Synchronizing Two Distinct Nano‐Circular Sliding Motions in Six‐Component Machinery for Double Catalysis

Abstract

The heteroleptic multi‐component double slider‐on‐deck system DS3 exhibits tight coupling of motional speed of two distinct nano‐circular sliders (k ₂₉₈=77 and 41 kHz) despite a 2.2 nm separation. In comparison, the single sliders in DS1 and DS2 move at vastly different speed (k ₂₉₈=1.1 vs. 350 kHz). Synchronization of the motions in DS3 remains even when one slows the movement of the faster slider using small molecular brake pads. In contrast to the individual DS1 and DS2 systems, DS3 is a powerful catalyst for a two‐step reaction by using the motion of both sliders to drive two catalytic processes.

The speed synchronization of two slider‐on‐deck systems, each activating a catalytic unit through circular motion, leads to synergistic effects in the catalytic turnover of a double catalysis.

Synchronized motion in machines plays a ubiquitous role in nature to maintain life ^[1] by acting as energy transducer, ^[2] e.g. in the synthesis of ATP as a high‐energy product and fuel by the multi‐component F_oF₁‐ATP synthase. Via efficient intercomponent communication, conformational changes in the F_o subunit lead to cyclic structural changes in the F₁ unit which mediate a phase‐shifted conversion of ADP and P_i to ATP. ^[3]

While motional speed in artificial machines has been a topic of great interest,[ ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹ ] lately even in combination with catalysis,[ ¹⁰ , ¹¹ , ¹² , ¹³ ] we decided to pursue the concept of synchronized movement[ ¹⁴ , ¹⁵ ] in catalytic multicomponent machinery alike that in enzymes. Coupling of motion in individual self‐assembled machines with multiple moving parts is very rarely reported,[ ¹⁶ , ¹⁷ , ¹⁸ ] and even less regarding distinct catalytic behavior. ^[19] Mostly coupling is achieved by geared motion in molecular arrangements.[ ²⁰ , ²¹ ] Hence, the challenge to synchronize motion in a self‐assembled multi‐component machine and to exploit the coupled movement for double catalysis remains unexplored.

Here, we report on three slider‐on‐deck systems, DS1–DS3, that are readily assembled using orthogonal interactions ^[22] and self‐sorting (HETPYP, N _py→ZnPor and homoleptic ^[23] [Zn(terpy)₂]²⁺ {terpy=2,2′ : 6′, 2′′‐terpyridine} complexations, see Figure 1) to enable two sliding motions with distinct exchange rates. Lastly, we will couple two catalytic functions to the stochastic motion of the slider‐on‐deck DS3.

Figure 1.

Ligands and interactions[ ²⁴ , ²⁵ ] used in this study.

Mixing ligands 1, 2 and [Cu(MeCN)₄]⁺ (Figure 1) in a 1 : 1 : 3 ratio quantitively assembled the three‐component HETPYP slider‐on‐deck DS1 (Figure 2), which was fully characterized by spectroscopic data (¹H, ¹H‐¹H COSY, ¹H DOSY, ESI‐MS) and elemental analysis. Figure 3a–c shows the pronounced ¹H NMR changes that accompany the formation of DS1 from its respective components. The binding and sliding motion of the lutidine terminals of 2 across the [Cu₃(1)]³⁺ unit causes drastic shifts and broadening of the proton signals i,p‐H from 7.05 ppm in 1 to 6.65 and 6.85 ppm in DS1. Similarly, notable shifts and broadening of signals k,n‐H located on the diarylphenanthroline unit from 7.97–7.99 ppm in deck 1 to 8.85 and 8.22 ppm and the emergence of the f′‐H signal at 2.07 ppm in DS1 (2.54 ppm in free 2) corroborate the assembly formation. Broadening of NMR signals suggests dynamic sliding in DS1, for instance, by the circular motion of 2 between three degenerate [Cu(phenAr₂)]⁺ units on 1.

Figure 2.

Stepwise conversion between the HETPYP slider‐on‐deck DS1, N _py→ZnPor slider‐on‐deck DS2 and double slider‐on‐deck DS3.

Figure 3.

Comparison of partial ¹H NMR (500 MHz, CD₂Cl₂, 298 K) spectra of a) slider‐on‐deck DS1=[Cu₃(1)(2)]³⁺, b) free biped 2 in CDCl₃, c) free deck 1, d) [Zn(3)₂]²⁺, and e) the N _py→ZnPor slider‐on‐deck DS2=1⋅[Zn(3)₂]²⁺.

