id
stringlengths 24
24
| idx
int64 0
402
| paragraph
stringlengths 106
17.2k
|
|---|---|---|
651a9bb1ade1178b2484ea17 | 6 | The ability to replicate higher-level results is further supported by the very strong correlation in the bottom panels of Figure with R 2 coefficients of 0.98 and 0.99 for MCAs and MAAs, respectively. These results actually outperform the ML models of Tavakoli et al. , which achieved 10-fold cross-validation R 2 coefficients of 0.92 ± 0.02 and 0.94 ± 0.02 for MCAs and MAAs, respectively. Of course, the better performance comes with a higher computational cost, although the median wall time for this data is less than two minutes using eight CPU cores (Intel(R) Xeon(R) CPU X5550 @ 2.67GHz). In fact, the timings can be further improved due to the handling of structures and conformers being embarrassingly parallel. Alternatively, omitting the single-point r 2 SCAN-3c SMD(DMSO) calculations will greatly reduce the computational cost without impacting the above results significantly as seen in Figure in the supporting information. It should be noted that the molecules in the bottom panels of Figure have on average ∼ 10 heavy atoms and ∼ 6 identified electrophilic and nucleophilic sites. |
651a9bb1ade1178b2484ea17 | 7 | To highlight the applicability of the workflow, we will first demonstrate how MCAs and MAAs can provide insights into the selectivity of chemical reactions. In Figure , we show two examples of synthesizing the antiretroviral drug Raltegravir that is used to treat HIV infections. The first example is an experimentally reported procedure from the work of Caputo et al. and the second example is a predicted retrosynthetic route by Manifold, which is a CASP tool from PostEra . The most nucleophilic and electrophilic sites for each structure are highlighted in green and blue, respectively, and additional values can be found in the supporting information. |
651a9bb1ade1178b2484ea17 | 8 | The results in Figure show that by locating the highest MCA and MAA among the two reactants (the values marked in bold), it is possible to predict the selectivity of the reaction. The most electrophilic site is the carbonyl carbon of 1a with a MAA of 395 kJ/mol, which is 89 kJ/mol higher than the second-highest MAA. The MAA of the carbonyl carbon is marked with a star ("*") as the chlorine atom acts as a leaving group during the geometry optimization when attaching the methyl anion to the carbonyl carbon. The most nucleophilic site in Figure is the primary amine of 2a with a MCA of 449 kJ/mol, which is 56 kJ/mol higher than the second-highest MCA. Hence, the experimentally observed nucleophilic acyl substitution reaction can be correctly predicted by locating the highest MCA and MAA despite the two reactants 1a and 2a having a total of 27 nucleophilic and 20 electrophilic sites identified by the workflow. Now, we will turn to Figure to analyze and validate the reaction steps proposed by Manifold and demonstrate how the workflow can be applied as a post-filtering tool. A CASP tool like Manifold usually outputs many different retrosynthetic routes, but some of the steps in the retrosynthetic routes may not be feasible due to selectivity issues. We propose using MCAs and MAAs as a tool to predict chemical selectivity and flag steps that are potentially incorrect. The retrosynthetic routes can then be ranked based on the number of warning flags similar to how retrosynthetic routes are commonly ranked based on the number of synthetic steps and the total price of building blocks. The first reaction step in Figure involves an ester amidation between 1b and 2b. However, based on the highest MCA and MAA, a more favorable reaction would be an ester amidation between two 2b compounds with MCA and MAA of 478 and 242 kJ/mol for the primary amine and the carbonyl carbon, respectively. The latter is marked with a star ("*") as the proton from the phenol group moves to the carbonyl oxygen during the geometry optimization when attaching the methyl anion to the carbonyl carbon. The MAA of the carbonyl carbon in 1b is 186 kJ/mol, which is 56 kJ/mol lower than the highest MAA. In fact, another site in 1b has a MAA that is 35 kJ/mol higher than the MAA of the carbonyl carbon. This reaction step should therefore be flagged as the chance of this reaction step being a success is low. Instead, one could imagine a similar retrosynthetic route starting with a nucleophilic acyl substitution similar to the one shown in Figure by replacing 1b with 1a. |
651a9bb1ade1178b2484ea17 | 9 | The second reaction step involves a Chan-Lam coupling reaction between the product of the first reaction step (3b) and methylboronic acid (4b). This reaction employs a copper catalyst, which makes the reaction mechanism more complex, and validating the reaction solely based on the highest MCA and MAA is probably not sufficient. However, looking at the MCAs and MAAs we see that none of the proposed reaction centers have the highest MCA and MAA. In terms of the MAAs, the workflow did not identify any electrophilic sites in 4b, and the highest MAA is therefore found in 3b. This MAA is marked with a star ("*") as the attached proton moves to the neighboring nitrogen atom during the geometry optimization. The highest MCA is 428 kJ/mol, which is only 17 kJ/mol higher than the MCA of the proposed reaction center. Removing the proton from the proposed reaction center would make it the most nucleophilic site. However, based on predicted pKa values by MarvinSketch from ChemAxon the phenol group is more acidic, and a deprotonation of this phenol group would lower the MCA of the proposed reaction center as seen in the supporting information. |
651a9bb1ade1178b2484ea17 | 10 | The last reaction step involves an ester amidation reaction between the product of the second reaction step (5b) and 4-fluorobenzylamine (6b). The most nucleophilic site is the primary amine of 6b with a MCA of 489 kJ/mol, and this site is also the proposed reaction center of 6b. The most electrophilic site is the highlighted carbon atom next to the fluorine atom of 6b with a MAA of 310 kJ/mol. The MAA is marked with a star ("*") as the fluorine atom acts as a leaving group during the geometry optimization when attaching the methyl anion to the highlighted carbon atom. The second most electrophilic site is the carbonyl carbon in 5b, which is the other proposed reaction center. This site is marked with a star ("*") as the proton from the phenol group moves to the carbonyl oxygen during the geometry optimization when attaching the methyl anion to the carbonyl carbon. In summary, the use of MCAs and MAAs to flag retrosynthetic steps for further inspection seems promising but will require further work. |
651a9bb1ade1178b2484ea17 | 11 | A second application of the automated quantum chemistry-based workflow is the ability to predict the reactivity of covalent inhibitors. In the work of Hermann et al. , calculated activation energies of different warheads reacting with CH 3 S -were used to estimate the reactivity towards cysteine. As a faster alternative, they also showed that a warhead-associated electrophilicity index could be used to predict the reactivity of some warhead classes. The electrophilicity index (ω) is defined as |
651a9bb1ade1178b2484ea17 | 12 | where ε HOMO and ε LUMO are the energies of the highest occupied and lowest unoccupied molecular orbitals. Unfortunately, employing the canonical HOMO and LUMO energies to calculate the electrophilicity index does not result in high predictability with respect to the reactivity of covalent inhibitors. Instead, Hermann et al. showed that by analyzing the highest occupied and lowest unoccupied orbitals and selecting those that are associated with the warhead (i.e., the left-hand side of the structures in Figure ), it is possible to calculate a warhead-associated electrophilicity index that strongly correlates with the reactivity of covalent inhibitors. |
651a9bb1ade1178b2484ea17 | 13 | In Figure , we compare the calculated activation energies to the warhead-associated electrophilicity index and calculated MAAs for various acrylamides (top), propargylamides (middle), and 2-chloroacetamides (bottom). The calculated MAAs are obtained for the leftmost carbon atom in each of the depicted structures with the chlorides (Cl -) being removed for the 2-chloroacetamides. |
651a9bb1ade1178b2484ea17 | 14 | The results of the acrylamides show strong correlations for both the warhead-associated electrophilicity index and MAAs with R 2 coefficients of 0.87 and 0.80, respectively. However, computing the warhead-associated electrophilicity index requires an analysis of the FMOs to select suitable orbitals, whereas the MAAs are straightforward to calculate. In fact, the R 2 coefficient for the acrylamides without adjusting the electrophilicity index to the warhead-associated HOMO and LUMO energies (i.e., simply relying on the canonical MO-based HOMO and LUMO energies) is only 0.60, which is significantly worse than calculating MAAs. |
651a9bb1ade1178b2484ea17 | 15 | The propargylamides only include nine different structures shown in Figure . When considering all of them, the R 2 coefficients are 0.64 and 0.67 for the warhead-associated electrophilicity index and calculated MAAs, respectively. Excluding the red entries in Figure , viewed as outliers in the work of Hermann et al. , results in a significantly better R 2 coefficient of 0.89 for the warhead-associated electrophilicity index. Yet, the MAAs for these structures align well with the other entries. On the other hand, the blue entry is relatively far from the black regression line. This entry corresponds to a structure with bulky groups on both sides of the triple bond, and the transition vector could be pointing toward the neighboring SP-hybridized carbon atom. Unfortunately, the transition state structures are not available. Instead, we can calculate the MAA for this neighboring carbon atom resulting in a strong correlation with an R 2 coefficient of 0.85. This approach is only possible for the atom-specific MAAs as the warhead-associated electrophilicity index uses FMOs primarily localized on both SP-hybridized carbon atoms. |
651a9bb1ade1178b2484ea17 | 16 | The results of the 2-chloroacetamides show no correlation for both the warhead-associated electrophilicity index and MAAs. This behavior is extensively studied in the work of Hermann et al. , and their arguments reflect the change from the Michael-type nucleophilic additions to an S N 2 reaction. Specifically, they show that the LUMO energy correlates with the bond strength of both the C-Cl and C-SMe bonds. Thus, the electrophilicity index fails to capture the energetics of the reaction due to the simultaneous bond formation and rupture. However, the calculated MAAs behave similarly despite not depending on FMO energies, which raises questions about their reasonings. Furthermore, when using CH 3 S - instead of CH - 3 , we again find no correlation with an R 2 coefficient of 0.01 as seen in the supporting information. This result is surprising as it contradicts the Bell-Evans-Polanyi principle (i.e., the change in activation energy for similar reactions being proportional to the change in reaction enthalpy), indicating that further analysis of the transition state structures should be carried out. |
651a9bb1ade1178b2484ea17 | 17 | We present a fully automated quantum chemistry (QM)-based workflow that automatically identifies nucleophilic and electrophilic sites and computes methyl cation affinities (MCAs) and methyl anion affinities (MAAs) to quantify nucleophilicity and electrophilicity, respectively. The workflow shows strong correlations against experimental data from Mayr's database with R 2 coefficients of 0.84 and 0.94 for the comparison of MCAs and MAAs to experimental N • s N and E values, respectively. Furthermore, the workflow achieves similar performance as higher-level PBE0-D3(BJ)/DEF2-TZVP COSMO(∞) calculations with R 2 coefficients of 0.98 and 0.99 for MCAs and MAAs, respectively, while having a median wall time of less than two minutes per molecule. |
651a9bb1ade1178b2484ea17 | 18 | Additionally, we highlight two different applications of the workflow. The first application is within computer-aided synthesis planning (CASP), where the workflow can be used to predict chemical selectivity and detect potential problems in a retrosynthetic route. This is demonstrated using experimental data from the work of Caputo et al. and a retrosynthetic route by Manifold from PostEra for the synthesis of Raltegravir. The workflow correctly predicts the reported reaction from the work of Caputo et al. by locating the highest MCA and MAA despite the two reactants having a total of 27 nucleophilic and 20 electrophilic sites. However, some of the steps in the retrosynthetic route by Manifold are found problematic suggesting that another route should be preferred. |
651a9bb1ade1178b2484ea17 | 19 | In the second application, we show that the workflow can be used to predict the reactivity of covalent inhibitors. We report a strong correlation between MAAs and calculated activation energies for various acrylamides and propargylamides similar to the work by Hermann et al. using a warhead-associated electrophilicity index. The results of the 2-chloroacetamides showed no correlation for both the warhead-associated electrophilicity index and MAAs, which could be due to errors in the calculated activation energies. The advantage of the MAAs over the warhead-associated electrophilicity index is that the MAAs are atom-specific and completely straightforward to calculate. Whereas, the warhead-associated electrophilicity index requires a selection of the highest occupied and lowest unoccupied orbitals that are associated with the warhead to match the performance of the MAAs. Future work will use the QM-based workflow to calculate MCAs and MAAs for a large set of diverse molecules and train an atom-based ML model for each property similar to the one presented in Ree et al. . The ML models will then be used to predict the chemical selectivity for a large set of experimentally reported reactions to provide a statical basis for using the MCAs and MAAs to predict chemical selectivity. An example of the visual output can be seen in Figure . |
655f8e0529a13c4d47d5b2e6 | 0 | The stability of colloidal particles in aqueous suspensions is intrinsically connected with their surface charge density, which is controlled by the pH of solution. Similarly the activity of many biologically relevant proteins and polyelectrolytes is controlled by solution's pH and ionic strength . Quantitative understanding of charge regulation in such complex systems is, therefore, of a paramount importance in a wide range of industrial and medical applications. For some simple colloidal systems with a regular distribution of surface active groups, one can use the Poisson-Boltzmann theory with the charge regulation boundary condition to study the particle protonation state . However, this approach breaks down for suspensions containing multivalent counterions or when dealing with flexible molecules, such as proteins or polyelectrolytes, whose three dimensional conformation is intimately coupled with the protonation state of the molecule. For such systems one is forced to rely on computer simulations . |
655f8e0529a13c4d47d5b2e6 | 1 | pH is defined as the negative decadic logarithm of activity, a H = c H e βµ ex H , of hydronium ions, pH=-log 10 (a H /c ⊖ ), where c ⊖ = 1 M is the standard reference concentration, β = 1/k B T , and µ ex H is the excess electrochemical potential. The constant pH (cpH) Monte Carlo simulation method is a widely used approach for generating titration curves in systems undergoing protonation/deprotonation reactions . However, an indiscriminate application of this method poses a fundamental problem. In cpH simulations, entities such as proteins, colloidal particles, or polyelectrolytes are confined within a simulation box, while protons and ions have the freedom to exchange with an acid and salt in an implicit external reservoir . Consequently, the cpH simulation method is inherently semi-grand canonical. During the course of a cpH simulation, a proton is introa) Electronic mail: [email protected] duced into the system from an external reservoir held at a predetermined pH. To maintain charge neutrality within the simulation cell, one of the cations or protons within the bulk of the cell is arbitrarily removed. However, this arbitrary removal lacks adherence to the principle of detailed balance, potentially yielding inaccurate outcomes , except in cases where the system contains a substantial amount of salt and is highly diluted in polyelectrolyte/protein. Fortunately, it is easy to rectify the standard cpH algorithm by incorporating a protonation step alongside a simultaneous grand canonical insertion of an anion. Conversely, a deprotonation step can be paired with a simultaneous grand canonical removal of an anion. This adjustment restores the detailed balance of the cpH algorithm, ensuring its internal consistency. We notice, however, that the simulation cell of such semi-grand canonical system will have a different mean electrostatic potential from that of the external reservoir. This is know as the Donnan potential. Therefore, the cpH simulation methods will allow us to predict the charge of polyelectrolyte only as a function of pH in the reservoir. Due to the presence of the Donnan potential, however, the pH of the reservoir, can be significantly different from that of an isolated system, for an identical protonation state . Thus, if one compares the titration curves, in which, say, the charge of colloidal particles is plotted as a function of the pH of the reservoir, with the titration curves of an isolated (canonical) system, there can be very large difference between the twoin particular for large volume fraction suspensions of low ionic strength . The difference between the two ensembles disappears in the limit or large ionic strength. This, perhaps, is the reason why this problem was not noticed previously -since most cpH simulations of proteins are performed at physiological concentrations and low protein volume fractions, when the difference between the two ensembles disappears. In fact, one can easily relate the pH c of a canonical system, in which the number of ions and protons is the same as the averages obtained using a semi-grand canonical simulation with a reservoir of pH gc , using equation |
