 remains relatively stable in the presence of PDA but gradually reacts with it at RT, Pd2 LA 2 LC 2 shows almost no conversion to PdLB 2 at all. Instead, it predominantly undergoes disassembly, releasing free LC as well as the forming Pd(II) complex of PDA and some unidentified species as side products (Figures S53 and S54). Analysis of the reaction mixtures containing heteroleptic cages by ESI‐MS did not reveal any additional intermediate species beyond those observed for the reaction with the Pd2 LA 4 complex, indicating that in this case the reaction most likely also proceeds via a mononuclear complex. Under heating conditions (70°C in CD3CN), homoleptic cage Pd2 LA 4 (with BF4 – counterion) converts to PdLB 2 with approximately 80% yield within 24 h. At the same time, Pd3 LA 2 LD 4 yields only about 60% of PdLB 2 after 7 days under the same conditions (Figure S55). For Pd2 LA 2 LC 2, complete consumption of the starting cage is observed after 1.5 days at 70°C, but only approximately 30% PdLB 2 is formed, with the remainder undergoing decomposition pathways.
2.5. Ligand Conformation Analysis and Conformation Effects in Cages
A comprehensive analysis of the X‐ray and DFT‐optimized structures of all homo‐ and heteroleptic assemblies reveals several distinct conformational states that LA can adopt within Pd(II) supramolecular architectures (Figure 6 and Table S4). These conformations not only populate to fit the geometry of the assembly but also seem to define the reactivity of LA within the respective assembly as well as the overall assembly stability.
FIGURE 6.

Heat map of relaxed scan calculated energies of the dihedral angle between the phenyl rings of the benzil backbone (top); views of various conformations of ligand LA with corresponding dihedral angles between phenylene arms and ligand LB in PdLB 2 (middle); and calculated relative LA ligand strain energy [ΔE strain = E(LA assembly) − E(LA free)] in homo‐ and heteroleptic Pd(II) assemblies (bottom).
In the mononuclear trans‐PdLA 2 complex, LA adopts a folded conformation with a dihedral angle of approximately 70° between the phenylene arms. This geometry places the dicarbonyl moiety almost in a syn‐arrangement, which is optimal for condensation with PDA [59], yielding the trans‐PdL B 2 complex that retains a similar inter‐arm angle of 72°. This structural compatibility is consistent with the faster initial kinetics of PdLA 2 condensation relative to the Pd2 LA 4 cage. In contrast, the helical Pd2 LA 4 assembly (standing in equilibrium with PdLA 2 in solution) adopts a more open conformation, showing dihedral angles between ligand arms of approximately 126° in the DFT model and 124–141° in the solid‐state structures. While the dicarbonyl groups are oriented outward from the cage cavity, allowing sterically unhindered nucleophilic attack by PDA, this conformation departs more strongly from the syn‐arrangement than in the mononuclear chelate species. At the same time, the structure of Pd2 LA 4 is highly dynamic and can adjust its size and shape (e.g., upon guest binding, Figure 5b), which may enable LA to adopt a more favorable conformation for condensation with PDA. Notably, the reactivity of the cage can also be modulated by guest binding. Although the dicarbonyl dihedral angles in the solid‐state and calculated structures of the host‐guest complexes with G1–G3 do not significantly deviate from those of the free Pd2 LA 4 host (Table S4), our experiments show that the reactivity of Pd2 LA 4 can be reduced upon guest encapsulation (especially for G3). This is attributed to stabilization of the strongly twisted cage structure, which hinders the central dicarbonyl unit from adopting the syn‐conformation required for efficient conversion into the trans‐PdLB 2 complex.
The introduction of a second type of ligand can further alter the geometry of LA , increasing the dihedral angle between its arms to 210° in the optimized Pd2 LA 2 LC 2 structure (approximately 190° in the crystal) or further to 238° and 261° upon incorporation of LD in the optimized structure of the Pd3 LA 2 LD 4 assembly. Additionally, the symmetry‐breaking incorporation of a second type of ligand is expected to rigidify the assembly structure. In Pd3 LA 2 LD 4, the dicarbonyl groups adopt a similar arrangement to that in Pd2 LA 4, but they point toward the interior of the cage, where steric effects may hinder interaction with PDA and should further reduce the reactivity of the assembly.
In contrast, the Pd2 LA 2 LC 2 cage shows a dicarbonyl orientation closer to an anti‐conformation, which is the least favorable for condensation with PDA. Also, cage‐internal dynamics that would lead to twisting of the LA ligand backbones and move the carbonyl substituents into syn‐arrangement for at least brief periods is not likely to occur in this non‐helical, low‐symmetry assembly. We thus hypothesize that this geometric arrangement accounts for the lower reactivity of this cage and explains why it mainly undergoes decomposition through Pd(II) abstraction by PDA, accompanied by release of the free ligands.