Analogously, the slider‐on‐deck DS2 was assembled by mixing ligand 1 and the homoleptic bis‐terpyridine complex [Zn(3)₂]²⁺ (1 : 1). Formation of DS2 was ascertained by spectroscopic methods (¹H, ¹H‐¹H COSY, ¹H DOSY NMR) and elemental analysis. The large upfield shift of β′′‐H from 7.42 ppm in [Zn(3)₂]²⁺ to 5.67 ppm in DS2 along with the upfield shift of the r‐H signal from 10.34 ppm in 1 to 10.25 ppm in DS2 were both attributed to the axial N _py→ZnPor coordination (Figure 3c–e) exchanging between all three ZnPor sites.

After preparing the individual slider‐on‐deck systems DS1 and DS2, we assembled the double slider‐on‐deck DS3 by mixing 1, 2, Cu⁺, [Zn(3)₂]²⁺ in a 1 : 1 : 3 : 1 ratio. Quantitative formation of DS3 was confirmed by spectroscopic data (¹H, ¹H‐¹H COSY, ¹H DOSY NMR) and elemental analysis. A comparison of DS3 with DS1 and DS2 indicated that both dynamic interactions, i.e., HETPYP and N _py→ZnPor, were preserved in DS3. Further, all three assemblies DS1‐DS3 were characterized by their diffusion coefficients (r _s (DS1)=14.5 Å, r _s (DS2)=15.5 Å, r _s (DS3)=13.7 Å) obtained from DOSY NMR experiments confirming their quantitative formation (Supporting Information, Figure S26–S28).

To quantify the sliding dynamics, we undertook variable temperature (VT) ¹H NMR studies that confirmed the dynamic coordination of both lutidine terminals of biped 2 to all three degenerate [Cu(phenAr₂)]⁺ stations of deck 1. The g‐H proton of 1 in DS1 was chosen as the diagnostic signal because at 298 K it appears as a sharp singlet (6.54 ppm) and separates at 243 K into two singlets (1 : 2) at 6.46 and 6.41 ppm. The signal at 6.41 ppm is assigned to phenyl protons g‐H at lutidine‐coordinated [Cu(phenAr₂)]⁺ units, whereas the broader signal at 6.46 ppm represents the g‐H signal of the freely rotating phenyl ring adjacent to the vacant [Cu(phenAr₂)]⁺ site. A kinetic analysis provided the exchange frequency (k) of the sliding motion at various temperatures (Figure 4a) as k ₂₉₈=1.1×10³ s⁻¹ (at 298 K). Activation parameters are ΔH ^≠=66.5±1.1 kJ mol⁻¹ and ΔS ^≠=36.3±4.0 J mol⁻¹ K⁻¹ furnishing a free activation energy for exchange at 298 K as ΔG ^≠ ₂₉₈=55.7±2.3 kJ mol⁻¹.

Figure 4.

Experimental and theoretical splitting in VT ¹H NMR spectra (600 MHz). Proton a) g‐H signal of slider‐on‐deck DS1 and b) a‐H signal of DS2 in CD₂Cl₂ furnishing rate data at different temperatures.

Additionally, we recorded the VT ¹H NMR of DS2 that confirmed the dynamic coordination of the two pyridine terminals of the di‐topic biped [Zn(3)₂]²⁺ to the three degenerate zinc porphyrins in DS2. The proton signal a‐H of the porphyrin nucleus was chosen because its sharp singlet at 298 K separated at 208 K into two broad singlets (2 : 1) at 2.66 and 2.62 ppm. While the signal at 2.62 ppm corresponds to the pyridine‐coordinated zinc porphyrins, the freely rotating third porphyrin furnishes a broader signal at 2.66 ppm. A kinetic analysis provides the frequency (k) for sliding (Figure 4b) with k ₂₉₈=3.5×10⁵ s⁻¹ and the activation parameters ΔH ^≠=48.2±0.9 kJ mol⁻¹, ΔS ^≠=23.3±3.8 J mol⁻¹ K⁻¹ as well as ΔG ^≠ ₂₉₈=41.2±2.0 kJ mol⁻¹.