655f8e0529a13c4d47d5b2e6 | 2 | where φ D is the Donnan potential between the semigrand canonical system and its external reservoir. We note, however, that the standard implementations of cpH simulations do not provide us with the value of the Donnan potential since it cancels out in the pair insertion/deletion moves used to preserve the charge neutrality during the simulation . Recently, however, we have developed a new reactive grand canonical MC-Donnan (rGCMCD) method, which allows us to determine the Donnan potential directly within the simulation , allowing us to calculate both the titration isotherms of canonical and semi-grand canonical systems simultaneously -showing that for systems of low colloidal volume fraction and low ionic strength, the number of deprotonated groups can be 100% larger in an isolated system 40,41 compared to a system connected to a reservoir of exactly the same pH. |
655f8e0529a13c4d47d5b2e6 | 3 | To perform rGCMCD simulations requires knowledge of the chemical potential of all ions present in the reservoir. This can be obtained using Widom's particle insertion method or by performing a separate grand canonical MC simulation just for the reservoir. There is also an additional complication that the Donnan potential must be calculated self-consistently during the simulation. Clearly, it is desirable to be able to obtain the titration curves directly for an isolated (canonical) system -without going through a semi-grand canonical algorithm. The difficulty is that in a canonical reactive MC simulation, one does not control the pH of the system, instead the total number of ions and protons present inside the simulation cell is specified. The simulation then determines how many of the protons will remain free and how many will be associated with the polyelectrolyte monomers. After the equilibrium is established, one can use Widom's particle insertion method to calculate the excess chemical potential of protons: |
655f8e0529a13c4d47d5b2e6 | 4 | where ∆E is the energy difference between a system with a virtual proton and without. The subscript 0 on the brackets indicates that the sampling for calculating the average is performed using the unperturbed system, without the virtual proton. To obtain pH, however, one also needs the average concentration of free hydronium ions inside the cell. For intermediate and large pH, however, there might not be any free hydroniums present inside the simulation cell at all, preventing us from accurately calculating the pH of the system. To overcome this difficulty, in this paper we will introduce a new surface Widom insertion algorithm to easily and accurately calculate the pH of a canonical system undergoing protonation/deprotonation reactions. The paper is organized as follows: In section II we briefly review the canonical reactive MC algo-rithm , in section III we preset a new surface Widom insertion method and discuss the modification of the usual Ewald summation necessary to properly account for the electrostatics of an infinite charge nonneutral system. We will also compare the titration isotherms calculated using the canonical simulation algorithms with the ones obtained using rGCMCD method. Finally, the discussion and conclusions will be presented in the section IV. |
655f8e0529a13c4d47d5b2e6 | 5 | Consider a polyelectrolyte or a colloidal particle with monomers that can undergo a protonation deprotonation reaction: In a canonical reactive MC simulation there are two types of movements: the bulk movements in which positions of ions are randomly changed with the acceptance probabilities given by the usual Metropolis algorithm; and reaction protonation/deprotonation moves, see Fig. . To construct a MC algorithm for the reaction moves, we first observe that the acid dissociation constant is the inverse of the two body partition function for the formations of a HA molecule. Thus, when a proton is transferred from the bulk to the surface, where it will react with a surface group A -, there are two changes that occur in the system: change of electrostatic energy ∆E and change in the chemical energy k B T ln(K a /c ⊖ ). The probabilities for the old (o) and new (n) configurations during a protonation move are proportional to: |
655f8e0529a13c4d47d5b2e6 | 6 | If during the deprotonation move the new coordinate falls into the interior of a colloidal particle, the ∆E is counted as infinite, and the move is rejected. The change in electrostatic energy during each move is calculated using Ewald summation with tin foil boundary condition. The Coulomb energy of a periodically replicated charged system is: |
655f8e0529a13c4d47d5b2e6 | 7 | L n 3 ) are the reciprocal lattice vectors for the cubic simulation box of side length L and volume V = L 3 . The prime on the sum indicates that i = j term is excluded from the summation when n = 0. The electrostatic energy is invariant with respect to the damping parameter κ e , which we set to κ e = 5/L, where L is the side length of the cubic simulation cell. With this choice of κ e , the sum over n can be replaced by the simple periodic boundary condition for the short range (erfc term) contribution to the electrostatic energy. |
655f8e0529a13c4d47d5b2e6 | 8 | To perform the simulations we use a cubic simulation box of side length L that contains either colloidal particles, polyelectrolyte, or protein molecules in a completely deprotonated state. In the present discussion we will use a primitive model, which treats water as a uniform dielectric continuum of Bjerrum length λ B = q 2 /k B T ϵ w = 7.2 Å, where q is the proton charge and ϵ w is the dielectric constant of water. There is, however, no difficulty to modify the algorithm to account for the explicit water or to combine it with a molecular dynamics simulation. The simulation cell also contains fully dissociated salt and acid ions -H 3 O + , Cl -, and Na + . We start the simulation with the number of H 3 O + equal to the number of negatively charged polyelectrolyte monomers. We then run the reactive canonical MC algorithm described above to obtain the equilibrium number of protonated groups and the number of free hydronium ions, see Fig. . Note that in a canonical simulation we do not have a direct access to the pH, which will be determined by the activity of hydronium ions in equilibrium. This can only be obtained using a separate Widom like particle insertion simulation that will allow us to probe the electrochemical potential of hydronium ions after the equilibrium has been established. To change pH inside the system, we can add a base such as NaOH. Since the spontaneous hydrolysis of water is so weak, addition of 1 base molecule will result in the formation of one water molecule, and an appearance of a Na + ion inside the simulation box. The net effect is, therefore, a replacement 1H + → 1Na + . We can then rerun the simulation to obtain the new protonation state. Repeating this process until all hydroniums are replaced by Na + , we can cover the full pH range. |
655f8e0529a13c4d47d5b2e6 | 9 | The crucial part of a canonical titration MC is the calculation of pH after the system has equilibrated. The standard Widom particle insertion method is usually not very practical, since to calculate the pH we need the whole electrochemical potential inside the simulation cell, and not just its excess part. At moderate to high pH, the interior of the simulation cell might not have any free hydronium ions at all, preventing us from accurately calculating the electrochemical potential and the activity of hydronium ions. On the other hand, since the condensed protons are in equilibrium with the hydronium ions in the bulk and, therefore, have exactly the same electrochemical potential, we can use them to accurate calculate the pH inside the system. There is, however, an additional complication when working with infinitely replicated Coulomb systems. The Ewald summation, effectively leads to a macroscopic crystal composed of replicated microscopic simulation cells. In general, a simulation cell will have a net electric dipole moment M M M = i q i r r r i and a finite second moment of the charge density tensor. From the electrostatics it is well known that such uniform polarization is analogous to the surface charge density M M M •n n n/V , where n n n represents the unit normal to the boundary of the macroscopic spherical crystal, see Fig. . This effective surface charge, will lead to electric field in the interior of the crystal. Similarly the fact that in general the simulation cell has a non-zero moment of the charge density tensor, results in a dipolar layer at the surface of the macroscopic crystal -so that interior of the crystal has different mean electrostatic potential compared to the exterior . This potential difference is known as the Bethe potential and for a charge neutral systems is given by: |
655f8e0529a13c4d47d5b2e6 | 10 | The derivation of this results within Ewald formalism is provided in the Appendix A. When the macroscopic crystal is "wrapped" in a tin foil, the induced image charge will kill off the surface contribution to the electrostatic potential. In general, it is known that for liquid state systems, calculations based on Ewald summation with tin foil boundary condition tend to be more accurate than the ones based on "vacuum" boundary conditions. However, in order to implement the Widom particle insertion method the system must be "unwrapped" from the tin foil, so that a proton can be brought from outside into the simulation cell. As the proton enters the crystal it will experience a jump in the electrostatic potential given by qϕ B . Note, that the Bethe potential is not constant -it depends on the instantaneous positions of all the charges inside the simulation cells. |
655f8e0529a13c4d47d5b2e6 | 11 | To make the discussion more concrete, consider a colloidal particle with Z active surface groups placed at the center of a cubic simulation cell, see Fig. . The cell also contains some number of H 3 O + , Cl -, and Na + ions and is overall charge neutral. We now run the reactive MC simulation to calculate how many of Z surface groups will become protonated. After the system has relaxed to equilibrium, we find that on average N of the surface sites are protonated. The canonical partition function of a colloidal particle with N protonated sites can be written as: |
655f8e0529a13c4d47d5b2e6 | 12 | where the Tr refers to the trace over all the microstates of both ions and protons inside the system. The electrochemical potential of a proton is the difference in free energy of two systems: one in which the colloidal particle has N protonated sites and the other N + 1 protonated sites, |
655f8e0529a13c4d47d5b2e6 | 13 | where ∆E is the difference in electrostatic energy between systems with N and N + 1 protonated groups, and ϕ B is the Bethe potential that the virtual proton gains after entering into the Ewald "crystal". Since Ewald sum periodically replicates the whole simulation cell, it will also replicate the virtual proton. A periodic charge nonneutral system will have an infinite energy. To avoid this, together with the virtual proton, in the calculation of E, we also introduce a uniform neutralizing background, which regularizes the electrostatic energy calculation, see Appendix A. The average in Eqn. ( ) is calculated using the ensemble average of the unperturbed (without virtual proton) system. We notice that the left hand side of Eqn. ( ) is c ⊖ /a H + . Taking the decadic logarithm of the two sides of this equation, we finally obtain |
655f8e0529a13c4d47d5b2e6 | 14 | where the subscript 0 on ⟨...⟩ 0 indicates that the sampling for the averages and the system evolution between the virtual proton insertion events are performed using the energy of the unperturbed system. The surface Widom insertion method brings a virtual proton from infinity into contact with a randomly selected colloidal active site; if the site is empty (has charge -q), it "reacts" with the virtual proton and its charge changes to 0 and the difference in electrostatic energy between the protonated state and the original deprotonated state, ∆E, is calculated using Eqn. (A9). The average in Eqn. ( ) is obtained using 5,000 uncorrelated insertion events. If the site is already protonated, the virtual proton will overlap with the real proton, resulting in infinite ∆E. The virtual protonation process does not affect the actual state of the site and is used just to probe the chemical potential. FIG. : Spherically replicated simulation cell, forming a macroscopic crystal. Each cell has a net electric dipole moment and a non-zero second moment of the charge density tensor. This results in a dipole layer and surface charge at the crystal boundary, producing an electrostatic potential in the crystal's interior. |
655f8e0529a13c4d47d5b2e6 | 15 | To validate the new approach, we used it to calculate the titration isotherms of 11% volume fraction suspension of nanoparticles of radius 60 Å with Z = 600 surface sites. For simplicity, we placed only one nanoparticle into the simulation cell, however, there is no conceptual difficulty in putting as many particles as is desired into the simulation cell. The simulation was performed inside a cubic cell, which in addition to the nanoparticle also contained 600 hydronium ions and four Na + and Cl -ions, corresponding to the concentration of 1mM of salt. To calculate the titration isotherm, we ran the reactive MC simulation with the acceptance probabilities for protonation/deprotonation moves given by Eq. ( ). After equilibration, the number of protonated sites was determined. To make sure the system was well equilibrated, we used 1 million particle moves. To check equilibrium we also monitored the energy of the system. After equilibration, we performed insertions of a virtual proton to calculate the pH using the surface Widom insertion method, Eq. ( ). Virtual insertions were performed at interval of 10,000 particle moves, to make sure the events were uncorrelated. We then replaced one of the initial 600 hydroniums by Na + and repeated the calculation -resulting in a slightly more negatively charged nanoparticle and a slightly higher pH. We repeated this procedure until almost all the hydronium ions were replaced by Na + , resulting in a nanoparticle with no protonated surface groups. We can summarize the sequence of calculations as follows: |
655f8e0529a13c4d47d5b2e6 | 16 | The titration isotherm calculated using this procedure is presented in Fig. . As a benchmark to check the accuracy of the new canonical titration method, we compared our results with the ones calculated using rGCMCD simulation . We see a perfect agreement between the two methods. We have then repeated the calculation for a system with 50mM of salt. Again the agreement between the two simulation methods is excellent, see Fig. The solid green curve is the benchmark calculated using rGCMCD simulation method of ref. . The solid green curve is the benchmark calculated using rGCMCD simulation method of ref. . |
655f8e0529a13c4d47d5b2e6 | 17 | We have introduced a new Monte Carlo approach to calculate the titration isotherms in a canonical ensemble. Unlike conventional constant pH (cpH) simulation methods, which inherently operate in the grand canonical ensemble and assess the protonation state of molecules solely with respect to the pH within the reservoir, our canonical titration method works directly with the isolated system. The simulation method employs a reactive Monte Carlo algorithm to determine the protonation state of macromolecules in relation to the total number of protons present within the canonical simulation cell. To compute the pH of a fully equilibrated system, we have developed a new surface insertion Widom algorithm, which effectively circumvents the challenges associated with the bulk Widom particle insertion, particularly for extremely low hydronium ion concentrations. To accurately account for the long-range Coulomb forces, we have adopted the Ewald summation method, highlighting the significance of the Bethe potential in the precise calculations of pH of canonical systems. Although the present simulation method was developed within the framework of the primitive model, in which water is treated as a dielectric continuum, there is no conceptual difficulty of extending it to more realistic atomistic simulations. In this respect, the canonical approach is much easier to implement than the alternative grand canonical methods, which require a simultaneous insertion of an anion together with a protonation move, in order to preserve the overall charge neutrality . Clearly in a dense system with atomistic water, most of insertion attempts will be rejected. On the other hand, one can easily combine the present canonical approach with a molecular dynamics (MD) simulation -so that the evolution of the system is performed using standard MD algorithms with a suitable thermostat -combined with titration moves in which one of the hydronium ions is transformed into a water molecule, with a proton transferred to one of the polyelectrolyte sites. The acceptance probability for such a protonation move will be given by Eq. ( ), and similarly for a deprotonation move. The pH calculation introduced in this paper can also be combined with standard MD algorithms, so that the sampling needed to perform the average in Eq. ( ) can be performed using a MD simulation. The implementation of the present approach to atomistic system will be the subject of the future work. |
61941a8e64a707355272903d | 0 | Tuberculosis is one of the leading global causes of mortality, and it is believed that one third of the population has a latent case of the disease. Pretomanid is a therapy for treatment of tuberculosis, and was recently approved by the US FDA under the Limited Population Pathway (LPAD Pathway) for treatment of pulmonary extensively drug resistant (XDR) tuberculosis in combination with bedaquiline and linezolid. It works as a respiratory poison against bacteria by releasing nitric oxide under anaerobic conditions. |