Considering the conformational landscape of the diphenyl‐1,2‐dicarbonyl backbone (evaluated by a relaxed scan of the dihedral angle between the phenyl arms; Figure 6), the conformations adopted by LA in the optimized structures of the free LA ligand (dihedral angle of 253°) and of Pd3 LA 2 LD 4 cage (dihedral angles of 238° and 261°) lie closest to the calculated energy minima. In terms of ligand strain, the extended conformations in Pd3 LA 2 LD 4 are less favorable than the free ligand by approximately 7.2–8.9 kJ/mol, likely due to additional strain in the isoquinoline and phenylene moieties arising from adjustment of the ligand geometry to the cage structure. The open conformation in Pd2 LA 4 is, as expected, even more strained relative to the free ligand, with a strain energy of 12.9 kJ/mol. Even higher strain energies are observed for LA in PdLA 2 and Pd2 LA 2 LC 2, with ΔE strain values of 14.5 and 16.7 kJ/mol, respectively (Figure 6 and Table S7). In both cases, the LA ligand experiences steric repulsion, either between the phenylene moieties in the folded PdLA 2 structure or between the phenylene and carbonyl moieties in the Pd2 LA 2 LC 2 cage. The calculation results are consistent with experimental observations: the lower ligand strain in Pd2 LA 4 correlates with its higher stability and greater population relative to PdLA 2 in solution. Among the heteroleptic cages, the less strained Pd3 LA 2 LD 4 species displays, accordingly, a higher stability than the Pd2 LA 2 LC 2 cage.
3. Conclusion
In conclusion, we have shown that the intrinsic flexibility and reactivity of the benzil‐based ligand LA enable it to adopt multiple conformations, giving rise to various homoleptic and heteroleptic Pd(II) assemblies with different nuclearities, topologies, and stimuli‐responsive transformation pathways. In homoleptic systems, LA can take folded or open conformations, forming the mononuclear complex PdLA 2 and the Pd2 LA 4 helicate, which—depending on the solvent—can coexist in solution. Incorporation of additional diverging ligands, LC or LD , bearing rigid fluorenone or phenanthrene backbones (either through direct Pd(II)‐mediated self‐assembly or via cage‐to‐cage transformation), induces dihedral angle expansion of LA upon formation of the classic heteroleptic cis‐Pd2 LA 2 LC 2 assembly or the rare, low‐symmetry isosceles triangular Pd3 LA 2 LD 4 ring topology. Both heteroleptic cages can be partially interconverted through addition of the corresponding ligands and Pd(II) ions, and both can also undergo partial guest‐induced retro‐cage‐to‐cage transformation back to homoleptic species. Furthermore, LA can be conformationally locked by condensation of its 1,2‐dicarbonyl backbone with 1,2‐phenylenediamine, generating the pincer‐shaped and rigid quinoxaline‐backboned ligand LB , which forms only the trans‐PdLB 2 chelate complex in the presence of Pd(II). Notably, this transformation can be performed in a post‐assembly manner, using either homoleptic or heteroleptic LA ‐based architectures as precursors. We could show that the conformation adopted by LA within the different supramolecular assemblies strongly influences both its reactivity toward condensation with 1,2‐phenylenediamine and the overall assembly stability. In homoleptic cages, guest binding can further modulate the reactivity of LA within the assembly, with higher‐affinity guests exhibiting a stronger inhibitory effect on the reaction.
Overall, our work demonstrates how ligand flexibility and reactivity can be strategically harnessed to control supramolecular geometry, nuclearity, and topology in Pd(II) coordination assemblies, enabling multiple interconversions between different states using ligand, guest, and reagent as stimuli. Observed relationships between ligand conformation, assembly structure, guest presence, and the course of the studied chemical transformations also highlight how hierarchical supramolecular organization can modulate functional group reactivity. Such control over the conformation and accessibility of specific reactive sites can be compared to how nature steers the reactivity of chemical functionality through nanoscopic confinement, conformational restriction, and allosteric regulation. Furthermore, the obtained Pd(II)‐based assembly system can serve as a versatile platform for the design of adaptive, chemically addressable metallosupramolecular architectures, with potential applications in information transduction and processing, selective guest recognition, and stimuli‐responsive catalysis under confinement.
Author Contributions
G.H.C. and A.S.M. conceived and designed the project. Z.Z., A.P.G., M.A. performed synthetic work, compound characterization, experiments and analysed the data. A.S.M. performed further experiments, theoretical and crystallographic studies, analysed the data, and prepared the initial manuscript draft. B.S. performed DOSY and VT NMR measurements, and processed crystallographic data. G.H.C. and A.S.M. supervised the work. All authors discussed the results, and the manuscript was written through contributions of all authors.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Supporting File 1: The authors have cited additional references within the Supporting Information [73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89].
Supporting File 2: anie73516‐sup‐0002‐SuppMat.zip.