Subsequently, we also recorded the VT ¹H NMR of DS3 to study the two orthogonal dynamic motions on deck 1. The proton r‐H signal of the zinc porphyrin showed a sharp singlet at 298 K but split at 243 K into two singlets (2 : 1) at 10.28 and 10.24 ppm (Supporting Information, Figures S33–S35). Whereas the signal at 10.28 ppm corresponds to the pyridine‐coordinated zinc porphyrins, the freely rotating third porphyrin exhibits a broader signal at 10.24 ppm. A kinetic analysis provided k ₂₉₈=4.1×10⁴ s⁻¹ for the N _py→ZnPor sliding along with the activation parameters ΔH ^≠=49.6±0.4 kJ mol⁻¹, ΔS ^≠=9.4±1.9 J mol⁻¹ K⁻¹ and ΔG ^≠ ₂₉₈=46.8±1.0 kJ mol⁻¹. To quantify in the same VT ¹H NMR spectra the N _lut→[Cu(phenAr₂)]⁺ (HETPYP) dynamics of DS3 we focused on proton g‐H, since its sharp singlet (at 298 K) separated at 218 K into two broad singlets (2 : 1) at 6.31 and 6.34 ppm. While the signal at 6.31 ppm corresponds to the two lutidine‐coordinated [Cu(phenAr₂)]⁺ stations, the freely rotating [Cu(phenAr₂)]⁺ furnishes a broader signal at 6.34 ppm. The kinetic analysis provided k ₂₉₈ (N _lut→[Cu(phenAr₂)]⁺)=7.7×10⁴ s⁻¹ along with the activation parameters ΔH ^≠=46.2±0.4 kJ mol⁻¹, ΔS ^≠=4.2±1.6 J mol⁻¹ K⁻¹ and ΔG ^≠ ₂₉₈=45.0±0.8 kJ mol⁻¹ (Table 1).

Table 1.

Summary of activation parameters and kinetic analysis of sliders‐on‐deck (DS1, DS2, DS3, DS3′ and DS3′′) categorized by the two individual rate‐determining steps involving N _py→ZnPor and N _lut→[Cu(phenAr₂)]⁺ dissociation. DS3′=DS3+5‐bromo‐2‐pyridine carboxaldehyde. DS3′′=DS3+2‐pyridine carboxamide.

Notably, the above kinetic analysis of the dual slider‐on‐deck DS3 (Table 1) suggests that the N _py→ZnPor binding of biped [Zn(3)₂]²⁺ is enhanced at the cost of the N _lut→[Cu(phenAr₂)]⁺ binding of biped 2. Most surprising is the drastic increase of the exchange frequency for the lutidine biped from k ₂₉₈=1.1 kHz in DS1 to k ₂₉₈=77 kHz in DS3 accompanied by a drastic decrease in the exchange rate of the pyridine biped from k ₂₉₈=350 kHz in DS1 to k ₂₉₈=41 kHz in DS3 (Table 1). The close matching of both frequencies suggests that a single event determines the rates. To check whether the N _lut→ [Cu(phenAr₂)]⁺ dissociation is rate‐limiting, we slowed the motion of the lutidine slider of DS3 applying small molecular brake pads, namely 5‐bromo‐2‐pyridine carboxaldehyde and 2‐pyridine carboxamide. Both compounds, if applied in a 1 : 1 stoichiometry w.r.t DS3, should bind as chelate ligands at the free [Cu(phenAr₂)]⁺ site on DS3, requiring that for any motion of the lutidine slider 2 in DS3 the biped has to kick out the brake pad. The latter will then associate to the newly liberated [Cu(phenAr₂)]⁺ site slowing again the next motion of the lutidine biped 2, a concept that has recently been tested in related dynamic slider‐on‐deck systems. ^[26] Importantly, if there is no coupling between the two sliding motions in DS3, the movement of the pyridine biped should be unaffected.

Accordingly, we assembled DS3 including the brake pad by mixing 1, 2, Cu⁺, [Zn(3)₂]²⁺ and the chelate (brake pad) in a 1 : 1 : 3 : 1 : 1 ratio. The DS3 assemblies including 5‐bromo‐2‐pyridine carboxaldehyde and 2‐pyridine carboxamide will be hereafter referred to as DS3′ and DS3′′, respectively. VT ¹H NMR investigations and subsequent kinetic analyses were performed for DS3′ and DS3′′ as described previously. Table 1 summarizes the activation parameters and rate data for each individual slider‐on‐deck system.