61941a8e64a707355272903d | 1 | Given the large quantities of drug substance that could be required, cost-effective syntheses are needed. A key feature of pretomanid is the dihydro-1,3-oxazine, containing an oxygen-substituted asymmetric center on the C3 unit (Figure ). One could foresee installation of this fragment from an (S)-glycidol derivative, and not surprisingly many of the current pretomanid routes make use of functionalized glycidols. Glycidyl pivalate appears to be a particularly important variant. 3 Enantiomers of optically active glycidol are of considerable expense, however, construction from less expensive precursors would be desirable. Epichlorohydrin is a feedstock chemical, and its pure enantiomers are highly available in comparison to those of glycidol. As a result, (R)-epichlorohydrin is approximately 5-6% of the cost 4 of (S)-glycidol and could thus form the basis for a more costeffective route to this intermediate. Numerous reports describe reaction of epichlorohydrin with carboxylates, particularly hindered carboxylates, as the ensuing glycidyl esters are used in alkyd resins, paints, coatings, and acrylate monomer compositions. Fewer reports detail the reaction of enantiopure epichlorohydrin with carboxylic acid derivatives. This work describes development of a practical route to (S)-glycidyl pivalate from low-cost and readily available (R)-epichlorohydrin and pivalic acid. |
61941a8e64a707355272903d | 2 | Our investigation began with a screen of typical conditions used to couple acids with racemic epichlorohydrin (Table ). Equivalents of starting material, preformation of the carboxylate, solvent, temperature and time were explored. In our hands, introducing an excess of epichlorohydrin was advantageous (Entries 4-6). Furthermore, removal of exogeneous solvent led to the best results, giving glycidyl ester 3 in greater than 95% assay yield (AY). While a high stoichiometry of epichlorohydrin was employed, we were encouraged that these conditions could be rendered economical if excess starting material were to be recovered. We subsequently shifted our focus on isolation of the desired compound from the reaction mixture, and the reaction scale was increased to 20 g of pivalic acid and 182 g (10 Eq.) of (S)-epichlorohydrin (Scheme 1). We expected that the high assay yield and simplicity of chemical inputs would facilitate isolation. |
61941a8e64a707355272903d | 3 | The reaction of sodium pivalate with epichlorohydrin produced one equivalent of sodium chloride which was easily removed via filtration as a result of its low solubility. Next, the epichlorohydrin (bp 118 °C) was evaporated and collected. A high proportion of the theoretical amount was collected, an important consideration in rendering an economically viable synthesis (143 g, 87%). The residual crude glycidyl pivalate (33 g, 6% epichlorohydrin) was distilled twice at 50-70 °C under high vacuum (~6-10 Torr), resulting in 74% isolated yield of the pure glycidyl pivalate. The compound appeared to be temperature sensitive at high concentration, and thus heat history was minimized. The product showed high optical activity (-21.87, CHCl3, 25 °C) as compared to literature values for (S)-glycidyl pivalate (20.7); however, the sign of rotation was inverted indicating that the undesired (R)enantiomer had been made. |
61941a8e64a707355272903d | 4 | Starting from (R)-epichlorohydrin led to (S)-glycidyl pivalate samples with [α]D values of 18.8 ° and 18.9 °. Attack of the pivalate anion on the epoxide rather than the primary chloride rationalizes this observation. Despite these highly encouraging results, analysis of the recovered epichlorohydrin revealed that the epoxide racemized over the course of the reaction. |
61941a8e64a707355272903d | 5 | Our approach was that either epichlorohydrin epimerization would need to be fully suppressed or that consumption of epichlorohydrin would need to be decreased in order to negate the requirement of starting material recycle. We first explored suppression of epimerization with the thought that at lower temperatures, the rate of substrate racemization might be significantly slower. The esterification was run at 60 °C in high assay yield. At this temperature, the enantiomeric ratio increased from 50:50 to 90:10 (Table , Entries 1-2). While this was a positive development, further improvements were required. The high assay yield was maintained at 50 °C, and enantiomeric ratio was elevated to 95:5 (Entry 3). This moved the conditions toward economic viability, however, even slight erosion of optical activity limits the ability to recycle epichlorohydrin. |
61941a8e64a707355272903d | 6 | Removing the need for the epichlorohydrin recycle would be preferable as it would simplify the system. If the epichlorohydrin excess could be reduced, the economic driver for recycle of the starting material would be eliminated. However, simply reducing the equivalents of epichlorohydrin led to much lower yields, and a large amount of decomposition was observed (Entries 4-5). The root cause was believed to be a heat sensitivity, where bimolecular degradation of the product most likely accelerated at elevated concentrations. To evaluate this hypothesis, the reaction run with 3 or 6 equivalents of epichlorohydrin was simply diluted with inert chlorobenzene to a volume equivalent to that of 10 equivalents of epichlorohydrin. This did indeed provide a significant increase to yield at and above 80% (Entries 6-7). Decreasing temperature to 60 °C was found to be the best solution as it further increases yield, avoids the need for exogenous solvent, greatly increases throughput of material, and renders the system highly economical as compared to glycidol. With the revised synthetic methods now in hand, we then sought Glycidyl pivalate isolation methods. Two solutions struck us as feasible: |
61941a8e64a707355272903d | 7 | 1. Prepare an in situ solution of glycidyl pivalate 2. Access a higher purity variant via distillation Each solution has benefits and drawbacks. The first option is desirable in that it avoids the distillation of the epoxide. Some temperature sensitivity was noted for the epoxide, and production of a reactive solution could maximize yield by limiting the heat history and concentration of the epoxide. However, this alternative does not provide a means of purifying the glycidyl pivalate, and the excess epichlorohydrin must still be removed. If successful, the second option provides a means of removing byproducts from the glycidyl pivalate to obtain a more highly controlled and pure product. |
61941a8e64a707355272903d | 8 | Production of an in situ solution was explored first (Figure ). A solvent swap to toluene was desired as toluene could be used in the ensuing synthetic chemistry. In first attempts toward this goal, epichlorohydrin was directly distilled from the reaction mixture under vacuum, and then toluene was added intermittently to make-up the volume lost from epichlorohydrin evaporation. Volatiles were then fully removed to give a glycidyl pivalate residue. The process was repeated three times. This led to a loss of active glycidyl pivalate in solution as observed by decrease in the assay strength (10-15%) and also the observation of unidentified by-products (Figure ). Again, heat and concentration sensitivities were suspected to cause the loss in activity. If the solvent swap could be conducted while maintaining constant volume, perhaps the decomposition could be avoided. This was accomplished by adding toluene continuously to a stirred solution of the glycidyl pivalate reaction mixture which was under vacuum (Figure ). Performing the solvent swap in this manner largely prevented the loss of active glycidyl pivalate to decomposition products. The reaction mixture had a 94% AY at the end of reaction (EOR) and a 90% AY after removal of epichlorohydrin through a toluene solvent swap. This yields a reactive solution which could be used for the subsequent alkylation step. 2b Next, we attempted to isolate glycidyl pivalate in good purity by direct distillation 9 (Table ). First, sodium chloride was removed from the reaction mixture via filtration, and then epichlorohydrin was evaporated from the filtrate. Care was taken to evaporate epichlorohydrin at minimal temperature under strong vacuum (<10 torr) so that vapors of epichlorohydrin do not exceed 60 °C. After evaporation, the assay yield of the crude glycidyl pivalate residue was 87%. |
61941a8e64a707355272903d | 9 | The product was then distilled. Again, it was important to apply maximum levels of pressure so that the temperature of glycidyl pivalate did not exceed 70 °C. At higher temperature, lower yields were observed as a result of product decomposition. Conditions used in this work were to distill at 50 °C and 6 Torr. The distillation was also scaledependent and yield increased significantly as glycidyl pivalate volumes grew. This is likely a function of system configuration which can be further optimized upon implementation, and might benefit from a continuous distillation system such as a thin-film evaporator so as to minimize thermal history of the heat sensitive compound. Isolated yield reached 76% with material of 95 wt% purity. Optical activity of the epoxide samples was confirmed through derivatization with 4-nitro-2-bromo-imidazole. The samples synthesized from optically active epichlorohydrin were compared against those of racemic epichlorohydrin. The SFC traces indicated an enantiomeric ratio 97:3, 10 which was consistent with the high optical activity observed from the specific rotation. |
66bfff4df3f4b0529023946e | 0 | The 2021 Nobel Prize in Chemistry has recognized the enormous potential of asymmetric organocatalysis as an environmentally benign enabling technology. While the pioneering studies of the field were based on the reversible covalent interaction of the substrate with the catalyst (via enamine or iminium formation), much recent work has focused on organocatalysts that interact with the substrate(s) in a non-covalent manner (e.g. thiourea catalysts). A privileged class of catalysts that operate via non-covalent interaction are chiral BINOL (1,1'-bi-2-naphthol) phosphorus(V) Brønsted acids (Fig. ). Typically, these catalysts activate the substrate by protonating a Brønsted-basic nitrogen atom and subsequently coordinate the activated species by ion-pairing and/or hydrogen bonding, thus guaranteeing stereoselectivity. A notable trend in BINOL-derived phosphoric acid organocatalysis is that they have been endowed with very large aromatic groups in an effort to increase stereoselectivity by generating a confined reaction space. Similarly, two BINOL-moieties have been covalently linked in imidodiphosphates, which exhibit fine-tuned pKa values and highly confined reaction sites whose structural complexity resembles that found in enzyme pockets. In light of this trend toward architectures featuring confined reaction sites, we wondered whether the integration of a phosphoric acid into the cavity of a shape-persistent macrocycle could lead to improved reactivity or selectivity. The cycloparaphenylenes (CPP) seemed particularly suitable for this purpose as they are highly strained and comprise only aromatic units, which makes them exceptionally rigid and susceptible for - interactions. While their optoelectronic and supramolecular properties have led to applications in organic lightemitting diodes (OLEDs), organic field-effect transistors (OFET) and organic photovoltaics (OPV) , their use in synthesis has been restricted to stoichiometric (active) template effects and the non-specific initiation of asymmetry amplification in a Soai reaction. Herein, we report the synthesis and investigation of chiral nanohoop organocatalyst PO4H-[9]CPP. By closing a strained CPP macrocycle around the privileged BINOL phosphoric acid, we create a unique reaction site that exhibits a well-defined cavity as well as ortho-substituents that are bent towards the cavity and conformationally restricted. We chose the asymmetric transfer hydrogenation of quinolines as benchmark reaction to test the new organocatalyst, because pioneering work by Rueping established the beneficial role of large aromatic ortho-substituents on enantioselectivity, which was attributed to a confinement effect (Fig. ). Moreover, Niemeyer and coworkers used this reaction recently to investigate the effect of another complex topology on organocatalysis: in a highly flexible [2]catenane architecture, the interplay of two Brønsted acid sites was found to increase enantioselectivity (Fig. ). We now show that the rigid nanohoop approach, despite offering only one reaction site, leads to much higher enantioselectivity than the conventional BINOL catalyst (with comparable orthosubstituents). |
66bfff4df3f4b0529023946e | 1 | To introduce the chiral BINOL unit into a cycloparaphenylene nanohoop, we treated enantiomerically pure MOM-protected BINOL (S)-1 with dibromide 2, a common precursor used in CPP synthesis, to close the ring by Suzuki-Miyaura cross-coupling (Fig. ). Following the macrocyclization which proceeded in 24% yield, (S)-3 was subjected to a freshly prepared solution of sodium naphthalenide to facilitate aromatization. Deprotection of the MOM groups using Amberlyst ® resin produced the macrocycle (S)-OH- [9]CPP in 52% yield over two steps as a pale yellow solid. Phosphorylation of BINOL nanohoop (S)-OH- [9]CPP with POCl3 with the subsequent hydrolysis of the intermediate phosphorus chloride yielded the final BINOLderived phosphoric acid (S)-PO4H- [9]CPP in 98% yield over the last two steps. Synthesis of the other enantiomer (R)-PO4H- [9]CPP was performed following the same procedure from the R-isomer of the BINOL boronic ester (R)-1 in similar yields (see S.I. for further details). All new compounds were fully characterized by 1 H and C NMR spectroscopy and high-resolution mass spectrometry (Supplementary Figs. 39 -59). Single crystals of nanohoop (R)-OH- [9]CPP were successfully grown by diffusion of n-hexanes into a saturated solution of the compound in 1,2-dichlorobenzene (Fig. ). Single crystal X-ray diffraction (SC-XRD) reveals that the introduction of a BINOL moiety via meta-functionalization into a cycloparaphenylene results in a structure that is relatively similar to the parent nanohoop [9]CPP. For instance, both (R)-OH-[9]CPP and [9]CPP exhibit, in the solid state, oval shape with near-identical internal cavity diameters of 12.8 Å (vs. 12.8 Å for [9]CPP) and 11.2 Å (vs. 11.3 Å), respectively. This structural similarity is possible because of the torsion angle of 73 ° within the BINOL skeleton, which is also responsible for the chirality of the nanohoop (Fig. and Supplementary Figs. ). Based on the Flack parameter we were able to confirm the absolute configuration of the nanohoop (R)- |
66bfff4df3f4b0529023946e | 2 | CPP in the crystal structure. In comparison to related BINOL-derived macrocycles, the new nanohoop exhibits similar structural parameters, which is even true when comparing the BINOL torsion angle found in (R)-OH- [9]CPP with those observed in larger, less strained nanohoops. This finding indicates that the unusual catalytic activity described below are likely less due to a structurally deformed BINOL scaffold and rather due to confinement. We also carried out DFT calculations of the phosphoric acid macrocycle (S)-PO4H- [9]CPP and its precursor (S)-OH- [9]CPP at the B3LYP/6-31 G(d) level of theory (for further details see the S. I.). The optimized structures of (S)-PO4H- [9]CPP revealed the expected endo orientation of the phosphoric acid (Supplementary Fig. ), thus creating the desired confined reaction site. |
66bfff4df3f4b0529023946e | 3 | The strain energy of nanohoops (S)-PO4H- [9]CPP and (S)-OH- [9]CPP was determined using the homodesmotic reaction model, providing nearly identical values of 40 and 42 kcal mol -1 , respectively. The difference in strain energy between the phosphoric acid derived nanohoop ((S)-PO4H-[9]CPP) and the BINOL derived molecule ((S)-OH-[9]CPP), while relatively small, is likely real, because it is and reflected in the deviation of the BINOL torsion angle in the ring versus the relaxed geometry, which is larger for the non-tethered (S)-OH- [9]CPP macrocycle (Supplementary Fig. ). |
66bfff4df3f4b0529023946e | 4 | Having achieved the synthesis and characterization of the enantiomerically pure phosphoric acids (S)-and (R)-PO4H-[9]CPP, we proceeded with their application in asymmetric organocatalysis. We chose the stereoselective hydrogenation of quinolines as a model reaction using the classic Hantzsch ester as a hydride donor (Table ). Our initial optimization was conducted with 2-phenylquinoline (4a) as substrate with (S)-PO4H- [9]CPP at 10 mol% catalyst loading under conditions well-established in the literature. 51 To our delight, we observed an excellent enantioselectivity of 95% ee, with the (S) catalyst favoring the formation of the (R)-hydrogenated quinoline (5a, Table , entry 1). We found that decreasing the catalyst loading from 10 to 5 and 2.5 mol% did not significantly affect the enantioselectivity or the yield of the reaction, provided that the reaction was performed over a sufficient period of time (Table , entry 1-3). For the remaining experiments, we opted for a catalyst loading of 5 mol% as a reasonable compromise between catalyst quantity and reaction times (typically 24 hours). We next compared our catalyst (S)-PO4H- [9]CPP to the simpler BINOL derivatives (S)-PO4H-Ph and (S)-PO4H-Biph, which possess phenyl and biphenyl ortho-substituents, respectively (see Table ). We found that the absence of a confined "pocket" in reference catalysts (S)-PO4H-Ph and (S)-PO4H-Biph (as an extended version, more closely resembling the constitution of the nanohoop), led to a pronounced decrease of the enantioselectivity to 48% and 51% ee, respectively, under otherwise identical reaction conditions (Table , entry 4-5). As a final experiment, we confirmed that the other enantiomer of our catalyst, (R)- |
