id
stringlengths 24
24
| idx
int64 0
402
| paragraph
stringlengths 106
17.2k
|
|---|---|---|
674755f77be152b1d019291d | 8 | AL(DIPY)3 13k (Fig. ) crystallized with one molecule in the asymmetric unit. The molecules were tightly packed with no solvent (Fig. ) and formed weak interactions between the ester groups [C(56)-H(56 42-2.47 Å between the H and O atoms. The dihedral angles between the dipyrrin plane and the substituents were 58-77°. In contrast the acid 16 (Fig. ) crystallized with 1/2 of the molecule in the asymmetric unit which displayed a disordered benzoic acid motif. Standard intermolecular hydrogen bonds between the carboxylic acid moiety and water molecules )] with distances of 1.73-2.49 Å between H and O atoms were observed. Crystal cavities were partially occupied with solvent molecules of hexane and water. The phenyl rings were orientated out of the plane of the dipyrrin by 70-80°; however, the -COOH groups remained nearly coplanar with the phenyl ring. The acid dimer formation is well known for the organization of frameworks, and similar tris(dipyrrinato) octahedral chelates have been reported with rhodium and cobalt metal centers. Figure . View of the molecular structure in the crystal of 13h and 16 (top) and packing diagram with solvent channels of 16 viewed normal to a-axis (bottom). Thermal ellipsoids show atomic displacement at 50% probability. |
674755f77be152b1d019291d | 9 | The BODIPY appended complex 20 crystallized with one molecule in the asymmetric unit. The respective dihedral angles between the dipyrrin plane (of the ALDIPY unit) and the phenyl substituent were 69-70°. The separation between the aluminum and boron atoms was 16. 35-16.50 Å. The crystal structure showed the formation of a framework structure with very large cavities (Fig. ). Similar to 16 this represents a promising route towards nanomaterials, e.g., via using PSs as drug delivery platforms, enhancing the biocompatibility or finding use in different fields. Figure . View of the molecular structure in the crystal of 20. Hydrogen atoms and disordered positions have been omitted. |
674755f77be152b1d019291d | 10 | To conclude, all the complexes presented herein were characterized by an octahedral configuration of the aluminum center (see Table ). The dipyrrin cores were relatively planar and the substituents were twisted out of its plane between 30-98° indicating that the groups to the phenyl ring play a significant role in the final configuration. Chelates 13b, 13h, 13i and 13j displayed the largest tilt of the meso-aryl ring versus the dipyrrin. The largest variation in this twist angle was found for 13h where two 2,4,5-trimethoxyphenyl moieties were almost perpendicular to the dipyrrin, while one showed a 63° angle. 13i and 13j had almost orthogonal configuration of the dipyrrin and the substituent. On the other hand, 13f showed the smallest distortion. A similar trend of the dihedral angles was found for the DIPY precursors: 12f (53°) < 12a (68°) < 12g (83°) < 12h (93°). The Al-N bonds were in the range of 1.97-2.02 Å (ca. 2 Å), which is similar with the rhodium and gallium analogues but is slightly longer than the cobalt analogues and shorter than the indium analogues. Finally, looking at the crystal packing and specifically that of 13d, 16 and 20, we can observe that the carboxylic acid and the fluorine groups play a role in the crystal structure giving a packing with voids and prominent porous networks. |
674755f77be152b1d019291d | 11 | Steady state and transient absorption spectroscopy were employed in order to investigate the ground and excited state properties of the homoleptic ALDIPYs. The photophysical data of these complexes may resemble other trivalent metal complexes (ML3) of Co(III), Ga(III), In(III), Rh(III). . Fluorescence studies have been reported but singlet oxygen determination and triplet state properties have yet to be explored. Notably, the triplet excited state properties of the aforementioned complexes have remained unexplored until now. |
674755f77be152b1d019291d | 12 | Ground state properties. The absorption spectra of the AL(DIPY)3 complexes are displayed in Figure . For clarity the spectra were grouped and distributed in three graphs as their features are overlapping. All complexes displayed two pronounced characteristic bands at ca. 450 and 500 nm (Table ). Lower values of molar absorption coefficient (ε at 450 nm band) ranged from ~40,000 to ~60,000 cm -1 M -1 for 13a, 13j, 16, 17, and 18; whilst for 13c, 13f, 13h and 13i ε was two times higher (range of ~90,000 to ~140,000 cm -1 M -1 ). The BODIPY-AL(DIPY)3 conjugate 20 displayed the highest absorptivity ~ 250,000 cm -1 M - 1 at 500 nm due to the strong absorbance of the BODIPY moiety. The difference between the absorption spectra of the single dipyrrin chromophores (Fig. ) and the AL(DIPY)3 complexes is apparent, with the former displaying a broader single band at ~430 nm with lower molar absorption coefficient and the latter displaying a double, red-shifted band (Fig. ). |
674755f77be152b1d019291d | 13 | The two bands in the absorption profile of the AL(DIPY)3 complexes in the region of 450-500 nm correspond to the excitonic states as a result of exciton coupling, well defined by Kasha et al., which is the apparent splitting of the absorption bands: Davydov splitting. The absorption data match other reported UV-Vis features of similar coordination complexes. Co(III), Rh(III), and Fe(III) tris(dipyrrinato) complexes have a broader band splitting compared with the In(III) or Ga(III) counterparts, which have a more profound double peak, and the same characteristic appears in the aluminum complexes with two bands at 450 -500 nm. Hence, metal coordination can affect the absorption profile as it can alter the excited state features. Another effect that impacts the features of the absorption spectra is the atomic radius of the elements. The excitonic state energy is inversely proportional to the distance between the molecules. Excitonic splitting of the respective energy state is leading to the case where the lowest exciton state is closer to the triplet excited state and therefore can efficiently result in higher intersystem crossing yields, efficiently producing triplet excited states, which is the aim of PDT. Moreover, comparing the UV-Vis of the parent dipyrrins (Error! Reference source not found.) and the AL(DIPY)3 complexes (Error! Reference source not found.) it is evident that not all the dipyrrins display a band in the range of 300-400 nm; however, all the complexes display a band in this range, with some of them exhibiting higher molar absorption coefficient than others. For example, the anthracene moiety absorbs in this region, displaying four peaks as a characteristic feature of both 12j and 13j. Consequently, bands that appear at the region 300-400 nm are related to dipyrrin based charge transfer transitions (intramolecular charge transfer ICT) along with the co-occurrence of the n-* and 1 -* transitions by the various mesosubstituents. Additionally, a possible metal to ligand charge transfer transition has been previously assigned for both low-and high-energy bands in relevant complexes, and may be expected to co-occur with the associated charge transfer within the dipyrrins and the 1 -* transitions. Presumably, in this case the metal to ligand charge transfer is not possible. |
674755f77be152b1d019291d | 14 | Lastly, regarding the influence of meso-substitution there was only a marginal difference in the absorption spectra of the AL(DIPY)3 complexes and thus different substituents do not greatly affecting the absorption profile. This is in accordance with reports of similar compounds. DFT and TD-DFT calculations were conducted for representative complexes (13a, 13d, 13h, 13i, 17) by using a hybrid B3LYP functional and a LANL2DZ basis set. The aim was to visualize the frontier molecular orbitals (FMOs) and the electron-density distribution within the complexes. Additionally, the theoretical singlet S1 ( 1 E00) and triplet T1 ( 3 E00) excited levels of the AL(DIPY)3 complexes were determined along with the singlet-triplet gap (ΔES-T). As shown in Error! Reference source not found. there is no apparent electronic distribution in the metal center and indeed it was localized within the dipyrrin moieties confirming the inter-ligand excitation profile and the charge transfer between the ligands. This result is in accordance with the literature. DFT calculations of Ga(III) and In(III) analogues attributed the electronic distribution mainly as ligand centered in the intra-ligand charge transfer. Kusaka et al. investigated the heteroleptic tris(dipyrrinato)indium(III) coordinated complexes and showed that there is no significant influence of the metal center in the FMOs; the HOMO and LUMO are localized on the dipyrrin ligands. Ultimately, the computed triplet excited energy was found between 1.60-1.75 eV following the trend: 13d < 17 < 13i < 13h < 13a, which satisfies the prerequisite for a PS to generate singlet oxygen (T1 > 0.98 eV, the lowest excited singlet state of oxygen). |
674755f77be152b1d019291d | 15 | Singlet excited state properties. Fluorescence emission spectra of AL(DIPY)3 complexes were recorded, and the fluorescence quantum yield was calculated by using rhodamine 6G or coumarine 153 as reference compounds. The emission profile of the fluorescence spectra of the AL(DIPY)3 complexes is rather similar displaying one broad band which is relatively weak. The emission band of the AL(DIPY)3 occurs between 530-590 nm, and spans across a range of more than 100 nm (Error! Reference source not found., SI S3.2), being a mirror image of the absorption band at ~500 nm. Conjugate 20 displayed one broad peak ~570 nm indicating its origin from the emission of precursor 13c, rather than that of the BODIPY 19. Excitation either at 450 or 500 nm had no influence on the emission profile of 20, which means the typical strong BODIPY emission at λmax ~520 nm is attenuated, since no emissive BODIPY feature is observed. Fluorescence quantum yields for the majority of the compounds were calculated in the range of 0.01-0.07 (Φf) following an ascending order: 16, 17 < 13d, 13e, 13k, 20 < 13a, 13c, 13f < 13h < 13i < 13j < 13b (Table ). The highest yields were displayed by 13b, 13i, and 13j probably due to the bulky 5substitution and steric hindrance. It has been previously reported that mesityl substitution on the meso position and restriction of the internal rotation can drastically increase the fluorescence quantum yields of dipyrrinato metal complexes. Together with the trends observed in In(III) and Ga(III) homoleptic complexes this asserts that Group 13 complexes are less luminescent than the zinc or boron complexes. For 17, attempts to use DMSO, THF, or 2-propanol were unsuccessful, whilst in ethanol or DMF a weak band appeared with negligible yield (< 0.001). Presumably, in polar media there is a prominent possibility of either solvent fluorescence quenching or formation of non-emissive ICT states which lead to a more efficient triplet formation via charge recombination and nonradiative decay as the dominant pathway. On the other hand, in non-polar solvents the ICT state will not stabilize and therefore, the excited state S1 will undergo ordinary ISC (in competition with the IC). Indeed, all the aluminum complexes display moderate triplet state lifetimes in air equilibrated solutions (see below). The emission peaks are equating to Stokes shifts between a minimum 1308 cm -1 for 13b and a maximum 3114 cm -1 for 13k (Table ). Complexes 13b, 13i, and 13j have smaller Stokes shifts pointing to a reduced number of molecular rearrangements reflecting less conformational freedom in the ground and/or excited states due to the steric hindrance by the substituents. The considerable Stokes shifts and the moderately low fluorescence quantum yields support that non-radiative decay pathways are the most dominant processes (ISC and IC). TCSPC was performed and fluorescence lifetimes were determined to assess the singlet photophysical properties of the complexes (Table ). The singlet state lifetimes of the AL(DIPY)3 complexes in toluene (16 in THF) ranged from 1.3 to 4.7 ns with the ascending order: 13e, 16 < 13k, 20 < 13c < 13f < 13a < 13d < 13h < 13i < 13j < 13b. Error! Reference source not found. displays all the fluorescence decays. Comparing 13b with the respective mesityl-dipyrrinato Ga(III) and In(III) complexes, aluminum complex appears to have slightly increased Stokes shifts and longer singlet excited lifetime (+1 ns) whilst maintaining a moderately high fluorescence quantum yield equal to the In(III) counterpart: Ga(III) complex: Φf = 0.02, τs = 3.75 ns, Δν = 1220 cm -1 and In(III) complex: Φf = 0.07, τs = 1.93 ns, Δν = 1113 cm -1 (in hexane). As expected, the singlet state lifetime of 17 was not detected neither in DMSO nor ethanol or THF (< 1 ns). |