Upon inspection of the data in Table 1 we see several mechanistically surprising trends. Firstly, the exchange frequencies k _lut and k _py in DS3, DS3′ and DS3′′ merge in between 1.1 kHz (as in DS1) and 350 kHz (as in DS2) with the numbers slightly declining upon addition of the brake pads. Such a finding suggests that there must be strong strain effects in DS3 (despite a 2.2 nm separation of the bipeds, Supporting Information, Chapter 7) that reduce k _py and increase k _lut. ^[27] Secondly, the lutidine slider, which is extremely slow in DS1, is accelerated by more than 60‐fold in DS3 and is now the faster process (k _lut>k _py). Thirdly, the ratio of the k _lut/k _py in DS3, DS3′ and DS3′′ is approximately constant, roughly by a factor of 2, as also reflected by a constant increment ΔΔG ^≠ ₂₉₈=1.9 kJ mol⁻¹ for the two free energy activation barriers in DS3, DS3′ and DS3′′. A rationalization of the constant increment ΔΔG ^≠ ₂₉₈ in DS3, DS3′ and DS3′′ has to consider (a) that both exchange processes have a different rate determining step, although they appear synchronized, and (b) that the rate of the N _py→ZnPor dissociation must be dictated by the lutidine slider, as the brake pads can only influence the motional speed of the latter.

For a discussion of the mechanism to delineate the cause of the apparent synchronization, we must first identify the global minimum structure(s) of DS3. On one hand there is the anti‐configuration “A” (altogether six degenerate structures) where the bipeds 2 and [Zn(3)₂]²⁺ are connected to different arms of the deck 1 (Figure 5). The notation for any of the anti‐structures describes the position (i.e. 1, 2 or 3) of biped 2 on deck 1 as the subscript, e.g. in A₁′′′, and the position (i.e. ′, ′′ or ′′′) of [Zn(3)₂]²⁺ on deck 1 as the superscript in A₁′′′.

Figure 5.

Representation of some degenerate structures A and B formed in the motion of DS3. Representative of other degenerate arrangements, the motion will start from A₁′′′ and A₂′′′.

The syn‐structures B comprise three degenerate structures (B′, B′′, B′′′) where the bipeds 2 and [Zn(3)₂]²⁺ are connected to the same two arms of deck 1. Hence the notation (e.g. B′) only requires the specification of the position of one of the two sliders. The legend in Figure 5 describes the notation used for discussions in the succeeding paragraphs.

If we begin by postulating that the syn‐structures B were the ground state, then any single‐step motion of the lutidine slider 2 would inevitably lead to the higher‐energy anti‐structures A, which should not (or hardly) be visible at equilibrium. In such a situation, the NMR would reflect the exchange B′′′⇆B′⇆B′′, requiring that both individual sliding processes have a single rate‐determining step and by extension an identical rate constant for k _lut and k _py. The experimental data presented in Table 1, however, do not support such model.

Hence, the above experimental findings lead us to postulate that A are the ground‐state structures. Suppose the motion starts from A₁′′′ and A₂′′′ (Figure 5). Interestingly, any exchange A₁′′′⇆A₂′′′ already allows the lutidine slider to visit all three copper(I) phenanthroline stations. Such process, although it does not involve syn‐structures B, may appear in the NMR like a full exchange since different positioning of one slider should not impact on the ¹H NMR shifts of the other slider due to their large spatial separation (ca. 2.2 nm).

Furthermore, there are a couple of experimental findings that need to be explained in a plausible mechanism after A has been identified as the ground state structure: (a) Since ΔΔG ^≠ ₂₉₈ (k _py vs k _lut) is constant by ca. 2 kJ mol⁻¹ in DS3, DS3′ and DS3′′, the motion of the pyridine biped A₁′′′→A₁′′ or A₂′′′→A₂′ must be impeded by a high N _py→ZnPor dissociation barrier. Otherwise, the pyridine slider would also be able to visit all three ZnPor sites (=full exchange) and furthermore its motion would become independent of the lutidine slider. (b) Consequently, motion of the pyridine slider is only possible in state B′′′, which has first to form in an endergonic step from A₁′′′, A₂′′′ by lutidine migration. As B′′′ is higher in energy and thus more strained than A₁′′′, A₂′′′, there is a good rationale why the strain‐driven N _py→ZnPor dissociation of the pyridine biped in B′′′ leading to structures A₃′ or A₃′′ is now faster than the reverse motion of the lutidine slider. (c) Importantly, such a mechanism readily explains why the motion of the pyridine biped will always depend on the lutidine slider's migration to the higher energy structures B, either with or without the brake pads.

By computation one can test part of this mechanistic scenario, because it predicts that B must be thermodynamically higher in energy than A. Indeed, the computational energy calculations predict that structure A is more stable than B by 26.7 kJ mol⁻¹ (Supporting Information, Figure S48–S50). With the computational and experimental data at hand, we are able to sketch the energy profile of the synchronized sliding. From the energy hypersurface, we realize that the process A′′′→B′′′ is endergonic (e.g. A₂′′′→B′′′ in Figure 6), which increases the activation free energy of lutidine sliding by ΔΔG ^≠ ₂₉₈=ca. 2 kJ mol⁻¹ compared to that for A₁′′′⇆A₂′′′. Assuming that the follow‐up motion of the pyridine slider in B is now faster (e.g. B′′′→A₃′) than the return to any of the two A′′′ (e.g. B′′′→A₂′′′) the system progresses to A₃′ or A₃′′ (Figure 6).