66bfff4df3f4b0529023946e | 5 | To follow up on this promising initial result, we explored the scope of the reaction with different 2-substituted quinolines (4a-f) (Table ). Both the reactivity and the enantioselectivity remained at the high level observed for the phenyl derivative when using more electron-donating or electron-withdrawing aromatic substituents (Table , entry 2-3). The same was found for the more sterically demanding biphenyl and t butyl substituents (Table , entry 4-5), indicating that the nanohoop catalyst does not increase enantioselectivity at the cost of a loss of activity for larger substrates. A reasonable explanation for this result is that the catalytic site in the macrocyclic catalyst is not confined in all three dimensions, but only in two. This reasoning is in agreement with what is known about the mechanism of the reaction, which features two consecutive hydrogenation events. Following the initial hydrogenation of the C-C bond, the subsequent hydrogenation of the C-N bond determines enantioselectivity. To achieve high ee, the Hantzsch ester needs to favour the Re or Si approach toward the complex of the substrate with the phosphate (Table ). It is plausible that the macrocyclic architecture allows enough space for the formation of this ternary complex between catalyst, substrate and Hantzsch ester, while offering a large difference between the Re and Si face due to its exceptionally high rigidity as well as secondary - interactions (Table ). In contrast to the wide scope observed for aryl groups, we found that the reactivity dropped drastically when subjecting 2-alkyl-substituted quinolines to transfer hydrogenation with our catalyst (Table , entry 6-7). Both 2-n-pentyl and 2-methyl quinolines reproducibly gave rise to sluggish reactivity and low levels of enantioselectivity, which is not the case for simple BINOL-based phosphoric acids, as reported previously by |
66bfff4df3f4b0529023946e | 6 | Rueping and successfully reproduced by us. The observation that quinolines bearing small aliphatic groups are excellent substrates for simple catalysts, yet no good substrates for the confined nanohoop catalyst, is counterintuitive and suggests that specific non-covalent effects dictate reaction outcomes. In an effort to gain a tentative understanding of the unexpected, low reactivity towards alkyl quinolines, we carried out several experiments. Given that the strong binding of substrate and/or product to the active catalytic site is a plausible mechanism for catalyst deactivation, we tested |
66bfff4df3f4b0529023946e | 7 | for "product inhibition" by performing the transfer hydrogenation of the phenyl substrate 4a in the presence of 1 equivalent of n-pentyl product 5f . After 24 hours, we observed full conversion of 4a to the respective product 5a with 93% ee, which is why we can rule out product inhibition as the origin for the low reactivity of alkyl substituted products (see further below for a study on the relative affinity of 4a, 4f, 5a and 5f for the catalyst). In a slightly different competition experiment, we combined the phenyl substrate 4a with the n-pentyl substrate 4f and observed no detectable transfer hydrogenation of 4a, along with only 2% conversion of 4f into 5f after 24 hours. This experiment indicates that the presence of the alkyl substrate 4f in the reaction medium "poisons" the catalyst for the otherwise reactive substrate 4a, for which a possible explanation is that the alkyl substrate binds to the binding pocket with higher affinity, yet with no turnover (Supplementary Fig. ). Further 1 H and P NMR experiments allowed a comparison of the affinity of substrates 4a and 4f, as well as the hydrogenated products 5a and 5f, towards the catalyst (S)-PO4H-[9]CPP. In sequential titration experiments between these four species, we determined that quinoline 4f exhibited by far the highest affinity with entire affinity series being: 5a < 4a < 5f < 4f (Supplementary Fig. -14; independently corroborated by ESI-MS, see Supplementary Fig. ). Thus, the products bind less strongly to the catalyst than the substrates, which is a prerequisite for high turnover numbers in supramolecular catalysis that is easily plagued by product inhibition. The particularly strong binding of alkyl substrates to the binding pocket of the nanohoop catalyst (S)-PO4H- [9]CPP however represents a problem, because it prevents catalytic turnover. At this point, we can only speculate on the specific reason is for the lack of turnover in the nanohoop cavity. A complexation-induced pKa shift could prevent the very first step in the catalytic cycle, which is the protonation of the quinoline and the formation of a tightly bound ion pair, yet this seems unlikely, because alkyl-substituted quinolines are more basic than their aryl-substituted counterparts. Moreover, the structure of the ternary complex with the Hantzsch ester or its corresponding pyridine may be significantly different for alkyl substrates. However, we did not observe indications to this end in qualitative experiments focusing on the Hantzsch ester concentration and the spiking of reactions with the pyridine. Further mechanistic work on specific non-covalent effects in the nanohoop cavity is clearly needed and currently ongoing. |
66bfff4df3f4b0529023946e | 8 | Finally, we examined the optical and chiroptical properties of the synthesized nanohoops ((S)/(R)-OH-[9]CPP) (Fig. and Supplementary Figs. 67 -74). The absorption spectra of (S)and (R)-OH- [9]CPP exhibit a broad absorption band reaching a maximum at around 330 nm with a molar absorption coefficient of ~1.1 10 5 M -1 cm -1 . Both enantiomers of the nanohoop are emissive and exhibit bands with two maxima at 452 and 473 nm, which are responsible for the distinct blue emission that is typical for symmetry-broken [9]CPP-type nanohoops (Fig. ). Electronic circular dichroism (ECD) spectra show the expected mirror-imaged spectra featuring pronounced Cotton effects thanks to the rigid conjugated architecture (Fig. ). The circularly polarized luminescence (CPL) spectrum of (R)-OH- [9]CPP manifests the same negative sign at higher energies as the circular dichroism spectra with a maximum of polarization at 432 nm with a glum value of -2.6•10 -3 . (S)-OH- [9]CPP exhibits very similar and mirror shape behavior but positive glum value of +2.8•10 -3 (Fig. ). Overall, the chiroptical properties fall into the useful and typical range for highly emissive conjugated macrocycles. We show that unusual selectivity and scope is observed in asymmetric transfer hydrogenations catalyzed by a BINOL phosphoric acid that is integrated into a shape-persistent CPP nanohoop. |
66bfff4df3f4b0529023946e | 9 | 50% to ca. 95% when comparing the new nanohoop organocatalyst with structurally comparable, simple BINOL phosphoric acids. While the scope for 2-aryl substituted quinolines was wide and surprisingly even included sterically rather demanding substrates, we found that 2-alkyl quinolines were not good substrates for our catalyst, which is in stark contrast to simpler catalysts that are highly effective for this type of substrates. Diverse studies on this unexpected finding point towards a highly specific effect that is due to the exceptionally strong binding of alkyl substrates to the macrocycle in conjunction with the lack of turnover. We expect that this work will inspire researchers with an interest in shape-persistent macrocycles to test compounds that are relatively easy to synthesize as catalysts. Likewise, researchers with an interest in organocatalysis may be well-advised to tackle their hardest problems with catalysts that are rigid, yet topologically non-trivial (e.g. macrocyclic, bowl-shape, cage-like, 69,70 mechanically interlocked 71 ). |
66bfff4df3f4b0529023946e | 10 | cooled down to -78 °C. The freshly prepared solution of sodium naphthalenide (3.0 mL, 0.5 M, 1.50 mmol, 12 equiv) was added dropwise and the solution stirred at -78 °C for further 2 h. The reaction was quenched by adding a solution of iodine in THF. An aqueous solution of Na2S2O3 was added and the reaction was extracted with DCM and the combined organic layers were dried over MgSO4 and concentrated under reduced pressure. The crude product was then dissolved in anhydrous THF (20 mL) and methanol (10 mL) and stirred for 1 d at 70 °C together with Amberlyst ® 15 (0.20 g). After cooling down to room temperature, Amberlyst ® 15 was filtered of, the solution was concentrated under reduced pressure and purified by column chromatography (petroleum ether / DCM 3:2 → 1:1) to yield (S)-OH- [9]CPP as yellow solid (54 mg, 66 μmol, 52%). |
66bfff4df3f4b0529023946e | 11 | Under N2, (S)-OH- [9]CPP (20 mg, 24 μmol, 1.0 equiv) was dissolved in anhydrous pyridine (2 mL). POCl3 (18.7 mg, 12 μL, 0.12 mmol, 5.0 equiv) was added to the solution and the mixture was heated at 60 °C during 24 h. After that time, H2O (2 mL) was added and heated at 60 °C for another 24 h. Then, the reaction was allowed to cool down to r.t. and the solvent was evaporated under reduced pressure. The crude was extracted three times with DCM using HCl (3 M) as aqueous phase and the combined organic layers were concentrated under reduced pressure without drying the organic phase. The resulting yellow solid was repeatedly washed with toluene in order to remove any traces of HCl and evaporated under reduced pressure to co-evaporate with the remaining traces of water. Final compound (S)-PO4H- [9]CPP was recovered as a light yellow solid (20 mg, 23 μmol, 98%). |
64ec41273fdae147fafe0a45 | 0 | Single-molecule magnets (SMMs) exhibit retention of magnetization under certain conditions, and therefore have potential applications in high density binary data storage. However, the viability of even the best-performing SMMs in such applications is currently limited as the desired magnetic properties have not yet been observed at temperatures that are suitable for practical usage. Some of the most promising high-temperature SMMs that have been studied so far are those involving dysprosocenium cations and related derivatives; the remarkable magnetic properties of this family of complexes has been attributed to the incorporation of substituted cyclopentadienyl (Cp R ) ligands, which: i) provide the steric bulk required to achieve axial ligand fields, enhancing the magnetic anisotropy of the Dy 3+ cation and hindering Orbach relaxation; and, ii) are rigid and thus impede molecular vibrations, suppressing Raman relaxation. Further stabilization of the ground mJ state of Dy 3+ in dysprosocenium cations, and concurrent destabilization of the least magnetic mJ state, could be achieved by increasing the strength of the axial ligand field through the reduction of Dy … Cp R distances by using smaller Cp R ligands. The isolation of the decamethyldysprosocenium cation, [Dy(Cp*)2] + (Cp* = {C5Me5}), has therefore long been a desirable target for the SMM community. The effective energy barrier to relaxation of magnetization (Ueff) value of the hypothetical [Dy(Cp*)2] + cation was calculated to be above 1000 cm -1 by Gao and co-workers in 2016. Recent work by Reta et al. concluded that although the linearity of the [Dy(Cp*)2] + cation and therefore its Ueff value would be reduced compared to the [Dy(C5 i Pr5)(Cp*)] + cation, the vibrational modes may be shifted off-resonance with electronic states that could lead to comparatively slower relaxation rates. However, it has been shown that the relatively small steric bulk of the Cp* ligand (cf. C5 i Pr5 and Cp ttt , C5H2 t Bu3-1,2,4) is not sufficient to block equatorial ligand interactions and allow the isolation of the [Dy(Cp*)2] + cation, even with weak donor ligands and weakly coordinating anions (WCAs). All SMMs containing {Dy(Cp*)2} fragments that have been reported to date have shown suppressed SMM properties compared to isolated dysprosocenium cations. Following our recent reports of the heteroleptic dysprosocenium complex [Dy(Cp ttt )(Cp*)][Al{OC(CF3)3}4] and its mono-halobenzene-solvated adducts, we resolved to apply these methods to {Dy(Cp*)2} analogues. Here we report the synthesis of the putative WCA-bound complexes "[{Ln(Cp*)2}{Al[OC(CF3)3]4}]" (Ln = Y, Dy), together with a family of halobenzene-bound {Dy(Cp*)2} + cations with the same WCA derived from this starting material. We found that the reduced steric bulk of the {Dy(Cp*)2} fragment permits a greater number of transverse ligand interactions, and increases the propensity of the cation to promote C-X activation and form decomposition products. The halobenzene adducts of {Dy(Cp*)2} + are characterized by single crystal (SCXRD) and powder X-ray Diffraction (PXRD), elemental analysis, ATR-IR spectroscopy, elemental analysis and SQUID magnetometry. The magnetic data are rationalized by complete active space self-consistent field spin-orbit (CASSCF-SO) calculations. We find that coordination of haloarenes promotes magnetic relaxation via structural distortions and an increased transverse field, but we observe a Ueff value of 930 (6) cm -1 for [Dy(Cp*)2(PhF-κ-F)2][Al{OC(CF3)3}4] that sets a new benchmark for {Dy(Cp*)2}containing SMMs. |
64ec41273fdae147fafe0a45 | 1 | For 1-Ln, the ATR-IR spectra feature aliphatic C-H stretches within the expected wavenumber range (3000-2800 cm -1 ). Additionally, one broad band is present in the region corresponding to a bridging bidentate BH4 -group (ca. 2250 cm -1 ); no peaks for terminal B-H stretches are observed. DFT calculations were performed on the diamagnetic Y analogs of the cations of 2-Dy, 3-Dy and 4-Dy |
64ec41273fdae147fafe0a45 | 2 | For axial Dy 3+ complexes, the exclusion of equatorially coordinating ligands is inherently challenging due to the relatively large size of the highly Lewis acidic Dy 3+ ions, the lack of directionality in Ln bonding regimes, and the tendency to maximize contacts with hard ligand donor atoms. The difficulties encountered in the isolation of axial dysprosocenium cations bending of the Dy-X-Cipso angle is observed with halogen size; this can be attributed to the anisotropic electrostatic potential distribution around the halide atoms. The variance in angle is greatest between PhF and PhCl, which is likely due to the much longer Dy-X distances Ueff is higher than previously reported for [Dy(Cp*)2] + (ca. 1000-1100 cm -1 ), which can be attributed to different geometries used and methods of determining Ucalc. In contrast, samples Halobenzene coordination promotes all relaxation mechanisms, although the observed rates also depend on induced structural distortions to the {Dy(Cp*)2} + core. These data confirm that for the [Dy(Cp*)2] + cation to be isolated and free of equatorial ligand interactions, a more weakly coordinating counter-anion needs to be employed or a significantly different experimental approach needs to be followed, e.g. encapsulation of [Dy(Cp*)2] + in an appropriate host-guest matrix. |
676acca881d2151a02c71391 | 0 | As the growth of synthetic organic chemistry fields was tremendously enabled by various forms of catalytic principles, modern protein bioconjugation strategies also rely heavily on the power of catalysis for fundamental development and various applications. Catalytic transformations have revolutionized industries of small molecule synthesis over the century. Because there have been growing interests in chemically modified proteins as renewable resources and materials (e.g., batteries, supercapacitors, anti-bacterial materials, and enzymes ), catalytic chemical modification of proteins would hold great promise for industrial production of such useful bioconjugates to address the global challenges regarding energy and environments. In addition to the potential improvement of production efficacy of bioconjugates, catalytic platforms can also be leveraged for site-specific modification of proteins for production of therapeutically important agents such as antibodydrug conjugates. Furthermore, a range of chemical tools in biochemistry and chemical biology fields is predicated on catalytic systems enabling target-specific labeling in living systems. Those protein modification applications have been achieved by addressing challenges of protein-targeting catalysis such as aqueous environments, low reaction concentrations (often nM to μM), catalyst deactivation by protein functional groups, and mild reaction conditions (e.g., 37 °C or rt). Despite the substantial development of protein bioconjugation through catalysis, systematic analysis of catalytic bioconjugation has been simply lacking to date, even though numerous reports reviewed the advances of protein bioconjugation either in broader fashions including non-catalytic systems or with emphasis on specific catalysis types such as transition-metal catalysis. This review article is focused on impacts of design, selection, and applications of catalysts on protein bioconjugation purposes and categorized into seven types of catalytic transformation mechanisms. The scope of the review article is chemical modification of natural proteins comprised of 20 canonical amino acids, particularly methods without genetic engineering. Methods that can be applicable at a protein level are primarily discussed, and relevant work has been chosen based on arbitrary molecular weight cutoff of 5,000 Da. Importantly, the main focus is on catalytic strategies rather than residue-specific organization found in many existing protein modification review papers. Nonetheless, recognizing the practical importance of a list of catalytic bioconjugation methods organized by target amino acid residues and other factors, we included an excel file containing lists of the discussed literature as Supporting Information, in which readers can sort and analyze the collection of papers with a parameter of interest. Among the diverse catalytic systems, seven catalytic mechanisms were used for organization of the article (Figure ) as the catalysis types are serving as common organizing topics of major catalysis-focused journals such as Nature Catalysis, ACS Catalysis, and ChemCatChem. Although, strictly speaking, only catalytic amounts (i.e., less than the amounts of reactants) should be ideally used for a given catalysis system, bioconjugation processes that are mechanistically catalytic even with the necessary use of excess amounts of a catalyst have been included in this present review. Though bioconjugation often indicates bond formation for functionalization purposes, catalytic bond cleavage reactions were also discussed. Because the focal point of the article is nonenzymatic processes, we refer readers to enzymatic bioconjugation-specific reviews. |