Acknowledgments
The authors thank the Deutsche Forschungsgemeinschaft (DFG) for support through GRK2376 (“Confinement‐controlled Chemistry,” project number 331085229) and Germany's Excellence Strategy EXC2033, project number 390677874 (“RESOLV”). This work was funded by the European Union under the Horizon Europe Marie Skłodowska‐Curie Postdoctoral Fellowship Programme, project CageComm, Grant Agreement 101208423. Zhiwei Zeng acknowledges the China Scholarship Council (CSC) for visiting PhD funding. A. Priscila Gia is grateful to MCIN for the FPI grant PRE2021‐099665. Diffraction data of Pd2 LA 4, G2@Pd2 LA 4, G3@Pd2 LA 4, and Pd2 LA 2 LC 2 [90] was collected at PETRA III, DESY (Hamburg, Germany), a member of the Helmholtz Association (HGF). The authors thank Dr. G. Pompidor and Dr. S. Chatziefthymiou for assistance at synchrotron beamline P11 (R‐20240690) and Dr. J. J. Holstein for help with the collection and processing of x‐ray data.
Open access funding enabled and organized by Projekt DEAL.
Zeng Z., Gia A. P., Mikherdov A. S., Acar M., Schmidt B., and and Clever G. H., Angewandte Chemie International Edition (65, no. 36 (2026): e9050553, 10.1002/anie.9050553
Dedicated to Prof. Dr. Bernhard Lippert on the occasion of his 80th birthday.
Contributor Information
Alexander S. Mikherdov, Email: aleksandr.mikherdov@tu-dortmund.de.
Guido H. Clever, Email: guido.clever@tu-dortmund.de.
Data Availability Statement
The data that support the findings of this study will be openly available in the RESOLVdata repository in machine‐readable form at https://data.tu‐dortmund.de/dataverse/resolv under the link https://doi.org/10.17877/RESOLV‐2026‐TBVPLR.
References
- 1. Hagerman P. J., “Flexibility of RNA,” Annual Review of Biophysics and Biomolecular Structure 26 (1997): 139–156, 10.1146/annurev.biophys.26.1.139. [DOI] [PubMed] [Google Scholar]
- 2. Teilum K., Olsen J. G., and Kragelund B. B., “Functional Aspects of Protein Flexibility,” Cellular and Molecular Life Sciences 66 (2009): 2231–2247, 10.1007/s00018-009-0014-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Nussinov R., “Introduction to Protein Ensembles and Allostery,” Chemical Reviews 116 (2016): 6263–6266, 10.1021/acs.chemrev.6b00283. [DOI] [PubMed] [Google Scholar]
- 4. Marin‐Gonzalez A., Vilhena J. G., Perez R., and Moreno‐Herrero F., “A Molecular View of DNA Flexibility,” Quarterly Reviews of Biophysics 54 (2021): e8, 10.1017/s0033583521000068. [DOI] [PubMed] [Google Scholar]
- 5. Crowley J. D., Lisboa L. S., and van Hilst Q. V. C., “7.06 ‐ Supramolecular Systems: Metallo‐Molecular Machines and Stimuli Responsive Metallo‐Macrocycles and Cages,” in Reference Module in Chemistry, Molecular Sciences and Chemical Engineering, edited by Reedijk J. (Elsevier, 2021), 174–205, 10.1016/b978-0-08-102688-5.00042-8. [DOI] [Google Scholar]
- 6. Pearcy A. C. and Crowley J. D., “The Biologically Inspired Abilities of Metallosupramolecular Architectures,” Chemistry—A European Journal 29 (2023): e202203752, 10.1002/chem.202203752. [DOI] [PubMed] [Google Scholar]
- 7. Yang Y., Du Y., Heard A. W., and Nitschke J. R., “Allosteric Regulation in Metal–Organic Cages,” Nature Synthesis 4 (2025): 537–551, 10.1038/s44160-025-00799-3. [DOI] [Google Scholar]
- 8. Pan M., Wu K., Zhang J.‐H., and Su C.‐Y., “Chiral Metal–Organic Cages/Containers (MOCs): From Structural and Stereochemical Design to Applications,” Coordination Chemistry Reviews 378 (2017): 333–349, 10.1016/j.ccr.2017.10.031. [DOI] [Google Scholar]
- 9. Ling Q.‐H., Lou Z.‐C., Zhang L., et al., “Supramolecular Cage‐Mediated Cargo Transport,” Chemical Society Reviews 53 (2024): 6042–6067, 10.1039/d3cs01081c. [DOI] [PubMed] [Google Scholar]
- 10. Fang Y., Powell J. A., Li E., et al., “Catalytic Reactions Within the Cavity of Coordination Cages,” Chemical Society Reviews 48 (2019): 4707–4730, 10.1039/c9cs00091g. [DOI] [PubMed] [Google Scholar]
- 11. Wiester M. J., Ulmann P. A., and Mirkin C. A., “Enzyme Mimics Based Upon Supramolecular Coordination Chemistry,” Angewandte Chemie International Edition 50 (2011): 114–137, 10.1002/anie.201000380. [DOI] [PubMed] [Google Scholar]