674755f77be152b1d019291d | 16 | Consequently, the radiative and non-radiative rate constants were determined (n: kf = Φf / τs, knr = (1 -Φf) / τs) and are consistent with the fluorescence profile of the AL(DIPY)3 complexes. The radiative rate (fluorescence) and non-radiative decay rates (IC, ISC, ICT) are of the order of 10 6 s -1 and 10 8 s -1 , respectively, supporting the prevalence of the non-radiative decays (Table ). Similar values have been reported of dipyrrin complexes showing that increasing the number of dipyrrin moieties in proximity influences the emissive properties and enhances the triplet state or internal conversion processes to the ground state. Lastly, the singlet excited energy state was determined experimentally via the intersection of the normalized absorption and emission spectra and was found to be in the range of 2.3 -2.4 eV (Table ). |
674755f77be152b1d019291d | 17 | The possible charge separated states result in non-emissive charge separated states which can undergo IC (S1 → S0) and ICT or ISC (S1 → Tn/1) as the most probable decay pathways. Additionally, the non-radiative relaxation from a higher to lower excitonic state and the weak emission profile suggest that the prominent electronic transitions are ligand centered. These observations indicate that the chromophores undergo excitedstate non-radiative relaxations (conformational, vibrational, electronic) prior to the observed steady-state emission spectrum with the radiationless ISC process (with rate constants of the order of 10 8 s -1 ) expected to lead to an efficient triplet formation. Any possible metal-to-ligand charge transfer that might co-occur could be merged with the broad peak of the emission spectrum since it is already red shifted (500-700 nm) where CT emission can be expected. Kusaka et al. considered an equilibrium between the emissive 1 -* transition of the ligand and the non-emissive charge separated states of tris(dipyrrinato)In(III) complexes. They surmised that in homoleptic complexes (with β substitution) there was a larger contribution of the charge separated states among the ligands leading to low fluorescence quantum yields, whilst in heteroleptic complexes a smaller contribution of the charge-separated states resulted to the highest fluorescence quantum yield. |
674755f77be152b1d019291d | 18 | Triplet excited state properties. Nanosecond transient absorption (TA) spectroscopy was employed to explore the triplet excited states and the proposed non-radiative pathways of the AL(DIPY)3 chelates (SI S3.5). A relative estimation of the triplet state quantum yield between the complexes was made. The desired long-lived triplet excited states were observed, and the triplet excited lifetimes (τT) were defined in air equilibrated solutions by simple monoexponential fitting. TA spectra were recorded in non-polar toluene solution and polar ethanolic solution. |
674755f77be152b1d019291d | 19 | (for all TA spectra with time trace fittings see Fig. ). It is apparent that TA spectra are superimposed by the ground state bleaching resulting in an absorption maximum at ~400 nm with two prominent negative absorbance signals at ~450 and ~500 nm reproducing the exciton split bands of the ground state. These are the characteristic peaks of the ground state electronic transitions of the complexes and provide information that the triplet-triplet absorptivity is weaker than the singlet absorptivity, although occurring in decent yields. Most TA spectra display a broad band between 550-700 nm. Complex 13d shows nearly the same intensity of the two ground-state bleaching bands indicating a weaker triplet-triplet absorption at 500 nm. Conjugate 20 maintained moderate triplet state lifetime displaying one ground state bleaching band at ~500 nm, repeating the absorption pattern typical of the BODIPY moiety in nanosecond TA experiments. Figure . TA spectra at ambient conditions of 13a in ethanol (40 ns tinc; 450 nm λexc); 13d in toluene (40 ns tinc; 450 nm λexc); 13h in toluene (80 ns tinc; 450 nm λexc); 13i in ethanol (60 ns tinc; 450 nm λexc); arrows pointing from blue to red color show the decay from the maximum intensity in the successive steps, respectively. |
674755f77be152b1d019291d | 20 | The corresponding triplet state lifetimes of the AL(DIPY)3 complexes in toluene were in the range of 290-400 ns with the longest triplet lifetimes displayed by 13b and 13h (Error! Reference source not found.). The results are similar to previously reported triplet excited state lifetimes of In(III) and Ga(III) chelates. However, to date only few studies focused on the photoexcitation processes of this class of tris(dipyrrinato) complexes. A comparable trend occurs in ethanolic solutions where the triplet state lifetime underwent a significant decrease (where comparison was possible) of the order of ~90 ns, except for 13e which showed a minor difference. The latter, together with 13h and 13i, displayed the highest triplet state lifetimes in ethanol ~290 ns. Compounds 13c, 13d, 16, and 17 exhibited a triplet state lifetime of around 250 ns with an overall lower value than in toluene solutions. Lastly, we can observe that the respective errors of the fitting curves are smaller in toluene than in ethanol and along with the higher triplet state lifetime values (in toluene) this is indicative that solvent significantly affects the results. Another complementary feature that can be estimated from TA spectra, and the intensity of the ground state bleaching, is the triplet state relative efficiency (Φisc). Since the UV-Vis absorption at the excitation wavelength was set to the same range (0.4-0.5), the relative triplet state yield can be estimated comparing the intensity of the TA spectra of each of the AL(DIPY)3 complexes (preferably by experiments on the same day). The ascending order of the relative triplet state yields in toluene is: 17 < 16 < 13e < 13a, 13f, 13k < 13b, 13c, 20 < 13d < 13i < 13h < 13j; while the order in ethanol is: 13d < 13e, 16, 17 < 13c < 13h < 13i. Therefore, we can postulate that 13h and 13i are more likely for efficient triplet sensitization since they persistently exhibit higher triplet state yields in both solvents; however, 13b and 13k may have the same potential since they display relatively high triplet state lifetimes in toluene whilst they both have moderate fluorescent characteristics. In comparison to toluene, complexes 16 and 17 illustrated a greater triplet absorptivity in ethanol and 13d absorbed less. TA spectra in DCM exhibited intense peaks with longer triplet lifetimes, but as there is degradation we did not proceed with other experiments. The triplet excited state lifetimes obtained were calculated as follows: 13h: 650±40 ns (Fig. ), 13i: 870±38 ns (Fig. ), 13k: 760±34 ns (Fig. ). These differences of the triplet state lifetimes among the solvents support the fact that photoexcitation processes are affected by the nature of the solvent both in the triplet and singlet excited states and imply that non-radiative decay T1 → S0 is also significant. |
674755f77be152b1d019291d | 21 | Additionally, to test whether oxygen is a key element upon photoexcitation, the triplet state lifetimes under oxygen-free conditions were calculated by simple monoexponential fitting (unless stated otherwise). TA spectra and time trace fittings are shown in Fig. . Solutions were subjected to five freezepump-thaw cycles prior to photoexcitation. TA features of the AL(DIPY)3 complexes were similar, with triplet excited state lifetimes being much longer, in the range of ~80-200 μs, giving testimony that oxygen significantly alters the outcome of the excitation. This condition is consistently observed for all the AL(DIPY)3 chelates (Error! Reference source not found.). Lastly, the rate constant for quenching of the triplet state by oxygen was also determined to be in the range of 1.60-1.90 × 10 - 9 M -1 s -1 ; with [O2] in ethanol and toluene at 20 °C: 2.1 × 10 -3 M and 1.8 × 10 -3 M, respectively. Singlet Oxygen Generation Singlet oxygen phosphorescence was determined at 1270 nm and the singlet oxygen quantum yields were calculated relative to 5,10,15,20-tetraphenylporphyrin (H2TPP) in toluene or erythrocin B in ethanol as standard reference compounds (Table , SI 3.6, Fig. ). Additionally, a ratio of the ΦΔ of AL(DIPY)3 chelates and ΦΔ of the reference (erythrocin B) was defined in another attempt to demonstrate the relative singlet oxygen generation. Singlet oxygen determination can be affected by the oxygen concentration and its lifetime in different solvents. There is a significant variation in the singlet oxygen quantum yield determined for 13h between ethanol (ΦΔ = 0.15) and toluene (ΦΔ = 0.70) solutions, while for 13i the opposite trend occurs: ΦΔ = 0.35 in ethanol; ΦΔ = 0.07 in toluene. Since singlet oxygen phosphorescence is a sensitive method, and taking into consideration that oxygen concentration in air equilibrated solutions is mainly the same in both solvents, the discrepancies between the two solvents can be ascribed to the difference in the singlet oxygen lifetimes: singlet oxygen lifetime in toluene (27 μs) is almost two times greater than that in ethanol (15 μs). The combination of the absolute values and the calculated ratios leads shows that chelates 13d, 13f, 13g, 13i and 13j can be efficient singlet oxygen generators. Therefore, their respective triplet state energy level should be higher than the lowest energy of molecular oxygen (0.98 eV) to generate singlet oxygen. To conclude, since exciton coupling splits the singlet excited state into two excitonic states with higher (X′′) and lower energy level (X′), it is expected that the X′ state is closer to the triplet excited state. Therefore, a decrease of this energy gap can enhance the ISC efficiency and result in higher singlet oxygen generation. The occurrence of a high IC yield from non-radiative decay to the ground state (S1/T1 → S0) cannot be eliminated and it can be higher than the ISC in some of the complexes, therefore further investigation is needed. The quantum yield of ISC (or ICT) and IC should be following the general trend: ΦΔ < Φisc + Φic < 1 -Φf. An approximation of the radiationless processes could be given by the singlet oxygen quantum yields and the triplet state information indicating that 13d, 13f, 13h, 13i and 13j may possess an enhanced intersystem crossing yield which agrees with the increased knr. Moreover, these chelates display moderate singlet excited state lifetimes. Additionally, the large Stokes shifts, the low radiative rates and the long triplet lifetimes confirmed the emission through the excited triplet state of the aluminum complexes. These findings can lead to the inter-ligand-based triplet excited state 3 p-p* (long-lived 3 IL) originating from the initial 1 p-p* states. All the above are extremely dependent on the polarity of the environment, which influences the processes and the competing pathways. The reduction of singlet oxygen efficiency has been observed in polar solvents (aqua-rich media) and is related to the low solubility of oxygen in water in comparison to other organic solvents as well to the lower singlet oxygen lifetime in water. Moreover, excited state quenching or aggregation in polar environments can occur resulting in singlet oxygen attenuation. For instance, encapsulation in a polymer matrix to generate nanoparticles or addition of poly(ethylene) glycol (PEG) substituents can enhance the water solubility and biocompatibility of the potential therapeutic agents. |
674755f77be152b1d019291d | 22 | Following the chemical and photophysical characterization of the AL(DIPY)3 complexes, their photobiological evaluation was assessed in view of their application as PSs in PDT. The AL(DIPY)3 chelates used for the in vitro studies were chosen in terms of their solubility in DMSO to obtain stock solutions with concentrations of 1-2 mM. Therefore, eight of the complexes were investigated with regards to their phototoxic effect against the CT26 mouse colon carcinoma cell line (13a, 13c, 13d, 13f, 13g, 13i, 16, 17). Cell viability was assessed by using resazurin assay where the cell survival is determined in comparison with the non-treated control sample which corresponds to 100% cell viability. In a first set of studies, the toxicity of the selected compounds was evaluated in the absence of light using the CT26 cell line. Most of the complexes showed no toxicity up to 20 μM except for 17 which displayed 40-50% of cell death at 20 μM (Error! Reference source not found.). Therefore, in the following phototoxicity studies, the highest concentration of 17 that was tested was 10 μM. |