676acca881d2151a02c71391 | 1 | Electrochemical synthesis, known for its low waste, high selectivity, and mild reaction conditions, has recently attracted attention as a clean approach for small molecule conversions. Because of these advantages, electrochemical synthesis has been studied as a method for direct late-stage modification of complex compounds. The potentials make it a promising technique for chemical modification of proteins, which often contain numerous reactive functional groups and can be challenging to purify from a reaction mixture with excess labeling reagents. |
676acca881d2151a02c71391 | 2 | At a peptide level, electrochemical synthesis offers two methodologies for modification. The first method involves generating an active electron acceptor from a stable precursor via electrode oxidation, which then modifies residues in a peptide through electrophilic substitution reactions. The second method directly generates radical species by single-electron oxidation of a specific functional group in a peptide, which are subsequently captured under suitable reaction conditions to produce a modified peptide. In terms of application to protein modification, the first method is currently applicable to proteins or long-chain peptides, while the second is limited to peptides with up to 10 residues. Therefore, this review focuses only on the first method through electrode oxidation. However, it is worth noting that in this approach, the modified residues themselves may also undergo electrode oxidation as the reaction progresses. |
676acca881d2151a02c71391 | 3 | In this subsection, we introduce a method for protein modification using electrophilic species generated by electrochemical reactions as reactive anchors (Scheme 1A). By applying an appropriate potential specific to the anchor precursor, active anchor sites are oxidatively generated in the reaction system, while preserving the functional groups of the protein. These electrophilic anchor sites modify proteins via electrophilic aromatic substitution reactions at Tyr and Trp residues. The Tyr-ene reaction of 4-phenyl-3H-1,2,4-triazole-3,5(4H)diones (PTAD) and Tyr residues has been extensively studied by Barbas and co-workers, and is regarded as a promising approach for protein modification. However, the oxidative generation of PTAD from 4-phenylurazole in such systems is not orthogonal to the various functional groups of proteins. Furthermore, competing side reactions, such as hydrolysis of PTAD, restrict its applicability in a broad context. In 2018, Gouin and co-workers reported the first protein modification employing electrochemical techniques. This method exploits the low redox potential of 4-phenylurazole (+0.36 V vs. SCE), enabling rapid Tyr-ene reactions while generating active PTAD species in situ, without oxidizing the aromatic rings of the protein (>0.45 V, Scheme 1B). Another example comes from Heptinstall and coworkers, who reported modifications of protein Tyr residues via iodination or nitration. These modifications are achieved through the electrochemical oxidation of KI or NaNO2, which activates the reagents. |
676acca881d2151a02c71391 | 4 | Trp residues can also be modified by aromatic electrophilic substitution reactions under appropriate conditions (Scheme 1B). In this approach, Trp residues are selectively modified by oxoammonium anchors generated by the polar oxidation of the stable organoradical (keto-ABNO). The redox potential of keto-ABNO is lowered when it complexes with Trp residues, enabling functional group tolerance and selective modification of Trp residues. |
676acca881d2151a02c71391 | 5 | In addition to the electrophilic reagents discussed in the previous subsection, electrochemical methods utilizing radical species as anchors have been explored for protein modification (Scheme 2A). In this approach, proteins are modified by radicals generated through one-electron oxidation of anchor precursors at the electrode surface. Radical reactions have been a powerful tool for protein modification because the approach can proceed efficiently in aqueous environments and exhibit low reactivity with polar functional groups. |
676acca881d2151a02c71391 | 6 | The Tyr-ene reaction shown in the previous subsection can also proceed under one-electron transfer conditions when 1methyl-4-phenylurazole or N-methyl phthalic hydrazide is used instead of 4-phenylurazole (Scheme 2B). In 2020, Nakamura and co-workers reported that nitrogen-centered radicals generated by electrode oxidation of N-methyl phthalic hydrazide can modify Tyr residues on variety of proteins. Similarly, Lei and co-workers reported that nitrogen-centered radicals generated by electrochemical oxidation of phenothiazine are also effective for modifying Tyr residues. In 2022, Weng and co-workers introduced a modification method targeting Trp residues using azidyl radicals (N3･, Scheme 2B). These radicals add to indole side chain, enabling the modification of peptides with 20 or more residues. Azidyl radicals are generated electrochemically from Mn II -N3, accompanied with C=N double bond formation. Weng and coworkers subsequently reported that trifluoromethyl radicals (CF3･) are also applicable to modify Trp residues. The methods described in the above two subsections are notable for specifically targeting Tyr or Trp residues for protein modification. These residues are appealing targets because of their low surface exposure, the controlled nature of their modification reactions, and the minimal impact on post-modified protein structure. However, the conversion of phenol and indole side chains typically requires stringent reaction conditions. Electrochemical methods are noteworthy as the strategies enable efficient modification of Tyr and Trp residues under mild conditions, serving as a powerful tool for protein modification. |
676acca881d2151a02c71391 | 7 | (1) oxidative protein modification through electron transfer to a photocatalyst, (2) protein modification activated by the redox cycle of a photocatalyst, (3) protein modification via 1 O2 generation through energy transfer, and (4) protein modification through the activation of a labeling reagent by energy transfer. Photocatalysis, as a robust strategy for activating small molecules, has emerged at the forefront of organic chemistry, experiencing rapid development throughout the 2010s. In these approaches, metal complexes and organic dyes engage in singleelectron transfer (SET) reactions or energy transfer reactions with substrates, converting visible light into chemical energy. In traditional chemical reactions, the process generally proceeds in a "thermodynamically downhill" direction, making a reaction drive toward an energetically favorable direction through energy release. In contrast, photocatalytic reactions enable the introduction of light energy from an external source, raising the energy level of reactants to facilitate "thermodynamically uphill" reactions, which are typically more challenging to achieve. Recently, an increasing number of studies have focused on protein modification through photocatalysis, reflecting a growing interest in this area. While transformations of biomacromolecules using UV light date back several decades including applications in photoaffinity labeling chemistry, approaches utilizing photocatalysis present the advantage of achieving reaction control with visible light, which is considerably more biocompatible than UV light. Reviews have been published, focusing on various perspectives, such as classification based on structural characteristics of catalysts, the application of photocatalytic chemistry in diverse fields, and photochemistry on a broad range of proteins, including photo-click reactions, nucleic acid modifications, and photo-uncaging. In this section, on the other hand, we will focus specifically on applications of protein modification based on different reaction/catalysis mechanisms. |
676acca881d2151a02c71391 | 8 | This subsection introduces methods for oxidative protein labeling mediated by electron transfer reactions between photocatalysts and substrates (Scheme 3A, 3B). In this mechanism, the activated photocatalyst extracts an electron from the substrate, generating a radical species. Based on the oxidation potential of the photocatalyst, radicals can form on both labeling reagents and tyrosine (Tyr) residues of the protein. Due to high reactivities of these radical species in general, rapid bond formation often follows the excitation process. With the loss of an additional electron and proton, a reaction proceeds-in which, formally, two electrons and two protons are removed-resulting in the formation of either C-C or C-N bonds. |
676acca881d2151a02c71391 | 9 | The precise mechanism determining whether the labeling reagent or the Tyr residue on the protein undergoes radicalization, as well as the specific reaction intermediates involved, is still not fully understood and remains open to further investigation. However, in tyrosine modifications using tyramide, the radicalradical recombination pathway has been shown to predominate over other pathways between neutral Tyr residue and radical. Additionally, a report indicated that the reaction can proceed even under conditions where the radical species of the labeling reagent is generated electrochemically at potentials that do not directly radicalize Tyr. Several reagents have been reported for Tyr modification, including tyramide, phenylenediamine, N-methylurazole, and phenoxazine, while pyrazole has been shown to label phenylalanine (Phe) residues. In the catalytic mechanism, for instance, Ru catalysts can be photoexcited, and with the presence of oxidants such as oxygen or ammonium persulfate (APS), transition to a Ru(III) state. This Ru(III) species subsequently oxidizes the substrate via SET as it returns to its ground Ru(II) state. The use of Ru-based photocatalysts has long been known for Tyr-Tyr cross-linking reactions. More recently, it has become possible to label proteins by mimicking tyrosine residues with tyramide derivatives as labeling reagents. Tyramide conjugated with tags such as biotin (often referred to as "biotin phenol") enables visualization of labeled proteins and proteomics analysis through enrichment using avidin-beads. Additionally, methods have been explored for switching reactions using complexes with quencher molecules on DNA oligomers. Apart from tyramide, ethylenediamine-type labeling reagents have been utilized for radical modifications of tyrosine residues. This labeling reagent is also capable of labeling not only Tyr but also Cys, when free Cys residues are located near the catalyst binding site. Additionally, N-methylurazole has been developed as a radical-based labeling reagent specifically for labeling tyrosine residues in proximity to Ru complexes. The ligand structure of these catalysts has been examined from multiple perspectives, including the minimization of nonspecific adsorption to proteins, as well as enhancement of ligand binding with proteins. The importance of such proximity effects is described in the Supramolecular catalysis section as well. |
676acca881d2151a02c71391 | 10 | Covalent bond formation between the catalyst and protein has also been reported, with the Ru(TAP)₂phen²⁺ complex (TAP = 1,4,5,8-tetraazaphenanthrene; phen = 1,10-phenanthroline) enabling tryptophan (Trp) labeling via a SET reaction. For photocatalysts such as flavin, acriflavine, and 2,4,6triphenylpyrylium (TPT), it is postulated that the excited photocatalyst abstracts an electron from the substrate. Flavin, in particular, can accept two electrons and two protons to achieve its reduced H₂-Fl state, after which it regenerates its ground state by donating electrons to oxygen. Lumiflavin-based photocatalysts have been applied in selective Tyr modification of proteins using a phenoxazine dialdehyde tag. Flavin-catalyzed Tyr modifications with tyramide have also found applications in analyzing cell-cell interactions by enabling controlled reactivity on cell membrane surfaces. Additionally, acriflavine, with its higher cell membrane permeability compared to organometallic complexes, has been adapted for controlling reactions within cells. TPT, known for its high oxidation potential (+2.55 V vs saturated calomel electrode (SCE)), has even been reported to facilitate modifications of Phe residues, which are typically challenging to activate. |
676acca881d2151a02c71391 | 11 | In protein modification through redox processes by photocatalysts, an electron transfers from a substrate to an excited photocatalyst (Scheme 4A, 4B). This electron transfer results in oxidation of the substrate while the photocatalyst is reduced (cat -). Chemical modification of proteins can be achieved at the α-position of the C-terminus, the β-position of Trp residues, and the methyl group of Met residues as details are described in the following paragraphs. In these instances, it was proposed that radicals formed on the protein are captured by Michael acceptors. Namely, the radical intermediates receive an electron from the photocatalyst in its reduced state, forming bonds as the photocatalyst returns to its ground state. The C-terminus can be preferentially reduced via a singleelectron process at a lower potential compared to the carboxylates of aspartic acid (Asp) or glutamic acid (Glu) in proteins (E1/2 red : ~1.25 V for Asp, Glu; ~0. 95 ethylidenemalonate and 3-methylene-2-norbornanone. Selective modification of the C-terminus using flavin photocatalysts in various peptides and insulin has been documented. Furthermore, for peptide substrates, use of an iridium photocatalyst (Ir[dF(CF3)ppy]2(dtbbpy)PF6, E1/2 red : 1.21 V vs SCE) alongside a polyaromatic photocatalyst (4CzIPN, E1/2 red : 1.35 V vs SCE) for C-terminal alkynylation with ethynylbenziodoxolone (EBX) reagents, and conversion of the C-terminus to N,Oacetals for electrophilic activation have been reported. Unique to Trp modification is that labeling the β-position of the Trp side chain using Ir[dF(CF3)ppy]2(dtbbpy)PF6 with Michael acceptor modification has been accomplished. In a report of Met modification, lumiflavin (E1/2 red = 1.5 V vs SCE), that is capable of accepting an electron from Met (Epa = 1.36 V vs SCE), was utilized, as reduced lumiflavin (HLF • ) facilitates proton transfer from the Met radical cation (pKa = ~3.5), acting as a base (pKa of HLF • = 8.5) and catalyzing the transfer of electrons and protons. The extension of these photocatalysts' redox cycles for peptides substrates to protein bioconjugation is based on similar mechanisms, necessitating thorough evaluation of reaction conditions to ensure orthogonality and mitigate side reactions with Michael acceptors and residues such as Cys and Lys. |
676acca881d2151a02c71391 | 12 | An exceptional example of photocatalytic protein bioconjugation involves using a quinolinone chromophore-based photocatalyst to accelerate the thiol-ene reaction between Cys and terminal olefins through hydrogen-atom transfer catalysis. Activation of electrophilic species employing SET and combinations of SET and hydrogen atom transfer (HAT) for Cys modification and histidine (His) modification 72 are also noteworthy. |
676acca881d2151a02c71391 | 13 | Numerous photo-responsive molecules known as photosensitizers activate molecular oxygen to generate highly reactive singlet oxygen, 1 O2 (Scheme 5A). 1 O2, which has a lifespan of only microseconds in water and limited diffusion, leads to oxidation reactions in close proximity to the photosensitizer. Histidine residues are primary targets for 1 O2induced oxidation, undergoing Diels-Alder additions that form reactive endoperoxide intermediates on the imidazole rings. |
676acca881d2151a02c71391 | 14 | Mechanistic studies have shown that these reactions do not proceed through stepwise oxidation and nucleophilic addition (Scheme 5B), indicating that the formed reactive species are electrophilic. There was a demand for nucleophiles that can efficiently capture these active species, as such nucleophiles enable various tagging applications. According to a few studies comparing different nucleophiles, 3-ethynylaniline demonstrated higher reactivity compared to other anilines, amines, and phenylhydrazines. Additionally, 1-methyl-4-arylurazole (MAUra), with a pKa of 4.7, 76 predominantly exists in its anionic (N -) form at neutral pH, enhancing its nucleophilicity and thereby facilitating the efficient capture of oxidized histidine. This strategy for protein modification, known as proximity labeling (PL), exploits the proximity-dependent nature of photocatalysis and the brief lifespan of 1 O2. By generating 1 O2 and capturing oxidative intermediates nucleophilically (Scheme 5C), it facilitates a wide range of applications. These include identifying RNA-binding proteins, site-selective modification of antibodies, controlling reactivity within cells for subcellular proteomic mapping, analyzing histidine on aggregated proteins after catalytic photo-oxygenation, examining metalbinding histidines, controlling surface reactions for cell-cell interaction studies, and interactome analyses in live mouse brains. While this review primarily focuses on protein bioconjugation, it is notable that the concept of using photocatalysts to generate reactive oxygen species, thereby oxidizing and degrading proteins, has been well-established in techniques such as chromophore/fluorophore-assisted laser inactivation (CALI/FALI). Scheme 5. Protein modification utilizing 1 O2 generation from a photocatalyst. (A) Generation of 1 O2 through energy transfer reactions between an excited photosensitizer and oxygen molecules. (B) A technique involving the capture of electrophilic intermediates produced by Diels-Alder reactions between His residues and 1 O2 using nucleophiles. (C) The structure and reactivity of the nucleophiles used in this method. |