- 12. Benchimol E., Nguyen B.‐N. T., Ronson T. K., and Nitschke J. R., “Transformation Networks of Metal–Organic Cages Controlled by Chemical Stimuli,” Chemical Society Reviews 51 (2022): 5101–5135, 10.1039/d0cs00801j. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Neukirch L. and Clever G. H., “Topological Variety and Self‐Sorting in Homo‐ and Heteroleptic PdnL2n Metallo‐Supramolecular Assemblies,” Chemical Science 16 (2025): 12242–12276, 10.1039/d5sc03203b. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Ebbert K. E., Sendzik F., Neukirch L., et al., “Solvent‐Controlled Separation of Integratively Self‐Sorted Pd2LA 2LB 2 Coordination Cages,” Angewandte Chemie International Edition 64 (2025): e202416076, 10.1002/anie.202416076. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Suzuki K., Kawano M., and Fujita M., “Solvato‐Controlled Assembly of Pd3L6 and Pd4L8 Coordination ‘Boxes’,” Angewandte Chemie International Edition 46 (2007): 2819–2822, 10.1002/anie.200605084. [DOI] [PubMed] [Google Scholar]
- 16. Algar J. L., Findlay J. A., Evans J. D., and Preston D., “A Switchable Palladium(II) Trefoil Entangled Tetrahedron With Temperature Dependence and Concentration Independence,” Angewandte Chemie International Edition 61 (2022): e202210476, 10.1002/anie.202210476. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Hu C., Chen D. W., Sudan S., and Severin K., “Palladium‐Based Coordination Cages as Dynamic Crosslinks in Acrylamide Hydrogels,” Chemical Science 16 (2025): 5559–5564, 10.1039/d5sc00335k. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Sun Q.‐F., Sato S., and Fujita M., “An M18L24 Stellated Cuboctahedron Through Post‐Stellation of an M12L24 Core,” Nature Chemistry 4 (2012): 330–333, 10.1038/nchem.1285. [DOI] [PubMed] [Google Scholar]
- 19. Han M., Engelhard D. M., and Clever G. H., “Self‐Assembled Coordination Cages Based on Banana‐Shaped Ligands,” Chemical Society Reviews 43 (2014): 1848–1860, 10.1039/c3cs60473j. [DOI] [PubMed] [Google Scholar]
- 20. Saha S., Regeni I., and Clever G. H., “Structure Relationships Between Bis‐Monodentate Ligands and Coordination Driven Self‐Assemblies,” Coordination Chemistry Reviews 374 (2018): 1–14, 10.1016/j.ccr.2018.06.010. [DOI] [Google Scholar]
- 21. Parbin M., Sivalingam V., Venkatachalam R., and Chand D. K., “Conformational Adaptability Enabled Higher‐Order Self‐Sorting Processes in Coordination Cages,” Chemical Science 17 (2025): 2535–2546, 10.1039/d5sc05568g. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Benchimol E., Ebbert K. E., Walther A., Holstein J. J., and Clever G. H., “Ligand Conformation Controls Assembly of a Helicate/Mesocate, Heteroleptic [Pd2L2L′2] Cages and a Six‐Jagged [Pd6L12] Ring,” Chemistry—A European Journal 30 (2024): e202401850, 10.1002/chem.202401850. [DOI] [PubMed] [Google Scholar]
- 23. Parbin M., Sivalingam V., and Chand D. K., “Highly Anisotropic Pd2Lab 2Lcc 2 and Pd2Lab 2Lcd 2 Type Cages by Heteromeric Completive Self‐Sorting,” Angewandte Chemie International Edition 63 (2024): e202410219, 10.1002/anie.202410219. [DOI] [PubMed] [Google Scholar]
- 24. Li S., Liu C., Chen Q., et al., “Adaptive Coordination Assemblies Based on a Flexible Tetraazacyclododecane Ligand for Promoting Carbon Dioxide Fixation,” Chemical Science 13 (2022): 9016–9022, 10.1039/d2sc03093d. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Kai S., Maddala S. P., Kojima T., et al., “Flexibility of Components Alters the Self‐Assembly Pathway of Pd2L4 Coordination Cages,” Dalton Transactions 47 (2018): 3258–3263, 10.1039/c8dt00112j. [DOI] [PubMed] [Google Scholar]
- 26. Kwan C.‐S., Walther A., Mikherdov A. S., Holstein J. J., Drechsler C., and Clever G. H., “Dual Cation/Anion Binding in Crown Ether‐Based Coordination Cages,” Journal of the American Chemical Society 148 (2026): 3230–3239, 10.1021/jacs.5c17231. [DOI] [PubMed] [Google Scholar]
- 27. Díaz A. E. M. and Lewis J. E. M., “Structural Flexibility in Metal‐Organic Cages,” Frontiers in Chemistry 9 (2021): 706462, 10.3389/fchem.2021.706462. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Tateishi T., Yoshimura M., Tokuda S., Matsuda F., Fujita D., and Furukawa S., “Coordination/Metal–Organic Cages Inside Out,” Coordination Chemistry Reviews 467 (2022): 214612, 10.1016/j.ccr.2022.214612. [DOI] [Google Scholar]