674755f77be152b1d019291d | 23 | Next, the phototoxicity effect of the chelates in CT26 cell line was assessed by using a broadband lamp (LED) with a light dose (L.D.) of 2.6 J.cm -2 . Accurate light doses were estimated considering the overlap between the LED and the photosensitizer spectrum. Final concentration of 13a-16 was in the range from 0.62 to 20 μM, whereas concentrations from 0.31 to 10 μM of 17 were used (Error! Reference source not found.). Our results showed that complexes 13a, 13f and 13i (Error! Reference source not found.A, D and F) are not phototoxic at the indicated light dose, while 13c (8B) showed a marginal phototoxicity at 20 μM. Complex 13d displayed a higher phototoxic effect, since it reduced cell viability by 50-60 % at 20 μM (Error! Reference source not found.C). |
674755f77be152b1d019291d | 24 | Interestingly, 16 and 17 had a very good phototoxic effect on the colon carcinoma cell line (Fig. and), whilst they were the most difficult to handle during the photophysical characterization studies. Their non-emissive singlet state properties led to an efficient triplet excited state formation and a low singlet oxygen generation (ΦΔ = 0.10 in ethanol), yet suitable to develop the desired phototoxic effect. It seems that the polar hydroxyl and carboxyl groups can enhance the phototoxicity effect since the post-functionalized 17 exhibits greater phototoxicity in comparison with its precursor 13d. The IC50 values of 16 and 17 are low with 1.8 μM and 3.5, respectively. |
674755f77be152b1d019291d | 25 | Lastly, 13h that contains three methoxy groups in the phenyl rings showed great phototoxic potential with approximately 100% of cell death at all the tested concentrations (0.62 to 20 μM). In contrast, 13f bearing only one methoxy group on each phenyl ring, did not exhibit any phototoxic effect (Error! Reference source not found.E). This can be attributed to the fact that 13h has a higher singlet oxygen quantum yield (ΦΔ = 0.70 in toluene) in comparison to 13f (ΦΔ = 0.38 in toluene). To calculate the IC50 value for 13h a new set of phototoxicity studies was carried out using lower concentrations and/or lower L.D. An IC50 value in the nanomolar range (80 nM) was determined using 13h was used in the range of 0.04-5 μM with a L.D. of 2.6 J cm -2 Error! Reference source not found.A). A slightly higher IC50 (0.18 μM) was calculated when 13h was examined under a lower L.D. of 1 J cm - 2 (0.08 -10 μM) (Error! Reference source not found.B). The low IC50 value is of great importance as it gives room for less exposure to the PSs, which can contribute to reducing unwanted side effects while maintaining the phototoxicity. Note, that 13h exists as an atropisomeric mixture; however, this is not a new concept for PS in PDT. E.g., Ru(III) complex TLD-1433 is a chloride salt of a racemic mixture of two isomers and can be effectively utilized as long as the toxicity profile is acceptable. The high phototoxicity mediated by 13h is consistent with its photophysical properties. Both in toluene and ethanol it can effectively generate singlet oxygen. The compound has a moderate fluorescence quantum yield (Φf = 0.03) taking into consideration that such molecules are normally poor emitters. Moreover, triplet excited state lifetimes were long and greatly affected by oxygen. Overall, the optimal photophysical properties of 13h are correlated with the significant phototoxicity and the lack of dark toxicity in CT26 cancer cells. Surprisingly, 13i which had improved singlet and triplet state properties, and singlet oxygen formation in polar solvent (ΦΔ = 0.34), did not induce any phototoxicity at 2.6 J.cm -2 . We hypothesize that the lack of activity might be explained by its poor internalization or possible aggregation. However, it should be noted that the applied L.D. was relatively low in comparison with other reports, which leaves room for exploration. Ultimately, we can surmise that the AL(DIPY)3 complexes with -OH, -COOH, -OCH3 groups in the phenyl moieties exhibit better biological compatibility and probably cell permeability. These polar groups may increase the amphiphilicity of the compounds which at the end can facilitate their cellular uptake. Our preliminary results demonstrated that 13h, 16 and 17 are promising photosensitizers as they exhibit high phototoxicity at low concentrations. Further studies are crucial to unveil the underlying mechanism of action of the AL(DIPY)3 complexes. |
674755f77be152b1d019291d | 26 | To conclude, we synthesized a library of novel tris(dipyrrinato)aluminum(III) complexes in a three-step synthesis. The complexes were stable under the basic condition of the ester cleavage and Pd-catalyzed reaction condition (Suzuki). X-ray single crystal analysis provided most of the corresponding structures with 13b, 13h, 13i and 13j displaying higher distortion of the aryl ring attached at the meso position of the dipyrrin. Noteworthy, conjugate 16 has the potential to form metal organic frameworks (MOFs). The complexes have excitonic and ICT states leading to non-radiative decays upon excitation. DFT calculations supported the inter-and intra-ligand electronic transitions that may occur and the marginal metal contribution on the transitions. However, long-lived triplet excited states were formed and allowed for singlet oxygen generation. In PDT, long triplet excited state lifetimes are envisioned, and this could be advantageous for (dipyrrinato)aluminum complexes breaking new ground in research towards photomedicine. |
674755f77be152b1d019291d | 27 | Namely, the chelates displayed fluorescence emissive properties with fluorescence quantum yields in the range of 0.01-0.07 in toluene; whereas in ethanol the signal was negligible. Chelates 13b, 13h, 13i and 13j displayed the higher yields in toluene. A similar trend was observed in the singlet excited state lifetimes ranging from 1.5 -5 ns. Consequently, because of the CT states, AL(DIPY)3 efficiently populate triplet excited states with long-lived triplet state lifetimes in air-equilibrated conditions (250-350 ns). Complexes 13d, 13h, 13i and 13j displayed the highest triplet state lifetimes in toluene and 13c, 13h and 13i in ethanol. Additionally, in the absence of oxygen, the triplet state lifetimes ascend up to the range of 50-200 µs, confirming that the complexes' reactivity with oxygen is of paramount importance after photoexcitation. Next, comparing the singlet oxygen formation in ethanol and toluene, we can conclude that in the polar solvent, luminescence is apparent but not significantly detected, while toluene proved more suitable for the measurements. Complexes 13d, 13f, 13h, 13i and 13j showed high singlet oxygen generation in both solvents. Preliminary results of the in vitro screening of eight AL(DIPY)3 against CT26 cells are promising, as these complexes were able to trigger cell death upon irradiation at safe nanomolar and micromolar concentrations. Specifically, we identified possible PS candidates since four of the complexes induced phototoxicity with the following ascending order: 13d < 16 < 17 < 13h, pointing to the latter as the best PS candidate. Complexes with polar groups attached might express improved amphiphilicity, allowing for a better cell internalization and permeability and thus exerting their phototoxic efficiency. To summarize, we conducted a comprehensive study on the development of novel homoleptic tris(dipyrrinato)metal(III) complexes with aluminum as earthabundant element. Together with the low toxicity, the investigation of their photophysical properties and their first screening of photocytotoxicity points towards possible uses as photosensitizers in medicinal bioinorganic chemistry. |
674755f77be152b1d019291d | 28 | General procedure for the synthesis of tris(dipyrrinato)Al(III) complexes. Synthesis of the tris(dipyrrinato)aluminum(III) complexes was adapted to literature procedures by using aluminum salt for the complexation. In a dry Schlenk tube the corresponding 5-substituted dipyrrin (3 eq.) and AlCl3 (1.2-1.4 eq.) were added and dried under vacuum for 1 h. Then, chloroform was added under argon atmosphere and the resulting slur/solution was degassed with argon for 30 min. DIPEA (3 eq.) was added dropwise and the reaction mixture was stirred overnight at 70 °C under reflux. Color changed from dark yellow/brown to dark red. Then the reaction mixture was allowed to reach r.t. and DCM was added. Organic phase was washed with brine (×2), NaHCO3 (×2) and water (×2), dried over Na2SO4, filtered, and the resulting solution was evaporated in vacuo to give a dark brown/red crude mixture. The product was purified via column chromatography. |
674755f77be152b1d019291d | 29 | Fluorescence emission spectra were recorded on a SPEX Fluorolog 3 fluorometer with double grating monochromators in the excitation and emission channels. Excitation light source was a Xenon lamp (450 W, Osram) and the detector was a Peltier cooled photomultiplier tube (Hamamatsu, R636-10). The fluorescence emission signal is collected in a right-angle geometry, and the fluorescence spectra are corrected for fluctuations of the excitation source flux. For the singlet oxygen emission at 1275 nm, a highly sensitive liquid nitrogen cooled InGaAs detector (Electro-Optical Systems DSS series cryogenic receiver, 2 mm InGaAs photodiode) was coupled to a Horiba Jobin Yvon Spex Fluorolog 3 spectrofluorometer. Maximum slits (excitation and emission) and long integration times (10 -15 s) were used. A 850 nm cut-off filter was used in the emission path, to prevent second order effects of the fluorescence compounds. Triplet state lifetimes were determined by nanosecond timeresolved absorption spectroscopy using an EKSPLA NT342B laser system in which the third harmonic of a Nd:YAG laser system was used to pump an optical parametric oscillator (OPO) at the corresponding excitation wavelengths for each experiment. The typical power was 1 -2 mJ per pulse and the laser system was operated at 5 Hz. Probe light, running at 10 Hz, was generated using a high-stability short arc xenon flash lamp (FX-1160, Excelitas Technologies) with a PS302 controller (EG&G). The probe light was split in a signal and a reference beam with a 50/50 beam splitter and focused on the entrance slit of a spectrograph (SpectraPro150, Princeton Instruments). The signal beam (A = 1 mm 2 ) was passed through a sample cell and overlapped with the excitation light on a 1 mm × 1 cm area, perpendicular to the excitation beam. A reference beam was used to normalize the signal for fluctuations in the flash lamp intensity. Both beams were recorded with an intensified CCD camera (PI-MAX3, Princeton Instruments) using 5 -20 ns gate depending on the time steps of the measurements. The timing was achieved with a delay generator 6 (DG535, Stanford Research Systems, Inc.) and the setup was controlled with an in-house written program (LabView). Time-resolved fluorescence decays were measured using a TCSPC technique. The relevant excitation wavelengths were generated by frequency doubling of the output of a tunable Titanium:Sapphire laser (Chameleon Ultra, Coherent). The repetition rate is decreased from the fundamental 80 MHz to a lower value (usually 8 MHz) using a pulse picker (Pulse Select, APE). The fluorescence was detected with a multichannel plate photomultiplier tube (R3809U-50, Hamamatsu) through a single-grating monochromator (Newport Cornerstone 260, f = 250 mm, grating 300 ln/mm blaze 422 or grating 300 ln/m blaze 750 nm M20, Carl Zeiss, 600 lines/mm). The overall instrument response is usually 20 -25 ps (FWHM) measured from a dilute scattering solution (Ludox) at the excitation wavelength. TCSPC histograms were recorded using ~10 4 counts in the peak channel using the SPCM program (Becker & Hinkle). |
674755f77be152b1d019291d | 30 | The absorption spectra were recorded in DCM, MeOH or THF at room temperature and molecular absorption coefficients were calculated from Beer Lambert's law A = ε×c×l, where A the absorbance of the molecule at specific wavelength; ε the molar extinction coefficient; c the concentration of the sample in the cuvette; l the length of the light path (the width of the cuvette was 1 cm). |
674755f77be152b1d019291d | 31 | The respective compounds were dissolved in toluene, THF, or DMF and their absorbance in the UV-Vis spectrum was adjusted to ca. 0.10 at the wavelength of excitation. Steady-state fluorescence emission spectra were obtained upon excitation at 445-500 nm. The fluorescence quantum yields (Φf) for the AL(DIPY)3 chelates were calculated with rhodamine 6G (Φf = 0.94 in EtOH), or coumarin 153 (Φf = 0.38 in EtOH) as standard references. The fluorescence quantum yield was determined by equation S1 (1 s integration time; 3 nm excitation slit; 3 nm emission slit). |