676acca881d2151a02c71391 | 15 | Developments have also been made in methods that transfer energy and activate labeling reagents (Scheme 6A). Such activation could proceed through Dexter energy transfer between an excited Ir-photocatalyst and a diazirine-based labeling reagent, for instance (Scheme 6B, 6C). This activation of diazirine in the proximity of the catalyst generates carbene, a highly reactive chemical species with a half-life of 2 ns. The photoactivation allows for precise control of protein modification reactions within a tightly restricted area less than 4 nm around the catalyst. The first-generation Ir photocatalyst based on Ir[dF(CF3)ppy]2(dtbbpy) (Ir-G1 cat) not only produces carbene but also activates arylazide, generating nitrene. This catalytic activity has been applied to analyze protein-protein interactions (PPI) on cell membrane surfaces, ligand binding site mapping, control labeling radius by altering labeling reagents, study binding proteins of sialylated glycoproteins, and dynamic analysis of phagocytic surfaces. Additionally, the second-generation Ir photocatalyst (Ir-G2 cat), which addresses cell membrane permeability issues of the first generation catalysts, has been employed for small molecular compound target identification, analysis of PPI in chromatin proteins, and application to stress granule components in cells. Activation of arylazide has been further explored using organic dyes such as Acridine Orange, fuorescein, rhodamine 123, a red light-activated osmium photocatalyst, and Sn IV chlorin e6 catalyst. These catalysts function through a SET mechanism rather than energy transfer, involving the reduction of arylazide in the presence of NADH through a stepwise reductiondissociation-oxidation pathway. |
676acca881d2151a02c71391 | 16 | As an alternative method, photocatalytic conversion of arylazide to aniline has been utilized in photo-uncaging techniques. This approach also produces o-thioquinone methide, an electrophilic species used for protein modification and subcellular proteomic mapping. 101 Scheme 6. Protein modification utilizing energy transfer from a photocatalyst. (A) Activation of modifiers through energy transfer from an excited photocatalyst. (B) Reactive species that can be generated by this method, capable of labeling various amino acid residues. (C) The structure of the photocatalysts used in this method. |
676acca881d2151a02c71391 | 17 | Metal-catalyzed protein modification has been enabled through a series of chemical strategies such as bioorthogonal chemistry and design/utilization of protein-compatible transition-metal complexes. A number of metal-mediated protein modification were achieved by use of noncanonical amino acids. For instance, copper-catalyzed azide-alkyne cycloaddition-a quintessential "click" reactionhas been widely used by introducing an azide or alkyne handle onto proteins through methods such as chemical modification and genetic/metabolic incorporation. Similarly, instillation of arylboronic acid or aryliodide allows for Suzuki-Miyaura coupling with a palladium catalyst. Olefin metathesis with a Grubbs-type ruthenium complex is also possible by introduction of noncanonical alkene groups with a thioether moiety. A large portion of metal-mediated modification of canonical amino acid residues, on the other hand, is often a non-catalytic system. Even for catalytic reactions of natural proteins, an excess amount of metal catalysts is necessary likely because of interaction of proteins with metal salts and because of challenges of catalysis in aqueous media. Indeed, there are many catalytic systems that uses organic solvents that are typically not compatible with protein substrates (but peptides). Examples described below tackled the challenges of protein modification in aqueous solutions by modulation of reactivity of a metal catalyst such as use of a coordinatively saturated complex to avoid undesired interactions with proteins. As there have been multiple review articles about metal-based protein bioconjugation in the past, this section will be focused on catalytic aspects of the processes. The metal catalysis leveraged by their Lewis acidic nature is discussed in the Acid catalysis section. |
676acca881d2151a02c71391 | 18 | Metallocarbene-and metallonitrene-based catalysis has been widely explored for chemical modification of natural proteins (Scheme 7A, 7B, 7C). The carbene and nitrene chemistry functions with the nucleophilic reactivity of amino acid side chains such as indole (Trp) and amine (Lys) through generation of electrophilic metal species from stable precursor molecules (e.g., diazo and sulfonamide compounds). Copper, rhodium, 118 and ruthenium 119 complexes have been shown to be useful for targeting various amino acid residues through metallocarbene intermediates (Scheme 7B). The majority of the metallocarbenebased strategies relied on discrete complexes such as a paddlewheel rhodium complex and ruthenium porphyrin complex, perhaps increasing the life time of the catalysts in aqueous solution in the presence of the nucleophilic biomolecules. Those metal complexes exhibits electrophilic nature, and the reactivity of the carbene complex can be enhanced through coordination of a buffer component to the metal center (Scheme 7D). Such a ligand binding can also cause the alteration of chemoselectivity of dirhodium carbene reactivity toward tryptophan to cysteine by use of a reagent bearing thioether group (i.e., biotin group). The rhodium carbenoid system can be coupled with a proximity-driven strategy to target many amino acid residues other than tryptophan (see the Supramolecular catalysis section). 122 More recently, copper nitrene complex was shown to act as a methionineselective protein modification method, where copper bromide salt and sulfonamide are the precursors of the nitrene complex. The unique chemoselectivity of the nitrene chemistry was ascribed to the thioether reactivity to the electrophilic metal center bound with acetonitrile ligands. As such, there has been a variety of carbeneand nitrene-based metal catalysis for protein modification since the initial report of the tryptophan-selective carbenoid method as one of the early examples of the modern protein bioconjugation study. In addition to the carbene and nitrene catalysis, a copper-catalyzed azide transfer reaction using sulfonyl azide to alkyl amine groups of proteins should be mentioned here, as the transfer mechanism has similarity to the diazo reagent preparation. 124 Scheme 7. Carbene and nitrene catalysis by copper, rhodium, and ruthenium complexes. (A) General reaction scheme. (B) Chemical structures of catalysts. (C) General depiction of mechanism of actions of catalysts. (D) Coordination of an additional ligand to a metal center as strategies to increase catalytic activity or alter chemoselectivity |
676acca881d2151a02c71391 | 19 | Cross-coupling reactions are another class of metal catalysis that has been one of the main focuses in the protein modification field (Scheme 8A, 8B). As various palladium cross-coupling methods emerged in the realm of synthetic organic chemistry during the past decades, the bioconjugation field has also been extensively examining their capabilities and utilities. The fundamental catalytic actions and reaction mechanisms of such cross-coupling bioconjugation approaches follows the same principle as that of the small molecule chemistry (e.g., oxidative addition, transmetallation, and reductive elimination). Often, electrophilic nature of a metal center after oxidative addition has been leveraged for both catalytic and noncatalytic protein bioconjugation strategies. A tyrosine-selective palladiumcatalyzed approach (Tsuji-Trost coupling) is one of the earliest examples among the palladium-based methods, driven by deprotonation of the phenol group making it a favorable nucleophile at high pH over other amino acid side chains. Another catalysis example is active-site cysteine selective through coordination of an auxiliary ligand to the palladium center. Other metal sources can be utilized for cross-coupling reactions of proteins including copper catalysis with use of boronate compounds as transmetallation reagents that modifies the amide backbone N-H group (Chan-Lam coupling). The original report by Ball and co-workers made use of a copper binding motif with a histidine residue-that is akin to known metalbinding peptide sequence called amino terminal copper and nickel (ATCUN) motif-to activate the backbone N-H and to facilitate the reductive elimination between the amide and boronate-derived group. Recently, use of a different solvent and different amino acid binding patterns 134 have been shown to achieve similar chemistry without aids of the histidine residues as well. Gold would be another metal source that allows for crosscoupling reactions on proteins targeting a tryptophan residue in aqueous acetonitrile solution. While this specific gold chemistry is the sole example of cross-coupling-based gold catalysis, several non-catalytic Au bioconjugation approaches have been reported to date, and together with its relatively lower toxicity compared to other transition metals, 138,139 development of gold catalysis may be merely in a nascent state for potential growth. |
676acca881d2151a02c71391 | 20 | While redox processes are often utilized for protein modification processes, iridium hydride-based catalytic reductive alkylation remains the sole example of the catalytic reduction for protein bioconjugation purposes in this metal catalysis section (Scheme 9A, 9B). Analogous to the traditional protein modification strategy using aldehyde and cyanoborohydride reagents, 141 the iridium-based hydride reduction would occur through the Schiff base formation from the protein amines followed by the reduction of the Schiff base to amine. The catalytic system reported by Francis and co-workers utilized activation of a pro-catalyst, watersoluble iridium Cp* complex through reduction with sodium formate (Scheme 9B). Electron-rich bipyridine ligands were found to be particularly effective, which perhaps is an indication of importance of nucleophilicity of the metal hydride species. While this iridium catalysis has been used in other reports, there has not been other metal-catalyzed reductive approaches developed for protein bioconjugation. The lack of the development may be due to the challenges of retention of reduction-sensitive S-S bonds in proteins, although mild and/or bulky reductants such as ascorbate and triarylphosphine 144 have been successfully utilized in the protein labeling strategies. Reduction-based metal catalysis may grow dramatically for protein modification fields, as reductive metal catalysis (e.g., metal hydride chemistry) has been useful in many chemical biology applications. Scheme 9. Iridium-catalyzed reductive alkylation/amination. (A) General reaction scheme. (B) The structure of the procatalyst (left) and active catalyst after reduction with sodium formate (right). |
676acca881d2151a02c71391 | 21 | Catalytic oxidation processes can be applied for protein modification through activation of amino acid side chains and labeling reagents (Scheme 10A, 10B). As oxidation of proteins is one of the fundamental processes in living systems, there is a range of oxidative reactions used for protein modification in noncatalytic manners. The two major approaches for oxidative catalysis for protein modification occurs through activation of either amino acid side chains (e.g., thiyl radical generation) or labeling reagents (e.g., diazene generation). An example includes alkenylation of a cysteine residue by a gold-and silver-mediated system using allene-based labeling reagents. The gold catalyst was proposed to be useful for not only the single electron oxidation of cysteine, but also activation of allene labeling reagent as Lewis acid (Scheme 10C). The necessity of silver triflate additive may limit the utility of this catalysis, as a silver salt is known to induce precipitation of proteins (e.g., common staining protocols for protein gel known as silver staining). An oxidation process can be utilized for catalytic activation of labeling reagents, as the hemin-catalyzed tyrosine modification takes advantage of such a mechanism. Inspired by biological oxidation processes of luminol in firefly chemiluminescence, this approach proceeds by formation of tyrosine-reactive diazene reagent through oxidation of N-N bond to N=N bond. A previous approach of tyrosine-selective modification by Barbas and coworkers necessitated preparation of an unstable diazene reagent prior to protein modification processes, and this catalytic oxidation strategy omits the technical challenges through in situ generation of the active species. It is interesting that the most effective reagent proved to possess N-methyl group that should not be able to form neutral N=N bond species as proposed (Scheme 10C). The high activity of the N-methyl reagent may be indicative of potential involvement of single-electron oxidation by the iron complex. |
676acca881d2151a02c71391 | 22 | Oxidative cleavage of the peptide backbone can be induced through metal catalysis, mimicking enzymatic processes in natural systems (Scheme 11A). Reactive oxygen species (ROS) is important biological species both for physiological and pathological conditions, where protein oxidation play pivotal roles. Through sophisticated design of a catalytic system with judicious choice of reaction components (catalysts, ligands, and oxidants), Oisaki, Kanai, and co-workers developed coppermediated backbone cleavage, selectively at serine residues (Scheme 11B). Single-electron oxidation of the primary alkylalcohol of serine through copper(II)-phenathroline complex and N-oxide reagent initiates the catalysis, and subsequent oxidation of the generated aldehyde group produces a hydrolytically unstable imide intermediate that eventually undergoes the bond cleavage. Though minor reactions at threonine (secondary alkylalcohol) would occur, the serineselective cleavage was achieved even for a small protein ubiquitin through this catalytic system. Another oxidative cleavage strategy is by a copper cluster that can generate ROS species with ascorbic acid to cause site-specific cleavage of lysozyme an enzyme (Scheme 11B). The three-dimensional structure of the copper cluster was attributed to the observed site-specific cleavage of the enzyme through binding interaction between the cluster and enzyme. As those reports showed a single example of a protein substrate for the catalytic cleavage, the future directions of the field are likely to expand the generality and scope of the methods to achieve enzyme-like catalysis. Serine-selective cleavage by a copper complex (top) and site-selective cleavage of lysozyme through reactive oxygen species (ROS) generation with a copper-and tungsten-based cluster (Cu-WD). The imide intermediate of the serine-selective cleavage can be hydrolyzed at both imide C=O groups (blue and pink), but only one of the two possible hydrolysis products is shown for the sake of simplicity. The image of ROS-mediated cleavage was reprinted with permission. Copyright 2023, American Chemical Society (). |
676acca881d2151a02c71391 | 23 | As acid-catalyzed processes are ubiquitous in enzyme active sites, numerous protein bioconjugation approaches have been also leveraged by a range of acid catalysts. In natural systems, enzymatic catalysis often depends on acid-mediated activation of weak electrophiles in proteins including proteolysis of amide backbones through interaction of acids to the carbonyl groups. While such enzymatic catalysis could often be substrate-specific processes, acid-catalyzed chemical modification of proteins can offer broader substrate scope with a potentially unique reactivity and selectivity paradigm. One of the traditional approaches for acid-catalyzed protein modification is to employ strong Brønsted or Lewis acid (e.g., zirconium (IV) chloride-derived acid and perchloric acid) to enable reactions of weak electrophiles such as amides and carboxylic acids, although such harsh conditions may not be compatible with many protein substrates. More recently, various chemical strategies (e.g., sophisticated ligand design, proximity-accelerated catalysis through reversible covalent bond formation, and nonaqueous systems) have been devised to overcome the challenges, as described in the following sections. It is noteworthy that the development of many of those acid-catalyzed protein bioconjugation methods has been driven by knowledge of synthetic organic chemistry including Lewis acid strengths, 163 metal affinity, and unique solvent properties. This section discuss catalysis by both Brønsted acid and Lewis acid including metal Lewis acid. |
676acca881d2151a02c71391 | 24 | A zinc salt was employed as Lewis acid for lipidation of cysteine through the catalytic activation of both nucleophiles and electrophiles (Scheme 12A, 12B). 168 S-lipidation is a naturally occurring post-translational modification that is relevant to various cell signaling events including synaptic transmission and GPCR protein signaling. In the report by Fairlie and co-workers achieving a chemical way for S-lipidation, zinc ions played catalytic roles in the SN2 reaction between thiol groups on cysteine residues and alkyl halides that contain fatty acid moieties. The zinc catalyst was suggested to have dual functions: increase of nucleophilicity of the thiol group and increase of electrophilicity of the alkyl halide. In other words, the catalyst would not only interact with the halide leaving group to enhance the electrophilicity of the alkyl halide, the nucleophile of cysteine residue can also be activated by the catalyst interaction lowering pKa of SH to thiolate ion. This design was perhaps inspired by a similar phenomenon found in natural zinc-containing enzymes (e.g. zinc-dependent transferases and zinc finger proteins ). As the catalysis proceeds with other divalent ions such as Ni and Cd 2+ , the affinity of the metal ions to the thiol group might be playing key roles in the system. Hexafluoroisopropanol (HFIP) can facilitate electrophilic aromatic substitution of tryptophan through activation of thiophene-ethanol labeling reagent by a metal, Lewis acidic catalyst (Scheme 12A, 12B). 172 HFIP, like other fluoroalcohols such as trifluoroethanol (TFE), is known to induce α-helical structures of polypeptide, and protein substrates may not always tolerate such conditions. The recent work by Ohata and co-workers demonstrated that increased protein compatibility of HFIP by ionic liquid additives. For instance, an anti-HER2 antibody, trastuzumab was shown to lose its selective antigenbinding activity after treatment in HFIP, but the activity and selectivity was retained when the antibody was treated with HFIP containing ionic liquids. Because HFIP and other fluoroalcohols are increasingly used for biomolecule modification recently, 111 the potential compatibility of the solvents with proteins motivates their application for catalytic protein bioconjugation. The HFIP-based bioconjugation work by Ohata and co-workers took advantage of Lewis-acid-catalyzed dehydrative alkylation reaction (Friedel-Crafts type process) 178 of tryptophan residues. Catalytic actions of the Lewis acid such as scandium ions were studied using density functional theory (DFT) calculations, suggesting that the acidity of HFIP is increased through the coordination of HFIP to scandium. The increased acidity of HFIP by the coordination was shown to cause protonation and liberation of the OH group of the labeling reagent (thiophene-ethanol) as shown in Scheme 12B. In other words, the preliminary computational study indicated HFIP as a proton donor for the dehydration process, and consistent with the observation, the same group also demonstrated use of Brønsted acid to catalyze the process as well. Although aqueous media have been often considered a requirement for useful protein bioconjugation methods, 180 this example in addition to a carboxylic acid-based serine modification method described below (the acid-catalyzed hydrolysis and alcoholysis section) may suggest practical usefulness of nonaqueous approaches. |