- 29. Chattopadhyay S., Durník R., Kiesilä A., et al., “Flexibility‐Aided Orientational Self‐Sorting and Transformations of Bioactive Homochiral Cuboctahedron Pd12L16,” Angewandte Chemie International Edition 64 (2025): e202513902, 10.1002/anie.202513902. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Zhang T., Zhou L.‐P., Guo X.‐Q., Cai L.‐X., and Sun Q.‐F., “Adaptive Self‐Assembly and Induced‐Fit Transformations of Anion‐Binding Metal‐Organic Macrocycles,” Nature Communications 8 (2017): ncomms15898, 10.1038/ncomms15898. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Hema K., Grommet A. B., Białek M. J., et al., “Guest Encapsulation Alters the Thermodynamic Landscape of a Coordination Host,” Journal of the American Chemical Society 145 (2023): 24755–24764, 10.1021/jacs.3c08666. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Montes‐Tolentino P., Mikherdov A. S., Drechsler C., Holstein J. J., and Clever G. H., “Control of Interlocking Mode in Pd4L8 Cage Catenanes,” Angewandte Chemie International Edition 64 (2025): e202423810, 10.1002/anie.202423810. [DOI] [PubMed] [Google Scholar]
- 33. Sudan S., Fadaei‐Tirani F., Scopelliti R., Ebbert K. E., Clever G. H., and Severin K., “LiBF4‐Induced Rearrangement and Desymmetrization of a Palladium‐Ligand Assembly,” Angewandte Chemie International Edition 61 (2022): e202201823, 10.1002/anie.202201823. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Percástegui E. G., “Guest‐Induced Transformations in Metal‐Organic Cages,” European Journal of Inorganic Chemistry 2021 (2021): 4425–4438, 10.1002/ejic.202100657. [DOI] [Google Scholar]
- 35. Walther A., Regeni I., Holstein J. J., and Clever G. H., “Guest‐Induced Reversible Transformation Between an Azulene‐Based Pd2L4 Lantern‐Shaped Cage and a Pd4L8 Tetrahedron,” Journal of the American Chemical Society 145 (2023): 25365–25371, 10.1021/jacs.3c09295. [DOI] [PubMed] [Google Scholar]
- 36. Sekiya R., Fukuda M., and Kuroda R., “Anion‐Directed Formation and Degradation of an Interlocked Metallohelicate,” Journal of the American Chemical Society 134 (2012): 10987–10997, 10.1021/ja303634u. [DOI] [PubMed] [Google Scholar]
- 37. Bloch W. M. and Clever G. H., “Integrative Self‐Sorting of Coordination Cages Based on ‘Naked’ Metal Ions,” Chemical Communications 53 (2017): 8506–8516, 10.1039/c7cc03379f. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Prusty S., Yazaki K., Yoshizawa M., and Chand D. K., “A Truncated Molecular Star,” Chemistry—A European Journal 23 (2017): 12456–12461, 10.1002/chem.201702264. [DOI] [PubMed] [Google Scholar]
- 39. Tessarolo J., Lee H., Sakuda E., Umakoshi K., and Clever G. H., “Integrative Assembly of Heteroleptic Tetrahedra Controlled by Backbone Steric Bulk,” Journal of the American Chemical Society 143 (2021): 6339–6344, 10.1021/jacs.1c01931. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Ebbert K. E., Benchimol E., Platzek A., et al., “Ring‐Size Control and Guest‐Induced Circularly Polarized Luminescence in Heteroleptic Pd3A3B3 and Pd4A4B4 Assemblies,” Angewandte Chemie International Edition 63 (2024): e202413323, 10.1002/anie.202413323. [DOI] [PubMed] [Google Scholar]
- 41. Benchimol E., Tessarolo J., Tarzia A., et al., “Adjacent Backbone Interactions Control Self‐Sorting of Chiral Heteroleptic Pd3A2B4 Isosceles Triangles and Pd4A4C4 Pseudo‐Tetrahedra,” Chemistry 12 (2026): 102780, 10.1016/j.chempr.2025.102780. [DOI] [Google Scholar]
- 42. Wu K., Benchimol E., Baksi A., and Clever G. H., “Non‐Statistical Assembly of Multicomponent [Pd2ABCD] Cages,” Nature Chemistry 16 (2024): 584–591, 10.1038/s41557-023-01415-7. [DOI] [PubMed] [Google Scholar]
- 43. Abe T., Sanada N., Takeuchi K., Okazawa A., and Hiraoka S., “Assembly of Six Types of Heteroleptic Pd2L4 Cages Under Kinetic Control,” Journal of the American Chemical Society 145 (2023): 28061–28074, 10.1021/jacs.3c09359. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Benchimol E., Regeni I., Zhang B., Kabiri M., Holstein J. J., and Clever G. H., “Heteromeric Completive Self‐Sorting in Coordination Cage Systems,” Journal of the American Chemical Society 146 (2024): 6905–6911, 10.1021/jacs.3c14168. [DOI] [PubMed] [Google Scholar]