674755f77be152b1d019291d | 32 | where Φ is the quantum yield; A(λ) is the absorbance of the solution at the excitation wavelength λ; I is the relative intensity of the exciting light at wavelength λ; n is the refractive index of the solvent and D is the integrated area under the corrected emission spectrum. Subscripts x and r refer to the unknown and reference solutions, respectively. |
674755f77be152b1d019291d | 33 | Most of the complexes were soluble in toluene and thus it was used for the measurements. Trials to determine the fluorescence yield in ethanol (for example for 13h and 13e were not successful. The emission band appeared quite noisy and weak. Additionally, the fluorescence quantum yield of 16 and 17 was difficult to calculate due to a very feeble fluorescence intensity. |
674755f77be152b1d019291d | 34 | For 16 the use of THF or ethanol were the appropriate solvents in terms of solubility; however, the fluorescence yield was calculated to be less than 0.01 in both solvents with a broad emission band at ~610 nm (Error! Reference source not found.). Similarly, a low yield was reported for the homoleptic 4-carboxyphenyl rhodium(III) complex derivative whilst counterparts of Co(III) ML3 are completely non-emissive. 3.3. Time-correlated Single Photon Counting (TCSPC) |
674755f77be152b1d019291d | 35 | The respective AL(DIPY)3 complexes were dissolved in toluene or THF and their absorbance in the UV-Vis spectrum was adjusted to ca. 0.10 at the wavelength of excitation. TCSPC was performed and the spectra were obtained upon excitation at 450 nm for all AL(DIPY)3 complexes. The detection wavelength was set to 580 nm. Fluorescence lifetimes for all the compounds were calculated at picosecond scale. |
674755f77be152b1d019291d | 36 | DFT calculations were performed in Gaussian 16; on the Dutch national e-infrastructure with the support of SURF Cooperative (lisa.surfsara.nl). Avogadro and Gabetid were used for the molecular editing and the MOs visualization. Density functional theory (DFT) and time-dependent DFT (TD-DFT) calculations were performed to investigate the ground-state and the excited-state properties of five representative AL(DIPY)3 complexes. Hybrid B3LYP functional was used and a LANL2DZ basis set to optimize the ground state (S0) geometry. The same methods were applied and TD-DFT calculations were performed to calculate the excited singlet (S1) and triplet (T1) states and visualize the respective molecular orbitals (MOs). The S1 ( 1 E00) and T1 ( 3 E00) energies, the difference between the first singlet excited and triplet energy state (ΔES-T) are shown in Table along with the energy difference between [HOMO -LUMO], [(HOMO-1) -LUMO], [(HOMO-1) -(LUMO+1)] and [HOMO -(LUMO+1)]. |
674755f77be152b1d019291d | 37 | The respective compounds were dissolved in ethanol, toluene or THF [for 16] and their absorbance in the UV-Vis spectrum was adjusted to ca. 0.10-0.20 at the wavelength of excitation. Direct detection of the luminescence emission of 1 O2 at 1275 nm was achieved upon excitation at 500 or 508 nm. Singlet oxygen quantum yields (ΦΔ) were calculated by equation 1, with erythrocin B (ΦΔ = 0.69 in ethanol), or 5,10,15,20-tetraphenylporphyrin (H2TPP) (ΦΔ = 0.68 in toluene) as reference compounds. The wavelength range of the emission was recorded from 1150 nm to 1350 nm; 10 or 15 s integration time; 14 nm excitation slit; 40 nm emission slit. D was calculated from the area under the curve between 1220-1340 nm. Note, that for singlet oxygen quantum yield calculations, the reference and unknown compound should be diluted and measured in the same solvent. |
674755f77be152b1d019291d | 38 | Cell culture and preparation of stock solutions The CT26 cell line (mouse colon carcinoma) was obtained from American Type Culture Collection Cells. Cells were kept in DMEM ((Dulbeccos' modified Eagle's medium, Sigma) with 10% heat inactivated fetal bovine serum (Gibco, Life Technologies) and 1% penicillin streptomycin (Sigma) in a humidified incubator with 5% CO2 at 37 °C. Cells were detached using trypsin-EDTA solution (Sigma), counted, and seeded at the necessary density in plates of the appropriate size. A stock solution of the respective compounds (ca. 1 mM) was prepared with DMSO Hybri-Max TM (Sigma) and stored at -20 °C. Before each experiment, the stock solutions were diluted with the cell culture medium at the desired concentration and then added to the cells. Owing to the inherent toxicity of DMSO, the final concentration of DMSO never exceeded 1-2% after the dilution in DMEM. |
674755f77be152b1d019291d | 39 | After 24 hours of incubation, a volume of 200 µL of the indicated concentration of the respective compounds was added into the appropriate triplicate wells with the final concentrations of: 0.625, 1.25, 2.5, 5, 10, 20 µM. CT26 cells were incubated with the compounds for 24 h, followed by cell washing with medium. To assess cell viability, resazurin blue (Sigma, diluted in PBS, phosphate buffer solution) was used (150 µL in each well, 0.1 mg mL -1 ). After ca. 1.5 h, the fluorescence of resorufin, the metabolic product that results from resazurin reduction, was recorded with a microplate reader (Biotek Synergy HT) using 528/20 nm excitation and 590/35 nm emission filters. The experiment was repeated three times, and the final cell viability was accessed after merging the mean values of all the independent experiments. The statistical significance analysis was applied by using Graphpad software. |
674755f77be152b1d019291d | 40 | After 24 hours of incubation, a volume of 200 µL of the indicated concentration of the compounds (from 0.625 to 20 µM) was added into the appropriate triplicate wells. The cells were then incubated with the compounds for an additional period of 24 h, In case of compound 17, which exhibited dark toxicity at 20 µM, the final tested concentrations were adjusted from 0.31 to 10 µM while compound 13h included concentrations from 0.04 to 5 µM or 0.08 to 10 µM. |
674755f77be152b1d019291d | 41 | After a washing step with medium to remove any non-internalized compound, 200 µL of a medium without red phenol were added (RPMI, culture media: Roswell Park Memorial Institute) and the cells were irradiated with a LED broadband lamp (400 -700 nm). The correction factor from the overlap of the absorption spectra between the LED and each group of compounds was calculated and applied in order to achieve accurate light doses. The cells were irradiated with a light dose of 1 J cm -2 (10 mW cm -2 broadband lamp, 8 min) or 2.6 J cm -2 (10 mW cm -2 broadband lamp, 48 min). After an additional incubation of 24 h, resazurin assay was applied to determine the cell viability (150 µL in each well, 0.1 mg mL -1 ). The experiment was repeated three to four times and the final cell viability was accessed after merging the mean values of the individual experiments. In the case of dyads the experiment was performed one time for each L.D. A statistical significance analysis was applied by using Graphpad software. Significance was evaluated with one-way ANOVA in comparison to the untreated cells with the Dunnett′s Multiple Comparison Test and it is displayed with stars in the respective figures; for the p-value: p < 0.05 for *, p < 0.01 for **, p < 0.001 for ***. 20% and C87 Cl11, Cl12 3% occupied. Compound 13k has weak diffraction at higher angle, leading to a high R(int) and 16 also displays disorder in one dipyrromethene benzoic acid group over a 2-fold axis. The benzoic acid moiety is modelled as a complete unit with half occupancy using restraints (ISOR, SIMU, SADI). There are partially occupied water molecules (2×25% occupancy in the ASU) as well as hexane (15% occupied in ASU) and all modelled with restraints (SIMU, DFIX, ISOR and DANG). Lastly, AL(DIPY)3 20 was more challenging with a diffraction limit = 0.97 Å due to weak diffraction. The diffuse contribution of solvent in the voids was accounted for by using the SQUEEZE routine in PLATON, with two voids of 2688 and 2689 Å 3 with 729 electrons each removed. Two arms (biphenyl-dipyrromethane) were disordered over two locations and modelled with a combination of rigid groups (AFIX 66 and AFIX 176) using restraints (SADI, SIMU, ISOR) with occupancies of 70:30% (C50:C50B) and 65:35% (C84:C84a). a Angle between the dipyrrin plane and the 5-substituent; b average of Al-N distance; c parameter which presents the sum of the deviation from 90° of the octahedron configuration. |
674755f77be152b1d019291d | 42 | Table shows key structural parameters for the ALDIPYs including the sigma parameter which represents the sum of the deviation from 90° of the angles in the coordination sphere (of the metal atom) reflecting the deformation of the octahedron. Sigma values determine the distortion of the complexes, the larger the value the higher distortion from the ideal octahedral conformation. Complexes 13b, 13d, 13j and 20 showed the higher distortion (sigma = 15-21), presumably due to their bulkier substitution that induces more strain within the molecules. However, all structures are very close to the ideal octahedral configuration. |
6731c55b5a82cea2fabf27d4 | 0 | . The human immunodeficiency virus (HIV) codes for the Viral infectivity factor (Vif) protein to activate human proteasome-degradation of human viral-inhibitor APOBEC3G (A3G). A3G is the most important single-stranded DNA deaminase enzyme inhibiting HIV by hypermutating its genome . Inhibition of Vif has been the target for abundant therapeutic research hampered by the lack of 3D models of Vif interaction with human proteins including highly insoluble A3G . |
6731c55b5a82cea2fabf27d4 | 1 | In the absence of HIV-Vif, human A3G, one member of the APOBEC3 (A3) enzymes, is the natural most potent anti-HIV inhibitor . During experimental HIV infections in the absence of Vif, human A3G binds RNAs in the cytoplasm of the host cells and is encapsidated into the new virions . In the HIV subsequent infection-replication cycle, the encapsidated A3G randomly induces dC lethal deaminations to dU (deoxyCytosine to deoxyUracil) during the reverse transcription of the HIV genome, reducing their replication and infection. |
6731c55b5a82cea2fabf27d4 | 2 | During years, the insolubility of A3G did not allowed to solve a 3D structure. By enhancing A3G solubility by sequence engineering (sA3G) and identifying optimal binding short RNA (RNA20), Vif-human complexes were solved by cryoEM as asymmetric hetero-dimers 9 (Figure ). Each of the two sA3G monomers binds: two Vifs (Vif blue binding the sA3G N-terminal RNA domain and Vif red binding the sA3G C-terminal deamination domain), two CBF and 1 RNA20 (Vif blue Vif red -sA3G-2CBF-RNA). Most probably, A3G binds RNA during its translation and contributes to the enhancement of affinities among the proteins on the complexes. Similar model interfaces have been predicted among other A3G complexes from different sources and laboratories . The proposed heterodimer model was further assessed because it included correct binding predictions to human ubiquitin-ligase . |
6731c55b5a82cea2fabf27d4 | 3 | Several studies have confirmed the amino acids and protein domains implicated in the hetero-dimer amino acid interfaces and/or A3G degradation . For instance, 127 W, 128 D, 130 D, 270 K of sA3G interactions with Vif Red 43 H, 70 W, 71 G, Y 9 , sA3G degradation-dependent 161 PPLP 22 or 40 YRHHY of Vif Red and A3G D128K mutation , viral package of A3G 124 Y, 127 W residues , etc. |
6731c55b5a82cea2fabf27d4 | 4 | Vif have been one of the most studied targets for HIV therapeutic strategies . Many antiVif inhibitors including complementary peptides, have been proposed . However, most of the maximal affinities reported for antiVif inhibitors have been > 100 nM. On the other hand, the newly defined interfaces of Vif-human protein asymmetric hetero-dimers 9 remain unexplored for inhibitors. |
6731c55b5a82cea2fabf27d4 | 5 | The present computational search was limited to docking candidates predicting the highest affinities to cavity interfaces of Vif blue Vif red sA3G±RNA (Graphical Abstract). Explorations were limited to highest docking-affinities to reduce possible in vivo active concentrations and side-effects. To produce new molecules predicting high docking-affinities, tens of thousands of raw-children were randomly generated from selected parents and thousands selected for best Vif blue Vif red sA3G±RNA cavity-fitting (fitted-children). For such explorations the DataWarrior Build Evolutionary Library (DWBEL) java-based co-evolution algorithms were employed as alternative to screening extralarge drug-like molecular banks or predictive learning docking models from protein sequences . Similar co-evolutions had predicted ligand-affinity improvements when targeting different protein-ligand pairs. For instance, new antibiotics for resistant Staphilococcus 42 , Abaucin-derivatives against Acinetobacter 43 , nonhuman anticoagulant rodenticides , monkeypox Tecovirimat-resistant mutants , inner-cavity SARS omicron , inflammatory coronavirus ORF8 protein , prokaryotic models of human K + channels 48 , inner-cavity of VHSV rhabdovirus , malaria circumsporozoite protein or RSV resistant mutations . |