676acca881d2151a02c71391 | 25 | Acyl imidazole derivatives have been applied as activitybased sensing probes to detect intracellular metal ions by leveraging catalytic activity of the ions (Scheme 12A, 12B). Metal ions serve as chemical signals in cells, and chemical probes that detect these ions are useful for understanding their biological functions. In particular, reactivity-based sensing approaches offer advantages that traditional reversible binding sensors do not possess. An acyl transfer-based approach originally developed by Hamachi and co-workers relies on coordination of tetradentate acyl imidazole derivatives to metal ions (Lewis acid), which makes the carbonyl group more electrophilic to trigger a substitution reaction of the complex with nucleophiles on intracellular proteins. The original report developed a zinc-selective probe with dipicolylamine as the zinc-binding site. Although the binding affinity of the probe to Zn 2+ was shown to be quite strong (0.7 nM) for this particular system, the approach could function as catalysis as long as there is a certain degree of dissociation process. Later, Chang and co-workers designed a copper-selective sensor through the incorporation of thioethers enabling detection of labile brain copper. The same group recently applied the catalytic system to multiplex imaging of Cu(I) and Cu(II) by stimulated Raman scattering using isotopically lableled nitrile vibration tags. The development of the zinc-and copper-targeting acyl imidazole sensors indicates that ligand design would be able to modulate the binding selectivity to produce new catalytic chemical probes toward different intracellular metal ions. |
676acca881d2151a02c71391 | 26 | Lewis acidic metal complexes can serve as a catalyst to cleave amide bonds in proteins at specific amino acid sequences (Scheme 13A, 13B). Proteases such as trypsin or pepsin are enzymes to catalyze proteolysis (i.e., hydrolysis of the peptide backbones) and are widely used for proteomic studies. Although the proteolysis processes with such natural enzymes can proceed efficiently in mild conditions, their exclusive cleavage recognition patterns (e.g., Lys and Arg for trypsin) are not always compatible with a peptide or protein of interest, and development of non-natural, artificial proteolytic systems with alternative sequence recognition patterns has been actively studied. Metal catalysts that displays Lewis acidic nature can be used as alternatives to natural proteases, for example. The metal center of such Lewis acid catalysts would facilitate the hydrolysis of the amide backbone by interaction with the carbonyl oxygen of the amide backbone making the carbonyl group more susceptible to the nucleophilic addition. An example includes a report by Kanai and co-workers showing scandium (III) triflate catalyst-mediated serine/threonine selective cleavage. Kostic and co-workers reported a different type of amide backbone cleavage with histidine and methionine selectivity using palladium-based catalysts. In addition, N-terminal residue-selective cleavage through the chelation of the α-amino group can be achieved with cobalt(III) complex as well (Scheme 13B, 13C). In order to facilitate selective and efficient cleavage process by a metal complex, Suh and co-workers developed a system making use of reversible imine formation between aldehyde tethered to the metal-ligand and amine on proteins, which enabled specific cleavage at Gln(91)-Ser (92) and Ala( )-Thr( ) of myoglobin protein (Scheme 13B, 13C). Such a reversible bond-forming process can be also applicable to site-specific installation of functionality as well. Instead of the reversible covalent-bond formation, supramolecular catalysis has been leveraged to induce the proximity-driven effects for peptide cleavage as well. Those catalytic proteolysis examples demonstrated that choice of metals and ligand design can produce enzyme-like catalytic systems, and there have been many other reports of different metal systems as described in recent review articles. Serine residues in proteins can be catalytically modified in carboxylic acid-based non-aqueous media (Scheme 13C). Although serine undergoes a variety of enzymatic modifications in living systems, 203 chemical modification of serine remains to be one of the challenging tasks owing to the modest nucleophilicity of the side chain and the abundance of the OH groups in aqueous media. Encourage by the catalytic tryptophan labeling in nonaqueous medium (i.e., hexafluoroisopropanol or HFIP) as described above, Ohata and co-workers described that carboxylic acids could serve as a potentially protein-compatible reaction medium for serine-targeting modification, 202 which was motivated by the widespread use of many carboxylic acid-based compounds as biocompatible buffer components in biochemistry and protein science (e.g., acetate, glycine, and citrate buffers). The chemical modification of hydroxyl groups of serine residue proceeds with an acid catalyst (e.g., trifluoroacetic acid and dysprosium(III) triflate) in carboxylic acid media where the excess carboxylic acid serves as an electrophile and reacts with the hydroxyl group (i.e., Fischer esterification-type reaction). The method was shown to be able to label protein substrates as well, including an intact antibody (trastuzumab), concanavalin A, and chymotrypsin. As chemoselective serine labeling strategies have been simply lacking, this acylation chemistry may be of indicative of the power of the nonaqueous, catalytic approach for protein modification. |
676acca881d2151a02c71391 | 27 | Scheme 13. Acid-catalyzed hydrolysis and alcoholysis. (A) A general reaction scheme of acid-catalyzed hydrolysis/alcoholysis. (B) Structures of representative metal catalysts that mediate protein-backbone cleavage. (C) Chemical strategies that enhance the catalytic hydrolysis and alcoholysis. Top: Backbone hydrolysis accelerated by the copper center as Lewis acid and reversible formation of imine inducing proximity-driven effects. Bottom left: Nterminal specific cleavage by a CoIII catalyst. Bottom right: Catalytic acylation of alkylalcohols on proteins accelerated by an acid-activated carboxylic acid used as a solvent. |
676acca881d2151a02c71391 | 28 | Aldol reactions using a dual catalytic system of copper and aldehyde catalysts enabled labeling of protein N-termini (Scheme 14A). Site-specific protein labeling methods can be advantageous for producing well-defined protein conjugate compared to chemoselective approaches and are useful for various applications such as single-molecule localization microscopy, 206 preparation of polyethylene glycol-tagged (PEGylated) therapeutic proteins, 207 and production of antibody-drug conjugates. 208 N-terminal amino groups can be attractive sitespecific modification handles, as its decreased basicity compared to lysine amines can be utilized for N-terminal labeling through pH control. While there is a repertoire of N-terminal selective methods to date, Hanaya and co-workers reported a catalytic variant by utilization of copper and aldehyde catalysts activating N-terminal α-proton. The copper-catalyzed aldol reaction was proposed to occur through activation of an N-terminal amino acid by the copper and aldehyde catalysts, forming a nucleophilic Cu(II)-enolate intermediate (Scheme 14B). This nucleophilic activation of the protein would be followed by electrophilic activation of another aldehyde molecule, eventually leading to aldol-type reactions between the activated species to produce the product with a stable C-C bond. Interestingly, in contrast to other acid catalysis earlier, other metal catalysts such as Sc(OTf)3 did not function as effective catalysts for the aldol reaction, perhaps indicating the importance of the subtle control of Lewis acidity and affinity toward certain ligands in this catalytic system. The aldehyde catalyst/reagent (2-pyridinecarboxaldehyde) was previously reported to form a hydrolytically unstable imidazolidinone product on N-termini of proteins, and the catalytic aldol process can be useful to produce more stable reaction products by alteration of the outcome of the aldehyde reaction with N-terminal amines by introduction of the copper catalyst (Scheme 14C). Whereas the hydrolytically unstable N-terminal product can be of use for reversible elimination of the chemical modification in a certain context, 212 this catalytic example represents the importance to develop an alternative approach for production of protein conjugate with a different property. The same research group recently reported another copper(II)-mediated N-terminal modification by leveraging the hydrolytically stable intermediate (Scheme 14A). Kanemoto and co-workers reported copper-catalyzed [3+2] cycloaddition between metalated azomethine ylide (bidentate) on the glycine Nterminus of peptides and maleimides in organic solvents. Hanaya and co-workers expanded the scope of this approach to proteins by the use of pyridyl-aldehydes, allowing reactions in the aqueous buffer through formation of a more stable intermediate, tridentate azomethine ylide (Scheme 14B, 14C). The reaction was applicable to peptides and proteins with various N-terminal amino acid residues. Site-specific modification of an antibody, trastuzumab with this modification method prepared antibodydrug conjugates with uniform drug-antibody ratio (DAR), which was applied for mice cancer models. Scheme 14. Dual catalytic aldol reaction (left) and copper-catalyzed cycloaddition (right). (A) Schematic illustrations of copper-/aldehyde-catalyzed aldol-reaction (left) and copper-catalyzed cycloaddition (right). The pyridyl-aldehyde acts as both the reagent and catalyst for the aldol reaction. (B) Catalysis mechanism of the copper-catalyzed aldol reaction (left) and copper-catalyzed cycloaddition (right). The pyridyl-aldehyde acting as a catalyst is depicted in red and the one as a regent in black. (C) Comparison of reaction products for N-terminal modification by choice of a labeling reagent and catalyst. Top: Alteration of the reaction product by the absence (reversible condensation reaction, 211,212 left) and presence (Irreversible aldol reaction, right) of the copper catalyst. Bottom: Irreversible copper-catalyzed cycloaddition. |
676acca881d2151a02c71391 | 29 | Organocatalytic protein bioconjugation has been achieved by modifying target proteins through organic ligand-assisted reactivity. Organocatalysis has been rapidly growing in the synthetic chemistry fields since the late 1990s. One of the major motivations for development of such catalytic strategies is to overcome the limitations of existing metal catalysts such as toxicity and high cost. For example, proline is a readily available, non-toxic natural amino acid that serves as a catalyst for aldol reaction and Mannich reaction. Even though a wide variety of organocatalytic transformation has been reported for small molecule substrates decades after its inception, it is interesting that protein bioconjugation driven by organocatalysis has been exclusively by acyl transfer reactions through proximityaccelerated chemistry to date. As also described in the supramolecular catalysis section, such proximity-driven chemistry functions through association and dissociation of an affinity ligand that interacts with a protein of interests. The reported organocatalytic bioconjugation processes generally proceed in two steps (Scheme 15A): ( ) association of a ligand to a target protein accelerates nucleophilic attack of ligand-tethered nucleophilic catalyst to an acyl donor reagent. ( ) Another set of nucleophilic attacks by proteins to the activated acyl donor occurs, followed by dissociation of the ligand-catalyst. The following paragraphs focus on evolution of acyl transfer-based organocatalytic bioconjugation methods and brief demonstration of their utilities and applications. |
676acca881d2151a02c71391 | 30 | Even in the first-generation labeling systems, organocatalytic acylation reactions through protein-ligand interaction demonstrated their usefulness for site-specific modification of target proteins for live cells and tissue samples (Scheme 15A). The seminal work of the organocatalytic bioconjugation was by Hamachi and co-workers, which demonstrated acylation of a glycoprotein-binding protein, lectin with a dialkylaminopyridinebased catalyst (i.e., dimethylaminopyridine or DMAP-type catalyst, Scheme 15B). The DMAP catalyst was tethered with saccharide-ligand, and thiophenyl esters were employed as acyl donors for the site-specific labeling of lectins. By the virtue of proximity-driven effects, target proteins can be selectively modified even in the presence of other proteins (e.g., in cell and tissue lysates). In addition to saccharide ligands, the DMAPbased catalyst can be conjugated to proteins that bind to protein targets; for instance, DMAP-tethered lectin was utilized for the labeling of glycoproteins on live cell surfaces. As a similar approach, a DMAP-tethered antibody fragment was developed to selectively modify receptors on cell membranes, which enabled epitope mapping of antibodies. Another early example of ligand-directed catalysts is an organocatalyst based on dimethylalkylamine tethered with biotin, which was used for modification of carboxyl groups of aspartate and glutamate residues of avidin. More efficient catalysis at physiological pH than the firstgeneration DMAP-based catalysis was demonstrated with anionic catalysts and thiol-based catalysts with milder acyl donors (Scheme 15B). One of the challenges in the DMAP catalysis is that pKa of the conjugate acid of the DMAP-based moiety is 8.6, 228 and substantial portions of the catalyst could be protonated at physiological pH ranges. To this end, pyrydinium oximes and hydroxamic acids were proposed as alternatives because of their lower pKas (6.6 230 and 6.5 , respectively) that could allow faster protein labeling than DMAP at physiological pH. Thiol-based catalysts possess additional benefits, compared to DMAP-based catalysts, that are kinetically favorable thiol-thioester exchange between the catalyst and the acyl donor. As an independent approach from the thiol-based catalysis, Kanai and co-workers developed a unique proximity-driven approach through reversible boronate formation to facilitate the organocatalytic modification for site-specific modification of target within live cells (Scheme 15B). 229 Such a range of the second-generation catalysts allowed the use of moderate electrophilic acyl donors (e.g., acetyl-CoA) compared to the one used with DMAP-based catalysts (thioesters derived from thiophenol), attenuating off-target labeling. For example, in the DMAP-based catalysis, the reaction usually requires a low temperature (e.g., 4 °C) to minimize the nonspecific labeling arising from the high electrophilicity of the thioester acyl donor. Therefore, the use of milder acyl donors such as N-acyl-N-alkyl sulfonamides, 227 alkyl thioesters, and acyl imidazoles 236 can be beneficial through suppression of the labeling agents' off-target reactivities. A notable application of the second-generation approach is that a hydroxamic acid-thiolbased catalyst conjugated with histone-ligand has been used for acylation of lysine-120 (K120) of histone H2B with endogenous acyl-CoA as an acyl donor. Since acetylation of lysine residues of histone proteins is a naturally occurring post-translational modification that regulates gene expression (epigenetic regulation), 237 this approach could be useful for the chemical manipulation of the epigenetic regulation of gene expression. |
676acca881d2151a02c71391 | 31 | Supramolecular chemistry-based strategies such as metal-anion interactions, macrocyclic self-assembly, and ligand-directed affinity labeling can promote proximity-induced reactivities for protein bioconjugation (Scheme 16A). Supramolecular chemistry pertains to molecular assembly through a range of non-covalent interactions. The process can be viewed as host-guest chemistry that causes a covalent bond-forming reaction between a substrate and reagent through supramolecular forces including hydrophobic interaction, hydrogen bond, van der Waals force, ππ stacking, and ion-dipole effect. Enzymatic systems make use of a number of types of supramolecular chemistry to achieve site-selective modification. For example, a transpeptidase sortase utilizes its domain called β6/β7 loop that recognizes an LPXTG (X = D, E, A, N, Q, or K) motif of a target peptide/proteins for siteselective modification of their C-terminal positions. Supramolecular chemistry has been used for non-catalytic bioconjugation reactions including metal-anion ionic interactions by His6-tag/Ni system and DNA/RNA hybridization through base pairings, as such non-catalytic supramolecular bioconjugation reactions have been reviewed in a recent article. As described below, artificial host molecules such as gallium cluster and cucurbituril as well as ligand-directed affinity labeling strategy (see the organocatalyst section) were used as catalytic protein labeling strategies. |