- 45. Yu S., Shi J.‐Y., Chen H., et al., “A Rational Desymmetrization Approach for Enantioselective Synthesis of Inherently Chiral C 1‐Symmetric Metal–Organic Cages,” Journal of the American Chemical Society 148 (2026): 7300–7311, 10.1021/jacs.5c19421. [DOI] [PubMed] [Google Scholar]
- 46. Wu K., Zhang B., Drechsler C., Holstein J. J., and Clever G. H., “Backbone‐Bridging Promotes Diversity in Heteroleptic Cages,” Angewandte Chemie International Edition 60 (2020): 6403–6407, 10.1002/anie.202012425. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Benchimol E., Tessarolo J., and Clever G. H., “Photoswitchable Coordination Cages,” Nature Chemistry 16 (2024): 13–21, 10.1038/s41557-023-01387-8. [DOI] [PubMed] [Google Scholar]
- 48. Tayier F. and Furukawa S., “Redox‐Active Metal‐Organic Cages for Multielectron Redox Processes,” preprint, ChemRxiv, April 16, 2026, 10.26434/chemrxiv.15002065/v1. [DOI]
- 49. Roberts D. A., Pilgrim B. S., and Nitschke J. R., “Covalent Post‐Assembly Modification in Metallosupramolecular Chemistry,” Chemical Society Reviews 47 (2017): 626–644, 10.1039/c6cs00907g. [DOI] [PubMed] [Google Scholar]
- 50. DiNardi R. G., Douglas A. O., Tian R., et al., “Visible‐Light‐Responsive Self‐Assembled Complexes: Improved Photoswitching Properties by Metal Ion Coordination,” Angewandte Chemie International Edition 61 (2022): e202205701, 10.1002/anie.202205701. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51. DiNardi R. G., Rasheed S., Capomolla S. S., et al., “Photoswitchable Catalysis by a Self‐Assembled Molecular Cage,” Journal of the American Chemical Society 146 (2024): 21196–21202, 10.1021/jacs.4c04846. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52. Lara L. P., Bädorf B., Notheis M. J., Schnakenburg G., Grimme S., and von Krbek L. K. S., “White‐Light Powered Autonomous Molecular Ratchet Drives PdII Capsules Out of Equilibrium,” Chemical Science 17 (2026): 8959–8967, 10.1039/d5sc09472k. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53. Han M., Luo Y., Damaschke B., et al., “Light‐Controlled Interconversion Between a Self‐Assembled Triangle and a Rhombicuboctahedral Sphere,” Angewandte Chemie International Edition 55 (2016): 445–449, 10.1002/anie.201508307. [DOI] [PubMed] [Google Scholar]
- 54. Li R.‐J., Tessarolo J., Lee H., and Clever G. H., “Multi‐Stimuli Control Over Assembly and Guest Binding in Metallo‐Supramolecular Hosts Based on Dithienylethene Photoswitches,” Journal of the American Chemical Society 143 (2021): 3865–3873, 10.1021/jacs.0c12188. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55. Zeid J. B., Nicolas J., Dekhtiarenko M., et al., “Programming Palladium Cage Geometry Through Ligand Redox Modulation,” Angewandte Chemie International Edition 65 (2026): e5709785, 10.1002/anie.5709785. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 56. Roberts D. A., Pilgrim B. S., Sirvinskaite G., Ronson T. K., and Nitschke J. R., “Covalent Post‐Assembly Modification Triggers Multiple Structural Transformations of a Tetrazine‐Edged Fe4L6 Tetrahedron,” Journal of the American Chemical Society 140 (2018): 9616–9623, 10.1021/jacs.8b05082. [DOI] [PubMed] [Google Scholar]
- 57. Pilgrim B. S., Roberts D. A., Lohr T. G., Ronson T. K., and Nitschke J. R., “Signal Transduction in a Covalent Post‐Assembly Modification Cascade,” Nature Chemistry 9 (2017): 1276–1281, 10.1038/nchem.2839. [DOI] [Google Scholar]
- 58. Black M. R., Bhattacharyya S., Argent S. P., and Pilgrim B. S., “Structural Transformations of Metal–Organic Cages Through Tetrazine‐Alkene Reactivity,” Journal of the American Chemical Society 146 (2024): 28233–28241, 10.1021/jacs.4c08591. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 59. Sato N., “8.03 ‐ Pyrazines and Their Benzo Derivatives,” in Comprehensive Heterocyclic Chemistry III, edited by Katritzky A. R., Ramsden C. A., and Taylor R. J. K. (Elsevier, 2008): 273–331, 10.1016/b978-008044992-0.00703-3. [DOI] [Google Scholar]