6731c55b5a82cea2fabf27d4 | 6 | The co-evolution strategy was developed in several steps: i) select the best parent molecules and cavities for DWBEL co-evolutions targeting the simplified Vif blue Vif red sA3G model with know anti-Vifs or star-like home-designed molecules, ii) generate thousands of DWBEL non-toxic fitted-children targeting the cavity interfaces of the Vif blue Vif red sA3G model, iii) select the top-children displaying the higher affinity consensus docking between two very different docking algorithms (DWBEL and ADV), and iv) study the influence of RNA by comparing the consensus top-children by ADV blind-docking targeting Vif blue Vif red sA3G±RNA. |
6731c55b5a82cea2fabf27d4 | 7 | Despite finding possible ligand candidates, the consensus top-children predictions mentioned above are limited in their possible applications. For instance: a) the engineered and simplified Vif blue Vif red sA3G±RNA models may not yet accurately reflect all their possible interfaces, b) the accuracy of the searches limited to high affinity candidates, will require experimental confirmation, c) docking have been performed with fixed rather than side-chain dynamic docking-cavities, in the absence of water interactions, and any of both properties may interfere and/or d) there were a limited number of docking cavities explored on the Vif complexes compared to many other protein interfaces and to their vast numbers of their chemical-space alternatives. Despite these limitations, some of the newly identified top-children targeting the Vif blue Vif red sA3G±RNA models could be experimentally tested or provide some examples of similar strategies to apply to further antiVif explorations. |
6731c55b5a82cea2fabf27d4 | 8 | To target Vif blue Vif red sA3G interfaces (Supplementary Materials / GraphycalAbstract.pse), screening to select the most appropriated 3D-molecular parents and their targeted cavities were performed by ADV blind-docking . First a library of 10 small molecules were selected from those previously described to bind Vif (antiVif) with anti-HIV activity. To maintain antiVif 2D geometries, their energies were minimized by the DW mmff94s+ force-field (Supplementary Materials / 10antiVif-mmff94s+.sdf). This DW force-field was required because other tested(gaff, uff, ghemical, mmff94 and mmff94s), altered the expected docked 2D structures and affinities of the resulting conformers (for instance, misplacing double carbon or nitrogen bonds and changing ring sizes). The antiVif mmff94s+ conformers (one per antiVif molecule) were then ADV docked using a 60x60x60 Å grid surrounding the whole Vif blue Vif red sA3G model (blind-docking). Blind-docking was preferred, because it predicts the best docking cavities, by selecting the conformers fitting them with the highest affinities. Among other possibilities, the resulting ADV blind-dockings predicted the highest affinities to Benzimidazole-26 (Vif8) and IMC15 (Vif4) . Both of them targeted the Vif red sA3G cavity interfaces (Supplementary Materials / 10antiVif.pse). Therefore, Vif8 and Vif4 were selected for further work. |
6731c55b5a82cea2fabf27d4 | 9 | To explore wider chemical spaces, antiVif-unrelated parents were also selected among 36 molecules home-designed with 3-fold symmetric star-like structures of several molecular weights (Supplementary Materials / 36star-mmff4s+.sdf). These star-like molecules contained amino-carboxyl terminal atoms to maximize possible amino acid interactions with any possible cavity. Their corresponding ADV blind-docking results predicted the highest affinities for star18, 29 and 30. All selected star-like molecules targeted Vif red sA3G cavity interfaces (Supplementary Materials / 36star.pse). |
6731c55b5a82cea2fabf27d4 | 10 | DWBEL co-evolutions from the selected antiVif and star mmff94s+ parentconformers and their corresponding Vif red sA3G -cavities, randomly generated ~ 35000 to 45000 raw children per parent. More than 2000 non-toxic children per parent were selected from the raw children by best-fitting Vif red sA3G -cavities, except the Vif4-derivatives that were less abundant (Table ). |
6731c55b5a82cea2fabf27d4 | 11 | To increase the affinity accuracies, the DWBEL non-toxic fitted-children were ADV docked using 45x45x45 Å grids centered to the PyMol centerofmass of Vif8 and star30. The smaller and re-centered grids restricted the docking search to those conformers targeting the Vif red sA3G cavities. Compared by an arbitrary 46 nM threshold, the results predicted the highest affinities for 38.0 and 58.8 % of the Vif8-and star30-derivatives, respectively (Table ). |
6731c55b5a82cea2fabf27d4 | 12 | The highest affinities were employed as the main selecting criteria for the top-children derivatives, because they may allow for lower concentrations and reduced side-effects if tested for possible biological activities. The number of raw-children corresponded to the DWBEL total randomly generated new molecules after 3 consecutive runs per parent. The number of Non-toxic Fitted-children corresponded to the children best-fitting the Vif red sA3G cavities. ADV corresponded to the number of Non-toxic Fitted-children predicting ADV blind-docking (45x45x45 Å grids) < 46 nM. %, calculated by the formula, 100 * number of ADV / number of Non-toxic Fitted-children. Bold %, percentage of Vif8-and star30-derivatives with the highest affinities. |
6731c55b5a82cea2fabf27d4 | 13 | Despite their identical amino acid sequences and 2D-chemical structures, Vif red and Vif blue bound different A3G/sA3G domains in the cryoEM hetero-dimer model (Supplementary Materials / GraphycalAbstract.pse). Therefore, the Vif red -sA3G consensus top-children conformers predicted by targeting 45x45x45 Å grids, were also ADV blind-docked to include any Vif blue sA3G cavities targeting 60x60x60 Å grids. Because RNA may stabilize Vif blue Vif red A3G complexes, RNA (optimal RNA20 oligomer 9 ) was also targeted. |
6731c55b5a82cea2fabf27d4 | 14 | Consensus top-children ADV blind-docking surrounding the whole Vif blue Vif red sA3G ±RNA molecule were performed using 60x60x60 Å grids. Their predicted docking-cavities were analysed by PyMol (Figure ). Numbers, consensus top-children selected to study their targeted amino acids (Figure ) Red, docking to Vif red +sA3G cavities. Blue, docking to Vif blue +sA3G cavities (Figure ). Gray, docking to cavities behind sA3G (Figure ). ▲, Vif8-derived consensus top-children. The ADV blind-docking affinities of the consensus top-children were then compared after targeting Vif blue Vif red sA3G ± RNA (Figure ). In the absence of RNA, the highest affinities were predicted for Vif8 consensus top-children number 6 (Figure , blue circle) and for star30 consensus top-children number 26 (Figure , red circle). In the presence of RNA, the highest affinities were predicted for 119 (Figure , red triangle) and for 32 (Figure , red circle). Some consensus topchildren increased their affinities when in the presence of RNA (i.e., 119, 432, 32, 40), while others maintained (i.e., 189, 337, 1, 43, 45) or reduced (6, 26) their affinities. |
6731c55b5a82cea2fabf27d4 | 15 | The nearby amino acids and/or ribonucleotides were identified by homedesigned scripts (*.py) run on PyMol/Python. The list of nearby amino acid or ribonucleotide residues per consensus top-children were then drawn into a homedesigned Origin template using home-designed scripts (*.ogs). The resulting nearby amino acids targeted a newly described sA3G region (~ 160-180 residues), while the rest of positions were similar to those identified in this work when antiVif drugs targeted Vif blue Vif red sA3G + RNA (compare in Figure , sA3G and Grey rectangles above). The ~ 10-100-fold higher affinities predicted by some consensus top-children compared to previous antiVif drugs, could be explained by targeting the newly described sA3G region. 9 and other biochemical properties as critical for A3G binding and degradation (W70, G71, H43 and Y44) and for important A3G interactions (W127, D128, D130, and K270) . |
6731c55b5a82cea2fabf27d4 | 16 | Drug-like non-toxic low-nanoMolar affinity top-children have been computationally predicted targeting protein interface cavities on the simplified Vif blue Vif red sA3G ±RNA model . Consensus affinities between DWBEL coevolution and ADV docking co-evolved into two different ~ 3-fold star-like molecular structures. The new consensus top-children cross-docked amino acids on sA3G and Vif-sA3G new and similar to those previously reported for known antiVif drugs, but predicting ~ 10-100 fold higher affinities. These new candidates may offer examples for experimental evaluation of possible HIV inhibitions. Combinatorial chemical synthesis and/or chemical analogues may help to chemically synthesize some of them. Alternatively, different candidates may be evolved by applying similar strategies targeting other interface cavities. |
65fd9be79138d2316108ad9a | 0 | Many crucial biochemical processes rely on intricate networks of protein-protein interactions (PPIs). It is widely acknowledged that almost half of the proteome in eukaryotes consists of proteins containing intrinsically disordered regions (IDRs). These IDRs are characterized as contiguous segments of 20 or more amino acid residues that remain unstructured under native conditions. Notably, IDRs play a pivotal role in numerous PPIs. Frequently, an IDR undergoes a disorder-to-order transition upon binding to another protein, resulting in the formation of a low-affinity, high-selectivity protein-protein complex. It is well documented that PPIs play a crucial role in the pathogenesis of diverse and complex diseases, such as cancers, diabetes, and neurodegenerative disorders. Unsurprisingly, modulation of PPIs with small molecules is an appealing drug discovery strategy. However, targeting PPIs with small-molecule drugs has proven to be a formidable challenge, with only a limited number of such inhibitors successfully becoming approved drugs over the past quarter century. A deeper understanding of how to design small molecules that emulate PPI recognition mechanisms, such as folding of a protein IDR upon binding, may unlock new strategies for developing drug candidates, and new possibilities for effective interventions in complex disease processes. This report focusses on unravelling the molecular driving forces of a disorder-to-order transition of a protein IDR induced by a small molecule. Targeting the p53/MDM2 proteinprotein interaction is a well-established anticancer strategy and several small-molecule inhibitors have reached late-stage clinical trials. However, the emergence of drug resistance due to mutations in p53 or MDM2 has underscored the need for developing next-generation inhibitors. Most p53/MDM2 inhibitors bind to the N-terminal domain of MDM2, which consists of a core structured region, and an IDR referred to as a 'lid' (Figure ). AM-7209, a clinical candidate compound belonging to a family of piperidinone molecules, uniquely orders the N-terminal 'lid' of MDM2 upon binding (Figure ). Previous work from our group has elucidated why AM-7209 selectively bind to the ordered-lid conformation of MDM2, whereas other compound classes bind without ordering the lid. Here we achieve a deep understanding of the molecular driving forces that underpin AM-7209:lid recognition by performing a combination of molecular dynamics (MD) simulations, isothermal titration calorimetry (ITC) experiments and nuclear magnetic resonance (NMR) measurements. Significantly, we identify a crucial amino residue in the intrinsically disordered lid of MDM2 that modulates AM-7209 affinity by up to three orders of magnitude. These insights provide valuable design principles for targeting future IDRs with small molecules and highlight the potential role of IDRs in drug-resistance mechanisms. |
65fd9be79138d2316108ad9a | 1 | Previous experimental and computational work has demonstrated that the N-terminal lid region of MDM2 adopts a helix-turn-strand motif in the presence of several piperidinone ligands. A panel of single-site and double-site MDM2 mutants was prepared for in silico analyses to probe the contributions to the stabilisation of this motif of various lid-to-protein and lid-toligand contacts (Figure ). Mutations to a glycine were attempted at several locations, as this residue shows significant differences in backbone conformational preferences over other amino acid residues and would be expected to increase the inherent structural disorder of the lid. Each mutant underwent two independent 0.5 μs equilibrium MD simulations, and the stability of the helical motif observed at the base of the lid was assessed by post-processing of the computed trajectories. |