676acca881d2151a02c71391 | 32 | Lysine-selective reductive amination can be achieved catalytically by a supramolecular gallium cluster (Scheme 16B). For small molecule substrates, a gallium cluster Ga4L6 (L = N,N′-bis(2,3dihydroxybenzoyl)-1,5-diaminonaphthalene) host were known to cause catalytic Nazarov cyclization and aza-Darzens reactions through assembly of reaction components by hydrophobic effect and stabilization of cationic intermediates (i.e., electrostatic stabilization of cations by the polycationic host Ga4L6). Even for peptide and protein bioconjugation purposes, the same anionic gallium cluster can serve as a supramolecular host to catalyze reductive amination reactions on alkylamine groups of lysine residues. The specificity to lysine residues for the supramolecular catalysis contrasts with traditional reductive amination with a borohydride reagent (e.g., NaCNBH3), which often cannot differentiate lysine and N-terminal amines unless there is precise pH control. The selectivity mechanism was not studied in this report, but the supramolecular catalyst's preference toward sterically more accessible ε-amine may be the reason for the observed phenomena. |
676acca881d2151a02c71391 | 33 | A macropolycyclic catalyst, cucurbit [8]uril can induce proximitydriven Michael addition through the assembly of a tryptophan residue and bipyridinium derivative (Scheme 16B). Cucurbituril is a macropolycyclic compound assembled from glycoluril and formaldehyde. Cucurbituril acts as a host molecule for various guests such as hydrocarbons, saccharides, dyes, amino acids, and proteins through hydrophobic interactions, ion-dipole interactions, and dipole-dipole interactions. Through such supramolecular capability, cucurbituril-mediated chemistry has been employed for protein modification such as azide-alkyne cycloaddition reactions enhanced through the hydrogen bonding network. For modification of proteins with natural amino acid side chains, cucurbit [8]uril-a cucurbituril that contains eight glycoluril units-was used for supramolecular catalysis-based Michael addition reaction modifying a cysteine residue through inclusion of both a bipyridinium group on the labeling reagent and a tryptophan residue simultaneously. Through this approach, cucurbit [8]uril-facilitated modification of a KRas protein was achieved using dehydroalanine-bipyridinium reagent; cucurbit [8]uril served as a host to a tryptophan residue of KRas and bipyridinium for labeling, followed by a proximity-induced thia-Michael reaction between cysteine residue on the substrate and dehydroalanine on the labeling reagent. It should be also noted that supramolecular catalysis-based chemical backbone hydrolysis has been also achieved using a polymacrocyclic catalyst, although their substrate scope is limited to peptides. 197 |
676acca881d2151a02c71391 | 34 | Protein-ligand interactions have been utilized for a variety of catalytic transformation by enhancing inherently slow reactions through proximity-driven effects (Scheme 16B). Early examples of ligand-directed labeling strategies (non-catalytic) were shown to modify active sites of enzymes and antibody binding sites. Such early examples were not catalytic, as the bound ligand does not dissociate after the labeling process. More recently, catalytic ligand-directed strategies have been reported where types of ligands range from small molecules to peptide ligands, and various catalysis, as described in the previous sections (i.e., photocatalysis, transition-metal catalysis, and organocatalysis). The proximity effect is often utilized for sluggish reactions that do not proceed without the rate enhancement mechanism, including examples of ruthenium photocatalysis and DMAP-based organocatalysis. In other words, generally unreactive amino acid residues could be modified through the proximity effects, as modification of phenylalanine was achieved by dirhodium catalyst conjugated with STAT3 ligand. The proximity-driven rhodium catalysis was indeed shown to be capable of modification of half of canonical amino acids, including asparagine, phenylalanine, glutamine, and threonine, showcasing the power of the supramolecular chemistry. Scheme 16. Supramolecular catalysis. (A) Schematic illustration and mechanism of the protein modification with supramolecular catalysts. (B) Structure of supramolecular catalysts and catalysis mechanisms. Top: Reductive amination catalyzed through the supramolecular interaction between the gallium cluster, pyridine-borane, and imine intermediate. Middle: Thia-Michael addition to a cysteine residue facilitated by the interaction between a tryptophan residue of a protein and bipyridinium unit of the labeling reagent in the cucurbit [8]uril (CB [8]) as a supramolecular host. The image of X-ray crystal structure on the right was reproduced with permission. Copyright 2000, American Chemical Society. Bottom: Proximity-driven catalysis through binding of a ligand to target proteins. |
676acca881d2151a02c71391 | 35 | Heterogeneous catalysts could be beneficial for proteolysis applications due to their tunable properties and facile separation from products (Scheme 17A). One of the earliest documented heterogeneous catalysis is Faraday's oxidation reactions by platinum catalysts in the 1800s. The advantages of heterogeneous catalysts are reusability and easier separation of catalysts. While there are many heterogeneous catalytic systems for small molecule substrates, catalysts that can be applicable for protein substrates are quite scarce. As described in the following paragraphs, metal-organic framework (MOF)based platforms are one of the few examples that act as heterogeneous catalysis for protein substrates. MOFs are crystalline materials composed of metal ions and organic ligands. Properties of MOFs such as pore size, type of metal ions, and surface area are tunable. As described in the acid catalyst section, artificial proteases composed of non-biomolecule building blocks could be useful for the digestion of proteins because of unique cleavage sites of such approaches. While MOF catalysts are also useful for the same purposes (i.e., hydrolysis of the protein backbones), the protein-hosting ability of MOFs can be an additional advantage for the following two reasons: 260 (1) The space confinement effect of MOF mesopores could mimic enzymes active site, as demonstrated in a few reports. (2) The heterogeneous nature of MOFs can be advantageous for proteomics digestion purposes, as the catalysts can easily be separated from products after the reaction. As free metal ions could also be a catalyst for given reaction systems, 200 it is often important that a MOF catalyst possesses chemical and structural stability, so that fragments/components of MOF through decomposition would not induce unwanted processes. For instance, one of the early reports for MOF-catalyzed proteolysis described leaching of Cu(II) ions from the MOF, 260 although the Cu(II) ion was not ascribed to the catalytic activity of the system. |
676acca881d2151a02c71391 | 36 | Zirconium-based MOF and metal-oxo clusters were demonstrated to mediate backbone cleavage of proteins through their Lewis acidic actions (Scheme 17B, 17C). Parac-Vogt and co-workers introduced a zirconium(IV)-based MOF as a heterogeneous catalyst for proteolysis of model protein substrates. The hydrolysis was proposed to proceed through the activation of the amide backbones by Lewis acidic Zr(IV) centers incorporated in the MOF catalyst. Hexazirconium metaloxo cluster was a building block of the specific MOF catalyst, which was assembled by capping and interconnection by six benzene-1,3,5-tricarboxylate linker in a trigonal antiprism fashion. The water-soluble metal-oxo cluster, [Zr6O4(OH)4(CH3CO2)8(H2O)2Cl3] + showed superior hydrolytic activity compared to the zirconium MOF catalyst. 268 Some metaloxo cluster-based catalysts (e.g., cerium-based polyoxometalate 269 and molybdenum-based polyoxometalate 270 ) showed unique properties such as regioselectivity potentially arising from the enzyme-like non-covalent interaction with specific regions of proteins. Another explanation for such enzymatic behaviors is that the catalysts might make cleavage sites more accessible through partial unfolding of protein structures of proteins. As the hexazirconium metal-oxo cluster is a component of the MOFs, 268 the metal-oxo cluster component has been used for mechanistic studies showing the accessibility of the catalyst active site by protein substrates and its similarity of the catalytic action to the hafnium cluster as described below. |
676acca881d2151a02c71391 | 37 | Multinuclear hafnium metal-oxo cluster, [Hf18O10(OH)26(SO4)13•(H2O)33], would be an effective heterogeneous catalyst for proteolysis specifically at aspartic acid residues (Scheme 17B, 17C). 275 there have been several reports of Hf(IV)-based MOFs as heterogeneous catalysts for organic reactions that involve activation of carbonyl groups. The Hf18 polynuclear cluster [Hf18O10(OH)26(SO4)13•(H2O)33] is insoluble in water and possesses both Lewis and Brønsted acidic moieties (i.e., protic protons of the coordinated water on the Hf centers) that was shown to facilitate hydrolysis of the amide backbones of a protein. The catalysis displayed selective cleavage at aspartate residues (both Asp-Xxx and Xxx-Asp bonds where Xxx is an arbitrary amino acid residue), and the proposed reaction mechanism is by nucleophilic attack of the aspartate COOH to the amide backbone that forms an anhydride intermediate for Xxx-Asp bond cleavage and an imide intermediate for Asp-Xxx bond cleavage, followed by another set of nucleophilic attack by a water molecule to completes the process. It should be noted that presumably due to the Brønsted acidity of the catalyst, negatively charged regions of proteins were effectively cleaved, which was not achieved with the Zr(IV)-based metal-oxo cluster catalyst. Therefore, hafnium-based catalysis may offer an alternative selectivity for proteolysis applications. |
676acca881d2151a02c71391 | 38 | Catalytic transformations have shown the great utility in chemical modification of proteins from various viewpoints, and the collection and analysis of a set of literature in this review also underscored possible future directions of the realm of the catalytic bioconjugation research. Appearance of redox-mediated chemistry in a multitude of sections is notable, probably implying its growing interests across various catalysis fields. However, because oxidation and reduction reactions are common processes in several canonical amino acid side chains, 279 strategies to suppress unwanted side reactions may be the unavoidable tasks as literature precedents also tackled the issue already (e.g., reactions under inert atmosphere 280 and use of redox-sensitive additives 281 ). It became also obvious that some catalytic strategies are limited to only a certain reaction type or virtually non-existent for protein bioconjugation purposes. For instance, organocatalysis and heterogeneous catalysis have been realized only through acyl-transfer reactions and backbone hydrolysis, respectively. We were unable to find any examples of asymmetric catalysis and mechanochemical catalysis that have been used for bioconjugation at a protein level, though potential usefulness of some reactions for small molecule substrates in this areas are indicated in recent literature. In particular, the dearth of the asymmetric catalysis is striking given the enantiomeric/geometric importance of post-translational modification in nature (e.g., methionine oxidation 284 and lysine acetylation 285 ) as well as increasing studies on d-amino acid/dproteins. Dehydroalanine functionalization would be an example highlighting this trend, as there have not been examples of protein bioconjugation utilizing dehydroalanine in an asymmetric fashion even if a plethora of reports demonstrated asymmetric conjugate addition reactions at a small molecule level. Plausibly, this challenge may have been exacerbated by limited availability of analytical techniques that can be usable to differentiate the isomer forms of a particular amino acid residue in protein substrates. The examination of literature also showed the power of dual catalytic systems, especially those combined with supramolecular catalysis for sitespecific or target-specific labeling strategies. Indeed, hybrid catalytic systems have been frequently utilized in chemistry of small molecules and peptides, 289 and protein bioconjugation may benefit from such hybrid systems as well. Finally, it is noteworthy that some catalytic mechanisms would be only possible in polypeptide substrates but not simple small molecule substrates (e.g., copper-catalyzed backbone modification of amide N-H driven by a neighboring histidine residue 132 ), and proteins may serve as a platform to expand the boundary of the catalysis domain beyond the small molecule chemistry. Diverse fields spanning bioorganic chemistry, chemical biology, biomedical science, and material science necessitate development of protein bioconjugation to address various scientific and pragmatic challenges, and the chemical strategies and principles mentioned in this review paper may be a catalyst to transcend the limit. |
67d02e0a6dde43c908694072 | 0 | ). Below the scheme, photographs show 1a sandwiched between quartz plates before and after the reaction. in Fig. . The transformation proceeds through benzene ring dissociation from the cation, followed by coordination of three CN groups from the anion to form the tetranuclear complex. This coordination is evidenced by an IR spectral shift in the CN stretching vibration from 2201 cm -1 (uncoordinated) to 2227 cm -1 (metalcoordinated) (Fig. , ESI †). Furthermore, UV-vis spectrum shows increased absorption in the 300-450 nm region after photoirradiation, consistent with cyano-group coordination. Microscopic images of the sample revealed that photoirradiation produced a uniform microcrystalline film of 2a on the quartz plate (Fig. ). Partial volatilization of the released benzene occurred upon removing the top quartz plate. Despite containing alkyl chains, the product did not melt upon heating but decomposed at 273 °C (Fig. , ESI †), approximately 110 °C higher than 1a. This increased stability likely arises from the absence of a thermally labile arene ring and enhanced anion stabilization through metal coordination. |
67d02e0a6dde43c908694072 | 1 | The photoreactivity of 1b (R = butyl, Fig. )), which has a higher melting point, was also investigated. Although 1b is a solid near room temperature (T m = 48.2 °C), it remains a liquid once melted (Tg = -51 °C). In its liquid state, 1b behaves as a Newtonian fluid with a viscosity (1.44 Pa•s at 20 °C) comparable to 1a (1.31 Pa•s) (Fig. , ESI †), making it suitable for comparative studies in both liquid and solid phases. Photoirradiation of liquid 1b at 10 °C led to microcrystalline aggregation similar to 1a. However, the reaction ratio was lower than that of 1a, with 22% unreacted IL persisting after 1 h (Fig. ), and the product contained ∼35% unidentified complexes alongside 2b (Fig. ). This inefficiency is likely due to partial crystallization during photoreaction, as supported by the quantitative formation of 2b when irradiated at 40 °C. Photoirradiation of solid 1b for 1 h resulted in a substantially low conversion (28%), yielding a heterogeneous solid-liquid mixture. Of the reacted product, 57 mol% formed the cubane complex 2b (Tdec = 317 °C), while the remainder consisted of unidentified complexes. The occurrence of the photoreaction, albeit with low conversion, despite being in the solid state, is probably due to benzene release and the formation of a minor liquid component (Fig. ). |
67d02e0a6dde43c908694072 | 2 | The packing structures of cubane-type compounds are of considerable interest in crystal engineering. In this case, the size mismatch between the cubane core and typical alkyl chain spacing (∼4. However, in the pseudo-polymorphic γ-form (space group I-4, Z = 2), the alkyl chains are nearly aligned along the a-axis (Fig. ) and Fig. ), ESI †), creating a void (V = 244 Å 3 ) at the unit cell center, occupied by solvate molecules. This void is enclosed by four alkyl chains from two adjacent molecules arranged along the c-axis. The asymmetric unit in this polymorph contains a quarter of the tetranuclear complex. |
67d02e0a6dde43c908694072 | 3 | These findings demonstrate that alkyl chain flexibility enables diverse crystal packing, with significant variations in cubane unit arrangement across polymorphs (Fig. ). Despite these differences, the tetranuclear cages exhibit nearly identical, slightly distorted cubic structures (Ru•••B = 4.79-4.83 Å), resembling the structure of a previously reported cyanoborate-bridged Cr-carbonyl complex [{Cr(CO)3(B(CN)4)}4] 4-. 9 Powder X-ray diffraction (PXRD) analysis revealed that 2a obtained via neat photoirradiation of 1a appeared to be a mixture of the β-form and other unknown forms, possibly a benzene solvate. The 2a formed via photoirradiation in methanol solution exhibited poor crystallinity, as indicated by the presence of only a few PXRD peaks. These structural insights provide foundational data for potential application of this method in film formation. In summary, UV photoirradiation of organometallic ILs quantitatively yielded solid tetranuclear cubane-type complexes. This molecular design uses coordinating anions that function as both IL counter-anions and bridging ligands in the metal complexes. The solvent-free reaction was more efficient and higher-yielding compared to solution-phase synthesis. Furthermore, structural analysis revealed that the alkyl chain flexibility led to diverse crystal packing in the cubane complexes. This study extends our IL-based photofabrication method for coordination compounds to the synthesis of cage complexes. In contrast to the B(CN)₄⁻ anion, the limited coordination and flexibility of the substituent in the BR(CN)₃⁻ anion likely facilitate cage complex formation, highlighting the importance of anion design. Our ongoing research explores extending this solvent-free approach to synthesize diverse metal complexes. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.