Figure 6.

Semiquantitative energy profile for DS3.

Finally, after rationalizing the synchronization of both sliding exchange processes in DS3, we sought for a possibility to use the coupled sliding for a reaction that is promoted by two catalysts. Former work in our group[ ¹² , ¹³ ] had established individual catalytic slider‐on‐deck systems for both copper(I) phenanthroline (click reaction) and zinc porphyrin systems (catalyst release leading to various organocatalytic reactions), in which catalysis vastly profits from motion.

We thus decided to utilize DS3 with its ability to liberate an organic tertiary amine (=catalyst 1) from the ZnPor unit due to the motion ^[12] and the dynamically exposed Cu⁺ (=catalyst 2) ^[13] for a reaction requiring two catalysts. The amine was expected to deprotonate 4 thus facilitating the copper(I)‐catalyzed cyclization furnishing 5 (Scheme 1). To test the two‐step reaction, we mixed 4 (5.0 mM), 1,3,5‐trimethoxybenzene, DS3, N‐methylpyrrolidine (NMP) in a 20 : 20 : 1 : 1 ratio in CD₂Cl₂ : CDCl₃ (9 : 1) in an NMR tube and heated it at 50 °C for 4 h. The resultant mixture was subjected to ¹H NMR which revealed 49 % yield of product 5 (Figure 7a). Importantly, reaction 4→5 is poorly catalyzed by (DS1⋅NMP) since NMP is rendered mainly inactive due to its binding at the zinc porphyrin units resulting in only 3 % yield of 5. Use of DS2 as catalyst expectedly resulted in no yield of 5 at all due to the lack of copper(I).

Scheme 1.

Catalytic transformation of the reactions 4→5 and 6→7 using 5 mol % of [DS1–DS3⋅NMP]. NMP=N‐methylpyrrolidine.

Figure 7.

a) Results of the catalyzed reaction of 4→5 using 5 mol % of DS1–DS3. Kinetic evolution of product 5 over 4 h. b) Results of the catalyzed reaction of 6→7 using 5 mol % of DS1–DS3. Kinetic evolution of product 7 formation over 8 h.

The methodology has a wide scope as it was readily extended to the catalyzed conversion of 6 to 7 (Scheme 1, Figure 7b). Further, monitoring the kinetics of formation of 5 and 7 in all three sliders (DS1–DS3) revealed a v ₀(DS3)/v ₀(DS1)=7.4 for 5 and v ₀(DS3)/v ₀(DS1)=20.3 for 7 indicating a significant catalytic potential of simultaneous motion of both sliders on one deck (Figure 7).

Further, for both reactions in Scheme 1, the catalytic activity of the machinery shows a clear trend along DS3⋅NMP>DS3′⋅NMP>DS3′′⋅NMP suggesting that the faster motion of both bipeds in DS3 is relevant for the higher turnover (see Supporting Information, Figures S65 and S66).

In summary, we show an unprecedented case of synchronized motion using the five‐component slider‐on‐deck system DS3, in which two very different sliding motions are kinetically coupled and result in synchronized movements despite a spatial separation of 2.2 nm. If the dual slider‐on‐deck DS3 was treated with brake pads, both motions were slowed although the brake pad would only be able to affect the lutidine slider. An unmatched mechanism explaining the coupled motion and synchronization of speed via intercomponent communication, mimicking long‐range communication in natural systems via the protein backbone(s), ^[2] is provided. The mechanism conceptually has nothing at all in common with intimate steric interactions[ ¹⁴ , ²⁸ ] or interdigitated geared motion[ ¹⁸ , ²⁰ , ²¹ ] as in other man‐made devices. Further, we have demonstrated how the coupling of both motions liberated two catalysts in high effective concentrations, a feature that was used to increase the yields in the double catalysis of benzofuran and indole products.

Conflict of interest

The authors declare no conflict of interest.

Supporting information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supporting Information

Dedicated to Professor Hiriyakkanavar Ila

V. V. Rajasekaran, A. Ghosh, S. Kundu, D. Mondal, T. Paululat, M. Schmittel, Angew. Chem. Int. Ed. 2022, 61, e202212473; Angew. Chem. 2022, 134, e202212473.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.