65fd9be79138d2316108ad9a | 2 | Single substitutions closer to the N terminus (residues 14-18) did not significantly disrupt the helical motif, although some mutations (V14G, V14D, T16G) increased positional fluctuations of the extended segment (Figure ). Modifications of I19 that decreased the size of the side chain (I19G, I19A) significantly decreased helical propensity, and significantly increased positional fluctuations. This was explained by the loss of non-polar contacts with residues T16, L54, and Y100. Interestingly, replacement of I19 by E19 was tolerated as the helical motif was stabilized by compensating interactions between E19 and R97. |
65fd9be79138d2316108ad9a | 3 | Previous work suggested that the salt-bridge between E23 and R97 is critical to lock the lid in an ordered conformation. However, the present MD simulations suggested that a broad range of mutations at E23 (E23G, E23L, E23Q) cause only moderate loss of helicity and did not significantly increase lid flexibility. The double mutant E23R97:R23E97 was simulated to test the effect of swapping the residues involved in salt-bridge formation. This causes a moderate decrease in stability of the helical motifs. Other double mutants that included I19G were simulated to evaluate potential additive effects (I19G:E23G, V14G:I19G, T16G:I19G). These double mutants exhibited helical stability comparable to that observed in I19G, suggesting that mutations of I19 were key to destabilizing the ordered lid state. |
65fd9be79138d2316108ad9a | 4 | A panel of eight mutants, selected on the basis of the MD simulations, was subjected to biophysical measurements to further investigate the relationships between predicted lid helix propensity, lid flexibility, and ligand binding energetics (Figures ). The compound Nutlin-3a, which does not order the intrinsically disordered lid of wild-type MDM2, served as a control. As anticipated, the ITC-derived thermodynamic signature of Nutlin-3a binding to MDM2 remained similar for all constructs, irrespective of the lid mutations (Figure ). This suggests that any interactions of Nutlin-3a with the lid, if present, are energetically inconsequential. This aligns with the notion that Nutlin-3a does not induce ordering of the MDM2 lid region upon binding. |
65fd9be79138d2316108ad9a | 5 | In contrast, the binding thermodynamic signature of AM-7209 was markedly influenced by mutations in the lid region (Figure ). Generally, weaker binding mutants exhibited a more negative entropy of binding and a more positive enthalpy of binding. Mutations predicted to have little-to-no effect on helix stability in the lid region (V14G, Q18G, E23G, E23R:R97E) showed entropies of binding similar to wild-type. Mutations predicted to disrupt helix formation (I19G, T16G:I19G) exhibited entropies of binding similar to the lid-truncated construct. The T16G mutant behaved intermediately between the two other classes of mutants. This supports the idea that binding of AM-7209 to weaker mutants is entropically favored due to the absence of lid ordering but enthalpically disfavored because of a reduced number of contacts between the ligand and the lid. |
65fd9be79138d2316108ad9a | 6 | Remarkably, analysis of the ITC-derived binding constants (Figure ) reveals that the affinity of AM-7209 for I19G is decreased by 1,000-fold compared to wild-type, comparable with the affinity measured for the lid-truncated construct. Conversely, mutations of residues that directly contact AM-7209 have a modest effect (V14G, Q18G) or intermediate effect (T16G) on binding constants. The significant effect of the T16G mutation on AM-7209 affinity suggests that the MD simulations may have underestimated the structural perturbation of the lid IDR caused by this mutation. The double mutant T16G:I19G binds AM-7209 with similar affinity to the single mutant I19G, suggesting that interactions of T16 with the ligand depend on formation of the helical motif in the lid. The importance of the E23-R97 salt bridge for lid ordering was probed with the E23G mutation. The resultant moderate loss of AM-7209 affinity is consistent with the moderate loss of helical propensity in MD simulations. Interestingly, swapping of salt-bridging residues in the E23R:R97E double mutant does not maintain wildtype binding affinity. This is consistent with MD simulations indicating a decreased propensity for formation of the R23-E97 salt bridge versus the native E23-R97 salt bridge. Uncertainties are the standard deviation of the mean from triplicate measurements. |
65fd9be79138d2316108ad9a | 7 | Significant perturbations in amide ( 1 H, 15 N) chemical shifts were apparent throughout the protein, including for several lid residues, upon titration (up to a ligand:protein molar ratio of 1.2:1) of AM-7209 into 50 µM WT MDM2 (Figure , and Figures ). When the same concentrations of AM-7209 were titrated into 50 µM I19G significant chemical shifts perturbations were observed throughout the protein with the crucial exception of lid residues (Figure ). Displaying the combined 1 H, N CSP values on a 3D structure of MDM2, confirms that the large perturbations observed in the lid region of WT MDM2 are absent in the case of the I19G mutant (Figure and). Figures show data for the entire protein sequences. |
65fd9be79138d2316108ad9a | 8 | Heteronuclear ( 1 H-15 N) NOE values for lid residues 10-17 showed a significant increase following AM-7209 titration into WT MDM2, suggesting decreased mobility on the ps-ns timescale, consistent with ordering of this region (Figure ). Conversely, negligible or small increases in heteronuclear NOE values were observed for the equivalent lid residues following AM-7209 titration into the I19G mutant (Figure , see Figures for data for the entire protein sequences). In WT MDM2, residues 12-16 showed a trend towards negative secondary Ca chemical shifts upon AM-7209 addition, while residues 20-23 showed a trend towards positive Ca chemical shift (Figure , light purple and purple), consistent with increased b- strand and a-helical propensity, respectively. These trends were not observed when AM-7209 was titrated into a sample of I19G (Figure , light orange and orange). Upon addition of AM-7209 to WT MDM2 a trend towards negative secondary CO chemical shifts values was observed for residues 12-16 as opposed to the positive values observed for residues 20-23, which is again consistent with increased b-strand and a-helical propensity, respectively (Figure , light purple and purple). No such trends were observed when AM-7209 was added to the I19G mutant (Figure , light orange and orange). |
65fd9be79138d2316108ad9a | 9 | A complete picture of the effect of the I19G mutation on the energy landscape of MDM2 (6-125) was sought to complement the calorimetric and NMR measurements. Free Energy Surfaces (FES) for WT in the absence and presence of AM-7209 were computed using our previously described aMD/US/vFEP methodology (see Figures and SI Methods paragraph). The calculations suggests that the lid region of unliganded WT predominantly adopts a partially disordered conformation that occludes the ligand-binding site (label C1 Figure ). They further suggest that the disordered lid in unliganded MDM2 can also transiently adopt an alternative position that gives partial access to the ligand-binding site (label C2, Figure ). Full "opening" of the lid is energetically disfavoured (label C3, Figure ). After AM-7209 binding, the dominant lid conformational state appears to shift to an ordered helical and partially extended motif (label C4, Figure ). An ordered helical and fully extended motif (label C5, Figure ), consistent with X-ray crystallography data, is slightly less energetically favoured. Such a fully extended lid conformation can be stabilised in X-ray diffracted structures via crystal packing. Motif formation requires repositioning of the lid away from core helix a2 to avoid steric clashes with the ligand. A closed disordered motif (label C6, Figure ) with the lid positioned above the ligand is also energetically accessible according to these calculations. |
65fd9be79138d2316108ad9a | 10 | The calculations suggest that in unliganded I19G the energetically dominant lid state is disordered and occludes the ligand binding site (label C8, Figure ). A partially open, partially helical motif is energetically feasible (label C9, Figure ). Upon docking in AM-7209, conformational preferences shift towards adoption of this partially open and helical state (label C11, Figure ). Lid conformations that wrap over the ligand become energetically disfavoured (label C12, Figure ). |
65fd9be79138d2316108ad9a | 11 | A secondary structure analysis of the full conformational ensembles predicted by the above free energy surfaces indicate that binding of AM-7209 to WT significantly increases helicity of residues 21-25 (Figure , first and second rows). By contrast binding of AM-7209 to I19G is associated with only a small increase in helical propensity for residues 21-25 (Figure , third and fourth rows). As reported previously, the DSSP algorithm doesn't detect the strand component of the helix-turn-strand motif in the ordered MDM2 lid IDR (Figure , second row) around residues 12-16. Nevertheless adoption of an extended conformation for lid residues 12-16 is apparent from visualisation of the free energy surfaces (Figure ) and consistent with the increased b-strand propensity for residues 12-16 inferred from secondary chemical shift changes (Figures ). The helix propensity around residues 10-17 observed in Figure is associated with transient helices seen in closed lid conformations that wrap above the ligand (e.g. Figure label C12). |
65fd9be79138d2316108ad9a | 12 | Taken together these results suggest that formation of a stable helical motif in residues 21-25 at the base of the MDM2 lid region requires additional contacts between the methyl groups of I19, the methyl group of T16 and the chlorophenyl ring of AM-7209 (Figure ). Disruption of this network of hydrophobic contacts increases lid flexibility and unwinds the helical motif. This explanation also accounts for the significant loss of affinity of AM-7209 for the T16G mutant observed in ITC experiments (Figure ). Thus, these results suggest that ordering of the MDM2 lid into a helical turn-extended motif is driven by shielding of a cluster of protein and ligand hydrophobic moieties, which is reminiscent of hydrophobic collapse observed in protein folding. To support this interpretation additional simulations were carried out with a modified version of AM-7209 in which the chlorophenyl moiety pointing towards T16 was replaced by a phenyl ring (AM-7209-Cl, Figure ). The decreased contacts between the ligand and T16 were observed to destabilise the helical motif in a manner qualitatively similar to the effects of the I19G mutation (Figure ). These findings provide an experimentally testable hypothesis for the proposed hydrophobic collapse mechanism. |
65fd9be79138d2316108ad9a | 13 | This study combined calorimetric and NMR measurements with detailed atomistic molecular simulations of protein-ligand interactions to elucidate the molecular recognition mechanism that underpins ordering of the N-terminal MDM2 lid region upon binding of AM-7209, a small molecule ligand. This ligand-specific lid ordering is found to be driven by the shielding from solvent of a cluster of lid and ligand non-polar moieties in a manner reminiscent of the hydrophobic-collapse model of protein folding. Other lid stabilisation mechanisms, such as direct contacts between lid residue Val14 and AM-7209, or salt bridge formation between lid residue Glu23 and MDM2 residue Arg97, were shown to only play a secondary role. The computer simulations enabled identification of a single residue mutation at Ile19, that abrogates this ligand-induced protein disorder-order transition, and is associated with a concomitant loss of three orders of magnitude in affinity for the ligand. Our findings thus suggest that, in addition to loss-of-function mutations in the TP53 gene, mutations in MDM2's intrinsically disordered lid region have the potential to mediate resistance to some classes of therapeutic p53/MDM2 antagonists. More broadly, our results highlight the prominent role that IDRs can play in molecular recognition of small molecules, and the importance of accounting for the presence of IDRs in ligand optimisation efforts. Recent advances have enabled generation of sequenceto-ensemble representations for entire IDR proteomes. Given the high prevalence of IDRs, assessment of energy landscapes of IDRs flanking conventional structured protein regions through detailed atomistic simulations may offer a generalizable strategy to exploit cryptic binding sites for small molecules in therapeutic interventions. |