- 60. August D. P., Nichol G. S., and Lusby P. J., “Maximizing Coordination Capsule–Guest Polar Interactions in Apolar Solvents Reveals Significant Binding,” Angewandte Chemie International Edition 55 (2016): 15022–15026, 10.1002/anie.201608229. [DOI] [PubMed] [Google Scholar]
- 61. Bloch W. M., Horiuchi S., Holstein J. J., et al., “Maximized Axial Helicity in a Pd2L4 Cage: Inverse Guest Size‐Dependent Compression and Mesocate Isomerism,” Chemical Science 14 (2023): 1524–1531, 10.1039/d2sc06629g. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 62. Grimme S., Hansen A., Ehlert S., and Mewes J.‐M., “r2SCAN‐3c: A ‘Swiss Army Knife’ Composite Electronic‐Structure Method,” Journal of Chemical Physics 154 (2021): 064103, 10.1063/5.0040021. [DOI] [PubMed] [Google Scholar]
- 63. Mennucci B., “Polarizable Continuum Model,” Wiley Interdisciplinary Reviews‐Computational Molecular Science 2 (2012): 386–404, 10.1002/wcms.1086. [DOI] [Google Scholar]
- 64. Chai J.‐D. and Head‐Gordon M., “Systematic Optimization of Long‐Range Corrected Hybrid Density Functionals,” Journal of Chemical Physics 128 (2008): 084106, 10.1063/1.2834918. [DOI] [PubMed] [Google Scholar]
- 65. Caldeweyher E., Bannwarth C., and Grimme S., “Extension of the D3 Dispersion Coefficient Model,” Journal of Chemical Physics 147 (2017): 034112, 10.1063/1.4993215. [DOI] [PubMed] [Google Scholar]
- 66. Weigend F. and Ahlrichs R., “Balanced Basis Sets of Split Valence, Triple Zeta Valence and Quadruple Zeta Valence Quality for H to Rn: Design and Assessment of Accuracy,” Physical Chemistry Chemical Physics 7 (2005): 3297–3305, 10.1039/b508541a. [DOI] [PubMed] [Google Scholar]
- 67. Marenich A. V., Cramer C. J., and Truhlar D. G., “Universal Solvation Model Based on Solute Electron Density and on a Continuum Model of the Solvent Defined by the Bulk Dielectric Constant and Atomic Surface Tensions,” Journal of Physical Chemistry B 113 (2009): 6378–6396, 10.1021/jp810292n. [DOI] [PubMed] [Google Scholar]
- 68. Plett C., Stahn M., Bursch M., Mewes J.‐M., and Grimme S., “Improving Quantum Chemical Solvation Models by Dynamic Radii Adjustment for Continuum Solvation (DRACO),” Journal of Physical Chemistry Letters 15 (2024): 2462–2469, 10.1021/acs.jpclett.3c03551. [DOI] [PubMed] [Google Scholar]
- 69. Spicher S. and Grimme S., “Efficient Computation of Free Energy Contributions for Association Reactions of Large Molecules,” Journal of Physical Chemistry Letters 11 (2020): 6606–6611, 10.1021/acs.jpclett.0c01930. [DOI] [PubMed] [Google Scholar]
- 70. Platzek A., Juber S., Yurtseven C., et al., “Endohedrally Functionalized Heteroleptic Coordination Cages for Phosphate Ester Binding,” Angewandte Chemie International Edition 61 (2022): e202209305, 10.1002/anie.202209305. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 71. Martí‐Centelles V., Lawrence A. L., and Lusby P. J., “High Activity and Efficient Turnover by a Simple, Self‐Assembled ‘Artificial Diels–Alderase’,” Journal of the American Chemical Society 140 (2018): 2862–2868, 10.1021/jacs.7b12146. [DOI] [PubMed] [Google Scholar]
- 72. Clever G. H., Kawamura W., and Shionoya M., “Encapsulation Versus Aggregation of Metal–Organic Cages Controlled by Guest Size Variation,” Inorganic Chemistry 50 (2011): 4689–4691, 10.1021/ic200517r. [DOI] [PubMed] [Google Scholar]
- 73. Bloch W. M., Abe Y., Holstein J. J., Wandtke C. M., Dittrich B., and Clever G. H., “Geometric Complementarity in Assembly and Guest Recognition of a Bent Heteroleptic cis‐[Pd2LA 2LB 2] Coordination Cage,” Journal of the American Chemical Society 138 (2016): 13750–13755, 10.1021/jacs.6b08694. [DOI] [PubMed] [Google Scholar]
- 74. Jerschow A. and Müller N., “3D Diffusion‐Ordered TOCSY for Slowly Diffusing Molecules,” Journal of Magnetic Resonance, Series A 123 (1996): 222–225, 10.1006/jmra.1996.0241. [DOI] [Google Scholar]
- 75. Jerschow A. and Müller N., “Suppression of Convection Artifacts in Stimulated‐Echo Diffusion Experiments. Double‐Stimulated‐Echo Experiments,” Journal of Magnetic Resonance 125 (1997): 372–375, 10.1006/jmre.1997.1123. [DOI] [Google Scholar]
- 76. Stejskal E. O. and Tanner J. E., “Spin Diffusion Measurements: Spin Echoes in the Presence of a Time‐Dependent Field Gradient,” Journal of Chemical Physics 42 (1965): 288–292, 10.1063/1.1695690. [DOI] [Google Scholar]
- 77. Tanner J. E. and Stejskal E. O., “Restricted Self‐Diffusion of Protons in Colloidal Systems by the Pulsed‐Gradient, Spin‐Echo Method,” Journal of Chemical Physics 49 (1968): 1768–1777, 10.1063/1.1670306. [DOI] [Google Scholar]
- 78. Vonrhein C., Flensburg C., Keller P., et al., “Data Processing and Analysis With the autoPROC Toolbox,” Acta Crystallographica Section D: Biological Crystallography 67 (2011): 293–302, 10.1107/s0907444911007773. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 79. Kabsch W., “XDS,” Acta Crystallographica Section D: Biological Crystallography 66 (2010): 125–132, 10.1107/s0907444909047337. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 80. Sheldrick G. M., “SHELXT – Integrated Space‐Group and Crystal‐Structure Determination,” Acta Crystallographica Section A: Foundations and Advances 71 (2015): 3–8, 10.1107/s2053273314026370. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 81. Sheldrick G. M., “A Short History of SHELX,” Acta Crystallographica Section A: Foundations of Crystallography 64 (2008): 112–122, 10.1107/s0108767307043930. [DOI] [PubMed] [Google Scholar]
- 82. Bourhis L. J., Dolomanov O. V., Gildea R. J., Howard J. A. K., and Puschmann H., “The Anatomy of a Comprehensive Constrained, Restrained Refinement Program for the Modern Computing Environment – Olex2 Dissected,” Acta Crystallographica Section A: Foundations and Advances 71 (2015): 59–75, 10.1107/s2053273314022207. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 83. Dolomanov O. V., Bourhis L. J., Gildea R. J., Howard J. A. K., and Puschmann H., “OLEX2: A Complete Structure Solution, Refinement and Analysis Program,” Journal of Applied Crystallography 42 (2009): 339–341, 10.1107/s0021889808042726. [DOI] [Google Scholar]
- 84. Neese F., “Software Update: The ORCA Program System—Version 5,” Wiley Interdisciplinary Reviews: Computational Molecular Science 12 (2022): e1606, 10.1002/wcms.1606. [DOI] [Google Scholar]
- 85. Chemcraft , Graphical Software for Visualization of Quantum Chemistry Computations, Chemcraft, Version 1.8, Build 682. https://www.chemcraftprog.com. [Google Scholar]
- 86. Bannwarth C., Ehlert S., and Grimme S., “GFN2‐xTB—An Accurate and Broadly Parametrized Self‐Consistent Tight‐Binding Quantum Chemical Method With Multipole Electrostatics and Density‐Dependent Dispersion Contributions,” Journal of Chemical Theory and Computation 15 (2019): 1652–1671, 10.1021/acs.jctc.8b01176. [DOI] [PubMed] [Google Scholar]
- 87. Ehlert S., Stahn M., Spicher S., and Grimme S., “Robust and Efficient Implicit Solvation Model for Fast Semiempirical Methods,” Journal of Chemical Theory and Computation 17 (2021): 4250–4261, 10.1021/acs.jctc.1c00471. [DOI] [PubMed] [Google Scholar]
- 88. Bannwarth C., Caldeweyher E., Ehlert S., et al., “Extended Tight‐Binding Quantum Chemistry Methods,” Wiley Interdisciplinary Reviews: Computational Molecular Science 11 (2021): e1493, 10.1002/wcms.1493. [DOI] [Google Scholar]
- 89. de Souza B., “GOAT: A Global Optimization Algorithm for Molecules and Atomic Clusters,” Angewandte Chemie International Edition 64 (2025): e202500393, 10.1002/anie.202500393. [DOI] [PubMed] [Google Scholar]
- 90. Deposition Numbers 2544681–2544685 Contain the Supplementary Crystallographic Data for This Paper. These Data Are Provided Free of Charge by the Joint Cambridge Crystallographic Data Centre and Fachinformationszentrum Karlsruhe Access Structures Service.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supporting File 1: The authors have cited additional references within the Supporting Information [73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89].
Supporting File 2: anie73516‐sup‐0002‐SuppMat.zip.
Data Availability Statement
The data that support the findings of this study will be openly available in the RESOLVdata repository in machine‐readable form at https://data.tu‐dortmund.de/dataverse/resolv under the link https://doi.org/10.17877/RESOLV‐2026‐TBVPLR.