65fd9be79138d2316108ad9a | 14 | The gene (uniprot ID = A0A0A8KA17) of MDM2 (residues 6-125 for wild type and mutants) was inserted into a pET20b plasmid (ampicillin-resistant) with a six-His-tag in the C-terminal sequence. Proteins were produced in Escherichia coli strains C41 (DE3), C43, and BL21 (DE3), grown in LB broth for non-labelled proteins, and minimal medium supplemented with NH4Cl and/or 13 C6-glucose for labelled proteins. Purification was performed using IMAC (Immobilized Metal Ion Chromatography-5 mL) followed by size-exclusion chromatography). |
65fd9be79138d2316108ad9a | 15 | ITC was used to measure the dissociation constant (KD) of MDM2 ligands Nutlin-3a and AM-7209. All titrations (10 µM proteins in the cell and 100 µM Nutlin-3a or 150 µM AM-7209 in the syringe) were performed using a MicroCal Auto-iTC200 isothermal titration calorimeter from Malvern Panalytical, assuming one site of binding. The data were analyzed using the MicroCal PEAQ-ITC Analysis Software version 1.1.0.1262.13. For some constructs, measuring the binding affinity directly through titration proved difficult, and hence a competitive titration was used in these cases. |
65fd9be79138d2316108ad9a | 16 | Chemical shift perturbation measurements and relaxation experiments (T1, T2, heteronuclear ( 1 H, 15 N) NOEs) were performed, in the absence and presence of AM-7209, for WT and I19G (~1 mM and ~480 µM protein concentration respectively and complexes in molar ratio 1.2:1 ligand:protein). NMRPipe was used to process all the 2D and 3D spectra and subsequently analyzed using NMRFAM-Sparky and POKY. The same samples were used for backbone |
60c74ba74c89191b4aad34d4 | 0 | The room-temperature chemical synthesis of C2 was first reported in the form of a pre-print and has now appeared as a full paper. The core of the article asserts at its simplest that a transient intermediate 11 formed as -C≡C-I + -Ph by treating precursor 1a with a source of fluoride anion can fragment to singlet C2 and I-Ph at ambient or low temperatures (Scheme 1). |
60c74ba74c89191b4aad34d4 | 1 | The so-generated C2 can then be trapped in a variety of ways which are highly suggestive of this putative intermediate. Iodonium species are indeed known in the literature as alkynylation reagents, albeit with a proposed mechanism of action for inserting C≡C into molecules that does not involve free C2. Most of the trapping experiments in the present article are reported in solution, with an implied assertion that singlet C2 as a discrete species is insufficiently reactive to be captured by solvent rather than by a chemical trap. One experiment is claimed to produce C2 gas, but in this case the implication is that the C2 is insufficiently reactive to be trapped by the fritted glass filter through which it must pass. |
60c74ba74c89191b4aad34d4 | 2 | One can subject the reaction sequence in Scheme 1 to a reasonableness check based on bond dissociation energies (BDEs). A more quantitative assessment is available through higher level quantum mechanics. The authors themselves have not commented 2 on this aspect in their current article, which is based on purely experimental aspects. Addressing firstly the thermodynamics of the core equilibrium; -C≡C-I + -Ph ⇄ C⩸C + I-Ph Estimates of the experimentally derived BDE of the C-I bond in C≡C-I + -Ar iodonium salts are in the region of 70-80 kcal/mol. When the iodonium C-I bond cleaves, it is directly replaced by a fourth bond as represented by C⩸C, the BDE of which is experimentally estimated at the much lower value of ~17 kcal/mol. When allowance is made for a gain of ~10 kcal/mol of free energy resulting from increase in entropy, this implies that around 43-53 kcal/mol of bond energy must be recovered by the formation and trapping at ambient temperatures of C2 itself. |
60c74ba74c89191b4aad34d4 | 3 | To assess this aspect more quantitatively, the ωB97XD/Def2-SVPD density functional method, with solvation energies estimated using a continuum method set for dichloromethane, has been applied to the reaction shown in Scheme 1 (with NMe4 replacing NBu4). This suggests that the relative free energies ΔG298 of 1a, 11 and the assemblage labelled "C2" (C2 + I-Ph + Me3SiF + Me4N + BF4 -) are 0.0, 0.1 and +68.2 kcal/mol respectively. The computed energetics of C2 itself were calibrated against the two consecutive bond dissociation reactions |
60c74ba74c89191b4aad34d4 | 4 | for which the thermochemistry has been determined in the gas phase. This calibration suggests that the relative energy of C⩸C itself is too high by ~28 kcal/mol when computed using the ωB97XD functional and the Def2-SVP basis set. If this correction is applied to the ωB97XD results, then the computed free energy G298 of the reaction 1a → "C2" is reduced from +68.2 to ~+40 kcal/mol. This is in broad agreement with the simple argument advanced above from experimentally based BDEs (43-53 kcal/mol). |
60c74ba74c89191b4aad34d4 | 5 | A further, simplified, model at the CCSD(T)/Def2-TZVPPD/SCRF=dichloromethane level was computed, the dissociation of Me-I + -C≡C -→ Me-I + C⩸C. A solvation model is essential, since the ionic reactant is expected to be substantially stabilized by solvation compared to the non-ionic reaction products. At this level of theory, the energy of C2 itself is computed to be too stable by ~4.6 kcal/mol. With this correction applied, the overall reaction free energy emerges as G298 +47.1 kcal/mol, again in the range 43-53 kcal/mol. The former value corresponds to a half-life of a unimolecular reaction (Eyring theory) of ~10 18 hours at 298K. |
60c74ba74c89191b4aad34d4 | 6 | To add further insights, CCSD(T)/Def2-TZVPP model studies were conducted in which the I + -Ph leaving group is replaced by what must be the ultimate leaving group He + , itself formed by radioactive decay of tritium. Here, unlike the C-I bond, the BDE of the C-He + bond is tiny (~1 kcal/mol) and its replacement by C2 does indeed then result in a reasonably exo-energic equilibrium (G298 -42.2 kcal/mol), augmented again by entropy gain. This serves as a reminder that C2 itself is a very high energy species. |
60c74ba74c89191b4aad34d4 | 7 | How can a reaction shown in Scheme 1 and generating the proposed free C2 overcome a reaction endo-energicity G298 of +47 kcal/mol, the most accurate estimate obtained by the computations reported here. Eyring theory tells us that at 298K, reactions with respectively a half-life of 1 minute and 1 hour correspond to free energy barriers of 20.0 or 22.5 kcal/mol, significantly lower than the range of energies predicted above. There are several possibilities to consider. |
60c74ba74c89191b4aad34d4 | 8 | Firstly, an as yet un-identified mechanism recovers the energy. Is it possible that sufficient enthalpy can be recovered by say reorganisation of ionic lattice energies so that the resulting free energy barrier G298 ‡ could promote a sufficiently rapid (half-life <1 hour) reaction at room temperatures? This would allow C2 to be trapped by another species, but this would compete with ~barrierless reverse trapping by PhI. If so, by what type of mechanism could this recovered energy then be concentrated directly into the carbon-iodine bond in order to cleave it? That the range of thermochemical values for the reaction obtained by the three models above is wrongly predicted to be too high by 20-25 kcal/mol. Alternatively, one might consider that free C2 itself is not produced, but instead some other species which must account for the results of the trapping reactions. A possible check on the gas-phase trapping would be to condense whatever species emerges from the dry reaction flask onto a cold-finger at liquid helium temperatures in an argon matrix and subject this directly to spectroscopic (Raman or other) analysis as an alternative to chemical trapping. |
67b7c567fa469535b9379533 | 0 | Internal conversion, an electron-spin conserved non-radiative loss pathway, is a process of great importance and interest, mediated by non-adiabatic, or vibronic, interactions (the electronic-vibrational coupling between two electronic states). This loss mechanism interplays and competes with every other mechanism of interest; vast and ranging, they can be tracked to all the major fields of science. This includes photon harvesting mechanisms and applications , charge and proton transfer in biological systems , energy transfer between excited states , and thermally activated delayed fluorescence and phosphorescent systems . |
67b7c567fa469535b9379533 | 1 | Investigation and probing of internal conversion can be done from either an experimental or theoretical vantage, with their own profits and consequences. An experimental vantage can yield highly accurate findings, as well as allowing for significantly larger systems to be investigated. Experimentally, one is only limited by what can be made. The phase state, temperature, external fields, measurement technique, to name a few of many parameters, can all be altered to probe the physics of a given system. Characterisation is typically performed using ultra-short pulsed lasers , with femto-or attosecond resolution permitting direct observation of relaxation pathways. The limitation here is money; requirements for these measurements is often very expensive, and are therefore not available to many researchers. |
67b7c567fa469535b9379533 | 2 | Conversely, a theoretical vantage is limited by the level of theory used, but allows for an opportunity to estimate photophysical properties well before synthesis, and for distinguishability and separability of mechanisms. A good example of this can be found in density functional theory-based excited state exciton dynamics where each mechanism can be was calculated individually and later recombined for various quantum yeilds of excited states . Due to the difficulty associated with experimental measurement , it is sometimes considered more prudent to employ ab initio techniques. However, the strength of theory is also it's weakness: it is only as accurate as the level of theory used. |
67b7c567fa469535b9379533 | 3 | Lin , Siebrand , and Englman & Jortner are among the celebrated who are credited with developing the foundations for internal conversion that are used today. However, it was not just them; decades of blood, sweat, and tears, from an army of dedicated researchers working to build the foundations we all know. The renowned energy gap law, for example, could not have come to pass without Lin laying the ground work, the photophysical observations of Kasha , or the unique observations of Robinson & Frosch . Following all this work was the experimental validation over the years , solidifying the theory presented, which allowed for many others to improve on the work, like Plotnikov , Valiev and co-workers , Shaw and co-workers , and Manian and co-workers , to calculate explicit coupling terms and occupational quanta (quantum occupation of vibrational normal modes). Inspired by the theoretical development, cutting edge light-emitting diodes and solar cell devices demonstrate that full control over internal conversion is vital to truly perfect device efficiency. |
67b7c567fa469535b9379533 | 4 | One primary bottleneck for the quantitative treatment of internal conversion is the evaluation of nuclear-coordinatedependent wavefunction components . Formal generation of vibronic coupling requires the mixing of nuclear-electronic and nuclear-vibrational wavefunctions, and is a relatively nontrivial photophysical property to predict. A consequence of this complexity makes it so that many choose to instead infer the rate of internal conversion from other quantum chemical properties as opposed to calculating it directly. Erker & Basché infer the rate of internal conversion in dibenzoterrylene as a function of the photoluminescent quantum yield P LQY , cast as : |
67b7c567fa469535b9379533 | 5 | where W is the rate constant for a given mechanism, r here denoting radiative decay, and the sum of W is all the combination of all rate constants, radiative and non-radiative. However, the issue becomes when an implied non-radiative rate constant is reported as internal conversion loss, rather than a summed non-radiative rate. Recent works have shown that even small intersystem crossing (singlet-triplet non-radiative decay) rates can accumulate and contribute to the stabilisation of the P LQY . This is not the fault of any of those in the field using inferred methods; benchmarking of density functionals show the exact treatment of internal conversion to be both resource heavy and laborious . |
67b7c567fa469535b9379533 | 6 | Due in no small part to its complexity, internal conversion is effected by a myriad of different photophysical properties and system specifics. The energy gap, temperature, medium, degrees of freedom, streic bulk, Franck-Condon factors, thermal occupation, electronic charge; all can influence the overall rate of internal conversion through intricate parametric relationships. Duschinsky rotations, or the Duschinsky effect , where frequency of a particular normal mode will shift in energy drastically with respect to which electronic energy level it is observed due to mixing of modes with respect to the ground state, recognised as the reason for the subtle asymmetry between absorption and emission spectra in a variety of systems , can shift the Franck-Condon factors quite drastically . |
67b7c567fa469535b9379533 | 7 | As shown, there are now an abundance of methods one can choose from, each with their own profits and consequences. In colloquial terms, one now has 57 ways to skin the same cat . But which is correct for a given system? This comprehensive review will examine the various theories used to calculate internal conversion in numerous systems and phases of matter. Beginning with the early success and later deprecation of pioneering work by Lin, Siebrand, and Robinson-Bixon-Jortner, internal conversion is examined in the context of Fermi's Golden Rule, Arrhenius-like methods, and the Plotnikov-Robinson-Bixon-Jortner models for molecular systems. We also consider more complex cases such as internal conversion in the solution phase, condensed matter systems, and quantum dot and nanostructures. We then move away from molecules and begin to examine more complex systems such as those in complex solution and condensed matter. Finally, internal conversion will be explored in the context of more niche applications, such as in exciton-polariton coupled systems, deuteration effects, and exact conical intersections. |
67b7c567fa469535b9379533 | 8 | This comprehensive review is formatted in a non-standard way: upon the casting of an equation citations will be given for each manuscript which reports it. This is done in the hopes of minimising possible confusion due to the myriad of methods available. These lists are laborious and exhaustive, but likely incomplete. Many works for example are inaccessible due to the era in which they were written (language, USSR; cold war-era), or issues with physical vs digital copies of old texts. Hence, we have only cited those works which we could find either physically or digitally. However, for more information we recommend examining these works where available. It should also be emphasised that in the solid state and quantum dot literature, the term "charge recombination" is used rather than internal conversion. While charge recombination can refer to any method in which recombination occurs, and therefore also includes fluorescence, based on the coupling, we infer the mechanism of internal conversion for charge recombination. Therefore, in this review, we will use internal conversion as our primary terminology. |
67b7c567fa469535b9379533 | 9 | Beginning at the Born-Oppenheimer adiabatic approximation , whereby it is assumed that the nuclear mass is much larger than that of the electrons that therefore the velocity is much faster than that of the nuclei, the total Hamiltonian can be expressed in terms of operators for the electronic kinetic energy T (q) and nuclear kinetic energy T (q), and the total electronic potential energy U (q, Q) and potential energy of the nuclei L (Q) : |
67b7c567fa469535b9379533 | 10 | corresponding to the set of electronic coordinates q = {q i }, the set of nuclear coordinates Q = {Q k } with nuclear masses {M k }. In some treatments, one may also incorporate the spin-orbit interaction into the Hamiltonian, however for most systems without heavy atoms, first-order spin-orbit coupling is weak, and can therefore be ignored. In terms of the partial Schrödinger equation, the exact wavefunction ψ (q, Q) can be expressed as : |
67b7c567fa469535b9379533 | 11 | where φ n (q, Q) and χ n (Q) are the electronic and nuclear wavefunctions, respectively. The eigenstates of Equation 3 are not stationary, but rather the whole system oscillates between various quantum states ; that this what we interpret as the electronic transition, followed by a transition in the quantum states of nuclear motion. Therefore, to properly probe internal conversion, the perturbation resulting in the transition between electronic states needs to be modelled correctly. At the time, a method to calculate an explicit perturbation term was unavailable. However, the transition probabilities can still be obtained as highlighted by Kronig . |
67b7c567fa469535b9379533 | 12 | where E m (E n ) is the energy of the m th (n th ) electronic state. This is valid until vibronic states from different electronic configurations begin to converge, to which the approximation begins to fail. Known cases of this include the Jahn-Teller effect (degenerate electronic states), the pseudo-Jahn-Teller effect (near-degenerate electronic states), and the Herzberg-Teller effect (vibrational participation in electronic transitions). Following, a standard treatment of first order perturbation theory is performed, leading to an expression for the probability Ω n for locating the system in state |ψ n ⟩ : |
67b7c567fa469535b9379533 | 13 | For benzene, Hornburger and co-workers reported a rate of ∼10 6 s -1 . For naphthalene, Lin reported transition probabilities between for 10 8 -10 4 s -1 , with a heavy dependence on the energy gap between the initial and final electronic states, while Bixon & Jortner noted an internal conversion rate constant slower than 10 12 s -1 for the 1 B 3u ← 1 B 2u transition, with a reported energy gap of 0.423 eV. Rashev reported a rate of 4.16×10 6 s -1 for the 1 B 2u to ground state internal conversion pathway in zeroth-order vibronic benzene. Further internal conversion calculations have been carried out for NO. Lin examined the transition probability and found a probability of 10 -6 s -1 . Lin also calculated a transition probability of 1 s -1 for the a 4 π state. |
67b7c567fa469535b9379533 | 14 | Robinson & Frosch discuss internal conversions in terms of nonstationary states. Internal conversion can be reimagined discussed as the problem of energy transfer in the solid state; while the initial and final state may not be in resonance, a rigorous but reasonable model whereby the set of final states is constructed using simple mathematical properties can be used to liken the problem to small lattice vibrations. A series of assumptions can be made here: firstly environmental effects (eg. solvent) have no influence on the electronic coupling and a weak influence on the vibronic coupling through Franck-Condon factors, and that each vibration, in the zeroth-order, is near-degenerate and coupled to the 0-band. From here, the combination of each individual state will form a discrete band, with a width ℵ proportional to the occupational vibrational quanta (typically ∼1 cm -1 ). Importantly, as more vibrations contribute to the overall rate, those states closer to resonance make a more measurable contribution to the rate. Thus, over small timescales, the transition probability can be cast in terms of either near or far resonance : |
67b7c567fa469535b9379533 | 15 | where r and s are the initial and final vibrational states, respectively. Integration of Equation 9 across all coupled final states yields the total transition probability. Ω n (t, E) varies linearly with respect to t, at least when it is much smaller than the speed of vibrational relaxation. In this case, the integration limit can be set to infinity; an overall transition probability can then be cast : |
67b7c567fa469535b9379533 | 16 | where x mj (ν mj ) and x nj (ν nj ) are harmonic-oscillator vibrational wavefunction factors, denoted in Ref. 87 as a vibrational factor. In the case where there is a single directly coupled final state and the vibrational factor is roughly unity, Robinson & Frosch note that even weak coupling will result in a rate constant of around 4 × 10 7 s -1 , agreeing qualitatively with experiment. Burland & Robinson calculate a rate of 3.24 × 10 1 s -1 for the B 2u →A 1g transition in benzene. |
67b7c567fa469535b9379533 | 17 | Despite the success the Robinson-Bixon-Jortner model displayed, it's derivation was based on an oversimplified physical model. Bixon & Jortner discuss that an exact physical solution can only be modelled using the exact Born-Oppenheimer states, which at the time was impractical. However, it is also worth noting that the exact distribution of energy levels is not crucial for a full analysis of internal conversion. Further, the derived formalism was developed specifically to treat an isolated molecule in the gas phase, and therefore has minimal applicability to modern quantum chemistry. |
67b7c567fa469535b9379533 | 18 | That is not to say that the legacy models are, or were not, brilliant in their time. However, we simply understand and can calculate more using modern quantum chemistry techniques. Both Lin and Longeut-Higgins concluded that the current state of the art of internal conversion, based on the transition probability, fails whenever the adiabatic approximation is not applicable, or does not correctly describe the stationary state of the system. Sharf and Silbey noted that due to the strong dependence of the energy denominator on nuclear coordinates, the predicted coupling could be erroneous, smaller by ∼1 order of magnitude than compared to a result calculated using simple adiabatic functions as a zeroth-order basis set. The "crudeness" of the basis set is due to approximation employed where electrons are unable to "adiabatically" follow nuclear vibrations. In other words, the potential energy of the electrons is taken as a set of Coulomb attractions/repulsions (see . This results in negative charge centers, and therefore poor vibrational constants , at least compared to those obtained using an adiabatic basis set. However, the adiabatic basis should be appropriate for most cases; Nitzan & Jortner argue that as long as the basis set is complete with respect to the Hamiltonian and perturbative operators, then the issue is moot. |
67b7c567fa469535b9379533 | 19 | It should be noted that while many legacy works have extended their respective formalisms to incorporate anharmonic effects , for the time it was still a difficult task. Aside from publications by Siebrand and co-workers , only Burland & Robinson , Fischer , Standord & and Fischer , and Shimakura and co-workers reported ways to address anharmonic effects. While for many of these works anharmonicity was driven by cubic and quartic contributions to the oscillators, Shimakura in particular employed the use of a Morse potential. However, the dependence on the occupational quanta was yet to be addressed, and therefore the current state of the art was still lacking. |
67b7c567fa469535b9379533 | 20 | A major flaw in the formalism is that the role of other states on the transition of interest ie. Herzberg-Teller components, are not accounted for. This simplification was used across numerous studies during the legacy of internal conversion theory , since the model omits some auxiliary restrictions, such as the magnitude of the electronic or vibrational energy gap, but rather focuses on the nature of the weak-coupling regime, and considers the initial state energetically well spaced from all other states as well as maintaining that the adiabatic potential surfaces are harmonic. That being said, coupling strength between two states is proportional to the energy gap, due to the nuclear kinetic energy operator inducing transitions between the two adiabatic Born-Oppenheimer states. Therefore, in order to correctly model internal conversion in any system, it is necessary to go beyond the Born-Oppenheimer approximation of quantum mechanics. |
67b7c567fa469535b9379533 | 21 | If we integrate the probability of locating a system in a given state with respect to time, we arrive at Fermi's Golden Rule , a bread-and-butter expression which can describe some photophysical processes as some perturbation to the system. If we redefine our previous Hamiltonian (Equation ) as some base system H 0 with some small perturbations acting on it H x , we can expand our Hamiltonian with respect to a series of ordered terms: |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.